XT.i^M* JOURNAL OF AGRICULTURAL RESEARCH Volume III OCTOBER, 1914— MARCH, 1915 depar'tment of agriculture WASHINGTON, D. C. 78745'— 15 8 PUBLISHED BY AUTHORITY OF THE SECRETARY OF AGRICULTURE, WITH THE COOPERATION OF THE ASSOCIATION OF AMERICAN AGRICUL- TURAL COLLEGES AND EXPERIMENT STATIONS EDITORIAL COMMITTEE FOR THE DEPARTMENT FOR THE ASSOCIATION KARL F. KELLERMAN, Chairman RAYMOND PEARL Physiologist and Assistant Chief, Bureau of Plant Industry EDWIN W. ALLEN Assistant Director, Office of ExperimetU Stations CHARLES L. MARLATT Assistant Chief, Bureau of Entomology Biologist, Maine Agricultural Experiment Station H. P. ARMSBY Director, Institute of Animal Nutrition, The Pennsylvania ^State College E. M. FREEMAN Botanist, Plant Pathologist, and Assistant Dean. Agricultural Experiment Station of the University of Minnesota All correspondence regardinsf articles from the Department of Agriculture should be addressed to Karl F. Kellerman, Journal of Agricultural Research, Washington, D. C. All correspondence regarding articles from Experiment Stations should be addressed to Raymond Pearl, Journal of Agricultural Research, Orono, Maine. >7f%- CONTENTS Page. Relative Water Requirement of Plants. Lyman J. Briggs and H. L. Shantz I Heart-Rot of Oaks and Poplars Caused by Polyporus Dryophilus. George G. Hedgcock and W. H. Long 65 Decomposition of Soil Carbonates. W. H. MacIntirb 79 A Fungous Disease of Hemp. Vera K. Charles and Anna E. Jenkins 81 A More Accurate Method of Comparing First-Generation Maize Hybrids with Their Parents. G. N. Collins 85 Natural Revegetation of Range Lands Based upon Gro\vth Re- quirements and Life History of the Vegetation. Arthur W. Sampson 93 Pecan Rosette. W. A. Orton and Frederick V. Rand 149 A Nitrogenous Soil Constituent: Tetracarbonimid. Edmund C. Shorey and E. H. Walters 175 Apple Root Borer. Fred E. Brooks 179 Changes in Composition of Peel and Pulp of Ripening Bananas. H. C. Gore 187 Assimilation of Colloidal Iron by Rice. P. L. GiLB and J. O. Carrero 205 Coloring Matter of Raw and Cooked Salted Meats. Ralph Hoagland 211 Oil Content of Seeds as Affected by the Nutrition of the Plant. W. W. Garner, H. A. Allard, and C. L. Foubert 227 Studies in the Expansion of Milk and Cream. H. W. BearcE 251 Life History of the Melon Fly. E. A. Back and C. E. Pemberton . 269 Identification of the Seeds of Species of Agropyron. Robert C. Dahlberg 275 Observations on the Life History of Agrilus Bilineatus. Royal N. Chapman 283 Efifect of Dilution upon the Infectivity of the Virus of the Mosaic Disease of Tobacco. H. A. Allard 295 Moldiness in Butter. Charles Thom and R. H. Shaw 301 Susceptibility of Citrous Fruits to the Attack of the Mediterra- nean Fruit Fly. E. A. Back and C. E. Pemberton 311 Physiological Changes in Sweet Potatoes during Storage. Hein- rich Hasselbring and Lon A. Hawkins 331 Three-Comered Alfalfa Hopper. V. L. WildermuTh 343 Life History of the Mediterranean Fruit Fly from the Standpoint of Parasite Introduction. E. A. Back and C. E. Pemberton. . 363 lU IV Journal of Agricultural Research voi. ni Page. Relation of Simultaneous Ovulation to the Production of Double- Yolked Eggs. Maynie R. Curtis 375 Brachysm, A Hereditary Deformity of Cotton and Other Plants. O. F. Cook 387 Ability of Colon Bacilli to Survive Pasteurization. S. Henry Ayers and W. T. Johnson, Jr 401 Fitting Logarithmic Curves by the Method of Moments. John Rice Miner 411 Organic Phosphoric Acid of Rice. Alice R. Thompson 425 Two Clover Aphids. Edith M. Patch 431 Net Energy Values of Feeding Stuffs for Cattle. Henry Pren- tiss Armsby and J. August Fries 435 Air and Wind Dissemination of Ascospores of the Chestnut- BUght Fungus. F. D. Heald, M. W. Gardner, and R. A. Studhalter 493 Index 527 ERRATA Page 32, Table XX, " TrifoUum repens" should read " Trifolium pratense." Page 52, Table XXX, " Linum usatissimum" should read " Linum usiiatissimum." Page 52, Table XXX, " Cucumbis sativa" should read " Cucumis salivtis." Page 53, Table XXX, " Boutelona gracilis" should read " Bouteloua gracilis." Page 53, Table XXXI, "Cucumis sativa" should read "Cucumis sativus." Page 56, Table XXXIII, column head, "Ratio, 1913 to 1912, " should read "Ratio, 1912 to 1913." Page 56, line 4 from bottom, " 7S±2 " should read "75±i." Page 97, line 26, "Salix nutallii" should read " Salix nutialUi ." Page 106, Table II, " Sulanioti velutinum " should read " Siianion veluiinum." Page 165, Table III, insertlineafter"MgO" toread "FcoOj. .. .0.07 o.io 0.09 o.io." Page 165, Table IV, "FI2O3" should read "FcjOa." Page 165, Table IV, "Total 99. 94 98.36 " should read " Total 99. 94 99. 54 100.36 100.00. " Page 166, Table V, "Total 104.45 104-82 101.41 101.99" should read "Total 99' 73 loo- 04 9925 99.81." Page 256, line 10, "(Ci=i— ;,„,• C2=C,^— (C,=); „ A^=D< — (D,)™.)" should read "C,=t-i,n,- C,=C,^-(C,''),„; N=Dt-(Dt)„,." Page 256, line 11, "(i'C,='a+i'C,,/3C=JC,Af.)" should read " i'C,=a:-|-i'C,C2^= IC,N." Page 278, lines 37-38, "(fig. 2)" should read "(fig. 3)." Page 279, line 19, "as" should read "at." Page 279, legend for figure 2, "Side views of basal portions, etc.," should read "Fig. 2. — Detail drawings of ventral view of seeds of Agropyron spp.: A, Agropyron repens; B, A. smithii; C, A. tenerum. X 9." Page 280, legend for figure 3, "Detail drawings of ventral view, etc.," should read "Fig. 3. — Side views of basal portions of seeds of Agropyron spp., showing the relative projection of tlie rachilla: A, Agropyron repens; B, A. smithii; C, A. tenerum. X 9." Page 280, legend for figure 4, "Edge of rachilla in Agropyron spp., etc.," should read "Fig. 4. — Edge of palca in Agropyron spp., showing space and comparative size of bristles: A , Agropyron repens; B, A. smithii: C, A . tenerum. X 9. " ILLUSTRATIONS PLATES Relative Water REQxnREMENT op Plants Pase. Plate I. Fig. i. — General view of the plant inclosure used at Akron, Colo., showing the pipe framework covered with a hail screen, with the board base surmounted by a single width of cheesecloth to protect the plants against high winds. Fig. 2. — General view inside tlie inclosure, showing the arrangement of pots and general conditions of growth. Com and sorghums are shown in the foreground, small grain in the background. Fig. 3. — General view of the inclosure photographed shortly after the grain in some of the pots had been harvested 64 Plate II. Fig. i. — Pot planted with sugar beets, showing the wax seal around the plants and also the sealed holes where stand was not perfect. Fig. 2. — Weighing pots, showing spring balance, weighing support, and general procedure. Two men operate the weighing support, one of whom lifts the pot by means of a windlass, while a third reads the balance and records the weight. By this method weighings can be made at the rate of two per minute. Fig. 3. — Gnndelia squarrosa (gumweed) and Artemisia frigida (mountain sage), illustrating the growth of native plants used in the water- requirement meastu'ements 64 Plate III. Fig. i. — Kubanka wheat, grown May 9 to September 3 , 1912. Fig. 2. — White Hull-less barley, grown May 16 to August 12, 1912. Fig. 3. — Kubanka wheat. Set grown outside of shelter, May 9 to August 31, 1912. Fig. 4. — Emmer, grown May n to August 12, 1912. Fig. 5. — Swedish Select oats, grown May 17 to August 23, 1912. Fig. 6. — Kharkov wheat, grown April 27 to August 28, 1912 64 Plate IV. Fig. i. — Northwestern Dent com, grown JunS 9 to September 16, 1912. Fig. 2. — Hopi com, grown June 12 to September 26, 1912. Fig. 3. — White durra, grown June g to September 26, 1912. Fig. 4. — Red Amber sorghum, grown June 29 to September 27, 1912. Fig. 5. — Minnesota Amber sorghum, grown June 9 to September 26, 1912 64 Plate V. Fig. i. — Sudan grass. First crop, grown May 28 to July 26, 1912. Fig. 2. — Voronezh proso, grown June 5 to August 20, 1912. Fig. 3. — Kiu-sk millet, grown June 9 to August 20, 1912. Fig. 4. — Select Grimm alfalfa, grown in the open. May 24 to July 27, 1912. Fig. 5. — Select Grimm alfalfa, grown in the shelter. May 24 to July 26, 1912 64 Plate VI. Fig. 1. — Cowpea, grown June 17 to August 26, 1913. Fig. 2. — Hairy vetch, grown May 29 to July 18, 1913. Fig. 3. — Soy bean, grown June i to August 26, 1913. Fig. 4. — Cantaloupe, grown June 14 to September 13, 1913. Fig. 5. — Indian Flint com, grown June 7 to August 27, 1913. Fig. 6. — McCormick potato, grown June 5 to October 4, 1913 64 Plate VII. Fig. i. — Triumph cotton in .shelter, grov^-n May 29 to September 16, 1913. Fig. 2. — Bocbera papposa, grown July 25 to vScptember 17, 1913. Fig. 3. — Rice in shelter, grown June 12 to September 16, 1913. Fig. 4. — General view of the shelter, showing emmer at the left and White Hull-less barley at the right. Fig. 5. — General view in the shelter, showing com in the foreground 64 VI Journal of Agricultural Research voi. ui Heart-Rot of Oaks and Poplars Caused by Polyporus Dryophilus Paee. Plate VIII. Fig. i. — Quercus alba: Crescent-shaped "soak," the initial stage of the piped rot produced by Polyporus dryophilus; from Arkansas. Fig. 2. — Quercus alba: A radial view of the rot in a limb, showing delignification; from Arkansas. Fig. 3. — Quercus oblongifolia: A radial view of rot, showing delignification; from Arizona. Fig. 4. — Quercus alba: A final stage of the rot, radial view, with more complete delignification; from Arkansas. Fig. 5. — Quercus alba: A tangential view of the rot, showing delignification in pockets; from Arkansas. Fig. 6. — Querents alba: An end view showing a cross section from the same tree as the preceding: from Arkansas. Fig. 7. — Quercus sp.: A section of oak from Von Tubeuf, sent to the jmiior writer as a specimen of the rot caused by Polyporus dryadeus in Europe. Fig. 8. — Quercus sp.: The reverse side of the specimen shown in the pre- ceding. Fig. 9. — Quercus sp.: A section of oak from Europe, obtained by Von Schrenk, with a piped rot similar to that of Polyporus dryophilus . ... 78 Plate IX. Fig. i. — A sporophore of Polyporus dryophilus, tuberous form on Quercus gambelii; from Arizona. Fig. 2. — Sectional view of a sporophore of Polyporus dryophilus on Quercus gambelii, showing the hard granular core with whitish mycelial strands; also the pore layer; from New Mexico. Fig. 3. — A sporophore of Polyporus dryophilus on Quercus californica, show- ing the upper siu^ace with a faint zonation; from California. Fig. 4. — A section through a sporophore of Polyporus dryophilus on Quercus garryana, showing the structure of the hard granular core ; from California. Fig. 5. — A front view, showing the margin of the same sporophore as in figure 3, representing the ungulate form. Fig. 6. — A view of the pore surface of an applanate sporophore of Polyporus dryophilus on Quercus alba; from Arkansas 78 Plate X. Fig. i. — A sporophore of Polyporus dryophilus, front view showing the margin, on Populus iremuloides; from Colorado. Fig. 2. — ^A second sporophore from the same tree as figiu-e i, showing an imbricated form. Fig. 3. — A view of the upper surface of a sporophore of Polyporus rheades on Populus tremula; from Stockholm, Sweden. Fig. 4.^A sectional view of a sporophore of Polyporus corruscans on Quercus; from Upsala, Sweden. Fig. 5. — A side view of an imbricate sporophore of Polyporus dryophilus, applanate form on Populus Iremuloides; from Colorado. Fig. 6. — A sec- tional view of the same sporophore as in the preceding figure, showing the hard granular core and whitish mycelial strands. Fig. 7. — A view of the upper surface of an applanate sporophore of Polyporus dryophilus on Quercus alba; from Arkansas. Fig. 8. — The pore surface of a sporophore of Poly- porus dryophilus on Populus iremuloides; from Colorado 78 A Fungous Disease of Hemp Plate XI. A hemp plant, showing upper branches attacked by tlie fimgous Botryosphaeria marconii 84 Natural Revegetation of Range Lands Based upon Growth Requirements and Life History op the Vegetation Plate XII. Fig. i. — View of the lower grazing lands in the Wallowa National Forest. Fig. 2. — Characteristic open stand of western yellow pine and dense cover of herbaceous vegetation, mainly pine-grass (Calamagrostis pubescetis), Wallowa National Forest. Transition zone (yellow-pine asso- ciation). Fig. 3. — A burned-over area of lodgepole pine, with character- istic dense sapling stand 148 Oct.. 1914-Mar., 191S Illustrations vil Page. Plate XIII. Fig. i. — Dense stand of lodgepole pine, with undergrowth of red huckleberry {Vaccinium scoparium). Canadian zone (lodgepole-pine association). Fig. 2. — A flat eminence in the Hudsonian zone, showing the characteristic clumped growth of whitebark pine and Alpine fir. Fig. 3. — Irregular topography of the upper grazing lands. Hudsonian zone (whitebark-pine association) 148 Plate XIV. Fig. i. — Arctic-Alpine and upper-subalpine region, where forage is sparse, due to poor soil, short growing season, and unfavorable climate. Fig. 2. — Mountain range lands prior to the beginning of growth and germi- nation. Fig. 3. — Same view as shown in figure 2, but more in detail, showing the condition eight days later (June 30) 148 Plate XV. Fig. i. — Gantrast in the progress of the flower stalk production of mountain bunch-grass on portion of range which has been completely closed to grazing for a period of three successive years and on range which has been subject to continued early grazing. Fig. 2. — Western porcupine grass (Stipa occidentalis) , showing empty glumes and floret with the scale and its awned projection to the left; to the right the floret with glumes removed, showing the sharp-pointed, slightly curved seed tip. Fig. 3. — Average development of the root system and aerial portion of mountain bunch-grass at end of the first growing season 148 Plate XVI. Mountain bunch-grass, showing root development and aerial growth at the end of the second season 148 Plate XVII. Mountain bunch-grass in the spring of the third year of growth just before producing flower stalks, showing the natural position and length of the elaborate root development and aerial growth 148 Plate XVIII. Mountain bunch-grass at the end of the third year, showing three flower stalks and inflorescence 148 Plate XIX. Sickle sedge {Carex umbellata brevirostris) , showing offshoots from the rootstocks and flower stalks with fruit in the process of development . 148 Plate XX. Fig. i. — Station 4 on Stanley Range as it appeared on July 12, 1907. Fig. 2. — View of station 4 on July 15, 1909, after two years' protection from grazing animals. Fig. 3. — View of tfuadrat i, established on July 10, 1907. 148 Plate XXI. Fig. i. — Quadrat i, as it appeared on July 16, 1909. Fig. 2. — Area of mountain bunch-grass closed to grazing animals on July 8, 1907. Fig. 3. — View of open range contiguous to area shown in figure 2 148 Plate XXII. View of plot in the Transition (yellow-pine) zone which has been protected from grazing animals for three successive years, showing contrast in carrying capacity with contiguous open range 148 Plate XXIII. Fig. i. — View of portion of allotment at medium elevation where the destruction of forage seedlings due to grazing and trampling was studied. Fig. 2. — Dense stand of smooth wild rye {Elymus glaucus) and short-awned brome-grass (Bromus marginatus) seedlings 148 Pecan Rosette Plate XXIV. Fig. i. — One normal pecan leaf and two leaves with rosette from Dewitt, Ga. Fig. 2. — Pecan shoot with early symptoms of rosette. ... 174 Plate XXV. Resetted pecan leaf showing perforations due to the failure of part of the mesophyll to develop 174 Plate XXVI. Fig. i. — Pecan shoot in advanced stages of rosette. Fig. 2. — Normal pecan shoot for comparison with rosetted shoot 174 Plate XXVII. Fig. i. — Young orchard pecan tree with a moderate attack of rosette on the left side and seriously dying back from the disease on the other side. Fig. 2. — Young orchard pecan tree in advanced stages of rosette 174 VIII Journal of Agricultural Research voi. iii Page. Plate XXVIII. Fig. i. — Young orchard tree with severe attack of rosette. Fig. 2 . — Rosetted pecan tree cut off to the stump the preceding season , with the present season's growth again distinctly showing rosette. Fig. 3. — Two seedling pecan trees planted the same day from the same lot of seed- lings 174 ApplB Root Borer Plate XXIX. Figs, i and 3. — Sections of an apple root, showing burrows of the apple root borer {Agribts mttaticollis) . Fig. 2. — Cross section of the trunk of a young apple tree, showing burrows made by the larvse of the apple root borer in ascending the trunk to pupate 186 Plate XXX. Fig. i. — Agrilus vittaiicollis: I,arva(i;), pupa (6), and adult(a) of the apple root borer in the pupal cell. Fig. 2. — Xylophruridea agrili, a common parasite of the apple root borer. Fig. 3. — Section of trunk of young service tree, showing below the white egg and above the exit hole of the apple root borer. Fig. 4. — Xylophruridea agrili: Larvae of the para- site; one feeding on the larva and the other in the pupal cell of its host . . . 186 Plate XXXI. Fig. i. — Agrilus vittaiicollis: Egg on trunk of young service tree. Fig. 2. — Agrilus vittaticollis : Feeding form of larva. Fig. 3. — Agrilus ■vittaticollis: Contracted form of larva as taken from pupal cell. Fig. 4. — Agrilus vittaticollis: Pupa. Fig. 5. — Agrilus vittaticollis: a. Adult, or beetle; b, claw; c, antenna. Fig. 6. — Xylophruridea agrili, a parasite of the apple root borer: Pupa 186 Coloring Matter op Raw and Cooked Salted Meats Plate XXXII. Fig. i. — Oxyhemoglobin, ox blood. Fig. 2. — Oxj'hemoglobin , ox blood. Fig. 3. — NO-hemoglobin, ox blood. Fig. 4. — Methemoglobin, ox blood. Fig. 5. — Methemoglobin, ox blood 226 Plate XXXIII. Fig. i. — Oxyhemoglobin, sheep blood. Fig. 2. — Oxyhemo- globin, sheep blood. Fig. 3. — NO-hemoglobin, sheep blood. Fig. 4. — NO-hemoglobin, pig blood. Fig. 5. — NO-hemoglobin, pig blood 226 Identification of the Seeds of Species of Agropyron Plate XXXIV. Agropyron repens: Spikes showing degrees of variation which may occur. A , Typical spike 282 Plate XXXV. Agropyron smithii: Spikes showing degrees of variation. A, Tj'pical spike 282 Plate XXXVI. Agropyron tenerum: Spikes showing degrees of variation. A, Typical spike 282 Plate XXXVII. A^ro/^yron spp.: Typical seeds and spikelets. Fig. i. — Agro- pyron repens. Fig. 2. — Agropyron smithii. Fig. 3. — Agropyron tenerum . . 282 Observations on the Life History of Agrilus Bilineatus Plate XXXVIII. Fig. i. — Agrilus bilineatus: Eggs in position in, the bark of an oak tree. Fig. 2. — Agrilus bilineatus: Cluster of newly laid eggs. Fig. 3. — Agrilus bilineatus: Eggs shortly before hatching. Fig. 4. — Agrilus , bilineatTis: Newly hatched lar\'a. Fig. 5. — Agrilus bilineatus: Mature larva. Fig. 6. — Agrilus bilineatTis: Larva in its cell. Section made perpendicu- lar to the surface of the bark. A, Point at which adult will emerge; B, burrow stopped with frass. Fig. 7. — Agrilus bilineatus: Pupa in cell. Sec- tion made parallel to the surface of the bark. Fig. 8. — Agrilus bilineatus: Adult female. Fig. 9. — Agrilus bilineatus: Adult male 294 Oct.. 1914-Mar.. 191S Illustrations ix Page. Plate XXXIX. Fig. i. — Leaf showing work of four Agrilus beetles in 24 hours. Fig. 2. — Hole in bark made by adult Agrilus in emerging from pupal cell. Fig. 3.— Larvae of Agrilus bilineatus and their burrows. Fig. 4. — Complete burrow of a larva of Agrilus bilineatus. A, Point at which larva hatched; B, beginning of second instar; C, beginning of third instar; D, beginning of fourth instar; E, pupal cell 294 Susceptibility of Citrous Fruits to the Attack of the Mediter- ranean Fruit Fly Plate XL. Fig. i. — Orange infested with larvse of the Mediterranean fruit fly (Ceraiitis capitata). Fig. 2. — Orange infested with larvae of the Mediter- ranean fruit fly {Ceraiitis capitata) , showing two breathing holes of the larvae in the decayed area 330 Plate XLI. Cross section of shaddock No. i, showing the thick, loose texture of the rag with darkened area above and to the right showing the channels made by well-grown Mediterranean fruit-fly larvae 330 Plate XLII. Fig. i. — Cross section of the orange shown on Plate XL, figitfe 2. Fig. 2. — Orange containing 87 punctures in the rind 330 Three-Cornered Alfalfa Hopper Plate XLIII. Fig. I. — The three-cornered alfalfa hopper (S/!c/oce/)/!a/a/«'; ore spp.: A, Agropyron repens; B, A . smilhii; C, A . ienerum 278 2. Detail drawingof ventral viewofseedsofj49ro/>}'rorespp.: A, Agropyron repens; B, A. smithii; C, A. teneruin. (See "Errata.") 279 3. Side views of basal portions of seeds of Agropyron spp., showing the relative projection of the rachilla: A , Agro pyron repens; B, A. smilhii; C, A . ienerum. (See ' ' Errata. ") 280 4. Edge of rachilla in Agropyron spp. , showing shape and comparative size of bristles: A , Agropyron repens; B, A. smithii; C, A . ienerum 280 Moldiness in Butter Fig. I. Graph showing the effect of salt on molding 308 Susceptibility of Citrous Fruits to the Attack of the Mediter- ranean Fruit Fly Fig. I. Cross section of peach, showing egg cavity of the Mediterranean fruit fly with eggs 320 2 . Cross section of peach , showing the general shriveling of the walls of the egg cavity and the separation of the eggs 320 3. Section of grapefruit rind, showing two egg cavities, one in cross section . 321 Three-Cornered Alfalfa Hopper Fig. 2. Map showing distribution of the three-cornered alfalfa hopper {Sticio- cephalafesiina) in the United States 344 Ability of Colon Bacilli to Survive Pasteurization Fig. I. Curve showing results of heating cultures of colon bacilli for 30 minutes at various temperatures 404 Oct.. 1914-Mar., 191s Illustrations xiii Two Clover Apraos Page. Fig. I. Aphis brevis: Antenna of fall alate female collected from hawthorn ... . 432 2. Aphis brevis: Antenna of alate male 433 3. Aphis bakeri: Antenna of alate female collected from clover 433 Net Energy Values of Feeding Stuffs for Cattle Fig. I. Graphshowing the dry matter eaten and the increments of heat produc- tion due to standing, computed per 500 kg. live weight per 24 hours. 455 2. Graph showing the relation of heat production to dry matter consumed, computed per 500 kg. live weight 472 Air and Wind Dissemination of Ascospores op the Chestnut- Blight Fungus Fig. I . Map of chestnut coppice growtli at West Chester, Pa. , in and near which tlie experiments on wind dissemination of the chestnut-blight fungus were carried out 497 2. Map showing the location of some of the important outlying exposure- plate stations 498 3. Map showing the location of water spore-trap stations Nos. I to VI 520 ADDITIONAL COPIES OF THIS PUBLICATION MAT BE PROCUEED FEOM THE SUPEKINTENDENT OF DOCUMENTS GOVEENMENT PRINTING OFFICE WASHINGTON, D. C. AT 25 CENTS PER COPY SiTBSCEiPTiON PER Year, 12 Numbers, 12.50 A Vol. Ill OCTOBER, 1914 No. 1 JOURNAL OF AGRICULTURAL RESEARCH CONTENTS Pase Relative Water Requirement of Plants ----- i L. J. BRIGGS and H. L. SHANTZ Heart-Rot of Oaks and Poplars Caused by Polyporus Dry- ophilus --_--_____ ^5 GEORGE G. HEDGCOCK and W. H. LONG Decomposition of Soil Carbonates ------ 79 W. H. MacINTlRE A Fungous Disease of Hemp - - - - - -81 VERA K. CHARLES and ANNA E. JENKINS A More Accurate Method of Comparing First-Generation Maize Hybrids with Their Parents - - - _ _ gS G. N. COLLINS DEPARl^ENT OF AGRICULTURE WASHINGTON, D.C. WASHINOTON I OOVERhMtNT PBlKTlNQ OFnCE : 1»M PUBLISHED BY AUTHORITY OF THE SECRETARY OF AGRICULTURE, WITH THE COOPERATION OF THE ASSOCIATION OF AMERICAN AGRICUL- TURAL COLLEGES AND EXPERIMENT STATIONS EDITORIAL COMMITTEE FOR THE DEPARTMENT FOR THE ASSOCIATION KARL F. KELLERMAN. Chairman RAYMOND PEARL PhysMooist aiid Assistant Chuf. Bureau of Plant Industry EDWIN W. ALLEN Assista,il Director. Ofiite of Experiment Stations CHARLES L. MARLATT Assistant Chief, Bureau of Entomology Biologist. Maine Agricultural Experiment Station H. P. ARMSBY Director. InstUule of Animal Nuirilion, The Pennsylvania Slate College' E. M. FREEMAN Botanist. Plant Pathologist, and Assistant Dean. Agricultural Experiment StaUon of the University of Minnesota All correspondence regarding articles from the Department of A^culture should be aSressed to Karl F. KeUer^an, Journal of Agnc.ltural Research. "^^r^lSnLce re^ardins articles fro. Experiment StaHons sho^he addressed to Raymond. Pearl. Journal of Agricultural Research. Or.no, Maine JOMAL OF AGRICULTIAL RESEARCH DEPARTMENT OF AGRICUETURE Vol,. Ill Washington, D. C, October 15, 1914 No. i RELATIVE WATER REQUIREMENT OF PLANTS By Lym.\n J. Briggs, Biophysicist in Charge, Biophysical Investigations, and H. L. Shantz, Plant Physiologist, Alkali and Drought Resistant Plant Investigations, Bureau of Plant Industry INTRODUCTION The marked differences in the quantity of water required by different species of plants for the production of a given weight of dry matter when grown under the same environmental conditions is a matter of scientific interest and of great economic importance in regions of limited water sup- ply. The measurements which have heretofore been made have for the most part been limited to a few species and have been carried out under such varied environmental conditions that comparison is difficult. The writers have therefore undertaken the measurement of the water require- ment of representative species and varieties of the principal crop plants, grown at the same place and under as nearly unif onn conditions as to time as the temperature requirement and Hfe history of the different crops will permit. The first series of measurements were made at Akron, Colo., in 191 1 (Briggs and Shantz, 1913a)'. These measurements were ex- tended in 1 91 2 and 191 3 to include many species whose water requirement had never before been determined. The later measurements form the subject of the present paper. The writers desire to express their obliga- tion to Messrs. R. D. Rands, A. McG. Peter, H. Martin, F. A. Cajori, N. Peter, and G. Crawford for efficient and painstaking assistance in con- nection with these experiments. EXPERIMENTAL CONDITIONS The experimental procedure in 191 2 and 191 3 was similar to that in the earlier experiments. The plants were grown to maturity in large galvanized-iron pots holding about 115 kg. of soil. Each pot was pro- vided with a tight-fitting cover having openings for the stems of the plants, the annular space between the stem of the plant and the cover be- ing sealed with wax. The loss of water was thus confined almost entirely to that taking place through the leaves, and the entrance of rainfall was almost wholly excluded. The wax which has been found to be the most ' Bibliographic citations in parentheses refer to " Literature cited." p. 62-63. Journal of Agricultural Research, Vol. Ill, No. i Dept. of -Agriculture, Washington, D. C. Oct. 15, 1914 (l) G-32 2 Journal of Agricultural Research voi. iii, No. i satisfactory for sealing the openings about the stems consists of a mixture of four parts of unrefined beeswax with one part of tallow. Six pots of plants of each variety ' were used, and the water require- ment of each pot was determined independently, in order to provide a basis for the calculation of the probable error of the mean. In making this calculation, Peter's abridged method, based upon the sum of the departures, has been employed. - The term "water requirement," when employed in the following pages without further restriction, indicates the ratio of the weight of water absorbed by a plant during its growth to the weight of the dry matter produced, exclusive of the roots (Briggs and Shantz, 1913a, p. 7). The plants were dried to constant weight in a steam-heated oven, main- tained at approximately 110° C. When the plants produced grain, the water requirement based upon the weight of the dry grain is also given. The percentage of grain produced by the plants grown in the pots usually compared favorably with the field performance. Unless a normal per- centage of grain is produced, the water-requirement ratio based on grain production should not be applied to crops grown under field conditions. In a few instances the water requirement based upon the weight of roots or tubers has also been determined. SCREENED INCLOSURE To protect the plants from birds and severe hail and wind storms, it was found necessary to conduct the experiments in a screened inclo- sure. The inclosure used in 191 1 consisted of a wooden framework covered with wire netting of X-'nch mesh. This framework shaded the plants somewhat, being made sufficiently rigid to support a track above each row of cans, from which the cans were suspended during weighing. To reduce the shading effect, a new inclosure was provided in 191 2, the framework of which was made of i-inch galvanized-iron pipe with pipe posts 9 feet high at intervals of 8 feet. The framework to a height of 3 feet was covered with a wooden wall which came slightly above the top of the pots. The remainder was covered with No. 21 galvanized-wire netting of 3/^-inch mesh. General views of the inclosure are shown in Plate I. Although the new inclosure reduced the shading effect, pyrheliometric and total radiation measurements made inside and outside the inclosure still showed a measureable reduction in the radiation due to the shade of the screen. Measurements made with an Abbot silver-disk pyrheli- ometer (Abbot, 191 1) showed that the intensity of the direct radiation ' The recorded strains used in these measurements were obtained from the following offices of the Bureau of Plant Industry: Foreign Seed and Plant Introduction (S. P. I.); Cereal Investigations (C. I.); Alkali and Drought Resistant Plant Investigations (A. D. I.). 2 The formula used was RTn=o.84S ' — , , where Rnj=the probable error of the mean. Id=the sum of n\n~j the departures, and «=number of determinations. A probable error based upon six determinations does not necessarily represent strictly the actual fre- quency diagram, and this must be borne in mind in the consideration of probable errors. For a discussion of the probable error when the number of obser\'ations is small, see "Student" (1908). Oct. 15, 1914 Water Requirement of Plants from the sun was reduced about 20 per cent by the inclosure at midday in midsummer, while total radiation measurements made with a differ- ential telethermograph (Briggs, 191 3) gave approximately the same reduction. Simultaneous measurements of the water requirement of wheat, alfalfa, and cocklebur grown inside and outside the inclosure in 191 3 showed that the inclosure reduced the water requirement about 22 per cent. The water-requirement measurements must therefore be con- sidered relative rather than absolute. In this connection it should be recalled that plants growing under field conditions are also mutually shaded and otherwise protected to some extent. -The writers' measure- ments in 1 91 3 show that wheat grown in pots sunk in trenches and sur- rounded by a field of grain has a water requirement 10 per cent above wheat grown in the inclosure and 10 per cent below wheat grown outside the inclosure in a freely exposed wind-swept position. (See Table I.) The stand of wheat about the trench was below normal, owing to the dis- turbance of the plants in trenching and in caring for the pots. The potted plants in the trenches were consequently more exposed than if growing normally in a field of grain. The water requirement of the potted plants in the trench is therefore somewhat above that of plants normally pro- tected. From this comparison it appears that the inclosure measure- ments, at least in the case of wheat, are less than 10 per cent below the water requirement of plants exposed under field conditions. Table I. — Effect of the screened inclosure on the water requirement of wheat at Akron, Colo., in igij Water requirement based on — Plant and period of growth. No. Dry matter. Grain. Water. Grain. Grain. Dry matter. 1913- Grams. Grams. Kilos. Per cent. 7 162. 2 56-5 103.8 35 1,837 640 Kubatika, C. I. 1440 8 167.4 63-9 105. I 38 1,645 628 (Triticum durum), 9 143- 4 49.0 91.2 35 1,861 636 check series, May 10 151.6 51-4 93- I 34 i,8io 614 22 to Aug. 13. II 148.9 56-5 89-5 38 1,584 601 12 159-3 5°-3 102. 2 32 2,032 I, 795 ±43 642 Mean 627±5 I 142-5 46.6 81.4 Z7, 1,745 571 2 151. I 43-9 80.8 29 1,840 534 Kubanka, in field, 3 1.53- 4 48.9 82.9 32 1,694 540 May 22 to Aug. 13. 4 156. 0 42.8 88.4 27 2, 063 566 5 143-9 48.2 83- 3 5i 1,726 579 I 6 167.8 44-9 97-3 27 2, 167 580 Mean i,873±6i 562 ±6 f 7^ 294.8 106. 5 150-3 36 1,410 510 74 273- 4 loi. 4 135-9 37 1,340 497 Kubanka, in shelter, 75 257-° 97- I 122. 4 38 I, 261 476 May 23 to Aug. 13. 76 304-2 116. 2 156. 0 32 1,342 513 77 253- 3 93- 8 121. 3 37 1,293 479 I 78 299.6 116. 8 150. 2 39 1,286 I,322±l6 502 Mean 496 ±S 4 Journal oj Agricultural Research voi. ni, no. i WEIGHING AND WATERING The discarding of the overhead track necessitated the construction of a movable support for -weighing the cans. The weighing support used is shown in Plate II, figure 2. It was constructed of i-inch gal- vanized-iron pipe and consisted of a crossbar which spanned the row and which was supported at each end by two bent posts. These posts were fitted with floor plates secured to two wooden skids, which slid along the ground on either side of the row of cans. In the earlier weighings the pots were suspended from a rope running through pulleys to a small windlass located on one of the posts of the support (PI. II, fig. 2). The windlass was later located directly beneath the crossbar and was operated through a chain-and-sprocket drive. Each pot was provided with bale ears by which it could be suspended directly from the balance by chains. When the plants were not suffi- ciently high to come in contact with the weighing apparatus, pots could be weighed at the rate of two a minute if two men handled the support and a third recorded the weight. When 300 pots or more are to be weighed three times a week, as was the case at Akron, rapidity in weighing be- comes important. The initial and final weighings have been made with an accuracy of one-fifth of a kilogram, either with a platform balance or a sensitive spring balance calibrated and corrected for temperature. Intermediate weighings have been made throughout with a spring balance calibrated by means of a sealed check pot weighing 130 kg. The water in all cases has been added from calibrated 2-liter flasks (Briggs and Shantz, 1913a, p. 11). The neck of each flask is cut so as to deliver 2 liters of water when brimful. The flasks are filled by submersion. In some of the later work a tank with a framework arranged for keeping a number of flasks submerged has been used. No time is thus lost in filling flasks or in adjusting the contents to a fiducial mark. SOIL FERTILIZER Surface soil from the experiment farm was used for filling the pots. Since it is well known that the water requirement is increased by a deficiency in the plant food supply (Briggs and Shantz, 1913b, pp. 31-56), the same quantity of a complete soluble fertilizer was added to each pot at each station at intervals during the growth of the crop. The fertilizer in 1 91 2 was applied at the rate of 50 p. p. m. of PO4, 100 p. p. m. of NO3, and 65 p. p. m. of K, all based on an assumed dry soil mass of 100 kg. per pot.' The phosphoric acid was applied as sodium phosphate; the nitrogen and potash as potassium nitrate. This amount of fertilizer was divided into four equal portions and applied at intervals during the active growth of the crops, the first application being made soon after * Approximately one-half of this quantity was used in 1913 and was applied as in 1912. Oct. IS, 1914 Water Requirement of Plants the plants had become well established. In practice it was found con- venient to make up a large quantity of the fertilizer solution of such concentration that 2 liters contained one-fourth of the total quantity required for one pot. The addition of this quantity to each pot was followed immediately by 2 liters of water. To test the influence of the fertilizer, one standard set of six pots of Kubanka wheat was grown without fertilizer at Akron, for comparison with the fertilized sets. The detailed results are given in Table II. The water requirement of the unfertilized set was 4±2 per cent below that of the fertilized set when based on the production of dry matter, and i ±4 per cent above, when based on grain production. The results therefore indicate that the additional plant food was not needed at this station, the water requirement of the two sets agreeing (within the errors of the experiment) whether based on the production of dry matter or on the production of grain. In 191 1 the water requirement of the unfertilized set ^ at Akron was 6±3 per cent above the fertilized set when compared on the basis of dry matter, although the ratios based on grain production were the same. T.\BLE II. — Effect of fertilizer on the water requirement of wheat at Akron, Colo., in IQ12 Plant and period of growth. Pot No. Dry matter. Grain. Water. Grain. Water requirement based on — Grain. Dry matter. 1912. Kubanka, C. I. 1440 {Trilicum durum). May 9 to Sept. 3, fertilized I 2 3 4 5 6 Grams. 270. 0 252.7 279.4 288.8 291.8 261. 7 Grams. 108. 2 88.3 98. I 99-7 99-5 92.2 Kilos. 95- I 97.1 109.9 118. 6 122. 0 106. 2 Per cent. 40 35 32 35 34 35 879 1,099 I, 120 I, 190 I, 226 1,152 352 384 393 411 418 406 Mean I, iii±37 394 ±7 f 1 8 9 10 II 12 291.8 272-3 290.3 274.8 300.2 263. 2 102.8 100. 6 101. I 79-5 105-7 87.6 108.6 107.4 108.8 109.4 114. I 88.4 35 37 55 29 35 33 Kubanka, C. I. 1440, May 9 to Aug. 21, 1,056 I, 067 1,076 1,375 I, 080 I, 090 372 394 375 . 398 380 336 Mean I, I24±32 376±6 CLIMATIC FACTORS The instrumental equipment for the measurement of climatic factors included maximum and minimum thermometers and an air thermo- graph exposed in a standard shelter 4 feet above the ground surface, an anemometer, a psychrometer, a rain gauge, and an evaporation tank. ' Based on pots i to 6, unfertilized, wliich had tlie same exposure as pots 7 to 12, fertilized. Journal of Agricultural Research Vol. Ill, No. 1 The sunshine, the wind velocity, the combined sun and sky radiation, the wet bulb depression, and the evaporation were automatically recorded. The results of some of these measurements, combined in 5-day periods for the sake of brevity, though not without sacrifice, are given in Tables III and IV. The discussion of the influence of climate on water require- ment has purposely been restricted, since such correlations as may exist can best be determined when discussed in connection with the results from other stations established for this purpose. Table III- — Summary of c lima lie conditions at Akron, Colo., in igi2 J912 April. May. June. July. August . September. . Days (inclusive). I to s 6 to 10 II to 15 i5 to 20 21 to 25 26 to 30 I to 5 6 to 10 II to 15 16 to 20 21 to 25 26 to 31 I to 5 6 to 10 II to 15 16 to 20 21 to 25 26 to 30 I to 5 6 to 10 II to IS 16 to 20 21 to 25 26 to 31 I to 5 6 to 10 II to 15 16 to 20 21 to 25 26 to 31 I to 5 6 to 10 II to 15 16 to 20 21 to 25 26 to 30 Air temperature CF.). Average of— Means. 4S 44 44 40 46 49 50 S3 44 61 64 60 64 59 60 55 66 72 66 73 70 69 73 69 68 65 69 68 72 (-5 51 51 46 43 Maxi- mums. ^lini- mums. 62 60 56 49 57 62 67 66 53 77 78 77 78 69 72 65 80 88 89 86 82 33 29 30 29 32 35 35 41 36 47 47 44 49 49 49 41 5° 55 49 55 54 55 59 Maxi- mum. 80 57 79 58 79 5° 83 54 82 53 89 S6 88 56 87 54 80 51 61 38 6b 37 59 34 50 32 73 69 68 54 70 69 76 75 65 83 84 92 78 84 78 86 89 87 96 92 90 94 85 84 85 89 86 95 96 90 91 77 81 71 Mini- mum. 29 23 26 26 27 32 28 34 32 41 42 35 45 47 44 37 43 5° 46 51 51 50 55 53 54 48 50 51 52 53 47 47 32 31 22 26 Precipi- tation. Inches. 0.26 • 51 •36 •03 1-33 •35 .40 •t'r! Tr. I. 14 ■ 41 I- 03 '■51 •44 •03 •34 . 01 .72 1.63 •85 . 16 ■56 .16 Evapo- ration. •43 I. 26 Inches. 0.82 •71 •71 •47 ■ 78 I. 09 .90 .60 •99 1.32 2-31 1.32 ■74 •84 I. 00 I. 21 I. 64 I. 12 1-58 1.36 I. 06 I- 38 I. 12 I. 04 I- OS I- IS 1.04 I. 29 1.48 1-39 1.08 .61 .62 Wijid veloc- ity per hour. ■54 .41 Mites. 7-7 »-5 14^7 95 9.6 9^5 10. 4 83 8.6 6.9 7^1 93 7.8 8.1 4^5 4^3 5^S 6. I 5^1 6. I 5^6 6.1 $■5 4. I 7- I 38 5-2 3° 4.0 5- I 6.4 7^1 5.6 63 63 4-5 Oct. IS, 1914 Water Requirement of Plants Table IV. — Summary of climatic conditions at Akron, Colo., in jgij 1913- April. May. June. July. August . September. Days (inclusive). I to 5 6 to 10 II to 15 16 to 20 21 to 25 26 to 30 I to 5 6 to 10 II to 15 16 to 20 21 to 25 26 to 31 I to 5 6 to 10 II to 15 16 to 20 21 to 25 26 to 30 I to 5 6 to 10 75 79 II to 15 16 to 20 74 68 21 to 25 66 26 to 31 67 I to 5 6 to 10 II to 15 16 to 20 21 to 25 26 to 31 I to 5 7.^ 6 to 10 68 II to 15 62 16 to 20 53 2 1 to 2 ^ 45 26 to 30 50 Air temperature ("F.). Average of— Means. 49 33 45 S3 44 60 47 54 55 55 62 69 65 57 66 71 69 74 78 74 73 76 73 75 Maxi- mums. 66 46 64 69 55 78 58 68 69 70 77 86 81 67 81 88 86 91 92 96 92 83 78 87 95 90 90 93 90 91 90 85 79 68 59 60 Mini- mums. 31 24 29 40 34 40 35 43 43 42 46 51 51 45 51 54 54 60 56 60 56 56 56 49 61 59 59 61 57 59 56 53 45 39 32 41 Maxi- mum. 77 73 77 74 68 84 69 81 80 80 84 91 87 77 91 93 89 97 i°3 93 87 93 97 93 95 97 92 92 86 87 76 70 Mini- mum. 21 10 18 34 27 37 31 41 33 38 39 48 48 37 42 52 52 49 53 56 46 53 S3 43 57 54 56 56 53 54 54 48 40 29 27 37 Precipi- tation. Inches. o. 02 •94 .87 •36 Tr. .82 •54 Tr. . 02 .06 .26 Tr. . 16 Tr. ■ 51 • 42 . 02 I. 12 .61 . 10 .05 .24 .04 • 17 •45 . 10 Tr. •39 •97 Evapo- ration. Inches. 0.82 .78 •31 .76 •51 ^•15 .72 .84 .86 •93 •99 I. 50 i^iS 1.07 1. 17 1.27 2. 19 I. 71 1.88 1.78 1.27 I. 01 I. 61 1^75 I- 54 1-23 1.38 1.58 1.83 1-37 I. 27 1.30 .98 .61 ■S' Wind veloc- ity per hour. Miles. 7^4 12.4 3^6 7.2 II. o 7-1 6.7 8.9 7^4 7.6 5-7 5-6 6.2 103 9.0 6.6 5-8 10.3 5-8 6.2 4.9 5-0 5-9 5-9 4.8 6.7 7.6 9.2 5-6 4-7 The months of June, July, August, and September, 1913, were all warmer than in 191 2, the average diflference in the monthly means being 4° F. In only 2 of the 24 five-day periods into which these months are divided did the mean maximum temperature in 191 2 exceed that of 1 91 3. The character of the two seasons is best reflected, however, in the evaporation graphs shown in figure i . The evaporation for the two years was not essentially different up to the ist of June. From this time on the evaporation in 1912 averaged much lower than in 1913. 8 Journal of Agricultural Research Vol. HI, No. I The marked response of the plants to the different seasons is shown in the reduced water requirement in 1912. (See Tables XXXII, pp. 36-38, and XXXIII, p. 39.) WATER REQUIREMENT OF VARIOUS CROPS WHEAT The water requirement of six varieties of wheat, including emmer, was measured at Akron in 191 2. The results a'rranged in order of increasing water requirement based on dry matter are as follows : Variety of wheat Water requirement Turkey 364±6 Kharkov 365±6 Kubanka 394±7 Emmer 428±3 Bluesteni 451 i4 Spring Ghirka 4S7±3 a/ — — — — — — r 1 ~— ~— — 1 ~^ — — p "" • ; '« / \,.' • / / k ^ y, ; / / / I ', / / \ \ ,' ; /•' \ ' \ N. V \ ,' i ' 'i' \ \ / V / '"., > .2 / \ ; \ 1 , 1 ,/ \ / s. / \ 0 y \ '> ' \; / \ / "^ 1 / / ' s. .'/' ■' \ \ / ' / 1 \ \. K \ *, ' i \ D - "\ > !> \ / l\ <5 •v '. I o\ V ~ ^ r. / \ / 1 \ / \ ' \ h— 1 s t °\ I A' '\. ^\ \ , ^ 1 \ n <: 1 u -, ^ i ". f) ? 1 ^ t 5 >, r, t 5 « ■) ; : \j ^ C \ 1/ ■> < 5 w 1 ^ i 1 ^ r \ K ^ <: i K 1 ; f ^ K 1 ' s *< n ; : 1 ^ r U ^ 1 « ■^ $> c/d/AZ" 340 I, 261 1,342 1,293 1,286 510 497 476 S13 479 502 Mean I,322±l6 496 ±5 Only one variety of wheat, the Kubanka, was included in the measure- ments of 1 91 3. This variety gave a water requirement 26 per cent above the 1 91 2 ratio and 19 per cent above the 191 1 ratio. The water require- ment on the basis of grain production was 19 per cent higher than in 191 2 and 1 1 per cent higher than in 1911. Oct. 15, 1914 Water Requirement of Plants 1 1 OATS The four varieties of oats employed in the water-requirement tests in 1912 were the same as those used in the 191 1 experiments. The water requirement in 1912, based on the total dry matter produced, was as follows : Variety of oats Water requirement Canadian 399±6 Swedish Select 423 ± 5 Burt 449±3 Si.xty-Day 491 ± 13 The Canadian again proved to be the most efficient of the varieties tested. The differences exhibited by the first three varieties in the list are practically the same as in 191 1. Much trouble was experienced in obtaining a stand of Sixty-Day oats. The germination was very poor and a second and even a third planting failed to give a good stand, as is shown by the variations in the yield of the different pots. (Table VI.) This is, perhaps, the cause of the higher water requirement obtained for Sixty-Day oats, which in 191 1 ranked next to the Canadian in efficiency. The water requirement of the different oat varieties, based on grain production in 191 2, was as follows: Variety of oats Water requirement Swedish Select i, 103 ± 18 Sixty-Day 1, I72±i33 Burt I, 224±55 Canadian i, 4i5±iig The probable error is high in all the determinations, except in the case of the Swedish Select (PI. Ill, fif. 5), and the relative order of the varieties is consequently of little significance. It is, however, of interest to observe that the Canadian variety is the least efficient in the use of water from the standpoint of grain production, which is in accord with the 191 1 experiments. Swedish Select and Burt oats were also included in the 191 3 measure- ments at Akron. On the basis of dry matter produced, the two varie- ties were equally efficient in the use of water. In the measurements of 1912 and 191 1 these two varieties gave only slight differences, the 191 1 and the 1912 results being in accord when the probable errors are con- sidered. On the basis of grain production, the Burt was the more effi- cient in 1913, and the Swedish Select in 1912 and in 191 1. No real differences of importance are shown in these two varieties when the meas- urements of the three years are considered. 12 Journal of Agricultural Research Vol. III. No. I Table VI. — Water requirement of different varieties of oats at Akron, Colo., in Igi3 and 1913 Pot No. Dry matter. Grain. Water. Grain. Water requirement based on — Grain. Dry matter. I9I2. Sixty-Day, C. I. 165 (.\vena sativa). May f 67 68 69 70 71 I 72 Gravis. 206. 5 287.7 270.9 93-9 274.7 248.9 Grams. 72.8 147-9 i°7-3 143-4 Kilos. 119. 9 131. 8 133-6 41. 6 129-5 124.7 Per cent. 36 51 40 1,645 892 1,245 5S0 458 493 15 to Aug. 23 52 904 472 501 1,172=1=133 49i=ti3 73 1 76 77 78 227. I 302. I 233-5 276. 0 158.3 250.0 80.9 118. 7 5°-4 79.8 60. 5 91. I 126. 7 89. I 107. 8 59-3 107. 0 36 39 22 29 Canadian, C. I. 444 (Avena sativa), May 17 to Sept. 16 1,125 1,068 1,767 1,35° 401 419 382 39° 375 428 24 1,769 I, 4i6d:ii9 399 ±6 79 80 81 82 83 84 352-8 339-9 299-5 344-5 351-5 363-8 138.8 no. 5 93-8 132-9 147-8 143-3 153-5 158.4 135-8 15°- 9 160.3 162.7 39 31 39 42 39 Burt. C. I. 293 {Avena sativa), May 15 to I, 106 ',433 1,448 1,135 i,°85 1,135 435 466 453 438 456 447 i,224±55 449 ±3 85 86 87 88 89 t 9° 402. 2 395- 7 409- 5 412. 2 389. 8 366.8 167. 7 145-9 155-9 152-9 145-9 144- 5 167. 2 171- 5 171. 2 164. 0 169. 7 161. 0 42 37 38 37 37 39 Swedish Select, C. I. 134 {Avena sativa). May 17 to Aug. 23. . . 997 1,176 1,098 i,°73 1,163 1,113 416 434 418 398 435 439 I, io3=ti8 423:fc5 79 80 81 82 83 84 265-5 250. I 262. 4 296- 7 286. I 291.9 84. 2 83.0 77-4 106. 9 loi. 4 95° 171. 4 164. 2 158-5 175-3 174-3 174.0 32 29 36 35 i5 1913- Swedish Select, C. I. 134 {.Avena sativa), May 23 to Aug. i 2,035 1,979 2,049 1,640 1,719 1,831 646 656 604 591 609 596 i,876±55 6i7±9 85 86 87 88 89 I 9° 243-7 245-6 240.4 255-5 250. 0 254-9 93-4 loi. 8 87.6 89-3 94-7 94-5 148.8 151.1 156-5 157-0 155-7 149-3 38 41 36 35 38 37 Burt, C. I. 293 (Avena sativa), May 23 to Tulv 2K 1,594 1,484 1,787 1,758 1,644 1,58° 611 616 650 614 623 586 I, 641 ±33 6i7±s 1 1 Oct. 15, 1914 Water Requirement of Plants 13 BARLEY Barley is the most uniform in water requirement of the small-grain crops which the writers have tested. The four varieties grown at Akron in 1 91 2 showed only slight differences in their water requirement. The results obtained, based upon the production of dry matter, were as follows : Variety of barley Water requirement Beardless 403 ±8 Beldi 4i6±4 White Hull-less 439 ±1 Hannchen 443 ±3 These same varieties were also tested at Akron in 191 1 and were found to be in practical agreement as regards their relative water require- ment. The mean value of the water requirement was 27 per cent higher in 191 1 than in 1912. The results obtained with barley when the water requirement is based on grain production are less uniform than when the total dry mat- ter is employed. Reference to Table VII will show that this is often due to a single pot which for some reason fails to set grain as abundantly as the rest of the series. The Beldi, a dwarf variety, showed the highest efficiency in the use of water in grain production. The White Hull-less (PI. Ill, fig. 2) has a water requirement slightly above the other varieties, even when a correction is made for the naked character of the grain. Table VII. — Water requirement of different varieties of barley at Akron, Colo., in IQ12 Water requirement Plant and period of yrow-th. Pet No. Dry matter. Gram. Water. Grain. Grain. Dry matter. 1912. Grams. Grams. Kilos. Percent. 91 301-5 141-7 138. I 47 974 458 Hannchen, C. I. 531 92 282.9 116. 6 127.7 41 1.09s 452 {Horde um distiehon), 93 281. I 103- 5 124. 2 37 I, 200 442 May 16 to Aug. 28. . . 94 300-3 137- I ^33- 3 46 972 444 95 309.2 145-5 134-9 47 926 436 I 90 3°8-5 152. 0 130.9 49 860 424 Mean i,oo5±36 443 ±3 f 97 230-5 loi. 5 93-9 44 925 407 Beldi, C. I. 190 {Hor- deum -vulgare), May 16 to Aug. 12 98 240. I 107. 0 98-5 45 920 410 99 243.0 log. 2 99-5 45 910 409 100 189. 4 81.3 82. I 43 1,010 434 lOI 223.8 loi. 6 97.0 45 954 434 102 236.5 103. I 95-5 44 926 404 Mean 941 ±10 f 103 271.2 95-3 119. 7 35 1,256 441 White Hull-less, C. I. 104 276. 0 97. I 121. 4 35 I, 250 440 595 (Hordeum vul- 105 273- I 95-6 119-4 33 1,249 437 gare). May 16 to 106 268.8 92-9 "9-7 35 1,289 445 Aug. 12 107 273-8 299-7 99-5 no. 0 118. 4 131. 6 36 37 1, 190 I. 197 433 439 1 108 I, 239±ii 439 ± I 14 Journal of Agricultural Research Vol. Ill, No. I Table VII. — Water requirement of different varieties of barley at Akron, Colo., in IQ12- Continued Plant and period of growth. Pot No. Dry matter. Grain. Water. Grain. "Water requirement based on — Grain. Dry matter. 19J2. Beardless, C. I. 716 (Hordeum vulgare), May 16 to Aug. 23. . . ■ 109 no III 112 "3 I 114 Grams. 112. 0 144. I 276.3 227-5 210. 5 291-3 Grams. 20.7 21. 9 135-7 92.2 63-9 120. 2 Kilos. 49.2 56.1 III. 6 93-8 87.2 103-7 Percent. 18 15 49 40 3° 41 '■■'823 1,017 1,364 862 439 389 408 412 414 356 Mean i,oi7±83 403 ±8 RYE The measurement of the water requirement of spring rye at Akron in 191 1 showed a surprisingly high figure — 54 per cent above that of Kubanka wheat. The 191 2 measurements (Table VIII) gave 496 ±9 for the water requirement of rye when based on dry matter and i ,802 ± 62 when based on grain production. The 191 2 (dry matter) ratio is thus about 26 per cent above Kubanka wheat, a marked increase in the rela- tive efficiency in comparison with the 1911 ratio. In fact, rye exhibited the greatest reduction in water requirement of all the crops tested in 1912. A consideration of the water requirement of crops grown out of season in 191 1 showed that rye was unusually efficient during the cool fall period. This result as well as the increase in efficiency in 191 2 suggests that rye may be unusually responsive to climatic conditions and that it is relatively better adapted to low temperature than the other small grains. Table VIII. — Water requirement of rye at Akron, Colo., in igi2 Pot No. Dry matter. Grain. Water. Grain. Water requirement based on — Grain. Dry matter. I9I2. Rye, spring, C. I. 73 {Sccale cereale). May 16 to Aug. 23 116 117 118 119 120 Grams. 210. 6 216. 4 179.2 248.6 234-5 254- I Grams. 57-8 65.1 52.6 65.8 54-7 75-6 Kilos. 96. 2 118. 8 84-3 123-3 120. I 124. 6 Per cent. 27 30 29 26 23 3° 1,664 1,825 1,603 1,874 2,195 1,649 457 549 470 496 512 490 1, 802 ±62 496 ±9 Oct. 15, 1914 Water Requirement of Plants 15 RICE Rice was grown for the first time at Akron in 191 2. The crop was slow in becoming established, and the growth period was relatively long. No grain was produced. The water requirement based on dry matter was 5i9± 13 (Table IX). It thus appears that rice, although a crop normally grown with an abundant water supply and in a relatively humid climate, is about as efficient in the use of water as rye. Its relative position might be materially changed if the tests were made in a warmer climate. Rice was also included in the 191 3 measurements (PI. VII, fig. 3). The stand was good and the growth was uniform and luxurious, but the season was too short to produce grain. The water requirement in 191 3 was 744±i7, or 43 per cent higher than in 1912. Table IX. — Water requirement of 'rice at Akron, Colo., in IQ12 and l()lj Plant and period of prowth. Pot No. Dry mat- ter. Water. Water require- ment based on dry matter. 1912. Rice, Honduras, C. I. 1643 (Oryza saiiva), Mav ''7 to Sent. 2% 151 152 153 154 155 I 156 Grams. 248.4 253-1 232.5 168.9 194.0 225. 6 Kilos. 133- 5 141. 2 127.4 9°- 5 94.8 100. 0 538 558 548 536 488 445 5iQ±i3 157 158 159 160 161 162 276.7 261. 7 230.2 274.8 294.8 298.7 218. 2 211-5 176.9 186.6 204. 0 215.8 1913- Rice, Honduras, C. I. 1643, June :2 to Sept. 16 790 809 768 680 692 722 Mean 744±i7 Kl.AX Flax was included in the water-requirement measurements at Akron for the first time in 1913. Its water requirement was found to be very high, 905 ±25 based on dry matter and 2,835 ±52 when based on seed production. It will thus be seen to have a water requirement as high or higher than any of the legumes tested in 1913. This is in accord with the measurements made by Leather (191 1, p. 270) in India, in which flax was exceeded in water requirement only by chick-peas and rice. At Akron in 191 3 flax required 22 per cent more water than rice. The detailed results are jriven in Table X. i6 Journal of Agricultural Research Vol. Ill, No. I T.ABLE X. — Waler requirement of flax at Akron, Colo., in igij Plant and period of growth. Pot No. Dry matter. Grain. Water. Grain. Water requirement based on— Grain. Dry matter. 1913- Flax, North Dakota, No. 155 {Linum usiia tiss i mu m), June 3 to Sept. i. . . . 127 128 129 130 131 132 Grams. 189. 0 154-7 209.3 93-0 116. 7 187.8 Grams. 64.9 53-0 77-2 23-9 32-5 63-7 Kilos. 184.8 143-4 210. 6 76.3 93-9 169. 2 PercenU 34 34 37 26 28 34 2,845 2,704 2,728 3,190 2,886 2,658 978 927 1,006 811 804 902 2,835±52 905±2S SUG.AR BEETS The water requirement of the sugar beet was again measured at Akron in 1912. The ratio 321 ±8 was obtained on the basis of total dry matter and 524±23 on the basis of the dry root (Table XI). This is about 15 per cent below the 191 1 value, a reduction in water requirement similar to that shown by the other crops tested during the two seasons. The sugar beet is an efficient plant in the use of water, being the equal of the corn group in this respect. T.-\Bi.E XI. — Water requirement of sugar beds at Akron, Colo., in igi2 Pot No. Dry matter. Roots. Water. Roots. Water requirement based on- growttl. Dry roots. Total dry matter. 1912. Sugar beet {Beta vul- garis), June 9 to Oct. 12 169 170 171 172 173 I 174 Grams. 257-4 175-0 173-3 163.6 196.8 164- 5 Grams. 189.7 93-6 100. 0 107.3 123.0 97-7 Kilos. 76.0 55-7 49.6 52-5 68. 7 58-9 Per cent. 74 53 58 66 63 60 401 595 496 489 558 603 29s •318 286 321 349 358 Mean 524±23 32i±8 COTTON Cotton was tested at Akron for the first time in 1912. Numerous bolls set, though none opened. The plants grew slowly during the first part of the season, owing probably to the cool nights. The observed water-requirement ratio was 488 ±14 (Table XII). Oct. 15, 1914 Water Requirement of Plants 17 Table XII. — Water requirement of cotton at Akron, Colo., in igi3 and igij Plant and period of growth. Pot No. Dry matter. Water. Water require- ment based on dry matter. 1912. Cotton, Triumph {Gossypium hirsutum), July 19 to Sept. 21 . 181 182 183 184 i8s 186 Grams. 38.9 32.0 67.4 61.4 51.6 62. 4 Kilos. 19.2 18.5 27. I 28.2 25. 2 31- S 494 578 402 459 488 505 488 ±14 163 164 165 166 167 168 237-5 184.3 247.4 249.6 161. 4 187.0 172.4 IIS- 6 166. 1 160. I 106. 4 115. 4 Cotton, Triumph, May 29 to Sept. 16 726 627 671 642 660 617 657±ii The same variety was also included in the 1913 measurements. (PI. VII, fig. I.) The planting was made earlier, and a much larger growth was obtained. The water requirement was 657^11, or about one-third higher than in 191 2. In this connection it should be stated that at Akron cotton is far north of its natural range, which may have increased its relative water requirement. CORN AND TEOSINTE Six varieties of com (Zea mays) were tested at Akron in 191 2 (PI. VII, fig. 5). Three of these varieties, Northwestern Dent (PI. IV, fig. i), Iowa Silvermine, and Esperanza, had also been used in the 191 1 experiments. The three new varieties were furnished by Mr. G. N. Collins, of the Bureau of Plant Industry, and represent widely different strains. The Hopi variety (PI. IV, fig. 2) is grown by the Hopi Indians in northwestern New Mexico (Collins, 1914); China White is a variety from near Shanghai, China; while Laguna was originally from the State of Chihuahua, Mexico. The water requirement of each variety tested in 191 2, based on the production of dry matter, is as follows: Variety of corn Water requirement Esperanza ^39:^^5 Northwestern Dent 28o± 10 Hopi 285±7 Laguna 295 ±6 Iowa Silvermine 302 ±7 China White 3iS±7 The Esperanza, as in 191 1, leads all the varieties, so far as efficiency in the production of dry matter is concerned, and ranks with the sor- 60300°— 14 2 i8 Journal of Agricultural Research Vol. Ill, No. I ghums in this respect. The differences exhibited by the remaining varie- ties are without significance when the limitations imposed by the probable errors are considered, although the quick-maturing Northwestern Dent and Hopi varieties appear to be slightly more efficient than the others. The detailed data are given in Table XIII. The pollination was not adequate to give representative grain yields. Five varieties of corn and one of teosinte were included in the 191 3 measurements. The water requirement of each variety, based on the production of dry matter, is as follows : Variety of com or teosinte Water requirement Indian Flint com 342 ± 5 Hopi com 3So±8 Teosinte, Durango 390±ii Northwestern Dent com 399i 12 Bloody Butcher com 405 ± 7 China White com 41 5 ±4 The most efficient varieties were the Indian Flint (PI. VI, fig. 5), a small variety grown by the Indians of northern Michigan, and the Hopi, another dwarf Indian variety. Teosinte, Northwestern Dent com, and Bloody Butcher, a local variety of corn grown near Wray, Colo., showed only slight differences. The China White, as in 1912, proved to be the least efficient of all the varieties tested, having a water require- ment 20 per cent above that of the Indian varieties. In 1912 the water requirement of the Northwestern Dent com was in practical agreement with that of the Hopi, while in 191 3 the Hopi gave a considerable lower value. The China White required 32 per cent more water in 1913 than in 1912; the Northwestern Dent, 42 per cent; and the Hopi, 23 per cent. The water requirement of certain com hybrids was also measured at Akron in 1912 and 1913. The mean water requirement of each strain has been included in the tables in the summary (Tables XXIX to XXXII) . Table XIII. — Water requirement of different varieties of corn and teosinte at Akron, Colo., in igi2 and igi^ Pot No. Dry matter. Grain. Water. Grain. Water requirement based on — Grain. Dry matter. igi2. ■ 277 278 279 280 2S1 282 GraTiis. 299.0 344.2 368.5 649.0 440.0 491.0 Grams. 18. I 53-9 66. I 234.4 loi. 5 117. I KUos. 102. 6 100. 4 i°5-5 161. I III. 2 126. I Per cent. 6 16 18 36 23 24 343 292 286 Q)ra, Northwestern Dent {Zea mays). 248 253 257 June 9 to Sept. 16. . . Mean 28o±io Oct. 15. 1914 Water Requirement of Plants 19 Table XIII. — Water requirement of different varieties of corn and tcosinte at Akron, Colo., in igi2 and igij — Continued Pot No. Dry matter. Grain. Water. Grain. Water requirement based on — Plant and period of growth. Grain. Dry matter. 1912. f 283 284 285 286 287 , 288 Grams. 596.0 403.7 399-7 441- 5 477-0 437-2 Grams. Kilos. 167.9 129.9 133-4 129. 6 135-6 128.9 Per cent. 282 322 334 294 Com, Iowa Silvermine (Zea mays), June 8 to Sept. 26 295 Mean 302 ±7 ■ 295 296 297 298 299 300 433-0 519- 5 516.8 415-8 330-9 364-7 132-7 140.3 131-1 "3-4 100. 2 111.3 307 270 254 273 303 305 Com, Hopi (Zea mays), June 12 to Sept. 26. . 285±7 f 265 266 267 268 269 . 270 243- S 577- 0 319- 0 524. 5 660. 9 401.5 84-3 184.6 97-1 173-9 179.6 126. I 346 320 304 331 Com, China White (Zea mays), June 12 to Sept. 26 272 314 3i5±7 289 290 2gl 292 293 I 294 376.6 261. 2 268.9 448.1 457-4 429.2 112. 4 83.8 84.0 124.4 127.8 119. 4 298 321 313 278 279 278 Cora, Laguna (Zea mays), July 2 to Sept. 26 Mean 295±6 301 302 303 . 304 492-3 574.7 563-7 510.7 114- 3 133-7 141-7 122. 7 Com, Esperanza (Zea 232 233 252 mays), June 12 to 240 Mean 239±3 247 248 249 250 251 252 411.5 485.4 456.6 501.9 499-4 456.8 174. I 201. 5 188.6 193-0 183.6 195-4 1913- 423 415 413 385 Com, Bloody Butcher (Zea mays), June 7 ... . 368 428 Mean 40s ±7 253 254 255 256 257 I 258 333-4 397-9 380.6 292. 0 335- I 380.3 123.4 163.3 161. 7 loi. 6 135-2 182.6 114.9 130. 5 120. 6 108.2 118. 7 128.6 37 41 42 35 40 48 Com, Indian Flint (Zea mays), June 7 931 800 746 1,064 878 704 345 328 317 370 354 338 Mean 854 ±39 34i±S 20 Journal of Agricultural Research Vol. Ill, No. I Table XIII. — Water requirement of different varieties of corn and teosinte at Akron, Colo., in igi2 and IQIJ — Continued Plant and period of growth. Pot No. Dry matter. Grain. Water. Grain. Water requirement based on — Grain. Dry matter. -913- f 283 284 285 286 287 288 Grams, 351.2 392.3 336.6 285.9 389.9 382.0 Grams. 131. 6 117.5 100. 0 Kilos. 153.7 142.9 136.6 128.7 141-4 143-4 Per cent. 436 364 406 451 363 375 Corn, Northwestern 34 1,086 Dent {Zea mays). June 7 to Sept. 6 ,. . , 3° 26 1,203 1,434 Mean I, 241 ±77 399±i2 313 314 315 316 317 I 318 346.7 400. 0 472.5 417. I 405.3 619. 6 118. 8 135-8 170. 2 147.0 163-4 185.4 343 340 360 352 403 300 Com, Hopi (Zea mays). June 14 to Sept. 16. . Mean 35° ±8 301 302 303 3°4 305 I 306 554.9 487-5 492.6 589.2 478-7 523-1 228.2 210. 4 202.3 228.1 198.7 226. 6 41 I Com, China \^'hite {Zea mays), June 7 432 411 387 to Sept. 16. . .... 415 433 Mean 415 ±4 289 290 291 292 293 I 294 616. 4 534.5 624. S 567.3 520. 0 421.4 234-7 211. 6 194.0 231-5 214.4 183.2 ^So Teosinte, Durango (Euchlaena m e xi- 396 408 cana), June 14 to Sept. 16 412 435 Mean 390±ii SORGHUM The investigation of the water requirement of the sorghums (Table XIV) is of special interest, owing to the marked efficiency exhibited by this group of plants in the use of water. The eight varieties grown at Akron in 191 2, together with the water requirement based on the pro- duction of dry matter, follow: Variety of sorghum Water requirement Brown kaoliang 223 ± i Red Amber 237 ±4 Minnesota Amber 239±2 Milo 249±3 White durra 255 ±3 Blackhull kafir 259±5 Dwarf milo 273±4 Sudan grass 359±2 Oct. 15, 1914 Water Requirement of Plants 21 The Brown kaoliang gave the lowest water requirement. Red Amber and Minnesota Amber, forage varieties of sorghum (PL IV, figs. 4 and 5), gave practically the same ratio, which is but slightly higher than Brown kaoliang. Tabuc XIV. — Water requirement of different sorghums at Akron, Colo., in igi2 and IQI3 Pt.t No. Dry matter. Grain. Water. Grain. Water requirement based on — Grain. Dry matter. I9I2. Red Amber, S. P. I. ; 17543 (.^ndropogon sorghum), June 9 to Sept. 27 253 254 255 256 257 258 Grams. 400. 5 541.7 592. 0 660. 5 564.0 714.2 Grams. 43.8 41-5 58.1 95-4 44- 3 93-5 KUos. 103. I 133-8 135-5 144.8 '35-1 161. 0 Per call. II 8 10 14 8 13 2,351 3.221 2,331 i,5'8 3,050 I, 722 258 247 229 219 226 Mean 2,366±i94 237±4 247 248 249 250 251 I 252 666.6 370.6 543-0 461.3 456.6 425-3 272. 6 139.9 218.4 168. I 195-7 163-3 151. 2 8g. 6 128. I no. 6 no. 3 103. 6 41 38 40 37 43 38 Minnesota Amber, A. D. I. 341-13 (.\ndro- pogon sorghum), June 9 to Sept. 26. . . 554 645 586 658 564 634 227 242 236 240 242 244 607 ± IS 239 ±2 217 218 219 220 221 222 434-5 370.2 334-3 403-7 420. I 301. 2 83.0 79-9 55-3 78.5 84-5 64- 5 115. I 102. 0 90. 6 105.7 no. 2 91.8 19 22 17 19 20 21 Milo, Dwarf, S. P. I. 24970 (.Andropogon sorghum), June 9 to Sept. 27 1,387 I, 276 1,638 1,347 1,304 1,422 265 275 271 262 262 30s Mean i,396±34 273±4 223 224 225 226 227 228 475-4 440.4 472.5 488.9 499-9 509-9 60. 0 85-7 77-9 114. 0 116. 0 79.6 125. 6 103. 7 114- 5 123-3 123.0 128.8 13 19 17 23 23 16 Milo, S. P. I. 24960 {Andropogon sor- ghum), June 9 to Sept. 27 2, 092 I, 210 1,470 1,081 I, 060 1,618 264 235 242 252 246 253 Mean I,422±IIS 249 ±3 f 235 236 237 238 239 240 506.9 568.9 432. 9 384.6 406. 2 449- 6 138. 7 142.7 85.6 63. 5 78.. 108.6 129.3 142. I 106. 7 105. I 99-4 115.9 27 25 20 '7 19 24 Durra, White, S. P. I. 24997 {Andropogon sorghum), June 9 to Sept. 26 925 996 I, 246 1,656 1,273 1,067 255 250 247 273 245 258 Mean i,i94±7S 255±3 241 242 243 244 245 346 588.9 411. I 548-9 556.0 526. 0 57'. 8 144.8 108. 8 '25-3 126. 0 171. 0 108.7 134.0 94-8 120. 5 123.8 116. 8 123.9 25 26 23 23 3i >9 Kaoliang, Brown, S. P. I. 24993 (Andropo- gon sorghum), June 925 870 962 982 683 1, 140 228 230 220 223 217 Mean 927±38 223 ± I 22 Journal of Agricultural Research Vol. III. No. I Table XIV. — Water requirement of different sorghums at Akron, Colo., in igi2 and igi3 — Continued plant and period of growth. Pot No. Dry matter. Grain. Water. Grain. Water requirement based on — Grain. Dry matter. 1912. 229 230 231 232 233 234 Grams. 247.0 409. I 494-5 413-° 363-3 377-1 Grams. Kilos. 72. I 108.4 119- 3 lOI. 8 91.9 97-4 Per cent. 292 265 241 Kafir Blackhull S. P. I. 24975 ('^w- drofyogon sorghum), June 9 to Sept. 27 246 253 258 Mean 259±5 211 212 213 214 215 216 190. 6 209.4 181. 8 205. 8 215-9 222. 5 60.8 65.8 59-4 63-4 63.8 68.6 319 314 326 308 296 Sudan grass, S. P. I. 25017 {Andropogon sorghum aclhiopictts), first crop, May 28 to Tulv 26 308 Mean 3i2±3 211 212 213 214 215 216 106. .7 94- 0 95-6 106. 5 91. 0 88.0 46.1 43-5 44. 6 45-9 42,9 43-0 432 463 467 431 472 Sudan grass, S. P. I. 25017, second crop, July 26 to Sept. 6 489 Mean 459 ±7 211 212 213 214 215 I 216 297-3 303-4 277.4 312-3 306.9 310. 6 106. 9 109.3 104. 0 109.3 106. 7 III. 6 359 ^60 Sudan grass, S. P. I. 25017, combined crop, May 28 to Sept 6. 375 350 348 360 359 ±2 ' 265 266 267 268 269 270 557- 8 585-0 591.8 677-5 577- 9 643.0 213.0 256.9 235-4 263. 0 204. 0 150- 7 164- 5 iSi. 4 177-4 203. I 167- 5 189.7 38 44 40 39 35 23 1913- Sorghum, Minnesota Amber, A. D. I. 34T-13 (Andropogon sorghum), June 14 to Sept 15. . 772 706 754 773 821 295 310 299 301 290 295 Mean 765 ±12 298 ±2 f 277 278 279 280 281 282 689.7 670.7 636.2 644. I 682.3 749- 7 209.4 190. 6 160. 2 165.6 187.4 187.8 200. 5 196. 0 187.9 193-9 200. 5 225.2 30 28 25 26 23 25 Sorghum, Red Amber, S. P. I. I7S43 (An- dropogon sorghum), June 7 to Sept. 15. . . 958 1,028 1,172 1,170 1,070 1,199 291 292 295 301 294 301 I, ioo±3i 296±i Oct. IS, 1914 Water Requirement of Plants 23 The least efficient variety tested in the sorghum group is Sudan grass (PI. V, fig. i), a forage plant which has recently received consider- able attention in the southern Great Plains. Only one year's meas- urements are available for Sudan grass, but the results so far indicate that it is not the equal of other well-known varieties of sorghum in efficiency in the use of water. Sudan grass required 40 per cent more water than Brown kaoliang for the production of the first crop. The second crop was light at Akron and had a much higher water require- ment. On the basis of the two cuttings combined, the water require- ment of Sudan grass was 62 per cent higher than Brown kaoliang. As a forage crop, however, the shorter and more slender stalks of Sudan grass may offset the disadvantage of its higher water requirement. In the production of grain the Minnesota Amber' variety gave the lowest water requirement ratio so far recorded for a sorghum crop, viz, 607 ±15. The Minnesota Amber produced a pound of grain at Akron in 1 91 2 with less water than was required by alfalfa in the production of a pound of hay. The high water requirement for grain production in Red Amber sorghum, Dwarf milo, milo, and White durra (PI. IV, fig. 3) is largely due to an attack of aphids, which caused many of the flowers to fail to produce seed. The parasites were killed by spraying early enough to prevent any serious reduction in total growth. The 1 91 3 water requirement measurements of sorghum were confined to two varieties, Red Amber and Minnesota Amber, both of which were included in the 191 2 measurements. The two varieties gave in 191 3 practically identical water-requirement ratios — namely, 296 ±1 and 298 ±2. The results from individual pots were in excellent agreement as indicated by the small probable error. A similar agreement was observed in 191 2. Each variety in 191 3 showed an increase of 25 per cent in the water requirement as compared with 191 2. A series of water-requirement measurements were made at Amarillo, Tex., in 1913, for the purpose of determining the influence of climatic environment on the water requirement. These measurements also included a number of sorghum varieties, the water requirement of which had never before been determined. Plants have a higher water require- ment at Amarillo than at Akron, so that measurements of different plants at the two stations are not directly comparable. The water requirement of Red Amber and Minnesota Amber sorghum was meas- ured at both stations in 1913, and the ratio of these measurements affords a means for reducing the Amarillo values to the basis of the Akron meas- urements. The mean water requirement of these two varieties at Akron was 85 per cent of that at Amarillo. The Amarillo water-requirement measurements as given in Table XVI have been reduced accordingly ' This variety was represented by a strain selected for its drought resistance by Mr. A. C. Dillman. of the Oflice of Alkali and Drought Resistant Plant Investigations. 24 Journal of Agricultural Research Vol. Ill, No. I for comparison with the Akron measurements. The computed values for Akron are given in Table XV. Table XV. — Observed water requirement of varieties of sorghum at Amarillo, Tex., and computed water requirement for Akron, Colo., in IQ13 Variety. Dwarf BlackhuU kafir White kafir Early BlackhuU kafir. White mile KafirXdurra Feterita Observed Computed water water requirement requirement at Amarillo. for Akron. c>y^^2, 285±3 349±4 297±4 3S6±i5 302 ±13 373±3 3i7±3 378±5 32i±5 38o±4 323±4 It will be noted (Table XV) that Dwarf BlackhuU kafir and iMinnesota Amber sorghum were the most efficient in the use of water of the eight varieties of sorghum tested at Amarillo in 191 3. The least efficient was feterita. The kafir Xdurra hybrid had practically the same water requirement as feterita. The latter has been extensively featured recently as a drought-resistant crop particularly adapted to the South- west. It does not appear, however, that its drought-resistant qualities are ascribable to an efficiency in the use of water, this variety being the highest in water requirement of all the sorghums tested at Amarillo in 1913. Vinall and Ball (1913, p. 27) have suggested that the success of feterita during recent dry years has been due to a thin stand resulting in part from poor germination. When grown under identical conditions as to stand, it showed no greater drought resistance than milo or kafir. Table XVI. — Water requirement of sorghum at .Amarillo, Tex., in IQIJ Pot No. Pry matter. Grain. Water. Grain. Water requirement based on — Grain. Dry matter. 1913- Sorghum, Red Amber, S. P. I. 17543 <-'!«- dropogon sorghum). May 15 to Aug. 20. . . 43 44 45 46 47 48 Grams. 717-4 682.2 723.6 701.9 741-3 768.7 Grams. 159-3 137-9 205-9 "7-4 197.0 208.7 KUos. 267.8 262. 4 268.0 256. 2 270.7 272.9 Per cmt. 22 20 28 17 27 27 1,680 1,904 1,302 2, 182 1,374 1,306 373 385 371 36s 365 355 Mean I, 625±II2 369*3 49 5° 51 52 53 I 54 590. 2 588.3 612. 7 646. 0 577- I 649-5 233-2 295-4 291.9 308.9 264.5 288.9 196. 2 196. 6 199.7 208. 1 201. 0 211. I 39 5° 48 48 46 44 Sorghum, Minnesota Amber, A. D. 1. 341- 13 (Andropogon sor- ghum), May 15 to Aug. 8 841 666 685 674 760 731 332 334 326 322 348 325 Mean 726±i9 33i±3 Oct. 15, 1914 Water Requirement of Plants 25 Table XVI. — Water requirement of sorghum at Amarillo, Tex., in igij — Continued Water requirement based on — Plant and period of growth. Pot No. Dry matter. Grain. Water. Grain. Grain. Dry matter. 19I3. Grams. Grams. Kilos. Per cent. 61 464-5 80.4 173-5 17 2.159 374 Milo, White, C. I. 365 62 481. I 107-3 166.3 22 1.550 346 {Andropogon sor- 6,3 453- 0 75-6 171. 9 17 2,272 380 ghum), June 7 to 64 470.0 108. 4 176.4 23 I, 626 375 Aug. 22. . 65 66 449-3 471. I 79.6 112. 2 170.9 179.6 18 2, 148 I, 601 380 381 24 Mean I, 893 ±"3 373±3 67 587-4 182. 7 198.6 31 1,087 338 Kafir, Early Black- 68 577- 8 178.6 207.3 31 I, 160 359 hull, C. I. 472 {An- 69 599- 4 235-7 197.6 39 839 330 dropogon sorghum), 70 530-6 193-4 i6g. 8 36 878 320 June II to Sept. 16. . 71 440. 2 220. 2 206. 9 50 940 470 72 603.4 199.8 192. I 33 962 319 Mean 978±37 356±i5 f 1^ 592.5 254- 0 190. 8 43 752 322 Kafir, Dwarf Black- 74 621. 6 276. I 200. 0 44 724 322 hull, C. I. 340 (An- 75 562.7 205. 2 191- 5 36 934 340 dropogon sorghum). 76 586.7 260. 0 197.9 44 761 337 June II to Sept. 16. . 77 544-8 195. 6 191. 7 36 980 352 I 78 559-3 152-7 187.0 27 I. 223 334 Mean 896±57 335±3 f 7Q 555-3 185.9 201. 7 ii 1,086 363 Kafir, White, C.I. 370 80 546-2 106. 4 200. I 19 a 1,882 366 {Andropogon sor- 81 571-8 210.8 198-5 37 942 347 ghum), June II to 82 584. 2 248-7 198. 0 43 796 339 Sept. 22 83 569. 6 221. 0 194-3 39 879 341 84 579- 6 228.3 195-9 39 858 338 Mean 9i2±34 349±4 f ^^ 552- 0 226. 5 200. I 41 884 363 Kafir Xdurra, hybrid 86 538- 4 215-3 193- I 40 898 358 198-15-3 {Andropo- 87 539- 7 217. 6 199. I 40 916 369 gon sorghum) ]une II 88 515-° 206.6 194-7 40 942 378 to Sept. 22 89 51°- 7 179-5 197. I 35 1,098 386 I 90 471- 5 158. I 194. 0 23 I, 226 411 Mean 994±42 378±S f 91 547-8 181. 3 212. I 3Z I, 170 387 Feterita.C. I. 182 (.4 k- 92 .585- 4 235- 8 216. 0 40 916 369 dropogon sorghum) 93 619.3 256-9 226. 9 41 884 366 June II to Sept. 18. . 94 531- I 210. 9 206. I 41 979 388 95 562.8 210. 2 208. ■; 37 992 370 I 96 512.0 191-3 203.8 37 I, 064 398 Mean I, 001 ±29 380 ±4 0 Omitted in computing the mean. 26 Journal of Agricultural Research Vol. Ill, No. I MILLET AND PROSO These plants are remarkable in that they outrank all others so far tested as regards efficiency in the use of water (Table XVII). The four varieties grown at Akron in 191 2 gave the following water requirement, based on the production of dry matter: Variety of millet or proso Water requirement Kursk millet iS7±2 Voronezh proso 206 ± i Tambov proso 2oS± i German millet 248±7 Kursk millet (PI. Y, fig. 3) represented by a strain developed by Mr. A. C. Dillman, of the Office of Alkali and Drought Resistant Plant Investigations, gave the lowest water requirement so far recorded for any crop at Akron. The two prosos, Tambov and Voronezh (PL V, fig. 2), have a water requirement about 10 per cent higher than the Kursk, while German millet is 33 per cent higher than the Kursk. Aside from the German millet, all of the varieties tested have a water require- ment distinctly below the best of the sorghums, the group ranking next in efficiency. Table XVII. — Water requirement of different millets and prosos at Akron, Colo., in igi2 and ipij Pot No. Dr^' matter. Grain. Water. Grain. Water requirement ba^ed on — Grain. Dry matter. I9I2. Millet, Kursk, S. P. I. 34771 (Chaetochloa italica), June 9 to 205 206 207 208 209 . 210 Grams. 190. 7 320.0 241.7 192.3 178.9 332.6 Grams. 66. I 119. I 102. 6 78.4 66. 5 139.3 Kilos. 36.7 58.9 44- I 36-1 32.6 65. 0 Per cenL 35 37 42 41 37 42 555 494 430 461 490 467 192 184 182 1 88 182 195 Mean 483±ii i87±2 f 187 188 189 190 191 , 192 115. I 202. 6 140. 0 350-3 92.5 80.5 27-3 46.7 33-9 82.4 22.4 24. 2 237 231 242 Millet German S P I 26S45 (Chaetochloa italica), July 2 to Scot. 21 23s 242 3°i Mean 248+7 193 194 195 196 197 . 198 255-3 342.9 igo. 0 317-6 336.2 282. 7 111. 0 162.4 80.7 140.2 142.8 112. 6 54.8 68.0 39-7 66.5 70. I 58.9 43 47 42 44 43 40 Proso, Tambov, S. D. 366, Akron, 366-1-10 (Panicum miliaceum), June 8 to Aug. 12. . . . 494 419 492 474 491 523 214 ig8 209 209 208 208 482 ±9 208 ± I Oct. 15, 1914 Water Requirement of Plants 27 Table XVII. — Water requirement of different millets and prosos at Akron, Colo., in igi2 and igij — Continued Plant and period of growth. Pot No. Dry matter. Grain. Water. Grain. Water requirement based on— Grain. Dry matter. 1912. Proso, Voronezh, C. I. 16 (Panicum milia- cf Mm), June 5 to Aug. 199 200 201 202 203 . 204 Grams. 238- 7 296. 1 281.9 298.5 256. 1 273.0 Grams. 114. 2 143-7 135- 6 149. 5 "9-4 133-2 Kilos. 48.9 61.2 56.8 59-9 53-7 57- 0 Per cent. 48 48 48 50 47 49 428 426 419 401 450 428 205 207 201 20I 209 Mean 425±4 206 ± I 259 260 261 262 263 264 174-8 281. 7 17s- 4 183-3 250. 6 201. 8 37- I 76.1 71.6 31-9 33-4 74-0 48.9 80. 7 50-5 51-3 66.3 63-5 21 27 41 17 13 37 1913- Kursk, S. P. I. 34771 (Chaetochloa ilalica), June 14 to Aug. 26. . 1,318 I, 060 705 858 280 286 288 280 26s 315 Mean 286±4 In grain production the millets make a remarkable showing, the prosos leading in this respect. Measurements of the water require- ment of the three varieties, based on grain production, gave the follow- ing results : Variety of millet or proso Water requirement Voronezh proso , 425±4 Tambov proso 482 ±9 Kursk millet 483 ±11 Voronezh proso, according to these figures, is able to produce nearly 2 pounds of grain with the water required for the production of i pound of alfalfa hay. Kursk millet was also included in the 1913 measurements. Its water requirement was 286±4, or 53 per cent higher than in 1912. Many of the plants were broken by a high wind shortly before harvest, which greatly reduced the grain yield. The mean water requirement, based on grain production, has consequently been omitted. ivEGUMES The legumes tested at Akron in 191 2 included sweet clover, chick-pea, and two strains of Grimm alfalfa, one being a selected strain (PI. V, figs. 4 and 5) developed by Mr. A. C. Dillman. Both the alfalfa and the sweet clover showed a marked reduction in water requirement compared with the results obtained in 191 1. Three cuttings were made in the case of each crop, but the plants were not mature at the time 28 Journal of Agricultural Research Vol. III. No. I the last cutting was made. The following values (Table XVIII) were obtained for the water-requirement ratio: T.^BLE XVIII. — Summary of water-rajuiremcnt measurements of legumes at Akron, Colo., in igi2 Crop. Alfalfa, Grimm Alfalfa, Grimm, A. D. I. selection. Clover, sweet Chick-pea Cutting. 592 ± 13 6oo±i7 547 ±12 790±io 853 ±13 677±i4 Third. Combined. 506 ±5 421 ±10 598±i8 65g±6 657±ii 638±4 5io±i4 The two alfalfas and the sweet clover were planted on the same day, and the crops in each instance were all cut on the same day, so that the results in the Table XVIII are comparable. The A. D. I. strain of Grimm alfalfa gave a slightly higher ratio than the unselected Grimm during the second period, but lower during the third period, when it made a better growth. (See "Dry matter," column 3, Table XIX.) Sweet clover, as in 191 1, proved somewhat more efficient than alfalfa during the first and second periods. During the third period sweet clover was less efficient than alfalfa. The chick-pea proved the most efficient of the legumes tested. Its growth period does not coincide with that of the other legumes, but approximates a combination of the lirst and second periods. (See Table XIX.) It thus appears to be distinctly more efficient in the use of water than either alfalfa or sweet clover. Chick-pea has, however, a relatively high water requirement compared with the small grain crops, which is in accord with Leather's measurements (Leather, 1910, p. 156). Table XIX. — Water requirement of different legumes at Akron, Colo., in igi2 Pot No. Dry matter. Grain. Water. Grain. Water requirement based on — Plant and period of growth. Grain. Dry matter. I912. ■ 127 128 129 130 131 Grams. no. 3 118. 5 141-4 124 2 112. 4 156.8 Crams. Kilos. 60.3 77-3 86. I 748 59-9 95- 0 Per cent. 548 652 609 Alfalfa Grimm S P I. 25695 (Mcdicago sativa), first crop, May 23 to July 26. . . 533 606 Mean 592±i3 127 128 129 130 131 I 132 141. 2 112. 5 129. 1 125. 0 131. 2 168.3 112. 4 97.0 100. 8 98.8 100. I 125. 8 796 862 Alfalfa, Grimm, second 781 790 763 748 crop, July 26 to Sept. 6 Mean 7go±io Oct. IS, 1914 Water Requirement of Plants 29 Table XIX. — Water requirement of different legumes at Akron, Colo., in igi2 — Contd. Pot No. Dry matter. Grain. Water. Grain. Water requirement based on — Grain. Dry matter. I9I2. 127 128 129 130 131 132 Grams. 68. 2 65.7 59-9 51- 9 48.2 60. 2 Grams. Kilos. 36.5 32-4 29. 6 26. 9 24- 4 29-3 Percent. 536 493 494 518 Alfalfa, Grimm, third crop, Sept. 6 to Nov. 506 487 Mean 506 -Hs 127 128 129 130 131 I 132 319- 7 296.7 330-4 301. I 291.8 385-3 209. 2 206. 7 216. 5 200. 5 184.4 250. I 634 697 656 666 Alfalfa, Grimm, com- bined crop. May 23 to Nov. 4 632 649 Mean 659 ±6 133 134 135 136 137 i 138 131. 6 121. 4 141. 0 140. 6 141. 4 ri8. 0 89.6 73-9 82. 4 72- 5 79.8 76.6 680 Alfalfa, Grimm A.D.I. 609 584 S16 564 649 E-23-20-52, first crop. May 24 to July 26 Mean 6oo±i7 ^3S 134 135 136 137 I 138 131- 7 no. 0 138.6 133-3 135-2 120. 2 118. 8 93- 0 112. 4 115. 8 108. I 107.4 902 845 810 Alfalfa, Grimm, second crop, July 26 to Sept. 6 870 800 893 Mean 853±i3 133 134 135 136 137 I 138 73-7 64.7 80. 7 82.0 83.8 79-5 35-4 28. I 32-4 32.0 36.1 31- I 479 434 401 390 Alfalfa, Grimm, third crop, Sept. 6 to Nov. 431 391 Mean 42 1 ±10 ^33 134 135 136 137 I 138 337-0 296. I 359-6 355- 9 360.4 317-7 243-8 •95-° 227. 2 220.3 224-4 215.1 724 658 632 Alfalfa, Grimm, com- bined crop, May 24 to Nov. 4 . 627 678 Mean 657±ii 121 122 123 124 , 125 108.2 61.5 64.4 69-5 137-0 61. 4 36.8 33-6 37- 7 68. 9 Clover, sweet, S. P. I. 2i2i6{Melilotus alba), first crop, May 23 to 568 598 522 542 503 Mean 547 ±" 30 Journal of Agricultural Research Vol. Ill, No. I Table XIX. — Water requirement of different legumes at Akron, Colo., in IQI2 — Contd. Pot No. Dry matter. Grain. Water. Grain. Water requirement based on — Grain. Dry matter. I9I2. 121 122 123 124 , 125 Grams. 155-3 130- s 103.8 143-9 149.0 Grams. Kilos. III. 6 87-5 74-7 94-5 92- 3 Per cent. 719 670 720 657 Clover sweet second. crop, July 26 to Sept. 6 620 Mean 677±I4 121 122 123 I 124 23-9 19.4 15-3 14.8 12.5 12. 0 9-3 9-5 522 6ig 608 Clover, sweet, third crop, Sept. 6 to Nov. 7 . Mean 598 ±18 121 122 123 I 124 287.4 211. 4 183.5 228.3 185.5 136.3 117. 6 141-7 645 645 641 Clover, sweet, com- bined crop, May 23 to Nov. 7 638±4 f 157 158 159 160 161 162 239-9 275-5 238.3 181. I 272. 0 227. 0 97.0 126. 3 85.0 47.2 120. I 104.5 138.0 129. I 127. I 96. 0 129.4 108.6 40 46 36 26 44 46 Chick-pea, S. P. I. 24322 {Cicer arieti- num), June 3 to Aug. 1,423 1, 022 1,495 2,032 1,077 1,039 576 469 533 530 476 478 i,348±ii4 5io±i4 A number of different legumes were included in the 191 3 water- requirement measurements at Akron (Table XX). On the basis of dry matter produced, the results obtained are as follows: Kind of legume Water requirement Cowpea 571 ±3 Soy bean 672 ±9 Navy bean 682 ±4 Peruvian alfalfa 651 it 12 Hairy vetch 6go±8 Horse bean, S. P. I. 25645 772 ± 1 1 Mexican bean 773 i8 Canada pea 775 ±5 Horse bean, S. P. I. 15429 78o± 19 Red clover 789±9 Crimson clover 805 ± 8 Wild soy bean 8i5±25 Grimm alfalfa 834 ±8 Purple vetch 935 ±9 Oct. 15, 1914 Water Requiremeni of Plants 31 Cowpea (PI. VI, fig. i) was the most efficient of the cultivated legumes. The least efficient was purple vetch. The water requirement of the first crop of hairy vetch (PL VI, fig. 2) was 9 per cent less than for purple vetch, the period of growth being the same for both varie- ties. Some of the pots of hairy vetch produced a good second growth, but none of the pots of purple vetch were able to survive the first cutting. The water requirement of the combined crops of hairy vetch was 690^8, or about three-fourths that of purple vetch. Of the soy beans (PI. VI, fig. 3) the wild required 21 per cent more water than the cultivated variety, which in turn required iS per cent more water than cowpea. (See T.able XX.) Table XX.— Water requirement of legumes at Akron, Colo., in ipij Plant and period of growth. Pot No. Dry matter. Grain. Water. Grain. Water requirement based on — Grain. Dry matter. 1913- 55 56 11 59 60 6i 62 63 64 65 66 GramS' 120. 8 no. 0 99- 0 114. 8 114.9 no. 5 98.0 I04-S 104.7 97-5 96- 5 117. 6 Grants. Kilos. 104. 1 91.7 72-3 89.8 94.0 86.9 80.8 79-5 87.1 84.8 73-7 92. 1 Per cent. 862 834 730 782 818 Alfalfa, Grimm, E-23- 20-52 {Medicago sa- liva), first crop, June 7 to July 18 786 824 760 832 870 764 783 Mean 8o4±9 55 56 57 58 59 60 61 62 63 64 65 66 115.8 106. 4 108.7 114. 8 101. 1 95-8 102. 7 104. I 114. 0 8s-7 98.9 109. 6 106. 7 95-7 80. 7 96- 3 95- I 84.4 88.9 91.9 98. 2 80.3 86.1 89. 1 922 goo 82 s 838 940 881 Alfalfa, Grimm , second crop, July 18 to Aug. 866 882 861 937 871 8X2 •Mean 878±8 55 56 57 58 59 60 61 62 63 64 65 66 89-3 82.2 73-8 76.6 81.9 79-3 85-5 76-5 88.9 66. 9 71-3 87. 6 77- S 70.4 55-5 59- I 70.4 63.0 63.3 60. 4 76.8 56-4 61.7 70.7 868 856 752 771 860 Alfalfa, Grimm, third crop, Aug. 26 to Oct. 794 740 790 864 843 866 807 Mean 8i8± 11 32 Journal of Agricultural Research Vol. III. No. I TablB XX. — Water requirement of legumes at Akron, Colo., in igij — Continued Pot No. Dry matter. Grain. Water. Grain. Water requirement based on — Grain. Dry matter. 1913- 55 56 57 58 59 60 61 62 63 64 65 . 66 GraTjis. 325-9 298.6 281.5 306.2 297-9 285.6 286.2 285. I 307.6 250. I 266. 7 314.8 Crams. Kilos. 288.3 257-8 217-5 245.2 259- S 234-3 233-° 231.8 262. I 221. s 221. s 251-9 Per cent. 886 864 773 801 871 821 Alfalfa, Grimm, com- bined crop, June 5 to 814 Oct. 23 813 853 886 831 800 8s4±8 97 98 99 100 lOI 102 41-5 7°-3 66.0 45 0 78.8 87.9 26.3 45-5 38-9 27-3 54-8 64.4 634 648 Alfalfa, Peruvian, S. P. I. 30203 (Medi- 59° 606 cago saliva), first crop, June 7 to July 69s 733 65i±i6 97 98 99 100 lOI . 102 62. 0 104. 5 91.4 69.9 88.4 102.3 42-5 73-3 63- 9 47.0 69-5 82.5 685 Alfalfa, Peruvian, sec- 699 672 ond crop, July 19 to 786 806 Mean 725±i8 97 98 99 100 lOI 102 63.6 99.2 85.0 74-5 C) 108. 9 39-9 55-5 49-9 42. 6 58.8 69-3 627 559 587 572 Alfalfa, Penivian, third crop, Aug. 26 to Oct. 636 S96±i2 97 98 99 100 :o2 167. I 274. 0 242.4 189.4 299. I 108.7 174-3 152-7 116. 9 216. 2 650 636 630 •618 bined crop, June 7 to Oct 23 723 65i±i2 103 104 105 106 107 108 99- 7 71.0 89.6 92. 6 116. 3 112. 8 73-1 56. 0 60.3 63.6 82.0 81. I 733 789 673 687 70s 719 Clover red S P I 34869 {Trifolium re- pens), first crop, June Mean 7i8±ii « Missing, Oct. 15, 1914 Water Requirement of Plants 2,i Table XX. — Water requirement of legumes at .Akron, Colo., in IQIJ — Continued Plant aiid period of growth. Pot No. Dry matter. Grain. Water. Grain. Water requirement based on — Grain. Dry matter. 1 1913- 103 104 105 106 107 , 108 Grams. 62.7 66.0 39-5 62.8 52-5 46.6 Grams. Kilos. 53-9 53-0 36-0 53-9 45-7 44-4 Per cent. 860 Clover, red, second crop, July 19 to Aug. 26 803 911 858 870 953 Mean 876±i4 i°3 104 105 106 107 , 108 13-6 18.3 9- 7 8-5 37-7 10. 4 17-0 17.4 10-7 9-0 26- 0 12.9 ij 250 952 1,103 1,058 Clover, red, third crop, Aug. 26 to 689 1,040 Mean i,oiS±49 i°3 104 loS 106 107 108 176. 0 iSS-3 138.8 163.9 206. 5 169.8 144.0 126. 4 107. 0 126. s 153-7 138.4 818 Clover, red, combined crop, June 5 to Oct. 814 771 772 745 815 Mean 789 ±9 109 no in 112 "3 114 191-3 169. 2 186.2 146.7 109.8 no. 2 160.8 131. 0 150-5 n8. I 84-9 91. 0 841 774 809 806 Clover, crimson, S. P. I. 33742 (Tri/olium in- carnalum), June 5 to Aug. 26 773 82s Mean 805 ±8 f 181 182 183 184 185 . 186 119.4 97.8 98.1 n4. 5 121. 9 120. 2 93-3 76.9 95-9 99.6 98.9 106. 9 781 786 977 870 811 Vetch, hairy, S. P. I. 34298 ( Vicia villosa). first crop, May 29 to July iS 890 Mean 853 i 23 ■ 181 184 18s . 186 n4. 8 148.4 155-7 131-7 61.7 80. I 92.6 74-7 Vetch, hairy, second crop, July 18 to Oct. 537 540 S9S 567 Mean S6o±xo 181 184 18s 186 234- 2 262. 9 277-6 251-9 15s- 0 179-7 191-5 181. 6 Vetch, hair}', com- bined crop, June 5 to Oct 22 662 684 690 722 Mean 690 ±8 60300°— 14- 34 Journal of Agricultural Research Vol. Ill, No. I Table XX. — Il'a( , Colo., in iQij — Continued Pot No. Dry matter. Grams. 207. 0 202. 6 215-7 209. 2 210. 6 202.7 Grain. Water. Grain. Water requirement based on — Grain. Dry matter. 1913- Cowpea, S. P. I. 29282 {Vigna sinensis), June 17 to Aug. 26. . . 151 152 153 154 155 156 Grams. 69.9 71.4 80. I 77-7 75- 0 79-5 Kilos. 123. I 115. 6 121. 7 116. I 120. 0 116. I Per cent. 34 35 37 36 39 1,761 1,620 I. 519 1.495 I, 600 I, 460 595 571 564 555 570 573 Mean I, 576±32 57i±3 i6g 170 171 172 173 I 174 38-3 53-7 44.6 30. 0 10. 0 17.4 31.0 38.4 33-4 21.7 8-7 14. I 809 715 749 723 870 811 Bean, horse, S. P. I. 15429 {Vicia faba), June 12 to July 19. . . Mean 78o±i9 175 176 177 178 179 180 9.4 40.9 29.4 53-6 18. 5 10. 4 7-1 34- 0 21.8 41-3 13-5 8-3 755 831 742 774 730 798 Bean, horse, S. P. I. 25645 (Vicia faba). June 17 to July 9. . . . Mean 772±ii Neither variety of horse bean did well. The growth was fairly good during the early period, but during the warm days of July the plants wilted down badly, despite an ample water supply, and had to be harvested before they had reached maturity. The water requirement, notwithstanding this, is no higher than that of many of the other legumes, and compares favorably in this respect with hairy and with purple vetch. The navy bean, although not as efficient as the cowpea and the soy bean, is more efficient than the Mexican bean, which required 13 per cent more water. The Canada field pea and the Mexican bean were equally efficient.' Crimson clover, on the basis of the combined crop, required practically the same quantity of water as red clover. Crimson clover produced only one crop and grew slowly throughout the period, although in total pro- duction it was practically the equal of red clover. The water require- ment of red clover is slightly below that of Grimm alfalfa, while Peruvian alfalfa required only 78 per cent as much water as Grimm for the pro- ^ Peas and beans were included by Lawcs (1850, p. r,.\) in liis experiments at Rothamstead. HuKland. His measurancnts show beans lo be slightly more efficient than peas. No other measurements of pc.is and beans have been made, so far as the writers are aware. 36 Journal of Agricultural Research Vol. Ill, No. duction of a pound of dry matter. The total dry matter produced by Peruvian was, however, much less. The water requirement for each of the several cuttings made of these crops is shown in Table XXI. Table XXI. — Water requirement of different cuttings of legumes Crop. First cutting. Second cutting. Third cutting. Combined cutting. Alfalfa, Peruvian Clover, red Alfalfa, Grimm. .. Vetch, hairj' 651 ±i6 7i8±ii 804 ±9 853 ±23 725±i8 876±i4 878±8 56o±io 596±i2 i,oi5±49 8i8±ii 65i±i2 789 ±9 834±8 690 ±8 Taking the water requirement of the first cutting in each instance as a basis of comparison, the second crop of Grimm alfalfa required 8 per cent more water, Peruvian alfalfa 1 1 per cent more, red clover 22 per cent more, while hairy vetch required 34 per cent less. On the same basis the third cutting of Grimm alfalfa required 2 per cent more water than the first, 8 per cent more for Peruvian, and red clover 41 per cent more. Comparing the combined cuttings, Peruvian alfalfa and hairy vetch required 18 per cent less water than Grimm, and red clover 5 per cent less. Grimm alfalfa was the only one of the lupines grown at Akron during both 1912 and 1913. Its water requirement in 1912 was 659J: 16; in 1 91 3, 834 ±8, an increase of 27 percent. ALFALFA, SUDAN GRASS, AND MILLET GROWN DURING THE LATE SUMMER AND AUTUMN AT AKRON, COLO., IN 1912 In Table XXII are given the results of water- requirement measure- ments based on the total dry matter produced by crops grown during the late summer and autumn and planted in pots which had already produced a crop earlier in the season. The soil was not changed, and no additional fertilizer was added. These forage crops can not be said to have been grown out of season, except in that the plantings were made too late in the season to permit the plants to reach their full development before harvesting. Table XXII. — Water requirement of alfalfa, Sudan grass, and millet grown during the late summer and autumn at Akron, Colo., in Igi2, without additional fertilizer. Crop. Pot No. Drj matter. Water. Water re- quirenient based ou dry matter. Alfalfa, Gnmm. Aug. 7 to Nov. 1912. E-23-20-52 (Medicago sativa), 6, following Kharkov wheat .... 37 39 40 41 , 42 Grams. 40.9 43- I 33-2 32.0 34-3 Kilos. 40.4 37-3 31-7 30.2 34-9 988 866 955 944 I, 017 954±i6 Oct. IS, 1914 Water Requirement of Plants 37 Table XXII. — Water requirement of alfalfa, Sudan grass, and millet grown during the late summer and autumn at .ikron, Colo., in igi2, without additional fertilizer^ Continued Crop. Pot No. Dry matter. Water. Water re- quirement based on dry matter. Alfalfa, Grimm, following Tiu-key wheat 43 44 45 46 47 48 Grams. 40.4 45-3 21. 0 25. 6 30-7 33- 0 Kilos. 34-8 40.5 21.3 23.0 26. 9 3°-7 862 894 I. 013 899 876 930 Mean 9i2±i5 31 32 33 34 35 . 36 14-7 16.8 18. I 19. 6 13-8 12.9 7-5 8.5 8-3 8. I 6. I 6.6 Alfalfa, Grimm, Sept. 3 to Nov. 7, following Bluestem wheat 5'° 506 458 413 442 5" Mean 473 ±13 7 8 9 10 12 14-7 19.8 15.0 16. 4 16.7 9-3 10.8 9. I 9-5 9-5 Alfalfa, Grimm, following unfertilized Kubanka wheat 628 545 606 579 569 Mean 585±" [ 163 I 164 12.9 II. 8 7-5 6.4 Alfalfa, Grimm, Sept. 9 to Nov. 4, following 58X 542 Mean S64±i6 n 10. 0 10.8 16.5 4.6 5- I 8-3 Alfalfa, yellow flowered (Medicagofalcata), Sept. 3 to Nov. 7, following Kubanka wheat 460 472 503 Mean 478±io 55 S6 11 59 [ 60 "5 18. I 10. s 19- S 16.6 9.6 30 5-4 3- I 5-8 S-7 3-4 Sudan grass, S. P. I. 25017 (.Andropogon sorghum aethiopicus), Sept. 3 to Oct. i, following spring Ghirka wheat 261 298 295 297 343 354 Mean 308 ±10 97 98 99 101 102 20. 0 25.0 19- S 15- S 20. 0 4-4 5-8 3-7 4.0 4.6 Proso, Black Voronezh, S. D. 331 (Panicum miliaceum), Aug. 22 to Sept. 28, following Beldi barley 220 232 igo 258 230 Mean 2->6±7, 1 < r 38 Journal of Agricultural Research Vol. III. No. I Table XXII. — Water requircnicnl of alfalfa, Sudan grass, and millet grown during the late sum7ncr aiid autumn at Akron, Colo., in igi2, -without additional fertilizer — Continued Crop. Pot No. Dry matter. Water. Water re- quirement based on dry matter. Millet, Turkestan, S. P. I. 20694 (Chaetochloa italica) , Aug. 22 toOct. i, following Wliite Hull- ■ 103 104 105 106 107 . 108 Grams. 27.0 28.3 30-7 3°-4 33-5 35- I Kilos. 7.8 8-5 8.4 8.9 9.2 II. 7 289 300 273 293 274 333 Mean 294±6 116 "7 118 119 , 120 9-3 12.4 19-3 24.2 23- S 12.3 i.S 2-5 4.0 3-9 31 1.8 Millet, Kursk, S. P. 1. 30029 {Chaetochloa italica), Sent. ■? to Oct. t.. following sorine rve 194 202 207 161 132 146 i73±io 61 62 63 64 65 66 22. 2 18.4 23-4 21. 6 24.7 24-5 3-7 2.9 ^■l 3-8 3-7 3-8 Millet, Kursk, S. P. I. 34771, .Sept. 3 to Oct. i. 167 158 158 176 150 i6i±3 The effect of fertilizer added to the previous crop on the water require- ment of the next crop is shown in the following measurements : Water Crop requirement Alfalfa, following fertilized Bluestem wheat 473 ±13 Alfalfa, following unfertilized Kubanka wheat S85±ii Alfalfa, following fertilized Grindelia, following imfertilized Bluestem wheat 564±i6 The water requirement of alfalfa following unfertilized Kubanka wheat was higher than that following the fertilized Bluestem, and although this difference is not very marked (19 ±3 per cent) when the probable error is considered, it is sufficient to show a slight effect of two consecu- tive crops without fertilizer in increasing the water requirement. Two pots of alfalfa following Grhidelia squarrosa had a water requirement equal to that of the unfertilized set. Grindelia was started in pots which had already produced a crop of Marvel Bluestem wheat in 191 1 without fertilizer. Therefore alfalfa was the third crop from the same soil mass, and although the pots growing Grindelia were fertilized, the succeeding alfalfa crop gave a water requirement in accord with that following unfertilized wheat. Oct. 15, 1914 Water Requirement of Plants 39 The effect of time of planting is shown in the following determinations of the water requirement of alfalfa : Water Crop requirement Alfalfa, grown August 7 to November 6, following Kharkov wheat 954rt 16 Alfalfa, grown August 7 to November 6, following Turkey wheat. 912 ±15 Alfalfa, grown September 3 to November 7, following Bluestem wheat 473±i3 Alfalfa planted during the dry, hot days of August required almost twice as much water for the production of a unit of dry matter as it did when planted in September. The seven varieties and strains given in Table XXIII were included in these late-season experiments and were grown under comparable condi- tions. The water requirement of Grimm alfalfa, Sudan grass, and Kursk millet was also determined in midsummer (see column 3), so that it is pos- sible to reduce the water-requirement measurement of the other late-season crops to a midsummer basis. The seasonal water requirement of Grimm alfalfa was 39 per cent higher than that of the late-season crop. Assum- ing this ratio to hold for the yellow-flowered alfalfa, the seasonal water requirement of the latter would be 865±i8. The midsummer water requirement of Sudan grass and of Kursk millet (S. P. I. 34771) was in each instance 16 per cent above the late-season crop, and this ratio has been used in computing the other millets to a midsummer basis. The computed values (o) are given in the last column of Table XXIII. Table XXIII. — Water requirement of late-season crops Variety. Water requirement. Late-season crop. Midsummer crop. Alfalfa, yellow-flowered 478±IO 473 ±13 3o8±io 294 ±6 226±7 i73±io i6i±3 a 865 ±18 657±ii 359±2 ''34ii:7 0 262±8 " 201 ±12 i87±2 Sudan grass Millet, Turkestan ProsOi Black Vorotiezh Millet, Kursk, S. P. I. 30029 Millet, Kursk, S. P. I. 34771 o Computed. Of these seven varieties the }ellow-llowered alfalfa (Medkaqo falcata) gave a water requirement in practical accord with the Grimm selection grown during the same period. Kursk millet (S. P. I. 34771) gave the lowest water requirement and proved to be decidedly more cfllcient than Black Voronezh proso and Turkestan millet. Sudan grass required 91 per cent more water than the Kursk millet, which is in exact accord with the results obtained for these two crops from determinations based on a whole season's growth. 40 Journal of Agricultural Research Vol. III. No. I CUCURBITS On account of the large space required by crops which produce vines, the cucurbits were grown outside the inclosure at Akron in 1913. The reduction in water requirement produced by the inclosure amounted in the case of wheat, alfalfa, and cocklebur to approximately 20 per cent. These ratios for the cucurbits should therefore be reduced by this amount in comparing them with crops grown inside the shelter. The observ^ed water requirement outside and the computed water requirement within the inclosure, both based on dry matter, follow: Crop Outside Inside Watermelon 75o±i9 6oo±i5 Cantaloupe 778±34 621 ±27 Cucumber 89i±i4 7i3±ii Squash.... 936±io 748±8 Pumpkin i,043±2i 834±i7 The cucumber, cantaloupe (PI. VI, fig. 4), and watermelon did well in the pots. Squash and pumpkin produced very little fruit (Table XXIV), and the growth of vines was not normal. Watermelon and muskmelon proved to be the most efficient of the cucurbits. Pumpkin, the highest of the cucurbits in water requirement, is about the equal of alfalfa in efficiency. Table XXIV. — Water requirement of cucurbits at Akron, Colo., in igij Pol Dry Dry matter in fruit. Water. Fruit. Water requirement based on^ Fruit. Dry matter. I9I3- Squash, Hubbard (Cu- curbila maxim a ) , June 3 to Sept. 13. . . 337 338 339 340 341 342 Grams. 272. 0 280. 9 244-3 247-3 267.9 258.9 Grams. 46. I 85.6 33-4 18.4 68.6 7-3 Kilos. 260. 7 252. I 224.2 242.9 244.9 244-7 Per cent. 17 30 14 7 26 3 5,658 2,945 6,715 3,570 959 898 918 982 914 946 4, 720 ±690 936 ±10 343 344 345 346 347 I 348 205.9 222. 4 228.8 225. 4 190-5 179.7 5-9 45-9 2.8 2.8 215. 6 221. 0 211. 6 243-2 215-2 194.6 1,047 994 925 1,078 Pumpkin, common 3 20 I field (Cucur bita pepo), June 3 to Sept. I-? 1,130 1,083 2 Mean i,o43±2i 319 320 321 322 323 I 324 171. I 185.6 185-7 182.5 175-8 174.8 loi. 5 100. 0 96-3 112. 4 83-7 107.4 154-8 167.6 158.2 152-9 171. I 152-3 59 54 42 62 48 61 Cucumber, Boston pickling (Cucumis sativus), June 14 to Sept. I 1,525 1,676 1,642 1,360 2,045 1,4^7 905 903 852 838 974 872 I, 6ii±67 8gi±i4 Oct. 15, 1914 Water Requirement of Plants 41 Table XXIV.— Water requirement of cucurbits at Akron, Colo . , in igi3 — Continued Pot No. Dry malter. Dry matter in fruit. Water. Fruit. Water requirement based on — Fruit. Dry matter. 1913- Cantaloupe, Rocky Ford (Cucumismelo), June 14 to Sept. 13 . . 325 326 327 328 329 330 Grams. 314.7 285.3 211. 4 264.9 272.8 305-6 Grams. 198.9 155-2 92-5 132-5 160. 6 162. 4 Kilos. 201. 7 214-4 210. S 218. I 197-3 222.8 Percent. 63 ss 44 50 59 53 1,014 1,382 2, 276 1,646 1,228 1,371 641 752 996 824 723 729 Mean I,486±i20 778 ±-,4 331 332 333 334 335 1336 301.8 318. 5 334.5 314-2 267.9 320.8 193.2 216.8 225.9 2og. I 159-4 224. 2 238.2 215- S 241.4 232.8 232.1 226. 5 64 68 67 66 59 70 Watermelon, Rocky Ford {Citrullus vul- garis), June 14 to Sept. i^ 1,233 995 1,069 1, 112 1,457 I, 010 790 677 722 740 866 706 Mean i,i46±49 75o±i9 On the basis of the production of fruit the watermelon has proven to be exceptionally efficient. The water requirement, calculated on the basis of the dry matter in the melons and reduced to inclosure conditions, was 915 ±39. The green fruit contained 95 per cent of water. The water requirement on a green basis would therefore be 46. RAPE, TURNIP, CABB.^GE, AND POTATO Rape, turnip, cabbage, and two varieties of potato, the Irish Cobbler and the McCormick (PI. VI, fig. 6), were included in the 191 3 meas- urements (Table XXV). The water requirement, based on dry matter, was as follows : Crop Water requirement Cabbage 539±7 Turnip 639±3i Potato: Irish Cobbler 659±i5 McCormick 7i7±ii Rape 743±7 Cabbage and turnip are seen to have a lower water requirement than potato and to rank in efficiency with oats. Of the potato varieties the Irish Cobbler was the more efficient and produced the most tubers. The McCormick, a late-maturing variety, produced fine vines but practically no tubers. The water requirement of rape was practically the same as that of turnip during the same period of growth, but the second crop, although not a heavy one, had a water requirement so much higher than the first that the combined crop is approximately 16 per cent higher than for turnip. 42 Journal of Agricultural Research Vol. Ill, No- I Tabi,E XXV. — Water requirement of rape, turnip, cabbage, and potato at Akron, Colo., in igij Plant and period of growth. Tot No. Dry matter. Tubers or roots. Water. Tubers or roots. Water requirement based on — Tubers or roots. Dry matter. 1913- 217 218 219 220 221 222 Grains. 171- 2 156-3 162- 0 140-5 171-5 164. 6 Grams. Kilos. 112- I 102. 9 102. 5 97-4 108.0 no. 5 Per cent. 65s 658 633 693 630 671 Rape (Brassica napus) , June 3 to July ig-- . . Mean 657±6 217 218 219 220 221 222 44.6 68.5 58.3 63.8 66.4 69. 0 54-3 69. 2 58.0 53-2 62.8 60. 9 I, 217 Rape, second crop, July 19 to Sept. 13. . 995 834 946 882 Mean 98i±3S f 217 218 219 220 221 222 215.8 224. 8 220.3 204.3 237-9 233-6 166- 4 172. I 160. 5 150. 6 170.8 171-4 771 767 729 737 718 734 Rape, combined crop, June 3 to Sept. 13 . . . Mean 743 ±7 Turnip, Purple Top {Brassica rapa). May 29 to July 19 211 212 213 214 215 216 128.9 128.4 58.6 128. 5 73-8 90.4 69-3 64.6 17-3 58.8 37-6 29.4 78.1 90. 7 30-6 69- 5 51-2 68-9 54 5° 29 45 51 32 1,127 1,403 1,767 i,iSi 1,360 2,341 606 706 522 541 694 762 i,53°±i32 639±3I 205 206 207 208 209 210 284.5 281.9 359-8 33°- 7 344- 6 313-2 148.9 160.3 191- 5 176. 0 175-7 177.0 523 569 533 532 510 56s Cabbage , Early Jersey Wakefield (Brassica oleracca capitata), June 3 to Sept. 12 . . . 539 ±7 Potato, Irish Cobbler (Solanum tuber- osum), June 5 to Sept. 12 "5 116 "7 118 119 120 170. 2 203. 2 211. 0 222.3 2l8. I 201. I 40. 6 67.8 81.7 86.6 105-3 52-4 120. 4 145- I 127. 0 149-3 130.0 ^33-3 24 33 39 39 48 26 2, 966 2,140 1,555 1,724 1,234 2,542 708 714 602 671 596 662 2,o27±i97 659±iS 121 122 123 124 125 126 215.8 213- 7 215-5 219-3 201. 4 206. 9 . I 1-7 5-° 10.3 •3 1-9 145- S 154-9 146.9 164. 5 154.2 145-9 674 724 682 (Solanum tuber- osum), June 5 to Oct 4 2 5 75° 766 706 I Mean 1 7i7±ii 1 ' Oct. 15, 1914 Water Requirement of Plants 43 N.\TIVE AND INTRODUCED GRASSES AND OTHER NATIVE PLANTS Two native Colorado plants were included in the 191 2 measurements at Akron, Grindelia squarrosa, or "gum weed," and Artemisia frigida^ or "mountain sage" (PI. II, fig. 3). These plants were carried through the winter in the pots, and only two pots of each set were in good condi- tion in the spring. They both behaved as biennials, forming rosettes in 1 911 and flowering profusely in 191 2. The data (Table XXVI) given for the period from May 20 to August 26 include much of the stored dry matter of the rosettes and root systems elaborated during an earlier period, and consequently the water requirement is somewhat too low. In order to check this, the data based on the total period of growth, which includes the growth and water consumption in 191 1, have also been given. This method of computation increases the water require- ment less than 6 per cent. Table XXVI. — Water requirement of native plants at Akron, Colo., in igi2 Plant and period of growth. Pot No. Dry matter. Water. Water re- quirement based on dry mat- ter. I912. Grindelia squarrosa, May 20 to Aug. 26 / 163 I 164 Grams, 385-4 367-7 Kilos. 172. 0 180. 0 446 490 Mean 468 ±18 / 167 \ 168 336-3 293-9 Artemisia frigida. May 20 to Aug. 26 153-4 143-9 456 491 Mean 474±I4 / 163 I 164 399-6 381.2 182.4 200. 5 TOTAL PERIOD OP GROWTH. Grindelia squarrosa, Aug. 18, 1911, to Aug. 26, 457 526 Mean 492 ±29 / 167 \ 168 372.8 345-8 177-4 183.9 Artemisia frigida, Jan. 10, 191 1, to Aug. 26, 1912 ..... 476 532 Mean 504 ±24 Although these are typical native plants of the high plains, they re- quired about 20 per cent more water than Kubanka wheat and rank higher in water consumption than any of the cultivated grains except rye and rice. Grasses produce so slowly that it is somewhat difficult to make satis- factory measurements of their water requirement. The 1913 experi- ments (Table XXVH) included pure buffalo grass, mixed grama and 44 Journal of Agricultural Research Vol. Ill, No. I buffalo grass, western wheat-grass, brome-grass, and a Siberian wheat- grass. Buffalo grass, brome-grass, and wheat-grass each gave two crops, buffalo and grama mixed gave three cuttings, while western wheat- grass produced but one cutting. The combined crops afford the best basis for the comparison of these grasses. On the basis of the total dry matter produced throughout the season, the water requirement is as follows : Variety of grass Water requirement Buffalo 3o8±22 Grama and buffalo 3891b 12 Wheat-grass 7°S±27 Brome-grass i,oi6±26 Western wlieat-grass i, 076 ±29 Brome-grass and western wheat-grass are comparatively very ineffi- cient in the use of water, requiring from 22 to 29 per cent more water than alfalfa. Western wheat-grass made a slow growth. Wheat-grass is more efficient than alfalfa, requiring 15 per cent less water. The short grasses made a wonderful showing, the water requirement of pure buffalo grass being only about one-third that of alfalfa, and that of grama and buffalo mixed, 47 per cent. The water requirement of buffalo grass is only 8 per cent above millet, which is one of the most efficient of the introduced plants. The mixed buffalo and grama grass required 36 per cent more water than Kursk millet. These are the first of the native plants to show any marked efficiency in the use of water, although some of the weeds, as will be seen later, are also highly efficient. Table XXVII. — Water rcquiremeiil of native and introduced grasses at Akron, Colo.. in 1913 Plant and period of growth. Pot No. Dry matter. Water re- quirement based on dry matter. I9'.v Grama (Bouteloua gracilis) and buffalo grass (Bulbilis dactyloides), mixed, first crop, Jxme 3 to July 19 139 140 141 142 143 144 Grams. 50- I 26. 9 31.8 21. 9 22. I 25- 3 Mean. Grama and buffalo grass, mixed, second crop, July 19 to Aug. 26 139 140 141 142 143 L 144 Mean. 5 1- 7 35-4 38.8 37- Q 41. 2 26. I Kilos. 17.4 10. 7 12.4 7-5 Q.8 8.6 13- 2 15.6 14.4 16. I 9-7 347 398 390 342 443 340 377±ii 350 373 402 380 391 371 395 ± 7 Oct. 15, 1914 Water Requirement of Plants 45 Table XXVII. — Water requirement of native and introduced grasses at Akron, Colo., in igij — Continued Plant and period of tiro\vtb. Pot No. Dry matter. Water. Water re- quirement based on dry matter. 1913- Grama and bufialo grass, mixed, third crop. ■ 139 140 141 142 143 L 144 Grams. 14-3 II. 6 II. 8 IS- 9 15- 0 18.3 Kilos. 7-4 6.1 7.0 7-1 7.6 4-3 517 526 593 446 S°7 235 Mean 47i±26 ' 139 140 141 142 143 I 144 116. I 73-9 82.4 75-7 78.3 69.7 42-9 30. 0 35-° 29. 0 33-5 22.6 Grama and buffalo grass, mixed, combined crops, June 3 to Oct. 20 369 406 425 383 428 324 Mean 389±i2 ■ 145 146 147 148 149 I 150 10. 0 32-1 18.9 12.9 13- 2 8.1 3-5 g. 0 5-0 3-8 3-8 1-5 Buffalo grass (Bulbilis, dactyloides) , first crop, Jtme 18 to Aug. 26 350 280 264 295 288 185 Mean 277±i3 145 146 147 I 148 3-6 3-1 2-3 5-2 I- 5 .8 - 7 1-9 Buffalo grass, second crop, Aug. 26 to Oct. 18 . ■ Mean 417 258 304 365 336±23 ■ 145 146 147 . 148 13.6 35-2 21. 2 18. I 5-0 9.8 5-7 5-7 1 Buffalo-grass, combined crops, June'i8 to Oct. 18 . 368 278 269 315 Mean 3o8±i7 133 134 135 136 I 138 64-5 75.2 76. 0 60.8 79-5 62.8 74-6 70-3 58.6 79-3 Brome-grass. S. P. I. 29880 (Bronius inermis), first crop, May 2 3 to July 19 973 992 925 964 998 Mean 97o±9 K^3 134 13s 136 I 138 3°- 3 46. I 36.. 26. 9 40. I 28.4 59- 3 46.6 21.8 47-0 Brome-grass, second crop, July 19 to Oct. 22. . . . 937 1,286 I, 290 810 1,172 Mean i,099±76 46 Journal of Agricultural Research Vol. HI, No. I Table XXVII. — Water requirement of native and introduced grasses at Akron, Colo. in IQIJ — Continued Plant and period of growth. Pot No. Dry matter. Water. Water re- quirement based on dry matter. 1913- Brome-grass, combined crops, May 23 to Oct. 22 . ■; ■ 133 134 135 136 I 138 Grams. 94.8 121. 3 112. I 87.7 119. 6 Kilos. 91.2 133- 9 116. 9 80.4 126.3 962 I, 104 1,043 916 I, 056 Mean . I, oi6±26 ( 67 68 69 70 71 I 72 25-9 33-6 24-3 23.0 17. I 34-0 19. 2 22.4 16. I 19- 3 13.8 25-8 WheQ.t'gTQSs(Agropyroncrisiatu7n), S. P. 1. 19537. 741 667 663 839 808 759 746±2i f 67 68 69 70 71 72 49.0 77.0 53-6 58.0 46. I 66.2 37-9 45- I 29-3 40. 2 34.8 51-3 Wheat-grass, second crop, July 19 to Oct. 22 773 585 547 693 755 775 688±3i 67 68 69 70 71 I 72 74-9 no. 6 77-9 81.0 63.2 100. 2 .=;7- 1 67.5 45-4 59-.'; 48.6 77.1 " Wheat-grass, combined crop, Jtine 5 to Oct. 22 . . 762 610 583 735 769 769 7os±27 235 236 237 238 239 I 240 37-7 41.4 18.3 21. I 25. 6 16.7 45-° 43-1 15-2 23.0 28.9 19.7 Wlieat-grass, western {A gro p yro n SinWiii), June I-I93 I, 040 831 1,087 1,128 I. 179 i,o76±29 WEEDS A number of weeds were grown in pots at Akron in 191 3, in order to determine their water requirement. (Table XX\'III.) Most of these were planted late in the season, after the crops of grain had been removed. Pigweed and the annual sunflower were, however, started at the beginning of the season. Three crops of pigweed were produced, the water require- ment for the first, second, third, and combined cuttings, based on dry matter, being 32.s±io, 326±4, 278±7, and 320±7, respectively. Oct. IS, 1914 Water Requirement of Plants 47 Sunflower, which made an excellent growth, gave a water requirement of 705 ±8. Sunflower thus requires almost three times as much water as pigweed and 86 per cent as much water as alfalfa. A comparison of the water requirement of pigweed during the three periods of growth will show that the water requirement is not greatly affected by the period of growth. If this holds for the other weeds, no great error will be produced in comparing the water requirement of these plants without regard to the period during which they were grown. The water requirement is, how- ever, probably slightly less than it would have been if the plants had been grown in midsummer. The results obtained, based on the produc- tion of dry matter, are as follows: Variety of weed Water requirement Purslane (Portulaca oleracea) 292 ±ii Pigweed (Amaranthus retroflexus) 320it7 Cocklebur (Xanthium commune) 432 ± 13 Narrow-leaved sunflower from sand hills {Helianlhus pciiolaris) . . 570±ii Annual sunflower (Helianihus annuus) 705±8 Narrow-leaved sunflower from near Akron (Helianlhus peiiolaris). 774±20 Lamb's-quarters {Chenopodium album) 801 ±41 Fetid marigold (Boebera papposa) 881 ±26 Western ragweed (Ambrosia ariemisifolia) 948±66 Purslane and pigweed, two introduced weeds, appear to be excep- tionally efficient plants, their water requirement being only slightly higher than that of Kursk millet and in practical agreement with the sorghums. vSome of the indigenous weeds were also found to be fairly efficient, cocklebur, a plant found in stream beds and about ponds, having a water requirement 13 per cent less than wheat, while the narrow- leaved sunflower from the sand hills had a water requirement 3 1 per cent less than alfalfa. Lamb's-quarters, an introduced plant, and fetid marigold (PI. VII, fig. 2) and western ragweed, indigenous plants, have a slightly higher water requirement than alfalfa. It is evident, therefore, that the common weeds differ greatly in water requirement. A growth of weeds in a crop or on summer fallow repre- sents a tremendous loss of moisture, a thousand pounds per acre of the most efficient weeds representing a loss of at least 1.5 inches of stored rainfall, or from 4 to 5 inches of stored rainfall in the case of the weeds having a high water requirement. The latter figures represent about the maximum amount of moisture that can be stored in fallow land. It is therefore easy to understand how the whole of the stored n)oisture supply may be lost through the growth of a moderate crop of weeds, and these varieties having a high water requirement are especially to be dreaded. 48 Journal of Agricultural Research Vol. Ill, No. I Table XXVIII. — Water requirertient of weeds at Akron, Colo., in igij Pot No. Dry matter. Grain. Water. Grain. Water requirement based on — Grain. Dry matter. 1913- Sunflower {Helianthus annuus), June 5 to Sept I 5 271 272 273 274 275 I 276 Grams. 516.6 554- 0 502. 2 459- 2 512.5 507-2 Grams. 40. 7 70.5 51.8 58.4 63. 0 63.2 Kilos. 373-5 363-9 343-2 334-9 371-6 360-9 Per cent 8 13 10 13 12 12 9, 180 5, i6i 6, 625 5.725 5.900 5.710 6, 384 ±383 723 657 683 729 724 7" Mean . 705i8 175 176 177 178 179 I 180 196. 0 140.4 148.7 103.4 156.0 163-7 34-1 19. 6 23-3 16.7 21.8 25.2 135-5 103. I 109- 5 87.1 132.2 129. 6 17 21 16 16 14 15 Sunflower, narrow- leaved {Helianthus peliolaris), July 25 3.971 5. 261 4.69s 5.218 6,061 5. 141 691 735 736 842 848 792 5,o58±i32 774±20 79 80 81 82 83 I 84 17. I 32.3 34-6 14.4 25-4 30.5 10. 0 18.6 18.2 9-1 '3-5 17-3 585 576 526 632 531 Sunflower, narrow- leaved (Helianthus petiolaris), from Sand hills, Aug. 13 to Sept. 18 567 57o±ii 169 170 171 172 173 . 174 116. 0 102.8 106. 5 99-9 167-3 55-0 95-8 98.8 93-3 86.7 125-5 55-2 826 960 876 868 Marigold, fetid (Boe- bera papposa), July 25 to Sept. 17 750 1,004 Mean. . . . 881 ±26 211 212 213 214 215 , 216 64. 0 35-3 47-4 40. 0 51-6 64. 2 30-9 26. 9 46. 2 33-4 47-2 44-6 624 762 975 835 914 694 Lamb 's-quarters (Chcn- opodium allium), July 25 to Sept. 17. . Mean 8oi±4i 187 188 189 190 191 I 192 17. I 12-3 10. 6 20. 6 13- I 27-5 12-5 12.9 13-8 18.7 13-3 18.7 Ragweed, western {Ambrosia ariemisi- folia), Aug. I to Sept. 18 . 1,049 1.302 908 I. 015 680 948 ±66 Oct. :5. 1914 Water Requirement of Plants 49 Table XXVIII. — Water requirement of weeds at Akron, Colo., in 1913 — Continued Plant and period of growth. Pot No. Dry matter. Grain. Water. Grain. Water requirement based on — Grain. Dry matter. 1913- 91 92 93 94 95 I 96 Grams. 56-9 133-7 145-7 115. 0 112. 8 161. 5 Grams. Kilos. 18.8 50.2 44-6 33-7 33-5 55-7 Per cent. 330 377 306 293 Pigweed (Amaranihus retroflexus), June 12 297 34S 325±io 91 92 93 94 95 1 96 74-9 62. 2 50. 0 54-9 60. 9 50. 6 22.5 20. I 16. 4 17.9 20.5 17.2 300 323 328 326 337 340 Pigweed, second crop, July 19 to .Sept. I Mean 326±4 91 92 93 94 95 I 96 7-5 II. 6 II. 8 16.7 14.9 18.6 2. 2 3-6 3-4 4.2 3-9 4-9 293 310 288 Pigweed, third crop, Sept. I to Oct. 4 . , . . 251 262 263 Mean 278±7 91 92 93 94 95 96 139-3 207-5 207.5 186.6 188.6 230.7 43-5 73-9 64.4 55-8 57-9 77-8 312 356 310 299 Pigweed, combined crops, June 12 to i 307 337 Mean 32o±7 183 241 242 243 244 245 246 103. I 36.2 28.0 33-5 38-4 48.5 18.8 33-2 8.9 7-5 9-7 10. 9 12.9 7.0 Purslane {Portulaca okracea), July 18 to Sept. 17 322 246 268 Purslane, Aug. 21 to 2S9 284 266 Sept. 17 372 Mean 292±II RELATIVE WATER-REQUIREMENT MEASUREMENTS The relative water requirement of the different varieties of plants grown at Akron in 191 2 is summarized in Table XXIX, Kubanka wheat being used as a basis of comparison. Grimm alfalfa is seen to have the highest water requirement of the 42 species and varieties tested during 191 2, while Kursk millet proved the most efficient of all the plants tested. • The varieties arc also grouped on the basis of crop or genus, and their mean water requirement compared with the mean water rcquire- 60300°— 14 4 50 Journal of Agricultural Research Vol. m. No. I ment of the wheat varieties. The relative water requirement on this basis is given in the last column of Table XXIX. The quantity of water required by the various crops for the production of a unit amount of seed or grain at Akron in 191 2 is summarized in Table XXIV. Rye is seen to be the least efficient of the grain-producing crops tested, with Voronezh proso the most efficient. Table XXIX. — Summary of iiaUr-requirement lyieasurcnunts at Akron, Colo., in igi2 WATER REQUIREMENT BASED ON DRY MATTER PRODUCED Crop and variety. Botanical name. Water requirement ol— Variety. Relative, compared with Kubanka wheat. Crop (or genus). Relative, compared with wheat (Trttiatm spp.). I912. Alfalfa: Grimm, S. 25695- Grimm, A. D E. 23-20-52. Clover, sweet Rice, Honduras. . Chick-pea Rye .spring Cotton, Triumph. . P. I. Native plants Oats: Sixty-Day Burt Swedish Select. .. Canadian Barley: Hannchen White HuU-less.. Beldi Beardless Wheat: Spring Ghirka . . . Marvel Bluestem. Emraer Kubanka Kharkov Turkey Sugar beet : Kleinwanzleben . Com: China Wliite Iowa Silvermine. Laguna China White XLa- guna. Hopi Northwest em Dent. China WhiteXEs- peranza. Esperanza Medicago sativa . do.. Melilotus alba Oryza sativa Cicer arietinum ... Secale cereale Gossypium hirsu- tum. /Artemisia frigida . . \Grindelia squarrosa Avena sativa . ....do ....do ....do Hordeum distichon Hordeum vulgare . , ....do ....do Triticum aestivum. ....do..... Triticum dicoccum Triticum durum. . . Triticum aestivum . ....do Beta vulgaris . Zea mays. ....do.... ....do.... ....do.... .do. .do. .do. .do. 659±6 657±ii 638 ±4 5i9±i3 Sio±i4 496 ±9 488±i4 474±i4 468±i8 491 ±13 449 ±3 423±5 399 ±6 443 ±3 439 ± I 4i6±4 403 ±8 4S7±3 45i±4 428 ±3 394 ±7 365±6 364±6 32i±8 3i5±7 302 ± 7 295±6 289±4 285±7 28o±IO 250±2 239±3 67 ±0. o, 67 ± .04 62 ± 32± 29 ± 26± 24± 20± . 04 i9± .05 2.S± . 04I I4± . 02 07 ± . 02 01 ± . 02 . I2± .02 . II± .02 . o6± .02 . 02± .03 . i6ih . 02 . I4± . 02 . 09± .02 . 00 ■ 93± -02 . 92 ± .02 .8i± .03 . 8o± .02 . 77± .02 ■ 75 ± -02 . 73± .02 . 72± .02 . 7i± .03 .63± .01 . 6i± .01 638 638 519 510 496 471 441 425 410 286 I. 60 1.56 1.27 24 21 19 I- 15 1.08 I. 04 78 70 Oct. IS, 1914 Water Requirement of Plants 51 Table XXIX. — Summary of waler-requiremeni measurements at Akron, Colo., in igi2 — Continued WATER REQUIREMENT BASED ON DRY MATTER PRODUCED — Continued Botauical name. Water requirement of — Variety. Crop (or genus). Crop and variety. .\ctual. Relative, compared with Ktibanka wheat. Actual. Relative. compared with wheat (TrifjiMm spp.). 1912. Sorghum: .Sudan grass Milo, Dwarf Kafir, Blackhull Andropogon sor- ghum aethiopicus. Andropogon sorghum do ... 359±- 273±4 259±5 255±3 249 ±3 239±2 237±4 223±I 248±7 i87±2 208 ± I 206 ± I 0. 91 ±0. 02 . 69± . 02 .66± .02 .65± .02 .63± .02 . 6i± . 01 . 6o± . 01 •57± -0' . 63 ± .02 .47± -OI •53± -OI . 52± .CI 262 218 207 Durra, White do 0. 64 Milo .do Minnesota Amber. .. ..do Red Amber do Kaoliang, Brown. . Millet: do Chaetochloa italica . do ■ 53 Kursk • 51 Proso: Panicum miliaceum do Voronezh ■ 51 WATER REQUIREMENT BASED ON GRAIN OR SEED PRODUCED Rye, spring Chick-pea Oats: Canadian Burt Sixty-Day Swedish Select . Wheat: Marvel Bluestem. . Spring Ghirka. .. . Kubanka Kharkov Turkey Emmer Barlev: White Hull-less.. Beardless Hannchen Beldi Sorghum: Kaoliang, Brown. Minnesota Amber , Millet: Kursk Proso: Tambov Voronezh Secale cereale . . . Cicer arietinum. Avena sativa . ....do ....do ....do Triticiim acstivum. . do Triticum dunim. . . . Triticum acstivum. . do Triticum dicoccum . Hordeum vulgare . . . do Hordeum distichon Hordeum vulgare . Andropogon sor- ghum. do Chaetochloa italica . Panicum miliaceum do I, 802 ±62 i,348±ii4 I, 4i6±ii9 I, 224±55 I,i72±i33 I, io3±i8 i,S73±49 i,468±34 I, iii±37 I, 064 ±60 995±22 984±i8 i,23g±ii i,oi7±83 I, 005 ±36 94i±io 927*38 6o7±is 483±ii 482 ±9 425±4 I. 27db . 12 io± . 07 I. 05± . 12 . 9g± .04 I. 6i±Oi oS I. 2I± .11 I. 42± . 06 I. 32± . 06 I. 00 . g6± .06 . 90± . 04 .89± .03 I. ii± .04 . 92 ± .08 . 90± . 04 .84± .03 .83± .04 •55± ■°^ • 43± -02 ■ 43± -02 .38± .01 1,802 1,348 .229 I, I'l: 1,05' 7f'7 483 454 1-50 I- 13 1.03 .88 .64 .41 •38 52 Jouynal of Agricultural Research Vol. Ill, Xo. I The relative water requirement of the different varieties included in the 19 1 3 experiments will be found summarized on the basis of dry matter in Tabic XXX and on the basis of grain production in Table XXXI- Fifty-five species and varieties were included in these measurements. Reference to these tables will show that a number of plants had a higher water requirement than alfalfa, heretofore the most inefficient in the use of water of any plant included in these experiments. On the other hand, millet maintains its supremac)- as the most efficient plant so far included in the water-requirement measurements. Table XXX. -Summary of water-rcquiremeyit measurements ai Akron, Colo., in IQIJ, based on dry matter produced Crop aiiH variety. Botanical name. Water require- ment. Relative, as compared with Kubanka wheat. I9I.3- Wheat-grass, western Bromc -grass Rag\\'eed .western Vetch , purjilc Flax Marigold, fetid Pumpkin Alfalfa, Grimm Soy bean, wild Clover, crimson I.amb's-quarters Clover, red Bean, horse, S. P. I. 15429. . . Pea, Canada field Sunflower, narrow leaved . . . Bean, Mexican Bean, horse, S. P. I. 25645 Squash Rice, Honduras Rape Potato, McConnick Cucumber Sunflower, annual Wheat-grass Vetch, hairy Bean, navy Bean, soy, cultivated Potato, Irish Cobbler Cotton, Triumph Alfalfa, Peruvian Turnip Cantaloupe Oat: Swedish Select Burt Watermelon Cowpea Sunflower, narrow-1 e a v e d , from sand hills. Cabbage Wheat, Kubanka Cocklebur Agropyron Smithii Bromus inermis Ambrosia arteraisifolia . Vicia atropurpurea . . . . Linum usatissimum. .. . Boebera papposa Cucurbita pepo Medicago sativa Glycine soja Trifolium incamatum . . Chenopodium album. .. Trifolium repens Vicia faba Pisum sativum Helianthus petiolaris. . Phaseolus vulgaris. . . . Vicia faba Cucurbita maxima. . . . Orj'za sativa Brassica napus Solanum tuberosum . . . Cucumbis sativa Helianthus annuus. .. . Agropyron cristatum. . , Vicia villosa Phaseolus vulgaris . . . . Glycine hispida Solanum tuberosum . . . Gossypium hirsutum . . Medicago sativa Brassica rapa Cucumis melo Avena sativa do CitruUus vulgaris Vigna sinensis Helianthus petiolaris . Brassica oleracea capitata, Triticum durum Xanthium commune i,076±29 I, oi6±26 948 ±66 93S±9 905^25 881 ±26 834±i7 834±S 8i5±25 805 ±8 801 ±41 789 ±9 78o±i9 775 ±5 774±2o 773±8 772±ii 748 d= 8 744±i7 743 ±7 7i7±ii 7i3±ii 705 ±8 70S±27 600 ±8 682 ±4 672±9 6S9±i5 6s7±ii 65i±i2 639^31 62I±27 6i7±9 6i7±5 6oo±i5 571 ±3 570±ii 539 ±7 496 ±5 432 ± 13 i7±o. 06 io± . 06 i.9i± i.89± I. 82± I. 78± i.68± i.68± I. 64± I.62± I. 62± i-59± '•57± I. 56± . 02 I. 56± . 04 I. 56± . 02 I. 56± .03 I. 51 ± .02 I. 5o± . 04 I. 5o± . 02 i-45± -03 I. 44± -03 I. 42± . 02 I. 41 ± i-39± i-38± i-35± ^■33± i-3i± I. 29± I. 2S± 13 °3 °S 06 04 02 °S 02 08 02 04 06 02 02 02 °3 03 03 06 06 I. 24± . 02 I. 24± . 02 I. 2I± .03 1. 15± .01 1. 15± .03 i.o9± I. 00 .87± °3 Oct. 15. 1914 Water Requirement of Plants 53 Table XXX. — Sttmmary of water requirement measurements at Akron, Colo., in Igij, based on dry matter produced — Continued Crop and variety. 1913- Com: China White Bloody Butcher Northwestern Dent . . Teosinte, Durango Grass, buffalo and grama. China WhiteXTeosinte . Com: Hopi China WhiteXHopi. Indian Flint Pigweed Grass, buffalo Sorghum: Minnesota Amber. . . Red Amber Purslane Millet, Kursk Botanical name. Zea mays . . . .do do Euchlene mexicana Bulbilis dactyloides and Boutelona gracilis. Zea mays ...do ...do Amaranthus retroflexus. Bulbilis dactyloides Andropogon sorghum . . ...do Portulaca oleracea. . . , Chaetochloa italica . . Water require- ment. 4I3±4 405 ±7 399±i2 3go±ii 389±i2 o76±4 350±8 345±3 342 ±5 320±7 3o8±i7 298 ±2 296 ± I 292±II 286 ±4 Relative as compared with Kubanka wheat . o. S4io. or 82 ± .02 8o± .03 79± . 02 78± -03 -6± . 01 71 ± .02 70rt . 01 6g± .01 05 ± .02 61 zb .04 6o± . 01 59± .or 59 ± .02 58± .01 Table XXXI. — Summary of water-requirement measurements in IQTJ lictsed on grain, tubers, roots, or fruit produced Crop and variety. I9I3- Sunflower; Annual Narrow-leaved Flax Pea, Canada field Bean, soy Potato, Irish Cobbler Bean , Mexican Oats, Swedish Select Cantaloupe Bean, Nav-y Oats, Burt Cucumber Cowpea Turnip UTieat, Kubanka Com, Northern Dent Watermelon Sorghum, Red Amber Com, Indian Flint Sorghum, Minnesota Amber Botanical name. Hclianthus annuus. .. Helianthus petiolaris Linum usitatissimum Pisum sativum Glycine soja Solanum tuberosum . Phaseolus vulgaris . . Avena sativa Cucumis melo Phaseolus vulgaris . Avena sativa Cucumis sativa Vigna sinensis Brassica campestris . . Triticum durum Zea mays Citrullus vulgaris . . . Andropogon .sorgum. . Zea mays Andropogon sorghum. Water require- ment. 384 ±383 o58±i32 83 s ±52 322±I2I 053 ±51 027±i97 888 ±62 876±55 824±237 646 ±36 641 ±33 61 1 ±67 S76±32 53o±i32 322±l6 241 ±77 146 ±49 ioo±3i 8S4±3i 765 ±12 Relative water requirement compared with Kubanka wheat. 4-83±c 29 3-83± . II 2. I4± •OS I. 76± . 1 1 i-55± .04 i-53± . 16 i-43± •05 I. 41 ± .04 i-38± .18 I.25± ■03 I. 24± •03 I. 22± •o.S I. I9± ■03 I. i6± . 10 I. 00 .04± .06 .87± .04 •«3± •03 .05± . 01 .58± 01 54 Journal of Agricultural Research voi. m. No. i COMPARISON OF THE WATER REQUIREMENT OF CROPS AT AKRON, COLO., IN 1911, 1912, AND 1913 Climatic conditions at Akron during the summer of 1912 were less severe than during the preceding summer. The rainfall in 1 91 2 was much greater than in 1 91 1 , the temperature was lower, and the evaporation was less. These conditions were apparently due in part to a marked reduc- tion in the intensity of the solar radiation at the earth's surface following the eruption of Mount Katmai, Alaska, early in June, 191 2, the dust from which produced a haze in the upper atmosphere. Abbot and Fowle (1913) observed a maximum reduction in the solar radiation of about 20 percent at Bassour, Algeria, and at Mount Wilson, Cal., Kimball (1913a, b) reports an average reduction of 17 per cent in the intensit}' of the solar radiation at Mount Weather, Va., during the last half of 1912, while Briggs and Belz (1913) have shown that there was a general reduction in the evaporation from a free-water surface during the summer months follow- ing the eruption. It is consequently of interest to determine whether the diminution in the intensity of the solar radiation was accompanied by a reduction in the water requirement in 191 2. Such a comparison is possi- ble in connection with the Akron experiments, since a large number of the varieties employed in the experiments of 191 1 were also included in the 1 91 2 measurements. All varieties showed in 191 2 a marked reduction in the water requirement as compared with 191 1. The measurements for each year are given in Table XXXII, together with the ratio of the 1912 to the 191 1 measurements. The 1912 measurements show an aver- age reduction in the water requirement of 21 ±2 per cent for the 25 varie- ties tested during both years. The individual ratios fluctuate somewhat, doubtless owing in part to errors of experiments,' but in part also to the different response of individual varieties to changed climatic conditions. 1 It should be mentioned here that the plants were fertilized in 1912 and not in 1911. This is a matter of importance in this connection, because it is well established that any deficiency in plant food increases the water requirement. The effect of the addition of fertilizer on the water requirement was measured both years. The use of fertilizer resulted in a slight reduction (6±2.3 per cent) in the water requirement of Kubanka wheat at Akron in 1911, comparing pots 7 to 12, fertilized, with pots 1 to 6. unfertilized. These pots stood side by side in the inclosure. (6Ve Briggs and Shantz, 1913a, p. 19.) In 1912 the fertilized Kubanka wheat plants showed a slight increase in water requirement — namely, 5 ±2.3 per cent, compar- ing pots I to 6 against pots 7 to 12. The differences in each instance are without significance when the errors are considered and are furthermore of opposite sign, so that the addition of fertilizer may be con- sidered to have had no effect on the water requirement, so far as Kubanka wheat was concerned. Rich surface soil from the same source was employed in the experiments of both years. Oct. 15, 1914 Water Requirement of Plants 55 Table XXXII. — Comparison of water-requirement measurements at Akron, Colo., in igil and igi2 Crop. Wheat: Kubanka Bluestem Spring Ghirka Emmer Oats: Sixty-Day Burt Canadian Swedish Select Barley : Hannchen Beldi White Hull-less Beardless Rye, spring Com: Northwestern Dent Iowa Silvermine. .. . Esperanza Sorghum : Red Amber Milo, Dwarf Kafir, Blackhull . .. Durra, White Kaoliang, Brown . .. Millet, German Legumes: Alfalfa Clover, sweet Beet, sugar Mean water-requirement ratio for 1912 to 191 1 Mean evaporation ratio for June, July, and August, igi2 to 1911 Water requirement. 468 ±8 S3i±5 506 ±3 534±i4 60s ±5 639 ±7 598±i4 6i5±7 527±8 543 ±2 S42±3 544±9 724±7 368±io 420±3 3i9±S 298±4 333 ±3 278±5 32I±2 301 ±3 263±i5 i,o6S±i6 709 ±9 377±8 394±7 45i±4 4S7±3 428±3 491 ±13 449 ±3 399 ±6 423 ±5 443 ±3 4i6±4 439 ± I 403 ±8 496 ±9 28o±io 3°2±7 239±3 237±4 273±4 259±S 255±3 223±I 248 ±7 659±6 638±4 32i±8 Ratio. 191J to igii. O. 84 ±0. 02 .85± .01 . 9oi: . 01 . 8o± . 02 .8i± .02 • 7o± . 01 . 67± .02 . 69± . 01 . 84± . 01 . 76± . 01 .8i± .01 . 74± • 02 . 69± . 01 ■75± -03 . 72± .02 . 79± .01 . 8oi: . 02 . 82 ± .02 ■ 93± -02 . 79± . 01 . 74± -oi .94± -05 . 62 ± .03 . 90± . 02 .85± .03 79± .02 ■75± -03 Evaporation measurements from a free-water surface at Akron are also available for igii and 1912 (Briggs and Belz, 1910, p. 17). The ratio of the evaporation in 1912 to that in 191 1 by months is as follows: June, 0.69; July, 0.78; August, 0.79. The average ratio during these three months is 0.75 ±0.03. This corresponds to a reduction of 25 ±3 per cent in evaporation as compared with a reduction of 21 ±2 per cent in the water requirement. The change in the water requirement of these 25 varieties taken together is seen to be in approximate agreement with the change in evaporation from a free-water surface. A consid- eration of the individual ratios indicates, however, that different varie- ties may respond quite differently to the same change in climatic con- ditions. To investigate further the variation in water requirement due to differences in climatic conditions, a number of varieties were also 56 Journal of Agricultural Research Vol. III. No. I included in the measurements of both 1912 and 1913. The ratios of the water requirement of these crops in 191 2 to that in 191 3 are given in Table XXXIII. Similar ratios are also given for crops grown in 191 1 and 1913. It will be seen that the mean 1913-1911 ratio approxi- mates unity; in other words, the mean water requirement of the crops under investigation was practically the same for both 3-ears. The water requirement in 191 2 is seen to be far below the 1913 value, the mean ratio being 0.75 ±0.01. The mean ratio of the monthly evapora- tion for June, July, and August, 1912, compared with i9i3,is o.8o± 0.02, which is in approximate agreement with the ratio of the water requirement of crops grown during the two years. The crops at Akron as influenced by climatic conditions in 1911, 191 2, and 1913 may then be summarized as follows: Table XXXIII. — Comparison of the -water requirement of the same crops at Akron, Colo., 1911, 1912, and IQI3 Water requirement. Crop. 1911. 1913. I9U. Ratio. 1913 to 1911. 1913 to J912. Wheat, Kubanka 468 ±8 6i5±7 639 ±7 368±io 394±7 423 ±5 449 ±3 28o±io 285±7 3i5±7 239±2 237±4 i87±2 657±ii 496 ±5 6i7±9 6i7±5 399±i2 358±8 4r6±4 298 ±2 296±i 286±4 834±8 775±S 659±i5 657±ii 744±i7 I. 06 I. 00 ■97 1.08 0.80 Oats: Swedish Select .69 Burt ■ 73 Com: Northwestern Dent Hopi • 7° .80 China White •76 Sorghum : 80 298±4 •99 .80 Millet, Kursk .66 Legumes : i,o68±i6 8oo±i7 44S±ii •77 •97 '•47 •79 Potato Irish Cobbler Cotton 488 ±14 496 ±9 ■ 7S±°t Rice Mean water requirement I. 04±. 04 .94±^o4 Evaporation in inches for three summer months . 28.46 21.42 26.75 . 80+02 The conditions during 1911 and 191 3 were such as to give rise to prac- tically the same water requirement. The water requirement of crops grown in 1912 was on the average only 79±2 per cent of crops grown in 1911 and 75±2 per cent of crops grown in 1913. Therefore, in order to determine the relative water requirement of the different crops, it appears justifiable to increase the 191 2 water requirement ratios by the reciprocal of 0.77 — namely, 1.3. This procedure has been followed in the summary Oct. IS, 1914 Water Requirement of Plants 57 table (Table XXXIV) , which places the water requirement of all crops upon the basis of years similar to 191 1 and 191 3. When the water- requirement measurement of any particular crop extended over more than one year, the mean value of the several water-requirement deter- minations is given in Table XXXIV. SUMMARY This paper deals with the measurement of the water requirement of plants at Akron, Colo., in the central portion of the Great Plains. The term "water requirement" is here used to express the ratio of the water absorbed by a plant during its period of growth to the dry matter pro- duced. The plants were grown to maturity in large galvanized-iron pots having a capacity of about 1 15 kg. of soil. Each pot was provided with a tight-fitting cover having openings for the stems of the plants, the annular space between the stem of the plant and the cover being sealed with wax. The loss of water was thus practically confined to that taking place through transpiration, and the entrance of rainfall was almost wholly prevented. The pots were weighed two or three times weekly to deter- mine the amount of water required to maintain normal weight. Water was delivered from 2-liter calibrated flasks through stoppered openings in the middle of the cover to a 5-inch flowerpot sunk in the soil immediately beneath the cover. To protect the plants from birds and severe wind and hail storms, it was found necessary to conduct the experiments in a screened inclosure. Pyrheliometric measurements showed that the inclosure reduced the radiation about 20 per cent. Water-requirement measurements con- ducted simultaneously with the same plants inside and outside the inclosure showed that the inclosure also reduced the water requirement approximately 20 per cent. Rich surface soil was used in the pots, and the pots were also fertilized to insure an adequate supply of plant food. Six pots of plants of each variety were used, and the water requirement of each pot was deter- mined independently, in order to provide a basis for the determination of the probable errors of the experiment. The detailed results given in the paper comprise measurements of 44 species and varieties in 1912 and 55 in 1913. The writers' 191 1 measure- ments have also been included in the summary table. The years 1911 and 1 91 3 were similar in character, and the same plants grown during both years gave practically the same water requirement. The 3'ear 191 2 was cooler and the evaporation and light intensity were much lower. These conditions had a marked influence on the water require- ment, the mean water requirement in 1912 being only jj per cent of that in 191 1 and 1913. In order to place all of the determinations upon a comparative basis, the 191 2 measurements have accordingly been increased 30 per cent in the summary table (Table XXXIV). 58 Journal of Agrictdtural Research Vol. Ill, No. I T.\Bi,B XXXIV. — Sutnmary of water-requirement determinations at Akron, Colo., in IQII, IQI2, and 1913, based on the prodxiction of dry matter Plant. Botanical name. Num- ber of ob- serva- tions. Water requirement. Of species or variety. ^Ican of genus. GRAIN CROPS. Proso: Voronezh, C. I. 16 Tambov, S. D. 366 Black Voronezh, S. D. 334 Millet: Kursk, S. P. I. 30029 . . . . Kursk, S. P. I. 34771 Kursk, S. P. I. 22420 German, S. P. I. 26845 . . . . Turkestan, S. P. I. 20694. . , Sorghum: Kafir, Dwarf Blackhull, C. I. 340. Kaoliang, Brown, S. P. I. 24993- Kafir, White, C. I. 370. . . Red Amber, S. P. I. 17563 Kafir, Early Blackhull, C. I. 472- Minnesota Amber, A. D. I. ,■541-13- Kafir, Blackhull. S. P. I. 24975- Milo, White, C. I. ^65 KafirXDurra, C. I. 198- „ ^5-3- Fetenta, C. I. 182 Milo, S. P. I. 24960 Durra, Wliite, S. P. I. 24997. Milo, Dwarf, S. P. I. 24970. Sudan grass, S. P. I. 25071. Com: Esperanza China WhiteX Esperanza, Indian Flint China White XHopi Hopi China WhiteXLaguna. . . . Northwestern Dent Laguna Bloody Butcher Iowa Silverminc China White Teosinte : China WhiteXTeosinte . . . Teosinte Wheat: Turkey, C. I. 1571 Kharkov, C. I. 1583 Kubanka, C. I. 1440 Galgalos, C. I. 2398 Emmer, C. I. 2951 Spring Ghirka, C. I. 1517. . Marvel Bluestem, C. 1. 30S2 Panicum miliaceum . ....do ....do Chaetochloa italica . ....do ....do .do. .do. Andropogon sorghum. .. ..do .do. .do. .do. .do. . .do. . . .do. . ..do. .do. .do. ..do... ..do... ..do... Zea mays . do . . . .do. . . . ....do.... ....do.... ....do.... ....do.... ....do.... ....do.... ....do.... ...do.... Euchlaena mexicana . Triticum aestivum. ....do Triticum durum. .. . Triticum aestivum. Triticum dicoccum. Triticum aestivum. ...do Years. 268±i 270±I 341 ±1° 26i±i5 265±3 2S7±2 293±9 444±9 285±3 296±2 297±4 301 ±2 3°2±I3 305 ±2 3o8±4 3I7±3 32i±5 323±4 324±4 327±2 344 ±3 467^9 3i5±3 325±2 342 ±5 345±3 36i±6 376±S 377±7 3S4±8 405 ±7 407 ±5 4i3±S 376±4 390±ii 473 ±8 475±8 492 ±4 496 ±4 545 ±7 S5o±3 559±4 293 322 > 368 383 513 Oct. IS. 1914 Water Requirement of Plants 59 Table XXXIV. — Summary of waier-requiremeni determinations at Akron, Colo., in igil, IQ12, and IQ13, based on the production of dry matter — Continued Plant. Botanical name. Ntun- ber of ob- serva- tions. Water requirement. Of species or variety. Mean of genus. GRAIN CKOPS — continued. Barley: Hannchen, C. I. 531 Beardless, C. I. 716 Beldi, C. I. 190 While Hull-less, C, I, 595, . Buckwheat Oats: Canadian, C. I, 444 Swedish Select, C, I, 134. Burt, C. I, 293 Sixty-Day, C. I, 165 Rye, spring, C. I. 73 Rice, Honduras, C. I. 1643. • • • Flax, North Dakota, No. 155. . OTHER CROPS. Beet, sugar; Morrison-grown Klein- wanzleben. Potato: Irish Cobbler McCormick Crucifers: Cabbage, Early Jersey, Wakefield. Turnip, Purple-top Rape Cotton , Triumph Cucurbits: Watermelon, Rocky Ford. . Cantaloupe, Rocky Ford. . Cucumber, Boston Pickling Squash , Hubbard Pumpkin, common field . . Legumes: Covvpea, S. P. I. 29282 Chick-pea, S. P. I. 24322.. Bean, navy Bean , Mexican Bean, soy, S. P. I. 21755. . Bean, soy (wild), S. P. I. 25138. Clover, sweet, S. P. 1. 21216 Pea, Canada field, S.P.I. 30134. Pea, Canada field, S.P.I. 22637. Vetch, hairy, S. P. I. 3429: Bean, horse, S. P. I. 25645 Bean, horse, S. P. I. 15429 Vetch, purple, S. P. I. 18131- Clover, red, S. P. I. 34869 Clover, crimson, S, P. I. 33742. Hordeum distichon Hordeum vulgarc ... .do do Fagopyrum fagopyrum. Avena sativa ....do ....do ....do Seeale cereale Orj'za sativa Liuum usitatissimum . , , Beta vulgaris. Solanutn tuberosum. ...do Brassica oleracea capitata Brassica rapa Brassica napus Gossypium hirsutum. Citrullus vulgaris. . . Cucuniis melo Cucumis sativa Cucurbita maxima. Cuciu'bita pepo Vigna sinensis Cicer arietinum. . . . Phaseolus vulgaris. ....do Glycine hispida . . . Glycine soja Melilotus alba. . Pisum sativum . ....do Vicia villosa Vicia faba ...do Vicia atropurpurea. Trifolium repcns Trlfolium incamatum. 502 ±4 534±7 542 ±3 556±2 S78±i3 559±8 .';94±4 6i3±3 622rt9 685 ±7 7io±i5 905^25 397 ±6 554±9 l 7i7±ii I 539±7 639±3i 743 ±7 646 ±11 600 ±15 621 ±27 7i3±ii 748±8 834±i7 S7i±3 663 ±18 682 ±4 773±8 672±9 8i5±2S 77o±5 77S±S 8oo±i7 690 ±8 772±ii 78o±io 935 ±9 789 ±9 805 ±8 534 578 597 685 710 9°S 397 636 640 646 600 667 791 571 663 728 744 770 788 794 797 6o Journal oj Agricultural Research Vol. III. No. I Table XXXIV. — Summary of ■water-requirement determinations at Akron, Colo., in 1911, 1912, and igi3, based on the production of dry matter — Continued Botanical name. Num- ber of ob- serva- tions. Water requirement. Of species or variety. Mean of Genus. OTHER CROPS — continued. Legumes — Continued . Alfalfa, Peruvian, S. P. I. 30203. Alfalfa, Grimm, A. D. I. E-23-20-52. Alfalfa, yellow-flowered. . AUalfa, Grimm, S. P. I. 25695- Grasses: Wlieat-grass Brome-grass NATIVE PLANTS. Tumble weed Pigweed Piuslane Grass, buffalo Thistle , Russian Grass, buffalo and grama. Cocklebur Gum weed Sage , mountain Sunflower f Akron) Sunflower (sand hills) . . . Sunflower Lamb's-quarters Marigold, fetid Ragweed, western Wheat-grass, western Medicago sativa . ....do Medicago falcata. Medicago sativa. . Years. Agropyron cristatum. Bromus incrmis Amaranthus graecizans. . Amaranthus retroflexus . Portulaca oleracea Bulbilis dactj'loides Salsola pcstifer Bulbilis dactyloides ,Bouteloua gracilis Xanthium commune Grindelia squarrosa Artemisia frigida Helianthus petiolaris do. Helianthus annuus. Chenopodium album. Boebera papposa. Ambrosia artemisifolia. Agropyron Smithii. 65i±i2 844±8 865±i8 963 ±9 7o5±27 I, oi6±26 277±4 297 ±4 292 ±11 3o8±i7 336±S 389±i2 432 ±13 608 ±23 6i6±i8 774±2o S70±ii 705 ±8 801 ±41 881 ±26 948 ±86 i,o76i:29 831 861 287 292 308 336 389 432 608 616 683 801 881 948 I, 076 The results given in the summary table (Table XXXIV) therefore represent the water requirement of the plants for years similar to 191 1 and 1 91 3, when grown in a screened inclosure, which reduces the solar radiation to 80 per cent of its normal value. According to measurements made with wheat, alfalfa, and cocklebur, the removal of the inclosure would increase the water requirement as given in the table by 25 per cent. The plants grown outside the inclosure were isolated and freely exposed, while plants under field conditions mutually protect and shade one another to some extent. Comparison with wheat plants grown in pots sunk in trenches indicates that the inclosure measurements, at least in the case of wheat, are less than 10 per cent below the water requirement of plants exposed under field conditions. The measurements in Table XXXIV represent the relative water requirement of the different plants tested, subject to the limitations imposed by the difference in the growth period of the plants. The Oct. n, 1914 Water Requirement of Plants 61 measurements in 191 2 would also indicate that the character of the season influences the water requirement of some plants more than others. For this reason, where the results of two or more years have been combined in the table, the probable error has been confined to the errors of experiment and does not include the fluctuations in water re- quirement due to the season. In order to facilitate comparison, the plants have been arranged under three main heads: Grain crops, other crops, and native plants. Under "Grain crops" are also included certain sorghums and millets which are usually grown for forage. Under the heading "Other crops" are included principally the legumes, cucurbits, crucifers, sugar beets, cotton, and potatoes, as well as some of the introduced grasses. Under the heading "Native plants" are listed indigenous species, as well as certain introduced species which have become thoroughly established. The grain crops fall rather naturally into two sections: Those of low water requirement — proso, millet, sorghum, and com — and those of high water requirement — wheat, barley, oats, rye, and flax. The plants with a comparatively low water requirement are late-maturing crops, which make their best growth during the hottest and driest portion of the summer. The plants having a comparatively high water requirement mature during midsiunmer and make their best growth during the earlier, cooler period of the year. The range in water requirement of the first group is from 261 for Kursk millet to 468 for Sudan grass, while the range in the second group is from 473 for Turkey wheat to 905 for flax. Representing the water requirement of proso as i, the water require- ment of the grain crops is as follows: Millet, 1.06; sorghum, i.io; com, 1.26; teosinte, 1.34; wheat, 1.76; barley, 1.83; buckwheat, 1.98; oats, 2.04; rye, 2.34; rice, 2.42; and flax, 3.38. In other words, flax requires more than three times as much water and rice more than twice as much water as proso and millet in producing a pound of dry matter. In the second group sugar beet ranks first, having a water requirement almost as low as com. Potato ranks next, followed by crucifers, cucur- bits, legumes, and grasses, in order. A wide range is shown in each of these families. The groups as a whole show a range somewhat less than the grain groups. Representing the water requirement of sugar beet as i, the values for the "Other crops," exclusive of the legumes, are as follows : Cabbage, 1.36; Irish Cobbler potato, 1.39; watermelon, 1.5 1 ; cantaloupe, 1.57; tur- nip, 1.60; cotton, 1.63; cucumber, t.8o; wheat-grass, 1.85; rape, 1.87; squash, 1.89; pumpkin, 2.10; and brome-grass, 2.56. The cowpea was the most eflicient of the legumes. Representing its water requirement by i , that of the other legumes is as follows: Peruvian alfalfa, 1.14; chick-pea, 1.16; soy bean, 1.18; navy bean, 1.20; hairj' vetch, 1.21; sweet clover, 1.35; Mexican bean, 1.35; horse bean, 1.36; red clover, 1.38; Canada field pea, 1.38; crimson clover, 1.41; wild soy 62 Journal of Agricultural Research voi. m, no. i bean, 1.42; select Grimm alfalfa, 1.48; yellow-flowered alfalfa, 1.5 1; purple vetch, 1.64; and unselected Grimm alfalfa, 1.69. The native plants show a range in water requirement greater even than the cultivated crops. Amaranths, buffalo and grama grasses, purslane, and Russian thistle have a low water requirement and com- pare favorably with millet and sorghum, while sunflower, fetid marigold, western ragweed, and western wheat-grass have a high water requirement about equal to that of alfalfa. Representing the water requirement of tumbleweed (Amaranihus graecizans) as unity, the water requirement of these native plants is as follows; Purslane, 1.05; pigweed, 1.07; buffalo grass, i.n: Russian thistle, 1. 21 ; buffalo and grama grasses, 1.40; cocklebur, 1.56; gumweed, 2.20; mountain sage, 2.22; sunflower, 2.56; narrow-leaved sunflower, 2.80; lamb's-quarters, 2.89; fetid marigold, 3.18; western ragweed, 3.42; western wheat-grass, 3.89. Varieties of the same crop often differ widely in water requirement. In the case of barley, the variety having the highest water requirement was 8 per cent above the lowest ; oats, 1 1 per cent ; wheat, 1 8 per cent ; proso, 27 per cent; com, 31 per cent; vetch, 35 per cent; alfalfa, 48 per cent; sorghum, 60 per cent; and millet, 70 per cent. This wide range in water requirement among the varieties of many crops encourages the belief that strains may yet be secured which are still more efficient in the use of water than those now grown in dry-land regions. LITERATURE CITED Abbot, C. G. 1911. The silver disk pyrheliometer. In Smithsn. Misc. Collect., v. 56, no. 19, 10 p., I pi. • and FowLE, F. E. 1913. Volcanoes and climate. In Smithsn. Misc. Collect., v. 60, no. 29, 24 p., Briggs, L. J. 1913. A mechanical differential telethermograph and some of its applications. In Jour. Wash. Acad., Sci., v. 3, no. 2, p. 33-35, i fig. and BELZ, J. O. 1910. Dry farming in relation to rainfall and evaporation. U. S. Dept. Agr. Bur. Plant Indus. Bui. 1S8, 71 p., 23 fig., i pi. 1913. Evaporation in the great plains and intermountaia districts as influenced by the haze of 1912. In Jour. Wash. Acad. Sci., v. 3, no. 14, p. 381-3S6. and Shantz, H. L. 1913a. The water requirement of plants. I. — Investigations in the Great Plains in 1910 and 1911. U. S. Dept. Agr. Bur. Plant Indus. Bui. 284, 49 p., 2 fig., II pi. 1913b. The water requirement of plants. II. — A review of the literature. U. S. Dept. Agr. Bur. Plant Indus. Bui. 285, 96 p., 6 fig. Collins, G. N. 1914. A drought-resisting adaptation in seedlings of Hopi maize. In Jour. Agr. Research, v. i, no. 4, p. 293-302, 2 fig., pi. 29-32. Oct. 15, .<,-4 Water Requirement of Plants 63 Kimball, H. H. 1913a. The effect of the atmospheric turbidity of 1912 on solar radiation intensi- ties and skylight polarization. In Bui. Mt. Weather Observ., v. 5, pt. s, p. 295-312, I fig. 1913b. The unusual atmospheric haziness during the latter part of 1912. In Jour. Wash. Acad. Sci., v. 3, no. 10, p. 269-273. Lawes, J. B. 1850. Experimental investigation into the amount of water given off by plants during their grow-th, especially in relation to the fixation and source of their various constituents. In Jour. Hort. Soc. London, v. 5, p. 38-63, illus. Leather, J. W. 1910. Water requirements of crops in India, /n Mem. Dept. Agr. India, Chem. Ser., V. I, no. 8, p. 133-1S4, pi. 3-19. 1911. Water requirements of crops in India. — II. In Mem. Dept. Agr. India, Chem. Ser., v. i, no. 10, p. 205-281, illus. "Student." 1908. The probable error of a mean. In Biometrika, v. 6, no. i. p. 1-25. ViNALL, H. N., and B.\ll, C. R. 1913. Feterita, a new variety of sorghum. In U. S. Dept. Agr. Bur. Plant Indus. Circ. 122, p. 25-32. PLATE I Fig. I. — General view of the plant inclosure used at Akron, Colo., showing the pipe framework covered with a hail screen, with the board base surmounted by a single width of cheesecloth to protect the plants against high winds. Photographed on July 15, 1912. Fig. 2. — General view inside the inclosure, showing the arrangement of pots and general conditions of growth. G)m and sorghums are shown in the foreground, small grain in the background. Photographed on July 3, 1912. Fig. 3. — General view of the inclosiu'e photographed shortly after the grain in some of the pots had been harvested. Photographed on September 4, 1912. (64) Water Requirement of Plants Plate I Journal of Agricultural Research Vol. III. No. 1 Water Requirement of Plants Journal of Agricultural Research Vol. III. No. I PLATE II Fig. I. — Pot planted with sugar beets, showing the wax seal around the plants and also the sealed holes where stand was not perfect. Fig. 2. — Weighing pots, showing spring balance, weighing support, and general pro- cedure. Two men operate the weighing support, one of whom lifts the pot by means of a windlass, while a third reads the balance and records the weight. By this method weighings can be made at the rate of two per minute. Fig. 3. — Grindelia squarrosa (gumweed) at left (pot 164), and Artemisia frigida (moxmtain sage) at right (pot 167), illustrating the growth of native plants used in the water-requirement measurements. Photographed on July 29, 1912. Water require- ment of Grindelia, 468±i8; of Artemisia, 474±i4. 60300°— 14 5 , PLATE III Fig. I. — Kubanka wheat (pots i to 6), gro\vn May g to September 3, 1912. Photo- graphed on July 19, 1912. Water requirement, 394±7. Fig. 2. — White HuIl-Iess barley (pots 103 to 108), grown May 16 to August 12, 1912. Photographed on July 19, 1912. Water requirement, 43g±i. Fig. 3. — Kubanka wheat (pots 12 to 18). Set grown outside of shelter, May 9 to August 31, 1912. Photographed on July 26, 1912. Water requirement 507^=13. Fig. 4. — Emmer (pots 61 to 66), grown May 11 to August 12, 1912. Photographed on July 19, 1912. Water requirement, 428±3. Fig. 5. — Swedish Select oats (pots 85 to 90), grown May 17 to .\ugust 23, 1912. Pho. tographed on July 26, 1912. Water requirement, 423±5. Fig. 6.— Kharkov wheat (pots 37 to 42), grown April 27 to August 28, 1912. Photo- graphed on July 29, 1912. Water requirement, 365±6. Water Requirement of Plants Plate III Journal of Agricultural Research Vol. III. No. 1 Water Requirement of Plants Plate IV ^^iS^SV Journal of Agricultural Research Vul. III. No. 1 PLATE IV Fig. I. — Northwestern Dent corn (pots 277 to 282), grown June 9 to September 16, 1912. Photographed on September 6, 1912. The lower leaves were picked off as they became dry and placed in bags attached to the pots, thus avoiding loss of dry matter. Water requirement, 28o± 10. Fig. 2. — Hopi corn (pots 295 to 300), grown June 12 to September 26, 1912. Pho- tographed on September 9, 1912. The lower leaves were picked off as tliey became dry and placed in bags attached to the pots. Water requirement, 285 ±7. Fig. 3. — White durra (pots 235 to 240), grown June 9 to September 26, 1912. Pho- tographed on September 6, 1912. The lower leaves were removed as soon as they became dry and placed in bags attached to the pots. Water requirement, 255±3. Fig. 4. — Red Amber sorghum (pots 253 to 258), grown June 29 to September 27, 1912. Photographed on September 7, 1912. Water requirement, 237 ±4. Fig. 5. — Minnesota Amber sorghum (pots 247 to 252), grown June 9 to September 26, 1912. Photographed on September 9, 1912. Water requirement, 239±2. PLATE V Fig. I. — Sudan grass (pots 211 to 2161. First crop, grown May 28 to July 26, 1912. Photographed on July 25, 1012. Water requirement, 3i2±3. Fig. 2. — Voronezh proso (pots 199 to 204), grown June 5 to August 20, 1912. Pho- tographed on July 29, 1912. Water requirement, 2o6i:i. Fig. 3. — Kursk millet (pots 205 to 210), grown June 9 to August 20, 1912. Pho- tographed on July 30, 1912. Water requirement, i87±2. Fig. 4. — Select Grimm alfalfa (pots 139 to 144), grown in the open, May 24 to July 27,1912. Photographed on July 26, 1912. Water requirement. 745±22. Fig. 5. — Select Grimm alfalfa (pots 133 to 138), grown in the shelter, May 24 to July 26, 1912. Photographed on July 26, 1912. Water requirement, 6oo± 17. Water Requirement of Plants Plate V g^imyny^ Journal of Agricultural Research Vol. Ill, No. 1 Water Requirement of Plants Plate VI 1* IpW'-^ ■i^^S'^r WWi^ ij.,. J¥" ^- ■^- \lA h>'-^ Journal of Agricultural Research Vol. III. fJu. 1 PLATE VI Fig. I.— Cowpea (pots 151 to 156), grown June 17 to August 26, 1913. Photographed on July 26, 1913. Water requirement, S7i±3. Fig. 2. — Hairy vetch (pots 181 to 186), grown May 29 to July 18, 1913. Photographed on July 17, 1913. Water requirement, 672±9. Fig. 3. — Soy bean (pots 193 to 198), grown Jime i to August 26, 1913. Photographed on July 26, 1913. Water requirement, 690±8. Fig. 4. — Cantaloupe (pots 325 to 330), grown June 14 to September 13, 1913. Pho- tographed in place on July 26, 1913. Water requirement, 778±34. Fig. 5. — Indian Flint com (pots 253 to 258), grown June 7 to August 27, 1913. Photographed on July 26, 1913. Water requirement, 342 ±5. Fig. 6. — McCormick potato (pots 121 to 126), grown June 5 to October 4, 1913. Photographed on July 26, 1913. Water requirement, 7i7±ii. PLATE VII Fig. I. — Triumph cotton in shelter (pots 163 to 168), grown May 29 to September 16, 1913. Photographed on September 15, 1913. Water requirement, 657±ii. Fig. 2. — Boebera papposa, gro«-n July 25 to September 17, 1913. Photographed on September 15, 1913. Water requirement, 881 ±26. Helianthus petiolaris in back- ground. Fig. 3. — Rice in shelter (pots 157 to 162), grou-n June 12 to September 16, 1913. Photographed on September 15, 1913. Water requirement, 744±i7. Fig. 4. — General view of the shelter, showing emmer at the left and White Hull-less barley at the right. Photographed on July 3, 1912. Fig. 5. — General view in the shelter, showing com in the foreground. Photographed on September 6, 1912. Water Requirement of Plants Plate VII Journal ut At;rlcultural Research Vol. Ill, No. 1 HEART-ROT OF OAKS AND POPLARS CAUSED BY POLYPORUS DRYOPHILUS By George G. Hedgcock, Pathologist, and W. H. Long, Forest Pathologist, Investi- gations in Forest Pathology, Bureau of Plant Industry INTRODUCTION The oaks {Qtiercus spp.) of the United States are diseased by a number of species of fungi which attack the heartwood. Von Schrenk and Spauld- ing (1909)' briefly described some of these diseases and also a piped rot of the heartwood of oaks and chestnuts {Castanea dentata) the cause of which was unknown to them. In 1 909, the senior writer found Polyporus dryophihis constantly associated with a whitish piped rot of several spe- cies of oaks in the southwestern and western United States. This rot was much like that described by Von Schrenk and Spaulding and was identical with that of specimens in oak collected by them. Later observations by the senior writer established the causal relation of Polyporus dryophilus to this piped rot. The junior writer in 1913 found a second form of piped rot caused by Polyporus pilotae in the heart-wood of the root and basal portion of the trunks of oaks and also in chestnuts. This was identical with the rot in chestnut trees figured and collected by Von Schrenk and Spaulding. The oaks of the southwestern and western United States are not used to any extent for lumber and timbers and are, as a rule, valuable only for fuel. This is due to the rotted condition of the heartwood in the larger and older trees. For example, the trunks of the valley oak {Quercus lobata),- which attains a large size in the valleys of central California, are usually either badly decayed or hollow and are of no value except for the poor grade of fuel they furnish. The senior writer in 1909 ascertained that Polyporus dryophilus was the chief cause of the deterioration of the oaks of the western United States. Meinecke (191 4) reports a destruc- tive heart-rot of oaks caused by this fungus in California and Nevada, and data by him will be cited in the section on the distribution of the fungus. In Arizona and New Mexico the oaks are diseased in the heart- wood nearly as badly as in California and Oregon, and P. dryophilus is the common cause of decay. In these States oaks are usually small and are valuable only for fuel. In Texas and the adjacent States of Oklahoma and Arkansas the piped rot produced by this fungus is very common, and among other ' Bibliographic citations in parentheses refer to " Literature cited, " p. 77. 2 The nomenclature for trees used in this paper is that of George B. Sud worth ( 1898). Journal of Acricultural Research. Vol. Ill, No. i Dept. of A;:.'iculture, Washington. D. C Oct. 15, 1914 G-34 (65) 66 Journal of Agricultural Research voi. in. No. i species the valuable white oak (Quercus alba) is commonly attacked. To the east and north the fungus has been found less frequently, but it occurs in many sections. From observations and estimates Polyporus dryophilus ranks with the most common heart-rotting fungi which attack the oaks. In 191 2 the senior writer found aspens {Populus treniuloides) in Colorado attacked by this fungus. It apparently is not commonly found on this host. PIPED ROT CAUSED BY POLYPORUS DRYOPHILUS The whitish piped rot caused by Polyporus dryophilus has been found by the writers to be directly associated with the sporophores of this fungus in the following 15 species of trees: Quercus alba, 0. arizonica, Q. calijornica, Q. digitata, Q. cmoryii, Q. gambclii, Q. garryana, Q. marilandica, 0. minor, Q. prinoides, O. prinus, Q. texaita, 0. velutina, Q. virginiana, and Populus tremidoides. PIPED ROT IN THE WHITE OAK MACROSCOPIC CHARACTERS The first indication of the whitish piped rot in white oak is a discolora- tion of the heartwood, which assumes a water-soaked appearance (PI. VIII, fig. i). This "soak" may extend from i to 10 feet beyond the actually rotting region where delignification is occurring. When dry, this water- soaked heartwood becomes hazel to tawny in color. The next stage of the rot is one of delignification, which usually begins alongside of and fol- lowing more or less regularly the medullary rays, thus producing a mottled appearance of the wood in radial view (PI. VIII, figs. 2, 5, and 6). This type of the rot is very common in the medium-sized branches (6 to 12 inches in diameter) and in the early stages of the disease in the bole of the tree. In final stages the diseased wood is firm, has a white, stringy ap- pearance (PI. VIII, figs. 3 and 4) and consists of white cellulose strands of delignified wood fibers and other wood structures bounded by areas of apparently sound but actually slightly diseased and discolored heartwood. Cinnamon-brown areas are scattered throughout the oldest rotted wood (PI. VIII, fig. 3). These areas are especially common and abundant in the vicinity of sporophores and along checks or openings through the sap- wood. The rot immediately adjacent to a sporophore is therefore often cinnamon brown to russet in color. No cavities large enough to be seen by the naked eye are produced by this rot, but much of the white cellulose is finally absorbed, leaving minute irregular cavities in the wood. MICROSCOPIC CHASACTERS Delignification usually begins in the wood fibers lying next to the vessels in the spring wood and adjacent to the large medullary rays. The sol- vents secreted by this fungus apparently are able to delignify all of the Oct. IS, 1914 Heart-Rot of Oaks and Poplars 67 elements of the wood. All, or only the outer rows, of cells of the large medullary rays may be delignified, the middle lamellae dissolved, and the completely delignified cell membranes partially absorbed. Isolated areas between the large medullary rays may also be delignified. The cells of some of the medullary rays and of the wood parenchyma often contain starch grains even after the absorption of a portion of the inclosing cell walls. A ferruginous substance is also present in many of the cells of the small medullary rays, in the lumen of the wood fibers, and even in some of the other wood structures. Many of the vessels adjacent to each large medullary ray contain hyaline branching hyphae 0.5 to i/t in diameter. The association of the delignified areas with the medullary rays is readily seen in a cross section of the wood where delignification is just beginning, but later in the more advanced stages of the rot this association is not so evident when the delignification of the wood fibers has become general throughout the rotting area. The early absorption of portions of the delignified tissue prevents the forma- tion of long continuous strands of cellulose fibers, although in a tangential view irregular white lines may be seen which consist of fragments of the delignified cells (PI. VIII, fig. 5). In very advanced stages of the rot near the center of the tree white longitudinal lines are seen in a radial view (PI. VIII, fig. 4). These usually consist of remnants of partially absorbed cellulose fibers bound together by strands of white mycelium, which also fill the vessels and the minute cavities left by the absorption of the delignified tissue. PIPED ROT IN CHESTNUT OAK The rot produced by Polyporus dryophilus in the chestnut oak {Quercus prinus) is slightly different from that in white oak. The diseased wood is hazel in color, with very narrow concentric zones of ivory-yellow cellulose. These zones are adjacent to the large spring vessels of each year and consist of the delignified wood fibers of this tissue. The large vessels in radial-longitudinal view are seen, even under a hand lens, to be filled with cobwebby strands of colorless hyphae. It is in the tissue adjacent to such hyphae-filled vessels where the deUgnification is most pronounced. PIPED ROT IN THE WESTERN OAKS The rot caused by Polyporus dryophilus in these oaks differs but little from that found in the white oak. The mottled appearance of the rot in its earlier stages is not so pronounced. In the final stage of the rot, after a very large proportion of all the elements is delignified, there is but little apparently sound heartwood. In the older rot in the center of the heartwood the white color by far exceeds the brown, of which there is very little. 68 Journal of Agricultural Research voi. in. No. i PIPED ROT IN EUROPEAN OAKS Robert Hartig (1878), in his epoch-making work on the true nature of the rots of woods, described a whitish heart-rot of the oak, which he attributed to Polyporus dryadeus. A careful study of Hartig's figures, and the description of the sporophore which he found associated with the white heart-rot so accurately described by him, is sufficient to con- vince anyone who is familiar with the true P. dryadeus that Hartig's fungus was not P. dryadeus. It is undoubtedly identical with the heart-rotting fungus known in America as P. dryophilus and found by the senior author to be associated with a whitish piped rot in oak. Through the kindness of Dr. Von Tubeuf the junior writer obtained a piece of the original rot (PI. VIII, figs. 7 and 8) which Robert Hartig (1878) ascribed to P. dryadeus. A careful study of this specimen showed that it is identical in every respect with the rot produced by P. dryophilus in the white oak. There is also another European specimen (PI. VIII, fig. 9) of this rot in oak in the Laboratory of Forest Pathology, of the Department of Agriculture, which has all the characters of the rot pro- duced by P. dryophilus . CHARACTERS OF PIPED ROT COMMON TO ALL SPECIES OF OAKS The rots produced by Polyporus dryophilus in all the species of oak examined had the following characters in common: (i) A water-soaked discolored area in the first stage (PI. VIII, fig. i) ; (2) a general association of the earUer delignification with the medullary rays (PI. VIII, figs. 5 and 6) ; (3) later a more general delignification of all the wood fibers (PI. VIII, fig. 3) ; (4) the formation of white mycelial longitudinal lines (PI. VIII, fig. 4); (5) the presence of cinnamon-brown areas in the older rotted wood (PI. VIII, fig. 3). These brown patches, ranging from 2 by 4 mm. up to 10 by 35 mm. in size, consist of fragments of wood interwoven with ferruginous, thick-walled, septate hyphae, which easily break into short pieces. The hyphae are about 3// thick, have many short (3 to 8;u) branches, and are mixed with various sizes of hyphae down to i/x or less in diameter, the smaller of which are hyaline. HEART-ROT PRODUCED BY POLYPORUS DRYOPHILUS IN ASPEN The description of the heart-rot which follows was made from the diseased wood of a dead aspen {Populus tremuloides) bearing the sporo- phores of Polyporus dryophilus . MACROSCOPIC CHARACTERS The general color of the diseased wood varies from a light buff to a maize yellow. In a cross section the rotted wood shows alternating concentric zones of light buff and ochraceous tawny. The light-colored zones consist of the vessels and wood fibers which have been the most vngorously attacked by the solvents of the fungus. The ochraceous- Oct. 15. 1914 Heart-Rot of Oaks and Poplars 69 tawny zones consist of vessels, cells of wood parenchyma, and other elements of the wood in the cells of which a ferruginous amorphous substance has been deposited. These cells are not as strongly attacked by the fungus as are those of the light zones. The rotted wood easily splits into concentric layers, the cleavage usually occurring along the boundary between the white and dark zones. In a tangential view, small, more or less isolated areas of delignified wood fibers may be seen. These delignified fibers are most abundant in the older, rotted portion. In the vicinity of the sporophores the typical cinnamon-brown areas seen in the oak are also present. The rotted wood is soft, almost silky to the touch, is very light in weight, and is easily broken into fragments between the fingers. MICROSCOPIC CHARACTERS The vessels in the light-colored zone have very thin walls, owing to the action of the fungus; the bordered pits are often eroded until only large irregularly shaped holes are left and the middle lamellae of the vessels and of the wood fibers in this region are dissolved. The wood fibers and some of the adjacent cells are finally delignified and absorbed. The delignification occurs most rapidly along the boundary lines between the light-colored and dark-colored zones, along which the cleavage com- monly occurs. The small amount of delignified fibers present and their rather rapid absorption prevent the formation of the large areas of white cellulose which are so common in the rot produced by this fungus in oak. In the zone of cleavage cobwebby masses of white mycelium occur which fill the vessels and the small cavities left by the absorption of the wood fibers. The medullary rays are readily attacked by the solvents of this fungus and usually have completely disappeared by the time the final stage of the rot is reached. ENTRANCE OF THE ROT IN THE HOST Polyporus dryophilus, so far as known to the writers, gains entrance in the wood of the host trees only through wounds in which the heartwood is exposed. The most common point of entrance is a broken or dead limb, although in the western and southwestern United States it also frequently enters through fire scars and other basal wounds. In Arkansas and eastward, where the species of oaks differ from those in the West and Southwest, the rot caused by this fungus is apparently confined chiefly to the branches and upper portion of the trunk. This may be due to the fact that often there are one or more large dead branches in the crown of the tree, while there are very few on the lower part of the trunk. The fungus has therefore little or no opportunity to enter the bole of the tree below the crown. When the fungus enters the stub of a broken limb, it grows downward through heartwood of the stub till it enters the trunk, when it spreads yo Journal of Agricultural Research voi. iii, no. i both upward and downward through the heart of the tree. When it enters near the base of the tree, it sometimes spreads upward throughout the heart of the entire trunk. This occasionally was noted in the white oak in Arkansas, and such trees were worthless for lumber. In Oklahoma and to the west oaks frequently have large dead branches at any point on the trunk of the tree. Through these the fungus may enter. The rot therefore is not confined as closely to the upper half of the trees as it is in the oaks of Arkansas and to the east. Probably 50 per cent of the western oaks attacked by this fungus have the rot through- out the entire trunk. The sporophores of Polyporus dryophilus when growing on oak are usually found only on living trees; however, specimens have been col- lected growing on the boles and large branches of trees which had been cut for at least three years, and in one instance a sporophore was found growing directly on the top of an old oak stump. The fungus apparently continues to grow slowly in the infected trees after they have been cut, but rarely fruits under such conditions. There is no evidence at hand concerning the possibility of infection by P. dryophilus after the death of the tree. In no instance in Arkansas has the junior writer found this fungus entering a tree through fire scars or other wounds on the butt of oaks, even where fiire scars were common. The rot always originated at some point above the base of the tree, and if a tree was found in which the rot had reached the collar of the tree it came from above and not from below. All of the sporophores of this fungus found on specimens of Populus were growing on dead or dying trees. In this case the fungus is able to fruit abundantly on both Hving and dead trees. This fungus on Populus seems to be truly parasitic, to some extent at least. It attacks the trunks of the trees chiefly, entering the heartwood through dead limbs after they are broken ofiF. The trees die by either breaking off or in some cases apparently from the direct effect of the fun- gus, which attacks the sapwood when the disease becomes far advanced. Several instances were found in oak where the fungus had apparently penetrated and killed small areas of the sapwood and formed its sporo- phores at these points. No positive evidence was found indicative of the age of the fungus in either oaks or poplars or of its rate of growth in the infected tree. Appar- ently trees of all ages are susceptible to this rot, provided the branches are old enough to have formed heartwood. SPOROPHORE OF POLYPORUS DRYOPHILUS Polyporus dryophilus has a hard, granular, sandstone-like core, a character that is unique and not possessed by any other polypore known to the writers. The sporophore of this plant, represented by numerous specimens collected by the writers in various portions of the United Oct. 15, 1914 Heart-Rot of Oaks and Poplars 71 States, in every instance shows this hard granular core (PI. IX, figs. 2 and 4) exactly as figured and described by Hartig (1878) in case of his P. dryadeus. This core extends back some distance into the tree in oaks; it is usually irregularly cylindrical while in the tree, but on its emergence from the tree it swells into a tuberous or spheroid mass and finally occupies the central and rear part of the sporophore (Pis. IX, fig. 2, and X, fig. 6). If the sporophore is formed from a large branch hole, it is usually of the applanate type, with a small core, but when the sporo- phore forms directly on the body of the tree, as it usually does, the shape is tuberous, unguliform, or even subglobular (PI. IX, figs. 2 and 4), with the bulk of the sporophore composed of hard, granular core. This core usually has white mycelial strands (PI. IX, fig. 4). The sporophore of P. dryophilus, therefore, has normally three distinct kinds of structures (PI. X, fig. 4): (i) The hard, granular core; (2) the fibrous layer which surrounds this core except at the rear; (3) the layer of tubes on the lower surface. Specimens are often found, however, especially from the western part of the United States, in which this fibrous layer may be entirely absent between the tubes and the granular core (PI. IX, fig. 4). Polyporus dryophilus is known in Europe under at least five different names: Polyporus julvus Fries, P. friesii Bresadola, and P. corruscans Fries for the form on oak, and P. ridpinus Fries and P. rheades Persoon for the form on poplar. The identity of P. dryophilus with the P. cor- ruscans Fries (PI. X, fig. 4) and with P. rheades Persoon is based on the specimens of these plants found in the Lloyd Herbarium at Cincinnati, Ohio. If these specimens are correctly determined, then the American plant is identical with the European plants named above. Authentic specimens of the form of P. dryophilus found on species of Populus were seen by the junior writer at the New York Botanical Gardens in collec- tions from Finland and Sweden and also from Maine. In the Lloyd Her- barium at Cincinnati, Ohio, are collections under the name of P. rheades on Popidus trcmida from Sweden (PI. X, fig. 3) and Denmark, and a collection from Austria on Quercus ilex. In the Cryptogamic Herbarium of Harvard University there is a collection on Populus grandidentata Michx. from New Hampshire, while in the Laboratory of Forest Pathology there is a fine collection on Populus tremuloidcs Michx. (PI. X, figs, i, 2, and 5) from near Steamboat Springs, Colo. This fungus on Populus agrees in all essential characters with the form of Polyporus dryophilus found on oak. The sporophores are, however, somewhat smaller than those usually found on oak and approach the applanate type (PI. X, figs, i and 2). The hard granular core is always present, but is formed between the sapwood and bark (PI. X, fig. 4), as the fungus is able to rot the sapwood, as well as the heart of this host. It therefore does not have to depend on branch holes or other openings through the sapwood in order to form its sporophores as it does in the oak. 72 Journal of Agricultural Research voi. iii. No. i DESCRIPTION OF THE SPOROPHORE OF POLYPORUS DRYOPHILUS Pileus thick, unequal, smooth to irregular nodulose, often convex below, unguliform (PI. IX, fig. s), subglobose (PI. IX, fig. i) or even applanate (PI. X, fig. i), simple or rarely subimbricate (PI. X, figs. 2 and 5), rigid, 4 to 22 cm. broad by 3 to 13 cm. wide (measured from front to rear of sporophore) by 2.5 to 21 cm. thick (measured from pore surface to top of sporophore); surface at first densely tomentose, becoming scabrous to smooth with age ; tomentum rather stiff, deciduous, short, maize yellow to ferruginous; surface of weathered sporophoresafter the tomentum has partially disappeared, zonate, zones several, narrow, extending entirely around the pileus near its margin (PI. IX, fig. 3); margin in immature specimens thick, usually obtuse (Pis. IX, figs, i and 5, and X, figs. I and 2), concolorous or slightly pallid, entire or undulate; context dual, con- sisting of a hard granular core, surrounded except in the rear by a thin fibrous layer; core subglobose to pulvinate, 3 to 10 cm. thick, ferruginous to cinnamon brown, gran- ular, often with white mycelial strands ramifying through it (PI. IX, fig. 2); fibrous layer on upper surface of core a mere pellicle about 0.5 mm. thick, expanding in mature specimens into a border (PI. IX, fig. 4) i to 3 cm. wide and 5 to 15 mm. thick; fibrous layer between tubes and core thin, i to 15 mm., usually not over 6 to 8 mm., fibrous layer zonate, concolorous; tubes slender, concolorous or slightly paler than core in some specimens, rather fragile in age, 5 mm. to 3.5 cm. long, shorter near margin of sporophore, usually about i cm. long; mouths regular when young, but becoming somewhat irregular and angular at maturity (Pis. IX, fig. 6, and X, fig. 8), two or three to a mm., glistening, grayish when young, becoming hazel to russet with age, edges thin; spores broadly oval, smooth, ferruginous, 4.8 to 8 by 3.4 to 6.4/1, average size 6.54 by 4.85/1 when on oak (PI. IX, fig. 6), 4.8 to 6.4 by 3.4 to 5.6/1, average size 5.82 by 4.05/1 when on poplar (PI. X, fig. 8); cystidia none; hyphae ferruginous, 4 to 6,u. The sporophores found on oak in Arkansas and in the eastern portion of the United States often have shorter tubes (PI. IX, fig. 4), slightly smaller spores, and a more applanate pileus than those found in the Western States (PI. IX, fig. 2). DISTRIBUTION OF POLYPORUS DRYOPHILUS The rot caused by Polyporus dryophilns is very widely distributed in the United States, having been found in 23 States: Arizona, Arkansas, CaHfornia, Colorado, Ilhnois, Kansas, Louisiana, Maine, Maryland, Mis- sissippi, Missouri, Nebraska, New Hampshire, New Mexico, New York, North Carohna, Ohio, Oklahoma, Oregon, Pennsylvania, Tennessee, Texas, and Wisconsin. Authentic specimens of the fungus have also been examined from the following foreign countries: Austria, Denmark, Finland, France, Germany, and Sweden. The sporophores of the fungus are frequent and the rot caused by the fungus is exceedingly common in New Mexico, Arizona, and California. DISTRIBUTION IN EUROPE Polyporus dryophilus is known to occur in Europe as follows, the jimior writer having examined authentic specimens: Germany (?); On Quercus sp. (F. P. 12404)'. * " F. P." = Forest-PathoJogy Investigations number. Oct. 15. 19U Heart-Rot of Oaks and Poplars 73 Germany: On Quercus sp. — ROBERT HarTig (from Herb. Von Tubeuf); Pfeiffer (Herb. N. Y. Bot. Gard.), part of the type specimen for Polyporus friesii; Berlin, Lloyd (Herb. Lloyd). Austria: On Quercus ilex. — Travnik, Rev. E. Brandis (No. 08864, Herb. Lloyd). Denmark: On Populus (?) sp. — J. LiND (No. 06339, Herb. Lloyd). Finland: On Populus sp.— Murtiala, Sept., 1882 (No. 5724, Herb. N. Y. Bot. Gard.). Sweden: On Quercus robur. — Stockholm, Romell, Oct., 1903 (Herb. N. Y. Bot. Gard.), and a second specimen, collector unknown (Herb. N. Y. Bot. Gard.). On Quercus sp. — Upsala, Lloyd (No. 08936, Herb. Lloyd); Stockholm, Romell (No. 08936, Herb. Lloyd). On Populus tremula. — Stockholm, Romell, June 25, 1905 (Herb. N. Y. Bot. Gard.), and a second specimen, Murrill (Herb. N. Y. Bot. Gard.); Stockholm, Hagelund (No. 08985, Herb. Lloyd). On Populus sp. — Hagelund (No. 09375, Herb. Lloyd); Stockholm, Romell (No. 08414, Herb. Lloyd). France: On Pinus (?) sp. — Fontainebleau, P. Hariot (No. 08880, Herb. Lloyd); this speci- men from France was reported as on pine, and has spores similar in size and shape to those growing on species of Populus and a sporophore much like those found on species of Quercus. DISTRIBUTION IN UNITED STATES Polyporus dryophilus has been reported from and collected in the various States of this country as follows: Maine: On Populus iremuloides. — Piscataquis Co., Murrill, in 1905 (Herb. No. 1901, N. Y. Bot. Gard.). On Betula (?) sp. — Near Moosehead Lake, Von Schrenk, in 1899 (Herb. N. Y. Bot. Gard). New Hampshire: On Populus grandideniata. — Chocorua, Farlow (?), in 1904 (Herb. W. G. Farlow). New York: On Quercus alba. — Bronx Park, Murrill, in 1908 (F. P. 1416). Penns\'lvania: On Qu4!rcus (?) sp. — Kittanning, SuMSTiNE 32 (Herb. N. Y. Bot. Gard.). Maryland: On Quercus alba, Q. cocchiea, and Q. minor. — Takoma Park, Hedgcock, in 1910. Ohio: On Quercus (?) sp. — M. A. Curtis, "Ex. Berkeley" (Herb. W. G. Farlow); Preston (?), A. P. Morgan, in 1887 (Herb. N. Y. Bot. Gard.); Preston, A. P. Morgan (0598); and Akron, C. D. Smith (07556, Herb. Lloyd). Virginia: On Quercus prinus. — Elkins, Long, in 1913 (F. P. 12418). On Quercus (?) sp.— Falls Chtirch, LuTTrEll, in 1902 (Herb. N. Y. Bot. Gard.). North Carolina: On Quercus prinus. — Brim, Long. On Quercus veluiina. — Jonesboro, P. L. Buttrick, in 1913 (F. P. 15045). 74 Journal of Agricultural Research voi. in, No. i Tennessee: On Qxiercus alba and Q. veluiina. — Roan Mountain, Hedgcock, in 1913. Mississippi: On Quercus lyrata. — Sand Point, Hedgcock, in 1908. Louisiana: On Quercus lyrata, Q. marilandica, Q. michauxii, and Q. pliellos. — Near Bogalusa, Hedgcock, in 1908. Missouri: On Quercus alba. — Motintain Grove, Hedgcock. On Quercus imbricaria. — Near St. Louis, J. N. Gladfelter (No. 1214, Herb. Mo. Bot. Gard.). On Quercus marilandica. — Steelville, Spaulding; Mountain Grove, Hedgcock. On Quercus minor. — Webster Groves, A. H. Graves, in 1909 (F. P. 1617). On Querctis palustris. — Mountain Grove, Hedgcock. Illinois: On Quercus alba. — Near Plymouth, Hedgcock, October, igog. Wisconsin: On Populus sp. — Oakfield, in igoj (Herb. Univ. Wise). On Querctis macrocarpa. — Rockton, L. H. Pammel, in 1886 (Herb. N. Y. Bot. Gard.). Nebraska: On Quercus macrocarpa. — Near Nelson, Hedgcock, in igii. Okl.\hom.\: On Quercus alba. — Cache, Long, in igi2 (F. P. 12407). On Quercus marilandica. — Cache, Long, in 1912 (F. P. 12420). On Quercus minor. — Cache, Long, in 1912 (F. P. 12408, 12416, 12419, 12421). On Quercus prinoides. — Cache, Long, in 1912 (F. P. 12414). Arkansas: On Quercus alba. — Treat (F. P. 12102), Casteel (Ozark National Forest; F. P. 12137, 12140, 12142, 12154, I22ig, 12243, 12263, 12268, I22g6, 124C2, 12403, 12405, 12406, i24og, 12413, 12425); Bigflat (F. P. 12 158, 12 156, 12 160), Long, in igi2: Womble (F. P. 12413), Cedar Glades (F. P. 12422), Long, in igi3;Fayettevilleand Farmington, Hedgcock, in igo6. On Qttercus digiiata. — Casteel, Long, in igi2 (F. P. 12272). On Quercus minor. — Whiterock, Long, in igi2 (F. P. 12240). On Querctis iexana. — Mountain View, Long, in igi2 (F. P. 12415). On Qtierctis veluiina. — Casteel, Long, in igi2 (F. P. 12410). Texas: On Qtiercus marilandica. — Near Boeme, Hedgcock, in 1909 (F. P. 760). On Qttercus minor. — Austin (F. P. 12424) and Denton (F. P. 12423), Long in 1912. On Qicerctcs nigra. — Near Houston, Hedgcock, in igog. On Querctis phellos. — Near Houston and near Boeme, Hedgcock, in igog. On Quercus iexana. — Near Houston and near Boeme, Hedgcock. in igog (F. P. 762). On Qtiercus veluiina. — Near Boeme, Hedgcock, in igog. On Qtiercus virginiana. — Near Houston and near Boeme, Hedgcock, in igog (F. P. 320). Colorado: On Popultis tremuloides. — Steamboat Springs, Hedgcock, in 1912 (F. P. 3894). On Quercus gambelii. — Square Top Mountain (vSan Juan National Forest ; F. P. 9229); near Mancos (Montezuma National Forest) ; southeast of Delta (Uncompahgre National Forest); Hedgcock, in 1912. New Mexico: On Quercus arizonica. — Pecos, Long, in 1913 (F. P. 12412). On Quercus emoryii. — MogoUon Mountains Hedgcock, in 1911. Oct. 15, 1914 Heart-Rot of Oaks and Poplars 75 On Que reus gambe lit. — Sandia Mountains (Manzano National Forest), Hedgcock, in 1906 (F. P. 126,230); in 1908 (F. P. 270, 551-553, 558); near Pinos Altos (Gila National Forest), Hedgcock, in 1909 (F. P. 811, 812); in Alamo National Forest, L. L. Janes, 1909 (F. P. 1 142); MogoUon Mountains, Bear Creek Canyon, and Trout Creek (Gila National Forest), Hedgcock and Long, in 1911 (F. P. 9837); Cloudcroft, Long, in 1911 (F. P. 12015); Pecos, Long, in 1912 (F. P. 12426). On Quercus oblongifolia. — Near Mogollon, Hedgcock, in 1911. Arizona: On Quercus arizonica. — ChiricahuaMountains, H. D. BmRALL, in 1908; nearSedona (Coconino National Forest), Hedgcock, in 1910: Santa Catalina Mountains, Hedg- cock, in igii. On Quercus chrysolepis. — Sedona, Hedgcock, in igio. On Quercus emoryii. — Chiricahua Moxmtains, Bl-rrall, in 1908: Groom Creek and CrownKing(PrescottNational Forest), Hedgcock, in 1910; Santa Catalina Mountains, Hedgcock, in 1911. On Quercus gambelii. — Groom Creek (F. P. 4557), Crown King (F. P. 4877), Sedona (F. P. 4941), and near Flagstaff, Hedgcock, in 1910; Santa Catalina Mountains, Hedg- cock and Long, in 191 1 (P. P. 9801). On Quercus hypoleuca. — Near Pinos Altos, Hedgcock, in 1909. On Quercus oblongifolia. — Groom Creek and Crown King (F. P. 4876) and near Sedona, Hedgcock, in 1911; Santa Catalina Moimtaias, Hedgcock, in 1911. On Quercus toumeyi. — Santa Catalina Mountains, Hedgcock, in 1911. California: On Quercus californica. — Scott River Valley (Klamath National Forest), Hedgcock, in 1909 (F. P. 1886); near Mirror Lake (Yosemite Park), Clarks (Plumas National Forest), North Fork, and O'Neals (Sierra National Forest), Meinecke, Ln 1910; near El Portal and Yosemite (Yosemite Park), Hedgcock and Meinecke, in 1910 (F. P. 4794); near Kennett, Hedgcock and Meinecke, in 1911 (F. P. 9649). On Quercus chrysolepis. — El Portal and Yosemite, Hedgcock and Meinecke, in 1910; North Fork (Sierra National Forest), Meinecke, in 1910. On Quercus garryana. — Scotts River and Mount Marble (Klamath National Forest), Hedgcock, in 1910 (F. P. 1847). On Quercus lobata. — Stanford University, C. F. Baker, in 1902 (Herb. Univ. of Wis- consin); near Chico, Hedgcock, in 1909; Dobe and Italian Bar (Sierra National Forest), Meinecke, in 1910. On Quercus wislizeni. — El Portal, Yosemite, and near Raymond, Hedgcock, in 1910; near Kennett, Hedgcock, in 1911. On Quercus sp. — Crane Valley (Sierra National Forest), and El Portal, Meinecke, in 1910. Oregon; On Quercus garryana. — Near Mount Hood (Oregon National Forest) and Rogue River Valley, Siskyou National Forest), Hedgcock, in 1909 (F. P. 1717); near Medford, Hedgcock, in 1911 (F. P. 9611). On Quercus californica. — Rogue River Valley, Hedgcock, in 1909. From the foregoing data the following trees are attacked by the disease caused by Polyporus dryophilus: Quercus alba, 0. arizonica, Q. californica, Q. chrysolepis, Q. coccinea, Q. digitata, Q. emoryii, Q. gambelii, Q. garryana, Q. hypoleuca, O. imbricaria, Q. ilex, Q. lobata, Q. lyrata, Q. viacrocarpa, Q. marilandica, Q. michauxii, Q. minor, Q. nigra, 0. oblongifolia, Q. palus- tris, 0. phellos, 0. prinoides, 0. prinus, Q. robur, 0. texana, 0. lelutina, Q. virginiana, and Q. wislizeni; Populus grandidentata, P. tremula, and P. iremuloides ; Beiula ( ?) sp. , and Pinus ( ?) sp. 60300°— 14 6 76 Journal of Agricultural Research voi. iii, no. i CONTROL OF THE PIPED ROT OF POLYPORUS DRYOPHILUS The piped rot caused by Polyporus dryophilus is one of several impor- tant heart-rots of oaks in the United States. Suggestions made for its control will apply more or less to all of these. So long as oak trees are allowed to stand long past maturity in our wood lots and forests, heart-rots will continue to be common. The practice of leaving uncut in a lumbered area all the badly diseased trees, especially those with heart-rot, is radically wrong from the standpoint of proper forest sani- tation, for this practice enables heart-rotting fungi to maintain them- selves in the forest while the new generation of trees slowly develops and attains the age at which they form heartwood and thus become suscep- tible to the attacks of heart-rotting fungi. Trees diseased with heart- rot ought not to be left for seed trees wherever it is possible to leave healthy ones for this purpose. In hardwood forests it is often not neces- sary to leave seed trees, owing to the abundant sprout production, and the presence of young trees intermingled among the more mature ones. Trees in the wood lot should be inspected annually, and all trees evi- dently rotted at the heart should be removed. If the trunk of a tree diseased with heart-rot is struck with an axe, it does not ring with a clear sound. The presence of the fruiting body of Polyporus dryophilus on a tree also is evidence of the presence of the piped rot and of the necessity of removing the tree. Sporophores on trees should be removed whenever found. In large forested areas it is not possible to personally inspect the trees every year nor to search the forests annually for sporophores, although the present prices of good white-oak lumber nearly justify the expense necessary in a system of careful forest sanitation. It will certainly pay in lumbering tracts of oak and other valuable hardwoods to cut out all unsound or diseased trees, remove the parts that can be used, and bum the remainder. Many trees under the present methods of lumbering are left standing because they are decayed in the trunk near the butt. If cut do\\Ti, these trees would be found to contain enough lumber to pay for the cost of operation. Such a procedure will lead to a better and closer utilization of our gradually decreasing supply of hardwood lumber, especially of white oak. The destruction of all trees that are rotted in the heart in timber sales will be a step far in the direction of control for these diseases of timber. A new forest grown on areas lumbered with due regard to sanitation will be certain to be nearly free from heart-rot. Oct. IS. I9I4 Heart-Rot oj Oaks and Poplars 77 LITERATURE CITED Hartig, Robert. 1878. Die Zersetzungserscheinungen des Holzes der Nadelholzbaume und der Eiche ... p. 124-12S, pi. 17. Berlin. Meinecke, E. p. M. 1914. Forest Tree Diseases Common in California and Nevada. A Manual for Field Use. p. 48-49. Washington, D. C. Pub. by U. S. Dept. Agr., Forest Service. ScHRENK, Hermann von, and Spaulding, Perley. 1909. Diseases of deciduous forest trees. U. S. Dept. Agr., Bur. Plant Indus. Bui. 149, p. 39-40, pi. 5, fig. 1-2. SUDWORTH, G. B. 1898. Check list of the forest trees of the United States, their names and ranges. U. S. Dept. Agr., Bur. Forestry, Bui. 17, 144 p. PLATE VIH Fig. I. — Qucrciis alba: Crescent-shaped "soak," the initial stage of the piped rot produced by Polyporus dryophilus; from Arkansas. Fig. 2. — Quercus alba: A radial view of the rot in a limb, showing delignification; from Arkansas. Fig. 3. — Qticrcus oblongifolia: A radial view of rot, showing delignification; from Arizona. Fig. 4. — Quercus alba: A final stage of the rot, radial view, with more complete delignification; from Arkansas. Fig. 5. — Quercus alba: A tangential view of the rot, showing delignification in pockets; from Arkansas. Fig. 6. — Quercus alba: An end view showing a cross section from the same tree as the preceding; from Arkansas. Fig. 7. — Quercus sp.: A section of oak from Von Tubeuf, sent to the junior writer as a specimen of the rot caused by Polyporus dryadeus in Europe. Fig. 8. — Quercus sp. : The reverse side of tlie specimen sho«-n in the preceding. Fig. 9. — Quercus sp. : A section of oak from Europe, obtained by Von Schrenk, with a piped rot similar to that of Polyporus dryophilus. (78) Heart-Rot of Oaks and Poplars Plate VIII Journal of Agricultural Research Vol. III. No. 1 Heart-Rot of Oaks and Poplars Plate IX Journal of Agricultural Research PLATE IX Fig. I. — A sporophore of Polyporus dryof>hilus, tuberous form on Ouercus gambelii; from Arizona. Fig. 2. — Sectional view of a sporophore of Polyporus dryophilus on Quercus gambelii, showing the hard granular core with whitish mycelial strands; also the pore layer; from New Mexico. Fig. 3. — A sporophore of Polyporus dryophilus on Ouercus californica, showing the upper surface with a faint zonation; from California. Fig. 4. — A section through a sporophore of Polyporus dryophilus on Quercus garryana, showing the structiu-e of the hard granular core ; from California. Fig. 5. — A front view, showing the margin of the same sporophore as in figure 3, representing the ungulate form. Fig. 6. — A view of the pore surface of an applanate sporophore of Polyporus dryophi- lus on Ouercus alba; from Arkansas. PLATE X Fig. I. — A sporophore of Polyporus dryophilus, front view showing the margin, on Populus ircmuloides; from Colorado. Fig. 2. — A second sporophore from the same tree as figure i, showing an imbricated form. Fig. 3. — A view of the upper surface of a sporophore of Polyporus rheades on Populus Ircmula; from Stockholm, Sweden. Fig. 4. — A sectional view of a sporophore of Polyporus corruscans on Quercus; from Upsala, Sweden. Fig. 5. — A side view of an imbricate sporophore of Polyporus dryophilus, applanate form on Populus ircmuloides; from Colorado. Fig. 6. — A sectional view of the same sporophore as in the preceding figure, showing the hard granular core and whitish mycelial strands. Fig. 7. — A view of the upper surface of an applanate sporophore of Polyporus dryophilus on Quercus alba; from Arkansas. Fig. 8. — The pore surface of a sporophore of Polyporus dryophilus on Populus tremu- loides; from Colorado. Heart-Rot of Oaks and Poplars Plate X Journal of Agricultural Research Vol. III. No. I PRELIMINARY AND MINOR PAPERS DECOMPOSITION OF SOIL CARBONATES By W. H. MacIntire, Soil Chemist, Tennessee Agricultural Experiment Station Investigations recently conducted at the University of Tennessee Agricultural Experiment Station have led to the discovery that the composition of soils is such as to make inhibitory any long-continued occurrence therein of magnesium carbonate, and the conclusion has been drawn that magnesium carbonate does not exist as a solid mineral in our humid soils. The research has demonstrated that the affinity of magnesia for silica s such that soils long since alkaline from excessive treatments of calcium carbonate are able to dissipate the CO, of magnesium carbonate under sterile, moist conditions. It has been further demonstrated that the affinity of magnesia for silica is so great that precipitated magnesium carbonate is extensively decomposed by pure SiO, and also by the closely allied compound titanium oxid, which occurs almost universally to an appreciable extent in soils. Analyses to determine residual carbonates at the end of one year established the fact that precipitated magnesium carbonate in amounts chemically equivalent to 16,070 pounds of CaCOa per acre in excess of the quantity indicated by the Veitch method as necessary to correct acidity had been entirely decomposed by each of three distinct types of unleached soil. A loam, a sandy loam, and a silty loam were used in the study. In substantiating the work the loam soil was subjected to eight check treatments in field rim experiments. In each of these eight instances precipitated MgCOj equivalent to over 1 5 tons of a good grade of limestone per 2,000,000 pounds of soil had entirely disappeared at the end of eight weeks, when the first analyses were made for residual car- bonates. No drainage took place during this 8-week interval. Comparisons between residual carbonates from limestone and dolo- mite treatments showed at the end of 9 months' exposure to weather 22,000 pounds of CaC03 per 2,000,000 pounds of soil for the limestone treatment as compared to 1 1 ,000 pounds of CaCOj as a residue from the dolomite. The investigations have also determined that the absence of carbo- nates subsequent to applications of magnesium oxid has been erro- neously attributed to persistent causticity of the magnesia, which is shown to be very readily converted to the carbonate, while this in turn is decomposed by siliceous substances. The use of CaCOj as a check has shown that the affinity of lime for silica is far greater than has been supposed, and the work has demon- strated that the lime-silica reaction in soils is an important factor in the consen-ation of lime applied in practice. While the lime-silica reaction Journal of Aericultural Research. Vol. III. No. i Dept. of Agriculture. Washington, D. C. Oct. 15. 1914 Teen.— I (79) 8o Journal 0} Agricultural Research vo\. in. No. 1 does not approach the magnesia-sihca reaction in rapidity, it is shown by field data that the hme-siHca reaction continues long after the attain- ing of alkalinity and that the reaction is extensive. Toxicity due to excessive treatments of magnesium carbonate after its conversion to silicates was demonstrated by plant growth. The progress of the work of the writer and associates is reported in detail in Tennessee Experiment Station Bulletin 107. A FUNGOUS DISEASE OF HEMP By Vera K. Charles, Assislant Mycologist, and Anna E. Jenkins, Scientific Assistant, Office of Pathological Collections and Inspection Work, Bureau of Plant Industry In September, 191 3, the attention of the Office of Pathological Col- lections and Inspection Work was called to a fungous disease which had attacked a variety of hemp {Cannabis saliva) grown for experimental purposes by Mr. L. H. Dewey, Botanist in Charge of Fiber-Plant Investi- gations. Although the disease did not make its appearance until the plants were almost full grown, it was very rapid in its action, only about two weeks having elapsed between the time that the disease was first noted and the death of many of the plants. One of the early symptoms of the disease was the wilting and drooping of the leaves. The foliage turned brown and finally died, but remained attached to the plant longer than in the normal condition. In nearly all instances the fungus first attacked the outer ends of some of the upper, though rarely the highest, branches of the plant. In some cases the branches above and below the diseased area remained uninjured for some time. It was observed that the disease spread more above than below, but that the affection of the plant became general in about two weeks. Although the disease ap- peared to attack the outer ends of the branches first, the main stem below the base of the diseased branch became bleached and afterwards dark- ened by the formation of the perithecia of the fungus (PI. XI).' The hemp was grown from seed originally introduced from China, having been grown for experimental purposes during a period of 10 years. Its cultivation had been generally successful, and until the season of 1913 no difficulty had been experienced from fungus attack. All of the plants in the plots in which the disease was most serious were from the seed of one single selected plant, the third best of the crop of 191 2. This plant showed no evidence of disease and was remarkable for its purple-colored foliage. Selections had been continuous for 10 generations without any admixture of other strains. Three or four plants of this plot which were especially precocious were marked as soon as it was observed that they were pistillate, and each one of these plants was attacked by the disease. So general was the attack that among the 135 pistillate plants of this plot 128 were destroyed by the fungus, representing a loss of about 95 per cent of this plot. Later the disease appeared in a larger plot of 320 and in less than four weeks 66^ per cent of the plants had been attacked. A microscopic examination of the first diseased material collected on September 12, 1913, revealed the presence of small, black pycnidia, containing minute, hyaline spores on branched conidiophores. These characters, together with the absence of stroma, placed the fungus in the genus Dendrophoma. (Fig. i, E and F.) This appears to be the first occurrence of the fungus in America. A second examination of the dis- eased hemp about three weeks later showed pycnidia containing spores ' Most of the field observations were made by Mr. L. H. Dewey, who mentions the occurrence of this disease in an article entitled "Hemp" in the Yearbook, U. S. Dept, of Agriculture, for 1913, p. 283-346, fig. 17-21, pi. 40-46. 1914. Journal of ACTicultural Research, Vol. Ill, No. 1 Deiit. of Agriculture. Washington, D. C. Oct. 15, 1914 G-33 (81) 82 Journal of Agricultural Research Vol. III. No. I characteristic of the genus Macrophoma (fig. i, D). At the same date an immature ascomycete was observed on material which had been allowed to remain on the ground. The final collection on November 3 showed an ascomycetous fungus present in large amounts on the hemp which had been spread for dewretting, while the two other spore forms were absent, or present only in negligible amounts, having matured before the development of the ascospores. The asci were borne in perithecia similar in appearance to the pycnidia of the two other forms (fig. 1, B). Fig. I. — ^Microscopic characters of the hemp fungus Boiryosphaeria marconii. A, Sketch of a section of stroma from culture, showing developing perithecia: a. microconidial stage, b, ascosporic stage, X 840. B. An ascus with ascospores, X 840. C, Ascospores, X 840. D, Macroconidia. X 840. E, Conidiophores of the Dendrophoma stage, X 1920. F, Microconidia. X i9?o. (Drawing by J. Marion Shull.) The spores were hyaline to slightly colored, nonseptate, and fusoid (fig. I, C). A probable connection between these three forms suggested itself to the authors, and cultures were started to prove, if possible, that these stages are different phases in the life history of one fungus. The spores of the Dendrophoma form are designated as microconidia and those of the Macrophoma stage as macroconidia. Cultures were made on various media, but as the fungus developed luxuriantly and rapidly upon corn meal, that medium was adopted for Oct. IS, 1914 Fungous Disease of Hemp 83 the cultural work. The fungus developed in the same sequence as in nature, the Dendrophoma stage appearing first, regardless as to whether the cultures were made from microconidia, macroconidia, or ascospores. Sections of the pycnidia made at a later date demonstrated that the development of the macroconidia followed the microconidia in the same pycnidium. In sections made at a still later date asci were found develop- ing in the same locule with the mature macroconidia. The three spore forms of the fungus as developed in culture agreed perfectly in character with those found in nature. The variations observed in size and shape of macroconidia and shape of the asci were also exhibited by the fungus in nature. The one notable difference, however, was in the stronger development of stroma in the cultures. Since the Dendrophoma spores and the Macrophoma spores developed in the same pycnidia, the macro- conidia and ascospores in the same perithecia, and all three forms in the same stroma, it is definitely proved that these three forms represent the different stages in the life history of one fungus (fig. 1, A). From the critical microscopical study of the Dendrophoma stage of this fungus in nature and in culture it is shown to be morphologically identical with a specimen of Dendrophoma marconii described by Cavara in Italy in 1887.' No stroma is produced as the fungus occurs on the host, although a well-developed stroma is produced in culture. This stromatic development is suggestive of the genus Dothiorella, but it is not a constant character, and as the fungus agrees so closely with Cavara's description of Dendrophoma on hemp,^ the authors consider these two forms to be identical. During the course of the microscopic study of Dr. Cavara's material a second type of spore was found which corresponded exactly with the macrospores discussed in this paper. No mention of these was made in Dr. Cavara's paper, however, and the writers were unable to determine whether or not they had been observed by him. Among the few fungi described on hemp and related genera no species were found possessing the characters of the perfect stage of the fungus here discussed. In 1831 a fungus was observed by Schweinitz on hemp, and was called by him Sphaeria cannabis Schw.^ This species is of histor- ical interest only, for the description is too meager to be of any taxonomic value. The characters of the ascosporic stage place the fungus in the genus Botryosphacria as defined by Saccardo.* As the imperfect stage of this fungus is considered identical with the first described form, Dendro- phoma marcnnii Cav., the specific name is retained and the fungus is designated Bolryosphaeriamarconii (Cav.) Charles and Jenkins. Botryosphaeria marconii (Cav.) Charles and Jenkins. Perithecia globose, perforate, diseased area pale olive buff to gray, 140 to 160 /< in diameter; basidia bearing microconidia mostly dichotomously branched, septate, hyaline; microconidia polymorphic, ovate, elliptical, or terete, continuous, hyaline, 4 to 5K by i}^ to 2/(; macroconidia fusiform or ellipsoid, continuous, hyaline to glau- cous, 16 to 18 by 5 to 6 11; basidia of macroconidia slender, generally 12 to 15 /< in length; asci clavate, 8-spored, 80 to 90 by 13 to 15 n; paraphyses filiform; spores fusoid, hyaline to pale light grape green, 16 to 18 by 7' to 8 ft. Microconidia, macro- conidia, and asci produced in the same perithecium. On Cannabis saliva. ' Briosi, Giovanni, and Cavara, Fridiano. I Funghi Parassiti delle Piante Coltivate od Utili. no 20. Pavia. 1887. ExitccaUF. 2 Cavara, Fridiano. Appunti di patologia vegetale (alcuni (unghi parassiti di piante coltivate.) In Att i 1st. Bot. Univ. Pavia, Is. 2I. v. i. p. 426. iSSS. 'Schweinitz. L. D. von. Synopsis (ungorum in America boreali media degentium secundum observa- tlones. In Trans. Amer. Phil. Soc. n. s.. v. 4, p. 222. no. 1741. 1834. Saccardo, P. A. Sylloge Fungomm ... v. 2. p. 432. Patavii, 1883. * Saccardo, P. A. Sylloge Fungoruju . . . v. i, p. 456. Patavii, 1882. 84 Journal of Agricultural Research voi. ni. No. 1 In view of the serious nature of the disease and its sudden appearance in America it has seemed best to present this preliminary paper. The true parasitic nature of the fungus was evident from its effect on the growing plants, but its parasitism was further demonstrated by the suc- cessful isolation of the fungus from the interior tissue of thoroughly disin- fected stems. Owing to limited time and opportunity for extensive field observations, many questions relating to the pathological phase of the subject remain unsolved. Problems pertaining to the method of infec- tion by the fungus, its manner of dissemination, and control measures for the disease are still subjects of investigation by the Office of Pathological Collections and Inspection Work. PLATE XI A hemp plant, showing upper branches attacked by the fungus Bolryosphaeria marconii. Fungous Disease of Hemp Plate XI Journal of Agricultural Rnsearch Vol. III. No. 1 A MORE ACCURATE METHOD OF COMPARING FIRST- GENERATION MAIZE HYBRIDS WITH THEIR PARENTS By G. N. Collins, Botanist, Office of Acclimatization and Adaptation of Crop Plants, Bureau of Plant Industry INTRODUCTION That the crossing of two distinct varieties of maize usually results in an increase of vigor and larger yields in the first, or F,, generation has come to be generally recognized. The amount of the increase, how- ever, varies greatly in different hybrids, and in many cases the increase is not large enough to be determined by ordinary experimental methods, if it exists at all. So far as known, no case has been reported where a decrease below the mean vield of the parents has been adequately demonstrated. It is highly desirable to know the conditions under which significant in- creases occur, but thus far little light has been thrown on this important point. If really incompatible varieties exist, a study of their behavior in hybrid combinations should afford a favorable opportunity to learn something regarding the conditions necessary for large increases. One serious obstacle to learning the factors involved in the increased yields of first-generation hybrids is the difficulty of accurately comparing the vigor and yield of a hybrid with that of the parent varieties. Hybrids in maize are made either by hand pollination between indi- vidual plants or by planting in alternate rows the varieties to be hybrid- ized and removing the tassels from all the plants of one of the varieties. The customary method of comparing the behavior of a hybrid with its parents is to plant the hybrid seed in rows or blocks alternating with similar areas planted with the seed of the parents. If the series is re- peated a sufficient number of times, reliable averages may be obtained, but in actual practice the number of repetitions is usually limited by lack of seed or space. ^ In making a comparison between a hybrid and its parents where the hybrid is made by planting the varieties in alternate rows, the question arises as to what seed will best represent the parents. If the original seed of the parent varieties is used, it will be one year older than the hybrid seed, and the uncertain element of deterioration with age is in- troduced.- By saving the seed from the plants used as a source of pollen in making the hybrid, fresh seed of the male parent can be secured, but if fresh seed of the female parent is obtained, it must be grown at some distance from the place where the hybrid is made. To use seed grown under different conditions introduces an element of uncertainty that ^ In experiments with maize extendinK over a number of years in many different localities, we have found that with rows loo feet long and the series repeated lo times, it has seldom been possible to detect with assurance differences in yield of less than lo per cent. 2 For a discussion of this point, see Hartley. C. P., Brown. E. B.. Kyle, C, H.. and Zook, L. L. Cross- breeding com. U. S. Dept. Agr., Bur. Plant Indus. Bui. 21X, p. 13-10, 1912. One and two-year-old seed of selection No. iiga.the variety used as male parent in the Maryland experiments, occurred side by side 42 times. The average superiority of the new seed was 7 ± i.j per cent. Journal of Agricultural Research, Vol. Ill, No. i Dept. of Agriculture, Washington, D. C. Oct. 15, 1914 G-3S (85) 86 Jotirnal of Agricultural Research voi. in, No. i may be as serious as that incurred by the use of old seed. If the condi- tions of growth where the pure seed is produced are more favorable than those under which the hybrid seed is produced, natural selection will be less rigorous. There is also the possibility of direct effect of en- vironment on the yielding power of the seed and the possibility of new- place effect. A further disturbing factor lies in the differences between the indi- vidual plants that produce the hybrid seed and those producing the pure seed which is to represent the parental varieties. When seed from a large number of plants is used, these differences tend to counterbalance each other and give an average of value for practical purposes, but in- formation which might extend our knowledge regarding the nature and causes of the increase may be completely obscured by this method of averaging. Some of these difficulties can be avoided if the hybrid seed is obtained by hand pollination. By this means seed of both parent varieties of the same age as the hybrid seed and grown under similar conditions can be secured. Inaccuracies due to diversity among individual plants will, however, be increased, since the number of plants involved will necessa- rily be smaller, and, as before, differences in the behavior of individual crosses which might throw Ught on the nature of the increases will be masked if conclusions are based upon averages. To avoid this last diffi- culty, the individual hybrid ears may be kept separate and an ear-to-row method of making comparisons with the parents may be applied. Differences in the breeding value of individuals are now appreciated in the breeding of pure strains and have led to the adoption of the method of separating the offspring of individuals into progeny rows. The results here reported show a diversity among the hybrid ears that result from crossing different plants of the same parent varieties that is even greater than that usually found between pure seed ears of a single variety, and the evidence indicates that individual diversity in hybrids will be found as important as in pure varieties. In comparing an individual maize hybrid with its parents account must be taken of the fact that to behave normally maize must be cross-polli- nated, and to secure cross-pollinated seed of the parent varieties two plants of each variety must be used, but only one plant of each variety can be represented in an individual hybrid ear. To avoid in some measure these sources of inaccuracy the method followed in the experi- ments here described is suggested. DESCRIPTION OF METHOD To compare the behavior of two varieties, which may be called A and B, with that of a hybrid between them, two plants were selected in each variety, Ai and A2 in the one variety and Bi and B2 in the other variety. The following hand pollinations were made: A1XA2, A2XB1, B1XB2, and B2 X Ai . The result is two hybrid ears and one cross-pollinated ear of each variety. It is believed that the mean yield produced by seed from the two hybrid ears compared with the mean yield produced by seed from the two pure seed ears gives a fair measure of the effects of hybridization. By making two hybrids involving all the plants used in producing the pure seed ears individual differences that affect the yielding power of the pure seed ears are similarly represented in the hybrids. Thus, in both the Oct. IS. 1914 First-Generation Maize Hybrids 87 parents and the hybrids the average yield represents the mean yielding power of the four parent plants, the only difference being the way in which the individuals are combined. To secure the most accurate comparison of the yield of the four ears, one seed from each of the ears was planted in each hill. The different kinds were identified by their relative position in the hill. To place the seeds accurately, a board 4 inches square was provided with a small, pointed peg 2 inches long at each corner. These pegs were forced into the soil at each hill, making four holes, one for each of the four kinds, only one seed being planted in a hole. The board was always placed with two sides of the board parallel to the row. It was necessary to exercise extreme care in dropping the seeds to avoid changing the position of the kinds. The best way to obviate mistakes of this kind is to make all the holes of a row in advance and to go down the row with one kind of seed at a time. At harvest time the seed produced by each plant was weighed and recorded separately. All hills that lacked one or more plants were excluded and the comparison confined to hills in which all four kinds were represented. The method of handling the yields was to determine the mean yield of the four kinds in each hill and to state the yield of each of the four plants as a percentage of the mean of the hill in which it grew. The percentage standing of each kind in all the hills was then averaged to secure the final expression of the relative behavior of the four kinds. This method of comparison is similar to the ingenious plan originated by C. H. Kyle,' for use in ear-to-row breeding. Kyle's method is to plant each of the ears to be tested in a separate row and in each hill to plant one seed of a standard, or check, ear with which all ears are compared. Since comparative and not absolute yields are desired in the study of hybrids and with only four kinds to compare, the introduction of a check in the present experiment would have increased the space occupied by the experiment without lessening the experimental error. APPLICATIONS OF METHOD The hybrid tested by this method was a cross between the Egyptian, a white sweet com, and the Voorhees Red, a related sweet variety with red aleurone.^ The two hybrids secured in accordance with the foregoing method were designated PhgS and Phgj. The use of the Voorhees Red variety as one of the parents made the comparison unusually difficult. This variety produces a considerable percentage of albino seedlings, and since no albino seedlings reach maturity, the result was a large number of hills with less than the full complement of plants. Eighty-four hills were planted, but only fifty-eight matured plants of all four kinds. The comparative yield of the four kinds is given in Table I. To illus- trate the meaning of these determinations, let us take the yield of the Egyptian variety. The number 112,8 indicates that the yield of 58 Egyptian plants averaged 112.8 per cent of the mean yield per plant of all four kinds— that is, 12.8 per cent above the mean. * Kyle, C. H. Directions to cooperative com breeders. U. S. Dept. Agr., Bur. Plant Indus., B. P. I. — 564, 10 p.. 1910. *The strain of EgMitian com used in this experiment was from commercial seed secured from J. M. Thorhuni & Co. iu 191 1. Tlie original source of the Voorhees Red was an ear kindly supplied by Prof. Byron D. Halsted, of the New Jersey State Agricultural Experiment Station, in 1907. 88 Journal of Agricultural Research voi. m. No. i The mean of the two parents is S4.2±3.o per cent of the general yield. The mean of the two hybrids is ii5-9±3-3 per cent. The mean yield of the hybrids is thus 31.7 + 4.5 percent higher than the mean of the parents, and this increase is ascribed to the effects of crossing. Table I. — Yield and height of the Egyptian and Voorhees Red varieties of sweet corn and Izt'o hybrids between them [Detenninations expressed as average percentages of the mean of the four kinds.l Variety of com. Yield. Height. Egyptian Voorhees Red. Hybrid Phg6. . Hybrid Phg;.. Per cent, 112. 8±4. 6 55. 6±4. o 89. o±5. I 142. 8±4.3 Per cent. III-3±I.0 84. o± . 9 100. o±i. 2 103. 6± I. I A striking feature of the results obtained is the difference between the yield of the two hybrid ears, which amounts to 53.8 ±6.7 per cent. Had the ear Ph96 alone been taken as representing a hybrid between these varieties, the hybrid would have exceeded the average of the parents by only 4.8 per cent, a difference upon which no reliance could be placed. If, on the other hand, the ear Ph97 had been taken, the difference in favor of the hybrid would have appeared as 58.6 per cent. The relative height of the four kinds was determined in the same manner as the yield — that is, the height of each plant was compared with the mean height of all the plants of the hill in which it grew, the latter being taken as 100. The average heights expressed in this way are given in column 2, Table I. The average height of the parents is 97.6 ±0.7 per cent of the general mean. That of the hybrids is ioi.8±o.8 per cent. The difference is 4.2 ±1.1 per cent. There is, then, a distinct increase in the height of plants as a result of crossing, but the increase is much less than the increase in yield, and the difference between the two hybrids is nmch less than was the case with the yield. It has usually been found that the increase that follows crossing affects the vegetative characters even more than the reproductive. If height be taken as an index of vegetative vigor, the reverse would seem to be true in the present cross. Increased vegetative vigor may have resulted in an increase of the branches rather than of the main stalk. To definitely settle this point, it would have been necessary to weigh or measure all of the suckers. This was not done, but the number of suckers was recorded for each of the kinds, and the difference, though small, indicates that a part of the increased vegetative vigor of the hybrids was expressed in the production of suckers. A total of 18 suckers was produced in the two pure-seed rows, while 35 were produced in the two hybrid rows. The association between vegetative vigor and yield is further shown by the fact that the hybrid Ph97 exceeded the hybrid Ph96 both in yield and in the produc- tion of suckers. It should be borne in mind, however, that an increased yield and an increased production of branches may not always be thus associated. It is to be expected that under some conditions excessive Oct. IS, 1914 First-Generation Maize Hybrids 89 branching may result in a decreased yield. Hence, if some hybrids show reduced yields, this fact alone should not be taken as proving an excep- tion to the general rule that the first generation of a hybrid shows increased vigor. The method of comparison here used brings the plants into close com- petition, and it may be urged that the differences between the kinds are as a result unduly accentuated. With a view to detecting a possible effect of competition, the yield of the plants in hills with four plants was compared with plants of the same varieties in hills with less than four plants. In P9, P19, and Phgy the j'ield per plant was slightly higher in the 4-plant hills than in the 3-plant hills. The differences were, however, insignificant. In Phge the yield of the plants from the 3-plant hills exceeded that from the 4-plant hills by 67 grams per plant. The number of 3-plant hills was so small, however, that little confidence should be placed in the difference, which was but three times the probable error. An attempt was made to secure a more accurate comparison by cor- recting for the differences in the yield of the different kinds, thus making it possible to compare the yield per plant of all the 4-plant hills with that of all the 3-plant hills. The average yield per plant in the 4-plant hills was 211 ±7 grams. The average yield of the 3-plant hills was 227 ±10. The difference of i6±i2 grams is therefore not significant. With such a large experimental error it is of course not impossible that the crowding of the plants has a tendency to reduce the yield, but if so the difference is too small to be measured by the means employed, If crowding operated to accentuate differences, it might also be expected to retard the date of flowering. The average number of days to flowering was, however, the same in the 4-plant hills and in the hills with less than four plants, being 72.4 days in both. Thus there is no evidence that the growing of the four kinds close together affects the relative yield of the kinds, and when ample space is provided between the hills, viz, 4 by 5 feet, as in this experiment, it is believed that this source of inac- curacy is insignificant. The conditions of the experiment here reported constitute a severe test of the method of comparison by individual hills. The kinds tested were very dissimilar, while the soil of the experiment was unusually uniform. The gain in accuracy secured by using the hill as the unit of comparison, instead of averaging the yield of all the plants of a kind, may be measured by a comparison of the standard deviations or the coefficient of variability observed when the yields are compared by the two methods. When the yield of each plant was compared with the average of all the plants of the same kind, the coefficient of variability was 5.42±o.i7. When the yield of each plant was compared with mean yield of the hill in which it grew the coefficient of variability was 5.o5±o.i3. There is, thus, a slight gain in accuracy, notwithstanding the exceptional uni- formity of the soil where the experiment was tried. With less uniform soil conditions the advantages secured by making the comparison on the basis of individual hills would increase. The dates when the first staminate flowers opened and when the first silks appeared were recorded for all the plants. The average mmiber of days from planting to flowering is shown in Table II. 60300°— 14 7 90 Journal oj Agricultural Research voL m. Xo. I Table II. — Azera^ Hmej'rom planting to fitrxering of varieties of ■muaze Vaikiy. Xnndjer ct Xcmfaer oc dzs3tofii5t . days to Szsc poftTi ' silks. Egyptian Voorliees Red. Hybrid Pfagb . Hybrid PI197 . 72- 1= 72-0:=; .2 72. 5± .2 73- 9==- 3 77-1= .3 73-5= -4 73-6= .4 With resp)ect to the apipearance of the staminate flowers the only significant difference is the slightly later flowering of the Egyptian variety. With respect to the appearance of silk', the Voorhees Red, the low- yielding variety, was distinctly later. The average time between the opening of staminate flowers and the appearance of silks was less than one day in the Egyptian and five days in the Voorhees Red variety. Both hybrids were intermediate, with 1.5 and i.i days, resj)ectively, between the average time of the apf)€arance of poUen and silks. A further comparison of the hybrids with their parents with respect to minor characters brings to Ught a number of striking differences. A comparison of the characters measured is made in Table III. Table III. — Comparison of minor ckaracters of maize hybrids -iritk their futrents Chancto. ! ^t»f" niaize. ' ToodlBB Hybrid Hyfarid Average o£ Average of Rfri maizSL Ph96. Ph97. pare:!£s. TilThi^*^ Hei^it cm. .2=36 =L.t I" =2.6 1=3 =3-4 293 =2.9 1^2 =1- 5 iJQ =2. 3 Xumber o< snckes - .zii.ai ■tl±. -32 -14= -03 .48= -oS . X60= .023 .313— .aarents. A necessary step in this direction is to develop a reliable method of measuring the effect of crossing, apart from other factors that influence }-ield. The development of satisfactorv- methods of compearing the yield of first -generation hybrids with that of their parents has been retarded by (i) a failure to fuUy appreciate the importance of individual diversity in hybrids, (2) the abnormal beha\-ior of seh'-pollinated maize plants, and (3) the diMculty of securing for comp>arison hybrids and j>arents with identical ancestry. It is believed that the method here described avoids these difiaculties and affords more accurate means of comparing first -generation maize hybrids mth their parents. The method is illustrated by an exp>eriment in crossing two varieties of sweet com in which it was found that the progeny from one hybrid ear yielded nearly double that of the other hybrid ear involved in the experi- ment. To have taken either ear alone would have led to entirely erro- neous conclusions regarding the increase secured as a result of crossing. The increase in \ield due to crossing as measured by the method here proposed was 31 per cent. ADDITIONAL COPIES OF THIS PUBIJCATION MAT BE PROCURED FROM THE SUPERINTENDENT OF DOCUMENTS GOVERNMENT PRINTTNG OFFICE ■WASHINGTON, D. C. AT 25 CENTS PER COPY SuBSCHiPTioN Per Year, 12 Numbers, 12.50 Vol. Ill NOVEIVIBER, 1914 No. 2 JOURNAL OP AGRICULTURAL RESEARCH CONTENTS Natxiral Revegetation of Range Lands Based upon Growth Reqturements and Life History of the Vegetation . . 03 ARTHUR W. SAMPSON Pecan Rosette 2^9 W. A, ORTON and FREDERICK V. RAND Nitrogenous SoU Constituent 175 EDMUiro C. SHOREY and E. H. WALTERS Apple Root Borer 279 F. E. BROOKS DEPARTMENT OP AGRICULTURE WASHINGTON, D.C. «>«MmaTON I OOVtRNMENT PRIBTINO OFFICE : I9U PUBI^ISHED BY AUTHORITY OP THE SECRETARY OF AGRICULTURE, WITH THE COOPERATION OF THE ASSOCIATION OF AMERICAN AGRICUL- TURAL COLLEGES AND EXPERIMENT STATIONS EDITORIAL COMMITTEE FOR THE DEPARTMENT FOR THE ASSOCIATION KARL F. KELLERMAN, Chairman Physiologist and Assistant Chief . Burenu of Plant Indiistry EDWIN W. ALLEN Assisiant Director, Office of JExPerirMnt Statiotts CHARLES L. MARLATT Assistant Chief. Bureau of Entomology RAYMOND PEARL Biolooist, Maine Aortcuiiural Experiment Station H. P. ARMSBY Director, Institute of Animai Nutrition, The Pennsylvania Stale College E. M. FREEMAN Botanist, Plant Pathologist, and Assistant Dean, Ajricuitufal Extvriment Station of the University of Minnesota All correspondence regarding articles from the Department of Agriculture ^ould be addressed to Karl F. Kellennan, Journal of Agricultural Research, Washington, D. C. All correspondence regarding articles from Experiment Stations shovdd be addressed to Raymond Pearl, Journal of Agricultural Research, Orono, Maine JOM£ OF AGMCETURAL RESEARCH DEPARTMENT OF AGRICULTURE Vol. Ill Washington, D. C, November i6, 1914 No. 2 NATURAL REVEGETATION OF RANGE LANDS BASED UPON GROWTH REQUIREMENTS AND LIFE HISTORY OF THE VEGETATION ■ By Arthur W. S.\mpson, Plant Ecologist, Forest Service INTRODUCTION Ideal range management would mean the utilization of the forage crop in a way to maintain the lands at their highest state of productiveness and at the same time afford the greatest possible returns to the stock industry. To maintain the maximum productivity, the annual herbage crop must be used in a manner which will not retard the growth or pre- vent the perpetuation of the most desirable forage species. On the other hand, if the stock industry is to receive the greatest possible returns at all times, the annual forage crop should be used when it is most needed and when the herbage is palatable and nutritious. It is obvious from this that the requirements of the vegetation and the requirements of the stock are to a great extent antagonistic. Hence, un- restricted grazing, without regard for the vegetation or the locality, eventually results in decreased productivity, and often in denudation. The decline in carr3dng capacity of our western grazing lands was brought about in part by injury due to trampling, but perhaps in greater part by premature grazing and overstocking. The growing herbage might be called a laboratory where plant nutrients are prepared, and the repeated removal of the foliage year after year during the fore part of the growing season means the destruction of this laboratory, which in turn means lack of nourishment for the vegetation, resulting in lowered vitality and an inability to produce seed. The easiest way to overcome the deteriorating effect of premature grazing and overstocking, as well as of trampling, would be, of course, to eliminate grazing entirely. Obviously, however, such a procedure would be impracticable from the standpoint of the stock industry. Since this is so, the best means of solving the problem in a scientific manner is to Journal of Aericultural Research, Vol. Ill, No. 3 Dept of Agriailturc. Washington. D. C. Nov. i6, 1914 F-3 (93) 94 Journal of Agricultural Research voi. iii. No. i approach it as similar problems in farm practice are approached — that is, (i) by a careful study of the vegetation making up the forage crop, (2) by a study of the natural factors upon which depends the success or failure of the forage crop and its perpetuation, and (3) by a study to find a method of grazing which will both fully utilize the forage and at the same time protect it from deterioration. Such studies were undertaken by the Forest Service in cooperation with the Bureau of Plant Industry during the spring of 1907 in the Wallowa Mountains of northeastern Oregon. While the intensive investigations were confined in the main to this one grazing region, the results have been applied elsewhere with success, notably in the Hayden National Forest in Wyoming. It is possible, of course, that the repro- ductive capacity of various forage plants may v&ry in different localities and also that there may be a difference in the behavior of plants on ranges grazed by sheep and those grazed by cattle and horses, any of which may affect the measure of success obtained by deferred grazing, but not the principles involved in the system. The purely experimental studies were continued throughout the seasons of 1907, 1908, 1909, and 1910 and were followed by a practical application of the principles evolved to range management on lands within the Wallowa National Forest. The system developed as a result of the studies — a combination of de- ferred and rotation grazing — is now being applied with minor variations to range lands throughout the National Forests, and promises to be of the greatest value in bringing about the efficient utilization of the forage re- sources. This article gives in full the data upon which the new system is based. The area where the intensive studies were carried on is first described. Following this are given the life histories of the important forage species, including growth requirements and the factors influencing the establish- ment of reproduction. This in turn is followed by a discussion of the relative merits of different systems of grazing. Finally there is presented a rational and economical grazing system based upon the requirements of the forage plants and of the stock industry TOPOGRAPHY AND SOIL The Wallowa National Forest, within which the studies were carried on, is a region of high mountains, very irregular and broken. From the Grande Ronde and Wallowa Valleys, which bound the forest at about 3,000 feet elevation, the mountains rise to from 6,000 to 9,500 feet. On the upper reaches of the numerous domes, above the limits of forest growth, snow often remains throughout the summer. In this group of high, snowy peaks, within a radius of about 3 miles, rise nearly all of the streams from which the stock of the region are watered. Nov. 15, 1914 Revegetation of Range Lands 95 The forest exhibits three principal rock formations: Basaltic, granitic, and limestone. These give rise to as many different soil types, which in turn very largely determine the character and density of the vegetation. The best and most luxuriant vegetation is found upon the basaltic soils, which cover the greater part of the region, the comparatively recent lava flows from which they originate having buried the original formationsi n some places to a depth of several hundred feet. They are porous and very friable, admit of about average percolation, and retain water well. The granite and limestone soils, on the other hand, are poorly decom- posed and lose moisture rapidly through percolation and evaporation. In consequence the vegetation is usually sparse, only the more drought- resistant plants being able to establish themselves. The limestones, mixed with shales, are the oldest and most restricted of the three forma- tions. The granites, which are of a later period, form the peaks, crests, and soils of the very highest mountains. CHARACTER AND DISTRIBUTION OF THE VEGETATION Within the great altitudinal range between the lower valleys and higher mountains of the region wide differences naturally exist in the physical conditions which govern plant growth, and therefore in the character and composition of the growth itself. On the other hand, physical and climatic conditions, and consequently the vegetation, are strikingly similar within certain altitudinal limits, making it possible to divide the region into four climatic zones. Following Merriam's classi- fication' these are: Transition zone (yellow-pine association) 3,000 to 4,500 feet. Canadian zone (lodgepole-pine association) 4.500 to 6,800 feet. Hudsonian zone (whitebark-pine association) 6,500 to 8,500 feet. Arctic-Alpine zone (Alpine-meadow association) above 8,000 feet. The altitudinal limits of these zones are not absolutely marked, since altitude does not wholly determine the character and composition of the vegetation. Hence, the limits given above should be considered only approximate for a given locality in this latitude. The Transition Zone. — The Transition zone contains a number of coniferous tree species, the most characteristic being western yellow pine {Pinus ponderosa). Toward the upper limits of the zone yellow pine g[ives way to Douglas fir (Pseudoisuga taxijolia) and lowland fir {Abies grandis). As a rtile, the timber is open, with considerable undergrowth of average palatability and nutritiousness, if grazed relatively early (PI. XII). Among the most characteristic and abundant herbaceous species is pine-grass {Calamagrostis rtihescens). Other species which furnish a large * Merriam, C. Hart. Life zones and crop zones of the United States. U. S. Dept. Agr., Div. BioL Survey, Bui. lo, 79 p., 1 map. 1898. 96 Journal of Agricultural Research voi. m. no. j part of the herbage on the lower, sparsely timbered lands are big bunch- grass (Agropyron spicahim), little bluegrass (Poa sandhergii), big blue- grass {Poa scabrella), and blue bunch-grass (Festuca arizonica). Ger- mination and growth begin in the most exposed situations during the first week in April, and early in May the vegetation shows everywhere throughout the zone. Stock are usually not admitted before May 15. Canadian Zone. — The Canadian zone is characterized by lodgepole pine {Pinus murrayana), the predominant tree of the region. In many places the timber is so dense that there is little or no undergrowth of vegetation. Only the most tolerant shrubs and herbs can exist in the subdued light under the heavy timber, and such lands are of practically no value for grazing. In other places extensive areas of lodgepole pine have been burned over. Sometimes reproduction is established promptly, but where fire has consumed most of the organic matter in the soil, the reestablishment of vegetation of any kind is slow. Among the forerun- ners in the invasion of permanent species, fireweed {Chamaenerion angustv- /ol-inm), a valuable sheep forage, and pearly everlasting (Anaphalis margaritacea) , a plant of no forage value, are most common. Both growth and grazing begin in this zone fully 20 days later and end two weeks earUer thaii in the Transition zone below. HuDSONiAN Zone. — The Hudsonian zone, in contrast with the lands immediately below it, is open in character, the timber growing sparingly and in clumps. The predominating vegetation consists of grasses inter- mixed with various other palatable plants, as shown in Plate XIII, figures 2 and 3. This zone probably covers a larger area than the two lower zones com- bined and supports most of the sheep permitted in the Wallowa Forest during the summer growing season. On account of the demands made upon this desirable range and because of the character of the forage, the Hudsonian zone has suffered more serious depletion than any other, and it was here that the most intensive study of revegetation was made. The trees of the Hudzonian zone, most of which extend to the nonnal timber line, are Alpine fir {Abies lasiocarpa), whitebark pine {Pinus albi- caulis), Engelmann spruce {Picea engelmanni) , and mountain hemlock {Tsuga mertensiana). Whitebark pine is the most characteristic species, and its altitudinal distribution is so clearly marked that one can be cer- tain wherever it is met that the conditions there are those of the Hud- sonian zone. The timber grows in small, dense clumps, precluding an undergrowth of any but the most tolerant species. Aside from the timber, vegetation is distinctly herbaceous and consists mainly of grasses and nongrasslike plants, commonly termed weeds. While a great many of the species are grazed to a limited extent at one time or another during the season, about 40 furnish 90 per cent of the Nov. i6, 1914 Revegetation of Range Lands 97 range forage, are: These, arranged in the order of their local forage value, Mountain bunch-grass (Festuca -viridula). Little bluegrass (Poa sandbergii). Short-awned brome-grass (Bromus mar- ginatus). Western porcupine, or needle grass (Siipa occidentalis) . Smooth wild rye (Elymus glaucus). Tufted hair-grass (Deschampsia caespitosa). Wild celery {Ligusiicum oreganum). Onion grass, or mountain bluegrass {Melica bella). Red bunch-grass (Agropyron flexuosum). Mountain wheat -grass {Agropyron viola- ceum). Yarrow, or wild tansy {Achillea lanulosa). Spiked trisetum {Trisetum spicatum.) Butterweed {Senecio triangularis). Coneflower (Rudbeckia occidentalis). Wild buckwheat {Polygonum phytolaccae- folium). Alpine timothy {Phleum alpinum). Horsemint (Agastache urticifolia). Wood rush (Juncoides glabraium). Nuttall willow {Salix nutallii). Fireweed {Ckamaenerion angusti/olium). Mountain dandelion {Agoseris glauca). Mountain onion {Allium validum). Little needle grass {Stipa minor). Wild onion {.Allium plaiyphy litem). Wild onion {Allium collinum). Tall swamp-grass {Carex exsiccata). False hellebore {Veratrum viride). Valerian {Valeriana siichensis) . Alpine redtop {Agrostis rossae). Blue beardtongue {Pentstemon procerus). Elk-grass {Carex geyeri). Skunkweed, or Jacob 's-ladder {Poletmy- nium humile). Sheep sedge {Carex illota). Reed-grass {Cinna latifolia). Woolly weed, or woolly hieracium (Hiera- cium cynoglossoides). Onion grass, or mountain bluegrass {Melica spectabilis). Wire sedge {Carex hoodii). Tall meadow grass {Panicularia nervata). Slender hair-grass (Deschampsia elongaia). Rush {Juncus confusus). White foxtail {Sitanion velutinum). Parry 's-rush {Juncus parryi). Rush (Juncus mertensianits). Rush (Juncus orihophyllus). Throughout the Hudsonian zone mountain bunch-grass {Fesluca viri- dula) is by far the most abundant plant and the most desirable to reveg- etate. The relish with which several of the other species are grazed and their similar altitudinal range and abundance made it very difiScult to determine which one ranks next in value, and the arrangement presented was not finally decided upon untU after the third year's investigation. Growth usually begins in the Hudsonian zone about the last week in June, and stock are given access to the lands early in July. Arctic-Alpine Zone. — The Arctic-Alpine, or timberless, zone is not only unfavorable to tree growth, but to grazing plants as well. As shown in Plate XIV, figure i, the zone is confined to the very highest crests and peaks, where the soil is shallow and poorly decomposed, the season of growth short, and nightly frosts common. Owing to the virtual lack of grazing in this region, it was not thought necessary to measure accurately the climatic factors. The species of the Hudsonian zone are for the most part entirely absent in the Arctic- Alpine. Among the characteristic alpine plants are cat's-foot {Eriogonum piperi), whitlow-wort {Draba aureola), Hoorebekia 98 Journal of Agricultural Research Vol. III. No. 2 lyalli, hulsea (Hulsea 7ianii), Alpine phacelia {Phacdia alpina), and false strawberry {Sihhaldia procumbens) . Even these species occur sparingly. Growth does not usu- , ally begin until well ■zi ^3 ~z -,5 ^"^t ^ UL ^^^ ! ZC ^^5^ ^2>. ^^^r- ■■ .'\ ■^2 t - t^*^ - ••-S::^ ^S ■■ ^1 ■•• ^ S ^i2>^ -2^^ • t.^ -< A i^X t -^ t-^ - ^^v -• ^3 -i V -I \-% t ^t ^- -.--■%■' ■It t^ ^^ ^^^^--^ ■- -^ 5=> ■>"* ^ > ' •• 7.^ • ^H ^ -*>. ■••^L.^v^ •• ^ / - --"^ ■ 't ^ -^'^ ■■ ^t ■■• ^ ^ '■•^c^^ ^ ^,- 5 t: ■i s .•".*• -- ^ ^^' '^Z ■Z'-^^ . ■ > / 1 \ 1 ' ? p SO a II into July and ceases about September i. Naturally in this zone any species which suc- ceeds in maturing via- ble seed must be vig- orous and able to de- velop in a short time. CLIMATE Records of tempera- ture, precipitation, and air humidity were kept in the Canadian, Transi- tion, and Hudson ian zones during the main grazing season, in order to determine what dif- ferences in the season of growth they exhibit. TEMPERATURii. — Fig- ure I shows the mean temperatures of the re- spective zones derived from the daily maxi- mum and minimum during the main grow- ing season of 1909. While the mean tem- peratures in the three zones diflFered widely, there is a close relation- ship between the daily fluctuations. In all cases the mean is lower, usually by several de- grees, in the Hudsonian zone than in the two lower ones, while the highest naturally comes in the Transition zone. The extremes from which the mean temperature was obtained show that Nov. i6. 1914 Revegetation of Range Lands 99 the variation in the maximum temperatures in the three zones is fully as great as in the case of the minimum. During the month of July, for ex- ample, the maximum temperature in the Hudsonian zone was 84°; in the Canadian, 90°; and in the Transition, 104° F. The highest temperature in the three zones during the entire growing season was 91°, 97°, and 105° F., respectively. Precipitation. — The higher temperatures characteristic of the less ele- vated lands are associated in the region of the study with a minimum precipitation. In the valley surrounding the Wallowa Mountains — that is, in the Transition zone — at an elevation of 3,600 feet, the annual pre- cipitation is about 17 inches, the greater part coming in the spring, autumn, and winter. As a result, the vegetation often sufifers for lack of moisture. In the Hudsonian zone, on the other hand, the larger amount of precipitation received is ample, and, with the exception of seedling plants, the vegetation is not affected by drought. Figure 2 shows that the Transition and Canadian zones received 51.6 and 26 per cent less rainfall, respectively, than the Hudsonian. While PLACCOFRCADING TRANSITION ZON€ CANADIAN ZONE HUDSONIAN ZONE INCHES or PRECIPITATION J 63 s 51 7 i-^ 1 1 1 Fig. 2. — DiaKram showing the total precipitation in the Transition, Canadian, and Hudsonian grazing zones during July, August, and September, 1909, inclusive. the actual amount of precipitation during the growing season of 1909 was somewhat above normal, the ratio between the amount which fell in the different zones is similar to that of an average year. The greater amount of precipitation at the higher altitudes, together with the com- paratively late date at which growth begins, accounts for the continuous development of the forage. Comparative Air Humidity. — Since the Hudsonian zone has a rela- tively lower air temperature, a greater amount of precipitation, and a more humid soil than the lower zones, it would be natural to expect that the relative air humidity would be lower and transpiration less severe than in the lower areas. Figure 3, which gives the daily variations in air humidity derived from evaporation readings, shows this to be the case. It was unexpectedly found that in the Hudsonian zone the evaporation was greater than in the Canadian zone immediately below. While the temperature and even the relative air humiditv, a.-^ computed from psychrometer readings, were lower in the Hudsonian zone, the dense timber of the Canadian zone so interrupted the air currents as to materi- lOO Journal of Agricultural Research Vol. Ill, No. 1 ally diminish the evaporation there. In comparing the three zones, how- ever, the evaporation in the Hudsonian and Canadian zones was found to be, respectively, 13 and 35.6 percent less than that in the Transition zone. Owing to the relative great evaporation in the latter region, light showers, espe- cially those which fall early in the day and are succeeded by a clear sky, are soon evaporated and so are of little value to plant growth. During a period of drought, other things being equal, a plant could not live nearly as long in the lower as in the higher elevations, on account of the excessive transpiration in the former locality. To sum up the conditions peculiar to the Hudsonian zone as compared with the lower grazing type, the temperature is lower, the precipitation heavier, and the transpiration less. It is therefore easy to see that the relatively slow- growing and warmth-requiring vegeta- tion in the lower lands could not thrive in the cooler and shorter growing season typical of the Hudsonian zone. Hence, only the most plastic and adaptable species of the Transition zone occur in the Hudsonian, and then only on the warmest and most exposed south slopes. LIFE HISTORY OF THE FORAGE PLANTS The growth requirements of range plants can best be determined by a study of individual species and the factors with which they have to con- tend from the time that they begin growth in the spring through the various stages of development to seed maturity, and then on to the permanent establishment and seed- bearing stage of the vegetation i)ro- duced from seed of the original plants. The essential features of this almost double life cycle are: (i) Inception of growth, (2) flower-stalk Nov. i6, 1914 Revegetation of Range Lands loi production, (3) development and maturity of seed, (4) viability of the seed crop, and (5) establishment of reproduction. Owing to the relatively unfavorable conditions for plant growth and the demands made upon them for summer range, the high mountain lands are the ones most in need of revegetation. For this reason the life history of the forage plants and the factors affecting their activities were studied more intensively in the Hudsonian zone than elsewhere. INCEPTION OF GROWTH The time at whicn growth begins in the spring varies widely in the different zones. Thus, in the heart of the Hudsonian zone growth begins ordinarily about five weeks later than in the Canadian zone immediately below and seven weeks later than in the Transition zone. Even in the same zone the beginning of the growth period may vary by as much as ID days from year to year, depending chiefly upon the climatic condi- tions during the spring, especially in May and June, but also to a certain extent upon the amount of snow accumulated during the winter. One year the early season may be charactenzed by warm, sunny days, and as a consequence the snow cover, especially on the more exposed situa- tions, may disappear as early as June 20, though this is rather exceptional. In another year the snowfall in May and June may be as heavy as at any time during the winter. North and east exposures are always later in responding to growth than south and west aspects of the same eleva- tion. Considerable difference exists also in the time when growth begins on different portions of the same slope. The influence of local conditions on the inception of growth was well shown in the case of a north slope in the Hudsonian zone with an altitude at the crest of 7,850 feet and an incline of from 15° to 18°. With a suc- cession of warm, sunny days the snow began to disappear first from the crest, then from the slope, and finally from the base. As the snow melted, growth began almost immediately. Before it had fairly begun on the slope, however, the crest showed a sparse covering of green, while a similar relation later existed between the slope and the base. In the early part of the season there was a marked difference between the amount of soil moisture at crest, slope, and base, the average percentages being 23.2, 27.6, and 39.9, respectively. As the season advanced, how- ever, moisture conditions gradually became equalized, and in the latter part of the growing season the moisture content of the entire slope was practically uniform. For this reason, though growth began on the crest from three to six days earlier than upon the slope and from seven to nine days in advance of that on the base, toward the end of the growing period the vegetation as a whole presented a strikingly uniform appearance. The period of growth resumption in each zone lasts ordinarily about 20 days. In the Hudsonian zone this period usually begins about June I02 Journal of Agricultural Research voi. m. no. 2 25 and terminates about July 15. While climatic conditions have a direct influence upon the time when growth begins, the length of the growth-resumption period in a given locality is determined more by the vigor of the vegetation than by anything else. Where year after year the herbage has been removed prior to the time during which the nutri- ents necessary for spring growth are stored in the roots, growth begins several days later and vegetative development is strikingly less luxuriant than in the case of plants of the same species which have not been sub- jected to similar treatment. FLOWER-ST.^LK PRODUCTION Under ideal conditions flower stalks begin to appear from 10 days to 2 weeks after growth has started. The stalks make a vigorous height growth, and there is a profusion of inflorescence, which is fertilized at an early date. Actually, however, the period of flower-stalk production is often retarded by cool temperatures and other cUmatic factors, so that the time the stalks begin to show may vary in different portions of the range as much as several days, similarly to the inception of growth. Obviously the two are closely related, an early and prolific herbage pro- duction being followed by an early and luxuriant flower-stalk develop- ment, and a late, scanty growth of herbage by a correspondingly late appearance of a few weak stalks. The vigor of the vegetation and consequently the time and abundance of flower-stalk production are also strongly influenced by the way the lands are grazed. Close cropping, coupled with successive trampling prior to the full development of the plant, delays not only the flower- stalk production, but also the maturing of the seed crop in subsequent years. The period required for an overgrazed plant to regain its vigor depends on the amount of injury received and the situation in which it grows. Three seasons of protection from grazing are usually sufficient for herbaceous vegetation to recover its vigor fully. To determine the actual difference in the time of flower-stalk pro- duction on closely grazed, moderately grazed, and protected areas, as well as the time required for overgrazed plants to recover their lost vigor, observations were made during three successive seasons of mountain bunch-grass areas grazed in different degrees, and on others completely protected from stock. The range selected was at an elevation of 7,300 feet in the heart of the Hudsonian zone. The unprotected or open range was grazed according to the usual practice, the forage crop being removed early in August each year, prior to maturity. During the first year (1907) no difference was observed in the time of flower-stalk production on the protected and open ranges, since the vegetation on both had previously been weakened through grazing. In 1908 and 1909, however, there was a notable difference in flower-stalk production, as shown in Table I and Plate XV, figure i. Nov. i6, 1914 Revegetation of Range Lands 103 Table I. — Annual progression in the flower-stalk production on closed areas and open range Condition of range. Period of flower-stalk production. Estimated percentage of increase in flower stalks. 1907 igoS 1909 1907 1908 1909 Closed to grazing Adjacent range open to grazing. July 10 to Aug. 20. July 10 to Aug. 20. July 4 to Aug. 10. July 8 to Aug. 20. July I to July 28. July 8 to Aug. 13. 0 20 60 1 After the first year of protection mountain bunch-grass produced its flower stalks both earlier and more luxuriantly than on the adjacent range open to grazing. This was also true of other forage species. Though the advance in the time that the flower stalks appeared (4 and 7 days in 1908 and 1909, respectively) was not very great, the stalks were developed more uniformly, the total period required for the function on the protected area being 37 and 28 days in 1908 and 1909, respectively, and on the open area, 43 and 38 days. To compare flower-stalk production on areas grazed in the usual way and on others protected from grazing until the seed crop can ripen, Sev- eral small plots of mountain bunch-grass, i meter square, were clipped with shears just above the ground for three successive seasons. On half of the quadrats the herbage was clipped once each month, or three times during the growing season, while on the remaining plots cutting was not done until after seed maturity — about September i. During the fourth and fifth seasons the herbage was undisturbed. The results showed that in the case of the plots clipped monthly the vegetative growth decreased in abundance each successive season. In the fourth year the undisturbed herbage was exceedingly weak, short, and sparse. No flower stalks were produced until the vegetation had been given one full season of rest, and then only a few late weak stalks were sent up. On the plots clipped after seed maturity, however, the flower stalks were produced fully as early, as uniformly, and as profusely as in the case of the plants which had remained unmolested during the 5-year period. Herbage production was also equal to that on the pro- tected areas. On the open range, grazed early in the growing season, the flower stalks were produced at irregular periods, a few appearing early in July, the majority coming in August. On yearlong protected areas and on those protected until the seed crop had ripened, the stalks appeared early, practically all being in evidence before August 10. Early and abundant production of llovver stalks is of the utmost importance in seed production. Ordinarily from three to five weeks are I04 Journal of Agricultural Research voi. iii. no. » required for the proper development and "filling" of the seed after the flower stalks are produced. Climatic conditions usually become rather severe in the latter part of August, however, and flowers fertilized after the first week in August have not sufficient time to develop and mature their seed. Plants with low vitality are likely to produce flower stalks so late that the seed has no time to mature. A few species which repro- duce vegetatively seem to have an inherent tendency toward irregular and late flower-stalk production, and such delay should not be confused with that due to low plant vigor. Of these species, pine-grass and yar- row, or wild tansy, require the longest time to produce flower stalks. On the lowest border of the Transition zone these species actually send up their flower stalks as early as July lo, and on the higher areas continue until inclement weather — about September 15 — stops their activities. PERIOD OF SEED MATURITY Under the most favorable conditions the time required for the devel- opment of seed is about 25 days. The period varies slightly with the length of the growing season and is longest at the lower elevations. It also varies among different species, and the time given for ideal condi- tions should be considered as approximate. Naturally the same factors which promote or retard the growth- resumption period and flower-stalk production also determine the time of seed maturity, though the last named fluctuates least, since toward the end of the growing season physical conditions, especially soil mois- ture and air and soil temperature, tend to become uniform throughout the range. The length of the seed-maturing period varies widely from vear to year as the result of the presence or absence of killing autumn frosts. In 1907 practically no seed of the more valuable perennial herbaceous species were matured in the Hudsonian zone until August 20, while in 1908 the ripening period came at least five days earlier. From this it might appear that the seeds were matured more slov.Iy in 1908 than in 1907. This was not the case, however, the apparent difference being explained by the fact that in 1907 weather conditions after the first week in September were so unfavorable that the seeds which had not matured by that time were destroyed. In 1908, on the other hand, heavy frosts and low temperatures did not appear until September 20, and practically all the seed matured. In 1908 seeds of mountain bunch-grass and other important species began to ripen on August 15, and by August 25 fully two-thirds of the crop had matured. After September 5 practically no immature seed were to be found, even on the cool north slopes at the higher altitudinal limits of the species. The secondary grazing plants, almost without exception, had matured their entire seed crop by September 10. Nov. i6. 1914 Revegetation of Range Lands 105 In 1909 the seed-maturity period began eariier than in the two pre- ceding seasons. In the case of mountain bunch-grass and a few other species this difference amounted to as much as 10 days in the identical situations observed in previous years. As with the production of flower stalks, the latest period of seed maturity occurred in 1907. The seed of vegetation on the cool and moister north slopes invariably matured later than that on other exposures and on level land at the same altitude, the difference amounting to a week or 10 days. Elevation is, of course, influential in determining the time of seed maturity. Each increase of a thousand feet, other conditions remaining the same, brings about an approximate delay of a week. The chief factor in determining the time of seed maturity, however, as well as the size of the seed crop, is the vigor of the vegetation. Where the herbage had been grazed for several suc- cessive seasons when green and the vitality of the vegetation thus low- ered, no seed was produced, or else the period of maturity came so late as to be seriously interfered with by frosts and low temperature. In contrast to this, the seed-ripening period on yearlong protected lands and on those not grazed until after seed maturity was much earlier and more uniform, while the amount of seed produced was notably greater. On the unprotected range there was little difference in the time of seed maturity from year to year. On both the yearlong protected range and that grazed after seed maturity, however, the period came earlier each successive season, in direct ratio with the increase in vigor of the vegetation. The importance of keeping the vegetation thoroughly vigorous is further exemplified by the clipping experiments. Where the herbage had been removed monthly for three successive seasons, no seed was developed in the fourth year when the plots remained undisturbed. On the other hand, on the plots clipped annually after seed maturity the seed crop was fully as large and matured at the same date as on lands from which stock were excluded. The experiments show, therefore, that if the forage crop is left undis- turbed until the seed has ripened, at which time the plants will have ceased growing, it will produce as large and as early a seed crop the fol- lowing season as will vegetation on range not grazed at all. Clearly these facts are of the greatest importance in devising a system by which the forage may be grazed without interfering with seed production. VIABILITY OK THIi SEED CROP The germinative power of the seed of the leading range plants was determined, in order to ascertain what reproduction might be expected under favorable conditions. In Table II, which gives the results of the tests, the high-range and low-range plants are grouped separately. io6 Journal of Agricultural Research Vol. III. No. 3 Table II. — Average fertility of the seed of range plants from igoy to igog, inclusive" PLANTS OF THE HIGH RANGE Name of plant. Scientific. Seed fertility (germina- tion). Mountain bunch-grass Little bluegrass Short-awned brome-grass Tufted hair-grass Onion grass, or mountain bluegrass.... Reed-grass Alpine redtop Smooth wild rj'e Alpine timothy Western porcupine, or needle grass White foxtail Little needle grass Elk-grass Tall swamp-grass Sheep sedge Wire sedge Wood rush , Mountain onion False hellebore , Skunk weed, or Jacob 's-ladder Wild celery Blue beardtongue Wild buckwheat Horsemint Mountain dandelion Woolly weed Yarrow, or wild tansy Coneflower , Festtica viridula Poa sandbergii Bromus marginatus Deschampsia caespitosa Melica bella Cinna latifolia Agrostis rossae Elymus glaucus Phleum alpinum , Stipa occidentalis Sutanion vclutinum Stipa minor Carex geyeri Carex exsiccata Carex illota Carex hoodi Juncoides glabratum Allium validum Veratum viride Polemonium humile Ligusticum oreganum Pentstemon procerus Polygonum pliytolaccaefolium . Agastache urticifolia Agoseris glauca Hieracium cynoglossoides Achillea lanulosa Rudbeckia occidentalis Per cent. 12. 2 7.0 47.6 26. 4 4.0 86.8 36.0 21. 2 69-5 27. o 69- 5 29.8 21-3 15-2 27- S 37. 24- 42. 6. i8. 9. 22. 2 36.0 10. 9 27.8 21.8 Average germination. 28.3 PLANTS OF THE LOW RANGE Big bunch-grass Pine-grass Marsh pine-grass, orbluejoint. Mountain June-grass Slender hair-grass Soft cheat Tall meadow-grass Geranium Fireweed Average germination. Agropyron spicatum Calamagrostis rubescens Calamagrostis canadensis Koeleria cristata Deschampsia elongata Bromus hordeaccus Panicularia ncrvata Geranium viscosissimum . . . . Chamaenerion angustifolium . 24-3 69.6 71-5 15-0 41. 6 48.2 85.0 29- 5 21- S 45-1 1 In the case of a few species listed seed tests were made during two seasons only. It will be observed that the viability of the seed of most plants is low. In the case of the important mountain bunch-grass, for example, barely more than one-tenth of the seed germinates. In a few cases mountain Nov. i6. 1914 Revegetation of Range Lands 107 bunch-grass seed showed a fair viability, the maximum germination obtained being 25.2 per cent, but the average of a great number of tests made under various degrees of temperature gave the figure in the table. In general, seed of the less desirable species, such as white foxtail and reed-grass, show a higher percentage of germination than that of moun- tain bunch-grass, short-awned brome-grass, and others. The germina- tive power of the seed was generally lowest in 1907, owing to the low vitality of the vegetation due to previous early grazing. In subsequent seasons on areas protected entirely from stock or until the seed had matured there was a pronounced increase in the germinative power of the seed. It will be seen from Table II that, in general, seed fertility decreases with elevation, the average germination of the plants on the higher ranges being only 28 per cent, as against 45.1 per cent for those on the lower elevations. Even with the best conditions of growth and plant vigor, the vegetation of the region must struggle to mature its seed during a short and none too favorable growing season. The effect of exposure and of low vigor of the vegetation (as indicated by the date of maturity) on the germinative power of the seed is shown in Table III. Table III. -Effect of exposure and date of maturity upon germination of mountain bunch-grass Series No. Source of seed. South exposure West exposure . North exposure East exposure . . South exposure West exposure . North exposure East exposure . . Feet. 7,400 7,400 7>3oo 7,350 7,400 7,400 7.3°° 7. 3S° Date of maturity. Aug. 20 Aug. 22 Sept. I ...do Aug. 31 ...do Sept. 12 Sept. 14 Germina- tion. Per cent. 14. O 9-5 "■S II. o 7.0 4-5 1-5 The data in this table bring out two important facts: (i) There is no difference in the vitality of the seed of mountain bunch-grass ripen- ing before September i , provided the variation in the maturing period does not exceed about 10 days. (2) There is a pronounced difference in the viability of seed which reaches maturity by September i, as com- pared with seeds ripened September 10 or later, the latter showing prac- tically no germinative power. The same relationship between the germinative power of early and late-maturing seed was observed in the course of field sowing in the natural habitats, though in all such cases the germinative power of both classes of seed was higher. io8 Journal of Agricultural Research voi. in. no. j The essential conclusions regarding seed germination are : 1. Even under the most favorable conditions the viability of the seed of practically all the forage species is low, especially on the high mountain lands. 2. Late resumption of growth in the spring and low plant vigor, both of which can be traced directly to premature grazing, result in a decrease in the amount of seed produced and in the germinative power of the seed itself. 3. If viable seed is to be produced, the vegetation must not be habit- ually deprived of its leafy foliage during the critical growing and food- storing period. SCATTERING AND PLANTING OF THE SEED But little time elapses between seed maturity and dissemination. This fact is highly advantageous, in that grazing may begin almost immediately after the seed matures without danger of having the crop consumed. The distance the seed is carried from the parent plant depends chiefly upon the species and the wind. Grasses and grass-like plants, such as sedges and rushes, drop their seed near the parent plant. Those of plants hke fireweed {Epilobium spp.) and Crepis spp., which are provided with bristly capillary hairs and pappus, and those of false hellebore, which are winged, are carried relatively great distances by the wind. Fireweed and dandelion do not grow in as dense stands as the grasses, but, as a rule, are more widely distributed over the range. About 90 per cent of the forage species depend primarily upon w'ind and water for the distribution of their seed. The remaining 10 per cent, of which huckleberry is an example, depend very largely upon animals for dispersal. To insure reproduction of the forage plants, the seed must in some way get itself planted. Though nearly all seed will germinate on the surface of the ground where there is abundant moisture, the resulting seedling plants in a localitv where the soil dries out early in the season are unable to extend their limited root systems deeply enough to reach the moist lower strata and consequently die from drought. The size and character of the seed play an important part in the natural reproduction of range plants. The seed of some of the most important species, such as mountain bunch-grass, short-awned brome-grass, and wild celery, are large and light, and even though dropped promptly upon maturity in the autumn, months before germination takes place, are usually found uncovered on the ground in the spring. On the other hand^ the seeds of wild onion and some of the sedges and rushes are smaller and heavier and have less difficulty in working into the soil. Among the valuable grasses observed, only one had become planted through natural means. This one exception was western porcupine grass, which is becoming securely established on the range not only in localities where it is abundant, but often on the tightly packed soils of denuded trails, Nov. i6. 1914 Revegetation of Range Lands 109 on hillside terraces formed by the trailing of sheep, and, in fact, every- where that its seed is developed. In favorable situations i square meter of surface showed as many as 700 seedlings of porcupine grass in the spring of the year. This unusual aggressiveness is not due, as might be expected, to exceptionally strong seed habits, but chiefly to the morphology of the scale, or lemma, which closely envelops the seed. The scale is very rigid, with an awn about lyi inches long protruding from the apex. At maturity this awn is tightly twisted, as shown in Plate XV, figure 2, but when moistened it untwists vigorously, causing the bent, needle-like point at the lower end of the scale to bore into the ground, the stiff, backward-turning hairs holding it in the earth when once started. The repeated twisting and untwisting of the awn with variation in the mois- ture finally results in the complete burial of the seed prior to the germina- tion period. Thus, if the seed of the valuable forage species is not planted by arti- ficial stirring of the soil, undesirable species, such as white foxtail, may become established at the expense of the valuable range plants. GROWTH AND ESTABLISHMENT OF REPRODUCTION OF FORAGE PLANTS The production of a seed crop of high viability does not necessarily mean any material increase in the forage stand. The seedling plants are often seriously injured or destroyed in the fore part of the grazing season by low temperature and lack of soil moisture. Certain plauts are not sub- ject to as serious injury as others, and so the ultimate stand may consist of a single species. The growth and vitality of reproduction will be dis- cussed under three heads: (i) "Development and loss of forage seed- lings during first year;" (2) "Loss of forage seedlings during dormant period following first year;" and (3) "Growth and loss of forage seed- lings during second and subsequent seasons." DEVELOPMENT AND LOSS OF FORAGE SEEDLINGS DURING FIRST YEAR During the first season of growth in the Hudsonian zone, approxi- mately ID weeks long, the seedlings do not grow tall enough to produce forage, though the young plants are sometimes cropped to a limited extent in the autumn. The height attained by mountain bunch-grass, as well as the root development, is shown in Plate XV, figure 3. It will be seen that the depth of the root slightly exceeds the height of that por- tion of the plant above ground. This plant represents about the average development of a forage seedling on well-drained and drier situations during the initial year of growth. On the lower elevations, owing to the longer growing season, the seedling plants usually attain much greater development than the one shown. Observations extending over five successive seasons show that in nor- mal years the low temperatures characteristic of the Hudsonian zone are responsible for considerable loss of seedlings during the first year of G2697°— 14 2 I lO Journal of Agricultural Research Vol. Ill, No. growth. The extent of this influence is shown in figure 4. It will be observed that freezing temperatures occurred on three nights in July, 1909 — namely, July 12, 17, and 27 — the temperatures recorded being 30°, 23°, and 29° F., respectively. In August freezing tem- peratures occurred on the ist, 4th, 2oth, and 26th, the lowest being on the night of the 4th, when the temperature registered 28° F. Only on the nights of freezing temperature in July, however, was serious harm done, and then only to the young seedling plants. The greatest in- jury occurred on the more moist, but not marshy, situations, where the sur- face soil heaved as a re- sult of alternate freezing and thawing. This action of the soil exposed por- tions of the roots of the seedlings, leaving them at the mercy of the sun and wind. In a few excep- tional cases 50 per cent of the seedling stand was thus destroyed. The freezing temperatures during Au- gust were not destructive, since then the root systems were better developed and there was less heaving of the soil, on account of the lower moisture content of the surface layer. Because of the high el- evation of the Hudsonian zone, the maximum tem- perature rarely exceeds 90° F. and, as a rule, does not seriously hamper the activities of the vegetation, though, in so far as it influences Nov. 16^ 1914 Revegetation of Range Lattds III the relative air humidity and increases transpiration and evaporation from the soil, it also decreases the available soil moisture. With the limited amount of precipitation during the growing period in the region, the soil moisture gradually decreases as the season advances. Low temperatures and lack of moisture in the upper soil often work together to destroy seedlings, in that the seedling roots are first pushed np by the heaving of the soil during freezing to a point where there is not enough moisture available for growth. The seedhngs of no particular species seemed to suffer more seriously than any other in the same situation. Individuals which sprang from seed that had not foimd its way into the soil and which therefore had comparatively shallow root systems were, of course, the most seriously affected. Plants of the same species growing in the same situation often exhib- ited contrasts in their ability to withstand drought, and when plants growing in different kinds of soil were observed, the contrast was great. This is shown in Table IV. Table IV. — Water content of soil at time 0/ death of forage seedlings Name erf plant, soil typ€, and .situatkm. Mountain bunch-grass: Basaltic clay loam — Quadrat station 4 Seeded area Do Quadrat station 4, seedlings in flats. Do Do Seeded area Gravelly clay loam — Seeded area, seedlings in flats Do Do Do West slope Do South slope Do East slope Western porcupine grass: Gravelly clay loam — West slope South slope East slope Short-awned brome-grass: Gravelly clay loam — West slope South slope East slope Number of succes- sive days of wilting, (?) Date of death. July 22 20 24 19 20 21 23 19 19 20 18 19 18 19 21 21 21 19 19 18 Percent- age of □onavail- able water. 6.8 6.3 6.8 6.2 6. I 6.2 7-2 5-4 5-7 S-4 5-4 5-i 5-4 5.6 5.6 5-2 5-4 5-2 5-8 5-6 S-9 Average percent- age of nonavail- able water for each species. 6.51 5-5 S. 26 5-75 1 1 2 Journal of Agricultural Research voi. m. No. i The figures in Table IV show two important facts: (i) In the finer soils a higher percentage of water is required to maintain the life of a seedling than in soils of coarser texture; and (2) mountain bunch-grass requires but little more available water than western porcupine grass, one of the deepest rooted species, to become established in a given soil. Mountain bunch-grass, moreover, will thrive on a soil containing less water than would be required for short-awned brome-grass. To judge from the nature of the situations invaded by mountain bunch-grass, the plant may safely be classed among the drought-resistant native species. Numerous observ-ations in 1909, supplemented by many counts on small unit areas in different range types and on different soils, showed that nearly the entire loss of seedlings occurred before August 1 . After that date loss was prevented by cooler temperatures and by an increase in soil moisture resulting from precipitation. However, the loss before August I had been rather severe, varying from 20 to 70 per cent, accord- ing to the situation, with a general average of about 50 per cent. In 1 9 10 the seedling loss in the fore part of the season was approxi- mately the same as in 1909, but the loss in the latter part was much greater, and only 25 per cent of the original stand remained vigorous and active in the autumn. This extensive loss was due to continued dry weather and high temperatures. LOSS OP FORAGE SEEDLINGS DURING DORMANT PERIOD FOLLOWING FIRST YEAR Table V shows that, in general, the loss of forage seedlings due to physical conditions from October i, 1909, to July i, 1910, approximately, was practically negligible. It will be seen that the heaviest seedling loss occurred on steep slopes, particularly on those where the soil was coarse and gravelly and the vegetation sparse. Quadrats 14 and 16 show this strikingly. The soil in quadrat 14, whose slope is 28.5° to the west, is very coarse and gravelly, and only one-tenth of the ground was covered with vegetation. In quadrat 16, with a slope of 3° to the south, the soil is mainly of clay loam, with a small amount of gravel, and three-tenths of the ground was covered with the same kind of vegetation as quadrat 14. The loss of seedlings on the two quadrats was 28.4 and 1.4 per cent, respectively. These losses were not due to severe temperatures, but primarily to heaving of the soil and erosion before growth began. Additional contrasts can be seen in quadrats 3 and 5, and 31 and 55. Nov. 16. 1914 Revegetation of Range Lands 113 Table V. — Loss of forage seedlings during the dormant and winter period Number of each seedling species. 80 Mountain bunch-grass. . . . I Western porcupine grass. 244 Mountain bunch-grass. .. . 4 Elk-grass 58 Mountain bunch-grass. . . . 1 Sickle sedge 116 Mountain bunch-grass. .. . 61 Western porcupine grass. 7 Yarrow 116 Western porcupine grass 17 Mountain bimch-grass . . . 30 Smootti wild rye 7 Western porcupine grass. 27 Smooth wild rye 2 Little bluegrass 138 Little bluegrass 9 Smooth wild rye 6 Yarrow 2 Mountain bunch-grass. . . . 128 Smooth wild rye II Yarrow 9 Crepis (sp.?) 91 Western porcupine grass. Quad- rat No. 3 5 13 14 16 31 38 42 44 55 Slope and ex- posure. 25° west. . 18° west . . 16° west . . 28.5° west. 3° south. . fii° south- \ east. i2°south.. U° south- \ east. 11.5° south 4° south. 10° east. . Total number of vigorous seedlings remaining — Autumn of 1909. 81 248 59 n6 68 138 37 29 155 148 91 Spring of 1910. 71 238 3 55 o 83 60 7 92 6 29 5 24 o 134 126 10 Loss. Per cent- 12.3 2.8 7.0 2». 4 1.4 26.3 8. I 17.2 3- 1 2.7 10. 9 On steep hillsides where the original vegetation and network of roots had been seriously injured in the autumn by trampling, erosion carried the seedlings away or exposed portions of the more superficial (lateral) roots. Of the seedling loss during the resting period, 80 per cent was brought about in this way, though even this was nominal, averaging in the location studied only 7.3 per cent of the total stand. Low temperatures were apparently responsible for the loss not directly due to gullying, but such loss was evident only in exposed situations. Practically all of the mountain lands are covered with a heavy blanket of snow before severe temperatures begin, which prevents excessive loss of water through the plant tissues aboveground and eliminates loss due to alternate freezing and thawing. No particular species appear to be especially immune to loss during the winter months. The roots of little bluegrass and sickle sedge seem to be exposed somewhat oftener than those of mountain bunch-grass, porcupine grass, short-awned brome- grass, and other species in the same situations. Mountain bunch-grass seedlings developed a rather unusually elaborate root system during the first year, which assisted in protecting them against adverse conditions. 114 Journal of Agricultural Research voi. iii. No. j GROWTH AND LOSS OP FORAGE SBEDLINGS DURING SECOND AND SUBSEQUENT SEASONS During the second year nearly all species make vigorous growth, both above and below ground. Plate XVI shows the deep and spreading character of the roots of 2-year-old mountain bunch-grass (natural size) at the end of the second season. It \vill be seen that the roots are much longer than the leaf blades. The vertical roots reach well beyond the dry substratum during the most critical period of drought. The leaf blades, which number about 70 or 80 on the more vigorous individuals, are all basal, with an average length of about 4 inches, about half that of the deepest roots. This splendid root and herbage development pro- vides the plant with abundant food-storage tissue, so that in the following season vigorous growth begins promptly. During the third year the development of the plant is quite as marked as during the two preceding seasons (cf. Pis. XVI and XVII). Both roots and herbage grow rapidly from the first, though the growth of the former still greatly exceeds that of the leaf blades. Such development is essential, for the roots absorb moisture slowly, and where transpira- tion is great, nothing short of a well-developed root system can supply the plant with the moisture it requires. Owing to the depth and spread of the roots, the question of available soil moisture is not a serious one, since an ample supply exists 3 inches below the surface layer. At the start of the growing period there is a superabundant supply of moisture, and plants whose roots are well beneath the surface soil continue to extend them more deeply until the innumerable root hairs have worked themselves through the capillary spaces among the soil particles, thus insuring the plant against drought. By the end of the third year mountain bunch-grass and other species complete their life cycles, and cease to be seedlings. Flower stalks and seed are then produced, as shown in Plate XVIII. At this time the plants are often as tall as older individuals. The specimen shown in Plate XVIII exhibits the maximum development attained, the average growth being shown in Plate XVII. The flower stalks (there are seldom more than three) of the 3-year-old plants as a rule are put forth a few days later than those of the parent or older plants. Con- sequently the seeds are not matured as early as are those of the longer established individuals, though the variation rarely exceeds five days. So extensive is the development of the plants by the second year of growth that the loss during that and subsequent seasons owing to cli- matic factors is negligible. The facts derived from the study of the life history of the vegetation, which are important as a basis for a rational and practical grazing sys- tem, may be summarized as follows: Nov. i6. 1914 Revegetatioji of Range Lands 115 1. The flower stalks of the important grazing plants begin to appear about July 5 and are for the most part produced between that date and August ID. The more vngorous plants send up their flower stalks first. Plants weakened by annual close and early grazing do not produce flower stalks until late in the season, and then send up only a few. 2. The seeds begin to mature by August 15, and by September i the major part of the seed crop is ripened and disseminated. Plants weak- ened by close and early grazing do not mature seed unless the growing season is unusually long and exceptionally favorable. 3. The viability of the seed of most species is low. The germinative power varies with difi^erent species, but especially with the vigor of the plants. Those which make a weak vegetative growth produce seed of very low viability. 4. The seeds of the most valuable species lack means for working themselves into the ground, and, if reproduction is to be secured, they must be artificially covered. 5. In the Hudsonian zone the germination period begins about June 25, and growth begins generally by July 15. 6. During the first year of growth, a period of about 10 weeks, the forage seedlings make a vigorous development. Owing to the friability of the surface soil, however, and the superficial position of the roots at that time, there is rather a heavy loss of seedlings from freezing and drought during the spring period. 7. During the donnant periods there is virtually no loss of seedlings. The only factor causing loss is erosion. 8. In the second and subsequent seasons physical conditions are favor- able to rapid development and growth of the young plants. By the end of the third season viable seed is produced. DIFFERENT GRAZING SYSTEMS IN THEIR RELATION TO GROWTH REQUIREMENTS AND REVEGETATION From the facts brought out by the life-history studies it is plain that a rational method of grazing should (i) avoid weakening the vegetation through continuous grazing prior to seed maturity; (2) utilize, so far as practicable, the trampling of the animals in planting the seed; and (3) provide for protecting the reproduction against heavy grazing until it is firmly established. At the present time grazing on the National Forests is carried out under one of three more or less distinct systems: (i) Yearlong or season- long grazing year after year; (2) yearlong or season-long grazing com- bined with an occasional total restriction of stock during the entire year for the purpose of giving the forage plants a chance to reproduce; and (3) deferred grazing, which aims at a rotation in the time of using each por- tion of the range, each year allowing an area to reach seed maturity 1 16 Journal of Agricultural Research voi. iii. No. 3 before it is cropped, but grazing it after that period, in order to avoid loss of forage througli nonuse and to assist reproduction by trampling in the seed. In the following pages the comparative merits of the three grazing systems, from the standpoint of the requirements of the range plants for growth and reproduction, are discussed. YEARLONG GRAZING The term "grazing system" implies a definite plan of utilizing the forage crop in accordance with certain basic principles. Yearlong or season-long grazing, however, is characterized mainly by a lack of system, since it fails to provide for the removal of the herbage at any particular time in any locality. Its ultimate results to stock and the range are not considered. It was this unrestricted grazing on National Forest lands prior to their inclusion that so seriously reduced the carrying capacity of the choice ranges. After the creation of the National Forests overstocking was eliminated as rapidly as the stockmen could meet the necessary reduc- tions, and regular grazing seasons were established. Even under sea- sonal regulations, however, the prevailing practice of yearlong grazing has not been conducive to the most rapid improvement of the range. In northeastern Oregon sheep are permitted to enter the mountain grazing areas early in July, when mountain bunch-grass and most of the other palatable species begin to put forth their flower stalks. Up to August I the flower stalks are virtually as palatable as the leaf blades, and where the range is stocked to its full capacity, as it is in practically all cases, most of the stalks are removed prior to the formation of seed. Moreover, there is a tendency to graze the same lands prematurely each year, a practice which impairs herbage development. This not only prevents seed production, but also results in gradually decreasing the carrying capacity of the range through starvation of the forage plants. Prior to the time of seed maturity practically all of the range has been grazed over at least once, and, as a rule, only the vegetation on the inaccessible lands is allowed to mature seed. After about August i the flower stalks are not eaten as a general rule, except in the case of certain moisture-loving species, such as butterweed (Senecio triangularis), but the vegetative portion, especially of the grasses, is so closely consumed as to prevent the manufacture of the food so essential to the development of the plants and the production of seed. Since the main seed-developing period in the Hudsonian zone comes in August, lack of an abundant food supply during the growth period is reflected in the low viability of the meager seed crop produced. To determine which species are becoming established under the system of yearlong grazing, several typical areas, overgrazed in various degrees. Nov. i6, 1914 Revegetation of Range Lands 117 were studied in 1907 before the stock was turned on to them. Full notes were taken on about 300 plots, i meter square. The lands selected were of the open, parklike type, with a scattered growth of whitebark pine and occasional clumps of alpine fir. Mountain bunch-grass was the predominating herbage species. Certain portions of these ranges were seriously depleted, and the usual succession of early, aggressive annual weeds had replaced the original perennial type. At this altitude the annual plants are of little value for forage, though grazed to a limited extent in the spring when succulent and tender. On this account the annual plants are not included in Table VI, which gives the results of the two seasons' observations ii8 Journal of Agricultural Research Vol. HI. No. s 8 a 8 8 o s 8 o a 8 •-* .o 8 8 I ■I to H B mber seed- ngs per uare eter. o 00 o eo o " 23 C ■?? g^= sa M .4^ a t-. so o *t o tn 00 O Tota umb of seed- lings 00 iH o r- r- *■ M M a •a j^ >o *0 ON VI >o O C *0 00 Num- ber oi unit areas coimte T fO (*) M wi n -•O'S « . C o C O 0 c O 0 O o o ia S§s^«§ P-g »EI f- •J •a aj 1 i 2 1 I ii E be ■0 a ? 1 0, 1 ^i 1 t: ■a £ I Z : gas" .H 1 .§1 0 ^ * d'.c 0 St to •c o *" a u: *5 rt C c •33 •2" 'ii'o 1 t;2 Pil 2 j3 2 Cll •g 01 o « ^ MM ! 11 .1 •s §^5 •s '■°«^ m bi _t ^ a sis c act; 0 1 .^1 , X 1 11 S E I S2 9* 2 «• O c IIP 1 Ei:§^ ■^j :s s. as 'tf S :5& is WS si 0) u c d 4- f^n c cc 00 v6 «o 1^ > be M « 1 a " ■S" :ii M 6 « r-O « « ft. 1 ss i i t 1 E J 1 , 1 > 2 •s •a 1 .^ E J C |b5=3bc c s > sz > rt 0 , -ai: c rt i,> 5"C •o-o J3 O e c 3 1 .2rti! _o 1 tj 1 •a 1 .a 1 J 4- ,°2 s ( o" s c :^o 3 £ » c 1 1 !•' c ' So ( e o o<» ■^ lo " w 1 M W) » , M Irt a ■ll= c c c P! 11= C Q C c c 1 ^ o t: •c X 1 ^2 ■^ t •c •c T a- a (^ •c c 4» •c •u t p 5 c •sl ri ca « 1 c c 4-i % o c *-l *- *■ w •J •t *i 1 I i = El 1 = b 5) t i aa O a a 3 = = a 3 a 3 3 3 *i: < < < ■< < <: •< < -< •< < Nov. 16, 1914 Revegetation of Range Lands 119 *0 0 0 WJ M 0 n ^ 0 0 »0 ^ M to 1 «n «■ rt wt (>• K) •* »-J ^ ro ^ ^ M 1 " z « 0 •«- 0 M jl » o-o -c , , ^ 1 00 fn (^ (^ 0 1 ■* ro .5 IE: .9^ 4- i1 3 p (y « •a » n 01 "5 3 a" ^- W) 0 rt o« 3 '3 M 2 Sf3 3J3 3.2 •2 2'2 m 2 £ Si; (: : : a "i! - E eg m w ■< :s 5 s in ^ "h US ;^ n ID «fl "C 0 ro ^ t^ 0 HI 00 00 0 00 M i^ a 4 a 'O « , < ^ IH i » 1 M " ro " M " 00 0 w -o n 0 1 irt fo 0 ;; •*■ 0 u « * - <» ^ 06 0. « « M !>. > >, te a a « B u ^ u a 0 J 5 3 0 c a e J > 1' J 0 > 1 0 c V a u 0 0 0 n t 1 ^ i 1 1^ a J 1 j • e e 0 c S ^ ( J ,* 1 J Q •S i 2 a id G - ** ; H •" u — C: M s 0 0 s 1° s > 0 0 '„ "„ , "X ^ *l* ■, °- " 1 w *„ i °f. SI SI n .; B ■; "o2 "o2 ^|i 4^ ?: »;o c a c ? ?■ r: c a 0 a e c 0 1 P 2' ■c •c •a 3.t 2' •a •c •a T 1 ■c to > > > >. > > ' Ie 1 g g 1 1 ": 1 3 3 ' " 1 a < s < a z < 1 120 Journal of Agricultural Research voi. iii. No. 2 The data given in Table VI, supplemented by observations made in 1908 and 1909, show conclusively that on the typical lands studied the important perennial forage species are not being reestablished. For example, such important species as mountain bunch-grass, little blue- grass, and big bunch-grass gave a maximum count of 6.4 seedlings per square meter, as opposed to one of 26.96 for sickle sedge, a species of little value. Even on the lands which still support a fair stand of the original valuable forage virtually no reproduction is taking place. Moun- tain bunch-grass, which produces flower stalks at a relatively early date, and whose chances are therefore good for maturing a viable seed crop, shows no reproduction from seed where the ranges are grazed each year before the first week in August. Practically all the seedlings on these ranges are of inferior species. Sickle sedge, an unpalatable but aggressive perennial, forms not less than nine-tenths of the total perennial seedling stand. This sedge matures a strong seed crop at a relatively early date and, in addition, perpetuates itself abundantly by offshoots from the rootstocks, which later develop seed. Besides sickle sedge, there was an occasional seed- ling of western porcupine grass, little needle grass, short-awned brome- grass, and slender hair-grass. The first three species are fairly good range plants, but the last named is grazed only to a limited extent early in the season. Probably because of this fact slender-hair-grass seedlings were more in evidence than any of the others. The maximum seedling density occurred on old bed grounds, where the vegetative cover was exceedingly scarce. The average number of seedlings obtained per square meter for all counts made upon such lands was 26.96 and 23.6 in 1907 and 1908, respectively. This exceeds by about 50 per cent the seedling stand for any other type of range exam- ined. There are three chief reasons why the seedling stand is dense on bed grounds: (i) The unpalatability of the parent species, coupled with early maturity of the seed; (2) thoroughness with which the seed is planted; and (3) relatively high water content of the soil. The seed of sickle sedge usually matures and drops before August i, and in consequence the plant is neither weakened nor the seed production interfered with by foraging animals. Though on most bed grounds the soil is hard-packed, on the particular ones examined it was loose and porous, and the trampling assisted in conserving its moisture by pulver- izing the surface. During the main growing season in 1907 and 1908 the soil moisture content of the bed ground averaged 30.2 per cent, exceeding by 7.9 per cent that of any other locality studied, except the swales. To sum up, it may be said that season-long grazing continued year after year seriously interferes with the growth of the vegetation, de- creasing both the quantity and palatability of the forage crop. By the Kov. i6. 1914 Revegetation of Range Lands 121 failure of the forage plants to produce seed, reproduction is prevented, resulting in a gradual decline in the carrying capacity of the lands. Even under conservative use the carrying capacity of the range does not improve rapidly through reproduction of the more desirable species. YEARLONG PROTECTION To determine the practicabiUty of reseeding the range through year- long protection from grazing, five typical overgrazed areas, situated at various elevations from 3,000 to 7,500 feet, were selected for study. Each area was fenced in 1907, and observations were made during four successive seasons. The results from the areas in each zone are pre- sented separately. HuDSONiAN Zone. — The areas closed to stock in the Hudsonian zone had been subjected to close yearlong grazing for several seasons (PI. XXI, figs. 2 and 3), and because of the resultant low vitality of the vegetation practically no seed was produced during the first two seasons. In the third and subsequent seasons, however, a satisfactory seed crop of average viability was produced. The vegetative changes which took place in representative quadrats are shown in text figures 5 and 6. It will be seen that at the time of their establishment the quadrats contained no perennial forage seedlings. The first year passed without any making their appearance. In 1909, however, 7 seedlings appeared after the germination period, but, as shown in figure 6, only 5 survived the subsequent dry season. On the denuded quadrat (fig. 6, quadrat 2) 2 mountain-bunch-grass seedlings came in during 1908, both of which succumbed later. In 1909, 10 seedUngs were found in the spring, only 6 of which survived the season. Seed was produced in abundance each year, but for the most part remained on the surface of the soil. At the beginning of the study the quadrats were stocked with an inconspicuous and useless plant called knotweed [Polygonum ramosissimuni) , which is common on overgrazed ranges throughout the mountain-bunch-grass association. On the permanent quadrats this species no more than held its own, but on the denuded plots it increased prodigiously. The contrast in the aggressiveness of reproduction of the annual and perennial species on protected areas, as shown in the case of mountain bunch-grass and knotweed, holds generally. The only perennial species which reproduced well under yearlong protection was western porcupine grass, the seed of which, as already pointed out, is planted by means of an awn attached to the floral glume. The fact that practically no reproduction from seed was secured as a result of yearlong protection does not necessarily mean that such pro- tection will not bring about an increase in the carrying capacity of the range. As a matter of fact the carrying capacity was increased through the production by the original perennial plants of more and longer leaf 122 Journal of Agricultural Research Vol. Ill, No. a blades and by an increase in the size of existing tufts or hummocks of tussock-forming plants. The leaf- blade increment is shown in Plate XXI, figures 2 and 3, where contiguous areas protected and grazed aimually prior to seed maturity are compared. The leaf blades began to increase in number and length iiJter the first year of protection and continued throughout the four years. The greatest development in the foliage came in the second year of protection, when the plant for the first lime Nov. i6, 1914 Revegetation of Range Lands 123 was able to manufacture ample food. Many species of grass had doubled in length by the end of the fourth year. The increase in actual stand or ground cover was due almost entirely to the enlargement of the tufts, and text figures 5 and 6 show that even under season-long protection the bunch-grasses and other valuable plants o 2 TO O t «) D o 2 4- e TO, 3 a c c £" 0. I ( 1 ■ n <>. >. <» ti q Tt z "^t SI 0 . 9 ^ <^ n ^^ ^ ">. o."^ Q. Q. ^ ^ < ^ Q. q. . s \ J. v% <». I, n J Kl N^ ^ s^ L. t ' M ^ jjq 1 ^d 0 ( ^ L M 1. ^ 1 . ■5. U. > '^ ^ •^ L >. i 'i. <^ n ^^ ^ '^ ."^ I t» 0, t ( f^ >* Kj q iV N ^ ^\ ' 1^ C ^' S >. <> <^ 1. l-^ ^ <» ». <>. ^ ■ ; ( 1 I Q. 0. <* Q (i ■* . ^ ' tX 0. ^. ^ . •s) ^ 1 ^ *. A I. "^ ^ 1 f -A ► A, N^ 6s 1 |fl^ •, - ( ^i> \ ip M kg 1 /'■ V ?^ 1, «. ' <^ w M 1 «l { / l»' ^, Q l'^ "^ e f 1' m ''^ ■^ " 1 11, i ' ^ *» k s >. I. „^ ■ ^ <1 K' 1 1^ I . > ^n >. a. In ■^ x 1 z ( Q. ( '•a Q I i w +- < 1 ^ < «, ) i) d 1 ^ ^ i'^ ■* ■ lr< 1 ( fm ^ ^ 1 *i Ci %\ « 1 v^ ^ u -^.^ t 0 ^ '■0 "^ " rt i Q >-n q q. Cl ,^) ^^ ^ ^ ^^ ^ '^ n * I « "■«. ■i. y "!," <> ^^ ^ ^-^ tl « '^ ^ :^ ^ n n 4 c <» I nV ^1. r q. 1 h QIC jfcfc i.\ Li fc ' II a) X r- m O JD t£cQ 40 2 •a E ■^ Q. .y o to o I do not increase rapidly by this means. Planimetcr measurements of the tufts showed an average increase in diameter of only 18 per cent for the third year of complete protection. The young, small tufts were the only ones to increase noticeably in size. The tufts of mountain bunch-grass and most tussock-forming 124 Journal of Agricultural Research voi. in. No. 2 species seldom measure more than 12 inches in diameter, and with the approach toward full development the annual increase becomes less and less. At just what age the hummock or tuft reaches full development is not definitely known, the age doubtless depending upon the species and situation. Transition Zone. — Areas were carefully selected and fenced for study in the Transition zone in July, 1907. The range studied had been so seriously depleted that it was difficult to ascertain exactly what plant species constituted the main forage crop. The grasses found in more or less abundance at the time the plats were established were soft cheat, slender hair-grass, Olney's bluegrass, mountain June-grass, pine-grass, big bunch-grass, and western porcupine grass. Other important forage plants occurring sparingly were yarrow, alfileria (Erodium cicuiarium)^ arnica {Arnica cordifolia) , and geranium. Though the ground appeared to be nearly denuded, there was here and there a somewhat conspicuous stand of Erigeron aureus, stonecrop {Sedum douglasii), Clarkia pulchella, and Douglas knotweed {Polygonum douglasii). During the first year of protection there was practically no plant invasion of the permanent quadrats (PI. XX, fig. 3), but when the fall rains came — about September i — the early stages of invasion became apparent on denuded areas. The plants to enter first were knotweed, an inconspicuous annual weed, Clarkia pulchella, Erigeron aureus, alfileria, soft cheat, western porcupine grass, and slender hair-grass, named in the order of their abundance. In the second year practically the same species entered the quadrats, with the addition of a mixed association of three perennial grasses — mountain June-grass, big bunch-grass, and pine- grass. Geranium, arnica, and stonecrop were also noted. At this time soft cheat, an annual species, was the most conspicuous and aggressive. By the end of the third year, 1909, soft cheat had made such a rank growth that in certain portions of the protected area it had completely replaced the shorter annual weeds (cf. Pis. XX, fig. 3, and XXI, fig. i). In the denuded quadrats certain perennial species, mountain June-grass, big bunch-grass, Olney's bluegrass, geranium, and yarrow, had also begun to appear as forerunners in the permanent establishment of perennial species. As a result of protection from grazing, the carry- ing capacity of the protected area as a whole had increased 150 per cent (PI. XXII). This figure represents the increment in the ground cover within the quadrats rather than the weight of the forage produced. Fully four-fifths of the new growth was composed of annuals, with soft cheat easily predominating. Seeds of this species are small and, like those of western porcupine grass, work their way into the ground by means of special contrivances. The larger and Ughter seeds of the perennial plants were found on the surface of the ground at the time of germination, and in consequence practically no seedlings of these species had been established. The perennial species had reproduced vegeta- tivelv, but the increase was bv no means rapid. Nov. i6. 1914 Revegetation of Range Lands 125 On both the lower and more elevated lands the new forage resulting from yearlong protection consisted almost exclusively of annuals. A very small percentage indeed of the new stand was established through the reproduction of perennial plants. During the first two seasons of protection, even at the lower elevations, the perennials produced prac- tically no fertile seed. In the third season viable seed were produced. The perennial vegetation, which previous to protection from grazing made such a weak growth that its presence was not observed, finally became conspicuous after one or two seasons of rest. Even under the most favorable conditions reproduction by the perennial species is slow, since the seeds are large and unable to work themselves beneath the surface of the soil. The nonuse of the forage under yearlong protection is a serious matter and would have to be considered before this system could be pronounced practicable. Moreover, the accumulation of in- flammable material during the period of protection would result in increased fire danger. Under any circumstances it could not be carried out on a large scale without a radical readjustment of the stock industry. To sum up the facts regarding yearlong protection : The system is not an efficient one, because the most valuable perennial species fail to reproduce by seed. While the carrying capacity of the land is increased, this increase is slow and does not compensate for the waste of the forage crop during the long period necessary for revegetation. DEFERRED GRAZING Unlike the two grazing systems just discussed, deferred grazing is based upon the requirements of the vegetation through practically a double life cycle, as defined in the section on the life history of the range plants. To determine the effect of the deferred grazing system, the vegetation in the different grazing zones was studied for two successive seasons. Convenient areas which had been overgrazed in various degrees and which were large enough to support a band of sheep after the seed had ripened were closed to grazing from the beginning of the season until the seeds of the important species had matured. Upon maturity of the seed crop the range was grazed moderately by a band of sheep during the remainder of the season. The following year the same area was closed to sheep for the same period in order to give the seedlings an opportunity to develop a root system strong enough to withstand tram- pling and also to permit a second seed crop to be developed and dis- seminated in case the first year's seeding was unsuccessful. When the area had been sati.sfactorily reseeded, it was grazed early in the season, and a second area, large enough to maintain a band of sheep from the time of seed maturity to the end of the season, was reserved for deferred grazing. Results of a study to determine the abundance of seedling reproduction under the deferred grazing system, as compared with that under other systems, are shown in Table VII. 62697°— 14 3 126 Journal of Agricultural Research Vol. III. Nc. s 8 ■2 a 8 '^ s o < 3 S jj g °-a Sj; aH g.S = S8" 5 »-i O 1' 3^ •s is I a c g Si . n y n be M ° p^"* bey. j~, :^ as ■§?' -i •s b M o ") r~ w^o Ov o «0 co O -o fJ) ^ rJj -^ (J)>0 ^ ►^ I O C?-© "O vo «^ fO 9-0 -I > *J 4* > t. : ia fi - • rt a ■■£ 5 s : >> s ■ * I « 0 ^i.^ n ^■^1 ^ tj « ffl 10 O. at Sxi TJ-O S 0 °-°J3 i - o s r. = i«"-o ~ s !S - " « O fc. « O U tfi" - 3 OJ E 3 3 S.a t**^ ID O "^ r- . •a a s;_3 30 o !^ S'5 b3'S3 § 2 ~ a^J; 0 0x1 " ^3^-33 „M Nov. i6, I9M Revegetation of Range Lands 127 SEEDWNG RBPRODUCTION SECURED It will be seen that as a result of the deferred system of grazing the density of forage seedlings was from 2 to 10 times as great in 1909 as in 1908. Quite as important as the density was the identity of the seedling species. It will be recalled that where yearlong grazing was carried on, seedlings, aside from the annual weeds, consisted of two early-maturing annual plants, sickle sedge and slender hair-grass, together with a more valuable species, western porcupine grass. On the yearlong protected plots practically the only species were the few with small, heavy seeds, and the two porcupine grasses (Stipa minor and S. occidentalis) which are self-planting. Table VII shows that where deferred grazing was car- ried out the ground became stocked with seedling plants of all species which produce seed. The most valuable species, mountain bunch-grass, which failed completely under yearlong grazing and yearlong protection, responded exceptionally well. Its seedlings were found in all situations where there were parent plants to produce the necessary seed crop. In the Canadian zone several of the valuable species which failed to reproduce under either of the other systems regenerated more or less abundantly under deferred grazing. Although, under deferred grazing, forage seedlings were found wherever there were enough parent plants to produce the necessary seed, the propor- tion of the seedling stand which ultimately became established depended mainly upon the habitat and climatic conditions, as well as upon sufficient protection from grazing during the period of establishment. LOSS OF SEEDLING REPRODUCTION BY GRAZING To determine the extent to which moderate deferred grazing reduces the stand of valuable forage seedlings and whether the subsequent seed- ling stand resulting from the additional seed crop when thoroughly planted by trampling will offset the number of seedlings lost through grazing, observations were made on selected plots at medium and high elevations. The first observations recorded the character of the vegeta- tion in and around each quadrat for a radius of 10 feet; the density of the herbaceous vegetation within and without the quadrat; the character of the soil and the slope and exposure; and the total number and identity of the seedlings within the quadrat, and their health vigor at the time of observation. With such data it was possible to account for any unusual loss resulting from subsequent grazing. It was recognized, for example, that a seedling, even when deeply rooted, is much more likely to be destroyed by trampling if it is situated on an abrupt hillside, especially in a denuded gravelly soil, than if situated in a level glade between tufts of grass with intertwining roots. Again, a seedling growing under ad- verse moisture conditions does not develop as elaborate and deep a root system as one which has received enough moisture to furnish the neces- 128 Journal of Agricultural Research voi. m. No. . sary nutrients and so can not withstand as much disturbance of the soil. After the lands had been grazed and sufficient time had been allowed for the vegetation to recover, each quadrat was again observed and notes taken on the total number and identity of the forage seedlings which remained, the number of seedUngs found dead, the number unaccounted for, the number whose recovery was doubtful, and the condition of the remaining seedlings at the time of the recounts. At Medium Elevations. — The entire area studied slopes to the west and has a minimum altitude of 5,500 feet and a maximum which brings it into the lower Hudsonian zone. The topography is so irregular and there is so much down timber that the herbage can be grazed only under the most skillful open herding. The lower portion of the range is of the browse type, the much relished Nuttall willow predominating. The undergrowth consists of a host of weedy species, such as fireweed and its associates, with a scattering of smooth wild rye and short-awned brome- grass (PI. XXIII). At the highest limits of Nuttall willow, seedlings of these species, on account of the shorter growing season, were not developed to the same extent as at the lower altitudes. The forage stand in existence before and after grazing is shown in Tables VIII and IX. Nov. i6, 1914 Revegetation of Range Lands 129 8 8 ts J! 8 1 ■-^ n 3 2: C3 ^ Qy ,„ t, 0 O— CT *S +3 _g2 ^3 •d 5« §1 •a '2 3*0 .as ■P l| >f it "tti ^ t>iis iJ*2 tt3"o 11 sis O* "I'O *J- w O N oij g So &ii oii oii "g u S fe S S 0 o .Iti dS ^ S ca M til u ^ .s .2: 5 a il 2 o E 2 'g3 18 >■ E" a c .»■ 'S" 2 -g a a Cm is 2 „■ a'S '^1 ■S 9 r, El a oT o c S 5^ 2 = 5 - a-> ^ ju I- i = a- •"§ a> I/) d ™ > M « i J3 M So *=& i It ■So s: f^S H-" H H-- ■Ji U3 ^ -D 1: Ho, l§8 2-Sg S g£3 « §£■; 3 w rt o ^ •" a 6 ;:; ft Sl'^'S'o *^ a 11 "' to ja B o.ti 3 - HOHinO 5 « >- -o • I30 Jourtial of Agricultural Research Vol. III. No. 3 .1 o O s =9 s >*i> ■I I ■S 1 "S. a; w ►J n < a a 5E §2" C8.Q Qj t, 0 o— ' cr 3-!^ U 5t3 3 o .9 2 O « 4»"rt eg ~ a-5 •i :a U O bt:.a 2 O.M a.n.2 O U) O K 5 a-o sl I g> •g ^T3 d .u-a « aj 1- ■£ & '0 a& a Nov. i6, 1914 Revegetation of Range Lands 131 ■> CO .s a S s 8 I d d V.S a a 1 .g •0 2;3 ^ S - ^ -a SJ ii 'U'o ■■ a ■>-> ^ ^ *"- g 11 1 ! 5 1 s -g- : J= 2 S ^- 1 ;; 2 S 3-a>.na.Safc S >. ill "■3 a ; •2 ■ a i 1 is la 4> a 0 « ao n i- rt E be N w d 2 ■ 5 ° ^ si 1 3 S S 5 bl-S £ ft 0 4 < Tfi . (^ fei^ e^S^U^S^ 9^ ^ > s : pq : otal mber lose overy oubt- fO W M (OMMf^fO*^^ M -0 0 00 >o ^§^£.2 _^^^_^ . _ ■ ,^ g ho .s .g jj V 0* ja i^j3 0 W 0 W 4>^ fl 0 I* Oj3 0 >. 0 >» fc g *- a "" 9 *" •a Ti -a 13 tj 3 c fc c E i&^fi c £ •o , Ti •d a; T! T TITJT) - T •a a = Ji.ar=j3 -^.-z: ^i-r^: 4.-:=^ Si ^.B^-^^ ==■3.3 ^=5 ^a^-sl^g^^.a^^- s-S"*-? s ^^l^.a 1= ^ 0 XI «ja « S J3 (n rt,a rt j: rt J3 ra « M to ^ S OJ ^ Efl W »?\ W Cfi CO W W U} t/3 ;5 W U3 W r ■tgis SsSa COC/DCO^ t: Cfl ja 0 "J Oj^ a.ii B.a t/3t/}y}cr 1 E CO 2 0 ■"• "-r f-O *»0 N-O "T'-''OO0*O 0>M mOO w ,5„«- VD H M 0 0 t^ Or^{>MO»Or» o« 1^ 0 M 0 l«-u is§£ l«8 !» >< t> M to 00 i/^fOWV(Owi t^ -o ^, fill ls°- M (1 M oOMt-n'OvOf*) 0 0 M M M OS S3 .; -0 li M >h 6 Tf V, £l M MfOfoior^'OOi 00 r* w n 0 0 3s fO r> fO %0 t MM rf n M M j2i otal mber seed- js re- ning. g % H d- CI 2 0 °- s e ■ 4J be e §z > a 0 — M u^ < M c 1 t Wl a •■ •• " •• P 132 Journal of Agricultural Research voi. in. No. 2 In Table VIII it will be seen that the seedling stand before grazing was rather dense, the average for all quadrats being 100.8 per square meter. Fortunately the less valuable species, such as sickle sedge, were the ones most weakened by drought, their condition, as will be noted, corresponding closely with the character of the soil, the more porous types supporting the less vigorous plants. After grazing (Table IX) the average stand was reduced to 48.5 per square meter, a loss of 50.9 per cent. The heaviest loss was at the upper limit of the area, where the short growing season caused the seed- lings to be less deeply rooted. At the lower elevations the heaviest loss was where the young plants were cropped. Many of the lateral roots were pulled out or broken, and death followed. On account of the rather severe and uneven grazing, it was practically impossible to determine definitely what species were best able to with- stand trampUng. An examination of the root systems showed that smooth wild rye had almost invariably pushed its roots more deeply into the soil than any other species. Short-awned brome-grass also develops an unusually strong, deep, and spreading root, and showed ability to withstand trampling and to recover its vigor when portions of the root were pruned off below the surface, or even when segments of the rootlets were exposed to the air. Seedlings of mountain bunch- grass also withstood trampling comparatively well, notwithstanding the fact that at the time the range was cropped it was not so far advanced as the other species. At High Elevation. — Owing to a great variety of conditions at the high elevations, 62 quadrats were established late in August before grazing. The range, which has a minimum altitude of approximately 7,500 feet, is distinctly herbaceous, the growth consisting primarily of grasses, with mountain bunch-grass and western porcupine grass pre- dominating in the order named. (See text fig. 5.) In addition, there are several species of sedges and rushes, with a sprinkling of weeds and non- grasslike plants, especially in the moister situations. At the time of the first observations nearly all the seedlings were in good condition, though dead individuals were often found in the drier situations. In other cases the terminal portions of some of the leaf blades were dead, but this did not necessarily indicate a weakened con- dition of the plant. The seedling stand, before and after grazing, is shown in Tables X and XI. Nov, 16, 1914 Revegetation of Range Lands 133 J ^ > § •3 0 s a a n •5 « Is Is a to ^ o 01^ :S ^ S S C3 a ^ ^ to to ^.9 « bC & ti ec JJ ja J3 J3 ja J3 ja c ^a a U ^ a J3 £0^ ■9 S-g.9 iJ o-o a ■3^-3 "rt !i ■3 ■SS'S 32 3 2 !s3 3 0 a §3 3 ^wS S^y5^ St g SBS «1 H f^ C.r. H '? S irt r^ fO 0« i § t-* C j3 in— • Q" 'S-o 3 y .3 .3 .3 q °"d a -a a -a a T) •3 a """S *^"3 o^-2 i* S^ 33 33 S3 " 1 1^ "tS S y5 S "cfi S w w p » a a ^ s;^ 8 3 3 8 Is ■S ai =3 § D 0° 0 eg > .9 ^ §■3 -a feH a "' a 0) t m3 Si3 3& 8" -n & irt" & 2 ? j=3 ^a Sin V,S r- 1| 1.a s. ja !» •:. a o O o O ■OS ,§ ,D •2f .E .9 3 9=9 S'o .S_ S ^_ tnO *» a 2 °- ■ I'.aa '5> a y a o Q ii'a a Safe gas t-0 c3 55! d 01 p "as 3 .9 d o O I s <^ a S3 3a°-a a, o-a ; Di . o . M O to-O ••a C.-0 fl S ^ w E «2 ll :Sm bo CO •Si 3 B t.a 4- o 5) 2 a o 3 C ft CO CI. a-o 0-3 S! S'S » aa-o Ei 2 at; ? o ■:5 'J ^ ^ tJ V Q V : ^2 u O. O' a in' a 0.-0 ° 5 rt 2 S '" I O r*5 (*) r^ n (1 3 -S^ o ^ "" a > _ > > *; s I w ►J n • P a 4) S3 uT 3 g oi" ^ & »i" a iz Si u u SM.3 a M S§S3-H . jE5.a-3S„- ^ ^ S _ Si E-° &S 4 a.g „3 .if 3i 11 3 2 '5. w Bs §1 & - c In «2 nil 6.U a& . pA aj 3 « 3^ 4» g agoOloagg SsS2gS2B|i:.gS t> t> !> P , U CQ .g-l rt 3 o a S 3 a tifi Si l§ 1-9 •82 I" .gl-j 5 si Nov. i6, 1914 Revegetation of Range Lands 135 !^ 0. §^ ^ ■^ > 2 O 3 £ a K i^ 3-0 1^ S! 21 .. ^1 ja e^^ f. •a Q n .Sf m£: 0 u S > 0 « ^U ! - 1 ifl I/) - I , jn . n rt « . ] oj 1 ftO 60 ^ i •fa ■ f W S) • u a a u :§ :55§ : ■ dT) 0-0. OTJ — 4) 0/ f ti aj ^2 4; fco 3§ go 2o b ."! ^ PS' "5 "5 0 •- 2 " Qa I E 2 rt a ° S "^ s « & ffl " a " ^8 ^§S. 3 0.-O g 0.0 j t^ in ra u 3 ^- a a ■ ai.Sja -^ 3 irt p .- O O.J O 33 ■I O ji; o . . . _ c/5 c/i !> .^ c/3 cfi !> c/2 :^ g-3 S! s-3 G-p c: as, la wen at) .°-.g fe 2 f5 P* 3.3 S ■°S5s§ fe an U3 "■ TJ-rt SS 3^ mS. baQ •3 8 •A OS a bfi &b a 3 3 " g2 j3 . a '^ a «?& gS.5 ag-o 2g^- Nov. i6, 19x4 Revegetation of Range Lan Is 137 M La [O <« & d bub §31 n m in V) Oi-O N "< CO O 3 -s -g-l § §.§§ & CS & d H "5 ' .aS'.s'S .9 .s'S.9 2' • BHam 2 a^ac : _ -. Ui (fl I. ■aas-a-Sa-g-c §5:S§93§t XI o «.-a ■8-Ss -Ss "S a p.S a.S g 3 3 a 3 a 5 X3X3 3X3 Xi .5.9 ' " B d •Ss §-a •^§E . o V •S'O , J3 OJ ■^-^u •?;>.« - 1 ^S'S ■S'H! P2l f >.a r3-S« "•3.- cncow 0 « « 138 Journal of Agricultural Research V* >0 tt.Ii< 0.S ■| " o tie u -3 1= OJ o « 2 ° .9 =; ^ SB; a e. > s •S X n < "t a bo -I o ^ U be ^ & U o u D. u o n a a g c 3 3 ii 3 3 d g ag.g '3'c3 P'rt'ffl q a w p q bid 3 3 " 3 3 i 3 o 0.0^ o o=a o 3 g ■ d n 3 3 : 3 3 3 a^i d d "■3 lU-3'3 a OJ d d oSS S "•o d Siia 4^o o d n !.g.9 . I d d •^3 5 u o o •S'l ■Sid's ii g2i! gs ^ a s-^ass esEa! ^ O ?i d 4^ = (#>- W) W O « f^ N bo^ ba a* K 4» a ?g a-g'a eg3 CO o O-d =• ie-sE 01 g a S I ffi a o 2*" 3i« •o ij-i r* lo "O -o 00 M g 5 m i .6. 3>-3J 1^ (fl'3 cat: a a be s > f- "* tn '^ ui ^ 3 a 3 q ^ •C t. 3 i_ 3t3 fe « 3 w) 3 o O ro M o, M o &si fl-1 5ib - a J c^ Q o,c"o p 3 I §2 " &s 01 CI - (=5 u o 5 « =33 140 Journal of Agricultural Research Vol. III. No. 1 2 "S^ •S5 3 B a & 3- la ss V 3 tfi O .Ho 11 '= a J2 *0 S •a ^ •o V o J3 -1 3 « 2f S 3 o 3 s s ■■3 a'-ss 8 Sa £••0. ^ 1 C bo ^ a fc.t: 2S 3j3 e _ 4J o p n J-. g t 3 .t; P rt o z: c-r «>■; fc". *n n ™« □ .2 S &.2 CO J; 3 ca □ 'B. 3&. H-a. a^ji a ao •^3 2 .il o g O-O 3 .. _ _ 0-^ S 0 o ci; n n 02 CO a c t- ^ 3^ "^ >^ t o tig b&ti (J ^* O o Q^ n 3 n a a • 3 ^; 3 3 3 ■ ^ a caBTja-o gC3^tQc4c9(/)Cdwi p a S c B nii aii t 3 S 3 3 3.!^ 3.M CIJO.^000000 3 1 TJ H r~ O 1 IP « So V t~ ^ t- 00 r- 00 O j ^ i o a s S.2 ooag 9 a a s Nov. i6, 1914 Revegetation of Range Lands 141 C.3 "3 c B3a .•=3» O CO &« t a a § .2 H > D ^ 00 ■ > 1 W m tn 2 «" fog MM2 Mg t^ =■ t^ '-' =■ ^ 2" ■3 S'3'3 g'3 g ■"■3 ■o >o 00 00 & fr* 0 *n n 3 r^ f^ 0 00 1-1 ^ rn ^ 4 » VI Ol •r 0 ** 00 " i 62697°— 14- 142 Journal of Agricultural Research voi. iii, No. 2 It will be seen that both before and after grazing the more abundant seedlings were of mountain bunch-grass, smooth wild rye, short-awned brome-grass, sickle sedge, little bluegrass, and western porcupine grass. Table X shows that before grazing there was an average for all quadrats of 155.3 seedlings per square meter. After grazing (Table XI) this was reduced to 80.5 seedlings, a loss of 48.2 per cent. A comparison of the height and root development of the same seedling species at medium and high elevations discloses the fact that the shorter and later growing season of the high range had not been conducive to the rapid development made by the seedlings in the lower and warmer, though drier, situations. Though the high ranges were grazed much more moderately, the loss was practically the same as on the lower ones. The factors responsible for the heaviest loss of seedlings through grazing were (i) superabundance of soil moisture, (2) lack of soil moisture, (3) abnormally dense seedling stands, and (4) irregular topography. In wet situations the roots did not penetrate as deeply as in the more compact and drier soils, and so were more easily disturbed. Although in some cases the seedlings recovered, the loss on *he moist soils was relatively large, in fact occasionally five times that on the dry soils. In excessively dry situations the loss from grazing was often serious, owing to the relatively weak growth made by the seedlings and their poor recuperative power. In some places practically the entire seedling stand was destroyed. Doubtless many of the seedlings whose destruc- tion was charged to grazing would have perished in any event from drought, though the stand as a whole was not affected by this factor. The lack of \ngor in individual plants where the stand was unusually dense often caused a heavy loss. Quadrats with from 250 to 500 seed- lings almost invariably suffered more than the contiguous plots which carried a sparser stand. In general, it may be said that more than 200 seedUngs to a square meter is a heavier stand than most situations can support permanently. Competition is most severe between plants of the same species, since each plant makes the same demand upon the habitat. Where a dense stand occurred it was usually of a single species. The loss from trampling was much more severe on steep slopes than on level situations. This was largely the result of the coarser texture of the soils on steep slopes and of the greater extent to which they are shifted by grazing. Moreover, because of the lack of soil moisture on many of the steeper slopes, the plants growing there are generally less vigorous than those growing in more level places. TIME AND INTENSITY OF GRAZING AFTER THE FIRST YEAR Though the information here presented shows that the range upon which deferred grazing was practiced suffered heavy loss of seedlings when moderately grazed, it should not be concluded that in order to Nov. 16, 19M Revegetation of Range Lands 143 insure permanent improvement grazing must be suspended from the time the first seed crop is produced until the seedlings become established. Notwithstanding the fact that half the stand in existence in the autumn is likely to be eliminated by grazing, the planting of an additional seed crop will, as a rule, fully offset this loss. The time at which the seed crop of the established vegetation reaches maturity, marking the approximate limit of growth and occurring in the region studied about September i , is the beginning of the period when the range may be grazed with the least injury to forage seedlings. During the four weeks prior to this period the root system almost doubles its growth and strength. On account of the much more elaborately developed root system at the end of the second year of the seedlings' growth, the loss through grazing at that period perceptibly lessens. Even then, however, the range should not be grazed prior to the maturity of the seed crop. Restriction of grazing to the period following seed maturity will give both the i- and 2-year-old plants sufficient protection to insure the restocking of the range. To sum up the conclusions regarding deferred grazing, it may be said that the system has proved highly successful wherever an adequate seed crop was produced. Its advantages over yearlong grazing and yearlong protection are (i) the restoration and maintenance of the vegetation without the loss of the forage crop in any year, (2) the planting of the seed, and (3) the removal of the vegetation itself, thus minimizing the fire danger from an accumulation of inflammable material. Deferred grazing has all the advantages of complete protection, so far as the rejuvenation of the weakened plants is concerned; and, if overstocking and abusive management are guarded against, the system will work no material injury to forest reproduction or watersheds. It is believed, therefore, that the principles of deferred grazing, with what- ever modifications are necessary to meet local conditions, should he applied to the management of all ranges. APPLICATION OF DEFERRED GRAZING vSYSTEM TO RANGE MANAGEMENT WHERE APPLICABLE If grazing lands are to be fully revegetated within a reasonable time, the range lands must, of course, support at the outset at least a sparse stand of the species valuable for grazing and revegetation purposes. In the Wallowa Mountains, where mountain bunch-grass constitutes the predominating herbage, a satisfactory seed crop and subsequent seedling stand were secured where the original tussocks stood as far apart as 6 feet. Where grazing has been so severe as to destroy the major por- tion of the original vegetation, the remaining plants may not produce 144 Journal of Agricultural Research voi. in. no. 3 viable seed until after they have regained their lost vigor, a matter of one or two seasons. On other sparsely vegetated lands, however, a stand of from 15 to 30 seedlings per meter, which is a satisfactory density on most soils, has been secured after the first year of deferred grazing. The benefits of deferred grazing are not confined to areas which sup- port plants of strong seed habits or those where the climate is particularly favorable to growth. Though on areas near and above timber line, where most of the forage plants reproduce vegetatively instead of by seed, deferred grazing does not tend to augment vegetation as it does on areas where the plants reproduce by the latter method, it does result in a permanent increase in vigor of the range plants and so promotes vegetative reproduction, which otherwise would be held in check by the premature removal of the herbage each season. In short, given a sufficient number of the original plants, deferred grazing is applicable wherever the vegetation is palatable after the seed crop has ripened and where water facilities will permit the range to be used in the autumn. Before the deferred grazing system was thoroughly tried out, certain stockmen maintained that after seed maturity the palatability and nutritiousness of the herbage would be low and therefore that the sea- son's forage crop would not only be wasted, but stockmen might be induced to keep their animals in the mountains until so late in the season that on account of the resultant loss of weight they would not be able to market them direct from the summer range. To determine definitely the nutritive value of the forage after seed maturity, chemical analyses were made of the foliage of mountain bunch- grass, first, when the flower stalks were being produced, and again, at the time the seed ripened. The average of the tests showed that the young growing plant is 27.21 per cent richer in ether extract (fat) than the mature plant, while the latter slightly exceeds the former in protein (nitrogen). The mature plant also contains more crude fiber, but since the flower stalks are not consumed after the seeds are ripened that part of each specimen was eliminated from the tests. In com- parison with timothy hay, mature mountain bunch-grass contains 94.39 per cent more protein, practically the same amount of ether extract, and 50.45 per cent less crude fiber, the last-named material being prac- tically indigestible. Nearly all the leading range plants, particularly the grasses, are grazed during the autumn with relish. It can not be said, however, that they are eaten with the same gusto after seed maturity as when they are growing vigorously. It was found that the first time a band of sheep passed over a matured range of medium density only about half of the forage crop was grazed off. Not until the range was grazed a second or third time was the crop entirely consumed. The vegetation on similar ranges grazed a month earlier was in most cases entirely con- Nov. 16, 1914 Revegeiation of Range Lands 145 sumed the first time the stock passed over it. On ranges grazed after seed maturity the naked flower stalks, rising from leafless tufts of bunch- grass, remained after the stock had passed over them, but on ranges grazed when the forage was succulent and tender no flower stalks were visible after the passage of the stock. No appreciable amount of herbage remained on either area. Sheep from several allotments where deferred grazing was practiced made fully as good progress as other sheep in allotments not handled under deferred grazing. By the time the seed has ripened, the milk flow of the ewes is nominal, and though it may decrease slightly when the animals are placed on the semi-air-cured forage, the lambs by this time are 4 or 5 months old, and milk is secondary to the nourishment secured through cropping. Deferred grazing does not materially change the character of the for- age on mountain ranges after seed maturity, because by this time succu- lent forage everywhere has been reduced to a minimum, lea\'ing only the air-cured plants and on open grazing lands a small amount of second growth. By protecting part of the range until the last few weeks of the grazing season there is the possible advantage of having a reserve supply of solid feed upon which to harden the stock prior to the drive to market or to winter range. The water facilities of the range may be an important consideration in determining whether or not to adopt the deferred grazing system. Regardless of the palatabiUty of the forage, deferred grazing can not be carried out unless there is an adequate supply of water. On many ranges the water faciUties may be improved by the construction of dams, the development and protection of springs, and even by digging wells and building windmills. Springs and small mountain streams are often replenished by the autumn rains. SELECTION OF LANDS The amount of range needed for grazing under the deferred system depends upon (i) the time at which the seed of the important forage plants matures and (2) the portion of the grazing season remaining after seed maturity. In the mountains of northeastern Oregon one-fifth of the grazing season remains after seed maturity. Accordingly, one-fifth of the carrying capacity, but not necessarily of the total acreage, of each grazing allotment may be reserved annually for purposes of revegeta- tion. The lower the elevation the earlier, of course, does the seed mature and the greater the proportion of range which must be reserved for deferred grazing. Since the lands are usually grazed by camps, the carrying capacity of which is well known, the user will have no difficulty in determining what proportion of the range should be reserved. 146 Journal of Agricultural Research voi. iii. no. > MANAGEMENT DURING THE REVEGATION PERIOD Once the area in need of revegetation has been selected, no stock should be allowed to graze on it until after the seed has ripened. Efforts should then be made to have the stock pass at least once over the entire area reserved, in order thoroughly to plant the seed. In the second year of deferred grazing if a reasonably dense stand of forage seedlings has been secured, abusive herding must be avoided. While it may do no apparent harm, so far as future seed crops are con- cerned, to fully utilize the forage in the fall after the first year of protec- tion, the loss of seedlings, even when the range is only moderatelj' grazed, amounts to about 50 per cent. Close grazing and carelessness in per- mitting the stock to btmch and trail must necessarily increase this loss. Therefore, while close grazing after seed maturity the second year may result in increasing the forage seedling stand the following season, such an increase could only be temporary, since the practice causes severe loss among the seedlings already in existence. Moderate grazing after seed maturity also, of course, results in the destruction of a large number of seedlings, but the double advantage of utilizing the forage and plant- ing an additional seed crop readily offsets this loss. Moderate grazing should be practiced in the second and subsequent seasons until the plants have reached full maturity and are permanently established. In the case of perennial plants this period is three years. When the area selected has been thoroughly reseeded and the plants permanently established, another area in need of reseeding should be selected. This system should not come to an end when the range has been completely reseeded, but should be kept up in order to thoroughly maintain the vigor of the vegetation and allow for an occasional seed crop. During the season of 191 2 deferred grazing was in effect on 10 allot- ments in various portions of the Wallowa National Forest. In every case the carrying capacity of the range has increased materially, and the best interests of the stock industry seem to call for the adoption of the system generally. SUMMARY (i) Normally the spring growth of forage plants begins in the Hud- sonian zone about June 25. For each 1,000 feet decrease in elevation this period comes approximately seven days earlier. (2) In the Wallowa Mountains the flower stalks are produced approx- imately between July 15 and August 10, while the seed matures between August 1 5 and September i . (3) Even under the most favorable conditions the liability of the seed on summer ranges is relatively low. (4) Removal of the herbage year after year during the early part of the growing season weakens the plants, delays the resumption of growth, Kov. i6. 1914 Revegetation of Range Lands 147 advances the time of maturity, and decreases the seed production and the fertility of the seed. (5) Grazing after seed maturity in no way interferes with flower-stalk production. As much fertile seed is produced as where the vegetation is protected from grazing during the whole of the year. (6) Germination of the seed and establishment of seedlings depend largely upon the thoroughness with which the seed is planted. In the case of practically all perennial forage species the soil must be stirred after the seed is dropped if there is to be permanent reproduction. (7) Even after a fertile seed crop has been planted there is a rela- tively heavy loss of seedlings as a result of soil heaving. After the first season, however, the loss due to climatic conditions is negligible. (8) When 3 years old, perennial plants usually produce flower stalks and mature fertile seed. (9) Under the practice of yearlong or season-long grazing both the growth of the plants and seed production are seriously interfered \vith. A range so used, when stocked to its full capacity, finally becomes denuded. (10) Yearlong protection of the range favors plant growth and seed production, but does not insure the planting of the seed. Moreover, it is impracticable, because of the entire loss of the forage crop and the fire danger resulting from the accumulation of inflammable material. (11) Deferred grazing insures the planting of the seed crop and the permanent establishment of seedling plants without sacrificing the sea- son's forage or establishing a fire hai&rd. (12) Deferred grazing can be applied wherever the vegetation remains palatable after seed maturity and produces a seed crop, provided ample water facilities for stock exist or may be developed. (13) The proportion of the range which should be set aside for deferred grazing is determined by the time of year the seed matures. In the Wallowa Mountains one-fifth of the summer grazing season remains after the seed has ripened, and hence one-fifth of each range allotment may be grazed after that date. (14) The distribution of water and the extent of overgrazing will chiefly determine the area upon which grazing should be first deferred. (15) After the first area selected has been revegetated it may be grazed at the usual time and another area set aside for deferred grazing. This plan of rotation from one area to another should be continued, even after the entire range has been revegetated, in order to maintain the vigor of tlie forage plants and to allmv the production of an occasional seed crop. ^■^ PLATE XII Fig. I. — View of the lower grazing lands in the Wallowa National Forest. The timber is western yellow pine. The exposed situations are covered with a dense growth of big bunch-grass {Agropyron spicatum). Fig. 2. — Characteristic open stand of western yellow pine and dense cover of herba- ceous vegetation, mainly pine-grass {Calamagrostis pubescens), Wallowa National Forest. Transition zone (yellow-pine association). Fig. 3. — A burned-over area of lodgepole pine, with characteristic dense sapling stand. <,I48) Revegetation of Range Lands Plate XI Journal of Agricultural Research Vol. Ill, No. 2 Reveijetatinti uf R.-uv^e Lajids Plate XIII Journal of Agricultural Research Vol. Ill, No. 2 PLATE XIII Fig. I. — Dense stand of lodgepole pine, with undergrowth of red huckleberry (Vac- cinium scoparium). Canadian zone (lodgepole-pine association). Fig. 2. — ^A flat eminence in the Hudsonian zone, showing the characteristic clumped growth of whitebark pine and Alpine fir. The glade land was formerly densely vege- tated with mountain bunch-grass. Fig. 3. — Irregular topography of the upper grazing lands. On northerly exposiues snow often remains until August. Photographed on July 25, 1908. Hudsonian zone (whitebark-pine association). PLATE XIV Fig. I. — Arctic-Alpine and upper subalpine region, where forage is sparse, due to poor soil, short growing season, and unfavorable climate. Fig. 2. — Mountain range lands prior to the beginning of growth and germination. Photographed on Jime 21, igo8. Fig. 3. — Same view as shown in figure 2, but more in detail, showing the condition eight days later (Jiuie 30). The conspicuous plant in the foreground is spring beauty {Claytonia lanceolata), which closely follows the recession of the snow and announces the earliest approach of spring. Revegetatiuii ^f RarlJ;^r Lands Plate XIV Journal of Agricultural Research Vol. Ill, No. 2 Revegetation uf Rarit^e Lands Plate XV Journal of Agricultural Research Vol. Ill, No. 2 PLATE XV Fig. I. — Contrast in the progress of tlie flower stalk production of mountain bunch- grass on portion of range which has been completely closed to grazing for a period of three successive years and on range which has been subject to continued early grazing. Section of fence temporarily removed. Photographed on July 7, 1909, before grazing. Fig. 2. — Western porcupine grass (Stipa occidentalis) , showing empty glumes and floret with the scale and its awned projection to the left; to the right the floret with glumes removed, showing the sharp-pointed, slightly-curved seed tip. Natural size. Fig. 3. — Average development of the root system and aerial portion of mountain bunch-grass at end of the first growing season. Natural size. PLATE XVI Mountain bunch-grass, showing root development and aerial growth at the end of the second season. Natural size. Reveeetation of Range Lands Plate XVI Journal of Agricultural Research Vul. Ill, No. 2 Revegetatiori uf Range Lands Plate XVII Journal of Agricultural Research Vol. III. No, 2 PIRATE XVII Mountain bunch-grass in the spring of the third year of growth just before producing flower stalks, showing the natural position and length of the elaborate root develop- ment and aerial growth. Natural size. PLATE XVIII Mountain bunch-grass at the end of the third year, showing three flower stalks and inflorescence. Natural size. Revegetation of Rany,e Lands Plate XVlll Journal ot Agricultural Re^eaf^ VlI. in, N.. 2 Revegetatiun Dead. 62697°- 1 62 Journal of Agricultural Research [voi. rii, no. 2 As will be observed from the data given, the results were largely negative. There were no signs of recovery, and in the case of stable manure (plot 10), lime (plot 12), and cottonseed meal (plot 13), there was an increase in the number of resetted trees. It should be noted that most of the trees having rosette at the start were well-advanced cases. At the same time, the subsoil was dynamited around three 15-year-old resetted trees in the orchard of Mr. G. G. Gibbs. The results in this case were likewise negative as shown by observations after one and two years. The results of dynamiting the subsoil in these two orchards, together with obser\'ations showing the absence of pecan rosette in swampy land, seem to indicate that the disease is not due directly to lack of proper subsoil drainage. The soil around two resetted trees in the J. B. Wight orchard, Cairo, Ga., in the spring of 1907, was treated with copper sulphate and mag- nesium sulphate at the rate of i pound of each for every inch in diameter of the trunk. The trees were decidedly injured by this treatment, and the diseased condition continued as before. A negative result followed the use of 8 pounds of copper sulphate around a single 9- year-old resetted tree at Bacenton, Ga. The soil around each of 22 resetted trees in the 4- year-old Davenport orchard at Belleview, Fla., was treated in June, 191 2, with 2 to 4 pounds of copper sulphate, according to the size of the tree. By the following midsummer 7 trees had recovered, 7 were somewhat improved in appear- ance, and 8 were either in the same or in a worse condition than at the beginning. In the same 40-acre block, 117 out of 389 resetted trees had recovered in. the same period. In ether words 31 per cent of the trees treated with copper sulphate recovered and also 30 per cent of the untreated trees in the same block. A considerable number of the untreated trees were also improved in appearance. Both groups of trees were fertilized by the owner with a complete commercial fertilizer at the rate of 8 pounds to the tree. The copper-sulphate treatment has from time to time been recom- mended by a number of orchardists, and in a few cases observed by the junior author some apparently beneficial results have occurred. But the usual failure of the grower to run proper checks with an experiment, to- gether with the fluctuation of the disease without any treatment, lends a rather doubtful character to the results. At any rate, this treatment is not to be recommended, except in an experimental way, until further tested out. SPRAYING EXPERIMENT Five resetted trees in the J. B. Wight orchard at Cairo, Ga., were given three applications of Bordeaux mixture — in March, April, and May, 1903. No positive results from this experiment were discernible. Nov. i6, 1914) Pecan Rosette 163 ORCHARD RECORDS Rosette records for periods varying from 2 to 1 2 years have been kept in several orchards in Georgia, Florida, and South Carolina. Of 159 trees observed in the J. B. Wight orchard for three successive seasons (1902 to 1904), 24 were healthy for all three seasons, 17 were healthy at the beginning, but later contracted the disease, 13 had rosette, but recovered, 108 were rosetted throughout the period, and 7 fluctuated each season. Eighty trees in the same orchard were observed in 1902, 1903, 1904, 1912, and 1913. So far as these observations go, 30 trees remained nor- mal throughout the period, 7 were normal, but contracted the rosette, 12 were rosetted, but recovered, 8 were rosetted throughout, and 21 fluctu- ated back and forth between the normal and rosetted condition. Observations on 274 trees in the same orchard during 1912 showed 136 trees with the rosette, and during the following season 35 trees more had contracted the disease. The cultivation and fertilization had not been varied. In a 40-acre block of the Davenport orchard at Belleview, Fla., con- taining 1,069 trees, 389 had the rosette in 1912 and 256 in 1913. All had received the same treatment, except for the few trees used in the copper- sulphate test, and as previously noted this treatment gave negative results. Of the three varieties present, 19 per cent of the Van Deman, 25 per cent of the Stuart, and 28 per cent of the Teche pecans had rosette in 1912. Sixty-two soil borings were made to 6 and 9 feet from the surface and in the vicinity of both healthy and rosetted trees. Of 31 trees in a clay to sandy -clay subsoil, 13 had the rosette and 18 were normal. Of 23 trees in a sandy subsoil, 9 were rosetted and 14 were normal. Of 8 trees in a subsoil containing considerable quantities of a soft lime rock, all were healthy. In the first two groups the difference in the number of diseased and healthy trees was not conspicuous. On account of the entire absence of rosette in the third group, partial analyses of the subsoil around three trees were made. In sample No. i the percentages of lime, magnesium, and phosphorus, computed as the oxids, were 9.68, 0.82, and 8.32, respectively; in sample No. 2, 5.42, 1.09, and 6.03; and in sample No. 3, 0.58, 0.99, and 3.44. More or less clay was present in all the samples, and in No. 2 there was a considerable admixture of creolin. In a small orchard of 96 trees belonging to Mr. G. G. Gibbs, at Talla- hassee, Fla., 38 trees were rosetted in 1912, and during the following season 4 of these recovered. In another orchard of 1 16 trees on the same fann, with similar topography and apparently similar soil, 6 were rosetted in 1 91 2 and 12 the following season. In the G. W. Saxon orchard of 231 trees at Tallahassee, Fla., 125 trees liad rosette in 191 2 and 173 in 191 3. Most of the orchard had received 164 Journal of Agricultural Research (voi.iu.no.j no fertilizer for several seasons, and where fertilizer had been applied no distinct difference in the rosette could be detected. In the W. P. Bollard orchard near Albany, Ga., 291 out of 646 trees were rosetted in 1912. By the following season loi more trees had con- tracted the disease. Fertilization and cultivation were uniform for the two seasons. In a block of 233 trees belonging to Mr. H. K. Miller, at Monticello, Fla., 81 trees had the rosette during 1912 and 1913. Fertilization was uniform for both years, and the soil was a sandy loam underlain by a rather stiff, red, sandy clay. Observations were taken during 191 2 and 191 3 in three blocks belonging to the Standard Pecan Co., at Monticello, Fla. Out of 406 trees observed in the first block 231 had rosette in 1912 and 254 in 1913. In the second block of 450 trees 119 were rosetted at the first observation and 121 at the second. In the third block of 570 trees, 42 were rosetted in 1912, and 34 in 1 91 3. A part of all three blocks were in swamp land that had not been fully drained, and the absence of rosette here was conspicuous. The soil in the remainder of the three blocks was a sandy loam underlain by a red clay or sandy clay. For three seasons preceding 191 2 a com- plete commercial fertilizer was used. In 191 2 a small application of stable manure was made around each tree. In the small orchard of 100 trees belonging to Dr. R. B. Gamett, at St. Augustine, Fla., 3 were rosetted, both in 1912 and 1913, 7 were rosetted but recovered during the second year, and 6 new cases developed during the second year. The soil is a sandy loam underlain by a clear- sand subsoil, with the water table 3 to 3K feet below the surface. Fer- tilization was uniform over the orchard, except that the soil around the 10 trees showing rosette in 191 2 was treated with lime and copper sul- phate. Seven of these trees recovered, but no checks were run to verify the result. In a block of 371 trees belonging to Mr. M. O. Dantzler, at Orangeburg, S. C, 146 trees had rosette in 1912 and 137 in 1913. Fertilization and cultivation were uniform during the two seasons. It is evident from these records that rosette fluctuates from year to year without any variation in the treatment given by the grower and that diseased trees may apparently make a complete recovery and remain healthy for an indefinite period, or after a season or two they may again contract the disease. It should be stated here that in the majority of cases the trees recovering from rosette had not reached the staghorn stage. However, a considerable number of trees with terminals dying back from the disease have been seen to recover and remain normal through the one or more seasons they have been subsequently under observation by the authors. From the variations in rosette recorded from year to year under uniform cultivation and fertilization it seems Nov. i6, 1914] Pecan Rosette 165 highly probable that seasonal climatic changes, such as variation in the water content of the soil, may have at least an indirect relation to the prevalence of the disease. ASH ANALYSES' Complete ash analyses were made of resetted and healthy leaves and of resetted and healthy twigs from Cairo and Dewitt, Ga., and from Belleview, Fla. In each case the diseased and healthy material was from trees of the same age and variety and had received similar cultivation and fertilization in the same orchard. The pure-ash analyses are given in Tables III to V. Table III. — Ash analyses of leaves and twigs of the Stuart pecan from. Cairo, Ga., September, igia Constituent. Leaves. Twigs. Resetted. Normal. Resetted. Norma], Per cent. 5-59 3.60 15-47 23-36 .98 31-48 21. 50 3-25 Per cent. 5- 02 3.02 21. II ■17 18.35 •83 34-09 14.09 8.33 Per cent. 5-47 3-69 12. 26 .26 23.01 I. 18 34- 14 24-45 .96 Percent. 5-49 3- II 16.59 • 14 SO3 P„On ci K2O NajO CaO 48. 19 14-78 1.24 MgO SiO, Total 100. 04 100. 09 100. 04 Table IV. — Ash analyses of leaves and twigs of the Van Deman pecan from Belleview, Fla., September, iqij Constituent. Total percentage of pure ash SO, PA CI K2O NajO CaO MgO FI2O3 AI2O, MiijOj SiOj Total Leaves. Resetted. Normal. Per cent. 4.68 7-31 7-57 •38 18.76 •38 37-75 2o. 66 I. 90 .40 4.83 99.94 Per cent. 4. 61 5 8. 16 32 20. 14- Stems. Resetted. Normal. Per cent. 3.68 6. 19 8. 24 ■3° 21-35 2.23 44. 16 17.40 Trace. •23 .26 98-36 Per cent. 3-37 4.69 12. 24 .40 25.61 3-24 36.56 16. 71 Trace. .26 .29 ' The ash analyses here given were made by the Bureau of Chemistry, Department of Agriculture. 1 66 Journal of Agricultural Research [Vol. ni. No. 3 Table V. — -4rA analyses of leaves and twigs of the Schley pecan from Putney, Ga., September, IQ13 Constituent. Total percentage of pure ash SO3 P2O5 CI K,0 NajO CaO MgO FejOs AI2O3 MnsO, SiOj Total Leaves. Rosetted. Normal Per cent, 4. 72 6.76 6.23 .69 23-03 •49 35- 12 16.74 3-03 I. 19 6-45 104. 45 Per cent. 4.78 5-83 6. 52 .26 14. 12 •34 42.04 16. 42 4.84 2.44 7- 23 104. 82 Stems. Rosetted. Normal, Per cent. 2. 16 6.49 6. 40 •38 22.37 2.82 40. 04 16.9s 1-43 .61 I. 76 loi. 41 Per cent. 2. 18 s- 7- 19. 2 42 19. 28 32 03 S' 27 23 09 74 loi. 99 Most of the differences between the pure ash of healthy and of rosetted material were not constant in the three sets analyzed. In both leaves and twigs from Cairo, Ga., the magnesium content was much higher in the diseased material, but in the two other sets the percentage was nearly the same in both diseased and healthy material. The percentage of phosphorus was greater in the normal leaves of two sets and in the normal twigs of one set. In the other cases the percentage was about the same in both healthy and diseased material. The calcium content was greater in the normal leaves and twigs of two sets and considerabh' less in the remaining set of material. The percentage of potassium was greater in all the rosetted material, with the exception of one set of rosetted and normal twigs, where there was slightly more potassium in the normal. Other differences shown by the analyses are either slight or greatly variable. DISCUSSION OF RESULTS PARASITISM VERSUS NONPARASITISM The following experimental data have a bearing on the question of possible parasitism or upon nonparasitism. Out of 144 nuts collected in different localities from badly rosetted branches, 91 germinated and none of the seedlings gave symptoms of the disease during the two seasons under observation. These nuts were placed in the greenhouses of the Department of Agriculture at Washing- ton, D. C, which is far removed, both as to locality and environment, from the orchards where the nuts were obtained, but the conclusion from Nov. i6. 1914) Pecan Rosette 167 this experiment nevertheless seems justifiable that whatever the cause of the disease it is not transmissible through the seed (see p. 155). Inoculation of tender growing tips with bits of diseased tissue gave negative results. Attempts at isolation of micro-organisms from living rosetted material gave negative results in all cases. Various fungi and bacteria were isolated from partially dead material obtained in four dif- ferent orchards, but no constant form appeared and inoculations gave negative results (see p. 156). These results strongly support the view that pecan rosette is not of a parasitic nature. Mycorrhiza have been found promiscuously on both normal and rosetted trees, or perhaps more often the roots of both normal and rosetted trees have appeared to be free from any such fungous growth. It is probable that the disease is not due to the presence or absence of mycorrhiza (see p. 157). Where normal buds or cions were worked upon rosetted stocks they have invariably developed the disease during the same or the following season, except in a few cases where the trees used as stock had in the meantime themselves recovered. With few exceptions rosetted buds and cions worked upon apparently normal nursery and orchard trees de- veloped into normal shoots. Where rosetted shoots developed their percentage was no greater than in contiguous nursery rows or orchard trees worked to normal buds or cions (see p. 158). The results here tend to show that rosette is not caused by a perennial fungus mycelium, by bacteria, or by any infecting virus within the tissues of the pecan. Healthy nursery trees set in large pots of top soil and of subsoil taken around badly rosetted trees remained normal during the two seasons under observation (see p. 155). This series of tests was carried out in the Department greenhouses at Washington, D. C, small quantities of the soil being used under entirely different environment than present in the location from which obtained. The results are therefore not taken as conclusive, but merely as tending to indicate the nonparasitic nature of the disease. Transplanting of healthy nursery trees to holes from which rosetted trees had been removed gave a very high percentage of rosette in the replants, while transplanting badly rosetted trees to situations where less rosette or none at all had been observed gave a very high percentage of recovery (see p. 152). By comparison of the results of these two tests it appears that a soil relation is the important factor in causing rosette rather than any transmission of the disease from one tree to another. Thus, from the nontransmission by seed, the negative results of isola- tion and inoculation tests, the varying presence and nonpresence of mycorrhiza, the budding and grafting tests, and the transplanting work, the nonparasitism of pecan rosette is considered a reasonable assumption. 1 68 Journal of Agricultural Research [vo1.iii,no.2 RELATION TO THE SOIL The relation of pecan rosette to the soil is partially elucidated by the following experimental data. Severe pruning of rosetted trees or cutting back to a stump gave nega- tive results (see p. 152). The new growth was often healthy in appear- ance, but toward autumn or during the following season symptoms of rosette invariably reappeared (PI. XXVIII, fig. 2). The fact that in no case was the disease eliminated by this treatment is at least not unfavor- able to the view that it is contracted directly or indirectly through the soil. Transplanting of healthy nursery trees to holes from which rosetted trees had been removed gave a very high percentage of rosette in the replants (see p. 152). These results point toward the conclusion that some relation exists between rosette and the soil, either directly through the soil itself or through previous infection of the soil by rosetted trees. The transplanting of badly rosetted trees to situations where less rosette or none at all has been observed gave a very high percentage of recovery (see p. 153). All rosetted trees replanted at Washington, D. C, recovered, though many of them were dying back with the disease when first taken up (see p. 154). In the latter cases the entire change of both soil and climate is held accountable for the uniform recovery of all rosetted trees. All these transplanting tests with rosetted trees tend to indicate the soil relation as the direct cause rather than infection of the soil with a parasitic organism or virus from previously diseased trees. The results of the fertilizer experiments were partially negative (see p. 159). However, in a i6-plot test upon normal trees at Baconton, Ga., 9 out of II limed plots developed cases of rosette, while i out of 5 unlimed plots showed doubtful traces on two trees bordering a limed plot. Care- ful observation at the beginning of the experiment did not reveal a single case of rosette, and at the last observation no rosette was found in con- tiguous parts of the orchard. The plots receiving lime, acid phosphate, and nitrate of soda and those receiving lime, acid phosphate, and muriate of potash developed by far the most rosette. In fertilizer experiments at two other points considerably more rosette developed on limed than on unlimed plots. These results, while not in all cases very definite, have at least tended to show that some relation exists between rosette and the constituents of the soil and that the lime content alone or in combination with other substances may have a varying causative effect. Dynamiting the subsoil in two different orchards gave negative results (see pp. 161-162). This fact, together with observations showing the absence of rosette in swampy land (see p. 150), seems to indicate that the disease is not directly due to lack of proper subsoil drainage. Orchard records of individual trees over periods of 2 to 12 years show considerable fluctuations in the disease, irrespective of treatment and without any special treatment (see p. 163). Trees with a mild or mod- erate attack frequently recovered, and even when the staghorn stage Nov. i6, 1914] Pecan Rosette 169 had been reached, trees occasionally regained their normal appearance without any artificial treatment. From the seasonal variations in rosette recorded from year to year under conditions of uniform cultiva- tion and fertilization it seems highly probable that seasonal climatic changes, such as variation in precipitation, may have at least an indi- rect relation to the prevalence of the disease. In large orchards the more or less simultaneous appearance of rosette in patches, together with its usual limitation to these areas, clearly suggests some definite connec- tion with the soil conditions. In general, the ash analyses of resetted and healthy material showed slight to highly variable differences, so that very little positive light was thrown upon the problem by this part of the work (see p. 165). In both leaves and twigs from one orchard the magnesium content was much higher in the resetted than in the normal material, but in two other sets of analyses the percentage was nearly the same in both diseased and healthy material. The percentage of phosphorus was greater in the normal leaves of two sets and in the normal twigs of one set. In the other cases the percentage was about the same in both healthy and dis- eased material. The calcium content was greater in the normal leaves and twigs of two sets and considerably less in the remaining set. The per- centage of potassium was greater in all the resetted material, with the exception of one set of resetted and normal twigs, where there was slightly more potassium in the normal. Other differences shown by the analyses are either slight or greatly variable. It thus appears from the results of experiments in pruning and cutting back, transplanting tests, fertilizer experiments, results of subsoil dyna- miting, and orchard records that the disease is directly or indirectly caused by some soil relation. On account of their variable character, the ash analyses shed little light on the problem. COMPARISONS WITH OTHER DISEASES Leaf-hopper injury suggested itself at one time as a possible cause of rosette of pecans. The rather far-reaching effects of the work of this insect were known, and the demonstration of their causal relation to curly-top of beets ' lent further plausibility to this theory. However, extensive observations have failed to disclose any connection between this insect injury and rosette. Leaf-hopper injury has occasionally been seen on pecans, but its symptoms are distinct and it has occurred both in the presence and absence of rosette. The leaves are often yellowed around the margin, somewhat curled, and if attacked while young, their growth is considerably interfered with. But there are not the distinct yellow mottling over the whole blade between the veins, the raised ap- pearance of the latter, and the tendency to reduction of the leaf blade, followed by the dying back of the shoot from the tip. 'Shaw, Harry B. The curly-top of beets. U.S. Dcpt. of Agr., Bur. Plant Indus., Bui. 181,46 p., jfie., 9 pi. 1910, lyo Journal of Agricultural Research [voi.ni.No.^ Neither is rosette to be confused with sun scald, or "winterkill," which affects young trees especially and manifests itself in the death of the cambium at the base of the trunk. The roots are usually uninjured and healthy sprouts soon put out from the base. Again, frost injury may kill back the tips in a manner somewhat analagous to the staghorn stage of rosette, but here the foliage symptoms are lacking, and in rosette the death of the branches is usually most prominent in late summer, a season far removed from frost. Both the yellows and rosette of peach in a general way suggest pecan rosette. However, though some symptoms may be common to all three diseases, the complete clinical picture is distinct in each case. Some of the most prominent symptoms of peach yellows consist in the produc- tion of abnormally long, spindling branches in dense groups due to the putting out of normally dormant and possibly adventitious buds. Yel- lowing of the leaves always occurs at some time during the course of the disease; but in spite of the name "yellows," the leaves are often abnor- mally dark green in a case of recent attack. The pushing into growth of normally dormant buds and also possibly of adventitious buds is likewise characteristic of pecan rosette, but the axes are shortened rather than elongated as in peach yellows. In this shortening of the axes of growth the resemblance to peach rosette is seen. In the last-named disease the twigs are so much shortened that the leaves as they develop become clustered into a compact rosette. Peach trees affected with rosette usually die during the first and never survive the second season, and those suffering from yellows rarely survive more than five seasons, but rosetted pecan trees have been known to live for 15 years. More- over, no case is on record of a recovery from peach yellows or peach rosette, while recovery from a moderate attack of pecan rosette is fre- quent, and not rare, even for the later stages. Furthermore, from the experiments outlined in this paper it appears that pecan rosette is not transmissible either through the seed or by budding or grafting, while the infectious nature of peach yellows and rosette is well established through experimental bud transmission to healthy trees. A striking resemblance is to be observed between pecan rosette and ordinary chlorosis of various trees. In moderate cases of the latter dis- ease yellowing of the leaves occurs without notable change in form or size, and the conspicuous dying back of the branches from the tip is lacking, but all gradations occur between such cases and those where the symptoms closely simulate rosette of pecans. The spike disease of pineapples' bears some general resemblance to rosette of pecans both as to effect and apparent cause. The leaves of affected pineapples become contracted at the base so as to form a spike- like blade, and the color of the leaf also becomes modified. It is claimed that cottonseed meal, sulphate of ammonia, kainit, muriate of potash, ' Rolfs, P. H. Pineapple lertilizers. Fla. Ap. Exp. Sta. Bui. so, p. 97-99. "899. Nov. i6, 1914I Pecan Rosette 171 and acid phosphate aggravate the disease, while nitrate of soda and sulphate of potash seem to have a beneficial effect upon diseased plants. The fertilizer tests with pecans point toward the conclusion that rosette is also a nutrition trouble. PROBABICO CO^CO I II CO I II II NH — C — NH NH— CO— NH NH— CO— NH NHj Tetracarbonimid was prepared from uric acid according to the method of Scholtz and its properties compared with the compound obtained from the soil. The two preparations were found to be identical in crystalhne form and solubility, in being precipitated from solution by metallic salts, and in their behavior on heating. Their identity was further confirmed by the analysis of the barium salt and the determination of the nitrogen content. The preparation from the soil was found on analysis to contain the following percentage of nitrogen: Sample No. i, 32.59 per cent, and sample No. 2, 32.47 per cent of nitrogen, the theoretical nitrogen content of C4H4N404 being 32.61 per cent. The barium salt prepared by pre- cipitating a hot, saturated solution of the soil compound with barium hydroxid was found to contain 61.55 P^r cent of barium, the barium content of C^BajN^Oj being 62 per cent. These figures, together with the correspondence in properties, are sufficient to establish the identity of the soil compound as tetracarbonimid. Nov. i6, 1914I Nitrogenous Soil Constituent 177 Tetracarbonimid was first obtained from a sandy soil from Florida devoted to orange culture. This soil was one of the series of 16 samples referred to in a previous paper (1914), and in which vanillin and other benzene derivatives were found. Tetracarbonimid was found in three of these soils in suiEcient quantity to make its identity certain, and in six other samples crystals having, so far as could be ascertained, the same properties as tetracarbonimid, were obtained in small quantities by the same method. This compound was also found in three out of four soils from other locations than Florida which were examined for it. These soils were (i) a sample of Norfolk sandy loam from Virginia, (2) a Sassafras soil from the grounds of the Department of Agriculture, and (3) a sample of Elkton silt loam from Maryland. These results indicate that tetracarbonimid is not an uncommon or accidental soil constituent. The soil from the Department grounds contained more of this com- pound than the others examined, and from 18 kg. of this soil 30 mg. of pure tetracarbonimid were obtained. The loss in purification in this separation was at least 50 per cent, so that, assuming that the treatment with sodium hydroxid extracted all of this compound, there was present per acre-foot of soil approximately 7 pounds of tetracarbonimid, repre- senting 2.3 pounds of soil nitrogen. This soil had a total nitrogen content of 0.13 per cent, or approximately 5,200 pounds of nitrogen per acre-foot, and it appears that the quantity of tetracarbonimid nitrogen is at any one time but a very small part of the total. However, in this investigation some evidence has been obtained indi- cating that the quantity of tetracarbonimid fluctuates under varying conditions of cultivation or crop growth. If, then, this compound is formed from other compounds under certain soil conditions and its dis- appearance by oxidation or other means, such as use by plants, is influenced also by conditions that vary, it may, in spite of the small quantity present at any one time, be an important step in the trans- formations that organic nitrogen undergoes in the soil. Tetracarbonimid has so far not been reported in any plant or animal tissue and has been made only by the mild oxidation of uric acid. Uric acid has not been found in soils or plants, and while it might be added to soils in human excrement, it is not a common constituent of manures or fertilizers; therefore it would seem that the source of tetracarbonimid in soils must be sought in some other compound. The close relation of uric acid to the purin bases suggests these as possible sources of this compound, and there is no theoretical reason for assuming that any of the purin bases might not under proper condi- tions of oxidation yield tetracarbonimid. This possibihty is disclosed more clearly by consideration of the structural formulas. Uric acid Xanthin (2.6.8. oxypurin) (2.6. oxypurin) NH-CO NH-CO II II CO C-NH. CO C-NH. I II /CO I II \CH NH-C-NH/ NH C- N *^ 178 Journal of Agricultural Research voi.iu.No.ai Hypoxanthin Adcniii (6 oxypurin) (6 amino-purin) NH-CO N=C.NHj II II CH C-NH. CH C-NH. II II ^CH II II \CH N - C- N *^ N-C- N '^ Guanin ■ (6 ammo-6 oxypurin) NH-CO I I NH2.C C-NHv II II JCH N - C- N*^ Xanthin, hypoxanthin (Schreiner and Shorey, 1910), and adenin (Shorey, 1913b, p. 16) have been found in soils, and guanin has been found in a soil that had been heated (Lathrop, 1912). All four bases have been found in plants and may also be formed from nucleic acid, a constituent of all nucleated cells of both plants and animals. LITERATURE CITED Lathrop, E. C. igi2. Guanine from a heated soil. In Jour. Amer. Chera. Soc., v. 34, no. 9, p. 1260-1263. ScHiTiENHELM, ALFRED, and Wiener, Karl. 1909. Carbonyldiharnstoff als Oxydationsprodukt der Hamsaure. In Ztschr. Physiol. Chem., Bd. 62, Heft i. p. ioo-io6. SCHOLTZ, M. 1901. Ueber ein neues Oxidationsproduct der Harnsaure. In Ber. Deut. Chem. Gesell., Jahrg. 34, Bd. 3, No. 17, p. 4130-4132. Schreiner, Oswald, and Shorey, E. C. 1910. Pyrimidine derivatives and purine bases in soils. In Jour. Biol. Chem., V. 8, no. 5, p. 385-393. Shorey, E. C. 1913a. Recent work on the chemical composition of humus. (Abstract.) In Jour. Wash. Acad. Sci., v. 3, no. 9, p. 260. 1913b. Some organic soil constituents. U. S. Dept. Agr., Bur. Soils Bui. 88, 41 P-. I pl. 1914. The presence of some benzene derivatives in soils. In Jour. Agr. Research, V. I, no. 5. P- 357-363. APPLE ROOT BORER By Fred E. Brooks, Entomological Assistant, Deciduous-Fruit Insect Investigations, Bureau of Entomology INTRODUCTION While engaged in observations on the larval habits of the roundheaded apple-tree borer {Saperda Candida Fab.) in West Virginia during the year 191 1, the writer noticed in apple {Mains spp.) numerous burrows of some smaller insect associated with those of the former. Infested wood was collected and adults were reared, which were determined by Mr. E. A. Schwarz, of the Bureau of Entomology, as Agrilus viilaticollis Rand., a species of beetle which hitherto had not been recognized as an enemy of cultivated fruit trees. Further observations showed that the species is quite generally distributed throughout the Appalachian fruit region and that in places it is doing considerable damage to young apple trees. The literature pertaining to this insect is meager. The species was first described from Massachusetts by J. W. Randall in 1837 and received the name which it now bears. In 1841 it was redescribed by Gory, in his monograph of the buprestids, as Agrilus jrenahis. Le Conte, in 1857, lists the species as "A. jrenatus Gory," with the note, "unknown to me." Again, in 1859 the same writer, in his Revision of the Buprestidae of the United States, uses both names as synonymous, giving neither the preference. In 1871 Edward Saunders in his Catalogus Buprestidarum, gives Randall's name the preference over Gory's. Since that time the species has been mentioned occasionally in lists of Buprestidae as Agrilus ■viUaiicollis Rand. E. P. Austin, in 1875, seems to have been the first to associate the insect with a host plant, at which time he notes: "Found occasionally in various parts of the State [Massachusetts] and seems to live on shad- berry (Amelancbier canadensis)." In 1889 Frederick Blanchard, in his list of the Buprestidae of New England, records that in Massachusetts the beetle is found occasionally in June, feeding on the leaves of thorn, service, or shadbush {Amelanchier canadensis) and chokeberry. A specimen of the beetle in the United States National Museum was col- lected by W. F. Fiske in June at Tryon, N. C, on leaves of Oxydendrum. The species is also recorded from Michigan, Pennsylvania, and New Jersey, and it probably occurs throughout the greater part of the eastern United States. At the suggestion of Prof. A. h- Quaintance, in charge of deciduous- fruit insect investigations for the Bureau of Entomology, a further study of the insect and its habits was made, the results of which are re- corded below. NATURE AND EXTENT OF INJURY The injury to the trees is done by the slender, white larva of the insect, which bores through the sapwood and heartwood of the roots and lower trunk. The burrows through the roots frequently extend outward for Joun al of Agricultural Research. Vol. Ill, No. a Dept. of Asriculturc. Washington, U. C. Nov. i6. 1914 (179) K-i. i8o Journal of Agricultural Research Vol. Ill, No. a several feet and in badly infested trees are so numerous that the roots often die, causing a weakening of the whole tree. The work of the insect is obscure, there being no chips or castings thrown to the surface, as is the case with the common roundheaded apple-tree borer. The egg, which is placed rather conspicuously on the bark of the trunk, and the exit hole in the bark through which the adult escapes from the wood, are the only external marks made by the insect on the tree. In addition to the injury resulting from the dam'aged roots, the exit holes in the bark admit more or less water, which frequently induces decay of the heartwood. The harm one individual root borer is capable of doing to a tree is much less than that usually done by a single roundheaded apple-tree borer, but in the localities where investigations were made the former species outnumbered the latter. At French Creek, W.Va., 345 apple, pear, and service, or shadbush, trees, from yi inch to 5 inches in diameter were cut off a few inches above the ground, and the burrows made by the larvae of both species in ascend- ing the trunks to pupate were counted. The three species of trees are favorite food plants of both of these borers, and the results of the counts, recorded in Table I, give an idea of the relative abundance of the two borers in the one locaUty. Similar results were obtained in several other localities where fewer trees were examined. Table I. -Relative abundance oj burrows of Agrilus vittaticollis and Saperda Candida in apple, pear, and service trees Kind of trees. Apple Pear Service Total Number of trees. 125 20 200 34S Number of burrows of Agrilus. 3" 9 342 662 Niunber of burrows of Saperda. lOI o Most of the apple and pear trees cut for the foregoing counts were such seedlings as could be found in roadside fence corners and neglected orchards, and the service trees were in the woods. Only 36 of the 125 apple trees, or about 28 per cent, were free from the burrows of Agrilus and 87, or about 70 per cent, were free from burrows of Saperda. FOOD PLANTS The adult of this insect has been recorded by various writers as fre- quenting the foliage of service, wild thorn, and chokeberry. The ^vTiter has found the larv^a attacking apple, pear, wild thorn, wild crab, and service. It has been found rather abundantly in the following West Virginia localities: French Creek, Cherry Run, Sleepy Creek, Spring- field, Moorefield, Romney, Keyser, Flkins, and Junior. Mr. E. B. Blakeslee, of the Bureau of Entomology, has found the larva in service trees at Winchester, Hayfield, and Staunton, Va. Of the several larval food plants named, apple and service seem to be greatly preferred above the others. Nov. i6, 1914 Apple Root Borer 181 THE BURROWS The burrows of the apple root borer (PI. XXIX, figs, i, 2, and 3) are of great length, that of a single larva often extending through the trunk and roots for 5 or 6 feet and in some cases even 8 feet. Throughout its length the burrow is packed with dustlike particles of wood that have passed through the alimentary canal of the insect. Cross sections of the burrows are oval in outhne, those of full-grown larvae being about i by 3 or 4 mm. in diameter. Except for the discoloration of the dustlike packing, time does not change the appearance or form of the burrows in the wood until they are obUterated by decay. Fresh burrows are usually found within half an inch of the bark, and in large trees the position of a burrow in the wood gives some clue to its age. In the large number of old trees cut and examined during this investigation the finding of a large proportion of the burrows comparatively near to the bark is taken as an indication that this insect has been on the increase for the last 10 or 15 years. LIFE HISTORY OF APPLE ROOT BORER THE KGG The eggs (PI. XXXI, fig. i) are somewhat variable in outline, an average specimen being flattened, oblong, disklike, 1.3 by 1.9 mm. in size. The surface is irregularly corrugated, the ridges and depressions being more distinct near the margin. When first laid, the egg is white or creamy white, but in a few days it changes to grayish brown, similar to the bark in color, and resembling in a general way the shield of a small scale insect. After the dark color is assumed, the embryo shows as a whitish spot at the center of the egg. In the latitude of West Virginia the eggs are deposited in May and June. They are glued tightly to the bark of the trunk a few inches above the ground singly (PI. XXX, fig. 3, b) or, rarely, in pairs. They are usually attached to a smooth surface, but are sometimes inserted into a crack or beneath a scale of bark. When the lar\'^a hatches, it leaves the egg from the underside and enters directly into the bark, thrusting its castings backward into the discarded shell and so filling it that it retains its normal size and shape. The abandoned shell often adheres to the bark for a year or longer. THE LARVA The larva (PI. XXXI, fig. 2), before settling in the pupal chamber, is long, slender, distinctly segmented, flattened, and of nearly uniform width throughout, the first thoracic segment being slightly wider than the others. Full-grown specimens average from 3 to 4 by 30 to 36 mm. The color is white, except the small head and the anal forceps, which are bro\vn. The anal forceps are prominent and shade from light brown at the base to dark brown or black at the converging tips. After entering the pupal cell, the lar\'a contracts to about half its former length and in this condition (PI. XXXI, fig. 3) it remains during the second winter of its Ufe in the tree. The larva lives in the tree for nearly two years, and during this time its movements are substantially as follows: On leaving the egg it bores i82 Journal of Agricultural Research voi. m, no. 3 directly through the bark to the cambium, thence through the cambium down the trunk to the ground, whence it proceeds outward through a convenient root. After boring in this manner through the cambium for a distance of from 6 to 12 inches it burrows abruptly into the solid wood, where all the feeding throughout the remaining part of the larval stage is done. For most of its length the burrow through the cambium fol- lows in a general way the grain of the wood. It sometimes zigzags across the grain directly after leaving the egg and invariably winds around the root in a spiral course once or twice just before leaving the cambium to enter the solid wood of the root. The burrow in the cam- bium soon heals and is scarcely discernible a year after it is made. After burrowing into the solid wood of the root, the larva continues to feed outward from the tree. If the root is long enough, the burrow may continue toward the tip for a distance of 3 or 4 feet, after which it turns and is directed back toward the base. The larva spends its first winter well out from the trunk, often in a root not more than one-sixteenth of an inch in diameter. In penetrating these smaller roots it converts all the hard wood into powder during its outward and return journey. It is active late in the fall and early in the spring, and probably consid- erable feeding is done during the winter. With the coming of warm weather it feeds rapidly back toward the base of the root, and by mid- summer it has reached the center of the root system and has begun to ascend the body of the tree. The latter part of the summer and the fall are spent in boring upward through the trunk (PI. XXIX, fig. 2) and in fashioning a pupal chamber. In trees that are cjuite small pupation takes place within 5 to 10 inches of the ground, but in larger trees the lar\^ffi for some reason ascend higher before fonxiing the pupal cell. In apple and pear trees that are as large as 6 inches in diameter at the base of the trunk it is not unusual for the larvae to ascend 2 or 3 feet to pupate, and in one case an individual was found in a 12-year-old pear tree that had burrowed up from the roots and pupated in a branch 46 inches above the ground. The ascent through the trunk is usually made within half an inch of the inner bark, the larva occasionally approaching the bark but never entering it. The pupal chamber is a curved and enlarged terminus of the burrow, occupying a vertical position, with the convex side toward the heart of the tree. The upper end of the chamber curves abruptly to the bark, while at the lower end the curvature is more gradual and the chamber lacks one-sixteenth of an inch or more of extending to the bark. The two extremities are from i inch to iK inches apart, and when the chamber is completed both ends are packed with wood fragments, the insect occupying a space between. In forming its pupal quarters the lar\-a first extends a burrow of the usual transverse dimensions through to the bark at the upper end. It then recedes for about one-fourth of an inch and begins eating the wood from the side of the burrow. As this proceeds, the larva directs its excavation backward along the ventral side of its body and so doubles the diameter of the original gallery. The larva thus has its two ends folded closely together, and as the body moves forward the head and anus are soon contiguous. At this point the excavation ceases, and the head turns and follows the anus forward until both extremities are at the upper end of the chamber. Here the head rests, but the hinder Nov. i6, I9I4 Apple Root Borer 183 parts move downward until the axis of the body is again in line. At first the borer is much crumpled in accommodating itself to its restricted quarters, but a molt soon follows, after which all the segments of the body are so shortened that the space is adequate (PL XXX, fig. i, c). The larva does not get permanently settled in its cell until well into December, often after severe freezing weather has occurred. THE PUPA The pupa (PI. XXXI, fig. 4) is broad and flattened dorsally, the males being smaller than the females. An average female measures 3.5 mm. wide by 12 mm. long. The color is white, the eyes, mouth parts, and, later, most of the abdomen and spots on the thorax changing to black as the imaginal stage is approached. The pupa occupies a vertical position in the cell, with its ventral side toward the bark (PI. XXX, fig. i, b). Pupation takes place with the coming of the first few warm days of spring, occurring in the locality of the principal investigations from April I to 15. All individuals were found to have pupated in the valley of the Potomac River, near Moorefield and Springfield, W. Va., on April 20, 1912, while on the mountains a few miles distant, where the elevation was about 500 feet greater, no pupae could be found on the same date. In Randolph County, West Virginia, at an altitude of about 2,500 feet above the sea, pupation had occurred on April 25, 1913. The pupal stage lasts from three to four weeks. THE ADULT The adult (PI. XXXI, fig. 5) is a slender, elongate beetle, averaging 2,5 mm. wide by 10.5 mm. long. The elytra are black, usually with a tinge of purple ; the head and thorax are iridescent purple or cupreous in some lights. The sides of the thorax and a wide stripe on top extending over the front of the head are covered with a dense bronze pubescence; the underparts are black, with a metallic purplish or coppery luster; the dorsal surface of the abdomen beneath the elytra is dark blue. A line of bronze pubescence extends along the sides of the abdomen and is visible dorsally beyond the inflected edge of the elytra. The front of the head is deeply impressed; the thorax is broader than long, its sides being regu- larly curved; the disk has oblique depressions on each side; the elytra are broadened behind the middle, the tips being separately rounded and finely serrulate. The upper surface is finely and densely granulate. HABITS OF THE BEETLE The adult apple root borers escape through small, round holes which they gnaw in the bark (PI. XXX, fig. 3, a). They come forth in May, all the individuals of one generation issuing in a given locality during a comparatively short time. The beetles are active only by day, are rather quick to fly when disturbed, and have the buprestid habit of sunning themselves in exposed positions on bark and foUage. It is probable that the life of the beetle does not often exceed two or three weeks. This is indicated by observations in the field and by experiences with a large number kept in wire cages placed over small apple trees. All beetles confined in such cages died within about two weeks. 184 Journal of Agricultural Research voi. in, No. 2 HABITS OF OTHER MEMBERS OF THE GENUS AGRILUS The apple root borer, Agrilus vittaticollis , belongs to a group that is represented by several well-known enemies of cultivated and forest plants and trees. The raspberry gouty-gall beetle {Agrilus ruficollis Fab.), the two-lined chestnut borer {A. bilineatus Weber), the bronzed birch borer {A. anxius Gory), and the sinuate pear borer (^4. sinuatus Oliv.) are destructive pests whose food plants are indicated by their common names. The damage done by the different members of the genus is usually due to the larvae working in or about the cambium, quite often of small branches or twigs. The root-feeding habit is not known to be common, although there are at least two or three other species in the Eastern States whose habits have not been fully described, the larvae of which live in the roots of forest trees and shrubs. NATURAL ENEMIES So far as could be observed, the natural enemies of the apple root borer are confined to one species of hymenopterous parasite which attacks and destroys the larva and pupa (PI. XXX, fig. 4). The adult of this parasite was first reared in April, 191 2, and was described by Mr. H. L. Viereck' of the Bureau of Entomology, as representing a new genus, under the name Xylophruridea agrili (Pis. XXX, fig. 2, and XXXI, fig. 6). Two generations of this parasite occur annually, one brood of the adult appearing early in the spring and another late in the fall. Oviposition was not observed, but it is presumed that the female uses her strong external ovipositor to pierce the bark and wood, in order to get at her host. The spring brood of adult parasites oviposits on the root-borer pupae or on the larvae just before pupation, and the fall brood oviposits on the larvse at about the time the pupal cell is being formed. The parasitic larvae attack their host externally, those from the spring eggs developing rapidly, and, when full grown, constructing cocoons which occupy the host cell. This generation passes the greater part of the warm season of the year as larvae in the cocoon. The larvae from the fall brood of eggs develop less rapidly, some of them reaching full growth and constructing cocoons in the fall and early winter and others not entering the cocoon until very early in the spring. The adult parasites have difficulty in escaping from their transforming quarters, and many of them die in their effort to gnaw their way out through the bark. Parasitization is much more frequent with the borers that are in very small trees. This is doubtless due to the fact that the borers in such trees are forced to work nearer the bark, where they are more constantly within reach of the ovipositor of the parasite. On the whole, from 25 to 40 per cent of the root borers are destroyed by this parasite. REMEDIAL AND PREVENTIVE MEASURES On account of the larva's concealed method of feeding, the diggmg-out process, which is used effectively against several species of fruit-tree borers, is not practicable or possible in this case. Control measures must be directed toward the protection of the. trunk of the tree against the 1 Viereck, H. L. Contributions to our knowledge of bees and ichneumon flies. includiuR the descriptions of 21 new genera and 57 new species of ichneumon flics. Proc. U. S. Nat. Mus.. v. 42, p. 646. 1912. Nov. i6, 1914 Apple Root Borer 185 deposition of the egg rather than the killing of the borer after it begins feeding. Where paints, washes, or mechanical devices of any kind are used on trees as a preventive of injury by the roundheaded apple tree borer, equal protection may be had against the apple root borer by treating the trunks at about the time fruit is setting in the spring. The egg-laying season is of short duration, and temporary wrappers of paper or burlap, or any other material that will entirely cover the lower 2 feet of the trunk for a period of four or five weeks following the blooming season of apple, will in a large measure prevent eggs from being placed on the bark. Treatment with sticky adhesives or heavy paints that are not injurious to the trees will answer the same purpose. The apple root borers develop freely in the common service tree, and the proximity to apple orchards of woods in which this tree flourishes may always be regarded as a source of possible infestation to the orchard. PLATE XXIX Figs. I and 3. — Sections of an apple root, showing burrows of the apple root borer {Agrilus villalicollis). About natural size. Original. Fig. 2. — Cross section of the trunk of a young apple tree, showing biurrows made by the larvae of the apple root borer in ascending the trunk to pupate. Natural size. Original. (186) Apple Root Borer Plate XXiX Juurnal of Agricultural Research Vol. Ill, No. 2 Apple Root Borer Plate XXX Journal of Agricultural Research Vol. Ill, No. 2 PLATE XXX Fig. I. — Agrilus vitiaticolUs: Larva (c), pupa (6), and adult (a) of the apple root borer in the pupal cell. Slightly enlarged. Original. Fig. 2. — Xylophruridea agrili, a common parasite of the apple root borer. Greatly enlarged; natural size, i8 mm. Original. Fig. 3. — Section of trunk of young service tree, showing below the white egg and above the exit hole of the apple root borer. Natural size. Original. Fig. 4. — Xylophruridea agrili: LarviE of the parasite; one feeding on the larva and the other in the pupal cell of its host. Enlarged about one-third. Original. 62697°— 14 8 PLATE XXXI Fig. I. — Agrilus ■vittaticolUs: Fig. 2. — Agrilus vittaticolUs: Fig. 3. — Agrilus vittaticolUs: X 6. Original. Fig. 4, — Agrilus vittaticolUs: Fig. 5. — Agrilus vittaticolUs: Egg on trunk of young service tree. Original. Feeding form of larva. X 4- Original. Contracted form of larva as taken from pupal cell. Pupa. X 8. Original. a, Adult, or beetle; b, claw; c, antenna, a, X 5; b, c, highly magnified. Original. Fig. 6. — Xylophruridea agrili, a parasite of the apple root borer; Pupa. X 8. Original. Apple Root Borer Plate XXXI Journal of Agricultural Research Vol. III. No. 2 ADDITIONAL COPIES OF Tins PUBUCATION MAY BE PKOCUBED FROM THE SUPERINTENBENT OF DOCUMENTS GOVERNMENT PRTNTTNG OFFICE WASHINGTON, D. C. AT 2S CENTS PER COPY StmscRiPTioN Per Year, 12 Ntjmbees, $2.50 Vol. Ill DECEIVTBER, 1914 No. 3 JOURNAL OF AGRICULTURAL RESEARCH CONTENTS Page Changes in Composition of Peel and Pulp of Ripening Bananas 187 H. C. GORE Assimilation of Colloidal Iron by Rice .... 205 P. L. GICE and J. O. CARRERO Coloring Matter of Raw and Cooked Salted Meats . 211 RALPH HOAGLAITO Oil Content of Seeds as Affected by the Nutrition of the Plant 227 W. W. GARNER, H. A. AIXARD, and C. L. FOUBERT Studies in the Expansion of Milk and Cream . . . 251 H. W. BEARCE Life History of the Melon Fly 269 E. A. BACK and C. E. PEMBERTOR Identification of the Seeds of Species of Agropyron . . 275 ROBERT C. DAHLBERG DEPARTMENT OF AGRICULTURE WASHINGTON , D.a WASMMtaTOH S aOVERMMEUT PDWrNa OFFICE S I>I4 PUBI,ISHED BY AUTHORITY OF THE SECRETARY OE AGRICULTURE, WITH THE COOPERATION OE THE ASSOCIATION OF AMERICAN AGRICUL- TURAL COLLEGES AND EXPERIMENT STATIONS EDITORIAL COMMITTEE irOR THE DEPARTMEUT FOR THE ASSOCIATION KAItL F. KELLERMAN, Chairman RAYMOND PEARL Phrtiologisl and Assistant Chief, Bunau Biologist. Maine AarictUtural Experiment of Plant Industry EDWIN W. ALLEN Assistant Director, O^ice Of Bxptriment . Slatiotu CHARLES L. MARLATT Astidaml Chief, Bureau of Entomology Station H. P. ARMSBY Director, Institute of Animal Nutrition, Tfu Pennsylvania State Collece E. M. FREEMAN Botanist, Plant PaiholoQist, and Assistant Dean, Affricultural Experiment Station of the University of Minnesota All coJTESpondence regardmg articles fron;> the Department of Agriculture should be addressed to Karl F. Kellermaa, Journal of Agricultural Research, Washington, D. C. All correspondence regarding articles fron> Experiment Stations should be addressed to Raymond Pearl, Jotimal of Agricultural Research, Orono, Maine JOURNAL OF AGEETIIAL RESEARCH DEPARTMENT OF AGRICULTURE Vol. Ill Washington, D. C, December 15, 1914 No. 3 CHANGES IN COMPOSITION OF PEEL AND PULP OF RIPENING BANANAS By H. C. Gore Chemist in Charge, Fruit and Vegetable Utilization Laboratory , Bureau of Chemistry INTRODUCTION In attempting to determine the effect of temperature on the rate of respiration of the banana {Musa sapientum) , it was found that the rate of respiration increased very rapidly during ripening(i3) ', as, indeed, had been observ^ed by Gerber (12), and the bananas maintained themselves during ripening at temperatures distinctly above that of their sur- roundings, in consequence of the intensely active respiration. The phenomenon of self -heating suggested the need of experiments in a cal- orimeter to determine the amount of heat evolved in relation to carbon dioxid formed and oxygen consumed. The necessity of accompanying such studies with analyses of samples of the bananas under observation led to the analytical work recorded in this paper. The chemical changes which occur in the banana during ripening have been studied by Buignet (5); Corenwinder (8, 9, 10), Marcano and Muntz (17); Ricciardi (20); Doherty (11); Colby (7); Gerber (12); Bal- land (4); Atwater and Bryant (i); Chace, Tolman, and Munson (6); Bailey (2, 3); Prinsen Geerligs (18); Tallarico (21); Jahkel (14); Yoshi- mura (22); and Reich (19). The results have invariably been expressed in terms of the percentage of the pulp of the fruit when analyzed, and the data are, in consequence, on a constantly shifting basis, as the peel continuously loses weight and the weight of the pulp steadily increases. An accurate account of the chemical changes for use in exact biochem- ical studies was therefore lacking. Four ripening experiments were made. In two experiments bunches of green bananas were ripened in a large respiration calorimeter designed for experiments with man. In the third and fourth experiments studies were made, in a specially designed ripening chamber, of the uniformity 1 Reference is made by number to " Literature cited, " pp. 202-203. Journal of Agricultural Research, Vol. Ill, No. 3 Dept. of Acriculture. Washington, D. C. Dec. 15, 1914 (187) E-3 1 88 Journal of Agricultural Research voi. m. no. 3 with which ripening occurs in different bunches of bananas and of the rate of starch hydrolysis during ripening in relation to changes in the rate of respiration. The first and second experiments were made in cooperation with Messrs. C. F. Langworthy and R. D. Milner, Nutrition Investigations, Office of Experiment Stations, who measured the weight changes on ripening, the carbon dioxid, the water and the heat evolved, and the oxygen absorbed. The composition was determined before and after ripening. Green bananas selected from commercial shipments just received from the West Indies were used in all cases. FIRST EXPERIMENT Six bunches of green bananas were used. They were evenly matured, so far as could be judged from the comparative intensity of the green color. The top and the bottom of the stem of each bunch were trimmed smoothly, and all injured fruits whose pulp might have been infected as a result of mechanical injury were removed. For analysis, six fruits were cut from each of five bunches and eight fruits taken from the sixth bunch, which was larger than the others. The samples were taken from the inner and the outer portions of the "hands" at the top, the middle, and the bottom of each bunch, cutting them off where the short stem by which the bananas are attached to the stalks was most constricted. After ripening, samples were removed from the same "hands" from which samples had been previously taken, but they were cut from places removed from where the preceding samples had been attached, in order to avoid possible effects of local stimulus to adjacent fruits due to cutting. No indication of such effect, how- ever, was noticeable. The samples of green or ripe bananas were first weighed and then separated into peel and pulp. The peels of green bananas adhere closely to the pulp, and appreciable losses from evaporation are unavoidable. .\fter separation, the peel and the pulp were weighed, to determine their respective proportions and the losses from evaporation. To express the analyses in terms of the original bananas, it was necessary to correct for this evaporation. To this end the loss in weight on peeling was arbitrarily divided equally between the respective weights of peel and pulp, and the percentages of peel and pulp were then calculated. The samples of peel and pulp were ground by passing them through a food chopper. The methods of analysis used in this and succeeding studies were as follows : Solids were determined by evaporation in vacuo at 70° C. Ash and soluble alkalinity of ash were determined by the official methods. • ' Wiley, H. W., et al. OflBdal and provisional methods of analysis. Association of OSScial Agricultural Chemists. U. .S. Dept. Agr., Bur. Chera. Bui. 107 (rev.), p. j8, 102. 190S. Dec. 15. 1914 Changes in Composition of Ripening Bananas 1 89 Nitrogen determinations were made by the official methods' by Mr. T. C. Trescot, of the Bureau of Chemistry. Ether extract was determined by exhausting with alcohol, and de- termining the ether extract in residue and alcohol-soluble extract. In case of the alcohol-insoluble extract the alcohol was evaporated, the residue transferred to a separatory funnel by use of water and ether, and extracted repeatedly with ether. These extracts were then combined, washed with water, and evaporated in tared flasks. Carbohydrates were determined by treating weighed samples with 80 per cent alcohol in Soxhlet extractors, using several portions of alcohol so as to avoid giving the larger proportion of the sugars a prolonged heat treatment. The residues were dried and weighed, and the weighed por- tions used in the estimation of starch and pentosans. The extracts were mixed with a little calcium carbonate, evaporated on the steam bath in a current of air, avoiding evaporation to dryness, and then taken up in water, treated with lead acetate, and made up to a known volume. After filtering, the excess of lead was removed by use of dry sodium oxalate. Reducing sugar and sucrose were determined in this solution by applying the copper-reduction method of Munson and Walker,^ before and after inversion.^ Starch. To determine starch, 3-gram samples of the alcohol-insoluble material were transferred to 300 c. c. flasks, hydrolyzed with hydro- chloric acid as directed in the official method,^ and the dextrose deter- mined by the copper-reduction method of Munson and Walker.'^ Pentosans were determined in the alcohol-insoluble residues by the provisional method of the Association of Official Agricultural Chemists. ° Although at the time it was supposed that due precautions had been taken to prevent inversion of sucrose, unfortunately it is not improbable that more or less inversion may have occurred, owing to the failure to heat the alcohol extract to boihng. When removed, the bananas were of a bright clear yellow, with a little green showing at the tips. They were entirely free from decay. The fruit surfaces were moist and waxy, indicating that the humidity had been high. During ripening, the calorimeter record showed that the tempera- tures had varied from 18.8° to 20.4° C, average 20.1° C. The oxygen content of the air supplied varied from 20.4 to 10.8 per cent by volume, average 15.5 per cent, probably sufficient for normal ripening. The bananas weighed 137.71 kg. when placed in the calorimeter and 132.37 kg. when withdrawn, a loss in weight of 3.88 per cent. The ' Op. cit., p. 7. •* Op. cil., p. 41. 5 Op. cit., p. 241. 2 Op. cit., p. 241. * Op. cit., p. 53. "Op. cit.. p. 54. 190 Journal of Agricultural Research Vol. HI, No. 3 fruit consisted of 41.72 per cent of peel and 58.28 per cent of pulp when green and of 37.85 per cent of peel and 62.15 per cent of pulp when ripe. Calculated on the basis of the original bananas, the percentage of ripe peel was 36.38, a loss of 5.34 per cent of its weight. On the same basis the percentage of pulp was 59.73, a gain in weight of 1.45 per cent. Expressed in terms of the original bananas, the solids in the peel decreased from 5.21 to 4.88 per cent, the water from 36.51 to 31.50 per cent, about two-thirds of the starch passed into sugars, and the pentosans decreased slightly. The ash, the alkalinity of ash, the nitrogen, and the ether extract did not change materially. There was an apparent net loss in carbohydrates of 0.46 per cent. Expressed in the same way, the solids in the pulp decreased from 16.93 to 16.74 psi' cent. The water increased from 41.35 to 42.99 per cent. The starch content changed from 13.15 to 2.40 percent, the reduc- ing sugars from 0.37 to 10.34 P^r cent, and the sucrose from 0.48 to 1.52 per cent. There was a net loss in carbohydrates of 0.88 per cent, due to respiration. The pentosans changed from 0.40 to 0.18 per cent. As in the peel, the ash, the alkaUnity of ash, the nitrogen, and the ether extract in the pulp did not undergo marked changes in the amounts present. See Table I. Table I. — Composition of bananas before and after ripening in respiration calorimeter COMPOSITION EXPRESSED IX TERMS OF PERCENTAGE OF FEEI. OR POI,P WHEN" AKALYZBD 3 Composition. •s i. 0^ •^ . (A bt s: & * 0 Part of fruit. -o ao' X % "'■£ ■9 - 01 1 3 .c iH t. a Hk J3 S M P 1 .S.-2 1 Ph •I s-. < < £ w A % m < ^ Peel: Green bananas 41.72 87. « 12.48 1.32 0.82 0.61 0.89 0.70 0.23 4-84 8-98 0.92 37-85 86.59 13-41 1-54 1.04 .64 i.oS 3-71 1.94 6.24 .89 Pulp: Green bananas 58.28 70.95 29.0s •87 .49 1. 30 • 24 .64 .8, 22.56 26.28 • 69 62.15 71.98 28.02 ■8s ■ 47 1.24 •24 17-31 2-54 4.02 6.36 •30 COMPOSITION eXPRESSED IN' TERMS OP PERCENTAGE OP THE WHOLE GREEN BANANAS « Peel: Green bananas. Ripe bananas. . Pulp: Green bananas. Ripe bananas. . 41.72 36.51 5-21 o-SS 0-34 0.26 0-37 0.39 o-io 2-02 3-75 36-38 31- SO 4-88 .56 -38 ■23 -39 I-3S .04 •71 2.27 58.28 41-35 16.93 •51 .29 .70 .14 -37 .48 13- IS IS- 32 59-73 42.99 16.74 •51 .28 •74 .14 10-34 I-S2 2.40 3-80 0.38 -32 .40 .18 ,j Percentage of loss in weight on ripening. 3.J Dec. 15, 1914 Changes in Composition of Ripening Bananas 191 SECOND EXPERIMENT In the second experiment similar composition changes occurred. In addition, it was demonstrated that the banana stem does not contribute materially to the fruit during ripening. The details of this experiment are as follows: Four bunches of green bananas, consisting of the most evenly ripened fruit, judging by color, of a lot of eight bunches of the green fruit, were placed in the calorimeter after sampling and weighing. Six fruits had been removed as samples from each of two bunches of green bananas and eight from each of the others. During ripening the temperature varied from 20.9° to 24.2° C, averag- ing 23.1° C; the oxygen content of the air varied between 20.0 and 6.1 per cent by volume, averaging 14. i per cent; and the humidity remained high, as before. When removed, all of the fruits were thoroughly ripe, and indeed one bunch was slightly overripe. Several specimens on this bunch were beginning to spoil. The skins were bright golden yellow, and the fruit surfaces were waxy and moist. No browning occurred, except where the bananas were superficially injured. After weighing, the bunches were again sampled. In the analyses determinations of protein, ash, and ether extract were omitted, as the quantities of these substances present had been found to change but very slightly during ripening. The four bunches remaining after selection of the bananas for the respiration calorimeter were used in the study of the composition of the green stem. Mr. W. H. Evans, of the Office of Experiment Stations, suggested that possibly the stem contained reserve materials supplied to the fruit during ripening. The bananas were detached from the stems in the same manner as when taking samples for analysis, leaving on the stems the stubs of the short stems by which the bananas were attached. The percentage of stems was 5.04 per cent. They were finely divided in a shredding machine and analyzed, using the methods employed for the analysis of peel and pulp. At the conclusion of the ripening experiment in the calorimeter the weight of the stems of the ripened bananas was determined and the stems then ground and analyzed. The original weight of the four bunches of green bananas placed in the calorimeter after sampling was 75.90 kg. They suffered a loss upon ripening of 6.36 kg., or 8.38 per cent. The bananas after ripening consisted of 95.30 per cent of fruit and 4.70 per cent of stem. Assuming that the fruit and stalks lost weight in equal proportions, the respective weights of green fruits and stems when placed in the calorimeter were 72.33 and 3.57 kg. The propor- tions of pulp and peel of the green bananas were 58.22 and 41.78 per cent, respectively. The entire bunches of green bananas as placed in the calorimeter after sampling consisted therefore of 55.48 per cent of 192 Journal of Agricultural Research \ ol. in, No. 3 pulp, 39.82 per cent of peel, and 4.70 per cent of stem. When ripe, the bunches consisted of 63 per cent of pulp, 32.30 per cent of peel, and 4.70 per cent of stem. Based, however, upon the weight of the original bunches of green bananas, the proportion in ripe bananas was 57.72 per cent of pulp, 29.59 per cent of peel, and 4.32 per cent of stem, a total of 91.63 per cent. The analytical data are given in Table 11 in terms of percentage of pulp and peel as analyzed (corrected for loss in weight in peeling), in terms of the original whole bananas, and in terms of the entire bunch of bananas — i. e., including the stem. The ripening period was longer than in the first experiment in the calorimeter, more starch disappeared, and more sugars formed, while the gains of water in pulp and losses in peel were greater. The study of the composition of the stem before and after ripening showed that the changes which occur in it during the ripening process are so slight as to be insignificant. The percentage of loss in solids, on the basis of the original bananas, was but 0.04 per cent. Table II. — Composition of bananas before and after ripening in respiration caloriineler COMPOSITION EXPRESSED IN TERMS OF PERCENTAGE OF PEEL, PULP, AND STEM OF THE BANAN.\S BEFORE AKD AFTER RIPENING Percent- age of whole fruit. Composition. Part of fruit. Water. Solids. Reduc- ing sugar as invert. Sucrose. Starch. Alcohol- insoluble solids. Pento- sans. Peel: Green bananas Ripe bananas Pulp: Green bananas Ripe bananas Stem: Green bananas 41.78 33-89 58.22 66.11 "4.70 4.70 88. 23 Ss- 12 70.76 73- 00 90.19 88.97 II- 72 14. 88 29. 24 27.00 9.81 11.03 0. 76 4.42 ■71 13.81 •63 •30 0.02 . 12 .02 8. 50 ■ 17 .04 4-43 1.65 24. 10 I. 17 I-S9 1.83 8.S5 6.77 27.05 3-4= 7.02 8.72 0.86 .96 .68 •25 COMPOSITION EXPRESSED IN TERMS OF PERCENTAGE OF THE WHOLE GREEN BANANAS & Peel: Green bananas. Ripe bananas. . Pulp; Green bananas. Ripe bananas. . 41.78 36.88 4.90 0.32 0. 01 1. 8s 3-S7 31- OS 26.43 4-62 1-37 .04 ■SI 2. 10 S8.22 41.20 17.02 .41 .01 14.03 15-74 60.57 44.22 16.3s 8.37 S-iS • 71 2.07 0.36 • 30 COMPOSITION EXPRESSED IN TERMS OF PERCENTAGE OF THE WHOLE STEM OF GREEN BANANAS & Peel: Green bananas. Ripe bananas. . Pulp: Green bananas. Ripe bananas. . Stem: Green bananas. Ripe bananas. . 39. 82 29.59 55- 4S 57-72 4- 70 4-32 3S-I3 25.19 39.26 42.14 4.18 3-84 4-67 4.40 16.22 IS- 58 -52 .48 0.30 1-31 ■39 7-97 .03 .01 0.008 .04 .01 4.91 .01 .002 1.76 ■49 13-37 .68 .08 .08 3. 40 2.00 15.00 1.97 ■33 ■38 o It is here assumed that the same proportion of stem is present in the green as in the ripe bananas. ^ Percentage of loss in weight on ripening, 8.38. Dec. IS, 1914 Changes in Composition of Ripening Bananas 193 THIRD EXPERIMENT It now seemed well to observe the composition of the banana of com- merce repeatedly at several stages during its ripening, in order to de- termine the uniformity of the changes which occur in bananas from different bunches. Four bunches of bananas were ripened in a specially constructed humidity chamber and each bunch was sampled three times during its ripening. This experiment demonstrated that the changes in ripening in the four bunches were remarkably uniform, while the data secured in the first two experiments were repeatedly confirmed. In addition, the study gives a good idea of the composition of bananas when just ripe enough to be edible and also when very ripe. ©=:, ^^- ■ 'b .^^ J r- fe -r-'^ t i K'l T ;_, i v.L •■ 1 *• '■■f„ ° )B te ■=> ■ >, ^ — ^ ^ FiC. I. — Constant-lemperature humidor. The detailed account of the experiment is as follows: The humidity chamber (fig. i) consisted of a large case, of cold-storage construction, divided by an insulated wall into two compartments, A and B. Compartment A was so arranged that it could be kept at con- stant temperature and humidity at temperatures below that of the room. Ice was placed in compartment B. In A the air was circulated continuously by an electric fan. The course of the air current was as follows: From the fan it was deflected by the baffle C to play upon the surface of water in the tank D. The water was here kept at constant level by a device not shown and was warmed by an electrically heated immersion heating coil, also not shown. The air then passed up through the vertical cooling coil H. The air, moisture laden and cooled, was then delivered through holes in the false ceiling E to the rest of the compartment, whence it was drawn back to the fan through holes in the false bottom F. Detached bananas or other objects were placed on the false bottom F or on the slatted shelf G. When bunches of bananas were being ripened, shelf G was removed and the fruit was suspended from hooks in the false ceiling. 194 Journal of Agricultural Research voi. iii, No. 3 Compartment B was lined with galvanized iron. At the bottom the horizontal coil I was connected through the partition with the vertical coil H in compartment A. The ice rested directly upon coil I. To cool compartment A, tap water was supplied to coil I and the ice-cold water displaced in coil I passed into coil H. Automatic delivery of cold water to coil H was accomplished b)- the operation of the sounder J, which in turn was controlled by the thermostat (not shown) in compartment A. The form of the thermostat has already been illustrated (13). The low- tension circuit opened and closed by the thermostat operated a relay in which a iio-volt D. C. circuit was opened and closed. In the iio-volt D. C. circuit a 32-candlepower carbon-filament lamp and the sounder J were placed in series. When the circuit was closed, the lever of the sounder compressed a rubber tube through which a slow stream of tap water was passing to waste. The water was then delivered to coil I. The sounder was of a resistance of 4 ohms. The constancy of the apparatus is from 0.1° to '0.2° C, depending upon a number of factors, of which the external temperature and the rate of water flow are perhaps the most important. The humidity re- mained high and constant. The apparatus is capable of operating con- tinuously for weeks at a time in the neighborhood of 20° C. in hot weather, with little or no attention, except that of being supplied with ice. The cold-storage features of the humidity chamber, without which it would be of little value for the purpose of operating at temperatures below that of the room, were suggested by Mr. S. J. Dennis, of the Bureau of Plant Industry. Four large bunches of average-sized green bananas, carefully trimmed and sampled, were weighed and suspended in the humidity chamber and the bananas allowed to ripen. When their color had just changed from green to yellow, the tips of the fruit being still green, each lot was weighed, sampled as before, again weighed, and allowed to ripen further. The temperature was kept at 20° C. during ripening, and the air in the humi- dor was constantly renewed. When the bunches had become very ripe, most of the fruits showing slight superficial browning, the bunches were weighed and samples again taken. The pulp of the bananas was now soft, tender, and somewhat mealy. It was entirely free from translucent semiliquid portions and from decay. On weighing the bunches at this time many fruits broke off, and the experiment was discontinued. The records of the weights permit reference of the results to the basis of the original bunches, assuming, as in the previously described studies, that the stems lost weight at the same rate as the bananas themselves. The analytical data are shown in Table III. In all cases nearly all of the starch of the peel and the pulp gradually passed into sugar. For completeness the ash, the alkalinity of the ash, and the protein were determined in both pulp and peel. As in the previous study, the changes Dec. IS, 1914 Changes in Composition of Ripening Bananas 195 in the quantiries of these ingredients present were insignificant. In the peel the transformation of starch into soluble carbohydrates was more rapid during the change from green to yellow than subsequently. In the pulp also the transformation of starch was the most rapid during the change in color of the bananas from green to yellow. By the end of this period about two-thirds of the starch had passed into soluble carbohy- drates. During the subsequent ripening in each case a large proportion of the remaining starch in the pulp became converted into sugars. Some variation appears in the amounts of sucrose and invert sugar formed on ripening in the pulp of the bananas from the several bunches, but unfor- tunately a defect in the method existed (see p. 189), and more or less inversion of sucrose may have occurred during analysis. The figures for sucrose are probably slightly high and those for reducing sugars corre- spondingly low. Striking features of each set of analyses are the water changes. The peels lost water at uniform rates; the pulps all gained water — most rapidly after turning yellow. Table III. — Composition of bananas during ripening on Ike stem in the humidity chamber COMPOSITION EXPRESSED IN TERMS OF PERCENTAGES OF PEEI. .^ND PULP OF THE B.\NANAS BEFORE AND AFTER RIPENING ■u ■'1 'C n 0 Composition of peel or pulp. t. y a 0 ™ « =1 6 Part of fruit. ■3& 0 is ll .2 ll .3 4-> a Z i 1 .c ■M4 0 1-9 |3 b •3 1 IS 8-= ^ a^ 3 S 1 ^ S < <3 fi 1 5 5 Peel: Days. 1 Green bananas 0 0.00 40.30 89.06 10.94 1-45 0.95 0.64 0.90 0.23 4.20 8.31 Ripe bananas 6 4.46 36.76 87-64 12.36 1-71 1. 10 .68 3^36 .06 2>00 6.86 1065 Full ripe bananas. Pulp: 13 10.04 31-15 84.31 15-69 2.09 1.38 •87 3.83 •34 1.98 8.30 Green bananas 0 0.00 59-70 72.94| 27.06 .90 -S3 1.04 •77 • 37 31-54 34-88 Ripe bananas 6 4.46 63.24 72.94 27-06 -94 •52 •97 16.41 1.02 S-42 7-92 Full ripe bananas. 13 10.04 68. 85 75-98 24.02 .86 •49 •93 13^9« 5-35 • 76 3.B9 Peel: Green bananas 0 0.00 40.15 87-75 12.25 1-37 .86 •93 •69 .21 4-94 9-49 Ripe bananas S 3-33 37-60 87.09 12.91 1-55 -95 .88 3^14 •51 3-54 7-S2 lobb Full ripe bananas. Pulp: 13 9-4.1 3.-22 83.59 16.41 1. 91 1 .20 1.07 4^77 .26 1-94 8-08 Green bananas 0 0.00 S9-8S 72.21 37-79 .88 ■52 1.30 .64 •33 33-55 35-92 Ripe bananas 5 3-33 6 J. 40 71-76 28.24 •90 •50 1^25 7.69 9^37 6.49 9.00 Full ripe bananas IJ 9-43 68.78 75-70 24.30 .8> -47 1. 16 14-55 S-07 •75 3.76 Peel: Green bananas 0 0.00 40-35 88.13 11. 87 1-37 -87 .88 .96 .13 4-43 8-74 Ripe bananas 5 3'I7 37-45 87-32 12.68 1.52 .96 •79 3.81 •65 3-33 7-09 1067 Full ripe bananas . Pulp: ■3 8.90 30.80 83-53 16.47 1.96 1.33 1^21 4- SO .19 .-98 8.33 Green bananas 0 0.00 59- 6s 72. 12 27.88 .92 ■54 1-37 •41 •5' 33-43 26.0s Ripe bananas S 3-17 62-55 71.87 28.13 •95 • 50 x-33 ^36 9.17 6.56 8-95 Full ripe bananas. IJ 8.90 69- 20 75-40 24.60 .86 .48 • •33 13-13 6.70 -77 2-78 Peel: Green bananas 0 o.oo 43-44 88.74 11.26 1-32 .86 •73 •96 •IS 4.10 8-33 Ripe bananas 7 3-98 39-70 87.64 12. 36 I- SI .98 •74 2.96 • 11 3.06 6.62 1068 Full ripe bananas. Pulp: ■3 8.02 34-27 85-38 14.62 1.83 1.20 •93 3^S6 • 16 1-77 7-49 Green bananas 0 o.oc 56.56 71.90 28.10 •85 ■49 1.20 .29 ■43 33.51 26- 22 Ripe bananas 7 3-98 60.30 72-07 27-93 -90 .48 1.14 14.69 1.98 6.88 9.66 Full ripe bananas . '3 8. 03 65-73 74.92 25-08 .80 •45 1.06 9-91 6.81 1. 01 3^18 196 Journal of Agricultural Research Vol. Ill, No. 3 Table III. — Composition of bananas during ripening on the stem in the humidity chamber — Gjntinued COMPOSITION BXPRBSSED IN TBRMS OP PERCENTAGE OF THE WHOLE GREEN BANANAS ■S^ b. Composition of peel or pulp. § 1 ^^ 1 •3 1 Part of fruit. 11 ll .g 1. •J s 1 •a ;3 ■s < 6 1 it - 0 li •a.g •2" 1 w i OS J3J* 83 < Peel: Days. Green bananas 0 0.00 40.30 35-89 4.41 0.58 0.38 0-26 0.36 0.09 1.69 3^3S Ripe bananas 6 4.46 35- " 30-78 4-34 .60 •39 • 24 I- IS .02 .70 2.41 1065 Full ripe bananas. Pulp: 13 10.04 28.03 23- 63 4.40 •59 ■39 .24 1^07 • 10 • 56 2^33 Green bananas 0 0.00 59-70 43-54 16. 15 •54 ■3' .62 .46 • 22 12.86 14.85 Ripe bananas 6 4.46 60.42 44.07 16.36 •57 ■ 31 •59 9.92 • 62 3-27 4.79 13 10.04 61.94 47-06 14.88 -53 ■30 •S8 8-65 3-31 •47 1.79 Peel: Green bananas 0 0.00 40. IS 35-23 4.92 ■55 •35 •37 .28 • og 1.98 3^8l Ripe bananas S 3-33 36-35 31-66 4-69 •56 ■35 ■32 1.14 • 18 .92 2^73 1066 Full ripe bananas . Pulp: 13 9-43 28.28 23.64 4-64 •54 • 34 • 30 1-35 •07 •55 2.29 Green bananas... 0 0.00 S9-85 43-22 16.63 •S3 •31 •78 .38 .19 13-50 IS- 51 Ripe bananas 5 3-33 60.32 43-35 17-05 •S4 •30 .76 4.64 S.66 3-92 5-43 Full ripe bananas. 13 9-43 62. 29 47.16 15-14 •51 .25 ■72 9.06 3.16 •47 1-72 Peel: Green bananas 0 0.00 40.35 35-56 4-79 •55 ■35 •■10 •39 •OS 1.78 3-53 Ripe bananas S 3-17 36.26 31-66 4.60 •55 ■35 • 29 1^02 .24 .85 =-57 1067 Full ripe bananas . Pulp: 13 g.90 28.06 23-44 4-62 •S5 •35 •34 1.26 • OS .56 2-34 Green bananas 0 0.00 59.65 43.02 16.63 ■55 • 32 • 82 ■2S •30 13-38 15-54 Ripe bananas s 3-17 60.57 43-53 17.04 •58 •30 • 80 4.46 5^55 $•97 S^42 Full ripe bananas. 13 8.00 63.04 47-53 15-51 •54 .30 •78 X.2J 4.22 •49 1^75 Peel: Green bananas 0 0.00 43-44 38-55 4-89 •57 •37 •3= .42 • 07 1.78 3^62 Ripe bananas 7 3-98 38-12 33-41 4-71 •58 •37 • 28 '■I3 .04 •79 2.52 1068 Full ripe bananas . jPuIp: 13 8.02 31-52 26.91 4-61 ■S8 •38 .29 '■13 • 05 •56 3.36 Green bananas 0 0.00 ■;6.s6 40-67 iS-89 • 48 • 28 .68 .16 .24 I2^73 14.83 Ripe bananas 7 ■vos 57-9° 41-73 16-17 •52 .28 .66 8.50 I- 15 3^98 5-59 ^ Full ripe bananas . 13 8.02 60.46 45-30 15-16 .48 •27 .64 5^99 4.12 .61 1.9a A statement of the average composition of the pulps of the four bunches of bananas as analyzed, summarized from Table III, is given in Table IV. Table IV. — .Average composition in terms of percentage of pulp of bananas at different stages Stage. Green pulp. . . . Ripe pulp Very ripe pulp. Percent- age of whole fruit. 58.94 62, 12 68.14 72. 29 72. 16 75-5° Solids. 27-71 37.84 24.50 0.89 -92 .84 Alka- linity of ash. 0.52 .50 .47 Pro- tein. NX6.3S. 1.23 1. 17 Re- ducing sugars 0.53 11-54 12.89 0.41 5.39 5.98 I 32. 26 6.34 .82 FOURTH EXPERIMENT In the fourth experiment lots of bananas of the same initial ripeness and taken from the same stem were analyzed successively during ripen- ing. The rate of respiration of each lot was determined immediately before its analysis. It was thus possible to correlate the chemical Dec. : Changes in Composition of Ripening Bananas 197 changes with the changes in rate of respiration. The results show that the most rapid starch hydrolysis occurs at the time when the respiration rate is greatest. The following is a detailed account of the work. The fruits from a large bunch of green bananas were cut from the stem by first cutting off the "hands" and then separating the indi\adual fruits, finally removing from each fruit the small piece of adhering stem, cutting where the stem was most constricted, as in sampling bananas in previous studies. A few drops of sticky, turbid juice oozed from each banana, and care was taken not to allow the fruit surfaces to become wet with it. Fruits from the inside and the outside of each "hand" were divided as evenly as practicable into nine different lots, each lot receiving fruits from all parts of the bunch, and consisting of from 13 to 17 fruits. The lots were weighed and placed in the humidity chamber on the afternoon of May i. The analysis of the first lot was made upon the afternoon of the following day, immediately after the determination of its respiratory activity. The remaining lots were successively analyzed at intervals of from one to three days, except the ninth lot, which was kept longest, and spoiled. See Table V. Table V. — Composition of detached bananas during ripening in the humidity chamber COMPOSITION EXPRESSED IN TERMS OF PERCENTAGES OF PEEI. AXD PULP OF THE BANANAS BEFORE AND AFTER RIPENING Part of fruit. Inter- val la humid- ity cham- ber. Loss in weight in hu- midity cham- ber. Percent- age of whole fruit. Composition of peel or pulp. Solids. Alco- hol-in- soluble soUds. Total sugar Starch. 1912. May z 3 PEEL. Green bananas Slightly yellowing bananas. do Yellow bananas ...do ... do ...do Brown bananas PULP. Green bananas Slightly yellowing bananas. ...do Yellow bananas ...do ....do ...do Brown bananas. Days. 1.66 2.47 3-67 6.48 7.83 8-39 10. 56 11-32 1.66 2.47 3-67 6.48 7-83 8.39 lo. 56 11.32 38.22 37-54 35-41 32-57 30-88 29.35 27-95 26. 26 61.77 62.45 64.58 67.42 69. 12 70.65 72.05 73-74 89. 72.47 72.37 72.44 73-26 73-45 74-63 75-72 77-02 10-54 "•39 11.36 13- 16 14. 10 14.92 15-75 16.68 27-53 27-63 27-56 26-74 26-55 25-37 24.28 22.98 6.83 8.52 6.J8 6.09 6.67 7.08 7-57 8.12 21-95 16.46 11.04 5. 82 3-89 3- 13 2.88 3-17 1-39 2. 16 2.91 3- 64 3-89 3-98 4.07 4- 38 4-35 9-65 IS- 00 19.6s 20.08 20.61 20.05 18.21 3.28 2-93 2. II 1.63 1.69 1.82 i.8« 1.84 19.01 13-96 8.62 3-65 I- 73 I- IS •87 1. 01 COMPOSITION E.XPRESSBD IN TERMS OF PERCENTAGE OF THE WHOLE GREEN BANANAS May PEEL. Green bananas Slightly yellowing bananas . do Yellow bananas do ...do do Brown bananas I 1.66 37-59 33-63 3-96 2-57 0.53 3 2-47 36-61 32-44 4-17 3-12 .80 3 3-67 34-11 30.23 3.88 2. II •99 5 6-48 30.46 26.45 4.01 1.86 I. II 7 7-83 28.46 24-45 4.01 1.90 I. II 9 8-39 26.90 22-89 4.01 1.90 1.07 12 10-56 25.00 21.00 3-94 1.89 1.02 '5 11.32 23.29 19.40 3-89 1.89 1.02 • 72 • so .48 • 49 ■47 •43 198 Journal of Agricultural Research Vol. Ill, No. 3 Table V. — Composition of detached bananas during ripening in tlie humidity chamber- Continued COMPOSITION EXPRESSED IN TERMS OF PERCENTAGE OF THE WHOLE GREEN BANANAS — COntillued Part ol fruit. Inter- val in humid- ity cham- ber. Loss in weight in hu- midity cham- ber. Percent- age of whole fruit. Composition of peel or pulp. Date. Water. Solids. Alco- hol-in- soluble sohds. Total sugar as invert. Starch. 1912. May 2 3 4 PULP. Days. 13 IS 1.66 60.74 60.91 62. 21 63.0s 63.70 64.72 64.44 6s. 39 44- 02 44.08 45- 06 46.19 46.79 48.30 48.79 50.36 16.72 J6.83 17.15 16.86 16.91 16.42 15.65 15-03 13-33 10.03 6.87 3-67 2.48 2.03 1-86 2.07 2.65 5-88 9-33 12.39 12.79 13-34 12.91 II. 91 II. ^i Slightly yellowing bananas, do 2 3 6 7 8 10 47 67 48 83 39 s6 8 5 2 I 50 8 TO 13 .. . do ^6 6^ In order to determine correctly the respiration rate at interv'als on ripening, it was necessar}' to collect the carbon dioxid evolved during relatively short periods. The bananas were placed in a tubulated desiccator, kept in the dark at 20° C, and a rapid current of air passed through. The air was first freed from carbon dioxid by passing it through a long wide glass tube filled with soda lime. The carbon dioxid evolved by the bananas was collected by drawing the air from the desiccator through a Reiset scrub- bing tube containing soda solution. The Winkler method of titration of the absorbed carbon dioxid was employed. In operating the flask of the Reiset apparatus was charged with a mixture of 500 c. c. of distilled water and 100 c. c. of an approximately normal solution of sodium hydroxid. The distilled water had pre\dously been well aerated to remove carbon dioxid, and the titer of the soda solution (after addition of barium chlorid to precipitate carbonates) was known. After mixing, the Reiset tube was inserted, comiection made with the desiccator containing the bananas, and suction applied. After absorption, the contents of the Reiset apparatus were washed into a large precipitating jar with aerated distilled water, excess of barium chlorid added, and the solution titrated with normal hydrochloric acid, using phenolphthalein as indicator. Each cubic centimeter of normal alkali con- sumed equals 0.022 gm. of carbon dioxid. In titrating it was necessary to admit the acid under the surface of the solution and stir well to avoid escape of carbon dioxid freed by local momentary excess of acid. As the amounts of carbon dioxid expected were approximately known it was found convenient to use a slight excess of solution of barium chlorid of such strength that each cubic centimeter decomposes i c. c. of normal sodium carbonate.' Lead tubes were used in leading the air to and from the desiccator. The air entered it near the top and was withdrawn from near the bottom. ' For a criticism of volumetric methods of estimation of carbon dioxid, see Kuster (15). Dec. IS, 1914 Changes in Composition of Ripening Bananas 199 At the start of each experiment a current of carbon-dioxid-free air was led through the desiccator for about half an hour at the rate of from i to 2 liters per minute, this inter\'al constituting a fore period during which the carbon dioxid was not collected. The absorber was then inserted in the train and the air passed through at the same rate as during the fore period. A water gauge inserted between the desiccator and the Reiset tube served to indicate the relative rate of flow. After an interval of from one to two hours the absorption apparatus was replaced by a freshly filled Reiset tube and the carbon dioxid collected for a second interval. The rate of respiration was thus determined during two successive intervals. In general, well-agreeing duplicates were obtained. The figures showing the results of the study of the rate of respiration are given in Table VI. At the beginning the bananas were just on the point of turning yellow and were much more active (io6 mg. of carbon dioxid per kg. hour) than the green fruit used in the calorimeter studies (30 mg. and 30 mg., respectively). The respiration was most intense (146 mg. per kg. hour) when the rate of transformation of starch was greatest. It then slowly slackened, reaching at the end 91 mg. per kg. hour." As the rate of respiration of each sample analyzed was deter- mined immediately before its analysis, the total amount of carbon dioxid evolved during the ripening period, 3.776 per cent of the original bananas, was easily estimated by a summation process from the data given in Table VI. Table VI. — Rate of respiration of detached bananas ripening in tke humidity chamber Inter- Aver- val age rate Cab Inter- Aver- from middle of res- piration Num- Origi- dioxid col- lected. ior which col- lected. Respi- age of one from Date. Description. ber of nal ration respi- collec- middle fruits. weight. rate.l" ration tion of one rate.!! period period to mid- to mid- dle of dle of . next. next.ft Gm. Gw. Hcmrs. Gtn. Gm. Hours, Gm. May 2 "Wholly green, but on the } - /o. 3344 \ • 4S76 1-33 0. 104 > 0. 106 point of turning yellow. . . 2.410 r- 75 .108 1 3 Bananas just beginning to 1 /• 464 \ -656 f '* 0.136 turn yellow, sample ripen- 1 ^^ 2.SS2 1.25 •145 \ ■ 1461 ) ing uniformly. 1-75 .147 i ^ 4 Fruit yellower than on day 1 / .85- \ .669 , M .143 previous, much green still present 1 '■* 2,506 3.50 1.92 ■ 136 •139 } ..38jj 0 Fully yellowed, but many 1 /■337 \-675 J ...7| 48 .138 specimens are green at the tips 1 ^^ 3,300 ■•35 3-53 ■ "7 .116 8 Fully yellowed 16 3.83s /■S36 \ -352 1.70 I. 13 .109 .III } ...i > 48 • 113 . 108 10 Fruits beginning to brown } - /■4S3 1.47 . 106 f ^ shghtly at the surfaces 2,919 \ -383 I. 25 ■ 105 13 Skins brown, much yellow [ 70 .104 still present , '^ 2,530 •403 /■374 \-376 • f rtrt Tnil \ 16 Skins almost entirely ' ■ 1 .09. .096 brown, a little yellow M 1*2.535 1.78 .091 present at ribs 1.78 .091 / ^ ' a In the two calonraeter studies the respiration rate of the bananas reached maxima of 150 and 200 mg per kg. hour, respectively, and the respiratory activities at the end were no and 100 mg per kg hour respec- tively. 6 Grams carbon dioxid per kg. per hour. <^ Weight used in calculating was 1,990 gm.. three fruits having been rejected whose weight calculated to the original weight was subtracted from the original weight of the sample. "fWcight used in calculating was r 310 gm.. otic fruit having been rejected, its weight calculated to original and subtracted. 200 Journal of Agricultural Research Vol. Ill, No. 3 The analytical data are given in Table V. The study of the detached fruit permitted a longer period of observation than when bananas attached to the stem were used. The rate of transformation of starch into soluble carbohydrates was very rapid at first — 32.11 gm. per kg. per day. It then increased to 34.9 gm. per kg. per day. Starch hydrol- ysis, so far as revealed by analysis, nearly ceased six days before the end of the life history of the bananas. The analytical data confirm the facts developed in the earlier experiments. DISCUSSION OF RESULTS As the result of the foregoing studies, the author is in a position to state more exactly than has heretofore been possible the nature and extent of the changes in the composition of bananas during ripening. The most conspicuous change is the long-recognized conversion of starch into sugars. It is most rapid while the fruits are turning from green to yellow. During this period the respiration rate increases manyiold, becoming greatest at the time when the rate of starch hydrolysis is most rapid. Starch hydrolysis then gradually slackens, later ceasing alto- gether. The respiration rate, too, becomes slower, but still remains far more active than in the green fruit. Next to the starch and respiration changes, most conspicuous are those of water. The peel loses, while the pulp gains water steadily. The respective losses and gains in water of the peel and the pulp on ripening, expressed in terms of the original green bananas, are summarized in Table VII. Table VII. — Percentage of losses and gains in water 0/ peel and pulp of bananas on ripening Experi- ment No. Place of ripening. Calorimeter do Humidity chamber. .do.. Loss of water in peel. Actual gain of water in pulp. I. 64 3.02 3-52 3-94 4- 51 4-63 6.34 Gain of water in pulp cor- rected for water formed and absorbed in physiological processes. 2.4 3-5 6- 13s In the first, second, and fourth experiments it is possible to show how much water is formed or absorbed by the pulp in physiological processes. The water formed in respiration can easily be calculated if formed in consequence of the complete combustion of carbohydrates and if the amount of carbon dioxid evolved on ripening in consequence of this com- bustion is known. The respiratory quotient and the thermal quo- tient determined by the Office of Nutrition Investigations for ripening bananas X16) agree in showing that the carbon dioxid evolved on normal ripening is due solely to the complete combustion of carbohydrates. Dec. 15. 1914 Changes in Composition of Ripening Bananas 201 We are therefore justified in calculating the water formed by the equa- tion C8Hj20j + 6O2 = 6CO2 + 6H2O. From the water so formed is sub- tracted the water absorbed in the saccharification of starch. See Table VIII. Table VIII. — Percentage of water formed by the pulp of bananas in physiological processes Experi- Carbon dioxid found. Calculated water formed from equation I. Loss of carbohydrate found on ripening in — 0 Calculated water propor- tioned according to car- bohydrate loss in — Calculated water ab- sorbed in Peel. Pulp. Peel. Pulp. starch hydrolysis. I 2 4 1-334 2. 256 3-776 0. 546 -9=3 1-545 0. 46 .41 . 40 0.88 1-43 2. 84 0. 188 ■ 20s . 190 0-358 .718 1-355 I. 14 1-23 I. 16 a Probably slightly larger than actual on account of failure to completely estimate maltose. See p. 189. In the first two experiments absorption of water amounting to 0.782 and 0.512 per cent occurred as a net result of respiration and starch hydrolysis. In the fourth experiment, where the bananas became over- ripe, the water formed in respiration was greater by 0.195 P^r c^'nt than that absorbed in starch hydrolysis. The increases of water in the pulp during ripening are all derived from the peel, except when bananas become overripe, when the water formed in respiration may more than balance the water absorbed in starch hydrolysis. From the quantity of sugar formed in the pulp it is evident that the osmotic pressure of the pulp must undergo a marked increase, with a corresponding decrease of vapor pressure, during the ripening of the fruit. A possible operating cause of the water transfer from peel to pulp is obvious. From a knowledge of the carbon dioxid formed in respiration and know- ing from the calorimeter data that carbon dioxid results from the com- plete combustion of carbohydrates, it can be determined whether or not the carbohydrates consumed in respiration were accurately made known from the analyses. Carbohydrate losses found by analysis contrasted wath the expected losses from the calorimeter data are shown in Table IX. Table IX. — Comparison of carbohydrate losses with theexpected losses from carbon dioxid in pulp of ripening bananas Carbohydrate losses found (expressed as hexose). Expected losses from carbon di- oxid formed in respiration. Per cent. 1-34 1.84 3-24 Per cent. 0. 91 1-54 2-58 202 Journal of Agricultural Research voi. in. No. 3 By analysis somewhat greater losses appear than indicated from the calorimeter data. It is not improbable that the small dififerences are due to analytical error. SUMMARY (i) The usual carbohydrate changes — saccharification of starch, with formation of sucrose and invert sugar, and consumption of sugars in respiration — proceeded with uniformity in bananas of different bunches. (2) The period of most rapid respiration corresponded closely with that of most rapid starch hydrolysis. (3) The quantities of ash, protein, and ether extract underwent but slight changes during the ripening of the bananas. Pentosans decreased markedly in the pulp, but remained little changed in the peel. (4) Analyses of the peel and pulp of ripening bananas showed a steady transfer of water from peel to pulp during ripening. LITERATURE CITED (i) Atwater, W. O., and Bryant, A. P. 1899. The chemical composition of American food materials. U. S. Dept. Agr., Office Exp. Sta. Bui. 28 (rev.), 87 p., 3 fig. (2) Bailey, E. M. 1906. Studies on the banana. In Jour. Biol. Chem., v. i, no. 4/5, p. 355-361. (3) 1912. Biochemical and bacteriological studies on the banana. In Jour. Amer. Chem. See, v. 34, no. 12, p. 1706-1730. (4) Balland. 1900. Composition et valeur alimentaire des principaux fruits. In Rev. Intemat. Falsif., ann. 13, livr. 3, p. 92-100. (5) BinoNET, H. 1861. Recherches sur la matiere sucr^e contenue dans les fruits acides; son origine, sa nature et ses transformations. In Ann. Chim. et Phys., s. 3, t. 61, p. 233-308, pi. 3. (6) Chace, E. M., Tolman, L. M., and Munson, L. S. 1904. Chemical composition of some tropical fruits and their products. U. S. Dept. Agr., Bur. Chem. Bui. 87, 38 p. (7) Colby, G. E. 1894. Analyses of bananas and banana soils from the Sandwich Islands. In Cal. Agr. Exp. Sta. Rpt. 1892/1894, p. 275-279. (8) CORENWINDER, B. 1863. Recherches sur la composition de la banane de Br^sil. In Compt. Rend. Acad. Sci. [Paris], t. 57, no. 19, p. 781-782. (9) 1876. Recherches chimiques sur les productions des pays tropicaux. In Ann. Agron., t. 2, p. 429-446. (10) 1879. Sur la banane. In Compt. Rend. Acad. Sci. [Paris], t. 88, no. 6, p. 293- 295- (11) DohERTY, W. M. 1892. The analysis of the Cavendish banana (Musca Cavendishii) in relation to its value as a food. In Chem. News, v. 66, no. 1715, p. 187-188. Dec. 15, 1914 Changes in Composition of Ripening Bananas 203 (12) Gerber, C. 1896. Recherches stir la maturation des fruits charaus. In Ann. Sci. Nat. Bot., s. 8, t. 4, p. 1-280, pi. 1-2. (13) Gore, H. C. 1911. Studies on fruit respiration. U. S. Dept. Agr., Bur. Chem. Bui. 142, 40 p., 17 fig. (14) Jahkel, Paul. 1909. tjber Anatomie u. Mikrochetnie der Bananenfrucht und ihre Reifungs- ersclieiniuigen. 41 p. Kiel. Inaugural- Dissertation. (15) KtJSTER, F. W. 1897. Kritische Studien zur volumetrischen Bestimmung von karbonathal- tigen Alkalilaugen und von Alkalikarbonaten, sowie iiber das Ver- halten von Phenolphtalein und Methylorange als Indikatoren. In Ztschr. Anorgan. Chem., Bd. 13, Heft 2/3, p. 127-150. (16) Langworthy, C. F., and MilnEr, R. D. 1913. Some results obtained in studying ripening bananas with the respira- tion calorimeter. U. S. Dept. Agr. Yearbook, 1912, p. 293-308. (17) Marcano, v., and Muntz, A. 1879. Sur la composition de la banane et sur des essais d 'utilisation de ce fruit. In Compt. Rend. Acad. Sci. [Paris], t. 88, no. 4, p. i56-i58" (18) Prinsen Geerligs, H. C. 1908. Rapid transformation of starch into sucrose during the ripening of some tropical fruits. In Intemat. Sugar Jour., v. 10, no. 116, p. 372-380. (19) Reich, R. igii. Reife und tmreife Bananen. In Ztschr. Untersvich. Nahr. u. Ge- nussmtl., Bd. 22, Heft 4, p. 208-226. (20) RlCClARDI, L. 1882. Composition chimique de la banane a differents degres de maturation. In Compt. Rend. Acad. Sci. [Paris], t. 95, no. 8, p. 393-395. (21) TiU.LARico, Giuseppe. 1908. Gli enzimi idrolitici e catalizzanti nel processo di maturazione delle frutta. In Arch. Farmacol. Sper. e Sci. Aff., v. 7, fasc. i, p. 27-48; fasc. 2, p. 49-68. (22) YosmMURA, K. 1911. Beitrage zur Kenntnis der Banane. In Ztschr. Untersuch. Nahr. u. Genussmtl., Bd. 21, Heft 7, p. 406-411. 67235°— 14 2 ASSIMILATION OF COLLOIDAL IRON BY RICE By P. L. GiLE, Chemist, and J. O. Carrero, Assistant Chemist, Porto Rico Agricul- tural Experiment Station INTRODUCTION Previous work ' at the Porto Rico Agricultural Experiment Station has shown that pineapples and upland rice grown on moderately or strongly calcareous soils are affected with chlorosis and that the failure of these plants to make a successful growth on such soils seems to be due wholly or in part to a diminished assimilation of iron. A determination of the forms of iron available to rice is therefore important. An experiment on the assimilation of colloidal iron is reported here, as it bears on this prob- lem as well as on the properties of plant roots. MATERIALS AND METHODS OF THE EXPERIMENT The availability of colloidal iron was compared with that of ferric chlorid by growing upland rice in a nutrient solution. Water cultures were used in this work, as it is impossible to tell in what form the iron may be present in soil cultures. To prevent precipitation of the colloidal iron by other salts of the nutrient solution the iron was put in one flask (flask B) and the other salts in a second flask (flask A) . The plants were grown with part of their roots in each flask. The seeds were germinated over distilled water until they had de- veloped two or more roots. Two plants were grown in each pair of flasks. Two hundred c. c. Erlenmeyer flasks of Jena glass with their necks joined together by surgeon's tape were used. The formula for the nutrient solution used in flasks A, which gave excellent results with rice in previous work, was as follows: KNO3 o. 1017 gm. KH2PO4 0714 gra. NaNOs 2143 gm. NajSO, 0315 gm. CaClj o. 05 gm. MgCl., 05 gm. H.SO^ sec. N/io. Distilled water, 1,000 c. c. The plants were changed to fresh solutions every few days. The colloidal iron used was the ordinary dialyzed iron. It contained 0.0383 gm. of Fe and 0.0058 gm. of Clj per c. c. After salting out the colloidal iron with potassium sulphate, the filtrate was examined and found to contain the following, calculated as grams per c. c. of the original dialyzed iron: Clj, 0.00436 gm. ; Fe, hardly a reaction with potassium • Gilc. P. L. Relation of calcareous soils to pineapple chlorosis. Porto Rico Agr. Exp. Sta. Bui. ir, 45 P-. 2 pl.. >9>'- Gilc, P. L.. and Ageton. C. N. The effect of strongly calcareous soils on tlie growth and ash composition of certain plants. Porto Rico Agr. Kxp. Sta. Bui. 16, 45 p., 4 pi., 1914. Journal of ARricultural Research, Vol. III. No. 3 Dept. of Agriculture, Washington, D. C. Dec. 15, 1914 (205) B-3 2o6 Journal of Agricultural Research Vol. III. No. 3 sulphocyanate; acidity, equivalent to 0.002 gm. of Clj from hydro- chloric acid; ammonia, none with Nessler's reagent. The dialyzed-iron preparation after dilution was dialyzed for 24 hours with a parchment membrane without any iron appearing in the dialyzate. While no soluble or ionized iron appeared in these tests we must assume that some existed in the dialyzed-iron preparation because of the high chlorin content. The nonappearance of iron in the dialyzate after dialysis and in the filtrate after salting out the iron was probably due to adsorption by the colloid or precipitate and to the strong hydrolysis that dilute ferric chlorid undergoes.' From the chlorin content it appears that about one- twelfth of the iron could have been present as ferric chlorid, but we can hardly assume that it was there as such or that a certain quantity of ferric chlorid in a colloidal-iron solution would act the same as in a simple aqueous solution. We can simply assume that the soluble iron bore some proportion to the quantity of chlorin. In the following tests, then, a low availability of the dialyzed-iron preparation is not proof of the assimila- tion of colloidal iron. RESULTS OF EXPERIMENTS Experiment I. — In a preliminary test the plants were grown for 42 days. The dialyzed iron was used at the rate of 10.5 parts of Fe per 100,000 parts of water for the first 10 days and at the rate of 1.05 parts of Fe per 100,000 for the remaining 32 days. The ferric chlorid was used at the rate of 0.41 part of Fe per 100,000. The results are given in Table I. Table I. — Growth of rice with dialyzed iron and ferric chlorid — Experiment I Nos. of flasks. Solution in flasks A. Solution in flasks B. Green weight of tops. Oven-dry Gain over weight no-iron of tops. plants. s-s 9-12 13-16 1-4 Nutrient solution without Fe do Distilled water Gm. 1.03 3-25 4-02 5-82 Gm. 0. 17 •54 .67 •95 Gm. do • S<* Nutrient solution -f FeCIa . 78 The preparation of dialyzed iron had a certain availability, but rela- tively large amounts were less effective than the smaller quantity of ferric chlorid. Experiment II. — In a second experiment the dialyzed iron and ferric chlorid were both used to furnish 0.4 gm. of Fe per 100,000 c. c. of water. The plants were grown for 59 days. The results are given in Table II. ' This was borne out by the following test: To i c. c. of dialyzed iron 0.00275 gm. of Fe from FeCb was added; the solution was made to 100 c. c. and the colloidal iron salted out by KsSO*. The filtrate, tested for soluble iron (colorimetric method with KSCX), showed 0-00245 gm. of Fe had been lost. Of the O.0O24S gm. of Fe lost 0.00056 gm. was lost by adsorption by the precipitate. Precipitation by hydrolysis by water alone caused a loss of 0.00023 gm. of Fe, and hydrolysis in the presence ol K2S04 caused a loss of 0.001S9 gra, of Fe. Dec. IS, 1914 Assimilation of Colloidal Iron by Rice 207 Table II. — Growth of rice with dialyzed iron and ferric chlorid — Experiment J I Nos. of flasks. Solution in flasks A. Solution in flasks B. Green weight of tops. Oven- dry weight of tops. Average oven-dry weight of tops. Gain over no- iron plants. Nutrient solution with- out Fe. do Distilled water Gm. 3-38 4.41 4.83 4.71 6.40 5- 61 7.82 Gm. o. 60 .76 • 79 •78 1.07 1.05 1. 31 Gm. 0.60 Gm. 5-8 0.4 Ein. of Fe per 100,000 c. c. from dialyzed iron. do .... do. . .do .78 0. 18 do 0.4 gm. of Fe per 100,000 c. c. from FeCU. do do 25-28 do ....do ..14 •54 In this experiment, where equivalent and small quantities of iron were used, the dialyzed-iron preparation appeared to have an availa- bility of about three-tenths that of the ferric chlorid. Experiment III. — A third series was conducted, using equivalent small quantities of iron, ten times this amount of iron from dialyzed iron, and twice this quantity of iron from ferric chlorid. The results are given in Table III. Table III. — Growth of rice uith dialyzed iron and ferric chlorid — Experiment III Nos. of flasks. .Solution in flasks A. Solution in flasks B. Green weight of tops. Oven- dry weisht of tops. Aver- age CR^ en- dry weight of tops. Gain over no-iron plants. Aver- age oven- dry weight of roots in flasks A. Aver- age oven- dry weight of roots in flasks B. J-5 Nutrient solution without Fe. do Distilled water Gm. 3-93 •3-47 4-35 3-75 8.78 5-77 5- 51 6.69 8.90 9. 01 17.92 20.03 Gm. 0.73 .69 1.06 •93 I- 55 1.05 1.00 1.24 1-57 I. 55 2.81 3- 13 Gm. Gm. Gm. Gin. do 0.71 0.144 O.OS5 ii-is ..do Fe from dialyzed iron, 0.4 gm. per loOjOooc.c. do .do 1. 00 0.29 ■ 201 •"S 21-25 26-30 31-35 36-40 41-45 46-50 SI-5S 56-60 do do ...do do 1.30 •59 . 212 •194 do Fe from FeClj, 0.4 gm. per 100.000 c. c. do Fe from FeCb, 0.8 gm. per 100.000 c. c. do do 1. 12 ,41 .217 .077 do 1.56 •85 .270 .174 Nutrient solution -t-FeCl3. do Distilled water ... do 2.97 2.26 •S"l .190 Flasks Nos. 26 to 30 did not agree well with Nos. 21 to 25, so it is possible that the average of 1.30 gm. is too low. It is apparent from the preceding tests that the dialyzed iron was much less available than the ferric chlorid. On the basis of the Fe content, the dialyzed iron had to be present in at least five times the amount of ferric chlorid to produce the same yield. The percentages of iron in the dry substance of the tops was deter- mined in the plants from Experiment III. The percentages varied from 2o8 Journal of Agricultural Research voi. iii. No. 3 0.035 to 0.037 P^r cent of Fe^Og. Since the samples were small, it is felt that the figures merely show that practically the same percentages of iron were present in all the plants. Thus, the quantities of iron in the different lots of plants would vary as their dry weights. In all the experiments the plants which received no iron in either flasks A or flasks B were strongly chlorotic, the chlorosis commencing about 6 to 10 days after the plants were put in the solutions. The plants receiving dia- lyzed iron or ferric chlorid in flasks B were also strongly chlorotic, although they were somewhat greener than the check plants without any iron, and the chlorosis was later in appearing. The plants which had the ferric chlorid added to the other nutrient salts in flasks A were of a normal green color. The root development varied greatly in the different flasks. In all the flasks A which contained the complete nutrient solution without iron the root development was good, the main roots being long, with numerous long laterals. The roots in flasks B which contained only distilled water made very little growth and had few laterals. The roots in flasks B which contained only ferric chlorid made even less growth than those in distilled water. As soon as the roots penetrated the ferric-chlorid solu- tion, the root tip appeared killed and no roots penetrated the solution for any distance. The roots in flasks B which contained dialyzed iron developed much better than in distilled water. During the latter stages of growth particularly a few heavy roots developed in the dialyzed iron, but these roots carried very few laterals. The oven-dry weights of the roots in flasks B given in Table III do not really give a comparison of the root developments in the solutions, for the reason that the weights of the roots in flasks B, Nos. 31-60, were made up chiefly of root "stubs" — heavy roots which started out from the plants but did not develop in the distilled water or ferric-chlorid solution. The larger the top growth, de- pendent on the amount of iron obtained, the more root "stubs" devel- oped. For instance, flasks B, Nos. 41-46 and Nos. 36-50, had 0.174 gni- of roots, while Nos. 31-35 and Nos. 36-40 had but 0.077 gm. of roots. As a matter of fact, the root development in the solution was less in the first case than in the second. DISCUSSION OF RESULTS The poor development of the roots in the flasks B was, of course, due to the well-established toxicity of unbalanced solutions — in this case of single salts or distilled water.' Because of the injury resulting from the iron solutions it is impossible to draw a very sharp conclusion concerning the assimilation of colloidal iron. Although the roots developed better in ' True holds the injury from ordinary distilled water "but a spedal case of the general type of injury wrought on cells by unbalanced solutions."— True, R. H. The harmful action of distilled water. In Science, n. s., v. 39. no. 999. P- 29s- 1914. Dec. IS, 1914 Assimilation of Colloidal Iron by Rice 209 the dialyzed iron than in the ferric chlorid/ they assimilated less iron, even in solutions containing three to five times more dialyzed iron than ferric chlorid. It seems probable that the small amount of iron obtained from the dialyzed iron was not colloidal iron but soluble iron. It is true that the tests made of the dialyzed-iron preparation revealed little or no ionized iron, but we must assume the presence of some soluble iron from the chlorin content of the preparation. In view of the low availability of the dialyzed-iron solution it would seem that the often-mentioned but questionable acid excretion of roots was not operative, at least not in this unbalanced solution. With respect to fineness of division and contact with the roots the colloidal iron was especially favorable for assimilation by an acid if the roots had excreted such. An apparent objection to the conclusion that colloidal iron is not assimilable is the fact that dialyzed iron is sometimes used in nutrient solutions and that in Von Crone's nutrient solution ferric phosphate is emploj'ed. In such solutions, however, it is not at all certain that the plants utilize insoluble iron compounds. In fact, from further work in progress it appears that rice at least is capable of assimilating only solu- ble iron in nutrient solutions. Aside from the question of the assimilation of colloidal iron, the pre- ceding test is of interest in connection with the study of unbalanced solu- tions. The fact that while some roots of the plant were developing well in the balanced solution of the flasks A other roots of the same plant were injured by the distilled water or ferric chlorid of the flasks B shows that in case the toxicity of certain salt solutions or ordinary distilled water is due to injury in the root cells by extraction of other electro- lytes,^ this extraction takes place faster than the electrolytes can be supplied from other parts of the plant. For in this case the roots in flasks A had an abundance of the electrolytes to draw on. The idea that the toxicity of single-salt solutions is due to or accom- panied by the penetration of the salts finds confirmation in the preced- ing experiments. It is evident from Experiment III that the stronger the ferric chlorid solution the less the root development in fhe solution, but the greater the amount of iron absorbed (as shown by the top growth which was dependent on iron absorbed). Osterhout ^ has shown by electrical conductivity measurement that single-salt solutions penetrate cells of the kelp, Laminaria. * The reduction of the toxicity of distilled water by colloids and finely divided solids has been frequently noted. » True, R. H. Loc. cit. 8 Osterhout, W. J. v. The permeability of protoplasm to ions and the theory of antagonisms. In Sdeuce. n. s., V. 35, no. 890, p. 112-115, 1912. 2IO Journal of Agricultural Research voi. m, N0.3 SUMMARY The work reported would seem to show that rice can not assimilate colloidal iron. It is believed that the iron obtained from the dialyzed- iron preparation was soluble iron. It is apparent that the toxicity of ordinary distilled water or ferric- chlorid solutions for plant roots can not be overcome by supplying other roots of the same plant with a balanced.solution. The toxicity of the ferric-chlorid solution was accompanied by the penetration of iron into the root and transportation, to the leaves. COLORING MATTER OF RAW AND COOKED SALTED MEATS By Ralph Hoagland, Laboratory Inspector, Biocketnic Division, Bureati of Animal Industry INTRODUCTION The red color of fresh lean meat, such as beef, pork, and mutton, is due to the presence of oxyhemoglobin, a part of which is one of the con- stituents of the blood remaining in the tissues, while the remainder is a normal constituent of the muscles. When fresh meat is cooked or is cured by sodium chlorid, the red color changes to brown, owing to the breaking down of the oxyhemoglobin into the two constituents, hematin, the coloring group, and the proteid, globin. On the other hand, when fresh meat is cured by means of a mixture of sodium chlorid and a small proportion of potassium nitrate, or salt- peter, either as a dry mixture or in the form of a pickle, the red color of the fresh meat is not destroyed during the curing process, the finished product having practically the same color as the fresh meat. • Neither is the red color destroyed on cooking, but rather is intensified. The practical value of saltpeter in the curing of meats is so well known that its use for this purpose may be said to have become practically uni- versal ; such use is not confined to the commercial meat-packing industry, but it is used in the home curing of meats as well. It is only within comparatively recent j'ears, however, that anything very definite has been known concerning the nature of the color of salted meats or the process of the color formation. The work which is reported in this paper was undertaken for the purpose of obtaining more complete information concerning the color of raw and cooked salted meats. HISTORICAL SUMMARY Weller and Riegel (1897),' in the examination of a number of samples of American sausages, obtained a red coloring matter on extracting the samples with alcohol and other solvents, which color they concluded to be in some manner due to the action of the salts used in curing upon the natural color of the meat. On account of similaritj' of spectra, this color was considered to be methemoglobin. Lehmann (1899) observed that when fresh meat was boiled in water containing nitrites and free acid or in old meat broth the surface of the meat turned bright red in color, in contrast to the brown color which fresh meat takes on when boiled in water free from nitrites. The addi- ^ Bibliographic citations in parentheses refer to "Literature cited." p. 225- Journal of Agricultural Research, Vol. Ill, No. 3 Dept. of Agriculture, Washiitfiton, D. C. Dec. 15. 1914 (21 1 ) A— 10 212 Journal of Agricultural Research voi. m, no. 3 tion of nitrates to the water did not cause the production of the red color on the surface of the meat on boiling. This color was found to be soluble in alcohol and ether and to give a spectrum showing an absorption band just at the right of the D line, and a second band, often poorly defined, at the left of the E line. On standing, the color of the solution changed to brown and gave the spectrum of alkaline hematin. A color with similar properties was obtained on extracting hams and various kinds of sausages with alcohol and other solvents. The author named this coloring matter "hsemorrhodin." Kisskalt (1899) studied the production of red color in fresh meats on cooking and found that this color appeared when meat was cooked in bouillon which was several days old, or in water containing nitrites and free acid. Meat boiled in water to which saltpeter had been added did not take on the red color; but, on the other hand, if the meat was first allowed to stand several days in contact v.'ith saltpeter and then boiled, the red color appeared. Haldane (1901) made an extensive study of the color of cooked salted meat, which color he concluded to be due to the presence of the nitric- oxid hemochromogen resulting from the reduction of the coloring mat- ter of the uncooked meat, or nitric-oxid hemoglobin (NO-hemoglobin). This color was found to be the same as that resulting from the boiling of fresh meat in water containing nitrites and free acid. It exhibited a spectrum showing a distinct band just at the right of the D line and a faint band a trifle to the left of the E line. The color was found to be soluble in alcohol and in ether and to be quite resistant to the action of reducing agents. The color of uncooked salted meats was found to be soluble in water and gave a spectrum characteristic of NO-hemoglobin. The formation of the red color in uncooked salted meats is explained by the action of nitrites in the presence of a reducing agent and in the absence of oxygen upon hemoglobin, the normal coloring matter of fresh meats. Orlow (1903) states that the red color of sausages is due to the action upon the color of the fresh meat of the nitrites resulting from the re- duction of the saltpeter used in the process of manufacture. The present author (1908) studied the action of saltpeter upon the color of meat and found that the value of this agent in the curing of meats depends upon its reduction to nitrites and nitric oxid, with the consequent production of NO-hemoglobin, to which compound the red color of salted meats is due. Saltpeter, as such, was found to have no value as a flesh-color preservative. Glage (1909) is the author of a pamphlet concerning practical methods for obtaining the best results from the use of saltpeter in the curing of meats and in the manufacture of sausages. The fact that the value of saltpeter as a flesh-color preservative is dependent upon the reduction of the nitrate to nitrite is recognized, and directions are given for the Dec. IS. 1914 Coloring Matter of Salted Meats 2 1 3 partial reduction of the saltpeter to nitrites by heating the dry salt in a kettle before it is to be used. It is stated that this partially reduced saltpeter is much more efficient in the production of color in the manu- facture of sausage than is the untreated saltpeter. Humphrey Davy in 1812 (cited by Hermann, 1865) and Hoppe-Seyler (1864) noted the action of nitric oxid upon hemoglobin, but it appears that Hermann (1865) was the first to furnish us with much information as to the properties of this derivative of hemoglobin. He prepared NO-hemoglobin by first passing hydrogen through dog's blood until spectroscopic examination showed that all of the o.xyhemoglobin had been reduced to hemoglobin, then saturating the blood with pure nitric oxid prepared from copper and nitric acid, and finally again passing hydrogen through the blood to remove all traces of free nitric oxid. It was observed that the spectrum of hemoglobin had changed to one showing two bands in practically the same position as those of oxyhemo- globin. The blood saturated with nitric oxid was found to be darker in color than either arterial blood or that saturated with carbon monoxid, and on exposure to air or on treatment with ammonium sulphid it proved to be as stable as carbon-monoxid hemoglobin. NO-hemoglobin is mentioned but briefly in most of the recent texts on physiological or organic chemistry as being a hemoglobin derivative of but little practical importance. Abderhalden (191 1) and Cohnheim (191 1), however, describe this compound quite fully. EXPERIMENTS WITH NO-HEMOGLOBIN The following studies on NO-hemoglobin were conducted by the writer. Formation. — Nitric oxid was prepared by treating copper foil with concentrated nitric acid, and the production of the gas was carried on long enough to free the apparatus from higher oxids of nitrogen before conducting gas into the solution of hemoglobin. A simple apparatus was arranged so that the sample of blood or of hemoglobin to be treated was first saturated with pure hydrogen, then with nitric oxid, and finally again with hydrogen to remove all free nitric oxid. Great care was exercised to exclude air from the apparatus, since the presence of even a small amount of free oxygen results in the oxidation of nitric oxid to nitrogen peroxid, with the consequent production of nitrous and nitric acids, which act upon hemoglobin to form methemoglobin. NO-hemoglobin was prepared, in the manner described above, both from defibrinated blood and from a solution of oxyhemoglobin prepared by Hoppe-Seyler's method. It was found, as a rule, that if a solution of oxyhemoglobin or of defibrinated blood was treated with nitric oxid, with all precautions to exclude free oxygen, the NO-hemoglobin usually contained a small amount of methemoglobin. This fact has been noticed by other investigators and is due to the union of the loosely 214 Journal of Agricultural Research voi. m. N0.3 bound oxygen of the oxyhemoglobin with the nitric oxid to form nitrogen peroxid, which, as has just been noted, acts upon the hemoglobin to form methemoglobin. Some of the earlier investigators have found that small quantities of oxyhemoglobin could be reduced to hemoglobin by means of hydrogen; but it has been the experience of the writer that, when working with any considerable quantity of oxyhemoglobin, practically no reduction took place even after passing a current of hydrogen through an oxyhemoglobin solution for several hours. In practice it was found best to reduce oxyhemoglobin to hemoglobin by means of hydrazin hydrate before saturation with nitric oxid. Properties. — NO-hemoglobin in concentrated solution has a dark cherry-red color; in dilute solutions it has a light cherry-red color, in contrast to the bright-red color of oxyhemoglobin or to the purple-red color of hemoglobin in solutions of the same concentration. In solutions free from methemoglobin NO-hemoglobin is quite stable, and solutions of the compound have been kept in a refrigerated room at 32° to 35° F. for several weeks without apparent change. On boiHng a solution of NO-hemoglobin a brick-red precipitate is formed, in contrast to the dark-brown precipitate formed on boiling a solution of oxyhemoglobin or hemoglobin. NO-hemoglobin shows a characteristic spectrum consisting of a heavy band just at the right of the D line and a somewhat lighter and wider band a trifle to the left of the E line (fig. i). These absorption bands occupy practically the same positions as those of oxyhemoglobin, but are distinguishable from the latter in solutions of the same concentration by lower intensity, less sharply defined edges, and by the fact that when a solution of oxyhemoglobin is treated with a reducing agent — e. g., hydrazin hydrate — the characteristic single broad band of hemoglobin appears, while on treating a solution of NO-hemoglobin with the same reagent, no reduction takes place and the bands are not affected. NO-hemoglobin is practically unaffected on treatment with potassium ferricyanid in neutral solution, with vStokes's solution (ammoniacal ferro- tartrate), with sodium nitrite, or with hydrazin hydrate, but is grad- ually reduced by potassium ferricyanid in acid solution. When a solution of NO-hemoglobin is treated with ether in the pres- ence of a small quantity of alcohol or of dilute acid, a bright-red colored extract is obtained which shows a distinct absorption band just at the right of the D Une, and occasionally, in concentrated solution, a faint band at the left of the E line. In the absence of acid or alcohol no color is extracted by ether. In a previous paper the writer considered that the color extracted under the above conditions was NO-hemoglobin, but from work which he has done since that time it is evident that the color extracted by ether is a derivative of NO-hemoglobin, produced apparently by the reducing action of the alcohol or acid upon the NO- Dec. 15, 1914 Coloring Matter of Salted Meats 215 /4 e 1 2 c 0 3 1 S 6 7 4 a 9 10 /< III! 1 1 1 ^1 \ 1 1 1 1 iiiili 1 III! IIU III n 1 iiiiiiiiiiiiiiiiiiiiiiiiiiiiiii iiiiiiiiiMii 1 II iiiiiiii A 1 n B 1 !"ii!:;'ii| f I n C r' ii!e!ijj*iii'itiiiiiiiinii:ij!iiiiiii;:ii!;i;i,ni|i D E F ' D G 1 A B c 0 3 * £ 3 6? b 8 F 9 /o // 1 nil II IM nil III II nmnniniii I 1 III II Ijllll MMI II 1 II 1 111 nil III 1 1 1 II III! Fig. I. — Spectra of hemoglobin aod some of its derivatives: A , absorption spectrum of a solution of oxyhe- moglobin ;B, absorption spectrum of a solution of NO-hemoglobin; C. absorption spectrum of a solution of hemoglobin prepared by treating a solution of oxyhemoglobin with hydrazin hydrate; D, absorption spectrum of a solution of methemoglobin prepared by treating a solution of oxyhemoglobin with potassium fcrricyanid; E, absorption spectrum of an alkaline solution of hematin; F. absorption spectrum of a solu- tion of hcmochromogen prepared by treating an alkaline solution of hematia with hydrazin hydrate; G, absorption spectrum of a solution of NO-hemochromogcn. 2i6 Journal of Agricultural Research voi. in. No. 3 hemoglobin. In pure solutions nitric oxid is insoluble in ether. The nature of this ether-soluble derivative of NO-hemoglobin will be dis- cussed in connection with the color of cooked salted meats. Preparation of pure NO-hemoglobin in dry condition. — Twenty- five c. c. of a concentrated solution of NO-hemoglobin were cooled to 0° C, 6 c. c. of absolute alcohol previously cooled to the same tem- perature were added, and the dish was gently rotated. An abundant quantity of dark cherry-red crystals formed immediately. The dish was covered, placed in a refrigerated compartment for 24 hours at a temperature of —4° C, and the contents then filtered with the aid of suction in a room held at a temperature of -1- 1° to -1-4° C. There was obtained a quantity of reddish brown crystals which were partly soluble in water, giving a reddish brown solution which showed a spectrum of NO-hemoglobin contaminated by the presence of methemoglobin. The material was not sufficiently soluble in water to allow of recrystalliza- tion. A large number of trials were made with various methods of procedure in an endeavor to obtain pure NO-hemoglobin in dry condition, but without much success. CrystalHzation by a method using alcohol seems to change the NO-hemoglobin, in part at least, to methemoglobin. When crystallization was carried on in the presence of a reducing agent — e. g., hydrazin hydrate or Stokes's solution — moist crystals could be obtained which showed a spectrum of NO-hemoglobin free from methemoglobin; but on drying the crystals in vacuo over sulphuric acid a change to methemoglobin took place. Crystallization of NO-hemoglobin. — In general, the method used by Reichert and Brown (1909) for the crystallization of hemoglobin and oxyhemoglobin was followed. In brief, the procedure was as follows: A pure concentrated solution of NO-hemoglobin was prepared by the methods previously described, and was examined spectroscopically to determine its freedom from other hemoglobin derivatives. Ammonium oxalate, in the proportion of 2 gm. to 100 c. c, was then added to the solution, which was then shaken to dissolve the salt. All subsequent procedure was carried on in a refrigerated room at a temperature of -t- 1° to+5°C. A few drops of the NO-hemoglobin solution were placed on a micro- scopic slide and allowed to evaporate until a heavy, dry proteid ring had formed around the drop. The cover glass was then carefully applied so as to exclude air bubbles, and the edges were sealed with balsam. Microscopic examination was made immediately after mounting and at intervals thereafter, according to the rate of crystal formation. The formation of crystals started at the dried proteid ring and proceeded toward the center of the mount, although occasionally crystals of ammo- nium oxalate would form, and the hemoglobin crystals would start from this base. Dec. 15, I9I4 Coloring Matter of Salted Meats 2 1 7 The above method was followed for the crystallization of both oxy- hemoglobin and NO-hemoglobin from sheep, ox, and pig blood, and of methemoglobin from ox blood. It was found practically impossible to secure crystals of hemoglobin or of its derivatives from the blood of the above-mentioned animals without the use of ammonium oxalate. It was also found necessary to carry on the work at a temperature slightly above freezing, since the crystals would not form readily at room temperature. Ox-blood oxyhemoglobin. — Plate XXXII, figures i and 2, shows crystals of oxyhemoglobin from ox blood. These crystals correspond very closely with those described by Reichert and Brown. No attempt was made to make a critical study of the crystallography of any of the various hemoglobin compounds studied. Ox-blood NO-hemoglobin. — Plate XXXII, figure 3, shows crystals of this compound. It was found very difficult to obtain crystals of NO-hemoglobin, owing, apparently, to its greater solubility as compared with oxyhemoglobin. The crystalUne structure, it will be noted, is distinctly different from that of oxyhemoglobin. Ox-blood methemoglobin. — Plate XXXII, figures 4 and 5, shows crystals of this compound which were prepared by treating oxyhemo- globin with potassium ferricyanid. It will be noted that the crystals of this compound are very similar in structure to those of NO-hemoglobin. Sheep-blood oxyhemoglobin. ^Plate XXXIII, figures i and 2, shows crystals of this compound which correspond very closely with those described by Reichert and Brown. Sheep-blood NO-hemoglobin. — It was not found possible to secure a thoroughly satisfactory mount of these crystals, owing to their high solubility. Plate XXXIII, figure 3, shows fair crystals of this com- pound in the form of plates, which, however, are largely obscured by the large crystals of ammonium oxalate. It may be noted, however, that the crystals of NO-hemoglobin are distinctly different from those of oxy- hemoglobin from the same source. Pig-blood NO-hemoglobin. — Plate XXXIII, figures 4 and 5, shows crystals of NO-hemoglobin obtained from pig's blood. It was not found possible to secure a good mount of oxyhemoglobin crystals from pig's blood, but the NO-hemoglobin crystals shown in this illustration are very different in structure from the oxyhemoglobin crystals from pig's blood described by Reichert and Brown. The above work shows that NO-hemoglobin derived from ox, sheep, and pig blood is a crystallizable compound, with a definite structure, depending upon the species from which it has been obtained, and that its structure is different from oxyhemoglobin crystals from the same sources. 2i8 Journal of Agricultural Research voi. m. no. j COLOR OF UNCOOKED SALTED MEATS To a sample of finely ground fresh beef was added 0.2 per cent of potas- sium nitrate, and the material was placed in a refrigerated room at a temperature of 34° F. for seven days. At the end of that period the meat had a bright-red color, but gave evidences of incipient putrefac- tion. On extraction with 95 per cent alcohol, a faintly red-colored extract was obtained, which on spectroscopic examination showed a faint band just at the right of the D line. The color of the extract quickly faded. On extraction with water the sample yielded a dark-red extract, which retained its color without change for several days. On spectro- scopic examination the following spectrum was observed: Just at the right of the D line a very heavy dark band, and a trifle to the left of the E line a heavy band, but wider than and not quite so dark as the first band and with less sharply defined edges. These bands correspond exactly in position with those of oxyhemoglobin and NO-hemoglobin, but more closely resemble the latter. On treatment with sodium-nitrite solution the extract changed some- what in color to a reddish brown, but still showed the two absorption bands noted above, and in addition showed a methemoglobin band at the left of the D line. The presence of this band would indicate that the extract contained some oxyhemoglobin, which on treatment with sodium nitrite changed to methemoglobin. On treatment with hydrazin hydrate the extract was not changed in color, and showed the following spectrum: Just at the right of the D line a very heavy band; at the left of the E line a somewhat lighter and wider band with less sharply defined edges. This spectrum corresponds exactly with that of the original extract. The position of the absorption bands just noted, the fact that these bands were unaffected on treatment of the extract with hydrazin hydrate, and the solubility of the color in water are sufficient proof that the color of the meat under examination was due to the presence of NO-hemo- globin. The production of NO-hemoglobin in the sample of meat under exami- nation is easily explained. The potassium nitrate added to the meat had been reduced, either by bacterial or enzym action or by both, to potas- sium nitrite, nitrous acid, and nitric oxid, with the consequent formation of NO-hemoglobin. Samples of various uncooked meats and sausages in which saltpeter had been used as one of the curative agents were obtained on the market for the purpose of studying the coloring matter of these products, and the following results were obtained: Smoked pork shoulder. — The freshly cut surface of the meat was bright red in color. On extraction with water the finely ground lean meat gave a dull-red colored extract, which showed the following spec- trum : Just at the right of the D line a fairly heavy dark band. ■ No other Dec. 15, I9I4 Coloring Matter of Salted Meats 219 bands were visible. On treatment with potassium ferricyanid the red color of the solution changed to brown and the band disappeared. On extraction with 95 per cent alcohol a red-colored extract was obtained which showed the following spectrum: Just at the right of the D Hne a fairly heavy band, in the same position as noted with the water extract. On treatment with potassium ferricyanid the color of the extract changed to brown, and the absorption band disappeared. The addition of sodium nitrite or of hydrazin chlorid caused practically no change in the color or spectrum of the extract. Smoked pork ham. — The cut surface had a bright-red color. On extraction of the finely-ground lean meat with water a red-colored extract was obtained which showed the following spectrum: A fairly heavy band just at the right of the D line. The addition of sodium nitrite or of hydrazin hydrochlorid caused no change in color of the extract. Salami sausage. — This product is known as a dry summer sausage, and is cured by hanging in a dry and well ventilated room for some time ; it is also subjected to a light smoking. The cut surface of the meat had a bright-red color. On extraction with alcohol a bright-red colored extract was obtained which showed the following spectrum: A heavy dark band just at the right of the D line. On treatment with hydrazin chlorid the solution became somewhat milky in appearance, but the color and spectrum were not afifected. Treatment with sodium nitrite caused no change in the color or spectrum of the extract, while potassium ferri- cyanid changed the red color of the solution to brown, and the absorption band disappeared. Dried beef. — This product is first cured in a brine containing salt, sugar, and saltpeter, and is then smoked. The cut surface of the meat was dark red. On extraction with alcohol a bright-red colored extract was obtained, which showed the following spectrum: A heavy dark band just at the right of the D line. Treatment with reducing agents gave the following results: Potassium ferricyanid destroyed the red color, and the absorption band practically disappeared; sodium nitrite partially destroyed the red color, but the absorption band at the right of the D line was still visible; hydrazin hydrate did not affect the color or spectrum of the extract. On extraction with water a faintly red-colored solution was obtained. The meat residue after extraction had a bright-red color. Smoked cervelat sausage. — This product is prepared in much the same manner as salami sausage. On extraction with alcohol a light-red colored extract was obtained which gave an absorption spectrum show- ing a distinct band just at the right of the D line. Treatment with hydrazin hydrate did not affect the color or spectrum of the solution, while sodium nitrite, hydrazin chlorid, and potassium ferricyanid de- stroyed the red color of the extract, and the absorption band disappeared. 67235°— 14 3 220 Journal of Agricultural Research voi. m, N0.3 Treatment of the sausage with water extracted no color, while ether gave a light-red colored extract, the color of which faded rapidly. Spec- troscopic examination showed a distinct band just at the right of the D line. Cured pork shoulder, not smoked. — The freshly cut surface of the meat had a bright-red color. On treatment with water a light-red col- ored extract was obtained which showed the following spectrum: A dis- tinct band just at the right of the D line. The color of the extract faded rapidly. On extraction with alcohol the meat gave a red-colored extract, showing a spectrum with a distinct band just at the right of the D line. Treatment with sodium nitrite or with hydrazin hydrate caused practically no change in the color or spectrum of the extract. Corned beef.-^u extraction with water a rather cloudy, red-colored extract was obtained which gave a spectrum showing a fairly distinct band just at the right of the D line. Treatment with sodium nitrite caused no change in the color or spectrum of the solution. On extraction with alcohol the meat gave a bright-red colored extract which showed a fairly heavy absorption band just at the right of the D line. Treatment of the extract with sodium nitrite did not affect the color or spectrum, while hydrazin hydrate, hydrazin chlorid, and potassium ferricyanid destroyed the red color and the absorption bands of the solution. SUMMARY OF RESULTS WITH UNCOOKED SALTED MEATS The results of the examination of the samples of uncooked salted meats reported above may be summarized as follows : The color of the meats was generally soluble in alcohol and in some cases in water. All samples gave extracts which showed an absorption band just at the right of the D line. In general, treatment with hydrazin hydrate or with sodium nitrite did not affect the color or spectrum of the extract. Potassium ferricyanid and hydrazin chlorid generally destroyed the red color of the extract and caused the absorption band to disappear. The fact that all samples of meats examined gave a red-colored extract with alcohol while only a part of them yielded a red-colored extract with water may be explained in this way: NO-hemoglobin is readily soluble in water, but insoluble in alcohol or ether. However, if a little alcohol is first added to a solution of NO-hemoglobin — svifiicient to cause a slight precipitation of the proteid — and the solution is then extracted with ether, a bright-red colored extract may be obtained which gives a spectrum showing an absorption band just at the right of the D line. The coloring matter extracted by ether is undoubtedly a decomposition product of NO-hemoglobin. In the case of the meats examined it would appear that all samples giving with water a red-colored extract that showed an absorption band Dec. IS, J9I4 Coloring Matter of Salted Meats 221 at the right of the D line had as their coloring matter NO-hemoglobin. In the case of those samples giving a red-colored extract with alcohol and showing an absorption band at the right of the D line but giving no red-colored extract with water, the coloring matter of the meat was not NO-hemoglobin, but a derivative of that compound. In the case of those samples which gave red-colored extracts with both alcohol and water and which showed an absorption band just at the right of the D line, the cqlor is certainly due, in part at least, to NO-hemoglobin, and it may be due, in part, to a derivative of NO-hemoglobin present as such in the meat, or the alcohol may break down the NO-hemoglobin during extraction. The evidence is ample to show that the action of saltpeter in the curing of meats is primarily to cause the formation of NO-hemoglobin; but it is very possible that under certain conditions of manufacture or processing to which salted meats are subject, the NO-hemoglobin may undergo changes. RED COLOR OF COOKED SALTED MEATS Haldane has shown that the red color of cooked salted meats is due to the presence of NO-hemochromogen, a reduction product of NO- hemoglobin to which the color of uncooked salted meats is due. NO- hemochromogen is characterized by its solubility in alcohol, its resistance to the action of reducing agents, and by a spectrum showing a distinct band just at the right of the D line and a faint band a trifle to the left of the E line. While Haldane's work seems to show clearly that the color of cooked salted meats is due to NO-hemjjthromogen, it has seemed desirable to study the subject further and to determine especially if the NO-hemoglobin of uncooked meats be reduced to NO-hemochromogen under other conditions than by cooking. The fact that in the examina- tion of certain uncooked salted meats a coloring matter had been obtained similar to NO-hemoglobin yet not possessing all of the properties of that compound, as has already been noted, led the writer to believe that the coloring matter of some uncooked salted meats might be due, in part at least, to NO-hemochromogen. NO-hemochromogen is but briefly mentioned in the literature. The compound is described by Linossier (1887), Haldane (1901), and by Abderhalden (191 1). The following experiments in the production of this compound were conducted by the writer. A dilute alcoholic, ammoniacal solution of hematin was saturated, first with hydrogen to remove air, then with nitric oxid, and finally again with hydrogen. On treatment with nitric oxid the brown color of the solution gradually changed to reddish brown; the characteristic spectrum of alkaline hematin disappeared, and in its place appeared a spectrum showing a fairly heavy band just at the right 222 Journal of Agricultural Research voi. m. No. 3 of the D line and corresponding to the spectrum of NO-hemochromogen, as described by Haldane, except that the second band at the left of the E line was not visible. The addition of Stokes's solution caused no change in the color or spectrum of the solution. On standing overnight the red color of the solution had changed to brown and the spectrum had changed to that of alkaUne hematin. A dilute alcoholic, ammoniacal solution of hematin was treated with a small quantity of a concentrated solution of hydrazin sulphate. The brown color of the hematin solution quickly changed to the bright pink of hemochromogen, with a corresponding change in spectrum. The hemo- chromogen solution was then saturated with hydrogen, nitric oxid, and hydrogen in turn. On saturation with nitric oxid the pink color of the solution quickly changed to cherry red, and spectroscopic examination showed a heavy dark band just at the right of the D line. No other bands were observed, even though the solution was examined in a deep absorption cell. The two substances produced, either by saturation of a solution of hematin with nitric oxid or by similar treatment of a solution of hemo- chromogen, are apparently identical. The evidence seems to show that this substance is the compound NO-hemochromogen. The structural relation between NO-hemoglobin and NO-hemochro- mogen is simple. NO-hemoglobin is a molecular combination of nitric oxid and hemoglobin — the latter compound consisting of the proteid group, globin, on one hand, and the coloring group, hemochromogen, on the other. NO-hemoglobin and NO-hemochromogen differ from each other simply in that one contains the proteid group, globin, while the other does not. Apparently, then, a method of treatment which would split off the globin group from NO-hemoglobin should result in the pro- duction of NO-hemochromogen, provided, of course, that the procedure did not in turn change or destroy the NO-hemochromogen produced. As has already been noted by Haldane, it was found that when a solu- tion of NO-hemoglobin was heated to boiling a brick-red precipitate formed, in contrast to the dark-brown precipitate which formed on heat- ing a solution of oxyhemoglobin or of blood. The brick-red precipitate was filtered off and was then extracted with alcohol, which gave a light- red colored extract showing a spectrum with a fairly heavy band just at the right of the D line. This spectrum corresponds with that of NO-hemochromogen. On standing, the color of the extract faded rapidly. A solution of oxyhemoglobin was treated with one-hundredth normal nitrous acid and showed a spectrum corresponding to that of NO-hemo- globin. The color of the solution changed from the bright red of oxy- hemoglobin to the dark cherry-red characteristic of NO-hemoglobin. Treatment with sodium nitrite did not affect the color or spectrum of the Dec. 15, I9I4 Coloring Matter of Salted Meats 223 solution. The production of NO-hemoglobin by this method is un- doubtedly due to the action of the nitric oxid, liberated from the unstable nitrous acid, upon the hemoglobin. On heating the solution to boiling a brick-red precipitate was formed which, after filtration and on extrac- tion with alcohol, gave a light-red colored extract which showed a spectrum with a fairly heavy band just at the right of the D line. Treat- ment with sodium nitrite did not affect the color or spectrum of the solution. A piece of lean beef about 3 inches square and 1% inches thick was placed in a cold o.i per cent solution of sodium nitrite, and the sokition was brought to the boihng point and boiling continued for one-half hour. Immediately on placing in the cold nitrite solution the surface of the meat turned dark brown in color; but on boiling, the meat took on a bright-red color, similar to that of cooked corned beef. On cutting the cooked meat the cut surface showed a bright-red color extending from the surface for about one-fourth of an inch toward the center of the piece, the interior of the meat being dark brown. The meat was ground and extracted with alcohol, which gave a bright-red colored extract showing a spectrum with a quite heavy band at the right of the D line and a very faint band at the left of the E line. The color which Lehmann obtained on extracting with alcohol meat which had been cooked in water containing nitrites and acid, he named "haemorrhodin." Results obtained by Haldane and by the writer seem to show, however, that the compound described by Lehmann is none other than NO-hemochromogen. The following kinds of salted meats were used in studying the color of cooked salted meats: Salami, cervelat, Frankfurter, and Bologna sausages, corned beef, cured pork shoulder (unsmoked), smoked shoulder, and smoked ham. Bologna and Frankfurter sausages are cooked in the proc- ess of manufacture, so that these products are cooked salted meats. Portions of each of the samples of meats were cooked for some time in boiling water, then finely ground and extracted with alcohol. In all cases a red-colored extract was obtained, the intensity of the color vary- ing with the product. On spectroscopic examination the extract from each sample showed a distinct band just at the right of the D line, but in practically all cases a second absorption band was not visible. The addition of sodium nitrite, as a rule, did not affect the color or spectrum of the extract. The color of the cooked salted meats was soluble in ether, but insoluble in water. The evidence seems to show very clearly that the color of cooked salted meats is due to the NO-heniochromogcn resulting from the reduction of the NO-hcmoglobin of the raw salted meats on boiling. It was found that the alcoholic extracts from the same product, whether in raw or cooked condition, showed practically identical prop- 224 Journal of Agricultural Research voi. m. no. 3 erties. This does not signify that the coloring matter of raw salted meat is necessarily NO-hemochromogen, but rather it appears that treatment with alcohol reduces the coloring matter of uncooked salted meats, NO-hemoglobin, to NO-hemochromogen. In the case of certain uncooked salted meats NO-hemochromogen may be present as such before extrac- tion with alcohol. The best means of differentiating between NO- hemoglobin and NO-hemochromogen as the coloring matters of salted meats seems to be on the basis of their differences in solubiUty. NO- hemoglobin -is soluble in water but insoluble in ether and alcohol, while NO-hemochromogen is soluble in ether and alcohol and insoluble in water. In other respects the properties of the two compounds are very similar. In the case of certain kinds of uncooked salted meats and meat food products — e. g., summer sausage and dried beef — the coloring matter seems to be NO-hemochromogen rather than NO-hemoglobin. The color is insoluble in water, but is soluble in alcohol and exhibits other properties of NO-hemochromogen. It is very probable that in the case of meats which have been cured with saltpeter or of meat food products in which saltpeter has been used in the process of manufacture, the reduction of NO-hemoglobin to NO-hemochromogen takes place to a greater or lesser degree, depend- ing upon conditions of manufacture and storage. The two compounds are so closely allied that their differentiation in one and the same product is not a matter of great importance. CONCLUSIONS The results of the investigation reported in this paper may be briefly summarized as follows : The color of uncooked salted meats cured with potassium nitrate, or saltpeter, is generally due, in large part at least, to the presence of NO- hemoglobin, although the color of certain kinds of such meats may be due in part or in whole to NO-hemochromogen. The NO-hemoglobin is produced by the action of the nitric oxid resulting from the reduction of the saltpeter used in salting upon the hemoglobin of the meat. The color of cooked salted meats cured with saltpeter is due to the presence of NO-hemochromogen resulting from the reduction of the color of the raw salted meat on cooking. Dec. IS, i9'4 Coloring Matter of Salted Meats 225 LITERATURE CITED Abderhalden, Emil. igii. Stickoxydhamoglobin. In his Biochemische Handlexikon, Bd. 6. p. 212, Berlin. CoHNHEiM, Orro. igii. Chetnie der Eiweisskorper, 3d ed., 388 p. Braunschweig. Glage, Max. igog. Die Konservierung der roten Fleischfarbe. 27 p. Berlin. Haldane, John. igoi. The red colour of salted meat. 7« Jour. Hyg., v. i, no. 1, p. 115-122. Hermann, Ludimar. 1865. Ueber die Wirkungen des Stickstoffoxydgases auf das Blut. In Arch. Anat., Physiol., und Wissensch. Med., Jahrg. 1865, p. 46g-48i. ' Hoagland, Ralph. igoS. The action of saltpeter upon the color of meat. In V. S. Dept. Agr., Bur. Anim. Indus. 25th Ann. Rept., igo8, p. 301-314. Hoppe-Seyler, F. 1864. Ueber die optischen und chemischen Eigenschaften des Blutfarbstoffs. In Centbl. Med. Wissensch., 1864, no. 52, p. 818-821, no. 53, p. 834-837. KissKAXT, Karl. i8gg. Beitrage zur Kenntnis der Ursachen des Rothwerdens des Fleisches beim Kochen, nebst einigen Versuchen liber die Wirkung der schwefligen Saure auf die Fleischfarbe. In Arch. Hyg., Bd. 35, Heft i, p. ri-i8. Lehmann, K. B. i8gg. Ueber das Haemorrhodin, ein neues weitverbreitetes Blutfarbstoffderivat. In. Sitzber. Physikal. Med. Gesell. zur Wiirzburg, Jahrg. iSgg, no. 4, p. 57-61. Linossier, Georges. 1887. Sur une combinaison de I'hematine avec le bioxyde d 'azote. In Compt. Rend. Acad. Sci. [Paris], t. 104, p. I2g6-i2g8. Orlow, S. igo3. La coloration des saucisses et des jambons. /« Rev. Intemat. Falsif., ann. 16, liv. I, p. 36-38. Reichert, Edwin Tyson, and Brown, Amos Peaslee. igog. The differentiation and specificity of corresponding proteins and other vital substances in relation to biological classification and organic evolu- tion: the crystallography of hemoglobins. Carnegie Inst, of Wash., Pub. 116, 338 p., 411 fig., 100 pi. Weller, H., and Riegel, M. 1897. Zum Nachweise der Farbung von Wurstwaren. In Forsch.-Ber. uber Lebensm. [etc.], Jahrg. 4, p. 204-205. PLATE XXXII Fig. I. — Oxyhemoglobin, ox blood. Fig. 2. — Oxyhemoglobin, ox blood. Fig. 3. — NO-hemoglobin , ox blood. Fig. 4. — Methemoglobin, ox blood. Prepared by treating oxyhemoglobin with potassium ferricyanid. Fig. 5. — Methemoglobin. ox blood. Prepared by treating oxyhemoglobin with potassium ferricyanid. (226) Coloring Matter of Salted Meats Plate XXXII Journal of Agricultural Research Vol. III. No. 3 Coloring Matter of Salted Meats Plate XXXIil Journal of Agricultural Research V.jI. Ill, Nu. 3 PLATE XXXIII Fig. I. — Oxyhemoglobin, sheep blood. Fig. 2. — Oxyhemoglobin, sheep blood. Fig. 3. — NO-hemoglobin, sheep blood. Fig. 4. — NO-hemoglobin, pig blood. Fig. 5. — NO-hemoglobin, pig blood. OIL CONTENT OF SEEDS AS AFFECTED BY THE NUTRITION OF THE PLANT By W. W. Garner, Physiologist in Charge, H. A. Allard, Assistant Physiologist, and C. L. FoubERT, Scientific Assistant, Tobacco and Plant-Nutrition Investiga- tions, Bureau of Plant Industry INTRODUCTION Although oils and fats are very widely distributed in the plant world, the commercial supply of these products is derived chiefly from a com- paratively small number of species, and in most cases the seeds of these plants furnish the raw material. In general, the oil produced by the plant is usually stored in the seed or- other reproductive parts, and for this and other reasons the seed constitutes the most favorable material for a study of the quantitative production of oil in plants. The seed as a rule varies less in composition than other plant parts, as would be inferred when we consider its relatively small size along with the fact that it normally possesses the ability to reproduce in detail the distinctive characters of the parental type. Nevertheless, when grown under widely different conditions, seeds frequently show such marked changes in composition that their agricultural or com- mercial value is materially affected. The composition of certain seeds, more particularly wheat and other grains, as influenced by environ- ment, has been extensively investigated in so far as relates to their content of protein, carbohydrate, and ash; but no extensive investiga- tion of the oil content of seeds as affected by the various factors of nutrition has thus far been reported. There are on record, however, numerous analyses of oleaginous seeds grown in different regions, which indicate marked differences in oil content, presumably due, at least in part, to the varying conditions under which the seeds were produced. Some of our important crop plants are of value, primarily, for the oil contained in the seed; and it is a matter of practical importance to ascertain, so far as possible, the most favorable conditions for obtain- ing maximum yields of oil. It was with this object in view that an investigation was undertaken, some of the results of which are presented in the present paper. The true fatty oils, composed of glycerin esters of the fatty acids, are quite different chemically from the carbohydrates; and, in fact, the two groups of compounds have little in common, except that both con- tain only the three elements — carbon, hydrogen, and oxygen. On the other hand, there is a very intimate and significant physiological rela- Journal of Agricultural Research, Vol. Ill, No. 3 Dept. of Agriculture. Washington. D. C. Dec. 15, 1914 G-J7 (2»7) 228 Journal of Agricultural Research voi. m. no. 3 tionship between the carbohydrates and the oils in the seed and other parts of the plant. Outside of the living cell the transformation of carbohydrate into fat, or the reverse process, has not been accomplished; but both of these processes are readily carried out by the living proto- plasm. The researches of Miintz (1886),' Leclerc du Sablon (1895, 1896), Gerber (1897 a, b), Ivanow (1912), and others with the poppy (Papaver somniferum), flax {Linum iisitatissimum), rape {Brassica spp.), soy bean (Glycine hispida), castor-oil plant (Ricinus communis), walnut (Juglans regia), sweet almond (Amygdalus communis), hemp (Cannabis saliva), and sunflower (Helianthus a7muus) all go to show that the de-\-el- opment of oleaginous seeds is characterized by a progressive accumula- tion of oil accompanied by a corresponding decrease in carbohydrates. Under proper conditions this transformation takes place in unripe seeds detached from the mother plant, further indicating that the oil is derived from the carbohydrate. Although oleaginous seeds in general are rela- tively rich in protein and the accumulation of oil proceeds simultane- ously with that of protein, no evidence exists that there is any direct relationship between the two processes. From the researches of the investigators mentioned, together with those of Schulze (1910) and his pupils, it may be inferred that the plant during the period prior to blooming normally accumulates an adequate supply of nutrients, chiefly in the form of carbohydrate and protein, to insure the development of the seed. At or near the blooming stage there begins a general movement of nutrients in the form of the simpler sugars and soluble nitrogenous constituents through the stem toward the re- productive parts. During the irrst stages of development of the seed the carbohydrates are laid down largely in the form of cellulose and hemicellulose, in order to provide for the early development of the testa, or seed coat, which serves first as conductive tissue for the embryo and in later stages as a protective membrane but not as a depository of surplus food. Then follows in the seed of some species a marked accumulation of so-called reserve carbohydrates, mainly starch and hemicellulose, while in other species the nitrogen-free "reser\'e" food accumulates in the fonn of oil. Examination of the stem parts during the development of the seed has shown that the organic non-nitrog- enous nutrients flowing into oleaginous seeds are largely the same as for starchy seeds — namely, soluble carbohydrates. While there appears to be no doubt that the oil in the plant cell, at least in the higher plants, is derived from carbohydrate, the mechanism of the reactions involved is not understood, since the transformation is not known outside the living cell. 1 Bibliographic citations in parentlieses refer to " Literature cited," p. 249. Dec. 15. 1914 Oil Content of Seeds and Nutrition of Plant 229 PROBLEMS INVOLVED AND GENERAL METHODS OF PROCEDURE The growth, development, and composition of the individual plant depend on the combined action of heredity and enviromnent. It fol- lows that in any investigation of heredity or of environment in relation to plant life such methods should be followed as will make it possible to distinguish between these two forces. This principle of clearly dis- tinguishing between the effects of environment and those due to heredity is simple enough in theory, but in practice it is often extremely difficult to follow because of interrelations between the two forces which are not fully understood. Thus, nouheritable variations in the progeny of relatively pure strains grown under what would seem to be uniform environmental conditions become especially apparent when we come to deal with quantitative differences such as are involved in a study of variable chemical composition as influenced by the factors of nutrition. In our experiments on the nicotine content of tobacco, the oil content of seeds of various species, and other quantitative characters, efforts to insure the greatest unifonnity possible in environmental conditions have failed to effect anything like uniform composition in the progeny of individuals representing the purest strains available. Some of the strains of tobacco under study had been inbred for eight generations. It follows that in dealing with quantitative differences in composi- tion as influenced by environment a sufficiently large number of indi- viduals must be used to avoid misleading results due to variations that can not be brought under control. Where conditions have made it impracticable to grow relatively large numbers of plants for study, the alternative of repeating the experiment has been adopted. It is equally important to use caution in reaching generalizations based on results obtained with an insufficient number of species or varieties. In the course of our work on the oil content of seeds it has been fre- quently observed that in the case of the soy bean, for example, different varieties arc not always influenced in the same manner by the environ- ment. In such plants as the soy bean and cotton, therefore, as many varieties as practicable have been included in the experiments. In selecting plants for study those that are of special importance by reason of the oil content of the seeds and which are otherwise adapted to the work in view have been taken. Thus far we have utilized cotton (Gos- sypium spp.), soy bean, peanut (Arachis hypogaea), and sunflower. With reference to chemical composition, the commercial value of seeds evidently depends on the relative percentages of their several constituents, but from the standpoint of the grower the returns also depend on the quantity of seed produced — that is, on the number and size of the seeds produced by the individual plant or on a given acreage. Hence, in the 230 Journal of Agricultural Research voi. m. no. 3 present investigation the weight of the seed in addition to its percentage composition was determined in order to ascertain the actual quantity of oil which it contained. When practicable, data were collected as to the size and character of growth of the plant as a whole for purposes of correlation and more particularly in dealing with specific factors of the nutrition process as a measure of the relative effects produced by the special conditions of the experiment. The actual yield of seed has not received any special attention, since this is greatly affected by factors of nutrition which have little influence on the problems involved, and, moreover, the matter of crop yields constitutes a separate problem. In analyzing the seed the official method' for the determination of oil was followed, using anhydrous ether as the solvent. OIL CONTENT OF SEED AT SUCCESSIVE STAGES OF DEVELOPMENT It has been shown that the accumulation of oil in the seed does not set in actively during the very earliest stages in the development of the seed, and the work of Ivanow (1912) suggests that there is a period of very intense oil formation, which occurs about midway between bloom- ing and the final maturity of the seed. This raises the question as to the existence of a "critical period" in oil formation which would have an important bearing on the effects of external conditions on the quantity of oil produced. Samples of soy beans and of cotton seed were collected at different stages of maturity, to secure more definite information on this point. The weight of the seed also was obtained in each case, so as to ascertain the changes in the absolute as well as the relative oil con- tent. To arrest respiration promptly and thereby avoid changes in com- position, the immature seeds were dried in the oven at 70° to 80° C. In the case of the soy bean 5 to 6 pods were picked from each of 100 to 125 plants in taking the samples. The material was grown at the Arlington Experiment Farm, Va., the Peking soy bean in 1910 and the others in 191 2. The several pickings from each variety were taken from the same plants at stated intervals, but for obvious reasons the different pickings do not necessarily represent the actual growth made by the beans in the intervals covered. They are strictly comparable, however, as to the relation between the oil content and the size of the seed at the several stages of maturity. In the case of the cotton seed care was taken to obtain the same number of immature and mature bolls from each plant, and these were always taken from the same branch, about 12 plants being used for each pair of samples. The results obtained are shown in Tables I and II. ' Wiley, H. W.. et al. Official and provisional methods of analysis. Association of Official Agricultural Chemists. U. S. Dept. Agr., Bur. Chem. Bui. 107 (rev.), 272 p., 11 fig. 1908. Dec. 15, 1914 Oil Content of Seeds and Nutrition of Plant 231 Table I. — Oil content of soy beans gathered at various stages 0/ maturity at Arlington Experiment Farm, Va., in iQio and igi2 VARIETY, s. p. I. NO. 19981 Date har- vested. Weight of I.OOO beans. Moisture in beans. Oilin moist beans. Oilin 1,000 beans. Date har- vested, • Weight of 1.000 beans. Moisture in beans. Oilin moist beans. Oilin 1,000 beans. Gm. Per cent. Per cent. Gm. Gm. Per cent. Per cent. Gm. Aug. 13... 3-53 7.50 2.95 0. 104 Sept. II. . 356.8 6.60 18.05 64-4 20. . . 39.3 5.50 10. 65 4. 19 16.. 447.0 7.55 18.02 80. s 23. . . 96.24 5.80 13.40 12.9 21 . . 472.3 5-95 18.60 87.8 28. . . 146. 2 6. 10 15.94 23.3 25.. 498.7 7.44 18.37 91. 6 Sept. 3... 263.7 5.40 16. 70 44.0 3°-- 487.0 6.35 18.85 90.8 6... 282. 0 6-55 17. 00 47.9 Oct. 7 . . 479-3 7.55 ^1-11 85.2 VARIETY, S. P. I. NO. 21755 July 24. . 8.2 7.00 3.80 0.312 Aug. 13... 158.0 5.80 14-75 23.3 30... 32.3 5.50 9-38 3.03 20. . . 217.4 6. 00 15.83 34.4 Aug. 3... 8i.8 6.30 11.58 9.48 26. . . 210. 5 5.00 15.65 32.9 / • ■ ■ 98.3 5- 40 12.52 12.3 VARIETY, S. P. I. NO. 32907 Aug. 20 . . . 3.28 6. 70 3.38 0. no Sept. 4. . . 67.3 .5. 85 16.95 II. 4 23... 12.5 6. 50 7-65 .955 7... 72.1 s. 70 17. 20 12.4 27... 23.2 5.15 12. 10 2.81 II . . . 85.0 6. 00 18. 10 15.4 30. . . 49. 2 5- 15 14-95 7.36 20. . . 82.5 5-55 18.28 15.1 Table II. — Oil content oj immature and mature cotton seed grown at Thompsons Mills, Ga., in igio Variety and condition of seed. Lint. Weight of 1,000 seeds. Hulls. Meats. Moisture in meats. Oil in moist meats. Oilin whole seed. Oilin 1,000 seeds. Toole: Immature Percent. 41.8 38.1 31-8 30.8 35-3 33- 5 Gm. 78.4 93-3 102. 0 134-4 133-0 155-6 Percent. 44-6 38.8 45.6 42.4 40.4 .39.3 Per cent. 55-4 61.2 54.4 57.6 59-6 60. 7 Per cent. 3.80 4.05 3.72 4. 00 3. 95 4.05 Per cent. 37.65 38.03 34.60 36.17 35.80 37.93 Per cent. 20. 9 23.3 18.8 20. 9 21.4 22. 9 Gm. 16. 4 21.7 19.2 28.2 28. s 35.7 Mature Trice: Immature Mature Sam McCall: Immature Mature In all cases the relative weight of the seed rather than the date of har- vesting is to be taken as the more nearly correct index of the stage of development. The first pickings of the soy beans were made when the seeds were exceedingly small, and the final pickings represent the fully matured seed. The results are definite and conclusive. Except for the period immediately following blooming and that directly preceding final 232 Journal of Agricultural Research voi. 111.N0.3 maturity, there is throughout the development of the seed a gradual and rather uniform gain in the oil content as compared with the growth of the seed. There is no evidence of a definite " critical period " for the accumu- lation of oil during the development of the seed. Considering only the per- centage of oil, there is a very sharp increase during the first few weeks after blooming, and then only a slow gain until near the end of the ripening. During the final stage of ripening there is a decrease both in the size of the seed and in the oil content. This phenomenon, which was observed also by Miintz (1886), is probably due to continued respiratory activity after assimilation has ceased. In the case of the cotton seed the immature samples were taken when the green bolls had reached full size and had begun to show numerous brown spots. As in soy beans, the increase in oil proceeds somewhat more rapidly than the growth of the seed. OIL CONTENT OF SEED AS AFFECTED BY RATIO OF LEAF SURFACE TO QUANTITY.OF SEED PRODUCED It has been shown that the accumulation of oil in the seed proceeds throughout the greater portion of the period of development and ripening. It also has been pointed out that during the so-called vegetative period preceding blooming there is an accumulation of carbohydrates in the leaves and stems which is later utilized in the formation of the oil deposited in the seed. From these facts it might be inferred that the total quantity of oil stored in the seed would be affected by the relative extent of the photosynthesizing plant parts, more particularly the leaves, while, on the same basis, the percentage content of oil might be influenced by the ratio between the photosynthesizing parts and the quantity of seed produced. In other words, it is to be expected that premature shedding of leaves, such as often happens under adverse conditions, or the shedding of a por- tion of the blossoms would affect the accumulation of oil in the seed. As regards the quantity of seed produced, a diminished supply of accumu- lated carbohydrate might lead to the production of a smaller number of seeds or of smaller sized seeds, or possibly both. For the reasons indicated it is apparent that in any analytic study of oil production in the seed as related to factors of nutrition account must be taken of the possible effects of the nutrition conditions during the two principal life periods of the plant — namely, the vegetative and the repro- ductive. The production of oil by the plant requires favorable conditions for the accumulation of carbohydrate during the vegetative period and for the transformation of carbohydrate into oil during the second period, although there may be, of course, more or less overlapping of the two processes. As a special phase of the influence of the accumulation of car- bohydrate on oil fonnation, experiments were carried out with soy beans in which the normal distribution of the vegetative and reproductive plant parts was modified by partial defoliation and by removal of a portion of the blossoms or very young seed pods. The number of plants used in each Dec. 15. 1914 Oil Content of Seeds and Nutrition of Plant 233 experiment ranged from 35 to 50. Data were also obtained as to the effects of the two treatments on the general development of the plant, as indicated by the height, the weight of the air-dry root and stalk minus the leaves, and the total yield of seed. The results are presented in Tables III and IV. Table III. — Oil cotitent of soy beans as affected by partial defoliation Variety and treatment. Date of blooming. S. P. I No. 32907: Control Nimiber of leaves reduced to about 40 per cent of normal on June 29. July 16,22, 30, and Aug. 14 Number of leaves reduced to about 50 per cent of normal at same periods as above. ,' Control Number of leaves reduced to about 40 per cent of normal on Aug. I and 15 S. P.I. No. 2r755: Control Number of leaves reduced to about 40 per cent of normal on July 15 and 30. S. P. I. No. 30599: Control Niunber of leaves reduced to about 40 per cent of normal on July 15 and 2a S. P. I. No. 30745: Control Number of leaves reduced to about 40 per cent of normal on July IS and 22 S. P. I. No. 30593: Control Number of leaves reduced to about 30 per cent of normal on July IS and 75 percent Aug.i:; July 30 do., do.. ..do July 8 ..do ..do do., do., .do., do. .do.. Aver- age weight of stalk and root. Gm. 33-7 24.0 ,36.8 24.1 9-7 S-3 24-5 16.5 24- S 16.7 27.2 Aver- age height of plant. Aver- age yield of beans per plant Inches. 22. 2 20.4 25. 2 21. S II. I 9-7 27.0 21.5 24.1 19-2 24.4 S6.6 42.2 SI.2 30- 7 24.5 II. 5 61. I 55- 2 SS.6 32.1 59-4 33-8 Weight of 1. 000 beans. Gm. 89.8 8s-4 8s. 6 87.1 81.8 209.0 16S.S 179-4 167. S 192.6 180.0 190.3 160.4 Mois- ture in beans. P.ct. 7-4S 6.65 6.75 8.6s S-SS 6. 20 6. 25 7-95 6.05 6. 70 6. 40 Oil in moist beans. P.ct. 16.7s 17.80 17.20 17.85 16. 21 16.32 19-95 20.93 19.80 20.35 20.93 Oil in 1,000 beans. Gm. is-o 15-2 JS-o 14.6 33-9 27- S 35-8 3S-t 38-1 36.6 39-8 Considering, first, the effects of partial defoliation (Table III), in all cases the weight of the root and stalk, the height of the. plant, and the total yield of beans are decidedly reduced. The size of the beans, how- ever, is only slightly reduced, so that the decreased yield is due almost entirely to the smaller number of beans developed. It is an interesting fact that the small decrease in size of the beans is almost exactly offset by the increase in percentage of oil, so that the actual quantity of oil in the individual seed remains practically the same. This fact only holds apparently within certain limits, for in the case of the variety designated as " S. P. I. No. 30593," where the defoliation was nearly three-fourths complete, the decrease in. size of the bean was too great to be fully offset by the higher percentage of oil. The most striking excep- tion, however; is shown by S. P. I. No. 21755. Th's variety differs from the others in that the seed are matured a very short time after bloom- ing. Ill the present case the seed were fully ripe on August 26, whereas 67235°— 14 4 234 Journal of Agricultural Research Vol. in. No. 3 all of the remaining varieties were about five weeks later in reaching full maturity. It appears, therefore, that a degree of defoliation which but slightly modified the size and oil content of the seed in those varieties requiring a long period for the development of the seed brought about a much more decided effect in the variety which is able to fully develop rather large-sized seed in a very short period of time. Another reason for the greater effect of the defoliation on the size of the bean, as well as on the total yield of beans, in S. P. I. No. 21755 is that the fohage normally is less abundant than that of the other varieties. Table IV. — Oil content of soy beans as affected by partial removal of very young seed pods Variety and treatment. S. P. I. No. 32907: Control I^arge number of pods removed on Aug. 19 S. P. I. No. 21755: Control Larger portion of pods removed on July 15 and 22 S. P. I. No. 30599: Control Larger portion of pods removed on July 1 5 and 22 S. P. I. No. 3074s: Control Larger portion of pods removed on July 15 and 20 and Aug. 19.. S. P. I. No. 30593: Control More than two-thirds of pods removed on July 15 and 20 and Aug. 12 Aver- age Date of weight blooming. | of stalk and July 30. . ...do July 8. . . ..do.... ..do.... ..do.... ..do.... ...do.... ...do... do. Gm. 36.3 49-7 9-7 9.0 24-5 36.2 24.5 38.7 27-2 42.1 Aver- age height of plant. Aver- age yitld of beans per plant. Weight of I.OOO beans. Mois- ture in beans. Oil in moist beans. Inches. Gin. Gm. P.ct. P. a. 25-7 50.2 84. 1 6-45 17-15 27.2 55- 0 103. 1 6.6s 17-69 II. I 24. s 209.0 8.6s 16-21 10.8 9.8 204.6 8.00 15-83 27.0 61. 1 179-4 6.20 19-95 21-5 52.6 202.2 6.40 20-02 24.1 SS.6 192.6 7.95 19-80 24.2 39-1 227.8 6- 35 19- IS 24-4 59-4 190.3 6.70 20-93 23-3 25-7 244.4 6.50 •9-95 Oil in 1,000 beans. Gin. 14.4 iS. 2 33-9 32-4 35-8 40-4 38-1 43-7 39-8 48.8 It might be expected that the effects produced by removing a portion of the young seed pods would be largely the reverse of those produced by partial defoliation, and this is found to be true in part. Removing a portion of the pods resulted in much heavier root and stalk. The effect on yield of the reduction in the number of beans allowed to de- velop is offset to a considerable extent by a notable increase in the size of the beans- The increase in the size of the bean is not associated with a corresponding decrease in the percentage of oil; hence, the actual quantity of oil in the individual seed is considerably increased. Here, again, the early-maturing variety, S. P. I. No. 21755, stands out as an exception. Reducing the number of seed allowed to develop failed to increase the weight of the vegetative parts or the size of the seed and its oil content. Considering the two sets of experiments together, when the develop- ment of the seed extends through a relatively long period, a reduction of, say, 50 per cent in the normal proportion of leaves or photosynthetic organs leads to a decreased weight of the other vegetative parts, as well 'as of the total yield of seed, but the size of the seed is only slightly re- Dec. 15. 1914 Oil Content of Seeds and Nutrition 0} Plant 235 duced, and the quantity of oil in the individual seed is scarcely changed. A reduction in the normal proportion of the reproductive parts, on the other hand, leads to an increase in weight of the vegetative parts, and the size of the seed and its oil content are materially increased. These facts are not applicable to the variety which developed its seed within a short period of time. OIL CONTENT AS AFFECTED BY SIZE OF SEED In studying quantitative relationships of seeds one is at once confronted with the fact that in any particular lot of seed there is always considerable variation in the size of individuals, whatever the conditions under which the seed may be grown. This is true even of the seed from an individual plant. To ascertain whether there is any constant relationship between the oil content and the size of the seeds from the same plant, seeds of several varieties of soy beans grown in different localities were separated by hand into the larger and smaller sizes and their oil content was deter- mined. The results are shown in Table V. Table V. — Oil content of soy beans of large and small size from the same plant Variety and locality. Ogemaw (S. P. I. No. 17258): Pullman, Wash Do Amherst , Mass Do Wooster, Ohio Do Hansen (S. P. I. No. 20409): Pullman, Wash Do Statesville, N. C Do La Fayette, Ind Do Buckshot (S. P. I. No. 17251): Kingston, R. I Do Amherst, Mass Do size of beans. Large. Small . Large. Small . Large. Small. Large. Sm^l. Large . vSmall. Large . Small . Large. Small. Large . Small Weight of 1.000 beans. Gtti. 171 106 290 169 259 137 46 25 64 37 86 53 364 196 380 215 Moisture in beans. Per cent. 5.60 5-25 6. ^o 6.60 6. 40 6. 40 6. 10 5-85 7-35 6.80 6.75 6.75 6.30 6.65 6. 05 6. 40 Oil in moist beans. Per cent. 13. 88 14. 17 16. 92 16. 63 16.30 15-95 11. 72 12. 27 13-27 13-50 II. 72 II. 15 17-45 17. 00 17.87 16.95 Considering only the large and small beans from the same lot of seed, it appears that generally the percentage of oil is approximately the same, although there are some cases in which there are considerable differences. No fixed rule can be laid down as to the relative percentages of oil in large and small beans. It is evident that there are marked differences in the size of the beans in each lot, but a quantitative separation is impracticable, since there arc all gradations in size. The only practicable method of comparing different lots of seed, therefore, is to secure average values based on comparatively large quantities, simply counting out the seed as they come, without any attempt at separation into sizes. 236 Journal of Agricultural Research Vol. III. No. 3 OIL CONTENT AS AFFECTED BY LENGTH OF GROWING PERIOD Independently of any differences between early and late varieties as such, it might be expected that difference in oil content would be influ- enced by both the character and the length of the growing period. Sev- eral investigators have concluded that the length of the growing period is an important factor in determining the starch and protein content of wheat. In plants of determinate growth, in which all the seed are developed during approximately the same period, the effect of the length of the growing period on the oil content may be studied b}- making successive plantings at given intervals, since many of such species show a marked tendency to shorten the growing period when planted abnor- mally late. Mooers (1909) has called attention to this tendency in soy beans, and the present authors have determined the oil content in the seed of several varieties of this crop planted at stated inter\'als during the spring and summer months.' The results of the tests are shown in part in Table VI. Table \'l. — Oil content of soy beans planted at intervals of t-uo -ueeks in IQII Variety and date of plaut- ing. S. P. I. No. 21755: May I May 15 June I June 15 July I ■ Haberlandt (S. P. I. No. 17271); May I May 15 June I June 15 July I Buckshot (S. P. I. No. 17251); May I May 15 June I June 15 July I Medium Yellow (S. P. I. No. 17269): May I May 15 June I June 15 July I Numberof days from planting to bloom- ing. 40 35 37 32 27 61 54 50 45 44 40 40 39 39 56 49 50 45 41 Nmnberof days from blooming to full maturity. 62 58 50 54 58 83 76 74 74 72 69 68 58 64 102 94 77 .73 66 Nimiberof days from planting to full maturity. 93 8- 86 144 130 124 119 116 112 109 107 97 97 158 143 127 118 107 Weight of 1. 000 beans. 186 153 166 145 168 2IO 211 240 244 271 272 277 250 294 329 313 312 30s 334 ^foisture in beans. Per cent. 6. 90 7-05 6. 90 7. 10 . 7. 60 6.65 6. 70 6.35 6.80 7. 20 6. 40 6. 40 6. 20 6. 10 6. 40 6-45 6.25 5-70 5.80 6.30 Oil in beans. Per cent. 17-25 16.38 17. 40 17.27 15- 51 19- 73 19.47 20. 32 18.95 17-35 19- 15 19. 62 19-35 18.95 18.43 17. 60 18. 25 17-75 17-38 i6-43 oil in 1. 000 beans. Gni. 32-4 25-1 28.9 25.0 26. I 41.4 41. 1 44. I 45-5 42-3 51-9 53-4 53-6 47-4 54- 2 57-9 57- I 55-4 53-0 54-9 * In these experiments the authors are indebted to the Office of Forage-Crop Investigations, of the Bureau of Plant Industry, for samples of the soy beans which were grown at the Arlington Experiment Farm. Va., together with the data as to time of plaating and date of maturity . Dec. 15. 1914 Oil Content of Seeds and Nutrition of Plant 237 The different plantings of soy beans show marked variations in the size of the beans and in their oil content, but there is no definite relationship between these characters and the date of planting. In other words, the character rather than the length of the season in which the seed is devel- oped seems to be the important factor. The seed from the latest plant- ings contain a somewhat lower percentage of oil than the others, but this relationship is not verified in other tests which have been omitted from Table VI for the sake of brevity. These additional data all show lack of definite relationship between size and oil content of the seed and the length of the growing period. There is a remarkable difference between varieties as to the shortening of the period required for maturing seed when planted late, as illustrated by the Buckshot and the Medium Yellow varieties. VARIETAL DIFFERENCES IN OIL CONTENT OF THE SEED As a preliminary to the study of oil content as affected by nutrition, it was necessary to ascertain the relation of heredity to the quantity of oil produced in the seed in so far as relates to the comparative behavior of different varieties. This work has been limited largely to soy beans and cotton, since collections of varieties of the other oil-producing plants inves- tigated have not been available. The method of procedure has been to grow a number of varieties of each species under uniform conditions, to use the purest seed obtainable, and to repeat the tests for several seasons, using the precautions which have been discussed above in drawing samples for analysis. In Table VII are given the results obtained from material furnished by the Office of Forage-Crop Investigations, representing seven varieties of soy beans grown at the Arlington Experiment Farm in 1907, 1908, and 1910. Table VII. — Varietal differences in the oil content of soy beans grown at Arlington Experiment Farm, Va., in igoj, igo8, and igio. Variety and year grown. Shanghai (S. P. 1907 1908 , igio I. No. 14952): Average. Eda(S. P. I. No. 17257): '9°7 igo8. ,. ; 1910 Average. Weight of 1,000 beans. Cm- 215-4 186. I 217. I 206. 2 280.6 263.4 269.4 Moisture in beans. Per cent' 6.65 6. 16 6.80 6-54 6.30 6. 17 6- 15 6. 21 Oil in moist beans. Per cent. 19- 55 18.37 20. 20 19-37 19.70 19.90 21-55 20.38 Oil in 1,000 beans. Gm. 42. 1 34-2 43-8 40. o 53- 5 52-3 ;8. 2 54-7 238 Journal of Agricultural Research Vol. lU. No. s Table VII. — Varietal differences in the oil content of soy beans grown at Arlington Experiment Farm, Va., in igoj, igo8, and jp/o— Continued Variety and year grown. Weight ot 1,000 beans. Moisture in beans. Oil in moist beans. Oil in 1. 000 beans. Yosho(S. P. I. No. 17262); Gm. 182.2 170.2 202.8 Per cent. 6. go 6.35 6.65 Per cent. 16.78 16. 25 15-44 Gm. 30.6 27.7 31- 3 Average 185.0 6.63 16.16 29. 9 Amherst (S. P. I. No. 17275); 193.6 179-3 176. 0 6.65 5-5° 6. 05 20. 15 19- 5° 20-33 39- I 34-9 35-7 182.9 6. 07 20. 00 36.6 S. p. I. No. 19981: TQ07 335-8 3"-4 331-9 6. 60 4.85 6.45 20. 67 20. 20 21.86 69.6 1908 62.8 IQIO 72.8 Average . . . 326.0 5-97 20. 91 68.4 56.9 63.4 64.6 6.80 6. 18 6. 40 15.67 15-55 15-45 8.9 9- 9 10. 0 Average 61.6 6.46 15-56 9.6 S. P. I. No. 22312: 224. 2 185-3 214. I 7-3° 4.93 6.65 17.72 16. 00 18.90 39- 5 igoS 29. 6 40.4 Average 207.9 6. 29 17- 54 ^6. 5 In Table VIII are shown the results from six varieties of Upland cot- ton, also a foreign species (Hawasaki) grown in the Piedmont section of Georgia and the Coastal Plain region of South Carolina. The data pre- sented are the averages for the years 1909, 1910, and 191 1 in these two sections, and in each case all plantings were from the same original lot of seed. Dec. IS. 1914 Oil Content of Seeds and Nutrition of Plant 239 Table VIII. — Varietal differences in the oil content of cotton seed grown in northern Georgia and in the Coastal Plain region of South Carolina [Average for three years) V'priety of cotton. Cotton seed grown in — Northern Georgia. Lint. Weight of 1,000 seeds. Hvlls, Mois- ture in ker- nels. -- Oil in moist ker- nels. Oil 1. 000 seeds. Coastal Plain region of South Carolina. Lint. Weight of 1,000 seeds. Mois- ture in ker- nels. Oil in moist ker- nels. OU 1,000 seeds. King Russell Shine Toole Dixie Hawkins Average Hawasaki Per cent. 36.8 31-5 32.7 36.8 32.9 35- o Gm. 103.8 154.5 108.3 93- I III.o , = the density at any temperature /, Dx = the density at some standard temperature x, n-, ,i, and ;- = constants to be determined for each sample investigated. In practice it was found that for certain temperature ranges the expansion was represented within the limits of experimental error, by the above equation, with all terms above the second power omitted. By means of a "least squares" method the observations on each sample given in Table I have been reduced and the calculated values of Dx, «, and 3 are given in Table IV. Observ'ations and calculations of density for an average sample of cream are given in Table III. It will be seen from the closeness of the agreement between the calculated and the observed values of the density at various temperatures that the assumed equation comes very near expressing the actual rate of expansion of the diJBferent samples at the time the density determinations were made. T.\BLE I. — Observed densities of milk and cream 1913- July 12 . July II. Fat con- tent." Per cent. 0. 025 .025 .025 ■025 .023 I 025 ■ 2-5 2-5 2-5 2-5 2-5 2-5 1 2.5 Temper- ature. "C. 20 30 40 5° 10 20 30 40 5° Density. G.lc. c. I. 0381 I. 0368 I- 0356 I. 0322 I. 0284 I. 0236 I. 0348 I- 0343 I. 0348 I. 0317 I. 0292 I. 0258 I. 0212 1913- Junes. June: Nov. 19. Fat con- Temper- tent. ature. Per cent. °C. 3-5 0 3-5 10 3-5 20 I 3-5 30 f 35 20 3-5 30 I 3-5 40 [ 3-5 20 3-5 30 3-5 40 I 35 5° Density. G.jc. c, I. 0363 I. 0348 I. 0320 I. 0284 1.0311 I. 0290 1.0237 I. 0314 I. 0279 I. 0238 I. 0191 o The percentages of (at here given are as reported by the Bureau of .A.aimal Industry at the time the sampU^s were preparetJ , Dec. IS, 1914 Expansion of Milk and Cream 255 Tabi£ I. — Observed densities of milk and cream — Continued Date. Fat con- tent. Temper- ature. Density. Date._ Fat con- tent. Temper- ature. Density. I9I3- Per cent. °C. G.lc. c. 1913- Per cent. °C. G./c. c. 5 0 1.0344 2S 10 I. 0136 5 S I- 0357 25 20 I. 0070 5 10 I- 03S3 June 27 25 30 I. 0007 July 10 5 20 1. 0322 25 40 ■9944 5 30 I. 0299 I 2S 50 .9890 5 40 I. 0217 I 5 SO I. 0I7I f 2S 20 I. 0107 5 S s 20 1. 0268 Nov. 20 25 30 I. 0028 Aug. 26 30 40 I- 023s I. 0192 25 I 2S 40 SO •99155 .9911 5 SO I. 0144 30 10 I. 0108 7-5 0 1. 0327 30 20 •9997 7-S 5 1. 0316 June 26 30 30 .9940 7-S 10 1. 0316 30 40 .9884 J"iy9 7-5 7-5 20 30 1. 0274 I. 0246 I 30 50 .9827 7-S 40 I. 0209 30 20 .9966 Nov. 19 I 7-5 7-5 7-S 7- S 50 20 30 40 1. 0159 I. 0261 I. 0226 I. 0184 Aug. 23 30 1 30 I 30 30 40 so .9918 .9864 .9813 7.S S° 1. 0136 30 20 .9978 Nov. 20 30 30 •9933 10 0 I. 0304 30 40 .9886 10 5 1. 0310 I 30 SO • 983-' 10 10 1. 029s Julys 10 20 I. 0242 30 20 I. 0044 10 30 1. 0197 Nov. 21 30 30 .9960 10 40 I. 0152 30 40 .9896 10 SO I. 0106 I 30 SO .9842 IS 0 I. 0256 IS 10 I. 0216 35 0 I. 0129 July 2 IS IS 20 30 I. 0170 I. 0II4 June 26 35 35 10 20 I. OOS7 •9970 15 40 I. 0061 35 30 .9879 15 SO I. 0028 35 I 35 40 SO .9807 •9750 Aug. 25 15 20 I. 0I6I f 40 0 1.0088 Aug. 26 IS IS 30 40 I. 0II2 I. 0061 June 3 40 40 10 20 I. 0037 • 9931 I IS SO I. OOII I 40 30 .9838 f 20 0 I. 0242 June 3 { 20 10 I. 0208 f 40 20 .9940 I 20 20 I. 0137 June? \ 40 I 40 30 40 ■9845 -9763 [ 20 20 I. 0141 June 7 20 30 I. 0081 1 20 40 I. 0018 40 0 I. 0050 June 25 40 5 I. 0036 20 0 I. 0214 40 10 I. 0037 20 10 I. 0169 40 20 •9931 July 3 20 20 I. oio6 20 30 1.0052 40 30 •9837 20 40 •9995 Aug. 26 40 40 •9765 20 50 ■99SO 40 SO .9710 256 Journal of Agricultural Research Vol. Ill, No. 3 CALCULATION OF RELATIVE VOLUMES After having determined the density of the several samples of milk and cream at the various temperatures, the observations for each sample were reduced by the method of least squares, as already stated, and the value of £>35 , ,.= Z>j.. Sp. gr. '•035 1^034 1-033 1.032 I. 031 I. 029 I. 028 I. 027 I. 026 I. 024 1.023 I. 022 I. 020 I. 019 I. 017 I. 016 1-015 I. 014 I. 013 I. 012 I. on I. 010 I. 009 I. 008 . 007 . 006 I. 004 1.003 I. 002 I. 001 I. 000 •999 .998 .998 ■997 .996 •995 •994 •993 Sp. QT. 1.030 I. 029 I. 028 I. 027 1.025 I. 024 I. 022 I. 021 I. 020 I. oi8 I. 016 I. 015 I. 013 I. 012 I. 010 I. 009 1. 007 I. 006 1.005 I. 004 1.003 I. 001 I. 000 •999 .998 •997 .996 •994 •993 .992 .991 .990 .989 .988 .986 •985 .984 ■983 .982 .981 .980 -o. 00040 . 00040 . 00040 . 00041 . 00041 . 00041 . 00042 . 00042 • 00043 . 00044 . 00045 . 00046 . 00047 . 00048 . 00048 . 00049 . 00050 . 00051 . 00052 . 00052 • 00053 - 00054 . 00055 . 00057 . 00058 . 00050 . 00061 . 00062 . 00063 . 00064 . 00065 . 00066 . 00067 . 00068 . 00069 . 00071 . 00072 .00073 . 00074 • 0007 s . 00076 — o. 000005 . 000004 . 000004 . 000004 . 000003 . 000003 . 000003 . 000002 . 000002 . 00000 1 . 00000 1 . 000000 . 000000 . 000000 +. 00000 1 . 00000 1 . 00000 1 . 000002 . 000002 . 000002 . 000003 . 000003 . 000003 . 000004 . 000004 . 000004 . 000005 . 000005 . 000005 . 000005 . 000006 . 000006 . 000007 . 000007 . 000008 . 000008 . 000008 . 000009 . 000009 . 000009 . OOOOIO 258 Journal oj Agricultural Research Vol. Ill, No. 3 p-sSi-S m" •s 3 It n t- E £■-- 2 '<5 a w CQ . M O oo o o o •r ro f^ + o o o TtCO ►-• ir> rf »o SO O O 1 I L) . . . . » o o o o " f^ «^ fo ^ ^ r* XT-, CO 00 (> ON 5 d d o d d OO 8 ? dd M « ^ 00 00 do 6^<> ro CO ro ro 1 1 0 r2 too 0 ^o t^ (J •* d d d odd . O CO O CO (^ d c> cS _j_ ro <0 1. 00 6 ; ? 3 u, 4> « ?ai Boo .a a " w b ^ sl5 Dec. 15. 1914 Expansion of Milk and Cream 259 Table IV. — Observed and calculated densities of milk and cream at different temperatures and -with different percentages of fat ° Fat content. Ehi Per cent, 0-025 ■ 2. 5 . . . 3-5 • 7-5 •• 7-S IS IS I. 03039 I. 0276 I. 0259 — o. 00040 - .00035 — . 00041 I. 0259 - . 00053 I. 0214 I. 0229 I. 0206 . 00042 . 00038 . 00042 I. 0174 — . 00045 I. 0086 .0086 I. 0023 — . 00048 . 00050 , 00052 — o. 000004 - . 000005 . 000003 . 000006 . 000004 — . 000005 . 000003 4- . 000006 + ■ 000002 Tempera- ture. 20 30 40 50 20 30 40 50 20 30 40 50 20 30 40 50 20 30 40 SO 20 30 40 SO 20 30 40 20 30 40 50 20 30 40 50 20 30. 40 SO 30 40 50 Dt. Observed. Calcu- lated. Observed minus calcu- lated. I- 0356 1. 0322 I. 0284 1. 0236 '■ 0356 I. 0322 I. 0284 I. 0236 1. 0317 I. 0292 I. 0258 I. 0212 I. 0317 I. 0292 I. 0258 I. 0212 1. 0314 1. 0279 I. 0238 I. 0I9I 1.0314 I. 0279 1.0238 I. 0191 I. 0322 1. 0299 I. 0217 I. 0I7I I. 0327 I. 0285 I. 0231 I. 0166 I. 0268 I. 0268 I. 0192 1. 0144 I. 0234 I. 0193 I. 0144 1. 0274 I. 0246 I. 0209 1. 0159 I. 0274 I. 0247 I. 0208 I. 0159 I. 0261 I. 0226 I. 0184 1. 0136 I. 0261 I. 0226 I. 0184 I. 0136 r. 0242 I. 0197 I. 0152 I. 0106 I. 0242 I. 0197 I. 0152 I. 0106 I. 0170 I. 0171 I. 0114 I. 0061 I. 0028 I. 01 1 1 1.0063 I. 0027 I. 0161 I. oi6r I. OtI2 I. 0061 I. OIII I. 0061 I. 001 1 I. OOII I. 0106 I. 0052 •9995 •9950 1. 0107 I. 0050 •9997 ■9949 +14 -14 + 5 o + I — I o o — I + t o o o o o — I + 3 — 2 + I o + I o — t + 3 — 2 + I °- The temperatures from which the reductions arc made are the same for all samples — namely, 30°. jo', 40*, and 50** C. — and for that reason Ci. Cs, etc., will be the same In all cases. ^ Tlic lack of agreement between the observed and the calculated values of density indicates that one or more of the observed values are considerably in error. All determinations on this sample should be dis- carded. 26o Journal o} Agricultural Research Vol. in, No. 3 Table IV. — Observed and calculated densities of milk and cream at different temperatures and with different percentages of fat — Continued Fat content. D«,° a a Tempera- ^ ture. Dt. Observed minus Observed. Calcu- lated. calcu- lated. — o. 00060 r 20 +0. 000002 -^ 1 40 jl 5° I. 0070 I. 0007 0.9944 0. 9890 I. 0070 I. 0006 0. 9945 0. 9890 c + 1 — I 0 f -^0 I. 0107 I. 0107 0 — . 00065 + . OOOOOO 30 40 I. 0028 0. 9965 I. 0029 0. 9964 — I + 1 I 50 0. 991 1 0. 9911 0 20 0. 9997 0. 9997 0 - -00057 0 30 40 50 0. 9940 0. 9884 0. 9827 0. 9940 0. 9884 0. 9827 0 0 0 20 0 . 9966 0.9966 0 — . 00051 — . OOOOOI 30 40 50 0. 9918 0. 9854 0. 9813 0. 9917 0. 9865 0. 9812 +1 — I +1 20 0. 9978 0. 9978 0 — . 00049 — . 000002 30 40 so o- 9933 0. 9886 0. 9832 0. 9934 0. 9885 0. 9832 — I +1 0 — . 00067 0 20 30 40 I 5° I. 0044 0. 9960 0. 9896 0. 9842 I. 0044 0. 9962 0. 9894 0. 9842 0 — 2 + 2 0 - .00073 + . 000008 f 20 30 1 40 0. 9970 0. 9879 0. 9807 0. 9970 0. 9980 0. 9806 0 — I + 1 I 50 0. 9750 0. 9750 0 - . 00074 + . OOOOIO 20 30 1 40 o- 9931 o- 9837 0. 9765 0-9931 0. 9838 0. 9764 0 — I + 1 i 5° 0. 9710 0. 9710 0 Per cent. 25 25- 30- 30. 30- o- 9975 -9995 .9912 30- 35- 40. .9891 .9910 . 9926 .9841 -9799 These results were then plotted on coordinate paper of such a size that densities could be plotted and read to one in the fourth place, and « and /? to one in the fifth and sixth places, respectivelj'. From these curv^es the values of n and /9 for various densities of milk and cream are tabulated in Table II. The curves are shown on a reduced scale in figure i. In figure 2 is shown the relation between the density of the samples and the percentage of butter fat contained in them. Having determined the density of each sample and the rate of change of density with change of temperature, it was possible to calculate the volume of any sample at any temperature in terms of the volume at any other temperature within the limits covered. It was suggested by the Dairy Division of the Bureau of Animal Industry that 20° C. (68° F.) be chosen as the standard temperature and Dec. 15. 1914 Expansion of Milk and Cream 261 that the relative volumes at other temperatures be given on the basis of unit volume at this temperature. It was also suggested that for the -/O TO eo so ■fO JO To^ 0 c -tt'^*'''o^ -jj — «■ /9x/0^ 0 0 ^X'' D 0 ^ ^ Q^-^ ^''^^ 0 0 -a--^ J8^ 5 9930 9940 9950 9960 9970 •9975 9985 9990 1.0000 I. 0010 1.0015 36 9925 9930 9940 9950 9960 9970 ■9975 9985 9990 1. 0000 I. 0010 1. 0015 37 9925 9930 9940 9950 9960 9970 •9975 998s 9990 1.0000 1. 0010 l.oots 38 99IS 9925 9935 9945 9955 9965 •9975 9980 9990 1. 0000 I. 0010 I.OOIS 39 9915 992s 9935 9945 9955 9965 •9975 9980 9990 t.oooo I. 0010 1.0015 30 9915 9925 9935 9945 9955 9965 .9970 9980 9990 1.0000 1. 0010 1.0030 31 9915 9925 9935 9945 9955 9965 .9970 9980 9990 1. 0000 I. 0010 I.0030 33 9910 9920 9930 9940 9950 9960 .9970 9980 9990 1. 0000 1. 0010 1. 0030 33 99'° 9920 9930 9940 995° 9960 .9970 9980 9990 1. 0000 1.0010 I.0030 34 9910 9915 9925 9940 9950 9960 •9970 9980 9990 1.0000 l.OOIO I.0030 35 9900 9915 9925 9940 9940 9960 .9970 9980 9990 1.0000 l.OOIO 1. 0030 36 9900 9910 9920 9930 9940 9955 .9965 9980 9990 1.0000 l.OOIO i.ooas 37 9890 9910 9920 9930 9940 9955 .9965 9980 9990 1. 0000 l.OOIO 1.003S 38 9890 9910 9920 9930 9940 9955 •9965 9980 9990 1.0000 l.OOIO I.003S 39 9890 9900 991S 9925 9940 9955 •9965 9975 9990 1.0000 l.OOIO I.0035 40 9890 9900 99IS 9925 .9940 9950 •9960 997S 9990 1. 0000 t-OOIO I.OOIS t^The tabulated values arc given to the nearest o.ooo;. 67235°— 14 6 266 Journal of Agricultural Research Vol. Ill, No. 3 Table V. — Volume of milk and cream at various temperatures occupied by unit -volume at 6S° F. {20° C.)— Continued Per- Temperature > = F.). ceiit- T but- 74 76 7S I So 1 82 84 86 88 90 92 94 96 ter fat. \'oIume. c. 025 1. 0005 I 0010 I 0010 i-oois I 0020 I i 0025 I 0030 1.0030 1 0035 I 0040 I 0045 I.OOSO I 1.0005 I ooto I 0010 1.0015 I 0020 I 0025 I 0030 1.0030 1 0035 I 0040 1 004s 1-0050 2 1. 0010 I 0010 I oois 1.0020 r 0020 I 0025 I 0030 1-0035 I 0040 t 0040 I 004 s 1-0050 3 1. 0010 I 0010 I 0015 i.ooao I 0020 I 0025 I 0030 1.003s 1 0040 I 0045 I 0045 I- ooss A 1. 0010 I 0010 I 0015 1.0020 I 0020 I 0035 I 0030 1.0035' I 0040 I 0045 I 0050 1-0055 5 l.OOIO I 0015 1 C020 1.0020 I 0025 J 0030 1 003 s 1.0035 I 0045 I 004s 1 0050 I -ooss 6 1. 0010 I 0015 I 0020 1.0020 I 0025 I 0030 I 0035 1 . 0040 I 0045 I 0050 I 0050 1.0060 7 I. GOTO I 0015 I 0020 1. 002s I 0025 I 0030 I 003 s 1.0040 I 004S I 0050 I 00s 5 1-0060 S 1. 0010 I 0015 I 00 30 1. 002s 1 0030 I 0030 1 003 s 1.0040 I 004s I 0050 1 0055 1.0060 9 1. 0010 I oors 1 0020 I. 0025 1 0030 I 0035 I 0040 1.0045 I 0050 1 005s 1 0060 1.0065 10 I. 0015 I 0020 I 0025 I- 002s I 0030 I 0035 I 0040 I. 0045 1 0050 I ooss I 0060 1.0065 II i.oois I 0020 I 0025 1. 002s 1 0030 1 0035 I 0040 1. 004s I ooss 1 ooss 1 0065 1.0070 12 i.oois I 0020 I 0025 1. 0030 I 0030 I 0035 I 0040 1. 0050 I ooss I 0060 1 0065 1.0070 13 1. 0015 1 0020 I 0025 1.0030 1 0035 I 0040 I 0045 1.0050 I ooss I 0060 I 0065 1.0070 14 I.OOIS 1 Q020 1 0025 1. 0030 I 003s I 0040 I 0045 1-0050 1 005s 1 0065 I 0070 I- 0075 15 I.OOIS I 0025 I 0030 1.0030 1 0035 I 004c I 004s 1.0055 I 0060 I 0065 I 0070 1. 0075 16 I.OOIS I 002s I 0030 I- 0035 1 0040 I 004s I 0050 1.0055 I 0060 I 0070 1 0075 1.0080 17 1. 0015 I 002s I 0030 I- 003s I 0040 I 0045 I 0050 1. 0060 I 0060 I 0070 I 0075 1.0080 18 1.0020 I 0025 I 0030 1.003s 1 0040 I 0045 I OOS5 1. 0060 I 0065 I 0075 I 0080 1-0085 19 1.0020 I 0025 1 0030 I- 003s I 004s I 0045 I ooss 1.0060 I 0065 I 007s I 0080 1.008s JO 1.0020 I 0025 I 0030 1.003s I 0045 I 0050 I ooss 1.0060 1 0070 I 007s I 0085 1.0090 21 1. 0020 I 0025 I 0030 1. 0040 1 0045 I 0050 I 0060 1.0065 1 0070 I 0080 I 0085 1-0090 22 1.0020 1 0030 I 0035 1.0040 1 0050 I 0055 I 0060 1.0065 1 0075 I 0080 I 0090 1.0095 23 1.0020 I 0030 I 003 s 1.0040 I 0050 1 ooss I 0065 1. 0070 I 007s I 0085 I 0090 1-009S 24 1.0020 1 0030 I 0035 1.0040 I 0050 I 0060 I 0065 1.0070 1 0080 1 0085 I 0095 I-OIOO 2S 1.0020 I 0030 1 003s 1.0045 I 0055 I 0060 I 0070 1.007s I 0080 I 0090 I 0095 1.0105 26 1. 002s I 0030 I 0040 I. 004s I ooss I 0060 I 0070 i.ooSo I OC85 I 0090 I 0100 I. 0110 27 1. 002s I 0030 I 0040 1.0045 I ooss I 0060 I 0070 1.0080 I 00S5 1 0095 I 0100 I.OZIO 28 1. 0025 I 0030 I 0040 1.0045 I ooss I 0065 I 0075 i.ooSo I 0090 1 009S I 01 OS 1.0II5 ^9 1.0025 I 0030 I 0040 1.0050 1 0060 1 0065 I 007S I. ooRo I 0090 I 0095 1 0105 I.0II5 30 1. 002s I 003s I 0045 1.0050 1 0060 I 0065 I ooSo 1 . 0085 1 009s I 0100 1 Olio 1.0120 31 1.0025 I 003s I 0045 1 . 0050 1 0060 I 0065 I 0080 1.00S5 I O09S 1 0100 I 0110 i-oiao 32 1.0030 I 0035 1 0045 1.005s I 0065 I 0070 I 0085 1.0090 I 0100 I 0105 1 0115 1. 0125 33 1.0030 I 0035 1 0045 I. ooss I 0065 I 0070 I C0S5 1.0090 I 0100 I 0105 I 0115 1-0125 34 1.0030 I 0040 I 0050 I- 005s I 0065 I 00 75 1 0085 1.0095 1 0105 I Olio I 0120 1. 0130 35 1. 0030 I 0040 J 0050 1.0060 I 0070 I 007s I 0090 I. 0095 I 0x05 r Olio I 0120 1. 0130 36 I. 0035 I 004s I 005S J. 0060 I 0070 I 0080 I 0090 I.OIOO 1 Olio t OTIS I 0125 1-OI3S 37 1.003s I 0045 I 005s 1.0060 I 0070 I ooSo I 0095 l.OlOO I 0110 I 0II5 I 0135 1.013s 3S 1.0035 I 004s I 005s 1.0065 1 0075 I 0085 I 009s 1.0100 I OTIS I 0120 I 0130 1.0140 39 1-003S 1 0045 I ooss 1.0065 1 007s 1 008s I 0095 1. 0105 I ciis I 0120 I 0130 1- 0140 40 1.003s 1 0045 1 0055 1.0065' I 007s 1 0085 1 0095 1. 0105 1 oils 1 0125 I 0130 1. OI4S Dec IS, 1914 Expansion of Milk atid Cream 267 Table V. — Volume of 7nilk and cream at various temperatures occupied by unit volume at 6S° F. (20° C.)— Continued Temperature 1 "F.). Percentage of butter fat. 98 100 102 104 106 loS 110 112 114 116 118 \'oliimc. 0.02s I.OOS5 1.0060 1.0065 1.0070 1.0075 1.0080 i.ooSs 1.0090 I- 009s I- 0100 1. 0105 I 1. 0055 1.0060 1.0065 1.0070 1. 0075 1.0080 i.ooSs 1.0090 I.0O9S l.OIOO I.OIOS 2 1.005s 1.0060 I. 006s 1.0070 1.0075 1.0080 l-ooSs 1.0090 1-0095 l.OIOO 1.0110 3 r.oo6o 1.0065 1.0065 1.0070 1-007S I.ooSo 1.008s 1.0090 1.0095 I- 0100 I.OIIO 4 J. 0060 1.0065 1.0065 1.0070 1.007S 1.0080 1.0085 1-0090 1.009s l.OIOO I.OIIO 5 1.0060 1.0065 1.0070 1.0075 1.0080 1.0085 I.ooSs 1.0090 I. 009s l.OIOO I.OIIO 6 1.0060 1.0065 1.0070 1.0075 I.ooSo 1.0085 1.0090 1.0090 1.0095 l.OIOO I-OIIO 7 1.0065 1.0070 1.0075 1.0075 i-ooSo 1.0085 1.0090 1.0095 l.OIOO 1.0105 1-0115 8 1.0065 1.0070 1.0075 I.ooSo 1.00S5 1.0090 1. 009s l.OIOO 1.0105 I.OIIO I.OIIS 9 I. 006s 1.0070 i.ooSo 1.0080 1.0085 1.0090 1.0095 l.OIOO I.OIOS I-OIIO I.OIIS 10 1.0070 1-007S i.ooSo 1.00S5 1.0090 1.0090 1.0095 l.OIOO 1-0105 I.OIIO I.OIIS IZ 1.0070 I.007S 1.0080 1.008s 1.0090 I.009S 1-0095 l.OIOO 1. 0105 I-OIIO I.OIIS J2 1.007s 1.0080 1.00S5 1.0090 I. 009s 1.0095 I.OIOS I.OIIO I.OIIS 1. 0120 I- 0125 ^3 1. 007s 1.0080 1.0085 1.0090 I. 0095 l.OIOO I.OIOS I.OIIO I.OIIS I.0120 1.0125 14 i.ooSo 1.00S5 1.0090 I.009S l.OIOO l.OIOO i.ono I-0II5 1-0120 1.0125 1. 0130 IS 1.008c i.ooSs 1.0090 I. 0095 l.OIOO I.OIOS I.OIIO I.0II5 1. 0120 1-0125 1.0130 16 1.0085 1.0090 1.0095 l.OIOO 1. 0105 I.OIIO I. oils 1.0120 1. 0125 I- 0130 I.OI3S 17 1.008s 1.0090 1.009s 1. 0105 1. 0105 i.oiis 1. 0120 1. 0125 1. 0130 1. 013s 1.0140 18 1.0090 1.0095 l.OIOO 1. 0105 I. Olio I. 0120 I.012S 1.0130 I- 013s 1-0140 1.0145 19 1.0090 1.009s l.OIOO I. Olio 1. 0115 I. 0120 1. 0125 1. 0130 1.0135 1. 0140 I.OI4S 20 1.C095 l.OIOO I.OI05 I. Olio I. oils 1.0125 1.0130 1.013s 1.0140 1.014s 1.0150 21 1. 009s 1. 0100 1.0105 I.0II5 1. 0120 1.0I2S 1. 0130 1.0135 1-OI4S 1-0150 I- 0155 22 1. 0100 1. 0105 I-OIIO 1. 0120 1.0125 1. 0130 I-OI35 1. 0140 I- 0150 I- 0155 I- 0160 23 i.oios i-oios I.0II5 1. 0120 I- 0125 1.0130 1. 0140 1.0145 I. 0150 1. 015s 1. 0160 24 i.oros l.OZIO 1.0120 I.OIZ5 1.0130 1.013s 1.014s 1.0150 I. 01 55 1.0160 1.0165 25 I-OIIO I-OllS 1. 0120 1.0130 I. OIJS 1. 0140 1.0145 1.0150 1.0160 1.0165 I. 0170 26 i.oiis 1.0I20 I.OI2S 1.013s 1. 0140 I.OI4S i-oisS 1.0160 1.0165 1. 0170 1.0180 27 I. oils I. 0120 1.0130 I. 0135 1. 0140 1.0150 1.015s 1. 0160 1. 0170 1. 0170 1.0180 28 1. 0120 I. 0125 1.0130 I. 0140 I. 0145 1.0150 I. 0160 I. 0165 I. 0175 I.OI7S i.oiSs 29 1.0120 1.0130 I- 0135 1. 0140 1.0150 l.oiss I. 0160 I. 0165 1. 0175 1. 0180 t-0185 30 1.0125 1. 0130 I.OI3S I- 014s 1. 0155 I'OISS 1.0165 1.0170 I. 0175 i-oiSo 1.0190 31 I'OI2S 1.0I3S 1. 0140 1.014s I- 0155 1. 0160 I. 0170 1.0175 1.0180 1.0185 1. 0190 32 1. 0130 I- 0135 1.0140 1.0150 1. 0160 1. 0165 1. 01 70 1. 0180 1-0185 1. 0190 1.0I95 33 1. 0130 1. 0140 1.0145 I* 0155 1.0160 1-0165 1.0170 1. 0180 I. 018s I- 0190 I.OI9S 34 1. 0135 I. 0140 1.0150 1. 0160 1-0165 I. 0170 I- 0175 1.0185 1.0190 I.OI9S 1.0200 35 1-0I3S 1.0145 1. 0150 I. 0160 1.0165 1.0170 i.oiSo 1.0190 I. 0195 1.0200 I.030S 36 1.0140 I-OI4S I- 0155 I. 016s 1.0170 1.017s 1.0185 1.0I95 1.0200 1.0205 I. 0210 37 I.0M5 1.0150 I. 0160 I. 0165 1.017s I. 0180 1. 0185 1.0195 I.0300 1.0205 I. 0210 38 I. 0150 i-oiss 1. 0165 I. 0170 1.0175 1.0185 1.0190 1.0200 1.0210 1-0215 I. 0315 39 I. 0150 1. 0160 1.0165 I. 0170 1.0180 1.018s 1. 0195 1.0205 1.0210 r.02is 1.0230 40 I. DISS 1.0165 1. 0170 I.OI7S 1.0185 I. 0190 i.oaoo I- 0210 1.0215 1.0220 I. 0230 268 Journal of Agricultural Research Vol. III. No. 3 Table V — Volume of milk and cream at various temperatures occupied by unit volume at 68° F. (20° O— Continued Temperature (" F.). Percentage of butter fat. 120 122 124 126 128 130 132 134 136 138 140 Volume. 0.025 I- Olio I. 0120 1.0125 1.0130 1-OI35 I. 0140 1.0145 1-OI55 1.0160 1.0170 1.017s I I. Olio I. 0120 1.0125 I. 0130 I- 0135 I. 0140 1.0145 I.OISS 1.0160 I. 0170 1-OI7S z 1.0115 1.0120 I. 0125 1.0130 I- 0135 1.0140 1.0145 I -0155 1. 0160 I. 0170 1.0175 3 i-oiis I. 0120 1-0125 I. 0130 I-OI3S I. 0140 1.014s I.OIS5 1. 0160 1.0170 I.OI7S 4 :-oii5 I. 0120 I.OI2S 1. 0130 1-0135 1.0140 1-OI45 I-01S5 1.0160 1.0170 I.0I7S 5 I-OII5 1.0120 1-0125 1.0130 I- 0135 I- 0140 1.014S I-01S5 I. 0160 1 . 01 70 I.OI7S 6 I.OII5 1.0120 I- 0125 1.0130 I- 0135 1.0140 1-0145 I- 015s 1.0160 1. 0170 I.017S 7 I. 0120 1.0125 1.0130 i-ot35 1-0140 I. 0145 1.0150 1-0155 i.oi6o I. 0170 1. 01 75 S I. 0120 I. 0125 1.0130 1-OI35 1. 0140 1.0145 1.0150 I-OIS5 1.016s I. 0170 1-OI7S 9 1. 0120 I. 0130 I- 013s I. 0140 1.0145 1.0150 i'OIS5 1. 0160 i.oi6s 1.0170 1. 0180 10 1. 0120 1.0130 i-oioS I. 0140 1-0145 1.0150 I- 0155 1.0160 I. 0165 1.0170 1. 0180 11 I. 0120 I. 0130 I- 0135 1.0140 I- 0145 1.0150 I.OISS I. 0160 1.0165 1.0170 i.ozSo 12 1.0130 I- 013s 1.0140 1.014s 1-0150 I-OISS I. 0160 1.0165 I. 0170 1.0175 1. 0180 13 1. 0130 I-OI35 1-0140 1.0145 1-0150 I- 015s I- 0160 1.0165 1.0170 I- 0175 i.oi8o 14 I- 013s I. 0140 1.0145 I. 0150 I-OI5S 1.0160 1.0165 I. 0170 1.0I7S 1. 0180 1.0185 IS I- 013s I- 0140 1. 0145 1.0150 i.oiss 1.0160 1.0163 I. 0170 1-0175 i.oiSo 1-0185 16 1.0140 I- 0145 1. 0150 I- 0155 1.0160 1.0165 I. 0170 I- 0175 1.0180 1.0185 1.0190 17 1. 014s I- 0150 I- 0155 1. 0160 I. 0165 1.0170 I. 0175 1.017s I. 0180 1.0185 1.0190 18 1.0150 I- 0155 I- 0160 1.0165 1.0170 I- 0175 i.oiSo 1. 0185 1.01S5 i.oigo 1.0195 19 I. 0150 I. 015s 1. 0160 1.0165 1.0170 1.0175 i.oiSo 1.018s 1.0185 i-oigo 1.019s 30 i-oiss 1.0160 1.0165 1.0170 1.0175 1. 0180 1.0185 I. 0190 1-OI95 I.OZOO 1-0205 21 1.0160 i.oi6s 1. 0170 I-017S 1. 0180 1.0185 1. 0190 I. 0190 1. 019S I.0200 I . 020s 22 1.016s 1. 0170 I- 0175 1-0180 1.018s 1.D190 1.0190 1-0195 1.0200 1. 0205 1.0210 23 1.0165 1.0170 I- 017s 1. 0180 1.0185 1.0190 1. 0195 1.019s i.oaoo 1. 0205 I.0210 24 1.0170 1.0180 1.0185 I- 0190 1.0195 1.0200 1. 0205 1.0305 I.0210 1. 021s 1.0220 25 1.017s 1. 0180 i.oiSs I- 0190 i.oigs 1. 0200 1.0205 1. 0210 I. 0215 1-0220 1.0225 26 1.0185 1. 0190 1.019s I.0300 1.0205 I.0210 1.0215 1.0220 1-0225 1.0230 I- 023s 27 1.01S5 1.0190 1. 019s 1.0200 1.0205 1.0210 1-0215 1.0220 1.0235 1-0230 I- 023s 28 1.0190 z. 0200 1. 0205 I-03I0 1.0215 I. 0220 1.0225 1. 0230 I. 0235 1.0240 1.024s 29 1. 0195 1.0200 1.0205 I. 0210 1.021s I-0230 1.0325 1. 0230 I. 023s 1-0240 1.024s 30 1. 0195 1.0200 I. 0305 I. 0210 1.0215 1.0230 I.033S I- 023s 1.0240 I.034S 1.0250 31 1.0200 1. 0205 1. 0210 I.02IS 1.0220 1.022s 1.0230 1-0235 1. 0240 1.024s 1.0250 32 1.0305 I. 0210 I- 0215 I- 0220 1.0225 1.0230 I. 0235 1. 0240 1.024s 1.0250 I.02SS 33 1.0205 1. 0210 1.0215 1.0220 1.0225 1.0230 1.0235 1.0240 1. 024s 1.0250 I-02SS 34 I.02I0 1-0215 1.0220 1.0225 I. 0230 1.0240 1-024S 1.024s 1.0250 I- 0255 1.0260 35 I. 0210 1.0215 1.0220 1.022s 1.0230 1.0340 1.024s 1.0250 1.0250 1.02S5 1.0260 36 I-02IS 1.0225 1.0230 r-023S 1.0240 1.024s 1-0250 I- 0255 1.0260 1-0265 1-0370 37 1.0215 1. 0225 1.0230 1-0235 1.0240 I. 0245 1.0250 I- 0255 1.0260 1.0265 1.0270 33 1.0220 1.0230 1-0235 1.0340 1. 024s 1.0250 I-02SS 1.0260 1.0265 1.0270 1.0280 39 1.0225 I. 0235 1. 0340 1-024S 1.0250 I.02S5 1-0260 1.0265 1.0270 1.027s 1.0380 40 ^•0235 1.0340 1.0345 I-0-S5 1.0260 1.0365 1.0270 1.0275 1.0280 1.0285 1.0290 LIFE HISTORY OF THE MELON FLY By E. A. Back, Entomological assistant, and C. E. Pemberton. Scientific Assistant, Mediterranean Fruit-Fly Investigations, Bureau of Entomology INTRODUCTION Aside from the Mediterranean fruit fly, Ceratitis capitata Wied., there is no other insect in the Hawaiian Islands that is causing such financial loss to fruit and vegetable interests as the melon fly, Bactrocera cucurbitae Coq. The damages caused by its ravages are placed by some even higher than those caused by C. capitata. While B. cucurbitae was not officially recorded until November, 1898, when it was first discovered by Mr. George Compere in the market gardens in the environs of Honolulu, it had been knownlocally about that city many years before. Mr. Albert Waterhouse, Acting President of the Hawaiian Board of Agriculture and Forestry, states that less than 30 years ago excellent cantaloupes (Cucumis melo) and watermelons {Citrullus vulgaris) and many kinds of pumpkins and squashes were grown in profusion the year round. Since that time the spread of the melon fly has been so rapid that this insect is now found on all the important islands of the Hawaiian group, and cantaloupes and water- melons can not be grown except on new land distant from old gardens. More than 95 per cent of the pumpkin {Cu-curbita pepo) crop is annually ruined, not to mention the havoc caused among the more resistant cucumbers {Cucumis sativus). Not only does the adult melon fly oviposit in fruit that has already set, but more often — in the case of the pumpkin and the squash {Cucurbita spp.) — in the unopened male and female flowers, in the stem of the \ane, and even in the seedUng itself, especially in seedlings of the watennelon and the cantaloupe. The writers have observed entire fields of water- melons killed before the plants were 6 to 8 inches long by larvae boring into the taproot, stem, and leaf stalks. An examination of almost any pumpkin or squash field in the agricultural districts on the Island of Oahu at certain seasons of the year will show that very nearly all the flowers are affected before they have an opportunity to bloom. In about 95 cases out of 100 the anthers of the male bloom are either reduced to a mass of rot or more or less eaten before the bud becomes full grown, and the young ovaries of the female bloom are ruined by the burrowing mag- gots either before or shortly after the flower unfolds. While cucurbitaceous crops are the favored host fruits of the melon fly, certain varieties of leguminous crops, such as string beans and cow- peas, are often badly attacked. When preferred host fruits are scarce, even peaches, papayas, and similar fruits are attacked to a limited degree. No satisfactory remedy has yet been found to prevent the infestation of fruit. The Chinese gardeners save a small percentage of crops subject to the attacks of this pest by covering the young fruit with cloth or paper or, in the case of the curcurbits, by burying them in the soil until they become sufiiciently large to withstand attack. The female melon fly deposits her eggs in small batches just beneath the surface of the fruit, vegetable, or plant affected. From these eggs Journal of AcTicuItiiral Research. Vol. Ill, No. 3 Dept. of ApTicukiire, Wa^^hington. D. C. Dec 15 '1014 (269) 270 Journal of Agricultural Research Vol. HI, No. 3 maggots are hatched which feed and burrow about, causing the rapid destruction of the affected parts, and then leave the host to enter the soil, where they pupate. After a short time the adult emerges from the pupa and soon deposits eggs for the following generation of larvae. THE EGG From the data included in Table I it will be seen that the duration of the egg stage of the melon fly is very short. During the warm summer months, when the daily mean temperature is about 79° F., eggs hatch in 26 to 35 hours after deposition. The data indicate that hatching pro- ceeds most rapidly about 27 or 28 hours after the eggs are laid. At a mean temperature of 75.6°, 84 eggs hatched in from 31 to 38 hours, while at a mean of 75°, 96 eggs hatched in about 47 hours after deposition. At 73.6°, 88 eggs hatched about 52 to 54 hours after being laid. Table I. — Duration of the egg stage of the melon fly Number of eggs under ob- servation. Eggs deposited. Eggs hatched. Average mean Day. Period. Day. Period. tempera- ture for period. 17 75 45 12 Aug. 20 ...do ...do ...do do . . . 10.30 to 11.30 a. m. . . do do do do Aug. 2 1 ...do ...do ...do ...do.... °F. 79-5 . 79-5 79-5 79-5 79-5 79- S 79- S 79- S 79-5 79-5 79-5 79- 5 79-5 75-6 75- 0 75-5 75-5 73-6 73-6 2 to 2.15 p. m 2.30 to 2.45 p. lU. . . . 2.45 to3 p. m 3 t03.i5p. m 3.15 to3.3op. m... . 3.30 to 3.45 p. m.... 3.45 to 4p. m 4 to 4.30 p. m 4.30 to 6 p. m 6 to 8 p. m ...do do ...do 6 20 ...do ...do do do do do ...do ...do ...do ...do 9 ...do ...do 7 do .. do ...do ...do do ...do..... 8 to 9 p. m 2 do .do . . do . . 9 to 10 p. m 10 p. m. to 2 a. tn.. . 3 to 5.30 a. m About 9 p. m 10 p. m. to 3 a. ra.. . 84 96 13 62 May 19 May 14 May 13 do 3. 15103.30 p. m 4 to 6 p. m May 20 May 16 May 14 do 10 a. m. to I p. ni . . . .do 77 II May II ...do 11.30 a. m. to 2 p. m . do May 13 ...do 6 to 9.30 a. m THE LARVA The larva of the melon fly passes through three instars before being full grown. The data in Table II show that at a mean temperature of about 79° F. larvae can complete their development in from four days and four hours to seven days. The larvae recorded as feeding upon papaya {Carica papaya) were transferred several times a day from one small piece of pulp to a fresh piece ; hence, they probably pupated a few hours sooner than they would have pupated had they undergone their entire development in a single fruit. Larvae developing in thick-skinned fruits, such as water- melons and pumpkins, often remain in the fruit after becoming full grown several days longer before emerging to pupate than they would have done had they been less confined. During the cooler seasons of the year the length of the larval life will probably be found to be much longer. Dec. IS, 1914 Melon Fly 271 Table II.— Duration of the larval stage of the melon fly Number of speci- Approximate period of development. Host fruit. Number of hours in— Period in larval stage. Average mean tem- mens under obser- vation. Instar z Instar 2 Instar 3 perature for period of devel- opment. J Sept. 25 to 29 . do Papaya . . . .do 26 24 25 27 26 24 23 23- s 23-5 22. 5 25-1 24 2S 25 25 40 24 23- S 22 23-5 23 36 5° 54 51 60 11 60 60 60 60 60 Days. Hours- 4 7 4 7 4 10 4 16 4 22 4 12 4 10 4 10 4 II 4 II 5 I 5 6 6 7 'F. 78.2 79 79 79 79 78.2 78.2 78.2 78.2 78.2 78.2 79 79 79 79 79 .do. . . . . do . . Sept. 23 to 30. do do do do do Sept. 29 to Oct. 4. do do do do do do do do do Aug. 22 to 26. . ss 28 do Aug. 22 to 28. . Aug. 22 to 29. . .... do Cucumber . . . 42 ^6 Cucumber . . . THE PUPA At mean temperatures varying from 71.6° to 79.4° F., the pupal stage ranges from 7I4 to 13 days, as determined by observations on the 1,400 pupae recorded in Table III. As the mean winter monthly tem- peratures seldom fall below 70° F., 13 days is probably close to the maximum length of the pupal stage during the cooler seasons in littoral Hawaii. Table III. — Duration of the pupal stage of the melon fly'^ Date of pupation. Aug. 14, a. m . Do Do Aug. 17, a. m. Do Do Aug. 18, p. m. Do Sept. 17, a. ra. Do Sept. 20, a. ra. Do Sept. 19, a. m. Do Feb. 3, a. m. .. Do Feb. 4, a. m. .. Do Date of emergence. Aug. 22, a. m . Aug. 23, a. m . Aug. 25, a. ra . ...do Aug. 25, p. m. Aug. 26, a. m. ...do Aug. 27, a. m. Sept. 26, a. m. Sept. 27, a. m. Sept. 29, a. m. Sept. 30, a. m. Sept. 28, a. m. Sept. 29, a. m. Feb. 15, a. m. . Feb. 16, a. ra. . do Feb. 17, a. m. . Number Niunber of adults of days of emerg- pupal mg. stage. 5 8 25 9 I II 60 8 I 8-5 27 9 2 7-S 10 8-5 26 9 37 10 355 9 33 10 300 9 15 10 45 12 192 13 38 12 228 13 Average mean tempera- ture. °F. 78. 78, 79 79 79.2 79-3 79-4 79.2 79.2 79-4 79-3 79-4 79-3 71.6 71.6 71.6 71. 6 " Total number of pupnc imder observation: 1,400. 272 Journal of Agricultural Research Vol. Ill, No. 3 THE ADULT The adults of the melon fly have proved most interesting from the standpoint of general hardiness, longevity, and oviposition. Longevity.— At the present time (Aug. 30, 191 4) the writers have about 205 adults that emerged on February 17. They are, therefore, 6 months and 14 days old and are as strong and vigorous in appearance and action as when they emerged. Of the 248 adults ahve on June 19, but 15 females and 28 males have died to date. If the death rate con- tinues as low in the future, a few adults will probably live to be a year old. Sexual, maturity. — Neither male nor female melon flies are sexually mature when they emerge from the pupa. Out of about 200 individuals emerging on May 24, one pair was noted in coition on June 13, or 20 days after emergence. Among a second lot of adults, emerging on May 23, no adults mated until June 16, when two pairs were seen in coition. The majority of females in these lots did not mate until fully 25 days old. The daily mean temperatures for the period from May 23 to June 16 averaged 75.5° F. Sexual activity begins only at sunset. From sunset to dark copulation occurs and lasts in many instances until daybreak. Oviposition. — At mean temperatures averaging 75.5° F., females did not begin egg laying until about one month after eclosion. While fruit was placed in jars with about 1,000 adults which emerged from May 23 to 25, no attempts at oviposition were noted until June 23, when 1 2 punctures containing no eggs were made in a mango by females that emerged on May 24. The first eggs, 12 in number, were laid on June 25, or 32 days after eclosion. No eggs were obtained from females that emerged from May 25 until June 28, or 34 days after eclosion. Daily rate of oviposition. — While females do not begin ovipositing until about i month old, they continue to lay eggs for a long period. Thus, in Table IV is recorded the daily rate of oviposition of seven females during the first three months after emergence, while in Table V is recorded that of three females during the fourth, fifth, and sixth months of their life. Table IV. — Daily rate of ovipositioti of the melon flies thai emerged ort May 25 and were placed separately with fruit on June 25, igi4 Number of eggs deposited. Fly No. J. Fly No. J. Fly No. 3. Fly No. 4. Fly No. 5 Fly No. 6. Fly No. 7. Tulv 10 23 ZI 13 17 ix. !«;. . ... 14 12 16 17. . , , 9 14 18 19 19 6 '3 23 lO Dec, 15, 1914 Melon Fly 273 Table IV. — Daily rate 0/ oviposition of the melon flies that emerged on May 25 and -were placed separately ivitk fruit on June 2$, igi4 — Continued Number of eggs deposited. Date of observation. Fly No. I. Fly No. J. Fly No. 3. Fly No. 4- Fly No. s- Fly No. 6. Fly No. 7- 3 2C 26 23 29 28 20 21 10 15 2 15 25 5 16 •7 . 6 8 10 3 8 19 11 6 2 10 12 17 . 16 10 16 I 2 5 9 18 II 12 20 21 . . 8 24, 17 12 9 Total X5S 20 47 55 116 89 SI Table V. — Daily rate of oviposition of melon flies that emerged on February 17 and were f>laced separately with fruit on May 22, igi4 Number of eggs deposited. Date. Number of epgs deposited. Fly No. I. Fly No. 3. Fly No. 3- Fly No. I. Fly No. J. Fly No. 3. May 22 21 June 1 2C 4 26 e 6 17 28 8 13 xo . 14 4 2 2 0 31 10 274 Journal of Agricultural Research Vol. Ill, No. 3 Table V. — Daily rate of oviposilion of melon flies that emerged on February ly and ■were placed separately witb fruit on May 32, 1974— Continued Number of eggs deposited. Date. Number of eggs deposited. Fly No. 1. Fly No. 2. Fly No. 3- Fly 1 Fly No. I. > No. 2. 1 Fly No.3- June II Tulv 17 18 T'Z 10 .. 16 T< 9 21 2 16 I 22 12 18 14 2t; 20 26 17 13 10 8 28 22 28 4 24 30 4 13 26 Aug. I 28 10 ? 13 ■lO 18 Tulv I 6 -2 8:::::::::::: 8 17 8 10 c 10 21 6 12 10 21 8 0 7 14 II 16 12 12 13 3 18 14 Total IC 16 166 133 169 16 3 While the individuals whose rate of oviposition is recorded in Table V were not given an opportunity to oviposit during the first three months, it will be noted that they oviposited quite as freely at the end of the sixth month as did the youngest females. There is no reason to believe that the first group of melon flies (Table IV) will not continue to oviposit, since the second group (Table V), even before they had an opportunity to oviposit in fruit, oviposited on the sides of the jar containing them. It will be noted that the female melon fly oviposits very irregularly. She is inclined to deposit a large number of eggs at greater intervals, whereas the female of the Mediterranean fruit fly deposits a smaller number nearly every day. Thirty-six eggs is the largest number ever secured from one melon fly in one day. The greatest daily average thus far recorded is that of fly No. i of Table IV, which laid a total of 155 eggs during the first 46 days after she commenced ovipositing, an average of 3.4 eggs per day. IDENTIFICATION OF THE SEEDS OF SPECIES OF AGROPYRON By Robert C. Dahlberg. Seed Analyst, Agricultural Experiment Station of the University of Minnesota ' INTRODUCTION The identification of the "seeds" - of the species of Agropyron is an important problem to the farmer, the seedsman, and the seed laboratory. Agropyron repens (quack-grass) is recognized as a dangerous weed, and the similarity between the seed of this species and other common but more desirable species of Agropyron gives rise to confusion. Up to the present time no diagnosis has been discovered which appears to be en- tirely satisfactory for use in seed-laboratory practice. This paper deals only with the seeds of the species of Agropyron which are common in the Minnesota seed trade — viz, (i) quack-grass {A. repens (L.) Beauv.) ; (2) western wheat-grass {A. smithii Rydb.); and (3) slender wheat-grass {A. tenenim Vasey). The diagnosis here presented has proved to be not only sound but easily appUed in many hundreds of tests made at the Minnesota Seed Laboratory. HISTORICAL REVIEW The species of Agropyron are quite easily distinguishable from each other when characteristics of the root systems, leaf characters, spikes, and spikelets are taken into consideration, and these differences are de- scribed in standard works on the taxonomy of the flowering plants (i, 2, 3, 4).^ In so far as the seeds are concerned, however, these pub- lished descriptions are not sufficiently detailed for use as a basis for identification of seeds alone. Hillman (5) published detailed descrip- tions of Agropyron spikelets which agree in every way with the observa- tions of the writer, but do not include individual seed characters in suffi- cient detail for an accurate diagnosis of the seeds alone. Stevens (8, p. 113) describes in some detail the empty flowering glumes, the size and shape of the seeds, and the rachilla of several species of Agropyron. His descriptions, except as to the rachilla, agree essen- tially with those of the writer. The rachilla is described by Stevens as follows : Tlie footstalk (rachilla) in Agropyron repens is entirely smooth, while in Agropyron occidentale * it is rather variable, but commonly with scattered, short, stiff hairs. While this may appear to be true when seeds are examined under a low-power lens, yet the rachilla of seeds of A. repens when magnified ' The writer wishes to acknowIedRC the assistance of Dr. E. M. Freeman, Assistant Dean, and Chief of the Division of Botany and Plant Patholoi^y. and Mr. W. L. Oswald, Chief of the Seed L,aboratory, Min- nesota Experiment Station, in planninE the work and Eiving sURcestions. ' The word "seeds" is used in this paper in its commercial sense and includes the grain, or caryopsis, inclosed in its persistent glumes, lemma, and palca, with the persistent rachilla segments. " Reference is made by number to " Literature cited," p. 281. ^ Syn. Agropyron smilhii. Journal of Agricultural Research, Vol. Ill, No. 3 Dept. of Agriculture. Washington, D. C. Dec. 15, 1914 Minn. — 1 (27s) 276 Journal of Agricultural Research voi. m. No. 3 to about 32 diameters exhibits very definite hairy characters, as described below. Pammel and King (6, p. 170) published a brief account of the seeds of A. repens and A. smilhii in which the main points of diflFerence are pointed out as being found in the shape of the palea and in the hairs on its face and edge. While these distinctions are in the main correct, they are insufficient for a complete diagnosis. As to shape of the seeds, these authors make the following statement: The seed of quack grass is more slender and spindle-shaped, while that of western wheat grass broadens out somewhat toward the tip, after the manner of brome and some other grasses. According to the results of the examinations of a large number of seeds, the writer finds that the difference in shape above noted is not constant. Hence, its use as a single determining character is not warranted. In a taxonomic key of seeds of Agropyrpn issued by Sarvis (7, p. 2) the rachilla of A. repens is described as being "puberulent, each hair being glandular at the base." According to the writer's observations, the rachilla would more properly be described as hirsutulous. More- over, in order to discern the glands at the base of the hairs, a compound microscope is required, which makes the use of this character imprac- ticable for ordinary seed-laboratory methods. Even with such high power, the glands are not always clearly discernible. A glabrous palea in A. repens and a hispid palea in A. smithii are indicated by Sarvis as important characters in the determination of the Agropyron species. While this is true for the majority of seeds of these species, the writer has found many seeds in which this distinction does not hold. These characters intergrade to such an extent that they are not only unre- liable, but are misleading as a single diagnostic criterion. Sarvis also holds that the tip of the palea of .4. tenerum is very puberulent. This is true not only for the species above mentioned but also for A. repens and A. smithii. SEED CHAR.'VCTERS OF SPECIES OF AGROPYRON In the determination of seeds of Agropyron there are no absolutely fast and definite single characters by which a seed of one species may be unfailingly distinguished from the seed of any other species. Varia- tion is found not only in the seed but also in the other unit parts of the plants, particularly in the spikes and spikelets (Pis. XXXIV, XXXV, and XXXVI). Moreover, seeds growing in different localities may exhibit considerable variation. This variation necessitates a close study of numerous characters of each seed, and any diagnosis to be of value must be based on a large number of seeds collected from a wide range and under widely differing conditions. It is also obvious that the larger the number of seed characters which are studied the greater will be the possibility of making an accu- rate determination of the species under examination. A single char- acter may vary to such an extent as to be quite untypical of the species, and consequently a determination based upon only one character may be incorrect, and therefore misleading. The necessity for an inti- mate knowledge of several distinguishing characters is even more Dec. 15, 1914 Seeds of Agropyron 277 pronounced when Agropyron seeds become so mutilated that portions of the glumes are destroyed, as is frequently the case in commercial seed mixtures. In some cases the glumes are entirely gone, leaving only the grain, and determinations according to characters described below then become impossible. No satisfactory method of real prac- tical value has yet been worked out whereby the seeds without the glumes may be accurately determined, and it seems probable that in such cases one may be compelled to resort to microscopic sections. One characteristic difference may be noted, however — namely, that the color of the matured grain of .4 . tenerum is somewhat lighter than that of either A . repens or A . smiihii. The two latter approximate each other closely in color. SOURCE OF SEEDS Materials for this work were secured from as many sources as possible, as given in Table I. Only those samples have been considered which were obtained from and determined by a competent botanist, who was sure of the origin of the seed. Table I. — Sources of seeds of Agropyron s/^p. used in investigation Source of seed. A. repens. A. smithii. A. tcnemin. Canada (Manitoba;. . Illinois (Chicago) ^. . Iowa Michigan Minnesota New York (Geneva). North Dakota North Dakota b Russia * Washington (State). . Wisconsin 6 Wisconsin Wyoming (a n (a) (a) 8 ■(°y (a) " Sample received. ^ From Seed Laboratory of United States Department of Agriculture, Washington. D. C. LABORATORY METHODS OF IDENTIFICATION It is obviously necessary that all methods of identification, especially for use by farmers or seedsmen and even for seed laboratories, be as simple as possible and that they do not require elaborate or expensive apparatus. If, however, the distinguishing characters are not visible to the naked eye or with the aid of an ordinary magnifying glass, then it becomes absolutely necessary to use a higher power of magnification. In the identification of seeds of Agropyron spp. it is advisable to use a magnification of about 32 diameters for the best results. The Greenough binocular giving the stereoscopic view has proved very satisfactory. It is absolutely necessary when examining seeds under the lens to place them so that the base of the seed is toward the light, in which position the light will be properly reflected from the hairs, making them appear clear and well defined. 278 Journal of Agricultural Research Vol. Ill, No. 3 SHAPE OF SEED The seed of A. repens (fig. i, A) has its point of greatest divergence about midway between the base and the tip, differing in this respect from A . ienerum (fig. i , C) , which has its point of greatest divergence about one- third of the length of the seed from the tip. The lemma and the palea of the latter species flare out more or less at this point, thus making the seed look flattened and thin. In the majority of cases the seed is unsym- metrical in shape, the top portion of the glumes being affected by a lateral displacement, as shown in the illustration. This makes possible a quick and accurate deter- mination of a bulk lot of seeds of A. ienerum. The seed of A . smithii (fig. i , B) has the same general shape fi^k /■.■'•:'\ ^s that of A. repens, but it is larger and has a more ro- bust appearance. RACHILLA It is impossible to de- scribe very definitely the characteristics of the ra- chilla of the different species of Agropyron because of the variation . In a general way the sides of the rachilla of .4. repens are more nearly parallel, and the rachilla itself is more or less ap- pressed to the palea. In A. smithii the sides of the rachilla diverge noticeably more from the point of its attachment. The rachilla stands out more promi- nently from the seed, being materially different in this respect from A. repens (fig. 2). The rachilla of A. ten- erum has no particularly characteristic shape, varying from the slender to the short, stout, diverging type. A very good idea of relative size and shape of these seeds may also be gained by studying Plate XXXVII, which shows typical seeds, together with a typical spikelet of each of the three species. The hairs clothing the rachilla constitute a valuable character used in the determination of the seed. However, care and good judgment must be exercised because of the great variation which may occur. The characteristic rachilla of A. repens (fig. i, ^) is sparsely covered with short, minute hairs having a rather large base. Occasionally a glandular structure may be discerned at the base. This, however, can only be seen with a high-power lens and is not considered of sufficient Fig. I. — Detail drawings of dorsal view of Agropyron spp.: A Agropyron repens; B, A. smithii; C, A. ienerum. X 9. Dec. IS, 1914 Seeds oj Agropyron 279 importance to warrant its use as a determining character. No rachilla of A . repens has been found which had the hirsute character of A . smithii (fig. I, B) or the pilose character of A. tenerum (fig. i, C). The rachilla of A. smilhii is characterized by hairs of the same gen- eral shape as the hairs found on the rachilla of A. repens. They are, however, larger and stronger and the number is noticeably greater. This characteristic is fairly uniform. The rachilla of A . tenerum is characterized by hairs of a pilose nature. They are long as compared with those of A . repens and A . smithii, and may often be distinguished by this feature alone from these two species, as the pilose nature has never been observed on them. However, an absolutely au- thentic specimen of A . tenerum has been examined which had a rachilla much resembling fl f-^^A AA that of A . repens. The hairs were short, but were not as |!V / J 'i \ \ /(i large as the base. Other r'l\ I f \ lit. characters on the seed, how- ever, made it possible to place it accurately in the species tenerum. LEMMA Another distinguishing character and one which is reliable as to uniformity may be found at the base of the lemma on the ventral side of theseed. In A.te nerum (figs . 2, C, and 3, C) there is a line of hairs which extends from the base of the rachilla on the dorsal side of the seed around and entirely across the face of the lemma on the ventral side near the base of the seed. In some cases it may be impossible to distinguish the hairs on the middle of the lemma, but the surface of the lemma at this point is rough- ened sufficiently so that it character. The seed of A. repens (figs. 2, A, and 3, A) has no such characteristic line of hairs, but the basal portion of the lemma is entirely smooth and shiny. This character in the seed of A. smithii (fig. 3, B) is somewhat variable and is therefore not of much value. Most commonly, however, it is found that the ring of hairs extends part way around on either side, and on the middle of the lemma there is a space which usually is eatirely smooth. Fig. 2. — Side views of basal portions of seeds of Agropyron spp.. showing the relative projection of the rachilla: A, Aoropyron repens; B. .1. smilhii; C. A. tenerum. X 9. is noticeable. This is a fairly definite 28o Journal of Agricultural Research Vol. III. No. 3 Fig. 3. — Detail drawings of ventral view of seeds of Agropyron spp.; A, Agropyron repens; B, A. smiihti; C, A. Unerum. X 9. PALEA The part of the seed which discloses good and reasonably definite characteristic differences is the palea. The face of the palea in A . repens and A . tenerum (figs, i, A, and i , C) is practically glabrous, except near the tip, where it is puberulent. Occasionally there is a small number of hairs distributed over the face of the palea. Since the tips of the paleae in both of these species are always puberulent, this can not be used as a distingixishing character. The palea of the seed of A. smithii (fig. I, B) is quite hirsute over its entire surface. The hairs on the edge of the palea have a distinctive shape for each of the three species and are very useful as a determining factor. Those of A. repens (fig. 4, A) are rather short, stout, and somewhat blunt. Those of A. smithii (fig. 4, B) are about as coarse as those on A. repens, but are noticeably longer, thus making them appear more slender. On A. tenerum (fig. 4, C) the hairs are finer, closer together, and more acutely pointed than in the case of the two others. The palea of A. smithii (fig. i, B) has a very characteristic tip, which character runs fairly uniform through- out the species. The tip of the palea is definitely divided, making a well- defined cleft (PI. XXXVII, 2,0). In some cases it is rather difficult to dis- tinguish this cleft, as the lobes may be slightly overlapped. The tips of the palea of .4 . repens and A . tenerum are simply rounded or only slightly indented. SUMMARY It is possible by careful examination to distinguish in commercial seed mixtures the seeds of the three species of Agropyron: A. repens, A. smithii, and A. tenerum. There is no one character which can unfailingly be relied upon for this diagnosis, but the combined characters of lemma, palea, and rachilla are necessary for a safe determination. Probably the nearest approach to a single critical structure is found in the palea, which exhibits fairly definite characters in each of the species. The diagnostic differences are summarized in Table II. Fig. 4. — Edge of raciiilla in Agropyron spp., showing shape and comparative size of bris- tles; A. Agropyron repens; B, A. smithii; C, A. tenerum. X 9- Dec. 15, 1914 Seeds of Agropyron 281 Table II. — Diagnostic differences of Agropyron spp. Character. A. re pens. A. smithii. A. tenerum. Shape of seed 3oat-sbaped .^ Boat-shaped Widest one-third of dis- tance from the tip, which is more or less flattened. Rachilla Sides approximately parallel. Hair* few. Sides divergent. Hairs numerous, stout, but Variable in shape and size. Hairs niunerous. short, and stout. longer than those which characterize A. re pens. slender, and long. Paleai Face Puberulcnt at tip; other- Hirsute over entire face . , Puberulent at tip. Re- wise glabrous. maining surface gla- brous. Edges Characterized by short. Hairs stout, but longer Hairs fine, acute, and stout, and blunt hairs. than those of A. repens. close together. Tip Rounded ur indented . Cleft Rounded or indented. Lemma Smooth and shiny at Usually with a break in Line of hairs c?ctends base on ventral side. the line of hairs on ven- across lemma at its tral side at base of seed. base. LITERATURE CITED (i) Beal, W. J. 1896. Grasses of North America, v. 2. p. 636-637. New York. (2) Britton, N. L., and Brown, Addison. 1913. Illustrated Flora of the Northern .States, Canada and the British Posses- sions, cd. 2, V. I. p. 2S3, illus. New York. (3) Engler, Adolf, and Prantl, K. A. E. 1887. Die Natiirlichen Planzenfamilien. T. 2, Abt. 2, p. 79, fig. 91. Leipzig. (4) Gray, Asa. 1908. Gray's New Manual of Botany, rearranged and extensively revised by B. L. Robinson and M. L. Fernald. ed. 7, p. 166, fig. 185-189. New York. (5) HiLLMAN, F. H. 191 1. The distinguishing characters of the seeds of quack-grass and of certain wheat-grasses. U. vS. Dept. Agr.. Bur. Plant Indus. Cir. 73, 9 p., (6) Pammel, L. H., and King, Charlotte M. 1910. Results of seed investigations for 1908 and 1909. Iowa Agr. Exp. Sta. Bui. 115, p. 156-175. (7) S.^rvis, J. T. 1913. Suggestions for the identification of the "seeds" of quack grass (..Agro- pyron repens (L.) Beauv.), western wheat grass (AgropyTon smithii Rydb.), and slender wheat grass (.Agropyron tenerum Vaseyl. 5 p. Brookings, S. Dak. (8) Stevens, O. A. 1910. Work done under the seed control act. In loth Bien. Rpt. N. Dak. .Agr. Col., p. 93-118. 67235°— 14 7 PLATE XXXIV Aqropyron repens: Spikes showing degrees of variation which maj- occur. .4, typical spike. A side view of this spikelet may be seen at top portion of this spike. Natural size. (282) Seeds of Agropyron Plate XXXIV Journal of Agricultural Research Vol. Ill, No. 3 Seeds of Agropyron Plate XXXV Journal of Agricultural Research Vol. Ill, No. 3 PLATE XXXV Agropyron smilhii: Spikesshowing degrees of variation. .4, typical spike. Natural size. PLATE XXXVI Agropyron tenerum: Spikesshowing degrees of variation. .4, typical spike. Nat- ural size. Seeds of Agropyron Plate XXXVl Journal of Agricultural Research Vol. Ill, No. 3 Seeds of Agropyron Plate XXXVII Journal of Agricultural Research Vol. Ill, No. 3 PLATE XXXVII Agropyron spp.: Typical seeds and spikelets. Enlarged. Fig. I. — Agropyron repens. Fig. 2. — Agropyron smithii. Fig. 3. — Agropyron tenerum. ADDITIONAL COPIES OF THIS PCBUCATION MAT BE PROCUKEE FROM THE eUPERINTEITOENT OF DOCUMENTS GOVERNMENT PRINTING OFFICE WASHrNOTON, D. C. AT 25 CENTS PER COPY . SimscRiPTioN Per Year, 12 Numbers, J2.60. Vol. Ill JANUARY, 1915 No. 4 JOURNAL OF AGRICULTURAIv RESEARCH CONTENTS ' . ■ ^ Pag« Observations on the Life History of AiJrilus Bilineatus . 283 ROYAL K. CHAPMAN Effect of Dilution upon the Infectivity of the Virus of the Mosaic Disease of Tobacco 2Q5 H. A. ALLARD Moldiness in Butter 301 CHARLES THOM and R. H. SHAW Susceptibility of Citrous Fruits to the Attack of the Medi- terranean Fruit Fly . 311 E. A. BACK and C. E. PEMBERTON Physiological Changes in Sweet Potatoes during Storage . 33 1 HEIWEICH HASSELBRING and LON A. HAWKINS Three-Cornered Alfalfa Hopper 343 V. L. WILDERWLUTH DEPARTMENT OF AGRICULTURE WAS HINGTON , D.C. WAliMINOrOK I GOVERNMfXT MUNTINO OFFlOt I «1» PUBLISHED BY AUTHORITY OP/THE SECRETARY OF AGRICULTURE, WITH THE COOPERATION OF THE ASSOCIATION OF AMERICAN AGRICUL- TURAL COLLEGES AND EXPERIMENT STATIONS EDITORIAL COMMITTEE FOR THE DEPAHTMEITT FOR THE ASSOCIATION KARL F. KELLERMAN, Chairman RAYMOND PEARL Physiologist otuI Assista^ii Chief, Bureau Biologist, M<^ine Agricultural Experiment of Plant Industry ' EDWIN W. ALLEN Assistant Director, OffUe of E%periment Stations CHARLES L. MARLATT A ssistant Chi^f^ Bureau of Bntomoloay S,tation H. P, ARMSBY Director, Jvstiivie of Aniniol Nutrition^ The Pennsylvania State' Colleoi^ E. M. FREEMAN Botanist, Plant Pathologist, and Assistant Dean, Agricultural Experiment Station of the Uiiiversily of Minnesota All correspondence regarding articles from tlie Department of Agriculture should be addressed to Karl F. Kellertnan, Journal of Agricultural Research, Washington, D. C. All correspondence regarding articles from Experiment Stations should be addressed to Raymond Pearl, Journal of Agricultural Research, Orono, Maine JOIMAL OF AGRICULTURAL RESEARCH DEPARTMENT OF AGRICULTURE Vol. Ill Washington, D. C, January 15, 1915 No. 4 OBSERVATIONS ON THE LIFE HISTORY OF AGRILUS BILINEATUS By RoY.'U, N. Chapman, Graduate Assistant, Division of Entomology, Department of Agriculture, University of Minnesota INTRODUCTION At the present time the two-hned chestnut borer, Agrilus bilincatus Weber, is commonly associated with the death of many oaks {Quercus spp.) in the southeastern part of Minnesota. In 1885 reports called attention to the damage done by this insect on oaks in Massachusetts, and since that time they have been frequently mentioned as enemies of the chestnut (Castanea dcntaid) and oak. It was not until 1 897 that F. H. Chittenden * described the adult, the larva, and the pupa, in connection with a brief summary of its life history, so far as it was known at that time. In this article it was stated that the adults appeared in the District of Columbia from May to the middle of June and laid their eggs on trees and that the larvae worked under the bark across the grain of the wood, making a burrow from 6 to 10 inches in length, and by the next spring had constructed a chamber in the bark of living trees, where the pupal stage of about two weeks was passed. Although the name of this beetle has been common in current entomological literature, nothing of note has been added to the knowledge of its life history since Chittenden's article. During recent years in the neighborhood of St. Paul and MinneapoUs great numbers of oaks, many of them on valuable residence property, have been killed, and their death has commonly been attributed to this pest. In some of the outlying country districts areas of several acres in extent have been completeh' devastated, leaving the land treeless. The present work was begun at the University of Minnesota during the fall of 1 91 3 at the suggestion of Prof. O. W. Oestlund, of the Depart- ment of Animal Biology, under whose direction the problem was out- lined and the work on the larval and pupal stages begun. In the spring of 1914 the work was continued at the Minnesota Agricultural Experi- 1 Chittenden. F. H. Insect injury to chestnut and pine trees in Virginia and neighboring States. In U. S. Dept. Agr., Div. Entom. Bui. 7. n. s.. p. 67-71. 1897. Journal of Agricultural Research, Vol. HI, No. 4 Dept. of Agriculture, Washington, D. C. Jan. 15, 1915 Minn. — 3 284 Journal of Agricultural Research voi. iii.no. 4 ment Station under the direction of Prof. A. G. Ruggles, in charge of the Section of Spraying and Tree Insects, Division of Entomology. The aim of this paper is to report new observations on the Hfe history and ecologic relations of Agrilus bilineatus. The descriptions of the adults, larvae, and pupae, already referred to, are so well known to ento- mologists that they will not be repeated. The drawings reproduced in Plate XXXVIII of the beetles, pupae, larvae, and eggs were made at the Department of Animal Biology under the direction of the author. The photographs reproduced in Plate XXXIX were made at University Farm by the Experiment Station photographer. The egg-laying habits have been described in this article in some detail because they have not been known to literature, and the same is true of the leaf -eating habits of the adults. The same may be said of the eggs and newly hatched larvae. Some of the observations of the life history of the larvae have not agreed with previous descriptions, and these, together with a few additional notes on habits, are included, while others, together with the details of the pupal stage, will be deferred to a later report. ECOLOGY The four common species of oak in the southeastern section of Minne- sota, Quercus alba L., Q. macrocarpa Michx., Q. rubra L., and Q. coccinea Wang, are subject to infestation with Agrilus bilineatus. It seems that the members of the black-oak group are slightly more susceptible to attack than those of the white-oak group, but in localities where the infes- tation is severe none of the species is exempt. In some cases the adult borers appear to prefer trees of a certain locahty, or, in other cases, certain individual trees, to others even in the immediate vicinity. In general, this preference is associated with a weakened con- dition of the trees, but this is by no means universal. The env-ironmental conditions of certain localities, such as drought, crowded pasturage, or cultivation, have made nearly all the trees subject to infestation by the two-lined chestnut borer, possibly because of a general weakened condi- tion. In other cases individual trees weakened by injury or disease have been attacked, while in still other cases trees which show no signs of weakened vitality are attacked and killed by this insect. It has often been found that the shoestring fungus, Armillaria mcllca Vahl, has apparently been the cause of the weakened condition of the trees, and the chestnut or oak borers have followed it. In fact, it has sometimes appeared that A. mellea was the primary cause of the death of the trees and that the Agrilus beetle was of only secondary impor- tance. An example of this was found in the neighborhood of Lake Elmo, Minn., where a few dead trees were found with the fimgus Armillaria, but with no traces of lar\^3e of the two-lined chestnut borer. These obser- vations, together with others which showed the shoestring fungus on practically every tree on which the Agrilus beetles were ovipositing, made Jan. IS, 1915 Agrilus Bilineatus 285 it seem probable that the fungus was the primary factor in causing the death of the trees. In the vicinity of Robbinsdale, Minn., and in other locahties, the Armillaria fungus is present, but is not so apparent, and all the dead trees show traces of oak borers. Furthermore, many trees with dead trunks started up from the roots the following year and gave every evidence of suffering from the girdling of the cambium layer rather than from a root infestation with the fungus. At Robbinsdale a tree was observed on July 27 with its leaves wither- ing as if it had been scorched, a characteristic of trees infested with Agrilus hilineatus. Three days before (July 24) this tree had appeared perfectly normal, but on the 27th the entire trunk of the tree was being girdled by the beetles. It was grubbed and its roots were examined by Mr. F. J. Piemeisel, of the Division of Plant Pathology, who stated that the fungus Armillaria mellea was not present and could have nothing to do with the death of the tree and that the root system seemed perfectly healthy. Other trees examined since gave similar evidence. It has not yet been possible to show any relation between the amount of fungus present on a tree and the severity of the attack by the beetles. On land that was being cleared for cultivation near Lake Elmo 40 appar- ently healthy trees were attacked by Armillaria mellea, and none were found to be free from it. Furthermore, W. H. Long states * that a large percentage of the oaks examined by him in the eastern section of the United States have been found to be infested with the fungus. If this is true also in Minnesota and the presence of the fungus is taken as evi- dence of the low vitahty of the tree, all but a small percentage of the oaks in Minnesota are now susceptible to attack by Agrilus bilincalns, and in all these cases the beetles must be looked upon as of only sec- ondary importance. This, however, would possibly place imdue weight on the mere presence of Armillaria mellea. As mentioned above, in a few cases in the vicinity of Lake Elmo the death of the trees may be due to Armillaria mellea alone. In the majority of cases the fungus was present, but the trees were girdled by the two- lined chestnut borer, seemingly a hastening factor, at least, in the death of the trees. There still remains the possibility that the fungus was the primary factor. There are cases, as already mentioned, where the borers alone have attacked and killed apparently nonnal trees when the fungus was not present. The economic importance of this condition can hardly be overemphasized, for it means that the Agrilus beetles, in spite of their supposed preference for unhealthy trees, chose one healthy tree when many trees infested with the fungus were available, indicating tliat the interrelation between the Armillaria -mellea and the Agrilus bilinrahis may not be of such primary importance as would appear at first. * Long. W. H. The death of chestnuts and oaks due to Annillaria mellea. V. S. Dept. Art. BuI. 3q. p. 4. 1914. 286 Journal of Agricultural Research voi. m,No.4 THE ADULTS On the 17th of June the first observations of the adult Agrilus bilineatus were made in the vicinity of Lake Elmo, Minn. Early in the after- noon two or three could sometimes be seen at one time on a dying tree. On the same day an adult was taken from its pupal cell in the bark, and a lar\-a was found preparing to pupate. Since no adults were seen two days earlier while the author was collecting in the same locality, it seems that these observations fix the appearance of the adults quite definitely for this section in a normal year. The adult borers increased in numbers until they reached their greatest abundance about the ist of July, when as many as 10 were seen at one time on a small area of bark. In the afternoon of June 26, 80 specimens were collected near Savage, Minn. There was a noticeable decline in numbers after the first week in July, and by July 20 the last record of the adults was made. Continued careful search in the field after that time was not rewarded. During the last days of the adults' flight many instances of apparent feebleness became evident. On one occasion, while watching Agrilus beetles, a female was seen to fly slowly toward a tree apparently intend- ing to alight on it as usual, but the insect fell to the ground as if unable to cling to the bark. The borer then made a second attempt, which was also unsuccessful, and it was picked up from the ground where it was lying apparently exhausted. The behavior of other chestnut borers in the field and in the insectary leads to the belief that the last chapter of their history is marked by the feebleness of old age. Agrilus adults lived about 12 days in the insectary, where they were kept in a cage inclosing an oak tree. The conditions were so favorable in the insectary that many details of the various habits could be studied to great advantage, and continuous observations throughout the season were made possible in spite of weather conditions. The habits of the Agrilus beetles were governed with such regularity that a seemingly definite program was discovered which ser^-ed as a guide for later observations. At no time during the season were beetles found, even on the sunny side of the trunks, until nearly noon. On July 7 in a place near Savage, Minn., where the adult borers were very abundant, a careful search was begun shortly after 9 o'clock in the morning and but one specimen was found at 9.30 a. m., a few at about 10 o'clock, and not until 1 1 o'clock were the}' found in their usual num- bers. From this time until shortly after noon they increased in num- bers; then they gradually disappeared until few were to be seen late in the day. The latest field obser\^ation was made about 6 o'clock. In the insectary the beetles were inactive in the bottom of the cage until about 8 a. m., when they would begin to fly about and feed on the foliage of the tree. Many were also obser^'ed courting during the early forenoon. The absence of the beetles from the tree trunks, which they frequent laterin Jan. IS, 191S Agrilus Bilineatus 287 the day, and their feeding habits as observed in the insectary, indicate the probabiHty that the early part of the day is ordinarily spent in this way. It was definitely proved that the adults of Agrilus bilineatus feed on the foliage. They usually eat around the margins of leaves, but also tear off the epidermis and sometimes eat nearly the entire leaf, including the midrib. Plate XXXIX, figure i , shows a leaf on which four beetles had fed for 24 hours. They even ate the edge of a paper bag which was on the floor of the insectary, where most of the obsers'a tions were made. The difficulty of observing the feeding in the field was increased by the protective habits of the beetles. The flight to and from leaves or logs and tree trunks was ordinarily quite direct, but in some cases they hov- ered before lighting. When disturbed, however, they flew rapidly in a zigzag manner, which is probably of protective importance, as under such conditions it is almost impossible to follow them even with the eye. If they are startled while at rest, they fold their appendages and drop to the ground, feigning death. They dodge from side to side very quickly and run rapidly over any kind of surface. On one occasion a beetle was watched walking about inside a glass tube. It experienced no difficulty unless it tried to walk upside down upon the slippery concave surface, when it began to lose its footing. This was evidently not a new occur- rence, for it immediately put one of its front tarsi to its mouth and ap- parently moistened it; it then reached this appendage back to rub it against the posterior ones. When all the tarsi had been treated in this way, the beetle ran about until it began to slip again, when the process was repeated. On sunny davs during the entire season males and females were court- ing and mating. These performances were often noticed on logs or woodpiles and on the foliage of small plants at the base of trees, as well as on tree trunks. The courting was usually abbreviated and sometimes wanting. Males were seen to fly from the air directly to females on the tree. At other times a male was seen to side-step to within an inch of a female and then spread its wings as if to fly before advancing farther. In one case a male stood near a female until she finished ovipositing and then mated with her. But it was not always the males that made the advances, for it was not uncommon to see females courting males, in which case they often found themselves ignored. In mating, the sexes were together from 2 to 1 2 minutes, with an average of about 4 minutes. The females were ovipositing from June 19 to July 13. No record of ovipositing was obtained before 1 1 a. m. and but one after 5.30 p. ni. While it was not so common to see females actually ovipositing, they could be seen nearly all the time during the hours of activity with their ovipositors out searching for places to oviposit. They laid the eggs in the bottoms of cracks, but not every crack was suitable. The females went carefullv along, using their ovipositors as tactile organs, exploring everv crevice. The insect often appeared to be in a great hurry and 288 Journal of Agricultural Research voi. 111.N0.4 rushed from one crack to another until the proper place was found. During the first week of July one female took 21 minutes from the time she lighted on a tree to find a place to lay her eggs. Not a minute of this time was wasted, and many cracks were rejected before the favorable one was located. The cracks chosen for o\iposition were usually ones that were quite deep. In one case a female used a crack which had evidently been made by lightning, but in all other cases a crevice at the bottom of a deep crack between ridges of the bark was chosen (PI. XXXVIII, fig. i). Since the bark is usually rougher on the trunk and larger limbs, especially near the ground, more favorable places to oviposit are found on these parts of the tree. Observations were made 25 feet from the ground, but few beetles were seen and none were looking for places to oviposit. In ex- amining trees which had been killed, it was found in one case that an egg had been laid in a crevice at the axis of a small branch 41 feet 6 inches from the ground. On one tree which was very badly infested practically every branch more than i]4 inches in diameter had burrows on it. Many attempts were made to find whether the sunny side of trees was preferred to the shady side, but, so far as could be determined, the beetles showed no preference. From this evidence it seems that the eggs may be laid on any part of the tree which afiFords suitable cracks and that since such cracks are most numerous on the trunk, this is the usual place for ovipositing. The females settled down when a favorable crack was found and were apparently motionless during oviposition. The beetles stood in any con- venient position during the process, and there was no relation between the number of eggs laid and the length of time apparently spent in de- positing them. Oviposition lasted from i to 5 minutes, and from i to 10 eggs were laid. It is probable that the number of eggs in a cluster de- pends upon the favorableness of the crevices in which they are deposited, because the females usually hasten to find another place as soon as one cluster has been laid. Just how many eggs are laid in all by one indi- vidual is not known. THE EGG Plate XXXVIII, figure 2, shows a very typical cluster of four eggs, the average number, on a bit of bark taken from the bottom of a crack, just as they appeared within an hour after they were deposited. One of the eggs lies entirely exposed, showing the typical form of an undisturbed egg, while the others illustrate how nicely they mass together and fit the irregularities of the crevices. It will be noticed that the eggs are not plump but have an unfilled, wrinkled appearance, which makes it possible for them to fit into crevices of all shapes. A typical egg, such as the exposed one in the illustration, is somewhat oval and measures about 940/t in length, ^iofi in width, and about 30o/( in thickness. The newly laid eggs were covered with a glistening substance which stuck them firmly in place when dry. Jan. IS, 191S Agrilus Bilineatus 289 The eggs were hatched in the laboratory in from 10 to 13 days. It was found that the outer membrane became dry and shriveled and even cracked in from 3 to 6 days, while the inner membrane became brown and the embryo seemed to develop ajt one side of the egg, which became plumper than the other side (PI. XXXVIII, fig. 3). THE LARViE When the larvse hatched, they broke through the egg membrane on the side toward the bark and immediately began to burrow. As a result, they were not exposed to view, and the eggshells were found filled with the frass which had been burrowed out at the start. In a few cases the eggs had been entirely loosened from the bark for examination, and none of the larvse from these eggs succeeded in getting a burrow started, except in one case where the egg was artificially fastened to the bark before the lan,'a left it. Since the eggs are stuck to the bark in the depths of cracks, they offer a certain resistance to the lar\'a's efforts, which makes it pos- sible for its mandibles to get hold of the bark and start the burrow. The newly hatched lar\'ae (PI. XXXVUI, fig. 4) measure only from i to 1^-2 mm. in length, but one was found capable of reaching the cambium layer in 24 hours by burrowing for 2}^ mm. The fact that the eggs were laid in the depths of cracks made it possible for the larvse to reach the soft cambium layer by penetrating but a few millimeters of hard bark tissue. Having reached the cambium layer they started off in any con- venient direction. Obsen^ations show that burrows made during the first instar often go obliquely across the grain of the wood or with the grain, the larvae being indifferent as to whether they go up or down the tree. If care is taken in removing the bark when green, the tiny burrows can be traced to a widening of the burrow which marks the end of the first instar. The burrows measured showed that the larvae had bur- rowed for a distance of 60 to 135 mm. when the first molt took place. In most cases the widenings in the burrow occurred when the lar\-ae had gone into the wood, less often into the bark, and were again returning to the cambium layer. It is evident that these points marked the limits of instars and that the molting took place in these excavations in the wood, which were often two or three times the length of the larvae. Places were found, however, where the burrows showed that the molt had been made in the cambium layer. The longest larva found in a burrow of the first instar measured 4.6 mm. in length, and the average width of the burrows was 270/i. The burrows made during the second instar measured about 900/i in width and took about the same course through the cambium layer, but they were about twice as long. At the beginning of the third instar quite a different course was usually found, especially in green bark on the trunks of trees, where the burrows were almost always transverse to the grain of the wood. The burrows of the fourth instar were about 2 mm. in width and often attained the length of 500 or 600 mm. Where the 290 Journal of Agricullural Research voi. in, no. 4 bark was thick these burrows were quite generally transverse to the grain of the wood. This condition, as well as the oblique course of some of the smaller burrows, is well shown in Plate XXXIX, figure 3. At the close of the fourth instar the larva burrows out into the bark, if it is thick enough, and constructs a cell in which it hibernates. Here pupation takes place in the spring. These cells are found in the ridges of the bark on the trunk and larger limbs of the tree and in the wood on small, thin-barked trees and limbs. In constructing the cell, the larva burrows out to within a few millimeters of the surface of the bark, with- draws itself 2 or 3 mm., then turns about to one side and excavates around the posterior portion of its body until an oblong cell has been constructed. The portion of the burrow leading to the cell, as well as the short portion between the cell and the bark, is plugged with the frass loosened in mak- ing the excavation. It is evident that at least the portion of the frass at the outer end of the cell was never ingested, for if this were the case it would have passed out at the anus, which has been at the other end of the cell all the time. When the cell is complete, it measures about 10 mm. in length and about 2 or 3 mm. in width and contains the larva bent upon itself, ready for hibernation (PI. XXXVIII, fig. 6). The work of*the larval life is well illustrated in Plate XXXIX, figure 4, which is a reproduction of a photograph of a limb in which the entire burrow is traced, from the shells of the eggs in a crevice of the bark to the larv'a in its pupal cell. The burrow was carefully marked with india ink, so that the black lines represent the exact width of the burrow in every case, and at the places where the larva entered the wood to molt the holes have been marked about with white ink for the sake of contrast. The larva in this case burrowed into the wood for a short distance at first, then returned to the cambium layer, where it burrowed about until it reentered the wood to molt. From the point where the larva entered the bark to the place it emerged from the wood after the first molt the burrow measures 69 mm. in length and 270/1 in width. It will be noticed that in each succeeding instar the burrow is nuich wider and longer, so that in the second instar the length is 103 mm. and the width is goo/t; in the third instar it is 210 mm. long and 1.21 to 1.56 mm. in width; and in the fourth and last instar the length is 456 mm. and the width is 1.96 to 2.15 mm. The total length of this particular bur- row is 835 mm., or nearly 3 feet. Since the burrows made during the early instars are so small that they are hard to find, it seems likely that Chittenden's statement ' that the complete burrow is only 6 to 10 inches in length was based on observations made on burrows which represented the last instar. Even these could hardly have been complete, for burrow-s of this instar nearly 2 feet long have been found. His statement that they are for the most part transverse to the grain of the wood makes it seem even more evident that those of the last instar were described, for, as Plate ' Chittenden, F. H. Loc. cit. Jan. IS. IP'S Agrihis Bilineatus 291 XXXIX, figure 3, shows, the transverse direction is the usual one in the last instar, especially when the bark is thick, while during the earlier instars the burrows run in a more oblique direction. No evidence can be offered as to the duration of instars. Larvae which were in the first stage were found from July 21 to August 13, mature larvae were found in their pupal cells as early as August 7, while the intermediate stages were found throughout this period. It was found that the lar\'ae in the last instar burrow from 2 or 3 to 23 mm. in 24 hours. Larvae of the second instar burrow as far as 6 mm. in the same length of time. Upon consideration of these records it is not surprising that trees infested with Agrihis hilineatus appear to die suddenly when larvae are to be found as numerous as shown by the burrows in Plate XXXIX, figure 3, where each one may consume cam- bium tissue equal to nearly twice its own bulk every 24 hours. In the section shown there were nine larvae in approximately i square foot of bark, each burrowing across the cambium layer. It was also found that when the larvae are so numerous that they confront each other, one or the other is eaten through as if it were merely cambium tissue. This may become an important economic factor, for cases have been obser^'ed in bark which was crowded with grabs where a number of dead lan-ae were found with burrows passing through their abdomens or even their heads. The slowest burrowing was found to be -in the dry wood, where the tissue was evidently the toughest and of the least nutritive value. Trees with growing tissue offer the best opportunity for making extended burrows with great nutritive value to the lar\'3e and to the detriment of the tree. On the other hand, as soon as the tree dies from being girdled, the tissue becomes dry and offers more resistance to the burrowing and is of little nutritive value to the larvae, which may die. A tree which was grubbed up on July 30, at which time it had just died, was examined on August 13. The dried bark, especially on the side exposed to the sun, contained shriveled larvae, over 50 per cent of which were dead. Similar conditions were found in other trees that had died early in the season, when the dryness seemed to affect the larvae more than later when they are in the pupal cells. This may also explain the condition described by Chittenden, who stated that burrows were found in trees, but no lar^-«E were present.' In summarizing the work of the larvae of the two-lined chestnut borer it is also of economic interest to note the wide distribution of the bur- rows on the tree, from the small branches less than an inch in diameter and between 40 and 50 feet from the ground down even to the roots, where in one case a larva was found constmcting a pupal cell 1 1 inches below the surface of the ground. ' Chittenden, F. H. I.oc. cit. 292 J ojirnal of Agricultural Research voi. in. N0.4 THE PUPA The pupal stage of the life history has been studied for the most part in the laboratory. During the winter the larvae were collected in their pupal cells and placed in wash bottles, which were then covered with cheese- cloth. The larvae were found to contract and straighten out in such a way as to face the end of the cell which is next to the bark. When con- tracted ready for casting the lar\-al skin, the laws measured from 6 to 10 mm., instead of 18 to 24 mm., and were greatly swollen, with constric- tions marking the posterior limits of the head, prothorax, and the loca- tion of the appendages. Two or three days before the larval skin was cast the posterior segments were collapsed and empty. When the pupa was ready, it began a series of wavelike dorsoventral bendings, which caused the skin to break on the dorsal side of the head and prothorax a little to one side of the middle. As these movements continued the skin was slipped gradually backward, collapsing as it left the posterior end of the body, until it was entirely off, when the pupa came to rest (PI. XXXVIII, fig. 7). The mouth parts passed to the ventral side and seemed to act as a lever against the side of the cell in pushing the skin backward. The pupal stage lasted about 10 days, during the latter part of which pigmentation began with the eyes, then the mouth parts, head, and thorax. The pupal skin was shed in much the same way as the larval skin, and the adult folded the wings, remaining inactive until the elytra were entirely pigmented. At first the movements of the adults were slow and uncorrelated as contrasted with the great acti\nty later on. The beetle burrowed through the bark as soon as it acquired its full activity and escaped through a characteristic opening (PL XXXIX, fig. 2). The openings are always found on the ridges of bark and resemble the shape of the hole made by the larva when it first entered the bark. Larvse which had not yet pupated were collected as late as June 17, when adults were found making their way out from the pupal cells. From this it seems that the insect in this state normally pupates during the latter part of May and emerges from the cell about the middle of Tune. PARASITIC ENEMIES Two parasites were incidentally noticed. One, which was reared from the larvse, was identified by Mr. S. A. Rohwer as a species of the genus Atanycolus, while another, unfortunately mutilated, which was reared from an egg, was placed by Mr. J. C. Crawford in the family Tricho- grammidae. CONTROL OF THE OAK BORER The method of control heretofore recommended has been the cutting and burning of infested trees before the emergence of the adults in the spring. This is an effective method and needs emphasizing, for people are tempted to leave all the trees which show any .signs of life, with the jau. IS. I9IS Agrilus Bilineatus 293 hope that they will recover the next spring. These are the most danger- ous trees, for, as has has been pointed out, the trees with sap are more favorable to the insects than the dry ones in which the larvae are liable to dry up or starve. The need of other methods, however, seemed imperative. During the past season the trunks and large limbs of some trees were sprayed with an iron- sulphate and lime-sulphur mixture, while others were sprayed with a Bordeaux mixture. This was done as a preventive measure during the egg-laying season and it seemed successful, as no beetles were seen on the trees Nvhich had been sprayed, even though the trees had been covered with beetles the day previous to this treatment. In contrast to this beetles were seen in great numbers throughout the season on the unsprayed trees near by. Other experiments which are under way can not be reported until at least another season has passed, and greater opportunity has been offered to try out proposed methods of prevention and control. PLATE XXXVIII Fig. I. — Agrilus bilinealus: Eggs in position in the bark of an oak tree. '2 natural size. Fig. 2. — Agrilus bilinealus: Cluster of newly laid eggs. X6. Fig. 3. — Agrilus biliiieatus: Eggs shortly before hatching. X6. Fig. 4. — Agrilus bilinealus: Newly hatched lar\'a. X6. Fig. 5. — Agrilus bilinealus: Mature larva. X6. Fig. 6. — Agrilus bilinealus: Larva in its cell. Section made perpendicular to the surface of the bark. A, Point at which adult will emerge; B, burrow stopped With frass. X6. Fig. 7. — Agrilus biliiieatus: Pupa in cell. Section made parallel to the surface of the bark. X6. Fig. 8. — Agrilus bilinealus: Adult female. X6. Fig. 9. — Agrilus bilinealus: Adult male. X6. Drawings by Helen A. Sanborn. (294) Agrilus Bilineatus Plate XXXVIII Journal of Agricultural Research Vol. Ml, No.4 Agrilus Bilineatus Plate XXXIX J-'urnal .,f Agricultural R.^i^a;^!^, Vol. III. r-Jo. 4 PLATE XXXIX Fig. I Fig. 2 Fig- 3 Fig- 4 hatched — Leaf showing work of four Agrilus beetles in 24 hours. ■Hole in bark made by adult Agrilus in emerging from pupal cell. Lar\'£e of Agrilus bilitieatus and their burrows. Complete burrow of a larva of Agrilus biliiieaius. A, Point at which larv-a B, beginning of second instar; C, beginning of third instar; D, beginning of fourth instar: E. pupal cell. EFFECT OF DILUTION UPON THE INFECTIVITY OF THE VIRUS OF THE MOSAIC DISEASE OF TOBACCO By H. A. Allard, Assistant Physiologist, Tobacco and Plant- Nutrition Investigations, Bureau of Plant Industry In order to obtain some idea concerning the effect of dilution upon the infective power of the virus of the mosaic disease of tobacco in subsequent inoculations, the following experiments were made. A quantity of expressed sap from mosaic-diseased leaves was first passed through filter paper to remove the cell tissue, etc. Clean tap water was then used to bring the filtered virus to the required degree of dilution. All dilutions were accurately determined and inoculations immediately made from these. Young, vigorous plants growing in 3-inch pots in the greenhouse were used in all tests. In order to insure a thorough test of the infectivity of the diluted virus, a drop of the solution carried on the point of the needle was introduced with each puncture. Every leaf of any size on the plants, usually four or five, was inoculated in this manner at several points. The plants were kept under observation for a long period after the first appearance of the disease in those groups treated with the original undiluted virus and the lower dilutions. This is virtually a quarantine period for the disease, since experience has shown that the incubation period of the mosaic disease is very uniform for simultaneous inoculations under any given set of conditions. A complete tabulation of all dilution experiments is given in Table I. Table I. -Effect of dilution upon the infectivity of the virus of the mosaic disease of tobacco Date of in- oculation.a Number of plants. Variety. Degree of dilution. Effect. 1913- May 6 Do 10 10 10 10 10 10 20 20 30 10 " All dil Connecticut Broadleaf do Original undiluted virus , . 1 part virusto too of water. I part virus to i.ooo of water. 1 part virus to lo.ooo of water. 1 part vims to i.ooo.ooo of water. 6 mosaic on June i. 7 mosaic on June i. Do. Do do Do do 4 mosaic on June i. All healthy on June i. Do. Do do Do do do OriRinal undiluted virus . . I part virus to i.ooo of water. I part virus to i.ooo.ooo of water. Tap water alone . 13 mosaic on May 26. 10 mosaic on May 26. 1 mosaic on May 26. All healthy on May 26. were made. Do . ..do Do do Do .. ..do utions were prepared on the same day the inoculations i Journal o( A Dcpt. of Ag KriciiKura riculturc. \ Research, Vashington, D. C. Vol. III. No. 4 Jan. IS. 1915 G— 38 (=95) .*A 296 Journal of Agricultural Research Vol. in. No. 4 Table I. — Effect of dilution upon the infeciivity of the virus of the mosaic disease of tobacco — Continued Date of in- oculation. Number of plants. Variety. Degree of dilution. 1913. May 29. , Do... Do... J une 1 2 . Do... Do... Do... Do... June 21. Do... June 24., Do... Do. Do. Do. Do. I9M. April 2 . . Do... Do. Do. Do. Do. Do. Do.. April 3. Do.. Do. Do. Do. Do.. Do,. April 9 Do.. Do. Do. Do. Do. Do. Do. Do. Connecticut Broadleaf . ....do .do. .do. .do. ...do, ....do. 20 Cuban 10 Connecticut Broadleaf, 15 Cuban 15 Connecticut Broadleaf. . 18 Cuban 15 Connecticut Broadleaf. . /lo Cuban Vio Connecticut Broadleaf . . Cuban do >Original undiluted virus. . . .do., .do., .do., .do., Cotinecticut Broadleaf., do .do., .do., .do.. Nicotiana rustics. do Connecticut Broadleaf., do ..do Nicotiana rustica.. ..do .do.. .do., .do.. Connecticut Broadleaf.. ....do Original undiluted virus . . I part virus to 1,000 of water. I part virus to 1,000,000 of water. Tap water alone Original undiluted virus . . I part virus to 1,000 of water. I part virus to 10,000 of water. I part vims to 1.000,000 of water. Tap water only part virus to 1.000 of water. 1 part virus to 10,000 of water. J- Tap water only Ori^nnal undiluted virus. . . I part virus to 1.000 of water. I part virus to 10,000 of water. I part virus to 100,000 of water. Tap water only Not inoculated .do., .do., -do.. .do.\ .do.. ..do. .do.. Original undiluted virus. . . I part \'irus to 1.000 of water. I part virus lo 10.000 of water I part virus to 100,000 of water. I part virus to 1,000.000 of water. Original undiluted virus. . . I part virus to i.ooo of water. Tap water alone Original undiluted virus. . , I part virus to 1,000 of water. I part virus to 1,000.000 of water. Tap water alone. , Original undiluted virus. . , I part virus to 1,000 of water. do Tap water alone Original undiluted virus. . 1 part virus to 1,000 of water. 1 part virus to 5,000 of water. I part virus to 10,000 of water. I part virus to 20.000 of water. I part virus to 50.000 of water. I part virus to 100,000 of water. I part virus to 200.000 of water. I part virus to 500,000 of water. II mosaic on June 5. 17 mosaic on June 5. All healthy on June 5. Do. 11 mosaic on June 18. 8 mosaic on June 18. 7 mosaic on June 18. 1 mosaic on June 18. All healthy on June iS. 26 mosaic on July 6 (18 Cuban, S Connecticut Broadleaf). 20 mosaic on July 6 (11 Cuban, 9 Connecticut [ Broadleaf). 12 mosaic on July 6 (4 Cuban. 8 Connecticut Broadleaf). All healthy on July 6. T2 mosaic on July 6. 19 mosaic on July 6. 10 mosaic on July 6. 2 mosaic on July 6. All healthy on July 6. Do. 6 mosaic on April 17. 2 mosaic on April 17. 1 mosaic on April 17. Do. All healthy on April 17. 4 mosaic on April 17. Do. All healthy. 3 mosaic (m April 17. 5 mosaic on April 17. All healthy on April 17. Do. 8 mosaic on April 17. 7 mosaic on April 17. 2 mosaic on April 17. All healthy on April 17. 9 mosaic on April 25. 7 mosaic on April 25. 2 mosaic on April 25. Do. All healthy on April 25. I mosaic on April 25. All healthy on April 25. Do. Do. Jan. IS, "915 Mosaic Disease of Tobacco 297 Table I. — Effect of dilution uj>on the infcctivily of the virus of the mosaic disease of (ofcocco— Continued Date of in- oculation. Number of plants. Variety. Degree of dilution. Effect. 1914- April 9 10 10 10 10 10 10 10 10 10 10 10 10 10 Connecticut Broadleaf .... do I part virus to i. 000.000 of water. All healthy on April 25. Do. Do do N(»t inoculated Do. Do Do Maryland Mammoth do Original undiluted virus. . . I part virus to 1,000 of water. I part virus to 5.000 of water. I part virus to 10.000 of water. I part virus to 20.000 of water. I part virus to 50.000 of water. I part virus to 100,000 of water. I part virus to 200,000 of water. I part virus to 500,000 of water. I part virus to 1.000.000 of water. 9 mosaic on April 25. 6 mosaic on April 25. Do do 5 mosaic on April 25. Do do 2 mosaic on April 25. Do do All healthy on April 25. Do Do Do Do do... do.., do . ...do Do. Do. Do. Do. Do do Do. These tests show beyond question that the virus of the mosaic disease when diluted to i part in i ,000 of water is quite as efifective in producing infection as the original undiluted virus. A dilution of i part in 10,000, however, gives evidence af attenuation. At greater dilutions than this the chances of infection are very greatly reduced. In dilution experi- ments of this sort, where increasing attenuation of the virus is taking place, it is obvious that no sharp line of demarkation can be found beyond which chances of infection do not exist. Since the solutions are punc- tured into the leaves with a sharp needle, the quantity of virus taken up and actually introduced into the plant tissues must be exceedingly small, especially for the higher dilutions. It is of interest to consider this fact in connection with the enzymic theory of the mosaic disease, which has been advanced to explain the nature of the disease. This theory assumes that the mosaic disease develops when certain oxidizing enzyms normally present in the plant increase in amount or in activity as a result of various external conditions affecting nutrition and growth. It is somewhat difficult to reconcile this theory with the fact that a tiny drop of \nrus diluted to i part in 10,000 can readily produce the mosaic disease. It must be assumed that the immeasurably small quantity of oxidizing enzyms carried by this drop is sufficient to increase the normal oxidase content already present in the plant to the extent of a permanent pathological reaction resulting in the mosaic disease. This is highly im- probable, since it is well known that the oxidase content of healthy indi- viduals normally varies to a measurable degree in response to various en- vironmental changes. G0733°— 15 2 298 Journal of Agricultural Research voi. iii. No. 4 There seems to be no logical reason for considering that something ex- ists in the constitution of all normal tobacco plants which is always capable of producing the mosaic disease in response to suitable condi- tions. This conception does not harmonize with the fact that even when the virus of the mosaic disease is highly diluted and the infective sub- stance becomes immeasurably small it is still capable of initiating the disease when introduced into healthy plants. All evidence at hand points to something in the virus quite extraneous to the protoplasmic constitu- tion of healthy plants. Once introduced into the tissues of such plants, this foreign substance rapidly increases in quantity and becomes actively prejudicial to those physiological activities associated with normal nutri- tion and growth. In the opinion of Woods and Heintzel this substance constituting the active, pathogenic principle of the virus of the mosiac disease may be re garded as purely chemical, nonliving, enzymic, and a normal constituent of all healthy tobacco plants. According to Hunger, on the other hand, the disease is caused by a toxic ferment not normally present in the cells of healthy plants, but which develops in response to unfavorable condi- tions of nutrition and growth. These theories are in complete agreement in ascribing to the mosaic disease a spontaneous origin within susceptible plants under favorable conditions. The development of this conception is quite natural if a spontaneous origin is accepted, since this does not admit of a consistent explanation on the basis of parasitism. At the time these theories were evolved, little was known concerning organisms which are smaller than the visible bacteria and yet responsible for parasitic dis- eases. It was known that the visible bacteria could not pass through the pores of certain filters. It was also discovered that passing the virus of certain diseases through these filters did not necessarily deprive it of its power to infect, although visible parasites were no longer present. At this time this characteristic seemed to remove those diseases connected with a filterable virus from that class of diseases definitely established as bacterial in their origin. Until additional facts had been secured an enzymic origin was perhaps the most plausible explanation for those obscure diseases connected with a filterable virus and supposedly capable of a spontaneous origin within certain plants. The writer's experiments, however, indicate that the mosaic disease can not be induced to arise spontaneously in healthy plants by the opera- tions of cutting back, repotting, or otherwise subjecting the plants to un- favorable conditions. SUMMARY The virus of the mosaic disease when diluted to i part in i ,000 of water is quite as effective in producing infection as the original undiluted virus. Attenuation of the virus is indicated in dilutions of i part in lo.cxx) of water. At greater dilutions infection is not likely to occur. Jan. 15, 191S Mosaic Disease of Tobacco 299 The virus of the mosaic disease is highly infectious to all susceptible, healthy plants. Such plants remain free from this disease so long as all chances of accidental infection are excluded. All evidence at hand in- dicates that something is present in the virus of the mosaic disease which is extraneous to the protoplasmic organization of healthy plants. This substance greatly increases in quantity when introduced into susceptible plants and interferes with normal nutrition and growth. Although enzymic activities have been considered responsible for the mosaic disease of tobacco, parasitism, in the writer's opinion, offers by fai the simplest and most reasonable explanation of its origin. It may at least be said that the theory of a parasitic origin for the disease more con- sistently accounts for all the facts at hand than any enzymic conception yet evolved. It seems not only needless but illogical to abandon a simple, direct explanation for one which leads to complexity of thought and yet fails to correlate all the facts at hand. MOLDINESS IN BUTTER By Charles Thom, Mycologist, and R. H. Shaw, Chemist, Dairy Division, Bureau of Animal Industry INTRODUCTION References to mold in butter are not uncommon in dairy literature, but specific information is lacking as to what forms of mold occur on butter, the conditions which permit mold development, and the actual changes produced in the butter. As met in the market, losses from mold take two forms: (i) The growth of mold upon the tub, lining, or wrapper injures the appearance and salability of the package with- out seriously affecting the quality of its contents. (2) Mold develop- ment upon the butter itself when continued for a considerable period produces changes which can not be eliminated even by the renovation process. Such butter becomes an actual loss. The work reported here aims to cover the biological phases of this problem. The study of the chemical changes produced in the butter will be reported later. ORIGIN OF BUTTER SAMPLES Characteristic samples representing the range of conditions and appearances found in commercial butter were obtained through the inspection service of the Dairy Division. These were examined in the mycological laboratory. The number and variety of mold colonies upon each sample were noted and cultures were made to obtain the species represented. The samples were then taken to the chemical laboratory for analysis. Consideration of the known factors influenc- ing mold growth called for the determination of the quantities of water and protein available, together with the percentage of salt as a possible limiting factor. In Table I are given the analyses of samples of moldy butter from several sources. Table I. — Analyses of samples of moldy butter Sample No. 3529- 3530- • 3554g 3554r. Water. Salt. Curd." Per ctnt. Per cent. Per cent. 11.6s I. q8 1.48 12. 00 .65 i-.=!3 9.40 2. 19 .64 II. 09 2.66 .66 11.72 2. 13 .64 Sample No. 3S54S. 3546a 3546b 3546c, Water. Salt. Per cent. Per cent. 7-38 0.63 16. 02 3-05 18. 00 I. 40 16. 02 3-50 Curd .a Per tcnt. o 57 .68 •75 .6a " The term "curd," as used in this paper, means the amount of nitrogen multiplied by the factor 6.38. Journal of Agricultural Research, Dept. of Agriculture, Washington, D. C. (301) Vol. Ill, No. 4 Jan. IS, 1915 A— 12 302 Journal of Agricultural Research voi. iii. No. 4 Samples Nos. 3546a, 3546b, and 3546c were taken from a tub of butter containing three small chumings. The tub had been kept in a, refrigerator for three or four months and was very rancid. Since the butter was designed for packing stock to be used in experimental work on renovated butter, no particular care was taken in its manufacture. It happened that the top layer (3546a) and bottom layer (3546c) were heavily salted, while the middle layer (3546b) contained but a small percentage of salt and a considerably higher percentage of water. The top and bottom layers were free from mold; the middle layer showed areas typically representing each of the types of moldiness described in the following pages. TYPE-S OF MOLD FOUND IN SAMPLES From the study of these and other available samples of moldy butler three well-marked types of mold effects are distinguished: I. Smudged, or Alternaria, type. — In samples Nos. 3515, 3546b, and 3554s, dark, smoky, or rarely greenish colors occurred in patches which suggested soot or dirty-finger marks. Microscopic examination showed mold mycelium with dark-brown or green walls on or under the surface. Frequently these colonies are entirely submerged in the butter. Sometimes hyphae were observed 4 to 5 mm. below the surface. Spores were rarely found, but colonies transferred to culture media grew freely and fruited normally. The dark-brown or black hyphee were the most common and proved to be species of Alternaria. Where a greenish color was seen species of Cladosporium developed. These submerged areas suggest the appearances noted by Patterson (1900),' and attributed to Stemphyliiim butyri Patterson. The occurrence of Cladosporium in butter has been studied by Jensen (1900) and the organism found was named by him "Cladosporium butyri." The species of Cladosporium, however, are abundant upon all kinds of roughage fed to cows, and the spores find entrance to the milk from the handling of such feed by the milkers. One of the writers had access to cultures made from many samples of cream by the bacteriologists of the Storrs Agricultural Experiment Station some years ago. In these cultures colonies of Cladosporium were so abundant as to indicate that spores of these species remain with the cream after separation. Species of Alternaria are very common in the same circumstances and appeared in these cultures, but less abundantly. Their appearance as colonies in the butter, therefore, is due to the ability of these species to grow in the very severe conditions imposed by a mass of butler. One of the writers has found a species of Alternaria growing and fruit- ing in a box of shoe paste. Species of this group have also been iso- lated from various forms of fat when small inclusions of water occur. Few other graminicolous fungi seem able to produce colonies under these conditions, though spores of many kinds are undoubtedly present. ' Bibliographic citations in parentheses refer to '' Literature cited." p. 310. Jan. IS, 1915 Moldiness in Butter 303 In at least one sample, contributed by Dr. G. P. Clinton, of the Con- necticut Agricultural Experiment Station, and again in a sample of fat studied by Dr. C. N. McBryde, of the Bureau of Animal Industry, a fungus producing abundant orange to red mycelium and red blotches upon the butter was obtained. In butter and in the culture media used no spores have thus far been obtained. This organism grows under the same conditions as Altemaria. 2. Green-mold type. — Green molds were found more or less frequently upon all the samples tabulated except Nos. 3546a and 3546c. Cultures of these molds proved to be species of Penicillium. Three common species were often found. These were P. roqueforli, a variety or strain of P. expansion, and P. chrysogenum. Several other forms difficult to identify were occasionally obtained. Aside from P. roqueforti, these are identifiable only by careful culture and comparison. These molds form green patches on the surface and follow seams or cracks into the mass. In one tub (No. 3515), where extensive moldy areas were found in cracks and seams, the presence of P. roqueforti was suggested by a strong odor and flavor resembling that of Roquefort cheese. Marked physical changes in the fat itself were noticeable. Culture confirmed the identi- fication of the organism. So far as observed, no such extensive changes were produced by the other species. The storage of butter in tubs is accompanied by low percentages of free oxygen* in the butter sug- gestive of the conditions in Roquefort cheese (Thom and Currie, 1913). Mold is found upon the liners, upon the inside of the tub itself, and in the cracks of the butter. In all these places interchange of gases is very slow, thus favoring the dominance of Roquefort mold, which is more tolerant of such conditions than other species. 3. OiDiuM Type. — The third form produces various shades of orange- yellow discoloration, with little or no surface growth. Culture and microscopic examination show that these areas are produced by Oidium lactis. This organism grows to the depth of several miUimeters within the mass of butter as a complex mycelium with hyphae varying in diam- eter with the size of the spaces between the masses of fat. Some spores are fonned and at times surface-fruiting areas. Bacterial activity is commonly associated with the presence of this mold. Black molds, or mucors. — Where butter has been moist enough for loose masses of surface mycelium to grow, mucors are sometimes seen. These molds are found by culture to be present in many other samples in which no visible colonies are produced. EXPERIMENTAL WORK To study the conditions favorable to mold- growth in butter special samples of butter were prepared, some low in water content and some high in water content, some thoroughly washed to reduce the curd con- ^ Unpublished results of Dr. D. C. Dyer, of Uic Dairy Division. 304 Journal of Agricultural Research Vol. Ill, No. 4 tent and some with casein added to raise the protein content. One ounce of salt to I pound of butter was used in some samples; no salt in others. Slices of butter of each kind were put into Petri dishes and inoculated with a series of molds obtained from butter. Among these were Oidium lactis, Mucor sp., Aliernaria sp., and several species of Penicillium. The dishes were then allowed to stand in an incubator at the temperature and relative humidity of the laboratory for several days. Absolutely no surface growth of mold was obtained. Part of these Petri dishes were then placed in moist chambers and it was found that mold colonies developed upon every sample so placed. These growths included not only the species inoculated into the butter but other forms whose spores were present in the butter as made. At the low relative humidities pre- vailing in the laboratories of the Dairy Division from February to April^ 1914, no mold colonies were able to develop in butter representing a range in water content greater than the usual range of percentage in market butter. The addition of 2 to 3 per cent of water to butter containing but 14 or 15 per cent does not make the quantity of water present sufficient to support mold growth aside from conditions of high humidity. REL.^TION OF HUMIDITY TO MOLD GROWTH In moist-chamber culture comparison between samples containing normal and low protein with samples containing excess or added pro- tein showed that mold growth was more rapid and extensive when pro- tein was added. The failure of molds to grow in these same cultures under the ordinary humidity conditions of the laboratory proved that the essential factor in molding was not protein, but water. To define these humidity relations more closely, three desiccators were prepared in which definite relative humidities could be maintained. For this purpose the bases of the desiccators were filled with sulphuric acid standardized to the specific gravities — from Hastings's (1909) table — required to maintain, respectively, 90 per cent, 79.6 per cent, and 69.6 per cent relative humidity. Three samples were used: One sample of butter was made with low-salt content (0.55 per cent); one at normal salting (2.43 per cent) ; and one sample of butter fat, free from water, with skim-milk powder added. The composition of these three samples is given in Table II. Table II. — Composition of samples of btillcr used in mold-growth tests Character of sample. Water. Salt. Curd. Normal-salt butter Per cent. 15.6 None. Per cent, 2.43 None. Per cent. 0. 62 .62 Butter fat -|- dry skim mi Ik .48 Jan. IS, 1915 Moldiness in Butter 305 Three slices, one from each of these samples, were put into each one of a series of 24 Petri dishes. The three sUces in each dish were thus under absolutely the same conditions. Six species of mold were then selected and were heavily inoculated into the plated slices— each slice in four Petri dishes being inoculated with one species of mold. Four sets of sLx dishes each were thus available. One set of six Petri dishes was put into a moist chamber (approximately 100 per cent of relative humidity), and one each into the desiccators at 90, 79.6, and 69.6 per cent relative humidity. The cultural results are given in Table III. Table HI. — Effects of salt and humidity on mold growth in butter" Growth in moist GroTi'th under relative humidities of— chamber. 90.6 per cent. 79.6 per cent- 69.6 per cent. Mold. Salted butter. But- ter fat and curd. Salted butter. But- ter fat and curd. Salted butter. But- ter fat and curd. Salted butter. But- ter 0.55 per cent. 2-43 per cent. 0.55 per cent. 2-43 per cent. O.S5 per cent. 2-43 per cent. 0-55 per cent. 2-43 per cent. fat and curd. 0.9 60.6 •5 -5 -5 0 0 0 - 1(?) 0 0 0 0 .8 0 0.3 0 •5 •5 •4 ■3 0 0 0 .2 .2 0 0 0 0 0 ■3 0 0.6 0 •3 •3 0 0 0 0 0 0 0 0 0 0 . 2 0 0 0 •3 0 0 0 0 0 0 0 Penicillium roqueforti Penicillium chrysogenum. . Penicillium expansum 0 0 0 o A typical colony would be designated as i.o; lesser growths by decimal fractions, ft Submerged. Examination of this table shows that a single species, Penicillium chrysogenum, was able to produce a colony upon the butter fat plus the water obtainable from the air. Careful examination of the other samples showed no mold. In the low-salted butter, however, with 15.6 per cent of water marked growth occurs; the water in this butter is therefore to be regarded as an essential factor in the molding found here. The butter containing 2.43 per cent of salt shows determinable growth from but two species, P. roqueforti and P. chrysogenum. No growth of species of Alter- naria, Oidium, or Mucor was found upon this butter. The low-salted butter shows very appreciable mold colonies of all species except the Mucor. Growth was greatest in the moist chamber. Nearly as good growth was obtained, however, with a relative humidity of 90.6 per cent, and considerable growth in four of the species with 79.6 per cent. In the presence of 69.6 per cent there was very little visible mold, even in the low-salted sample. The individual samples of low-salted butter all showed the characteristic orange-yellow colors due to the development of spores of Oidium laciis, which were evidently present from the first in all slices. In this experiment the organism grew only in its submerged form; hence, it was little affected by the relative humidity to which the other species responded so clearly. Alternaria and Oidium developed 3o6 Journal of Agricultural Research voi. iii.no.4 only in the low-salted samples. Alternaria appeared in several places without inoculation. Rhizopus nigricans was found once in a moist- chamber sample. Aspergillus fumigatus and A. nigcr both appeared in one or more cases, but none of these species appeared where the per- centage of salt was 2.43. The same fact is illustrated on a larger scale by the tub of butter analyzed as No. 3546 in Table I. Of the three samples packed together in one tub the middle layer was low-salted and typically moldy, while the top and bottom layers were free from mold. These results harmonize fully with the data from the analysis of butter as found in Table I and with the preliminary cultural data as given in subsequent experiments. Two of the three types of moldiness, the smudged and the orange-yellow forms, occur only in butter containing less than 2 per cent of salt. Even with green molds under high humi- dities and at temperatures far above those used in storage, growth in these experiments was negligible in butter with a salt content of 2.43 per cent. THE SALT FACTOR IN MOLD GROWTH The salt factor in butter is calculated as follows: Thompson, Shaw, and Norton (191 2), in analyzing 695 samples of American creamery but- ter, found an average water content of 13.9 per cent; salt content, 2.51 per cent; and curd content, 1.18 per cent. This amount of salt in solu- tion in the water present forms, therefore, approximately a 13 per cent brine, which represents the brine formed by adding 18 parts of salt to 100 parts of water. If the same water content be assumed and the salt content found be i per cent, the brine present is 5.1 per cent (made by adding 7.1 parts of salt to 100 parts of water) ; with a salt content of 2 per cent, this strength would be 10.2 per cent; with 3, 15.3 per cent; and with 4, 20.4 per cent. For purposes of mold growth the strength of the brine found is one very significant factor. Another factor is repre- sented by the distribution of air and moisture throughout the mass of butter, and still a third by the relative humidities to which the butter is subjected. To obtain more complete cultural data for comparison, a series of cultures was made with media containing known percentages of salt. For this purpose 6.5 per cent of salt was introduced into one lot of Czapek's agar (Dox, 1910) and 14 per cent in a second lot. The first represents approximately the proportion of brine in butter with 1.3 per cent of salt, the second the brine with a content of about 2.8 per cent of salt. These cultures were grown in a moist chamber to eliminate the concentration of the brine by drying. The cultural results are given in Table IV. Jail. 15, 191S Moldiness in Butter 307 Table IV. — Cultural results with media containing known percentages of salta Altcmaria sp. 3515 Altemaria sp. 3513 Altemaria sp. 3546 Cunninghamiella sp. . Fusarium sp Mucor sp. 3513 Mucor sp. 3514D4 Mucor sp. 3514C1 . . . . Mucor sp. 3532 Oidium lactis Penicillium sp. 3529a. Penicillium roqueforti Percentage of salt. 6-5 14.4 0.7 0. 2 • / . 2 •7 . 2 .6 0 ■4 0 •7 0 .6 0 .6 0 ■ 5 0 . I 0 I. 0 * I. 0 • 9 6.4 Mold. Penicillium roquefotti 3SisC Penicillium expansum Penicillium stolonife- rum, var Penicillium chrysoge- num Penicillium purpuro- genum Red mold 3536.3 Rhizopus nigricans. . . . Trichoderma sp Percentage of salt. 6.S I. o ■9 ■3 • 7 •3 1) 1. o ■ 4 o A typical colony is designated as i.o; lesser growth by decimal fractions. b These cultures developed slowly, but finally reached the condition indicated. These cultural results agree with other data published recently (Thom, 1914). Two more series of cultures were made, containing approxi- mately 18 and 21 per cent of salt. In these such organisms as produced marked growth with a salt content of 14.4 per cent were carried, together with other species of Penicillium and Aspergillus. Penicillium chryso- genum, P. stoloniferum, Penicillium sp. 3529a and Aspergillus repens pro- duced considerable growth with 18 per cent of salt. Three other organ- isms produced slight growth. With 21 per cent of salt no colonies were obtained, although spores of P. chrysogenum germinated. Comparison with the results of butter inoculation shows that Czapek's solution sustained much larger growth than butter containing comparable per- centages of salt. To show the results of these culture series for the organisms obtained from butter, the graphic representation (fig. i) was prepared. The four series reported were calculated as representing approximately brine conditions in butter containing 1.3, 2.7, 3.4, and 4.1 per cent of salt. Even under the very favorable conditions offered by the culture media, temperature, and humidity used, the mold growth found in the second series was small and in the third series was negligible. DISCUSSION OF RESULTS From the data already given, mold is seen to attack the butter itself if unsalted or very lightly salted. Normally salted butter may be affected by green mold only if held under conditions very favorable to mold growth. In general such losses are not great. Both the species of Oidium with its orange-yellow patches and the smudges of Altemaria disappear promptly when even very moderate salting is practiced. These are the important factors in losses of unsalted butter as studied by Jensen (1901 and 1908). Since Oidium sp. penetrates the mass of butter 3o8 Journal of Agricultural Research Vol. m. No. 4 and produces marked discoloration, as well as bad flavors, salting, if practiced at all, should be heavy enough to eliminate this group of organisms. Green molds may damage normally salted butter if cracks and open spaces are left by bad packing. In most cases such mold will be confined to liners and containers if the packing is fairly well done. Rogers (1906) found that paraffining the tubs or boxes used prevented mold on both container and liner. The paraffin prevented the escape of water which would leave the air spaces necessary for mold growth, thus 35'S C,3S^/. tv 3Z 3539 a /.34 2.7^ 3.3 e ^.oe s 10 30 Fig. I.— Graph showing the effect of salt on molding. Cultural results with organisms obtained from butter. Nos. 35150, 3541W27, 3929a, 3515.12, 3515.18. and 3541.26 were species of Peniciliium. N0.3S14C1 was a species of Mucor and No. 3536.3 was the sterile red mold from butter. The dotted portions of the graph represent hypothetical courses for organisms disappearing at percentages not determined but limited by the next experiment. preventing also loss of weight from the butter itself. Previous papers have taken no account of the presence of mold spores in the butter itself. All possible treatment of containers will fail unless conditions are pro- duced which will prevent the growth of these spores. The same condi- tions which stop the growth of molds present on the paper and wood of the package also prevent the spores in the butter from growing. In all storage of butter the temperature factor must not be neglected. Mold growth is progressively reduced by low temperatures. Work else- where reported shows that species of Peniciliium (Thom, 1910, p. 92, 93, 105) grow very slowly as the temperature approaches the freezing point. Jan. 15, 1915 Moldiness in Butter 309 If, as in butter, the fluid present is a strong brine, the temperature must be actually carried considerably below the freezing point of water to elimi- nate danger from the growth of micro-organisms. Temperatures a few degrees above freezing accompanied, as they frequently are, by moist con- ditions are favorable to molding in butter. Unsalted butter is more sub- ject to deterioration from micro-organisms than salted butter. Success- ful storage of such butter is therefore even more dependent upon scrupu- lously clean dry refrigeration at low temperatures than is the case with salted butter. Cellars and ice refrigeration rarely furnish conditions which will prevent mold growth in unsalted or low-salted butter, although such growth may be delayed or reduced. Butter properly made and salted normally, as indicated above, will not show mold under reasonably careful handUng. SUMMARY (1) Mold in butter usually takes three forms: a. Orange-yellow areas with a submerged growth of mycelium are produced by Oidiuni lactis. h. Smudged or dirty-green areas either entirely submerged or with some surface growth are produced by species of Alternaria and Clado- sporium. c. Green surface colonies arc produced by species of Penicillium, or, more rarely, Aspergillus, either upon the butter, causing decomposition, or upon the container or wrappings, injuring the appearance of the sample in the market. (2) Species of Oidium, Alternaria, and Cladosporium can not develop in butter containing 2.5 per cent of salt. The occurrence of any of these forms in a sample of butter indicates low salting. (3) Excess of curd favors mold growth. Well-washed butter is less subject to mold. (4) lycaky butter — butter from which water of buttermilk exudes and collects in the wrappings or in the container — ^furnishes the best condi- tions for the beginning of mold growth. From these wet areas colonies may spread to the butter itself. (5) Wet surfaces, wet wrappings, or high humidity are essential to mold growth in butter. Mold will not grow upon the surface of a piece of butter exposed to humidities of 70 per cent or lower. The water in the butter is thus not sufficiently available to the mold to support the de- velopment of a colony, unless evaporation is reduced by high humidities. In closed packages, wet or damp cellars, or carelessly packed masses with cracks or fissures in which moisture collects, mold may seriously injure the appearance of butter packages or actually induce great changes in the butter itself. (6) Salt up to 2.5 to 3 per cent in butter is sufficient to eliminate mold or reduce it to negligible amount. This is equivalent to the use of a 12 to 1 5 per cent brine. 3IO Journal of Agricvltural Research voi. m, no. * LITERATURE CITED Dox, A. W. 1910. Intracellular enzyms of Penicillium and Aspergillus, with special reference to those of Penicillium camemberti. U. S. Dept. Agr., Biu". Anim. Indus. Bui. 120, p. 37. Hastings, M. M. 1909. Cold-storage evaporimeter. U. S. Dept. Agr., Bur. Anim. Indus. Circ. 149, p. 6. Jensen, Orla. 1901. Studien iiber das Ranzigwerden der Butter. In Landw. Jahrb. Schweiz, V. 15. P- 337- Patterson, Flora W. 1900. New species of fungi. In Bui. Torrey Bot. Club. v. 27, no. 5, p. 285. Rogers, L. A. 1906. Investigations in the manufactxire and storage of butter. II. — Preventing molds in butter tubs. U. S. Dept. Agr., Bur. Anim. Indus. Bui. 89, 7 p., I fig. 1908. ParafBnLng butter tubs. U. S. Dept. Agr., Bur. Anim. Indus. Circ. 130, 6 p., I fig. Thom, Charles. 1910. Cultural studies of species of Penicillium. U. S. Dept. .\gr., Bur. Anim. Indus. Bui. 118, p. 92, 93, 105. 1914. The salt factor in mold-ripened cheeses. In Storrs .\gr. Exp. Sta. Bui. 79, p. 387-394. and CurriE, J. N. 1913. The dominance of Roquefort mold in cheese. In Jour. Biol. Chem.. v. 15, no. 2, p. 249-258. Thompson, S. C, Shaw, R. H., and Norton, R. P. 1912. The normal composition of American creamery butter. U. S. Dept. Agr., Bur. Anim. Indus. Bui. 149, p. 10. SUSCEPTIBILITY OF CITROUS FRUITS TO THE ATTACK OF THE MEDITERRANEAN FRUIT FLY By E. A. Back, Entomological Assistant, and C. E. Pemberton, Scientific Assistant, Mediterranean Fruit-Fly Investigations, Bureau of Entomology INTRODUCTION Since the discovery in 1910 that the Mediterranean fruit fly (Ceratitis capitata Wied.) had become established in the Hawaiian Islands, the fruit growers, and especially the citrous fruit growers, of the mainland States have increasingly feared that this dreaded pest would be able to gain access to the mainland on some one of the many ships plying between Honolulu and the Pacific coast and would appear in the citrous orchards of California and Florida. In addition to this danger from the Pacific, there have been similar fears regarding imported fruits from the Bermudas and the Mediterranean regions. While investigations carried on by the Federal Horticultural Board have shown that the opportunity for entry and establishment of the fly from these trans-Atlantic countries is verv slight, there remains the ever-present danger that sooner or later this pest will reach the mainland from the Pacific, in spite of the increasingly rigid quarantine of Hawaiian host fruits. It is therefore opportune to record data secured in the Hawaiian Islands which tend to show that even if this fruit fly should obtain a foothold in the warmer portions of the United States, it probably would not be the serious pest to citrous fruits that previously published literature would indicate. HISTORICAL REVIEW This literature has been full of references to the havoc caused to citrous fruits by the Mediterranean fruit fly. The first published reference is by Latreille, who states (181 7) ' on the authority of Cattoire that the colonists of Mauritius could with difficulty obtain citrous fruits sound at maturity, on account of the attacks of a dipterous insect that deposited eggs in the fruit. MacLeay (1829) writes of this pest as an insect very destructive to oranges and states that fully one-third of the oranges arriv- ing in London from the Azores were in a decayed condition as a result of the attacks of this pest. He also secured the insect from citrous fruits in Madeira and the Cape Verde Islands. F. DeBreme (1842) speaks of this fruit fly as a pest to oranges near Malaga, Spain; and Westwood (1848), under the caption "The Orange Fly," mentions securing specimens from ' Bibliographic citations in parentheses refer to " Literature dted." p. 330. Journal of Agricultural Research. Vol. III. No. 4 Dept. of Agriculture, Washington, D. C. Jan. 15. 1915 K-13 (3") 312 Journal of Agricultural Research voi. in. No. 4 decayed oranges received at London from St. Michael. Villeneuve (1859) exhibited before the Entomological Society of France an infested orange from Algeria, and Laboulbene (1871) describes the injuries caused by the fruit fly to oranges in Algeria and quotes from notes furnished him by Boisduval to the effect that at Bildah and in all Algeria the orange crop was completely destroyed by the insect. On the other hand, Rondani (1870) writes that the species is rare in Spain and is found in Italy only in the southern part. While the purpose of this article is not to record the Hterature of this fruit fly, these few references are sufficient to show that much of the early literature greatly emphasizes the destructiveness of the Mediterranean fruit fly to citrous fruits and has laid little stress upon other fruits more susceptible to attack. It is also interesting to note that much of this older literature, which has been generously copied by later writers, records damage to citrous crops grown in very equable climates and in localities where presumably, as in the Hawaiian Islands, there are many host fruits whose commercial value was so small that they escaped the notice of these writers, who judged of the seriousness of the pest by the fruits arriving at their home markets or from common reports. It is also very possible, with our more exact knowledge of the causes of the decay of fruit in transit and of the wholesale shedding of citrous fruits in the field, due to several fungous diseases, to question the reliability of some of the earlier statements. HOST FRUITS Apparently the first obser\-er who did not entirely agree with Mac- Leay's statement that whenever a puncture is found in the rind of the orange "there is a worm concealed in the interior" is Laboulbene, who said that when he compared his observations on the damage done to oranges by the Mediterranean fruit fly with those recorded by others he found certain contradictory facts which needed further investigation. These contradictory facts, although Laboulbene did not know it, were concerned with what has been determined by the writers as an excessive mortality occurring among the eggs and lar\'se of the Mediterranean fruit fly in the orange rind. This mortality, which in the examination of 39 grapefruit that were yellow in color amounted to 99.7 per cent of 7,722 forms, as shown in Table I, will prove a very effective factor in checking this pest in the citrous regions of the United States, especially when com- bined with the climatic and floral characteristics of these citrous regions and the method of growing and harvesting the fruit. Jan. IS. 19>S Citrous Fruits and Mediterranean Fruit Fly 313 Table I. — Results of examinations of ripe citrous fruits infested by the Mediterranean fruit fly "■ Kind of fruit. Punctures. Eggs. Larvae. ber of Alive. Dead. [ruits Not ex- Emp- Nor- Abnor- amin- ed. ty. ty. mal. mal. In In In In In In ture. rag. pulp. ture. rag. pulp. 39 IJ3 3-8 959 5,882 20 I 0 534 325 0 so 380 I8S 093 729 29 0 0 339 15 0 SO ISO 218 34S 187 0 0 25 474 424 0 14 0 44 S 36 0 is; 55 17 38 0 14 48 237 6 38 S 2 3 40S 696 0 17 194 80 196 201 0 5 0 156 280 0 ss 2SI 452 287 397 55 0 17 1,237 1,652 0 2S S7 174 6t>4 1,026 20 9 231 347 59 I 8s I "5 207 286 0 8 383 8 ■4 17 Total num- ber of forms ex- amin- ed. Grapefruit Lemon Lime Shaddock No. x . Shaddock No. 2 . Kusaie lirae Sweet orange Sour orange Chinese orange. . 7,72a 1, 80s 1,455 303 1,155 838 •3,64s 2,337 923 o These examinations were made sufficiently long after the fruits were gathered to permit all eggs to hatch. All eggs recorded in tables are in reality dead, even though certain of them are marked " normal " in appearance. The excessive mortality referred to does not mean that citrous fruits are less attractive to the adult Mediterranean fruit flies or that the fruit of certain species of the Citrus family is not capable of becoming badly infested. Reference to Table I shows that the female fly freely oviposits in grapefruit, lemons, limes, shaddocks, and sweet, sour, and Chinese oranges. Whatever may be the degree of preference shown by the females for other fruits, it is not great enough, at least under Hawaiian conditions, to lead them entirely to ignore citrous fruits, even when these are grown in close proximity to such a favored host fruit as the peach. A study of the data in Table II shows that the female has a much stronger preference for the mango (Mangifera indica) and the ball kamani {CalophyUum inophyUum) than she has for the orange or lemon. While the data are very limited as to the amount and the number of fruits treated, they are indicative of conditions in the field covering a larger range of fruits. In Bermuda during December, 191 3, the senior writer found oranges unaffected while Thevetia and loquats {Eriobotrya japonica) were well infested. Unfortunately for experimental purposes there are in Hawaii no large Citrus orchards free from other host fruits. Instead there are growing a great profusion of host fruits, chiefly in city or suburban districts, which furnish a rapid succession of fruit flies. No matter, therefore, what preference the ovipositing females may show for noncitrous fruits, the flies are present in such large and constantly augmented numbers that the slowly maturing citrous fruits are bound to be attacked. This is especially true during the months of December, January, and February, when a comparatively small number of host fruits other than Citrus are in season. Like conditions also exist at other seasons of the year during the short intervals between the ripening of other host fruits. While many of these host fruits ripen quickly, the citrous fruits, with the exception of the Chinese orange, develop G9733°— 15 3 314 Journal, of Agricultural Research Vul. ni. No. 4 slowly and ofifer themselves for attack over a considerable length of time. Female fruit flies have been seen in Honolulu ovipositing in certain grapefruits and oranges over a period of two or three months. It is not therefore contradictory to the statement that citrous fruits are not the preferred hosts of the fly that we find so large a number of punctures recorded in Table I. Table II. — Host-fruit preference of the Mediterranean fruit Jly Experi ment No. /Orange, ripe but Rveen in color.. \Mango, partially ripe /Grange, partially ripe.. \Mango, partially ripe. . f Orange, partially ripe.. \Mango, partially ripe. . fOrausxe, ripe. \Maugo, ripe. . Combinatiou and conflition of fruits. (Orange, ripe. \Wango. ripe. . ! Lemon, green in color Mango, partially ripe Ball kamani. partially ripe. Lemon, ripe Mango, ripe 3aU kamani, ripe.. [Lemon, partially ripe ^Maugo, partially ripe iBall kamani, partially ripe. [Mango, nearly ripe \Ban kamani. ripe but sound iLemou, begirmiug to turn color. Orange, ripe .Mango, partially ripe ■JBall kamani, partially ripe ■._. . . . IRose-appIe (Caryophyllus jambos) nearly ripe. [Orange, ripe {Rose-apple, ripe. Mango, ripe i Lemon, beginning to turn yellow. Mango, partially ripe Ball kamani, mature but solid Number of punc- tures. Number of eggs. 105 I3S 33 84 ss ri4 54 376 243 24 HABITS OF MEDITERRANEAN FRUIT FLY For those unfamiUar with the Mediterranean fruit fly it may be briefly stated that this pest belongs to the order Diptera and the family Trypetidae. It is one of many species of this family that cause much injury by their attack upon various fruits. In the Plawaiian Islands the fly attacks over 30 different species of fruits and has. caused great financial loss. The adult female, which is about the size of the ordinary house fly, pierces the skin of the host fruit and forms an egg cavity beneath, in which she deposits eggs. The larvse which hatch from these eggs either burrow at once to the center of the fruit, as in the peach {Amygdaliis persica), or may feed in the outer portion, as in the star-apple (Chrysophyllum Jan. IS. 19.5 Citrous Fruits and Mediterranean Fruit Fly 315 cainilo). In either case the frviit is rendered worthless by the developing larvae before it is ripe. When the larvte become well grown, they leave the fruit (PI. XL) either before or after it has fallen and enter the ground or other protected places and transform to the pupal stage, from which the adult later emerges. In Hawaii the Mediterranean fruit fly requires in passing from the egg to the adult stage from i^l4 days in sum- mer to about 47 days during the coldest winter weather. In this paper the words "puncture" and "egg cavity" are often used synonymously. PROPORTION OK EGG PUNCTURES CONTAINING EGGS The data in Table I show that many of the punctures in the rind made by the female contain no eggs. In one of the most favored host fruits, the peach, practically all the punctures made contain eggs. Of 534 punc- tures made in 112 peaches but 13 were empty. The rind of lemon con- tains a much higher percentage of empty punctures than that of any of the other citrous fruits in Hawaii, except the Kusaie lime (Citrus limettd). In the 50 fruits examined 380 empty punctures were found, as compared with 185 with eggs. Practically all punctures in Chinese oranges contain eggs. (See Table I.) In the 85 fruits examined only 1 puncture out of 116 was empty. Grapefnut, or pomelos, shaddocks, and sour oranges seem to be preferred for oviposition to the ordinary budded or seedling oranges. It has been noted that adult fruit flies, especially the males, congregate in large numbers on citrous trees, and in the laboratory both sexes are quickly attracted to pieces of cut rind of citrous fruits. They seem to take pleasure in feeding upon the oils and other substances con- tained in the broken cells, and it is possible that in the field their liking for juices made available by the process of forming the egg cavity is so great that the females discontinue ovipositing and begin feeding. The large percentage of empty punctures in lemons and Kusaie limes, in par- ticular, can not be ascribed to a lack of ripeness, as in practically all in- stances the fruits examined were fully grown and a large percentage were colored and overripe. MORTALITY OF EGGS AND I.ARV^ Although many punctures in citrous fruits may be empty, others con- tain a sufficient number of eggs to infest badly a fruit not so well equipped by nature to withstand attack. Out of 13 punctures in one grapefruit 9 contained 76, 153, 32, 25, 18, 8, 46, 113, and 9 eggs, respec- tively. While this is a larger number of eggs than is usually found in a like number of punctures, it is sufficient when supplemented by the data from other citrous fruits to arouse interest in finding a reason why, with so many eggs deposited in citrous fruits, so very few flies succeed in reaching maturity. (See Table I.) Thirty-nine oranges, cither yellow or orange in color, picked from the trees on September 13, 1913, and containing an average of 32 punctures, with a maximum of 108 and a minimum of 7 3i6 Journal of Agricultural Research Vol. Ill, No. 4 punctures, developed no flies, and their pulp was in a sound though somewhat shrunken condition after they had been held in the laboratory for one month. That there takes place in citrous fruits a very great and pre\'iously unrecorded mortaUty among the eggs and larvae is clearly set forth in the data in Table I. This mortaUty is especially pronounced in grapefruit, lemons, sweet oranges, and Kusaie limes in Hawaii, is less in Hawaiian hmes and sour oranges, and very much less in Chinese oranges. It has been a common belief among many in Hawaii that citrous fruits are too acid to permit thelar\'setoUve in their pulp until ripe, in spite of the contradic- tory evidence that the quite acid Chinese orange is generally infested. The data in Table HI are here given in proof that no citrous fruit, not even the lemon, is too acid for the development of Mediterranean fruit-fly larv'ae. A study of the data shows that there is a high mortality among larvae transferred to citrous fruits. Too much importance, however, should not be placed upon this, as these fruits must be mutilated some- what in the process of transferring the larvae and therefore are more easily attacked by decay fungi, which bring about a condition not espe- cially desirable for the growth of larvae and often positively fatal to their development. The data are of special interest in proving that even first-instar larvte are able to reach maturity in well-grown though green lemons. The percentage of first-instar larvse maturing in green lemons was in several instances even greater than that of larvae maturing in ripe lemons. Table HI. — Development of lamce of Mediterranean fruit fly in citrous fruits Transference of larvae. Instar. Number of larvse. From — To- Date. Trans- ferred. Died. Ma- tured. Feb. 19 First . . ..do. 41 120 ■7 ASa 160 20 120 40 40 60 So 260 60 220 8c 20 120 240 60 60 20 20 40 140 40 60 1 30 40 90 40 5j 120 17 3SI 117 17 no 33 26 33 50 307 48 ■3S 71 '3 67 '37 37 33 13 20 16 6s 28 33 34 30 39 8 Winged kamani (TfTwina- lia catappa). .. .do Hrppti Ipmnn 28 ...do.. Mar. 12 to IS.. Feb. 20 ...do.. Second ...do.. 9« 43 Do Green lemon . do Chinese orange . . do i4toi3.. i8andi9 18 16 ...do.. ...do.. Third. ...do.. 8 Do Do la to 16. . ...do.. 18 ...do.. Do 18 ...do.. Calilomia grapefruit do aoand 21 First.. . do 9 Winged karaani Do 7 do 13 to 21. . 20 to 32. . 30 and 21 21 Second ...do.. Third. ...do.. do Do do WinRcd kamani .do 38 16 First. . ...do.. 8 do ...do.. Do . ... . .do 13 and 13 Second ...do.. 75 Papaya {Carica paPaya) . . do do 16 ...do.. 38 Winged kamani do 13 to 16.. Third. ...do.. 96 do do 16 ...do.. Ball kamani . ..do ...do.. 33 Jan. IS. i9's Citrous Frutts and Mediterranean Fruit Fly 317 The data in the tables make it evident that the cause of the mortality is not the acidity of the fruit. The figures are of interest in showing that the mortality occurs largely in the rind, either among the eggs in the punctures or among the newly hatched larvae in the egg cavity where they hatch or in the rag beneath. The percentages of mortality occurring among eggs and newly hatched larva? are given in Table IV. Table IV. — Mortality of eggs and larva: of Mediterranean f mil fly in the rind Kind of fruit. Grapefruit Lemons Linies Kusaie limes. . . . Sweet oranges. . . Sour oranges. . . . Shaddock No. i . Shaddock No. 2 . Chinese orange. . Total number of forms exam- ined. S, 222 991 1,054 838 3.635 2.357 J. 155 303 >.039 Mortality in percentages in rind. Among eggs. 90.5 76.1 43.1 47.6 19. o 71-7 3-8 ■3-5 3-7 Among larvae. 9-3 21.0 55.6 51.8 79.0 17-3 95-3 i8. I 46.5 Total. 99. s 97.1 98.7 99.4 98.0 S9.0 99.1 31.6 51- » MORTALITY AMONG EGGS As eggs deposited in such host fruits as the peach and loquat hatch with great certainty, the writers were of the opinion that the oil in the oil cells of the rind was an active agent in killing the eggs in citrous fruits. In puncturing the rind in the process of forming the egg cavity the female is likely to drill through one or several oil cells and the oil thus freed, though not of sufficient quantity to drive the female away, is sufficient in many instances to kill aU or many of the eggs deposited. The data in Table V indicate that there is no question that the oil causes the death of the eggs. Only 163 out of 1,600 eggs treated with oil hatched, as compared with 1,313 out of 1,600 eggs held as a check. The eggs under observation were dissected out of punctures in Cahfomia apples and placed on fresh foliage in moist jars. The treated eggs were not sprayed according to the usual method, but by bending over them a portion of the rind of fresh orange (in the first record) or fresh lemon (in the second and third records) so that the oil from the ruptured cells reached the eggs in that fine raistlike spray familiar to all who have eaten freshly gathered oranges. The much larger number of treated eggs that hatched in the last record is accounted for by the writers by their being fully 20 hours older than those in the second lot when treated. It should be stated that eggs removed from their host do not usually all hatch, as some sustain slight injuries and others may be infertile. See Table V. 3i8 Journal of Agricultural Research Vol. Ill, No. 4 TabliJ \'. — Effecl of oil fi urn rind of orange and lemon upon the haUhing of eggs of the Mediterra nean fruit fly Period of (Icpositing egg.s. 1.30 p. m., Mar. 23. to 9 a. m.. Mar. 23 9 a, m. to I p. m.. Mar. 37 9a. m., Mar. 27,10 9 a.m.. Mar. 28. ■ . Total Treated with oil. 800 400 400 800 400 400 Number of esKs hatched. Treated. Check, 3 119 609 342 364 Further evidence that the oil in the ruptured cells is the killing agent is the very small mortality among the eggs deposited in the Chinese oranges. Since in this fruit the rind is only about two twenty-fifths of an inch in thickness, the female is compelled to deposit her eggs either through the rind into the pulp or in a position between and parallel to the rind and pulp, but at a distance from the puncture that seems to be a protection from any oil set free by the puncturing process. Of 609 eggs thus deposited between the rind and the pulp 600, or 98.5 per cent, hatched, as determined by an examination of 85 fruits one week after they had been picked. A comparison of the 98.5 per cent hatched in Chinese oranges with the percentage of the mortality among eggs in other citrous fruits emphasizes the part the oil has in causing mortality among eggs. It is also interesting to note in passing that the eggs in Kusaie limes, the rind of which is sufficiently thick so that the eggs are deposited directly beneath tlie puncture, die with great regularity, while the eggs in Hawaiian limes, the rind of which may be sufficiently thin to permit the eggs being deposited as in Chinese oranges or so thick (accord- ing to the indi'vidual tree) that the eggs are laid either in the cavity in the rind or between the rind and pulp but directly beneath the puncture, suffer a degree of mortality between that of eggs deposited in Chinese oranges and Kusaie limes. While in Chinese oranges the eggs deposited between and parallel to the rind and pulp hatch with great regularity, those deposited through the rind into the pulp are subjected to a mortality caused either by exces- sive moisture or lack of air. Eggs thus laid are usually placed beneath the skin covering the pulp, and the fascicle which they compose appears, after the rind has been removed, as a dull white spot that is easily overlooked. Usually no trace of the opening through the skin covering the pulp through which the eggs have been deposited can be found. The eggs appear thoroughly sealed within the pulp. In some few instances the opening is distinct and occasionally an egg is left in it, half in and half out of the pulp. When these openings in the skhi occur, the eggs appear to hatch normally. Egg masses deposited entirely within the pulp may be located externally a few days after oviposition ja.i. .5. 19,5 Citrous Fruits and Mediterranean Fruit Fly 319 by a round white sunken area in the rind which varies in size with the passing of time up to an inch in diameter. Of 560 eggs found in the examination of the abovcymentioned 85 Chinese oranges 471, or 84.1 per cent, were unhatched and dead. This same kind of mortality occurs to a less extent in ordinary Hawaiian limes, but with no regularity, as the rind of these fruits in most instances is so thick that the female can not place her eggs within the pulp. MORTALITY AMONG LARV^ It has been shown that mortality among the eggs occurs in the rind and in the pulp. Lar\'al mortality occurs chiefly during the first instar, either in the egg cavity or in the rag beneath. Though mortality does occur in the pulp to a slight degree, no further notice of it will be taken, as it has little bearing upon the general purpose of this paper. The data in Table I show that of 6,571 larvae recorded as dead but 18 died in the pulp, while 3,166 died in the egg cavities where they hatched, and 3,387 in the rag of the rind. The causes for this mortality of lar\'ae in the rind are threefold : The oil from the ruptured cells, the texture of the walls of the pimcture, and the texture of the rag. In the treatment of the eggs with oil, as recorded in Table V, it was found that of the lar\'se hatching from the 163 eggs out of the lot of 1,600 eggs sprayed all died either before they were entirely out of the eggshell or before they had crawled much more than one-fourth of an inch. They exhibited a general weakness entirely lacking in normal larvae. Larvae hatching from check eggs were normal and crawled actively to all parts of the containing vials. As the oil sprayed on the eggs and foliage on which the eggs rested appeared to have entirely evaporated by the time of hatching, the writers believe that the few larvae that succeeded in emerging from the eggs died from weakness imparted to the developing embryo by the oil with which the eggs were sprayed rather than from the effect of any oil still on the foliage with which they came in contact on hatching. Subsequent experiments have shown this supposition to have been correct. The writers believe, therefore, that the very large per- centage of the deaths among newly hatched larva: occurring in the egg cavity is the result of the action of the oil liberated during the formation of the cavity — oil which is sufficiently abundant to weaken the developing embryo but not abundant enough to kill the egg. To such weakened larva; and probably to many other normal larvae hatching in egg cavities made without the rupture of oil cells the texture of the walls of the cavity present another difficulty. In many host fruits, such as the peach and loquat, the eggs are crowded into and completely fill the cavity made by the female, but shortly after oviposition, possibly as the result of the action of a fluid introduced by the female with the eggs, the flesh of the host slirinks considerably from the eggs, thus usually 320 Journal of Agricultural Research Vol. III. No. 4 Fig. I. — Cross section of peach, showing egg cavity of the Mediterranean fruit fly with eggs. Drawing made directly after oviposi- tion. Original. leaving the eggs well separated and with ample room, making conditions favorable for the newly hatched larvae. (See figs, i and 2.) On the other hand, in citrous fruits no such enlargement of the egg cavity takes place. Instead there occurs a general hardening of the walls of the cavity, and the eggs remain as tightly packed as when deposited. In many instances the cavity walls become almost woody. In fact the egg cavities in all the thicker skinned citrous fruits, such as grape- fruit, lemons, and sweet oranges, re- semble a gall the cavity of which is filled with eggs and the opening more or less clogged with a yellowish substance from the ruptured cells, and in Hawaii still further sealed by exudations of gummy secretions, especially in certain grapefruit, limes, and lemons. These gall-like cavities in the rind do not share in the general withering of the rind that takes place in citrous fruits after they have been picked for some days, but stand out from the general surface as small nodosities. (See fig. 3.) It is usual in host fruits of the fruit fly for the punctured surface to develop a depression. These tliickened and often woody walls of the cavity no doubt offer an obstacle to the lar%'ae reaching the rest of the fruit which they can not overcome; hence, the larvae are forced either to die in the cavity itself or to work their way out through the opening of the puncture to the surface of the fruit. It is probably seldom that lan.'^ leave the fruit by way of the opening of the puncture, but a few newly hatched larvas have been found by the writers with their bodies half way out of the fruit. Larvae that succeed in getting out of the cavity must burrow through the rag before reaching the pulp, and this is a difficult task, as evidenced by the fact that out of 3,345 newly hatched larvae that succeeded in reaching the rag, as shown in Table I,' 3,276, or 97.9 per cent, died in the rag. The larvae, after leaving the egg cavity, burrow in all directions, but seldom get more than i inch from the cavity and usually not that far. Often they are able to reach the skin covering the pulp or to burrow FiG. 2. — Cross section of peach, showing the gen- eral shriveling of the walls of the egg cavity and the separation of the eggs. Drawing made i'4 days after oviposition. Original. ' This number excludes shaddock No. i and sour and Chinese oranges, which are of no commercial value and are more easily infe.ted than grapefruit, lemons, limes, and sweet oranges. Jan. 15. 1915 Citrous Fruits and Mediterranean Fruit Fly 321 down between the sections of the fruit, but seem to be lacking in strength to penetrate the skin after they have reached it. There apparently is nothing in the rag itself as a food to cause the death of the larvae, as larvae can attain full growth when feeding on the rag of certain shaddocks. Whether larvae die or not seems dependent upon the degree of toughness of the rag, and the closeness with which the rag adheres to the skin covering the pulp. The toughest rag found was that of sweet oranges still hanging on the tree in March in a much overripe condition. These fruits had begun to be pithy at the stem end, and the rind, which was more or less russeted, had begun to wither and yielded no oil when sharply bent. These fruits were very much like the over- ripe, badly russeted seedling oranges frequently found on trees in Florida during April and May as "leftovers" from the winter crop. In the 20 fruits examined, containing an average of 7 punc- tures to the fruit, no larva was able to penetrate the rag. The coarsest rag or that with the loosest texture is that found in certain large shaddocks growing in Hilo, Hawaii (PI. XLI). These fruits were much overripe when gathered from the tree in March. An examination of 14 of these fruit- showed that out of 245 larvae, mostly in the third instar, present in the rag and pulp, 152 were alive in the rag, 55 alive in the pulp, and only 38 dead, all in the rag. Fourteen other shaddocks, apparently in the same state of ripeness but growing on another tree and so very much undersized as to resemble a medium-sized grapefruit, had a very much tougher rag. An examination of these fruits showed that but 5 very young larvae out of 701 found in the rag were alive and that no larvae had succeeded in penetrating the pulp. The data in Table I show that in the grapefruit, lemons, limes, and sweet oranges examined, 326, 15, 424, and 1,552 first-instar larvae, respectively, died in their attempt to puncture the rag, as compared with but i first-instar larva found alive in the rag of grapefruit and 17 third-instar larvae found in the pulp of sweet oranges. The ordinary sour orange of Hawaii, which is identical wiih that grown in Florida, possesses a loosely attached rind the rag of which is much looser in texture and from the standpoint of imperviousness to the young larvae seems half way between that of the ordinary sweet orange and large well-ripened shaddocks. After the larv-ae have succeeded in passing through the rag of these oranges they work their way between the rag and the skin and finallv enter the pulp, usually at the blossom end. "^m 1 1 Fig. 3. — Section of grapefruit rind, showing two egg cavi- ties, one in cross section. Drawing made one week after fruit was picked. Note conical elevation about tlie eyg cavities left by the withering of the rind; also the thickened walls of the egg cavity and the single larval channel in the rue- Original. 322 Journal of Agrictiltural Research v.4111. no. 4 Hawaiian limes possess less rag than sour oranges and the larvas reach the pulp more easily. In 1,692 ripe and yellow limes picked during April, showing an average of about 5 punctures to the fruit, the larvae succeeded in reaching the pulp in but 287 cases. Chinese oranges have been shown to be generally infested because they possess a very thin, loosely fitting rind, and for practical purposes may be said to possess no rag. Unfor- tunately the writers have had but little experience with tangerines (Citrus nobilis), as these are rarely found in Hawaii, but such few fruits as have come to their attention have been well infested, which is to be expected, because of their thin, loosely fitting rind and rag. PERSISTENT ATTACK LEADING TO INFESTATION OF THE PULP Laboratory experiments and field examinations have shown that the female fly seldom deposits more than six eggs in a puncture at one time. So well has nature equipped the average citrous fruit to withstand attack that it is doubtful whether such fruits as the grapefruit, lemon, or orange would ever become infested ' until very much overripe, if the female fly fonned a new puncture for each batch of eggs deposited, thus making it necessary for the lar\'£e hatching from each lot of eggs to face identical difficulties in reaching the pulp. This, however, she does not always do. As many as 153 eggs have been taken from a single puncture in grapefruit. A very large number of punctures contained more eggs than the female deposits normally at one oviposition. It is very evident, therefore, that females o\'iposit in a large number of instances in the same puncture rather than make a fresh puncture for each batch of eggs. Frequently freshly laid eggs have been found in egg cavities from which channels made by larvae from pre-viously deposited batches of eggs extend through, to, and into the pulp or in punctured areas of the rind showing dry decay which is known from observation to have been forming for fully one month. Usually the rag beneath a puncture develops a discolored area, no matter whether the puncture originally contained eggs or not, and very often this discoloration of the rag, which appears to be caused by a dry rot, extends to the outer rind and causes deadened, sunken areas to fonn about the punctures. Such blackened areas, which had been developing in the rind of well-punctured oranges held at the laboratory for one month after picking, arc shown in Plate XLII, figure 2. It has already been stated that many larvae die in the rag and that before dying some of these lar\-K channel through the rag in all directions. Often all the larvae escaping from a puncture will be found dead next the skin protecting the pulp; again they will be found dead at the heads of channels extending fully i inch from the puncture. In large shaddocks they may even channel 3 or 4 inches through the loose rag (PI. XLI). I The term "infested" is here applied to fruits which have larvse in the pulp, show decay, and become generally tmfit for consumption. lau. IS. 1915 Citrous fruits and Mediterranean Fruit Fly 323 It is evident, therefore, that larvae hatching from the successive batches of eggs deposited in punctures or in the decayed areas of the rind forming about the punctures find conditions increasingly favorable to their ulti- mate success in reaching the pulp. The longer the fruit is allowed to remain on the tree after it becomes ripe, the easier it is for the maggots to reach the pulp. The rind can not withstand indefinitely the persistent attack of successive lots of larvae and the work of decay fungi to which the punctures give entry (PI. XLII, fig. i). Thus, 39 sweet oranges showing an average of 32 punctures to the fruit, gathered from the trees in Sep- tember, 1913, at a time when they were just becoming ripe, developed no larvae. On the other hand, out of 784 sweet oranges gathered during March, 1914, in a very much overripe condition, 254 produced 2,272 lan,-£e, or an average of about 9 larvae to the fruit. On account of the looseness of the rind and rag of sour oranges and the greater ease with which the rind is destroyed by decay fungi, these fruits are more quickly infested by the fruit-fly larvae. While both sweet and sour oranges in an overripe condition ultimately succumb to the repeated attacks of the Mediterranean fruit fly if per- mitted to remain on the tree, lemons, both of the commercial smooth- skinned and the rough-skinned varieties, withstand these attacks with a constancy that is astonishing. Lemons are not grown in sufficiently large numbers in Hawaii to permit the writers to record observations on large quantities of fruit, but even in orchards where the fruit is heavily punctured infested fruits are very seldom found. In about two years' time only three infested lemons of the commercial variety and one of the rough-skinned variety have been seen by the writers or by fruit-fly inspectors. Out of 235 well-grown and for the most part ripe lemons of the commercial type, picked from the tree, only i developed lar\'ae (this contained 3), and this fruit when picked was partially decayed as a result of a thorn prick. Out of 161 lemons of the same variety, taken from the ground in a very much overripe condition, but 2 developed larvs — I and 5, respectively. No larvae developed in 434 ripe rough- skinned but badly punctured lemons picked from the tree. One partially decayed rough-skinned lemon taken from the ground produced 1 2 larvae. The thicker skinned grapefruit, such as the writers have had an oppor- tunity to study best, have shown a strong resistance to the repeated attack of larvae or fungi. Yet these fruits were all grown on less than a dozen trees and in one garden. Twenty-five fruits taken from beneath these trees in a very ripe condition and showing an infestation of the rind equaling that recorded in Table I of the 39 fruits picked from the same trees, produced no larvai in the pulp. However, lar\-ae have been found in a few thin-skinned grapefruit that were in a very much overripe con- dition. 324 J ourncd of Agriculhiral Research voi. ni.xo. 4 SECONDARY ATTACK OF CITROUS FRUITS BY INSECTS OTHER THAN THE FRUIT FLY AND BY FUNGI The excellent experimental work of the Bureau of Plant Industry carried on in Florida during the last few years has forcibly demonstrated the causes of decay of citrous fruits in transit from orchard to market. Mechanical injuries to the rind have been found to be a fertile source of trouble by furnishing entry for decay fungi. The writers believe that a great share of the decay of oranges en route to market, recorded in the early history of the Mediterranean fnut fly, was caused more by insanitary conditions in the holds of ships than directly by fniit-fly larvae. It is more than likely that the oranges shipped from the Madeira Islands and the Azores to London contained fruit-fly punctures which greatly aided the blue mold in its destructive work. Statements made by the early writers and even repeated in the Hawaiian Islands at the present time, that citrous fruits drop as soon as punctured, are untrue. There is no such thing as a general shed- ding of fruits following puncturing of the rind. Oranges and grape- fruit have been known by the writers to hang on the tree from two to three months after they were first punctured. It is probable that the wholesale shedding of fruit recorded by others was caused by fungi or physiological troubles. Species of Drosophila and Bruchus may usually be found ovipositing in breaks in the rind of Citrus made by the Mediterranean fruit fly. Their persistent attack, supplemented by decay fungi, causes an appre- ciable amount of decay in Hawaii. EFFECT OF ATTACK OF THE MEDITERRANEAN FRUIT FLY UPON CITROUS CROPS OF CALIFORNIA AND FLORIDA In the opening paragraph the writers made the statement that their investigations in Hawaii have led them to believe that even if the Mediter- ranean fruit fly should be introduced into the citrous regions of the United States it would not become a serious pest to citrous fruits. In the Hawaiian Islands, especially in the lowlands, climatic conditions are more favorable for the rapid increase of the fruit fly than they are in any section of the United States or of the Mediterranean regions where oranges are grown commercially. The monthly mean temperatures at Honolulu during 1912 and 1913 ranged from 69.6° to 79.2° F. During the hottest summer weather the fniit fly requires a minimum of about 14K days to complete its life cycle from egg to adult. During late December, 1913, and January and early February, 1914, it required many flies fully 47 days to reach maturity in common guavas {Psidium guajava). During March and April, 1914, the fruit fly required from 20 to 30 days to pass from egg to adult in half-ripe peaches and from 28 to 40 days in lemons. In Bermuda during December and January, Jan.. 15 191S Citrous Fruits and Mediterranean Fruit Fly 325 when the monthly mean temperature normally ranges from 62.5° to 64.8° F., the senior writer, with the kind assistance of Mr. E. J. Wortley, Director of Agriculture, Bermuda Agricultural Station, found that the length of the pupal stage was about 31 days, which would make the period for development from egg to adult about 58 days in favored hosts. In Honolulu the cold-storage experiments of the writers have shown that the fruit fly requires about 91 days to complete the same development at about 56° F. A temperature of 54° to 57° will not prevent adults from emerging from pupae in cold storage, although it lengthens the pupal stage from 8 days, a normal minimum required at Honolulu in warm weather, to 36 days. Very few eggs out of several hundred were able to hatch at a temperature of about 53° to 54°, while practically no eggs will hatch nor larvae mature at a temperature of 50° F. A continued temperature ranging from 33° to 46° F. will kill pupa; and larvae, although both may be subjected to these temperatures for short periods without apparent injury. Freezing temperatures have proved generally fatal to both larvae and pupae. At 45° larvae are not able to pupate, although some hardy specimens may become active and pupate if removed at the end of a month from this temperature to the normal Honolulu summer temperature. A total of 10,203 second and third instar larvaa kept at a temperature varying from 42° to 46° were all dead at the end of 45 days, except one third-instar larva which was probably moribund, while out of 10,959 second and third instar larvae kept at a temperature varying from 33° to 38° none were aUve after the seventeenth day. These data from the notes on file are given here to show that even the cool winter climate of the lowlands of Hawaii has a decided effect in checking the increase of the fruit fly, that temperatures as low as 56° F. greatly lengthen the life cycle, and that a temperature of 50° to 52° practically prevents eggs from hatching. Unfortunately no data are at hand on the effect of the temperature varying above and below a mean temperature ranging from 50° to 53°. Certain deductions, however, can be made from known facts regarding the development of the fruit fly in the Mediterranean region, especially in southern Spain, France, Italy, and Sicily, that show that the fly does not multiply, or at least undergoes an extremely slow development, -when the monthly mean temperatures range from 50° to 54°. During the spring and summer of 1913, Prof.'H. J. Quayle, of the University of California, investigated the status of the fruit fly in the Mediterranean regions for the Bureau of Entomology,' and his observations bore out the contentions of the present Italian entomologists that the fruit fly is not a serious pest to Citrus in Spain and Italy. The fact that Prof. Quayle found no evidence * Quayle, H. J. Citrus fruit insects in Mediterranean countries. U. S. Dcpt. Act.. Bui. 134. 33 p.. 2 fig.. 10 pi. 1914. 326 Journal of Agricultural Research Vol. III. No. 4 df fruit-fly infestation in oranges and lemons in Spain during March or in southern Italy and in Sicily during April, May, and early June is strong evidence that the fruit fly is prevented by the mean tempera- tures prevailing in these countries from becoming a pest during the winter and spring months. Loquats are a preferred host of the fniit fly, being badly attacked when flies are present, and the appearance of infested fruits is such that infestation is easily detected. Even during July at Valencia, Spain, Prof. Ouayle found but a slight infestation of peaches and overripe oranges. In August, near Palermo, Italy, peaches were found badly infested, but lemons growing in the midst of peach trees were not infested. Reports indicate that in Spain and southern Italy the fruit fly may cause some damage to ripening oranges during September and October, although this is slight and of short duration. Citrous fruits, especially oranges, are not usually punctured by the female fruit fly until they are well grown and about to turn color, and the period of time is short after they reach this stage of ripeness until cool weather renders the fly sluggish. The season of the year, there- fore, when the bulk of the citrous crops are best suited for fruit-fly attack coincides with the season of inactivity of the fly due to lower tempera- tures. The mean monthly temperatures given in Table VI indicate that the Mediterranean fruit fly would find conditions in the citrous regions of Florida and California quite similar to those in Spain and Sicily. The Florida temperature, especially in the citrous regions of southern Florida, is decidedly above the winter means of those of California. However, even if the temperatures were higher than they are, the writers feel that the fruit fly would assume a minor position as a citrous pest. The cool winter weather would have the same retarding eS'ect upon develop- ment in California and Florida that it has in the European countries. Table VI. — Monthly mean temperatures iti citrous regions Locality. Seville. Spain Malaga, Spain Naples, Italy Palermo, Sicily San Francisco, Cal Redlands, Cal San Diego, Cal Portersville, Cal Jacksonville, Fia Eustis, Fla Miami, Fla Myers, Fla Honolulu, Hawaiian Islands Prospect Hill, Bermuda. Janu- ary. Feb- ruary. March. April. May. June. July. Au- gnst. Sep- tem- ber. Octo- ber. No- vem- ber. cem- ber. 52-2 55- 9 59- S 63-9 69.6 78., 84.7 84-9 7.S.1 68.4 60. 1 52.9 5.va SS-4 57-6 6..S 65-7 71.4 76.8 77-2 -2-5 66.0 60.3 55-4 46.8 48.4 SI. 4 56.8 63-7 70-3 75.6 75- 0 69.8 63.1 .54-7 48.7 50.5 52. I S4-7 56.8 64.0 70.7 76.3 76.6 73-4 67-3 59-4 53-4 so. 0 52- 0 S4-0 SS-o 57- 0 S9-0 59- 0 59- 0 61.0 6o-o 56.0 SI.o 5I-0 52,0 SS-o 61. 0 06.0 74.0 78.0 78-0 72.0 65.0 59° S3-0 54- 0 SS-o 56.0 60.0 63.0 65.0 68.0 70-0 66.0 64.0 59-0 56.0 49.8 SI. 6 59-0 63.0 66 0 57-0 62.0 48.0 s6.o SS-o SS.o 68. 0 75- 0 80.0 82.0 82. 0 73.0 71.0 sS.o 61.0 67.0 70.0 77.0 81.0 »,.o 83.0 So.o 73- 0 66.0 60.0 65.0 67.0 71.0 74- 0 76.0 81.0 82.0 82.0 81.0 78.0 74-0 69.0 62.0 6s. 0 68.0 72.0 77.0 80.0 81.0 81.0 80.0 75- 0 70.0 64.0 71.4 70.8 69.6 72.6 74-6 76.0 77-4 78.2 78.2 77.6 74-6 74.0 62.5 62.2 63.9 66.1 71.4 77-7 79.8 81.0 78.0 73-7 68.6 64. S Jan. 15. 1915 Citrous Fruits and Mediterranean Fruit Fly 327 The general effect of retarded development of the fruit fly due to cold weather is to increase the mortality among all stages. Even pupse are subject to an increasing rate of mortality the longer they are subjected to lower temperatures. Adults seem more able to withstand prolonged cold weather than any of the other stages. One individual was kept alive bv daily feeding for 4K months during a Hawaiian winter and spring. However, during the cooler months the adults are more sluggish and fall more easily a prey to adverse climatic conditions, such as heavy winds and rains, and to predaceous insects. Mr. George Compere reports having seen adults sunning themselves on orange trees in Spain after a night during which the temperature dropped to freezing, thus showing that adults can withstand temporarily any cold snap likely to occur in a citrous section. However, the fact that adults do not succeed in thriving during the winter temperatures of southern Spain and Italy and in Sicily seems to be well proved by the fact that it is only during the summer and early fall that the fruit fly becomes a serious pest in favored host fruits and in overripe citrous fruits. If this were not so, fruits would become badly infested much earlier in the season than they do. The number of adults surviving the winter must be very small. Even the mild winters of Hawaii at Honolulu have a very noticeable effect upon the numerical abundance of the adult flies, as shown by trap experi- ments extending over one full year. In addition to this beneficial effect of lower winter temperatures, both California and Florida growers will receive further protection as a result of the conditions surrounding the growing of Citrus as a commercial proposition. In the Hawaiian Islands, especially about Honolulu, citrous fruits are subjected to the most severe attack imaginable under field conditions. They are attacked over long periods by an abundance of fruit flies that mature in many host fruits ripening at intervals through- out the year on all sides of isolated citrous trees. The number of wild fruits in which the fruit fly can breed in the citrous regions of California and Florida is, in comparison with Hawaii, so extremely small that the fly would find conditions unfavorable for rapid increase, even if weather condi- tions were more favorable. In many instances large acreages of Citrus oc- cur where vegetation is normally decidedly stunted unless irrigation is prac- ticed. With the excellent work of the horticultural inspectors in California a reduction of the noncitrous host fruits in and about citrous groves is a practical proposition. Even near-by orchards of drupe fruits are not the menace that they seem to many, inasmuch as their crops are unsuit- able for fly attack except during short periods of the year. The very scarcity of vegetation that can not be destroyed which produces fruits subject to fruit-fly attack makes it possible to attach a far greater impor- tance in California and Florida than in the Hawaiian Islands to the excessive mortality of the fly discussed in this paper. It has been shown 328 Journal of Agricultural Research voi. iii. no. 4 that it takes repeated attacks to infest grapefruit, lemons, and oranges in Hawaii and that the pulp of these fruits is infested usually only after the fruits are very ripe; in fact, not until they become much riper than commercially-grown oranges usually are allowed to become in either California or Florida, unless exception be made of such varieties as late Valencias. The relatively small number of adult fruit flies entering a block of citrous trees would find it very hard to establish themselves, since the numbers would be so insignificant as compared to the fruit surface suitable for oviposition that each female would be less likely to oviposit repeatedly in the same puncture. Her progeny would therefore meet with almost insurmountable difficulties in reaching the pulp. It has been stated that in Hawaii citrous fruits offer themselves for attack over several months and that they are not subject to serious attack until they have turned or are about to turn color. It is a well- known fact among horticulturists that in very equable climates the pulp of oranges may be ripe enough to eat while the rind is still very green. In Florida and California this is not so true. One has only to visit the packing houses in either State to be convinced that much fruit is gathered for the early trade in a semiripe condition, or at least when the rind is quite green in color. The writers feel safe in saying that market condi- tions are such that early fruit is placed on the market at the earliest possible moment, in order that high prices may be secured. The fear of unseasonable frost and freezes has made it difficult for those who have the interests of the citrous industry at heart to prevert the shipping of too green fruit. It would seem that with a reasonable expenditure of more care than labor, citrous groves in either Florida or California can be made so well protected from the Mediterranean fruit-fly attack that such few flies as enter them during the fall will find the early fruit, upon which they can work because of its degree of ripeness, picked before they are able to injure it to any extent. The cold weather will protect the later fruit by rendering the fruit flies inactive, and by the time the spring temperatures become suitable for fly activity the bulk of the fruit will have been marketed and the numerical abundance of the adult flies greatly lessened. In addition, if it becomes necessary, as a result of unfavorable condi- tions, to use artificial means of control, spraying with a cheap poisoned bait will be a practical method of reducing the number of adults. If the writers under most adverse conditions can reduce by spraying the num- ber of adult fruit flies over 50 per cent in one city block in Honolulu, into which it has been proved that adults are continually migrating, it is only reasonable to expect that the same good results as have been secured in South Africa, where fruit has been protected by spraying, will follow spraying in either Florida or California, where outside sources of infestation can be so easily controlled. Jan. .5. 191S Citrous Fruits and Mediterranean Fruit Fly 329 CONCLUSION Citrous fruits are not the favored host fruits of the Mediterranean fruit fly (Ceratiits capitata Wied.) that the earlier writers thought. While grapefruit, oranges, lemons, and many limes may become quite badly infested with well-grown larvae if allowed to remain on the tree long after they become sufficiently ripe for the market, nature has so well equipped them to withstand attack that larva; are seldom found in their pulp imtil they are much overripe. Oranges and grapefruit are generally eaten and found uninfested if gathered as they ripen. Indeed, in Honolulu, where conditions are very favorable to early infestation of the pulp, owing to the excessive numbers of adult flies breeding in a large number of host fruits ripening in rapid succession, it is doubtful whether grapefruit, oranges, and lemons would ever become infested until long after becoming overripe if the female fly formed a fresh egg cavity for each batch of eggs deposited, for the reason that the eggs and the young larvse found in the egg cavity and in the rag of the rind would then be forced always to face well-nigh insurmountable difliculties. The oil of the cells ruptured in the formation of the egg cavities kills a large percentage of the eggs and newly-hatched larvae. Larvae that succeed in entering the rag from the egg ca\ity are able to reach the pulp in astonishingly small numbers because of the imperviousness of the rag. It is only the persistent attack of successive lots of larvas hatching from different batches of eggs laid in the same puncture in which the oil has become inoperative that finally breaks down the barrier between the young lar\'ae and the pulp. The Mediterranean fruit fly is quickly affected by low temperatures. A temperature of about 56° F. has lengthened the time required by the fly to pass from the egg to the adult stage from 14K to 91 davs. A tem- perature ranging from 50° to 55° F. will either seriously check devel- opment or kill large numbers of the immature stages of the fly. The winter monthly mean temperatures of California and Florida are so similar to those of the citrous regions of southern vSpain and Italy and of Sicily that it is to be expected that the fruit fly, if introduced to the mainland, would not become a serious pest to Citrus spp. It happens that the very cold temperature necessary to bring citrous crops to that degree of perfection in which they are most susceptible to fruit-fly attack likewise renders the fly so inactive or sluggish that it may be disregarded as a pest for that period of the year. In addition to the assistance of adverse climatic conditions during that ])art of the year when they are most needed to protect citrous crops, the growers of California and Florida are still further protected — and most admirably so — from attack by the very scarcity of wild host fruits that can not be destroyed. It will be found a practicable under- taking to remove such a number of noncitrous host plants at present 697.33°— 15— 1 33° Journal of Agricultural Research voi. m. No. 4 growing about commercial citrous orchards that the succession of fruits in which the Mediterranean fruit fly can breed during the large portion of the year when citrous fruits are unavailable for attack because of their greenness will be reduced to a minimum, if not entirely done away with. It is under conditions such as can be secured in California and Florida that the excessive mortality occurring in the rind will become a valuable factor in preventing infestation or establishment of the pest, as each fruit will in reality become a trap for stray females. The scarcity of host fruits will also make spraying with poisoned baits a practical undertaking, should it become necessary to resort to artificial methods of control. Adverse climatic conditions at a season when citrous fruits are most susceptible to attack, solid plantings of Citrus in commercial orchards, a scarcity of noncitrous host fruits, the ease with which the fly can be reduced by spraying with poisoned baits, and the general practices fol- lowed in harvesting fruits make it possible for the citrous growers of California and Florida to rest assured that the discovery of the Medi- terranean fruit fly in either State will not bring about the ruination of the industry. Its presence will be a constant menace, but it can be successfully fought. LITERATURE CITED 1817. Latreille, p. a. Les Tephrites. In Ctivier, G.. I,a rfegne animal, v 3, p. 647. Paris. 1829. MacLeay, W. S. Notice of Ceraiitis citriperda, an insect very destructive to orange. In Zool. Jour. [London], v. 4, p. 475-492, pL 15, 1829. 1842. BrSmE, F. de. Note sur le genre Ceratiiis de M. MacLeay. In Ann. Soc. Ent. France, v. 11, p. 183-igo, pL 7, fig. 1-5, Stance du 4 mai, 1S42. 1848. Westwood, J. O. The orange fly. In Gard. Chron., 1848, no. 37. p. 604, 3 fig. Account of Ceratitis capitata and character of injury. 1859. GouREAU, Col. In Ann. Soc. Ent. France, ser. 3, v. 7, Bui., p. xlui-xi.v, Stance du 23 mars, 1859. Communication on Ceratitis hispanica reared from an orange sent to M. Gourcau by M, \'ille- neuve. 1870. RoNDANi, C. Ortalidinae italicae collectae, distinctae et in ordincm dispo- sitae . . . Dipterologiae italicae prodromi, Pars VIL fasc. 4. In Bui. Soc. Ent. Ital., V. 2, p. 29, 1870. Description of Petalopkora Macq. 1871. Laboulbene, A. Note sur les dommages causes par la Ceratiiis hispanica aux fruits des orangers dans nos possessions d'Alg^rie. In Ann. Soc. Ent. France, s6r. 5, V. I, p. 439-443, .Stance du 13 d^cerabre, 1871. PLATE XL Fig. I. — Orajige infested with hirvseof the Mediterranean fruit fly (Ccratitii capilaia). Note that the fruit looks sound, except about the irregular hole, through which a few well-grown lars ae have already left the fruit. Original. Fig. 2. — Orange infested with larvae of the Mediterranean fruit fly {Ceratitis caf^itiila), showing two breathing holes of the larvae in the decayed area. Original. Citrous Fruits and Mediterranean Fruit Fly Plate XL « Tk^' '* v'^-^. ^%r*ry Journal of Agricultural Research Vol. III. No. 4 Citrous Fruits and Mediterranean Fruit Fly Plate XLI Journal of Agricultural Research Vol. III. No. 4 PLATE XLI Cross section of shaddock No. i, showing the thick, loose texture of the rag with darkened area above and to tlie right showing the channels made by well-grown Med- iterranean fruit-fly larvse. Original. PLATE XLII Fig. I. — Cross section of the orange shown on Plate XL, figin-e 2. Note tliat in this instance the larvse have brought about decay in only one section. Often many sections are thus affected in very ripe fruits. Original. Fig. :;. — Orange containing 87 punctures in the rind. Photographed one month after being picked from the tree in a ripe condition. Note that the rind about many punctures is simkcn as a result of a dr>' black-rot. The pulp of this fruit was perfectly soimd. Original. Citrous Fruits and Mediterranean Fruit Fly Plate XLII Vti?tHtv*j*aii?v Journal of Agricultural Research Vol. Ill, No. 4 PHYSIOLOGICAL CHANGES IN SWEET POTATOES DURING STORAGE By Heinrjcu Hasselbring and LoN A. Hawkins, Plant Physiologists, Plant Physiological and Fermentation Investigations, Bureau of Plant Industry INTRODUCTION In resting storage organs of plants growing in northern and in tem- perate regions, carbohydrate transformations involving the disappear- ance of reserve starch during the colder months and its temporary reap- pearance in spring have been found to be of general occurrence. The disappearance of starch from the cortex of trees in winter and its reap- pearance in early spring was first noted by Miiller (1877),' who believed the absence of starch in the cortex resulted from its migration into the wood. Russow's investigations (1882-83), which included the examination of a number of tropical and subtropical greenhouse plants, showed that the total or partial disappearance of starch from the cortex of woody plants in winter was a phenomenon of widespread occurrence. He found, however, that the starch did not migrate into the wood, as Miiller sup- posed, for when pieces of cortex chiseled from trunks of trees were kept at a temperature of 14° to 17° R. (17° to 21° C.) starch grains began to reappear in 20 hours. In the tissues which were free from starch in win- ter he found oil and fats. He observed a correlation between the tem- perature and the disappearance and reappearance of starch, but since the processes occurred also in tropical plants in the greenhouse, he did not regard temperature changes or climatic conditions as the prime causes of the observed transformations. Later Grebnitzky (1884) and Baranetzky ( 1 884) showed that the starch of soft- wooded trees disappeared entirely from the wood, cortex, and rays in winter, and that oil appeared in its place, while in hardwood trees the starch disappeared from the cortex, but per- sisted in the wood. Fischer (1891), in his extended investigations on the physiology of woody plants, fully confirmed the observations of Russow (i 882-83) , Grebnitzky ( 1 884) , and Baranetzky (i 884) regarding the appear- ance of oil in place of starch in soft-wooded trees, and showed further that in hardwood trees glucose and tannin are present in the cortex after the disappearance of the starch, and that the glucose, but not the tannin, disappears when starch is regenerated. He found that the regeneration of starch takes place at a temperature only a few degrees above 0° C. The minimum temperature at which he observed the regeneration of starch in twigs was about 5° C, while at 10° to 20° the process went on very rapidly. ' Bibliographic citations in parentheses refer to "Literatme cited," p. 341. Journal of Agricultural Research, Vol. III. No. 4 Dcpt. of Agriculture, Washington. D. C. Jan. 15. 1915 (33^) G-" 332 Journal of Agricultural Research voi. iii, N0.4 That the periodic transformation of reserve starch is not restricted to the stem tissues of plants is shown by the observations of Haberlandt (1876), of Mer (1876), and of Schulz (1888), who found that the starch disappears from evergreen leaves in temperate regions in winter, while Haberlandt and Schulz noted also that it was re-formed in spring. The most thorough investigation of the carbohydrate transformations in ever- green leaves was made by Lidforss (1907), who found that the leaves of all evergreen plants in cold countries, except aquatic plants, lose their starch in winter, sugar appearing in its place, and that starch is regen- erated in the leaves in February and March when the temperature scarcely rises above 5° C. That similar changes occur in the subterranean parts of perennial plants in temperate regions was shown by Rosenberg (1896), who obser\'ed the disappearance of starch after leaf fall in the subterranean parts of Spiraea nlmaria, Scrophidaria nodosa, Plantago major, Poicntilla argentea, and Hepatica triloba, but did not determine what substances appeared in its place. By far the most complete account of carbohydrate transformations in dormant organs of this type is given by Miiller- Thurgau (1S82) in his classical researches on the accumulation of sugar in the potato (Solanum tuberosum) and other plant organs at low tem- peratures. Miiller-Thurgau found that an accumulation of sugar and a corresponding loss of starch occurred in potatoes kept at low temperatures (0° to 6° C), while, contrary to popular opinion, no sugar is formed in potatoes which have been actually frozen. He found that when potatoes which had become sweet as a result of exposure to low temperature are kept at a higher temperature (8° to 10° C.) the sugar disappears and the starch increases. Furthermore, he showed that the sugar formed con- sists mostly of reducing sugar with some cane sugar in the proportion of about 2.5 to I, and that similar transformations occur in other parts of plants. These phenomena are interpreted by Miiller-Thurgau (1882) as follows : The transformation of starch into sugar is an enzymic process which, although more rapid at high temperatures, occurs also at low tem- peratures. The respiratory activity which is almost at a standstill at 0° C. rises with the temperature so that at higher temperatures an increasingly greater amount of sugar is consumed by respiration. The amount of sugar used in respiration at higher temperatures is, however, small compared with that utilized by another process — i. e., the re-formation of starch from sugar, which takes place at tempera- tures somewhat above 0° C. and increases in speed with the rise of temperature. Appleman (1914) in his studies on the rest period of the potato also finds that the carbohydrate changes in the dormant tubers are entirely dependent upon changes of temperature. It appears, there- fore, that the carbohydrate transformations of the potato, although a Jan. IS, 1915 Changes in Sweet Potatoes During Storage 333 subtemperate plant and not capable of long withstanding temperatures much below freezing, resemble in their general trend those of subter- ranean organs of temperate plants. In the more strictly tropical sweet potato (Iponioea batatas) carbohydrate transformations of a similar na- ture have been observed. Thus, Harrington (1895) found that in stored sweet potatoes there was an increase of the total amount of sugar up to March 6, bej'ond which the experiments were not continued. Shiver (1901), whose experiments were somewhat more extensive, found that during the time of his experiments (up to April 17) there was a gradual decrease of starch and an increase of cane sugar, while the invert sugar showed but slight fluctuations. Neither of these writers described the conditions under which the potatoes were stored nor attempted to de- termine the effect of temperature on the metabolic changes. The storage of sweet potatoes is accompanied by considerable losses as a result of decay which is not wholly preventable by any of the methods of storage advocated at present. The decay is brought about by micro- organisms which invade the tissues. In the matter of susceptibility the internal changes in the roots must play an important part. These changes are affected by changes in temperature and other conditions to which the roots are subjected during storage. It is therefore a matter of practical importance, as well as of theoretical interest, to study the internal changes which take place in sweet-potato roots after har\-est and during storage, and to determine the effect of external conditions upon such changes. The work reported in this paper is a general study of the carbohydrate metabolism of sweet potatoes stored at different temperatures. PLAN OF THE EXPERIMENTS For the purpose of this work two varieties of sweet potatoes, the Jer- sey Big Stem, representing the sugary type, and the Southern Queen, representing the starchy type, were selected. The potatoes used in the experiments were a part of the general crop grown by the Office of Horti- cultural and Pomological Investigations during the summer of 191 1 in a series of variety tests which had been continued for a number of years. At the time of harvesting, a representative lot of about 15 bushels of each of the two varieties was selected in the field and packed in slat crates holding about a bushel each. These were placed with the rest of the crop in the sweet-potato cellar of the Office of Horticultural and Pomological Investigations, where all were subjected to the "sweating," or curing, process. During the period of curing, the temperature of the room was kept at approximately 27° C. for about 10 days, after which it was allowed to drop to the regular storage temperature, ranging in this case, except near the end of the season, between 11.7° and 16.7° C. Nine crates of each variety were left in the cellar at the above-mentioned 334 Journal of Agricultural Research voi. iii. no. 4 temperature, which was maintained by the aid of artificial heat when necessary. The remaining six crates of each variety were placed in a cold- storage room, which was kept at a fairly uniform temperature of 4° C, by means of circulating brine cooled by ice and salt. Thus, both the lot stored at the usual storage temperature and that placed in cold storage were submitted to the same preparatory curing process. In order to determine the carbohydrate changes which occurred in these lots during the season, samples were analyzed on the day the potatoes were dug and at intervals of about a month during the course of the experiment, from October to June. In these samples the water, starch, reducing sugar, and total sugar were determined. EXPERIMENTAL METHODS Sampling. — For each set of determinations, a random sample of 4 to 5 kg. was taken. The roots were rapidly washed and wiped with a towel. When the surface had become entirely dry the roots were cut up as quickly as possible and ground in a power-driven meat grinder having a face plate with holes 3.2 mm. in diameter. The operation of cutting and grinding required about 10 minutes. The mash thus obtained was thoroughly mixed on a glass plate and quartered twice. The final sample thus obtained was placed in a crystallizing dish and covered with a damp towel while the samples for sugar, starch, and moisture determinations were being weighed out. Moisture. — For the determination of moisture, samples of approxi- mately 10 gm. were transferred into tared weighing bottles and accu- rately weighed. The material was covered with 95 per cent alcohol, which was subsequently evaporated in vacuum desiccators containing sulphuric acid. The samples were then dried to their lowest weight in a current of hydrogen in a vacuum oven at 78° C. The drying required 15 to 18 hours, during which the bottles were weighed three or four times. Starch. — It was not possible to make the starch determinations immediately. Samples of 25 gm. correctly weighed to i cm. were there- fore transferred to Erlenmeyer flasks of 200 or 250 c. c. capacity and covered with 150 c. c. of 95 per cent alcohol. A little precipitated calcium carbonate was added to the flasks, which were then brought to the boiling point in a water bath. Subsequently the samples were washed with alcohol into tared porcelain extraction thimbles, 75 mm. high and 40 mm. in diameter, with perforated bottoms which were covered with filter paper cut to fit. Another piece of filter paper was pressed down upon the material and held in place by means of a cotton plug. The thimbles were supported well up in Soxhlet extraction ap- paratus and extracted with strong alcohol for 12 hours. After extrac- tion the cotton and filter paper were removed, and the thimbles were Jan. IS. 191S Changes in Sweet Potatoes During Storage 335 dried for 20 hours at 60° C. and subsequently were allowed to stand in the laboratory at least 48 hours, in order that the material might come to a state of moisture-equilibrium with the air. The thimbles were then weighed and the material was quantitatively transferred to a mortar and ground to a fine powder. The starch was determined as glucose by the acid-hydrolysis method (Wiley, H. W., et al., 1908) in two accurately weighed fractions of this powder, each representing about one-half of the extracted residue before it was ground. Sugar. — For the determinations of sugar, samples of 25 gm. were washed into 250 c. c. volumetric flasks with enough neutral 70 per cent alcohol to bring the volume up to about 200 c. c. About i gm. of calcium carbonate was added to each flask. The flasks were then boiled in the water bath for 10 minutes and on the following day were cooled to 20° C. and filled to the mark. After being stoppered they were allowed to stand for a few days, during which they were occasionally shaken to insure uniformity of concentration of sugars in the solid and the liquid portions of the contents. The solutions were subsequently treated essentially according to the method described by Bryan, Given, and Straughn (191 1). Reducing sugars and total sugars were determined according to the method of Allihn (Wiley, H. W., et al., 1908). The cane sugar was calculated from the difference between the total sugar and the reducing sugars. Temperature. — The temperature of the two storage rooms was recorded by thennographs. The curves obtained in the warm storage room were integrated with a planimeter to obtain the average weekly temperatures, which are given in Tables I and II. The average weekly temperatures for the cold-storage room were written down from inspec- tion of the records, since the tracings in this case were practically straight lines. EXPERIMENTAL DATA The data showing the seasonal changes in the composition of sweet potatoes stored in the farm cellar at a temperature varying mostly from 11.7° to 16.7° C. are given in Table I. The percentages of car- bohydrates have all been referred to the original moisture content of the potatoes. The loss of solid matter by respiration had, of course, to be disregarded. The numbers expressing the total content of carbo- hydrates were obtained by the addition of the numbers representing the starch (as glucose) and the total sugars (as glucose). 336 Journal of Agricultural Research Vol. in. No. 4 Tabi.K I. — Carbohydrate transformations in sweet potatoes stored in the farm cellar BIG STEM Oct. 30 Nov. 8. Dec. 6. Jan. 4.. Feb. I. Mar. I . Mar. 3o Mar. 26 Apr. 16 June I . Water. P.cl. 73-50 72.99 71.89 72.06 73.02 72.49 72.87 Starch. P.cl. 19.07 16.94 16. 42 13-44 14.47 14. 20 14. 63 Cane sugar. P.cl. 1.90 3-51 3-94 6.96 6.40 5. 61 6.03 5-85 Reduc- ing sugar as glu- cose. Total sugar as glu- cose. P.cl. 0.90 1.32 1.40 1.23 J. 10 .87 .90 .87 p.cl. 3. 90 S. 03 s-ss 8.04 s. 76 7.84 6.77 7-24 Total carbo- hy- drates as Glu- cose. p.cl. 34.09 23-85 23-79 22. 77 22.8s 23.03 I Gain or loss of starch (as glu- cose).a 2.36 ■58 -1-13 + -39 -t-l. 14 — .30 -I- .47 Gain or loss of sugar (as glu- cose).a Gm. H-3. 12 + -Si + -35 — I. 07 + -47 Average weekly tem- perature. "C. 26. 7 21. 7 16.7 16.7 15.6 15.0 IS-O 18.3 15.6 14.4 12.8 II. 7 12.S 16. I 15- O 610.3 16.7 144 ''150 15.0 16. I 16. I iS. 9 16.7 17.8 20.6 18.9 18.9 17.2 21. 1 21. I Oct. Week iding — 'ct. 30 -lov. 6 Nov. 13 Nov. 30 Nov. 27 Dec. 4 Dec. 11 Dec. iS Dec. 35 Jan. I Jan. 8 Jan. 15 Jan. 22 Jan. 28 Feb. 5 Feb. 13 Feb. 19 Feb. 26 Mar. 4 Mar. II Mar. iS Mar. 25 Apr. I Apr. 8 Apr. IS Apr. 23 Apr. 29 May 6 May 13 May 30 May 27 June 3 SOUTHERN QtJBEN' Oct. 23..., Nov. 10. . . Dec. 7 Jan. II. . . , Feb. 3.... t Feb. 2S. . , Apr. 8 May 4 June 4 71.69 68.41 22.09 19.87 0.3, • 77 1.64 3- 89 26. 18 26. 7 2.97 25-96 -2.47 -1-2. 2S / "-7 I 16.7 16.7 67.69 19.30 3-5° - 72 4.41 25-85 - -63 -I- -52 IS. 6 15.0 15.0 18.3 IS. 6 67.51 '9- 75 3. S3 ■75 4.46 26.41 + -SO -f .OS 14. 4 12.8 II. 7 12.8 63.03 19.22 3- 95 .60 4.7s 26. II — ■ 59 -1- .29 16. I 15.0 f* 1 2. -• 6S. OD 18.99 4- OS -53 4.80 25.90 - . lb ■f .05 16.7 14.4 ^ 15.0 15.0 66.71 30.3s 2.93 -Sa 3.61 36. 32 -l-i-S' -1. 19 16. I 16. t i.S. 9 16.7 69. 31 19. ;8 3-39 -SI 4.07 26.05 - .63 + .46 17.8 20. 6 18.9 18.9 68. IS 20. 15 3- 80 ■Si 3-50 35.89 + .41 - ■ 57 17.2 21. X 21. 1 Oct. 30 Nov. 6 Nov. 13 Nov. 20 Nov. 27 Dec. 4 Dec. II Dec. 18 Dec. 25 Jan. I Jan. 8 Jan. IS Jan. 23 Jan. 28 Feb. 5 Feb. 13 Ffb. 19 Feb. 26 Mar. 4 Mar. II Mar. 18 Mar. 25 Apr. 1 Apr. 8 Apr. IS Apr. 22 Apr. 29 May 6 May 13 May 20 May 27 June 3 a Per 100 gm. of material. ^ Record obtained for only one day of this wt-pk. Jan. 15, 191S Changes in Sweet Potatoes During Storage 337 The data in Table I show that under the conditions of this experi- ment the moisture content of the roots remains fairly constant. There is a slight decrease in the moisture content, more marked in the Southern Queen than in the Big Stem variety, during the curing process, but on the whole there is comparatively little change in the percentage of moisture. The loss of moisture is probably compensated in part by the water formed by respiration, while the loss of substance by respiration would increase the relative moisture content, thus tending to conceal actual water lost. The percentage of starch shows a rather sudden decrease immediately after the potatoes are dug. The subsequent decrease is more gradual, and continues until a minimum is reached in March. After that time there is a continuous rise in the percentage of starch until the last date on which the potatoes were examined. Concomitant with the changes in the percentage of starch there is an inverse change in the percentage of sugar. Corresponding with the first sudden decrease of starch, there is an equally sudden increase in sugar. Later the increase in sugar content is more gradual, and reaches a maxi- mum at the time of the starch minimum. After the sugar content has reached a maximum there is a gradual decrease, which, however, is not as marked as the increase during the first part of the season. ■ The course of the changes in the percentage of cane sugar follows that of the total sugar in both varieties, but in the Southern Queen the invert sugar after the initial rise shows an almost continuous decrease, whereas in the Big Stem the invert sugar content also shows a distinct maximum. The total carbohydrate content in both types remains fairly constant; consequently the numbers showing the loss (or gain) of starch between the successive dates of sampling show a fairly close agreement with those showing the corresponding' gain (or loss) of total sugar. The aberrations are probably to be attributed partly to the loss of substance through respiration, but mostly to nonconformity of samples. The data showing the carbohydrate transformation in sweet potatoes stored at low temperatures (approximately 4° C.) are given in Table II. Table II. — Carbohydrate transformations in sweet potatoes in cold storage BIG STEM, FIRST LOT" Starch. Cane sugar. Reduc- ing sugar as glu- cose. Total sugar as glu- cose. Total carbo- hy- drates as glu- cose. Gain or loss of starch (as glu- cose).'' Gain or loss of sugar (as glu- cose).'' Average weekly tem- perature. Mov. J. Dec. Q. . Dec. a I . P.cl. 72. V9 P.cl. 16.94 P.cl. J- SI 6.46 P. cl. ■•3> 1.60 P.cl. 5. 03 P.cl. 23. fia Gvi. -4. 03 -2. 79 + 3-»o Week ending — Oct. 33 Oct. 30 Nov. 6 Nov. 13 Nov. 20 Nov. 27 Dec. 4 Dec. 1 1 Dec. 18 Dec. 35 o The figures are alt calculated for the original water content of the roots, 73.50 per cent. ^ Per 100 gm. of material. 338 Journal of Agricultural Research voi. in. no 4 Table II. — Carbohydrate transformations in sweet potatoes in cold storage — Continued BIG STEM, SECOND LOT** Date. Water. Starch. Cane sugar. Reduc- ing sugar as glu- cose. Total sugar as glu- cose. Total carbo- hi- d rates as glu- cose. Gain or loss of starch (as glu- cose). Gain or loss of sugar (as glu- cose). Average weekly tem- perature. Mar. 27 P.cl. 72-19 73-32 P.cl. 12.99 9-74 P.ct. 6.41 8.74 P.ci. I. 6s 2.44 P.cl. 8.39 11.64 P.ct. 32.83 22.47 Gm. Gm, 'C. Week ending — -3-61 +3-»S SOUTHERN O^JEEN & Nov. 10. Dec. 8... Dec. 33.. 68.41 66.77 19.87 17.40 2-97 S-93 0.77 •59 3-89 6.83 25-96 26.16 7-8 7-2 S-6 5-6 4-4 4-4 4-4 -2. 74 -t-2-94 67-57 16.48 6.94 -65 7-96 26.28 — 1.02 ■fi-13 3-9 3.3 \ 2.8 Oct. 23 Oct. 30 Nov. 6 Kov. 13 Nov. 20 Nov. 27 Dec. 4 Dec. II Dec. 18 Dec. 25 o The figures are all calculated for the original water content of the roots. 73.50 per cent. *> The figures are all calculated for the original water content of the roots. 71.69 per cent. In these experiments three lots of potatoes were used. One lot of the Big Stem and one of the Southern Queen were placed in cold storage immediately after they had been cured. Another lot of the Big Stem variety which had been kept in warm storage until March 27 was placed in cold storage on that date. The cold-storage experiments were of short duration, since the potatoes invariably rotted after having been kept at the low temperature for about six weeks. These data show that at low temperatures the disappearance of starch and the accumulation of sugar in sweet potatoes take place more rapidly and proceed to a greater extent than at high temperatures. As to the relative proportion of the individual sugars, the two types of potatoes seem to differ somewhat. In both types the cane-sugar con- tent is markedly higher in cold than in warm storage. In the Big Stem sweet potatoes the invert-sugar content also is higher in cold storage, but in the Southern Queen the invert-sugar content is no higher in cold than in wann storage. In general, cane sugar is the chief product which accumulates at low temperatures. The total carbohydrate con- tent, with one exception, remains fairly constant, and the increase of sugar accounts for the loss of starch. The exception mentioned is the discrepancy between the loss of starch and the gain of sugar in the Big Stem potatoes during the inten'al from December 9 to December 21. The only explanations that can at present be suggested for this discrep- ancy are either that after long exposure to low temperatures the various phases in the process of the transformation of starch into sugar are influenced in such a way that intermediate products which escape detec- Jan. IS. 19-5 Changes in Sweet Potatoes During Storage 339 tion by the analytical methods employed accumulate to a greater extent than usual; or, inasmuch as many of the potatoes showed small rotten spots at the time of the last sampling, it is possible that although these were cut out and the flesh appeared otherwise entirely sound, the enzyms secreted by the fungus had brought about a partial transforma- tion of starch beyond the zone actually invaded by the mycelium. DISCUSSION OF RESULTS A striking fact brought out in the tables is the high starch content and the low sugar content of the sweet potato immediately after har- vesting. A number of analyses, not here reported, of potatoes dug at different times also showed that freshly dug potatoes contain only small quantities of sugar. However, as soon as the potatoes are dug, a rapid transformation of starch into sugar takes place. A number of experi- ments not given here showed that this sudden transformation of carbo- hydrates takes place over a v.'ide range of temperatures and that even at 30° C. the process is so rapid that sugar accumulates in excess of the quan- tity used in respiration, while at any subsequent period the accumulated sugar diminishes at that temperature as a result of respiration. This initial transformation is in such striking contrast with the later less rapid transformation that the two may almost be considered as distinct phases in the carbohydrate metabolism of the roots. It appears that during the period of active growth processes occur which prevent the accumula- tion of sugar in the roots. The elaborated materials from the leaves are almost wholly transformed into starch. The reverse process, which takes place as soon as the potatoes are dug, seems to be associated with the cessation of the flow of materials from the vines to the roots. The influx of materials from the vines therefore seems to determine the direction of the carbohydrate transformation in the growing roots. Subsequent to the initial period, the carbohydrate transfonnations in the sweet potato are greatly influenced by temperature. In warm storage there is a'continual accumulation of sugar in excess of the quantity used for respiration during the first part of the storage period. The corre- sponding disappearance of starch leaves no doubt as to the source of the sugar. During the latter part of the season the process is apparently reversed. The increase in the percentage of starch and the decrease in the percentage of sugar during this period suggests that during the latter half of the storage season a reformation of starch takes place, such as has been ob- served in twigs and woody stems and in the tubers of the common potato. It should be noted, however, that increased respiration during the latter half of the season, during which the temperature of the storage room rose gradually, may account for the loss of sugar. However, the constancy of the total carbohydrates and the increase in the percentage of starch seem 340 Jountal of Agricultural Research voi. lu. N0.4 to show that there is an actual transformation of sugar to starch during this period. In a general way the course of the carbohydrate transforma- tions in sweet potatoes seems to be correlated with the seasonal varia- tion in the temperature of the storage room. That the temperature may be the controlling factor in determining the direction of the carbohydrate transformation is shown by the continuous transformation of starch into sugar in the sweet potato as well as in other storage organs of plants at low temperatures and the reversion of the process at higher temperatures. Further experimentation is necessary, however, in order to determine whether temperature is the sole controlling factor. In some respects the behavior of the sweet potato is in marked contrast to the behavior of resting storage organs of plants of temperate regions. In general, it has been found that the accumulation of sugar as a result of starch transformation ceases at temperatures on!}' a few degrees above 0° C. Thus, Fischer (1891} found that in the cortex of trees the regenera- tion of starch takes place at a temperature a few degrees above 0° C, while Miiller-Thurgau (1882) found that in the common potato the accu- mulation of sugar practically ceases at 8° C. In the sweet potato a rapid transformation of starch into sugar in excess of the quantity used for respiration takes place in freshly dug potatoes at temperatures as high as 30° C. At later periods a marked accumulation of sugar takes place in the sweet potato at temperatures much higher than those at which the accumulation of sugar ordinarily ceases in resting storage organs.* The sweet-potato roots exhibit a further peculiarity with respect to the quantitative relations of the substances formed by the conversion of starch. With the exception of soft-wooded trees, where oil results from the conversion of starch, reducing sugars have been obsers^ed as the most usual and most abundant products resulting from starch transformation in resting storage organs. In the common potato Miiller-Thurgau (1877) found that cane sugar is present together with glucose in the proportion of i part of cane sugar to 2.5 parts of glucose, while Appleman (1914) reports in potatoes kept at a temperature around 0° C. for 2>2 months 3.94 per cent of total sugar and 2.40 per cent of reducing sugar. In the sweet potato cane sugar is the principal product formed by the conversion of starch, while the quantity of reducing sugar is small. In warm storage the cane-sugar content of the Big Stem sweet potatoes reached 6.96 per cent and that of the Southern Queen 4.05 per cent, while the maximum reducing sugar contents were, respectively, 1.67 and 0.77 per cent. In cold storage the cane-sugar content of the two types rose to 8.74 and 6.94 per cent, respectively, while the maximum reducing sugar content in the two cases was 2.44 1 The transitory solution initiating the process of translocation of starch in leaves and in the storage organs of plants about to resume active growth, of course, takes place at higher temiieratures. This process seenis to be somewhat different in its nature from the solution of starch in resting storage organs as a result of exposure to low temijeratures. Jan. IS. 1915 Changes in Sweet Potatoes During Storage 341 and 0.77 per cent, the invert sugar content of the Southern Queen having shown no increase in cold storage. In all cases the proportion was approximately 4 to 5 parts of cane sugar to i of reducing sugar. SUMMARY During its growth the sweet-potato root is characterized by a very low sugar content. The reserve materials from the vines are almost wholly deposited as starch. Immediately after the roots are harvested there occurs a rapid trans- formation of starch into cane sugar and reducing sugars. This initial transformation seems to be due to internal causes and is largely inde- pendent of external conditions. Even at a temperature of 30° C. both cane sugar and reducing sugars accumulate during this initial period in excess of the quantity used in respiration, while during subsequent periods the quantity of reducing sugar diminishes at that temperature as a result of respiration. These initial changes seem to be associated with the cessation of the flow of materials from the vines. In sweet potatoes stored at a temperature of 11.7° to 16.7° C, the moisture content remains fairly constant. There is a gradual disappear- ance of starch during the first of the season (October to March) and probably a re-fomiation of starch accompanied by a disappearance of cane sugar during the latter part of the season (March to June). The changes in reducing sugar are less marked than those in cane sugar. The changes in starch and cane sugar appear in a general way to be correlated with the seasonal changes in the temperature. In sweet potatoes kept in cold storage (4° C.) there is a rapid dis- appearance of the starch and an accompanying increase in cane sugar. These changes do not attain a state of equilibrium at that temperature, as the sweet potatoes invariably rot by the action of fungi before the changes have reached their maximum. At both high and low tem- peratures cane sugar is the chief product formed by the conversion of starch in the sweet potato. The quantity of invert sugar in the root at any time is comparatively small. LITERATURE CITED Appleman, C. O. 1914. Biochemical and physiological study of the rest period iji tlie tubers of Solanuni tuberosum. Md. Agr. Exp. Sta. Bui. 183, 226 p., 17 fig. Baranetzky, J. 1884. [Ueber die jahrlichc Periodc der Starkespeicherung in den Zweigen unscrer Baume.] In Bot. Centbl., Bd. 18, No. 18, p. 157. Bryan, H. A., Given, A., and Straughn, M. N. 191 1. Extraction of grains and cattle foods for the determination of sugars: A comparison of the alcohol and the sodium carbonate digestions. U. S. Dept. of Agr., Bur. Chcm. Circ. 71, 14 p. 69733°— 15 5 342 Journal of Agricultural Research voi.in. N0.4 Fischer, A. 1891. Beitriige zur Physiologic der Holzgewachse. In Jahrb. Wiss. Bot., Bd. 22, p. 73-160. Grebnttsky, a. 1884. Uber die jahrliche Periode der Starkespeicherung in den Zweigen unserer Baume. In Bot. Centbl.. Bd. 18, No. 18, p. 157. Haberlandt, Gottlieb. 1876. Untersuchungen iiber die Winterfarbung ausdauemder Blatter. In Sitzber. K. Akad. Wiss. [Vienna], Math. Natnrw. Kl., Bd. 73, Abt. i, p. 267-296. Harrington, H. H. 1895. Water and sugar in sweet potatoes as influenced by keeping. In Texas Agr. Exp. Sta. Bui. 36, p. 628-629. LiDFORSS, B. 1907. Die Wintergriine Flora. In Lunds Univ. Arssk., n. f. bd. 2, afd. 2, no. 13. 76 p., 4 pi. Mbr, E. 1876. De la constitution et des fonctious des feuilles hivemales. In Bui. Soc. Bot. France, t. 23, p. 231-238. MOllER, N. J. C. 1877. Vertheilung des Starkerestes in der Strombahn. /n Ai> Botanische Unter- suchungen, Bd. i, p. 241-245. Heidelberg. MOller-Thurgau, H. 1882. Uber Zuckeranhaufung in Pflanzentheilen in folge niederer Temperatur. In Landw. Jalu-b., v. 11, p. 751-828. Rosenberg, O. 1896. Die Starke der Pflanzen im Winter. In Bot. Centbl., Bd. 66, No. 24, p. 337-340. (For observations on the roots of trees, see Leclerc du Sablon, Sur la variation des reserves hydrocarbonfees dans la tige et la racine des plantes ligneuses. In Compt. Rend. Acad. Sci. [Paris], t. 135, no. 20, p. 866-868. 1902.) Russow, E. 1884. [tjber Tiipfelbildung und den Inhalt der parenchymatischen Eletnente der Rinde.] In Sitzber. Naturf. Gesell. Univ. Dorpat, Bd. 6, p. 350-389. 1884. Uber das Schwinden und Wiederauftreten der Starke in der Rinde der einheimischen Holzgewachse. In Sitzber. Naturf. Gesfell. Univ. Dor- pat, Bd. 6, p. 492-494. ScHULz, Ernst. 1888. Uber Reservestoffe in immergrtinen Blattem. In Flora, Jahrg. 71, No. 14/15, p. 223-241, 248-258. SmvER, F. S. 1901. Sweet potato. II. Changes in composition on storing ; III. Relative value of different methods of storing. In S. C. Agr. Exp. Sta. Bui. 63, p. 6-37. WaEY, H. W.,etal. 1908. Official and provisional methods of analysis, Association of Official Agricul- tural Chemists. U. >S. Dept. Agr. Bur. Chem. Bui. 107 (rev.), 272 p. PRELIMINARY AND MINOR PAPERS THREE-CORNERED ALFALFA HOPPER By V. L. WiLDERMUTH, Entomological Assistant, Cereal and Forage Insect Investigations, Bureau of Entomology INTRODUCTION The small triangular insect of the hemipterous family Membracidae on which this paper is based was first noted and described as Membracis festina by Thomas Say in 1831 (i);' and in 1869 Stal (2, 3) referred it to the genus Stictocephala. Since that time it has frequently been noted by entomological writers, who usually merely mentioned its occurrence in a new locality or repeated what had already been observed. In 1888 this insect was first noted in literature as being injurious (4). However, the species was not generally considered of economic importance until the winter of 1910, when Prof. Herbert Osbom, in a paper (11) read before the American Association of Economic Entomologists, called attention to the economic habits of the genus Stictocephala and gave special attention to the species 5. festina and its economic relation to alfalfa and clover. Prof. Osbom in his paper stated that little or nothing was known of the life history and habits of the species. It is the purpose of this paper to give a report of the same, together with other related data, as collected by the writer, assisted by Messrs. R. N. Wilson and T. Scott Wilson at Tempe, Ariz., and by Mr. Edmund H. Gibson at Greenwood, Miss. SPECIFIC IDENTITY OF THE THREE-CORNERED ALFALFA HOPPER The name "three-cornered alfalfa hopper," adopted for this insect because it is the common term applied to it by farmers throughout areas of heavy infestation, is applicable to both Say's (i) Stictocephala festina and Van Duzee's (10) Stictocephala festina, var. rufivitta. On several occasions Mr. Otto Heidemann has determined a few specimens as 5. festina, var. rufivitta, among material sent to the Bureau of Entomology for identification. As these were secured both by Mr. Gibson in Tennes- see and Mississippi and by the writer in Arizona, one is led to believe that the species and the so-called variety occur rather generally together. Since Van Duzee (10) bases his description of the rufivitta variety upon male specimens only, and since only male specimens among hundreds examined have exhibited the determining character — namely, that "the dorsal carinae are not evanescent before their point of meeting the ^ Reference is made by number to " Literature cited," p. 36a. Journal of Atrricultural Research, Vol. III. No. 4 Dcpt. of Acriculture. WashinEton. D. C. J m. 15, 1915 (343) 344 Journal of Agricultural Research Vol. Ill, No. 4 posterior carinae," both the writer and Mr. Gibson, who has made many observations on this point, feel that this variety has been founded on too slender grounds. Van Duzee's variety angulata has never been taken. DISTRIBUTION Osbom, in his paper (ii) on the genus Stictocephala, points out that S. festina has a wide distribution and is found throughout the southern and southwestern United States. It is certain that in these sections it occurs in the greatest abundance, but its range is not limited to them. (See fig. I.) Say (i) described the species from specimens secured in Florida. In 1889 Provancher (5) reported the species in Ottawa, Canada, and in 1890 Smith (6) gave New Jersey as a new locality. Later Fig. I. — Map showinE; distribution of the three-cornered alfalfa hopper {Stictocephala festina) in the United States. The densely dotted area shows region of injurious infestation; the sparsely dotted area shows region of occurrence in limited numbers. Original. than this (1894) F. W. Coding (8) gave the following localities: Virginia, Pennsylvania, Georgia, Florida, Missouri, Texas, Iowa, l^Iontana, and Colorado (Riley) ; New York and Connecticut (Van Duzee) ; New Jersey (Smith); Canada (Provancher). That the species occurs in very lim- ited numbers in the northern half of the United States is certain. Osborn found in his travels of 1909 and 1910 that 5. lutea has a southern boundary agreeing quite well with the northern boundary of 5. jestina. The writer had the pleasure of examining alfalfa sweepings made by Mr. R. N. Wilson at 12 different localities in Colorado and Utah during the summer of 191 1, and although several species of Membracidae were represented in the collections, not one specimen of 5. jestina was present. The writer has observed the species in abundance throughout the South- western States, and found it injuring alfalfa {Medicago sativa) at Yuma, Tucson, Casa Grande, Tempe, Phoenix, Buckeye, andGlendale, Ariz., while Mr. R. N. Wilson reported it in alfalfa sweepings at Sacaton, Ariz., a region Jan. IS, .915 Three-Cornered Alfalfa Hopper 345 isolated by a desert from any other cultivated area. Specimens were taken by the writer at Bard, Cal., another region remote from cultivated areas. Throughout the Imperial Valley in southern California the alfalfa hoppers were found in injurious abundance, while in Mexico, in the peninsula of Lower California, the pest was taken in numbers. In 191 2 and 1 913 Mr. Edmund H. Gibson found the species well distributed throughout the States of Mississippi and Tennessee and in several locali- ties in Alabama, as well as at Atlanta, Ga. EFFECT OF ALTITUDE ON DISTRIBUTION It is quite interesting to note here* that in Arizona and New Mexico the species is distinctly one inhabiting lower altitudes. During the summer of 1913 a great many observations were made on this point. The highest point, to the writer's knowledge, at which it has been taken is Fairbank, Ariz., where, at an altitude of 3,868 feet, on September 4, 1913, Mr. Harry Newton, then an agent of the Bureau of Entomology, found both nymphs and adults to be quite common on alfalfa. The writer made sweepings at Raton, Cimarron, and Las Vegas, N. Mex., all at an altitude above 5,000 feet, and while many alfalfa insects common in lower altitudes were taken, not a specimen of Stictocephala was secured. At Ute Park, Taos, Embudo, Bluewater, and Gallup, N. Mex., localities ranging in altitude from 5,000 to 8,000 feet, Mr. J. R. Sandige made sweepings from alfalfa and likewise failed to take a single specimen of Stictocephala, although it is known to occur in lower altitudes in the State. Possibly the most striking obser\'ations were those of Mr. R. N. Wilson, made while on a trip through Arizona for the express purpose of securing records on this species. He visited points varying in altitude from 2,000 to 7,000 feet, but never found the species above about 3,000 feet. His note, made on August 25, 1913, giving a summary of the trip, is as follows: The writer returned to-day from a trip over part of Arizona, including stops at Pres- cott, Camp Verde, Williams, Show Low, Pinetop, White River, and Gila River Valley points from Rice to Solomonsville and Miami. Special alfalfa sweepings were made at each of the above-mentioned places to determine whether or not Stictocephala festina occurred in that locality. The highest altitude at which the species was found was 3,000 feet, at Camp Verde. Altitudes varying from this to 7,000 feet were examined, but no trace of S.festina was found. When Rice was reached, where the altitude isonly 2,5oofeet, this species was again found in numbers, and all the way up the Gila River Valley to Solomonsville (altitude 2,985 feet) the Stictocephala were very common. FOOD PLANTS The alfalfa hopper lives on a great variety of food plants. Its general dis- tribution and the fact that it is found in such isolated places under cultiva- tion are doubtless due to the wide range of its food habits and probably, also, to the presence of native leguminous plants upon which, in all prob- ability, it lives. Its favorite foods without a doubt belong to the legume family, for it is particularly fond of alfalfa, cowpeas {Vigna sinensis) , a.nd the various clovers, but it has also been found feeding upon trees, shrubs, herbs, and grasses. The earhest recorded food plant is the tomato, which in 1888 was reported by Dr. Oemler (4) as being injured by this species. The next record we have is Prof. Cockerell's (9) in 1899, when he reported the hopper as feeding on alfalfa. He also mentions its occurrence on almond 346 Journal of Agricultural Research Vol. Ill, No. 4 trees, but does not say whether it was feeding thereon or not. Prof. Osbom (ii) reported it in 1910 as feeding upon both alfalfa and clover. The writer has found the species, feeding as well as breeding on Bermuda grass {Capriola dactylon), Johnson grass {Sorghum halepense), wheat (Triticum spp.), barley {Hordeum sativum), oats (Avena sativa), bur clover (Medicago denticulata) , yellow sweet clover (Melilotus officinalis), and alfalfa, which, as has been stated, is its principal food plant. Mr. T. Scott Wilson took specimens feeding on soy bean (Glycine hispida) at Sacaton, Ariz., and Mr. Edmund H. Gibson, besides reporting the species as feeding upon alfalfa, also finds it feeding upon vetch and Hordeum murinum at Tempe, Ariz., and upon red clover and cowpeas at Greenwood, Miss., and in fact doing its greatest damage to the last- named plant. Dr. A. W. Morrill, vState Entomologist of Arizona, has found the insect feeding upon beans and in some instances proving a pest to that plant. Late in the season one finds the insect resting upon many varieties of plants, but whether feeding on all these is unknown. Mr. R. N. Wilson found it upon the following plants: Sunflower, upon which it was doubtless feeding; cocklebur; A triplex truncata; Erigcron cana- densis and Eriger on sp.; mesquite and cottonwood, feeding on the former; Sporobolus airoides, and Trichlaris mendocina. DESCRIPTION OF THE THREE-CORNERED ALFALFA HOPPER THE ADULT The adults (PI. XLIII, fig. i) are about 6.16 mm. long and light green in color. The accompanying table of measurements (Table I) made by Mr. Gibson shows that the males are shghtly smaller than the females. Table I. — Length of live adults of the three-cornered alfalfa hopper Specimen No. Length of — Male. Female. I . . . ... Mm. 6.0 6. I 5-9 6.0 Mm. 6.0 6-3 6-3 6.0 6.6 ■2 C Average length ** 6.8 6. 24 > Average length of to adults, 5 males and s females, is 6. 16 i The males have a reddish line down the dorsum of the prothoracic shield. This marking, being absent in the female, is a sex character by which mature males and females are quite readily distinguishable. The insects are triangular in shape, presenting a broad solid aspect when viewed from the front (PL XLIII, fig. i, b). The following original de- scription, made by Thomas Say (i) in 1831, was evidently made from male specimens, because, as is mentioned above, the females do not have the "carina tinged with rufous." Thorax with a subacute line each side before, meeting behind the middle. Inhabits Florida. « Jan. IS. I9IS ■ Three-Cornered Alfalfa Hopper 347 Body yellowish-green: thorax unarmed, carinate behind; at tip attenuated, subu- late and complying with the general curvature; each side before a carinate line, meeting together at the carina behind the middle, with the carina tinged with rufous: front of the thorax not altogether fiat, but a little convex; hemelytra, three terminal cellules unequal; the two costal ones equal, as broad as long; the inner one not obviously larger than the others together, somewhat longer than broad. Length to tip of hemelytra one fifth of inch. The lateral prominent lines of the unarmed thorax, separate this species from all those I have described excepting goniphera, which, meet before the middle of the length of the back. THE EGG The egg (PL XLIII, fig. 2, b) is about 1 mm. (0.9 to i .3 mm.) long and 0.35 mm. (0.25 to 0.4 mm.) in diameter. It is white, rather oblong, slightly larger at one end, and with a greater curve on one side. The surface is smooth, except a portion on the larger end which is regularly covered with small papillae. THE NYMPH The nymphs are the same general shape as the adults, but instead of having the prothoracic shield as a body covering, the}' are regularly covered with prominent projections, spines, and hairs. There is one dorsal pair of these projections on the head, four pairs on the thorax, and seven pairs on the abdomen, the posterior pair on the anal segment being much reduced in size. Their general color is as follows : Head and thorax very light straw. Eyes with margin white and center cologne earth. Antennae white. Thorax with a regular cologne-earth patch on each side, widest on the raesothorax, where it reaches the darkest shade. Legs white, except tip of last tarsal joint, which is dark brown. Abdomen white, approaching light green, owing to food material within, with the irregular dark spot of the thorax extending narrowly across the first segment, widening greatly on segments 2, 3, and 4, and showing only on the posterior margin of the fifth, being widest on the posterior margin of segments 3 and 4. Anal segment light-straw color at extreme end. The different nymphal stages, of which there are five, are the same in general appearance, except that the main dorsal projections in the first stage have only one subspine, while in the second and remaining stages there are numerous branches. A second difference is the growth pos- teriorly of the prothoracic shield and the appearance of wing pads in the last three stages. These two differences, with the increase in the size of the body and the general darkening of colors in each successive stage, enable one to recognize any of the different stages. DESCRIPTION OF INDIVIDU.\L ST.^GES Stage I (PI. XLIII, fig. 2,0). — Length, 1.6 mm. (1.4 to 1.7 mm.), average of 10 speci- mens. Head, thorax, abdomen, and all appendages pale when first bom. After feeding the abdomen takes on straw color and the cologne-earth patch of the tliorax becomes faintly visible. Eyes white. Twelve pairsof dorsal hairlike projections with one upright spine. One pair on head, four on thorax (two on prothorax, one each on mesothorax and metathorax), and seven on abdomen. Spines colorless, pale. Uody regularly but sparingly covered with spines, conspicuous and large compared to size of body. Stage II (PI. XLIII, fig. 3). — Length, 2.1 mm. (1.9 to 2.3 mm.), average of 10 speci- mens. Head, thorax, abdomen, and appendages light straw colored. The cologne- earth patch of the thorax becomes more pronounced and extends back onto the abdo- men. Eyes in this and succeeding stages with white margin and brown center. The dorsal projections have become fleshy and bear several lateral spines, the upright spine 348 Journal of Agricultural Research voi. m. No. 4 being reduced in length. Other body spines much more numerous than in first st^e, but reduced in proportionate size. Stage III (PI. XLIII, fig. 4). — Length, 2.9 mm. (2.6 103 mm.), average of 10 speci- mens. Head, thorax, abdomen, and appendages dark straw color, cologne-earth patch in some specimens especially pronounced. Dorsal projections more fleshy and con- taining a greater number of lateral spines. Prothoracic shield beginning to develop. Wing pads faintly visible. Stage IV(Pl.XI,III,fig. 5). — Length, 3.8 mm. (3.5 to 4.1 mm.), average of lospeci- mens. Head, thorax, abdomen, and appendages greenish straw color. Dark patch on thorax and abdomen becoming dark brown, almost black. The color varies greatly, however, some specimens being light green and others ver>' dark throughout the stage. Dorsal projections in this and fifth stage quite fleshy, lateral spines numer- ous. Prothoracic shield with posterior projection extending nearly to end of thorax. Wing pads clearly defined. Stage V (PI. XLIII, fig. 6). — Length, 4.8 mm. (4.5 to 5 mm.), average of 10 speci- mens. Color same as in Stage IV, about the only difference between this stage and Stage IV being the enlarged size. Point of prothoracic shield extending over the first segment of the abdomen and wing pads extending to posterior part of second abdominal segment. LIFE HISTORY AND HABITS The observations on the three-cornered alfalfa hopper have been car- ried through two years and parts of two others at Tempe, Ariz., and through one entire year at Greenwood, Miss. The results, therefore, have been secured under widely differing conditions, the former place being in a hot, semiarid country with an annual rainfall of about 8 inches, while the latter is in a warm, humid country with an average aimual rainfall of nearly 50 inches. In the Salt River Valley of Arizona much difficulty was experienced in securing Itfe-history records during the months of June, July, and August, because of the excessive heat. In order to have the specimens under close observation, it was, of course, necessary to confine them in cages under more or less artificial conditions, and under such conditions the death rate among nymphs was very high. The combined lengths of the egg and nymphal stages under Arizona conditions varied with the tem- perature, being from 35 to 114 days, with an average of about 50 days for all conditions. In Mississippi Mr. Gibson found that a much shorter period was required. Here the variation, as shown by observations on a much smaller number of specimens, was from 26 to 37 days. Rec- ords of a larger number of specimens would doubtless have given a wider variation. EGG STAGE In Arizona the egg stage varies from a minimum of 12 days to a maxi- mum of 41 days, this variation depending upon the prevailing tempera- ture. The average for all records is 22 days. As may be seen in Table II, during an average mean temperature of 59° F. the time required for incubation varied from 33 to 43 days. At a mean temperature of 63° the variation was from 23 to 30 days, while with a mean temperature of about 85° the eggs hatched in a period ranging from 12 to 17 days. Jan. 15. 191S Three-Cornered Alfalfa Hopper 349 Table II. — Length of the egg stage of the three-cornered alfalfa hopper at Tempe, Ariz., in IQ12; host, alfalfa Eggs laid. Eggs hatched. Length incuba- tion. Tem- Date. Num- ber. Date. Num- ber. ture. It* za Feb. 6 6 Mar. 19 22 Apr. q 2 I 'Many. Many. Many. Mar. 7 7 Apr. 22 23 25 May I Apr 25 26 27 May I 2 3 5 2 I 2 I I 6 2 4 3 3 4 I I Days. 30 30 33 34 36 43 34 36 40 41 24 26 °F. n '=59 "59 } '= 16 17 .... 10 Many. 3 4 I 9 June 5 I 0 2 4 2 3 I 2 28 23 i 25 26 III 3° 13 63 May 24 25 June 21 July 5 Many, Many. Many. Many. 6 7 8 6 7 8 10 July 6 I 20 21 22 25 49 12 3 13 4 I 7 3 2 9 3 2 14 16 12 '3 14 16 15 16 17 15 16 83 84 85 87 j8 a Influenced by artificial heat. ^ No way of getting exact count without disturbing the eggs. c Average mean temperature. At Greenwood, Miss., Mr. Gibson was able to get eggs to hatch in the remarkably short time of four days. On June 30 eggs were deposited in cowpea stems, and on July 3 several. had hatched. The writer is unable to surmise the reason for this short duration of the incubation period. The temperature, although no records are available, could hardly have been higher than that recorded during July at Tempe. The amount of humidity may have had something to do with the hasty incubation; then, too, the different host plant, all records from Tempe having been made on alfalfa, may have been a factor in lessening the period. Table III gives the results of Mr. Gibson's experiments to determine the length of the egg stage. 350 Journal of Agricultural Research voi. iii, no. 4 Table III . — Length of the egg stage of the three-cornered alfalfa hopper at Greenwood, Miss.; host, cowpeas Date of oviposition of adults. June 30 . Sept. I. . 3- Date of egg fiatching. July 3 Sept. 8 Length of egg stage. Days. 4 7 5 Num- ber of eggs. Many 8 OVIPOSITION In alfalfa. — ^The egg is deposited beneath the epidermis through a long slit made in the stem of alfalfa by the female with her ovipositor. This slit is often several times the length of the egg. Measurements of a large number of slits displayed a variation in length of from 0.75 to 2.25 mm. The egg is placed either just below or to one side of this puncture, and occasionally, instead of being just under the epidermis, an egg may be found shoved deep within the plant tissues, even to the center of the stem or beyond. Usually only one egg is deposited through a single opening, but sometimes two or more are placed together. Quite often, however, a great many sUts are grouped side by side and the eggs laid singly but giving the appearance of having been bunched through the same opening. When this is the case, a large scar is made, and the place of oviposition would be quite noticeable if it were not for the fact that it usually occurs back of a sheath leaf or at the surface of the ground, where it is partially hidden. Females have been found with their abdomen extended down the stem of the plant and below the surface of the ground, and subsequently eggs have been found in the stems at such places. It has been observed that eggs are usually laid at night or early in the morning. These observations were made, however, during extremely warm weather; during cold weather the females would prob- ably pick out the warmer part of the day to display their activities and thus avoid the minimum temperature, as they doubtless avoid extreme temperature. In COWPEAS. — The method of oviposition in cowpea stems is considerably different from that in alfalfa, the texture of the cowpea plant evidently mak- ing possible the placing of a great many eggs in the stem through one open- ing. Mr. Gibson has found that the eggs are always laid in groups and in his field notes quite aptly refers to these places of o\'iposition as egg pockets. He has observed from i to 1 2 eggs in a pocket. A count of the eggs in six pockets showed respectively 6, 4, 12, 2, 3, and 5 to the pocket. Following oviposition, in about one-third of these pockets a gall formation develops. From their appearance these must be similar to the galls which develop on alfalfa stems following ringing and which are described in the para- graph on alfalfa injury. These naturally give the egg pockets a dis- tinctive and pecuHar appearance. They are often as large as the stem itself, sometimes as much as one-fourth of an inch in diameter, and well show the efforts of the plant toward the heaUng of the injured part. Mr. Gibson thinks that these galls are due to an overproduction of epider- mal cells caused by the physiological stimulus given to the plant by the injury, and states that they are of about the same texture and hardness jao.is, i9>5 Three-Cornered Alfalja Hopper 351 as the plant stem itself. It seems probable that the eggs in pockets where these galls have developed may be so interfered with that they can not incubate, but no definite observations were made on this point. NYMPHAL PERIOD The nymphal period comprises five stages, with a total length of from 22 to 69 days, depending upon the prevailing temperature. As is shown in Table IV, during the cooler spring month of March the total length varied from 42 to 69 days, while in the hot month of July the variation was from 22 to 37 days. The length of the different stages was found to be very unequal, the last two being found to average the longest, while the fifth might be prolonged almost indefinitely, provided food or other conditions were not right. This in itself is an important point, for the species would be able to survive for some time on only a minimum of food. There was found to be very little relation between the length of the periods and the sex of the individual. 352 Journal of Agricultural Research Vol. III. No. 4 ■ft. ■Cl. -a. "1?. "l^ I >■ < OS € 3& S2. 5^9 ►J -3 31 a •°a 2-5° C<1J3 3 •-fs i! 0*0 5 " a Q s " ►J-s ii Co S«a °.g. 0 2 0 +++ ttft^l^^W W^WIM ' I + r ft rri intn (^o j^o .(>.(<.(-. t>00 00 OC » M ■* t^ r^ O.O0 0^-«5«r^r-(0!2 rnuiOa-OOOOvC 1 1 »0 ^ lx>o♦^iO♦^^o^:.'o^^'o■o a 0*0 ^ oo* OKH »D o♦^^ 'SaO* Nj -x) 'TD f^-O OOCO^-OOO lOMiO M Oi 0000 « l*)M l/^10I--ll 1-o.Oifioo O*r-Oi 00 O Mi*i>ntrtt*«rtw ^OiOooa 5 1 1 = 5^3 Tt-aWMCOooco-o ^00 ^0000 wi'O »^r- II S3 1 t^ 0» t^ "^^ O T&^ftOlM 9i9m r. Oi r% o " a rt^ -o ^ r»j ^ * rroMMOOt^'iMO l>"t«' a < a oi »^oo <« oi t^ < s • aoo po »* f») »■ NoO "^-O "1 •'i 1 »j- Tt" m ^ ^ ^-O t **! »o f) - lI l: l: u Si ^^.^. r ^ « «*1 « f^ c (■ <*) r- M > o o S ^'0 t^tnu a 2-inch mesh. The larvae emerging from the fruit instinctively work their way downward and fall through the screen, and are carried thence by gravity and their own movements through the outlet below into the small container. This arrangement works well with very nearly all the host fruits likely to be used as a source of pupse, such as Mumsops clangi, the rose-apple {Etigcnia jambos), the kamanis {Terminalia cattapa and Calophyllum inophyllum), and the strawberry guava (Psidium cattle yanum) . Such fruits as the mango (Mangifcra indica) can be used, but they yield so much juice that the inside of the frame becomes so wet that the escaping larvae often pupate without falling through into the container below, and the sand in the latter becomes more or less saturated. Previous to the adoption of this method of securing pupae, there was in general use, both by the Hawaiian Board of Agriculture and Forestry and by the writers, a shallow box about 14 by 12 by 3 inches, ordinarily used by florists. The bottom of this box was covered to a varying depth with sifted sand, on which were placed the infested fruits. During the height of the season of 1913 from five to eight men were employed by the Terri- torial and Federal authorities in daily removing infested fruits from one box to another containing fresh sand, and in sifting for the pupae the sand over which the fruit had lain for 24 hours. Besides requiring much time, the prolonged sifting necessary to free the pupae injured many of them. The writers are now using a frame 6 by 3 feet and 3 feet in depth similar to that shown in Plate XLIV, figure 2. The use of a sufficient number of cheap wooden frames covered with tin well painted will make it possible for one man in several hours to do the daily work formerly requiring six to eight men. An adaptation of the old box system has been found useful when it has been desirable to keep separate the pupae from small lots of fruit gathered Feb. 15, 191S Mediterranean Fruit Fly 365 from various localitites. One box with a screen bottom of a sufficiently large mesh contains the fruit and is placed over a box of the same size containing sand. By this method the fruit does not need to be handled daily, and the sand below is kept so dry that the larvae falling through into it are easily sifted (PI. XLIV, fig. i). EMERGENCE OF hARVM AND PUPATION The larvse leave the fruit in largest numbers at or just after daybreak. Thus, on the 6th of July, 16,624 larvse emerged between 4 and 9 a. m.,as compared with 57 between 9 a. m. and 11 p. m. Siftings made at 6, 7, 8, 9, and 10 a. m., on July 7, yielded 1,006, 448, 171, 95, and 14 larvae, respec- tively, during these hourly intervals. On July 8, at 5.20, 6, 7, 8, 9, and 10 a. m., 152, 978, 369, 72, 31, and 14 larvae, respectively, emerged, as com- pared with 7 larvae during the rest of the day. The mean temperature for the period of larval emergence during which these observations were made ranged from 73° to 74° F. Nearly all puparia are formed in from one to two hours during warm weather. LENGTH OF PUPAL STAGE From the data included in Table I it will be seen that the minimum length of the pupal stage is 6 days when the mean temperature ranges from about 76° to 79° F. During the warmest Honolulu weather the larger proportion of any lot of pupae requires from 9 to 1 1 days before yielding adults. This period may be increased to at least 19 days when the daily means drop to about 69° to 71°. Table I. — Duration of the pupal stage of the Mediterranean fruit fly Date of pupa- tion. Date of emerg- ence. Number of adults emeig- ine. Pupal stage. Mean tempera- ture. Date of pupa- tion. Date of emerg- ence. Number of adults emerg- ing. Pupal stage. Mean tempera- ture. Days, °F. Days. °F. Jan. 16 Feb. I I 16 bg.b Apr. 17 Apr. 29 43 12 69.6 Jan. 18 Feb. 3 I 16 70.0 Do.. . Apr. 30 I 13 72.7 Jan. 19 Feb. 4 38 16 70.0 Apr. 20 May I 24 II 73-3 Jan. 20 ...do 18 IS 70.0 Do.... May 2 10 12 73-2 Jan. 23 Feb. 5 10 13 70.3 Apr. 22 ...do I 10 73-4 Jan. 26 Feb. 10 16 IS 71.2 Do.... May 3 II II 73- S Feb. 13 Feb. 27 130 14 70.8 Apr. 23 May 5 6 12 73-7 Feb. 14 Feb. 28 250 14 70.8 Apr. 24 ...do 9 II 73-9 Feb. 16 Mar. 2 300 14 71.2 Apr. 2S ...do 1 10 73-7 Feb. 18 Mar. 3 300 14 70.3 Do.... May 6 10 II 73-7 Feb. 27 Mar. 13 7 14 70.4 June 4 June 13 I 9 76.1 Do Mar. 14 40 IS 70.3 Do.. Jime 14 3 10 76.1 Do Mar. 15 15 16 70.4 Do.. . June IS 73 11 76.2 Do Mar. 17 17 18 70.9 June 8 June 16 3 8 76.2 Do Mar. 18 3 19 70.9 Do... June 17 IS 9 76.2 Feb. 28 Mar. IS 88 IS 70. 5 Do.. June 18 38 10 76.1 Do Mar. 16 56 16 70.9 Do.. . June 19 53 II 76.1 Do Mar. 17 14 17 71.0 Do.. June 20 33 12 76.1 Mar. I Mar. 16 29 15 70. 2 Do.... June 21 5 13 76.1 Do Mar. 17 9 16 71. 1 June ID June 16 5 6 76.4 Mar. 14 Mar. 26 I 12 73-9 Do.... June 19 14 9 76.3 Do Mar. 27 I 13 73-7 Do.... June 20 loi 10 76.3 Apr. II Apr. 22 4 11 JI.3 Do.... June 21 160 11 76.2 Do Apr. 23 8 12 71.4 Do... June 22 7 12 76.1 Do Apr. 24 3 13 71.5 Do... June 23 3 13 7(S-3 Apr. 13 ...do 6 II 72.0 July 13 July 20 I 7 79-2 Do Apr. 25 4 12 72.9 Do... July 21 3 8 79.3 Apr. IS Apr. 26 2 II 72.4 Do.... July 22 39 9 79.3 Do Apr. 27 3 12 72..'! Do.... July 23 43 10 79.1 Apr. 17 Apr. 28 '5 II 72.6 Do.... July 24 3 II 79-0 366 Journal of Agricultural Research voi. m. no. $ It was found that in Bermuda, when the monthly means range from 62.5° to 64.8° F., the pupal stage was lengthened to about 31 days under normal conditions. The writers have found that the Mediterranean fruit fly can pass from egg to adult if kept in the dark in cold storage at 56° to 57° and that at this temperature practically all pupse yield adults from 37 to 41 days after pupation. Pupae placed in cold storage in the light at a temperature varying between 58° and 62° were apparently unafiFected by the cold, except that the length of the stage was increased to from 29 to 31 days for pupse which were about 3 hours old when placed in cold storage. In carrying pupae from place to place for rearing purposes a temperature of less than 56° to 60° is not advised, as great mortality occurs. Thus, from about 300 pupae i day old placed in cold storage at about 50° on June 2 and removed to a normal summer tem- perature at Honolulu on July 22 only 8 adults emerged during the period from July 24 to 26. ADULTS The adults of the Mediterranean fruit fly emerge in largest numbers early in the morning during warm weather and more scatteringly during cool weather. CARE OF ADULTS As adults die in greatest numbers within 48 hours after emergence, or within 72 hours at the longest, if food is not given them, those required for future observations must be transferred to a place where they may be fed and cared for daily. In this work the writers have found glass jars 9 by 12K inches covered with cheesecloth very convenient. Such jars will hold from 200 to 300 flies in good condition. Fruit juices of almost any sort are eagerly eaten. Water slightly sweetened with pine- apple {Ananas ananas) syrup was used with good results by the writers for many months, but was replaced later by a mixture of water and finely di- vided parts of papaya {Carica papaya) . When fed with such diluted food, adults thrive best on two feedings a day, one in the morning and one late in the afternoon. The food may be applied in finely di\'ided drops to the sides of thejar by flirting the mixture forcibly against the cheesecloth covering by means of a snapping movement of the thumb and forefinger. The adults feed greedil}' and soon become distended. In this condition many fall and rest upon the bottom of the jar; hence, the less food falling on this portion of the jar the fewer will be the deaths resulting from entanglement in it. If the flies are not required for oviposition, they can be kept alive far more easily by suspending within the jar a juicy fruit upon which they may feed. One mango, the skin of which had been broken in numerous places, served to keep alive from 100 to 200 flies for one week. Feeding by flirting mixtures through the cloth co\'er- ing causes the sides of the jar to become soiled quickly and necessitates Feb. IS. 1915 Mediterranean Fruit Fly 367 the changing of adults to clean jars every two or three days. When fed with suspended fruit, the flies require no changing for at least two weeks. Mortality among the flies increases rapidly in badly soiled jars. LENGTH OF LIFE Well-fed Mediterranean fruit flies have been kept alive in these jars more than five months. One fly that emerged on December 31, 1913, lived until May 11, 1914, or 131 days. Other flies, which emerged on February 28, 1914, are alive at the date of this writing — August i. Usually about 50 per cent of the flies may be expected to die during the first two months after emergence. When the monthly mean tempera- ture averages about 76° to 79° F., comparatively few flies live to be more than 3 months old. When kept at a temperature ranging from 58° to 63°, the writers believe from accumulating data, that especially strong adults will live to be more than 6 months old. SEXUAL M.\TURITY Neither male nor female flies are sexually mature when they emerge from the pupa. Males show sexual activity often four days after emerg- ence, and copulation has been observed five days after emergence. When the daily mean temperature averages from 76° to 78° F., the larger per- centage of females is ready to mate from six to eight days after eclosion. Adults that emerged on May 23 and 24 and were placed on May 25 in the light in cold storage at 61° to 64° were not observed to mate until June 5, when 14 days old. Copulation may occur at any time throughout the day. OVII'oSITION Oviposition may take place in Hawaii as early as 5 days after emergence during very warm weather, but not for about 10 days when the tempera- ture ranges between 68° and 72° F. At mean temperatures above 74° various lots of adults will yield large numbers of eggs from 7 to 8 days after emergence. Adults oviposit best at temperatures varj'ing from 70° .upward, but have been observed depositing eggs at 65° to 67°, and in cold storage in the light at about 62°. It is impossible at this writing (August 1,1914) to state the full capacity for egg deposition possessed by females of this species. The number of eggs found at any time in the reproductive organs is no indication of the total number of eggs an individual female is capable of depositing, for new eggs are being formed continually throughout life. The data in Table II show that during the first 18 weeks of her life one adult deposited 499 eggs and was still in a thrifty condition. Two other females during the same time deposited 416 and 336 eggs, respectively. A fourth female living but 80 days deposited 312 eggs. Usually females die soon after 368 Journal of Agricultural Research Vol. Ill, No. 5 they cease to oviposit. Fly No. 9 in Table II is an exception, as she de- posited but 3 eggs in one puncture during her life of 68 days and lived without ovipositing for 35 days before she died. The data in Table II give the capacity for oviposition possessed by females up to 18 weeks of age. Those in Table III show that this capacity is fairly well main- tained by certain females at least during the fifth month after emergence. Thus fly No. 5 of Table II deposited on an average 4.5 eggs per day after she began ovipositing, while fly No. i of Table III deposited an average of 4.6 eggs per day for the first 24 days of the fifth month of her life. Table II. — Daily rate of oviposition of the Mediterranean fruit fly. Females emerged on April 4, IQ14, and were placed with fruit on April 14, IQ14 Date of o\nposition.'» Number of eggs deposited. Fly No. I. Fly- No. 2. Fly No. 3- Fly No. 4. Fly No. 5. Fly No. 6. Fly No. 7. Fly No. 8. Fly No. 9. Apr. 16 17 to 20. 20 to 22 . 22 to 25. 25 to 27. 27 to 29. 29 to 30 . May I to 3 . . . 3 to 5 . . . 6 7 8 9 10 II 12 13 14 IS 16 17 18 19 20 21 22 23 24 2S 26 27 28 29 30 31 June I 2 3 3 14 II o o 25 2 8 9 9 3 10 4 S o o o 9 I o o 17 0 4 4 14 0 8 4 8 0 .S 0 8 0 II o 7 13 6 o 4 (*) 0 0 0 0 20 2 19 2 6 24 19 13 16 17 25 12 O 19 9 7 o 7 10 7 10 5 II o 3 5 6 <» Dates on which none of the l> Died on this date. « Escaped on this date. o 4 18 4 9 IS 3 9 12 9 8 13 9 flies oviposited are 6 9 3 3 o 6 3 S 12 14 o 15 16 19 o 12 3 7 6 5 14 8 13 10 6 4 omitted from o 5 o 14 o 23 19 o 9 2 o 8 II 9 6 8 3 4 4 3 o 5 3 7 10 21 5 14 7 ' the table. o O 2 O 20 O 13 o o 3 o o o o o I 14 9 II 8 0 4 7 3 3 8 4 7 4 8 S 6 2 Mediterranean Fruit Fly 369 Table II.- — Daily rate of oviposition of the Mediterranean fruit fly. Females emerged on April ^, ip/4, and were placed with fruit on April 14, igi4 — Continued Date of oviposition. Number of eggs deposited. Fly- No. I. Fly Xo. 2. Fly No. 3- Fly No. 4. Fly No. 5 Fly No. 6. Fly No. 7. Fly No. 8. Fly No. 9. June 4. 5- 6. 7- 8. 9 10. II. 13- 14- 15- 16. 17- 18. 19. 20. 21. 22. 23- 24. 25- 26. 27- 28. 29. o°- July I 2 . 3- 4- S- 6. 7- 8. 9- 10. II . 12. 13- 14 'S- 17- 19. 20 , 23- 24- (a) IS 6 12 o 4 10 o o o 5 16 6 o 3 10 3 7 o o o (a) 4 2 0 7 0 9 0 6 0 0 0 6 4 4 0 10 2 9 7 9 6 14 5 2 7 3 6 2 o 9 2 5 2 4 o o 14 o II II 3 4 7 2 4 5 o 3 12 7 o («) Total . 191 34 86 314 '499 312 9 12 3 o 7 12 o 6 2 7 3 8 2 4 o 9 3 9 5 3 12 I 9 2 6 6 3 2 3 8 4 9 5 o 12 3 2 4 4 2 2 8 I I 6 3 o o 2 4 4 2 O O S 3 o 4 o o o &4I6 ^ii(> C) A Died on this date. ^ Aug. I, 1914: These females are still alive and promise to ovii>osit for some time yet. 370 Journal of Agricultural Research Vol. III. No. 5 Table III.- — Daily rate of oviposition of the Mediterranean fruit fl-y. Females emerged on February 28, IQ14; hence, were 4 months old on June 28, 1914 Date of oviposition. " July I 2 3 4 5 6 7 8 9 10 II 13 14 15 16 17 18 19 20 21 22 23 24 Total Number of eggs deposited. Fly No. 1. Fly No. 2. Fly No. 3. Fly No. 4. Fly No. 5. Fly No. 6. 10 4 5 "•So h 4 o 4 6 7 o 13 71 C) o 6 o o 12 6 13 2 4 {") 43 a Dates on which none of the flies oviposited are omitted from the table. b Died on this date. c Aug. I, 1914: These females are still alive and give promise of sexual activity for some time to come. Oviposition experiments are still in progress; hence, no estimate can be made of the maximum egg-laying capacity of the female Ceratitis capitata. It is evident, however, from the data in Tables II and III that the females lay small batches of eggs quite regularly throughout life. DIFFERENCES IN HABIT BETWEEN THE ADULT MEDITERRANEAN FRUIT FLY AND THE ADULT MELON FLY Those desiring to rear the melon fly {Bactroccra ciicurbiiae Coq.) or parasites of its eggs and young larvae will find marked differences in habit between this and Ceratitis capitata. These same differences will prob- ably be found to occur between species of Dacus and Ceratitis. There is very little difference found by the writers in the egg, larval, and pupal stages. The adults of the melon fly are far more hardv than adults of C. capitata. The writers have on hand many adults 6 months old which give every promise of living indefinitely, as very few have died during the last few months and those living are as active as when newly emerged. The adult melon fly exhibits no sexual activity for such a long period Feb. Mediterranean Fruit Fly 371 after emergence that one is likely to be discouraged in obtaining eggs. While the sexes of the Mediterranean fruit fly are sexually active through- out the day, the melon flies become active only at sunset. From sunset until dark copulation occurs and lasts in many instances until daybreak, inasmuch as numerous pairs have been observed in coition at midnight and at dawn, when all flies are very quiet. Adults issuing from pupae on May 24 did not mate until June 13, or 20 days after emergence, although they were observed every evening. The majority of females in this lot did not mate until 25 days old. The daily mean temperatures for the period from May 24 to June 13 averaged 75.5° F. The female Bactrocera cricurbitae is more irregular in her habits of oviposition. As shown by the data in Table IV, she lays more con- sistently a large number of eggs at one time. Table IV. — Daily rate of oviposition of the melon fly (Bactrocera cucurbitae). Emerged on May 2J atid placed separately with fruit on June 25, IQI4<' Number of eggs deposited. Date of oviposition.*' July 10. II. . 15 •• 17.. 18.. 19.. 21. . 22. . 23- 24.. 26.. 27.. Fly No. I. Fly No. 2. 0 0 14 0 0 0 19 0 0 0 0 0 13 0 2g Fly No. 3. o 13 o 9 o o o Fly No. 4. Fly No. s. 10 o o o Fly No. 6. 23 o 12 o o o 6 o o o 23 Fly No. 7. o o o o o 19 o o o o o . o a These 7 females were all alive on July 27. & Dates on which none of the flies oviposited are omitted from the table. The data in Table IV were secured from young females during the early period of sexual activity. Other data on file show that females over 5 months old deposit quite as freely. EGGS Eggs may be obtained most easily for experimental work by suspend- ing fruit on a string in a jar containing adults, after the latter have begun to mate. In Plate XLV, figure i, is illustrated this simple method of obtaining eggs during a known period. If the epidermis of the fruit is shaved off in several places oviposition will be made easier. The removing of eggs either from the body of the female or from the sides of the containing jar, as practiced by several workers, has not given good results. If it is desired to keep constant watch over eggs they may be dissected easily from the egg cavity and spread upon a section cut from any firm 372 Journal of Agricultural Research Vol. in. No. s leaf. The leaf should then be inserted into a small vial, an absorbent- cotton plug added to force the leaf well toward the bottom, and the vial with contents inverted and partially submerged in a jar of water. It has been found by checks that eggs, when handled in this manner, develop normally unless injured in the transfer. An ordinary moist chamber does not seem to serve the purpose so well. Table V. — Duration of the egg stage of the Mediterranean fruit fly Num- ber of eggs under obser- vation. Eggs deposited. Eggs hatched. Aver- age mean tem- pera- ture. 35° 264 13s 236 12 102 69s 28 3 50 102 176 243 13 2 44 90 72 77 60 12 63 24 134 128 20 74 18 2 lOI 10 Jan. 21-22, 4 p. tn. to loa. m . Mar. 9 Mar. 22-23, 2 p- m. to 9 a. m. . Do Do Do Mar. 27, 9 a. m. to I p. m. . . . Do Do Do Mar. 27, a. m May 19, 3 p. m. to 6 p. m . . . May 12-13, 3 p. m. to 12 m. . Do June 17-18, 4 p. m. to8a. m. . Do June 18, 1.30p.m. to 3.30 p. m June 19, 10 a. m. to I p. m . . Jime 19-20, 4 p.m. to 8.30 a. m June 20, 9 a. m. to 4 p. m . . . June 23, 10 a. m. to 4 p. m . . Do June 24, 1.30 p.m.to 4.30 p.m. July I, 1.30 p. m. to 5p. m. .. July 15. 3-3° P- m. to 4.30 p.m. Do Do Do July 15-16, 12 m. to 5 p. m. . Do Do Do Nov. 13-14, 4 p. m. to 9 a. m . Do Jan. 26, 6 a. m. to 3 p. m Mar. 12-13, 4o°P- m. to8a. m. . . Mar. 26, a. m Mar. 27, a. m Mar. 28-29, 8 a. m. to 6 a. m. . . . Mar. 29-30, 6 a. m. to 10 a. m ... . Mar. 30, a. m Mar. 30, a. m., to Mar. 31, a. m. . Mar. 31-Apr. 1,9 a. m.toSa. m. . . Apr. 1-2, 9 a. m. to 8 a. m Mar. 31, a. m May 22, 7 a. m. to 10 a. m May 15, a. m May 15-16, 2 p. m. to 8 a. m. . . . June 20, II a. m June 19-20, 6 p. m. to 8 a. m ... . June 2&-21, 6 p. m. to 8 a. m. . . . June 21-22, 6 p. m. to 8 a. m. . . . June 21-22, 6 p. m. to 9 a. m. . . . June 22-23, 6 p. m. to 730 a. m .. . June 25-26, 6 p. m. to 6 a. m . . . Jtme 26, 7 a. m. to 11.45 a. m . . . June 26-27, 4.30 p. m. to 6 a. m . . . July 3-4. 5 P- ™- to 6.30 a. m. . . . July 17, 4.30 p. m. to 6 p. m . . . . July 17, 6 p. m. to 8 p. m July 17, 9 p. ra. to 10 p. ra July 18, 10 a. m to 4 p. m July 18 July 19, a. m July 19, 9a.m.toip.m July 20, p. m Nov. 16, 2 a. m. to 6 a. m Nov. 16, 6 a. m. to II a. m 3>442 °F. 68.7 70. 2 71.0 71. o 71-3 71. o 7I' 71 71 71 71 76, 75' 75' 77' 77. 77. 76.6 77.0 77.0 76.8 76.8 77.0 77.0 78.9 78.9 78.9 78.3 80.0 79.8 79.8 79- S 7S-S 75- S Plate XLV, figure 2, is reproduced from a photograph of an apple that had been hung in a jar with flies for one day. Each dark spot repre- sents a puncture. The entire apple was estimated to contain over 2,000 eggs. As the females live over long periods and oviposit freely through- out life eggs may easily be obtained daily for parasitic work while experimenters are en route from one country to another. Apples are Feb. 15, 1915 Mediterranean Fruit Fly 373 probably the most satisfactory fruit for egg deposition during a voyage, as they may be had at almost all points and they keep for a considerable length of time. While the female shows decided preference for certain fruits, she will oviposit, when forced, in almost any fruit if oviposition is not prevented by physical conditions. Fruit in a hard and semiripe condition is better for oviposition than fully ripe fruit, as the latter is likely to be more juicy, and very juicy fruits often cause a high mortality among eggs and young lar\'ae. During very warm weather eggs hatch in about two days. It will be seen, however, from the data in Table V that the length of the egg stage is considerably increased by lower temperatures. At a mean temperature of 78.9° F. 134 eggs hatched between 49 and 50 hours after being deposited, although 12 eggs deposited at the same time did not hatch until from 66 to 72 hours. At a mean temperature of 71° F. 695 eggs hatched within 72 hours, while 3 hatched in from 120 to 144 hours, or about 6 days after deposition. Eight eggs hatched between 4 and 4>2 days after deposition at a mean temperature of 68.7''. At 59° to 62° eggs hatched in cold storage in from 5 to 7 days, and at 54° to 57° in from 7 to 14 days after deposition. LARV^ The larvae pass through three instars, which may be readily distin- guished. Of chief interest in connection with this paper is the length of larval life. The data in Table VI show that this may be as short as 5 or 6 days when the mean temperatures average about 77° F. One larva at this temperature required 14 days to become full grown. The character of the fruit often influences the length of the larval stage. In citrous fruits, especially in limes and lemons, it appears to be longer. Thus larvae require 14 to 26 days to reach maturity in a ripe lemon, as compared with 10 to 15 days in a green peach. Citrous fruits, however, are not desirable for rearing work with either flies or their egg parasites. For successful rearing work, where it is desired to prolong the length of larval life by slightly lowered temperatures, adults should not be permitted to lay more than 50 to 150 eggs in such fruits as the apple. The feeding of larvae in overinfested fruits brings about such a rapid decay that few become well grown. At 56° to 57° F. larvae have become full grown and emerged from slightly infested apples in a refrigerator over a period ranging from 36 to 53 days after hatching. 374 Journal of Agricultural Research \oi. Ill, No. 5 Table VI. — Duration of the larval stage of the Mediterranean fruit fly Num- ber of speci- mens under obser- vation. Approximate period of development . Host fruit. Instar Instar 3- Larval stage. Mean temper- ature for period of de- velop- ment. I 2 I I I I 12 I I I 2 7 6 4 3 i8 12 3 I I 12 '7 14 20 8 3 3 2 I I June 12 to i8 Do Do Do June 19 Do. June 19 I to 25 . June 19 June 19 June 19 June 22 June 22 June 26 June 26 June 26 Mar. 31 Mar. 31 Mar. 31 Mar. 31 Mar. 31 Mar. 13 Mar. 13 Mar. 13 Mar. 13 Mar. 13 Mar. 13 Mar. 13 Mar. 13 Mar. 13 Mar. 13 to 26 to 27 to July I . . to July 3 . . to July 2 . . to July 3 . to July 6. to July 2 . to July 3 to Apr. 10. to Apr. 1 1 . to Apr. 12. to Apr. 13 . to Apr. 15. to 27 to 28 to 29 to 30 to 31 to Apr. I . . to Apr. 2 . . to Apr. 3 . . to Apr. 4. . to Apr. 8. . Papaya. ...do.. ...do ...do ...do ...do ....do ...do ...do ...do Green peach ...do Hard peach Ripe peach .. do Green peach ... do ...do .. .do ...do California lennn. do do. do. do. .do. do. do. .do. do. Mrs. 38 36 36 48 26 48 48 48 48 48 Hrs. 36 48 48 48 30 24 24 30 24 27-5 Hrs. 48 48 48 48 72 72 96 96 216 264 Days, 5- I 5-S 5-5 6.0 5-3 6.0 7.0 7.2 12. o 14+ 10 II 9-5 6 7 10 II 12 13 IS 14 IS 16 17 18 19 20 21 22 26 °F. 77.6 77.6 77.6 77.6 76.4 76.4 76.6 76.6 77.0 77.1 77.2 77.2 77-4 77.8 77-7 69.6 69.8 70. o 70-3 71. o 70. 2 70-3 70-3 70.4 70. 4 70-5 70-5 70.4 70.4 70.4 PLATE XUV Fig. I. — Wooden boxes, 14 by 12 by 3 inches in size, used in obtaining pupae of fruit flies. The upper box, with a coarse screen bottom, contains the infested fruit; from which the larvs drop through the screen into the lower box containing dry sand. Original. Fig. 2. — Contrivance used for keeping the infested fruit free from the sand and bringing the emerging larvs to a central container where they may be gathered quickly. Original. Mediterranean Fruit Fly Plate XLIV Journal of Agricultural Research Vol. III. No. 5 Mediterranean Fruit Fly Plate XLV t Journal of Agricultural Research Vol. Ill, No. 5 PLATE XLV Fig. I. — Method of keeping adult fruit files alive over long periods. The jars in the lower row contain suspended fruits in which the females readily oviposit. Original. Fig. 2. — An apple after having been suspended for one day in a jar containing Mediterranean fruit flies. Each dark spot represents a puncture containing from i to 30 eggs. The apple is too heavily infested for practical work in rearing parasites. Original. RELATION OF SIMULTANEOUS OVULATION TO THE PRODUCTION OF DOUBLE-YOLKED EGGS' By Mavnie R. Curtis, Assistant Biologist, Maine Agricultural Experiment Station INTRODUCTION In an earlier paper ^ it was shown that double-yolked eggs differ greatly in the number of the normal egg envelopes common to the two yolks. The explanation was then offered that two eggs may come together at any level of the oviduct. If the union is anterior to the isthmus ring, a double-yolked egg results. It was also pointed out that while some of the doubling of eggs is no doubt due to the delay or the backward movement of the first egg, nevertheless an unusually rapid succession of ovulations is necessary to account for the occurrence of double-yolked eggs within a clutch. The purpose of this paper is, first, to present further data regarding the structural variations in double-yolked eggs and the relation of these variations to the functional divisions of the oviduct; and, second, to record observations bearing upon the relation between simultaneous ovulations and double-yolked egg production. CLASSIFICATION OF DOUBLE-YOLKED EGGS BASED ON THE NUMBER OF COMMON EGG ENVELOPES In the earlier paper ^ it was shown that yolks with separate vitelline membranes may have a complete set of common egg envelopes, including common chalazal membranes — " chalaziferous layers" (Pis. XLVI, fig. 3, and XLVII, fig. i); or they may have separate chalazal membranes, all their other envelopes being common (PI. XLVII, fig. 2); or they may have some separate and some common thick albumen layers (PI. XLVIII, fig. i); or filially they may have entirely separate thick albumen layers but common egg membrane and shell (PI. XLVIII, fig. 2). A large series of double-yolked eggs shows every possible stage, from cases where the two yolks are flattened together so tightly within the common chalazal membrane that they resemble a single large yolk to eggs in which the doubleness is visible externally by a depressed ring around the shell (PI. XLIX, fig. i). In such cases there is often a thin fold of membrane * studies on the Physiology of Reproduction in the Domestic Fowl. — XI. This paper is the eleventh in a series published in various biological journals and agricultural experiment station bulletins. Papers from the Biological Laboratory of the Maine Agricultural Experiment Station. No. 75. - Curtis. M. R. Studies on the physiology of reproduction in the domestic fowl. — VI. Double- and triple-yolked eggs. /» Biol. Bui., v. ?6. no. 2, p. 55-83. 1914. 'Curtis. M. R. Op. cit. Journal of Agricultural Research, Vol. in. No. s Dept. of Agriculture, Washington. D. C. Feb. ij, 1915 Maine — r 75012°-1.5 2 (37S) 376 Journal of Agricultural Research voi. m, no. 5 projecting into the albumen at the deepest part of the furrow (PI. XLIX, fig. 2). Although the eggs form a continuous series, they can be sepa- rated with reasonable accuracy into three classes or types: Type I. — Double-yolked eggs having the entire set of egg envelopes common to the two yolks. Type II. — Double-yolked eggs having all or part of the thick albumen common to the two yolks, but with separate chalazal membranes. Type III. — Double-yolked eggs in which the yolks have entirely separate thick albumen envelopes but a common egg membrane and shell. RELATION OF THE NATURE OF THE DOUBLING OF THE EGG TO THE SEVERAL FUNCTIONAL DIVISIONS OF THE OVIDUCT The most obvious interpretation of the various types of doubling observed in double-yolked eggs is that the two components may come together at any level of the oviduct from the ostium to the isthmus ring. The eggs classified as type I arise when the union of the two yolks occurs in that part of the duct where chalazal membrane is secreted (probably the funnel and funnel neck); type II, when the union is at any level in the albumen-secreting region more than the length of the first egg anterior to the isthmus ring; and type III, when the union occurs while the first egg is passing the isthmus ring. Observations of the relation of the egg envelopes were made on nearly every double-yolked egg produced at the plant of the Maine Experiment Station during the past year. Data were collected on 131 eggs with two normal yolks and a common egg membrane and a shell. Only 21, or 16. 03 per cent, were united in a common chalazal membrane, showing that they had passed practically the entire length of the oviduct together. In contrast to this, 93, or 70.99 per cent, had separate chalazal membranes but with all or part of their thick albumen common to the two yolks. These eggs showed every possible gradation from two yolks with entirely common thick albumen envelopes to cases where the two component eggs were contained in a very thin layer of common thick albumen. They had apparently come together at every possible level of the duct from a point in the funnel where the chalazal membranes are complete to a point near the lower end of the albumen-secreting region. Seventeen eggs, or 12.97 per cent, had entirely separate thick albumen envelopes. They had evidently come together while the first component was passing the isthmus ring. The large proportion of double-yolked eggs with all or part of their thick albumen common to the two yolks is easily explained by the length of the portion of the duct in which they may come together. A photograph of the oviduct opened out to show the various regions is reproduced in Plate L. There is no visible line of demarcation between the funnel Feb. 15, 191S Simultaneous Ovulaiion and Double- Yolked Eggs 577 (A), where the chalazal membrane and chalazas are probably secreted, and the albumen-secreting portion (B). There is, however, a difference in the general appearance of the glandular ridges and the microscopic character of the glands for a distance of 5 to 8 cm. in an adult laying barred plymouth rock fowl. There is a definite Une of demarcation, the isthmus ring (x) , where the egg membrane begins to be secreted. The length of the duct from the mouth of the funnel to the isthmus ring has been determined for 57 normal laying barred plymouth rock fowls. The mean length was 46.3 cm. The portion of the oviduct in which two yolks can unite and have all the egg envelopes common is probably within the length of the funnel — that is, from 5 to 8 cm. The portion where they will have all their envelopes separate, except the egg membrane and shell, probably roughly approxi- mates the length of the first egg when passing into the isthmus, or from 5 to 7 cm. The union of two eggs in any other part of the duct from the ostium to the isthmus ring (30 to 35 cm.) would result in the formation of a double-yolked egg with all or part of the thick albumen common to both yolks. If we take as the means of the above figures 7, 33, and 6 cm. and calculate the percentage that each is of the total length from ostium to isthmus ring we shall have the expected percentage of double-yolked eggs of each type, if the probability of the union of two yolks is equal at every level of the duct. The number and percentage of eggs of each type observed and the number and percentage expected on the above assump- tion are given in Table I. Table I. — Number and percentage observed of each type of double-yolked eggs and the number and percentage expected, if there is an equal probability of the union of the two components at every level of the oviduct Tyix-. Number. Percentage, Observed. Expected. Observed. Expected. I (all egg envelopes cominon) . 21 9.3 17 19-94 93-98 17.08 16.03 70-99 12.98 15. 22 71-74 13.04 II (separate chalazal membranes; all or part of thick albumen common) III (separate thick albumen envelopes; common membrane and shell) Total 131 131. 00 100. 00 100. 00 The close agreement between the data for the eggs obser\-cd and those expected in each type supports the conclusion that the union of the component eggs occurs indiscriminately at all levels of the oviduct. While the eggs observed formed a graded scries and the divisions of the duct are somewhat rough, nevertheless, variations in classification within any possible range of observation could not reverse the conclusion that the short distances of duct which could yield eggs of the first and 378 Journal of Agricultural Research voi. in, no.s third type and the comparatively long portion where the union of two eggs could result in an egg of the second type are obviously in accord with the observed percentage of the eggs of each type. THE QUESTION OF THE SIMULTANEOUS 0\XaATION OF THE TWO YOLKS OF A DOUBLE- YOLKED EGG On purely theoretical grounds Parker ' explains the origin of double- yolked eggs as the "simultaneous or almost simultaneous" discharge of two yolks from the same or separate follicles. He suggests that prob- ably when the two yolks are inclosed in the same vitelline membrane they come from the same follicle, but that when they are in separate membranes the}- are probably from separate follicles. Glaser - suggests that it is not necessary to assume simultaneous ovula- tion in the case of the two yolks of a double-yolked egg, since the first yolk may remain in the infundibulum until the next normal ovulation. The present author ■'' has further suggested, first, that this delay of the first egg may occur at any level of the duct anterior to the isthmus ring; second, that the first egg may be moved back up the duct by antiperi- stalsis; and, third, that a yolk ovulated into the body cavity may be later picked up by the funnel and inclosed with its successor in a double- yolked egg. It is probable that double-yolked eggs arise from some or all of these causes. However, as was pointed out in the previous paper,'' at least an abnormally close succession of ovulations is necessary to account for a daily succession of double-yolked eggs or of double-yolked eggs laid after a series of normal daily eggs. No bird belonging to the flock of the Maine Experiment Station during the last six )ears has produced double-yolked eggs on successive days, but a large number of fowls have produced double-yolked eggs within a series of normal daily eggs. For example, fowl No. 2K produced a double-yolked egg as the seventh egg of a lo-egg clutch. No. 37M one as the sixth egg of a 1.3-egg clutch, and No. 306K produced double-yolked eggs both as the second and fourth eggs of the same 6-egg clutch. In fact, in 43, or 36.44 per cent, of the 118 cases on which we have complete data the bird which produced a double-yolked egg had laid a normal egg on the preceding day. In these cases it seems certain that the period between ovulations must have been much shorter than the normal period. Further, in 17, or 14.40 per cent, of the 1 18 cases the bird laid normal eggs, on both the preceding and following days. In these cases the evidence for a heightened rate of fecundity is unmistakable, although the ov^ula- tions which furnished the >oIks for the double-yolked egg may not have been simultaneous. 'Parker, G.H. Double hens' eggs, /n Amer. Nat., v. 40, no. 469, p. ij-as, i 6k. 1906. Bibliography, p. »J-3S ' Glaser, Otto. The oriRiii of double-yolked eggs. In Biol. Bui., v. 24. no. 3, p. 175-186. 1913- ' Curtis, M. R. Op. cit. Feb. IS. i9's Simultaneous Ovulation and Double- Yolked Eggs 379 POSSIBILITY OF A RELATION BETWEEN THE RATE OF FECUNDITY AND THE TYPE OF DOUBLING OF THE DOUBLE- YOLKED EGG It has been shown in previous paragraphs, first, that one-sixth of all the double-yolked eggs show by their structure that the two yolks have passed practically the entire length of the oviduct together (16 per cent have the complete set of egg envelopes common to the two yolks); and, second, that in more than one-third (36.44 per cent) of the cases of double-yolked eggs the two yolks must have been ovulated at an abnor- mally short interval, since these double-yolked eggs were laid on days following the production of a normal egg. Or, to put the matter in another way, in 43 out of 1 18 cases the two yolks must have been ovu- lated in less than the normal time, while in only 19 cases did the two yolks pass the entire length of the oviduct together. The anal5'sis may be carried still farther: Of the 43 cases in which an egg had been laid on the preceding day only 7, or 16.28 per cent, were eggs with the two yolks inclosed in common chalazal membranes — that is, even where we have evidence from the egg record of the fowl that the ovulations must have been unusually rapid the structure of the egg indicates that they were simultaneous in only a small percentage of the cases. This is shown in the first column of Table II. This column also shows that in a few of these cases the entire thick albumen envelopes of the two yolks are separate. This indicates that the two eggs have not united until very near the end of the albumen-secreting portion of the oviduct. Table II. — Number and percentage of each type of double-yolked eggs occurring as single eggs, or preceded or followed within one day by a normal egg Typt. Cases in which a double-yolked egg occurred on the day following a normal egg. Cases in which a normal egg was laid on the day following, but not on the day preceding the double-yolked egg. Cases in which a double-yolked egg occurred as a single egg. Double- yolked eggsob- ser\'ed. Percent- age of total. Num- ber. Per cent. Num- ber. Per cent. Num- ber. Per cent. Total num- ber. I (all egg envelopes com- mon) 7 33 3 36.84 39- 76 I8-7S 6 20 2 31- .S8 24. 10 12. 50 6 30 11 31-58 36. 14 68.75 10 83 16 16. 10 II (all or part of thick albumen common) III (all of thick albumen separate) 70.34 13-56 Total 43 36.44 28 23- 73 47 39-83 118 ICO. 00 380 Journal of Agricultural Research voi. iii.no. s It should be borne in mind that the structure of the egg can not be used to measure the time between ovulations. It only registers the level of the oviduct where the two eggs come together. Further, the author '■ has called attention to the fact that it is necessary to assume a difference in the rate of the passage of the two eggs through the oviduct, in order to account for the union of two yolks which did not enter simultaneously. Nevertheless, it seems certain that in cases where the eggs do not unite at the upper end of the duct the time of entrance of the two yolks must have been separated by a measurable period. Table II also shows that some of each type of double-yolked eggs are produced as single eggs, some on the day after a normal egg was laid, and some on a day followed but not preceded by a normal egg. About one-third of the eggs of both type I and type II were produced in each of the three relations to the production of normal eggs, but more than two- thirds of the eggs of type III were single eggs. This suggests that, while eggs of all types, including type III, may result from a heightened rate of fecundity, the most usual cause of the doubling of eggs at the end of the albumen portion is an abnormal delay of the first egg in the oviduct. In this section it has been shown, first, that there is a certain, though small, percentage of the cases of double-yolked eggs in which, judging from the egg structure, the two yolks probably entered the oviduct practically simultaneously, and, second, that there is a still larger per- centage of cases where, as shown by the egg records of the bird, the two ovulations must have occurred within a few (threeorfour) hours. It seems certain that in all cases of the sinmltaneous ovulation of two yolks the complete set of egg envelopes is common to the two. Yet neither the structure of the egg nor the egg record of the bird can prove absolutelv that the time between ovulations has not been considerably reduced. OVARIAN RELATION OF THE TWO FOLLICLES WHICH FURNISH THE YOLKS FOR A DOUBLE-YOLKED EGG Glaser ^ described the pathological ovary of a bird which habitually laid double-yolked eggs. He concluded that in this case the double- yolked eggs arose from follicles secondarily fused. The secondary fusion of follicles resulted in a common blood supply which, when associated with a common state of permeability in the two ova, resulted in their simultaneous maturity and discharge. That the condition at autopsy of the ovary described by Glaser warrants the assumption that the second- ary fusion of the follicles resulted in a common blood supply which was an important factor in the synchronous maturity and ovulation of yolks may, perhaps, be questioned. But at least it offers a suggestion in regard to the origin of double-yolked eggs which is open to investigation. 'Curtis, MR. Op. cit. = Glaser. Otto. Op. dt. Feb. IS. I9IS Simultaneous Ovulation and Double- Yolked Eggs 381 Some observations have been made at this laboratory upon the relation of the two follicles occurring in normal ovaries which have furnished the yolks for double-yolked eggs. A young pullet, No. 1825 (1913 chick), laid her first egg at 1.30 p. m. on November 20, 191 3. This egg was double-yolked. The bird was killed at 4 p. m. the same afternoon, and the ovary was carefully removed and photographed (PI. LI, fig. i). It was perfectly normal with a gradu- ated series of seven enlarging yolks and two distinct follicles. There was no egg in the oviduct. The bases of the two follicles were separated by about 15 mm. of ovarian epithelium, which was covered with small yolks. Another pullet, No. 8053 (hatched in 1913), laid a double-yolked egg at 3.30 p.m. on November 26, 1913. This was her first egg. At 4.35 p. m. the bird was killed. There was no egg in the oviduct. Two nor- mal separate discharged follicles were present on the ovary (PI. LI, fig. 2). The bases of the two follicles were quite distinct, although not situated at a great distance from each other. They were supplied by separate arteries from the same branch of the ovarian artery. The ovary was apparently normal in every respect. It contained a series of six enlarging yolks. In both these cases the only yolks the bird had ever ovulated were the yolks contained in the double-yolked egg. In each case these yolks were from separate follicles which had separate blood supplies. The study of the structure of the egg and of the egg record of the bird has already led to the conclusion that it is not necessary to assume simultaneous ovulation or even an unusually rapid succession of ovula- tions, except in a small percentage of the cases of double-yolked egg production. It, therefore, is important to consider these points in a study of the ovarian relation of the follicles. In the two cases just discussed the double-yolked eggs were both of type II — that is, each yolk was inclosed in an envelope of thick albumen and then the two were inclosed in a very thin common envelope of thick albumen. Plate XLVIII, figure i, is a reproduction of a photo- graph of one of these eggs, that of bird No. 1825. The doubling of these eggs had evidently occurred far down in the albumen-secreting region of the oviduct. We should not therefore expect that the two yolks had been simultaneously ovulated. The cases in which there is good reason for suspecting simultaneous ovulations are those in which the two yolks are inclosed in a complete set of common envelopes. This shows at least that they have traversed the entire length of the oviduct together. An egg of this type was produced on October 18, 1914, by bird No. 139M (PI. XLVI, fig. 3). The conniion chalazal membrane has been partly torn away, in order to dcnionslratc that the two yolks have separate vitelline membranes. The bird laid a normal egg on October 382 Journal of Agricultural Research Vol. III. No. 5 19 and was "killed and examined on the 20th. At the time of autopsy there was an egg in the oviduct. The egg record of this bird from October i to 20 is as follows : October. . . . I 2 3 4 5 6 ' 8 9 10 II (a) 12 13 14 I IS I 16 T 17 18 61 19 I 20 Eggs I I I I I I Cl o Nested, but did not lay. ^ Double-yolked egg. ' Egg in oviduct at autopsy. Plate LII, figure i, shows the ovary of this bird. The ovary is perfectly normal, with a series of six enlarging yolks. As would be expected from the egg record, several discharged follicles were found, ranging in size, as in other normal ovaries, from the large one just dis- charged to those barely visible. The seven largest can be arranged according to size and associated with the eggs laid from October 14 to 20. Follicle A is much the largest and no doubt furnished the yolk found in the oviduct. Follicle B is next in size and evidently furnished the yolk for the egg laid on the 19th. Follicles C and C are practically equal in size and probably furnished the two yolks for the double-yolked egg on the 1 8th. Follicles D, E, and F continue the decreasing series and probably furnished, respectively, the yolks for the eggs on October 16, 15, and 14. All the other discharged follicles are distinctly smaller. Every follicle on the ovary was carefully examined for evidence of the fusion of folhcles or of a common blood supply. (The bird had been injected with starch solution.) All the follicles on the ovary had sepa- rate stalks and each had a single cavity. Follicles B, C, C, D, and E, in common with all the other follicles in that part of the ovary, were supplied by separate small branches from a single large branch of the ovarian artery. There is, then, in this case no evidence that the double-yolked egg has arisen from a fusion of follicles or from a common blood supply, although the structure of the egg indicates that the two yolks have passed the full length of the oviduct together. Since there is e\'idence of simultaneous ovulations in less than one- sixth of the cases of double-yolked egg production and since even in such a case the two follicles may be quite distinct, two simultaneous ovulations resulting from the fusion of follicles are at least a very unusual cause for the production of a double-yolked egg. NATURE OF THE FOLLICLE WHICH PRODUCED A LARGE YOLK WITH TWO GERM DISKS A study of all the abnormal eggs produced at the Maine Experiment Station shows that the doubling of an egg in the ovary is rare. An egg belonging to this class was laid on March i, 1914, by bird No. 31 iK. Externally this egg resembled a double-yolked egg. Its weight was 82.25 gni- It contained one single very large yolk (weight 30.12 gm.), Feb. 15. 1915 Simultaneous Ovulation and Double- Yolked Eggs 383 with two large and apparently normal germ disks (PI. XLVI, fig. i). There was a thin white line visible only part way around the yolk but passing between the two disks and standing out plainly in this region. The size of the yolk, the presence of two germ disks, and this line were the only evidences of doubling. The vitelline membrane was punctured and the yolk removed. The membrane was washed out carefully with salt solution. It contained a single cavity, with no suggestion of doubling, except the slight thickening seen as a white line on the surface of the yolk. The weight of albumen and shell, 44.47 and 7.66 gm., respectively, was comparable to the weight of these parts in a double-yolked egg where the two yolks are separate. This egg clearly belongs to the first class described by Immerman ■ in which the two yolks are in a single vitelline membrane. The bird which laid this egg was killed at 11 a. m. on the following day. A normal egg weighing 45.81 gm. was in the shell gland. The yolk of this egg was normal, having a single germ disk. It weighed 17.7 gm., which is about 59 per cent of the weight of the yolk with the two germ disks. The egg contained as yet only a small amount of thin albumen and no visible shell. It had evidently just entered the shell gland. The contained yolk had without doubt been ovulated the morning of the day the bird was killed, or about 48 hours after the ovulation of the yolk with the two germ disks. The egg would not have been laid until the second day following the one on which the large egg was laid. The ovary of this bird is shown in Plate LII, figure 2. Two of the enlarging yolks have been removed, in order to show the follicles. Only ihree large follicles were present. Two of these, A and B, were about the same size, A being somewhat larger. The stalk of the third, which was considerably smaller, is seen at C. Either A or B must have fur- nished the yolk with the two germ disks. This yolk was nearly twice the size of the normal yolk in the oviduct. In the time which elapsed between the two ovulations the follicle which had produced this enormous yolk had been resorbed to practically the size of the one just ovulated. Neither of these follicles, nor in fact any other on the ovary, showed any evidence that it was composed of two fused follicles. If the double yolk arose from the fusion of two oocytes, it seems probable that this took place at a stage earlier than the formation of the follicle. The practically double size of the yolk contained in a single follicle but with two germ disks suggests that the germinal vesicle may be an important factor in determining the quantity of yolk deposited within a yolk membrane. In the domestic fowl the fusion of follicles in an ovary capable of producing normal yolks must be of rare occurrence. The examination of several hundred ovaries of laying birds has not furnished a single case. * Immerman, Ferdinand. Ober Doppcleier bcim Huhn. 43 p., 5 fig. Basel, 1S99. Inaugural Disser- tatiun. 384 Journal of Agricultural Research voi. in, no.s In cases of general peritonitis accompanied by visceral adhesions, slight superficial adhesions are sometimes found between the follicles containing hardened yolk. These follicles could never produce normal yolks. DESCRIPTION OF A FUSION OF YOLKS WHICH MAY HAVE ARISEN FROM A FUSED FOLLICLE A small egg weighing only 19.88 gm. was laid on October 19, 1913, by bird No. 19K. This egg contained the double yolk shown in Plate XLVI, figure 2, and weighed only 1.45 gm. Neither part had a visible germ disk. The vitelHne membranes were fused at the point of contact and there was a communication between the cavities of the two yolks, if they were actually separate yolks. The bird had been laying normal eggs and continued to do so. Why this very immature pair of yolks was ovulated is difiicult to understand. Since the bird was not killed, the nature of the follicle which furnished them is not known. This and the preceding case are the only ones which have come under our observation in which the two yolks were inclosed in a common vitelline membrane. CONCLUSIONS The various kinds of evidence given in this paper lead to the con- clusion, first, that double-yolked eggs sometimes represent a heightened rate of fecundity and sometimes an abnormally low physiological tone of the oviduct; second, that, even in cases in which the rate of fecundity is high, the ovulations are not always simultaneous; third, from the above it is apparent that the production of a double-yolked egg can seldom be explained as a result of simultaneous ovulations; and, fourth, in cases in which we have the best of reasons for suspecting simultaneous ovula- tions, the two follicles may be quite distinct. It seems quite possible that a heightened rate of fecundity may result in every conceivable shortening of the period between ovulations con- sistent with the daily rhythm in the general physiological activities of the bird. Whether it results in the formation of a double-yolked egg is no doubt determined by the actual length of the period and the following response of the oviduct. SUMMARY (i) Double-yolked eggs with normal separate yolks may have all the egg envelopes common to the two yolks, or they may have some separate and some common envelopes. (2) They may be classified with reasonable accuracy into three groups: Type I. — Double-yolked eggs having the entire set of egg envelopes common to the two yolks. Type II. — Double-yolked eggs having separate chalaziferous layers but all or part of the thick albumen common to the two yolks. Feb. IS, 191S Simultaneous Ovulation and Double- Yolked Eggs 385 Type III. — Double-yolked eggs in which the j'olks have entirely separate thick albumen envelopes but a common egg membrane and shell. (3) Of the eggs studied 16.03 per cent belonged to type I, 70.99 per cent to type II, and 12.98 per cent to type III. (4) A large series of double-yolked eggs show all gradations within and between these groups. (5) The most probable interpretation of this phenomenon is that the two components unite at any level of the oviduct from the funnel mouth to the isthmus ring. (6) The conclusion that the union of the component eggs occurs indiscriminately at all levels of the oviduct is strongly supported by the fact that the percentage of eggs of each type is closely porportional to the percentage of the portion of the duct in which the union of two eggs would give double-yolked eggs of that type. (7) In 36.44 per cent of t"he double-yolked eggs the ovulations which furnished the two yolks must have been separated by an abnormally short interval, since a normal egg had been laid on the preceding day. (8) An examination of the egg structure, however, shows that the two yolks have passed the entire length of the duct together in only 16.28 per cent of the cases in which the ovulations are knowii to have been usually rapid. (9) While a heightened rate of fecundity may result in the production of an egg of any of the three types, 68.75 per cent of the eggs of type III are single eggs. It seems probable that many of them have resulted from the delay of the first egg in the oviduct. (10) The ovary of each pullet which had just laid a double-yolked egg as her first egg contained two normal separate follicles which had separate blood supplies. In these cases, however, the doubling of the egg had occurred near the end of the albumen-secreting region. (11) In a case in which there was evadence from the structure of the egg that the two yolks had passed the entire length of the oviduct to- gether the two follicles were also quite distinct, with separate blood supplies. (12) This, together with the fact that in only a small percentage of double-yolked eggs is there any evidence of simultaneous ovulation, indicates that the fusion of follicles and a resulting common blood supply is by no means the usual cause for the production of a double-yolked egg. (13) A simple normal follicle furnished the yolk with two germ disks; hence, the fusion of the oocytes (if this was the origin of the two germ disks) must have occurred before the formation of the follicle. PLATE XLVI Fig. I. — Large yolk (weight, 30.12 gm.) with two germ disks; found in a large hen's egg- Fig. 2. — Fused immature yolks (weight, 1.45 gm.); found in a small hen's egg. Fig. 3. — Type I double-yolked egg, showing two yolks with separate vitelline mem- branes but inclosed in a common chalaziferous layer. The yolks were slightly pressed apart before photographing, in order to show that the vitelline membranes were entirel)' separate. (386) Simultaneous Ovulation and Double-Yolked Eggs Plate XLVI Journal of Agricultural Research Vol. Ill, No. 5 Simultaneous Ovulation and Doubie-Yolked Eggs Plate XLVII Journal of Agricultural Researcli Vol. III. No. 5 PLATE XLVII Fig. 1. — Type I double-yolked egg, showing two yolks with separate vitelline mem- branes but inclosed in a common chalaziferous layer. The yolks were slightly pressed apart before photographing, in order to show that the vitelline membranes were entirely separate. Fig. 2. — Type II double-yolked egg, showing two yolks with separate chalazal mem- branes but common thick albumen. PLATE XLVIII Fig. I. — Type II double-yolked egg, showing two yolks with some separate and some common thick albumen envelopes. Fig. 2. — Type III double-yolked egg, showing two yolks with all the thick albumen separate. Simultaneous Ovulation and Double-Yolked Eggs Plate XLVIII Juutiialul At^ncultural Research Vol. Ill, N.,. 5 Simultaneous Ovulation and Doublo-Yolked Eggs PLATE XLIX Journal of Asni.ultural Rescanl Vol. Ill, No. b PLATE XLIX Fig. I. — Shell of type III double-yolked egg, which shows external evidence of its double nature by a seam in the shell. Fig. 2. — The inside of the shell shown in figure i, showing the fold of egg membrane which projected between the two component eggs. PLATE L Oviduct removed from a laying bird and cut open along the point of attachment of the ventral ligament. It is opened back, showing the characteristic glandular regions. .4,Fiinnel; B, albumen-secreting region; .Y, isthmus ring: C, isthmus; D, shell gland; and E, vagina. Simultaneous Ovulation and Double-Yolke'l Eggs Plate L ^^^B X ^^H'^fl ^H ^H '^^ B ^^■c ^B ^^H ^^^^^K' H ^''1 ^B ^^^y aM ^^^^^^^H ^^mw '^.^^^^^B * ^^^1 ^^^^B ^^^^^^^^^^^^B ■ ^^^^^^^V ^ ^1 H Journal oT Af.^rii.u llural Rt---,;':iri ti V. 1. 111. N. , S Simultaneous Ovulation and Double-Yolked Eggs Plate LI Journal of Agricultural Researcli Vul. Ill, No. 5 PLATE LI Fig. I. — Ovary of a pullet, showing the follicles which productd the yolks for the double-yolked egg shown in Plate XLVIII, figure i. Fig. 2. — Ovary of a pullet, showing follicles which produced the yolks for a double- yolked egg similar in structure to the one shown in Plate XLVIII, figure 2. 75012°— 15 3 PLATE LII Fig. I. — Ovary of a pullet, showing a series of resorbing follicles, two of which (probably C and C) produced the yolks for the doublc-yolked egg shown in Plate XLVI, figure 3. Fig. 2. — Ovary of a bird, showing the two largest resorbing follicles, one of which produced the yolk with two germ disks shown in Plate XLVI, figure i. Simultaneous Ovulation and Double-Yolked Eggs Plate Lll ^ 4 i^mSO #■1 JS^ cm W' ^^^F ' '- '^\ ^K 1 Journal of Agricultural Research Vol. III. No. 5 BRACHYSxM, A HEREDITARY DEFORMITY OF COTTON AND OTHER PLANTS By O. F. Cook, Bionomist in Charge of Acclimatization and Adaptation of Crop Plants and Cotton- Breeding Investigations, Bureau of Plant Industry The word "biachysm" is suggested as a name for abnormal variations of plants characterized by shortening of the intemodes, without corre- sponding reductions of other parts. Brachysm is to be distinguished from nanism, or true dwarfing, which involves proportional diminutions of many parts, if not of all. Genuine dwarfs, with consistent reductions of all of the organs of the plant, would have little value for agricultural purposes, but man}' brachytic varieties are highly prized. In spite of their shorter intemodes, brachytic varieties often bear leaves, flowers, and fruits as large as plants of normal stature, and sometimes larger. Most of the so-called "dwarf," or "bush," varieties of peas, beans, squashes, tomatoes, and other garden vegetables represent brachysm rather than true dwarfing. The "cluster" and "limbless" varieties of cotton belong to the same category, and the vSan Ramon coffee of Costa Rica affords another example of brachysm in a woody plant. A special interest may be claimed for the cluster and limbless varieties of cotton because they seem to throw light on the nature of brachj'sm and similar abnormalities as phenomena of heredity. As brachytic varieties arise by nuitation and show allernative inheritance in crosses, they illustrate two of the jjlienomena of heredity that have received much attention in recent years. The morphological and physiological relations of such characters must be understood before it is possible to appreciate their practical importance in breeding or their bearing upon general evolutionary problems. SPECIAL FEATURES OF BRACHYSM IN COTTON The special interest that attaches to the phenomena of brachysm in cotton arises from two general facts: That the shortening of the inter- nodes is usually confined to the fruiting branches and that it is usually accompanied by other abnormalities. On account of the specialized structure of the cotton plant, it becomes possible to learn more of the relations of brachysm to other phenomena of heredity than if all of the intemodes of the plant were affected and the other organs remained unchanged. Journal of Agricultural Research, Vol. IH, No. 5 Dept. of AKriruItllrr, Wn«;hincton, D. C. Keb. 15. 1Q15 O 40 (387) 388 Journal of Agricultural Research vonii, no. 5 The cotton plant has a definite dimorphism of branches, no flowers or bolls being produced on the main stalk or the vegetative branches.* The main stalk and the vegetative branches of cluster cottons are not short- ened, but are often longer in brachytic varieties than in those that have normal fruiting branches. The leaves of the main stalk and vegetative branches of cluster cottons are of normal form, but they are larger, of thicker texture, and have longer petioles than those of normal long- jointed varieties. The axillary buds of noncluster varieties usually remain dormant or produce long vegetative branches, but in cluster cottons they commonly develop into short branches and produce one or two bolls. These differences may be considered as direct results of the failure of the fruiting branches to make normal growth, for similar changes occur in noncluster varieties, in plants that have been severely pruned, and also in seedlings that lose their terminal buds through insect injuries or by abortion.- The abnormalities that accompany the shortening of the fruiting brandies in cotton are shown in the fonns of the leaves and the involucral bracts. The leaves of short-jointed fruiting branches often become smaller and more bractlike, while the bracts are often enlarged and leaf- like, and show definite indications of the stipulac and foliar elements that are completely fused in normal bracts. Each of the involucral bracts of a cotton plant represents a modified leaf. The specialized form of the bract results from having the stipules of the leaf greatly enlarged, the petiole entirely suppressed, and the blade reduced and united with the enlarged stipules. The abnormal leaflike bracts and bractlike leaves of cluster cottons show all the stages between normal leaves and normal bracts. In such abnormalities there is usually an obvious relation between the reduction of the blade or the suppression of the lobes of the blade and the enlargement of the stipules of the same leaf, and also between abnormalities of the leaf and of the involucral bracts of the same intemode. This can be understood by comparing Plate LIII, which shows an abnormal reduced leaf and an abnormal enlarged bract, with Plate LIV, which shows a leaf and a bract of normal size and proportions. The converse relation is that when the bracts take a more leaf-like fonn, with an enlargement of the middle or blade element of the bract, it is almost always accompanied by a reduction of the stipular elements, as shown in the leaf-like bract in Plate LIII. When one side of a leaf is reduced or has the lobe sup- • The dimorphic specializations of the branches and leaves of Uie cottott plant have already been de- scribed. (Cook, O. V. Dimorphic branches in tropical crop plants ... U. S. Dept. Asr. Bur. Plant Indus. Bui. 198, 64 p . t) tit;.. 7 pi. 1911. Cook, O. F. Dimorphic leaves of cotton and allied plants in relation to heredity. U. S. Dcpl. Aer. Bnr. Plant Indus. Bui. in, 59 p., 18 fig., s pi. 1911.) ' Abortion of terminal buds is a frequent result of a peculiar disorder of cotton seedlings previously described. (Cook. O. F. Leaf-cut, or tomosis, a disorder of cotton seedlines. In U. S. Dept. Aer. Bur Plant Indus. Circ. 1^0. p. 29-34. r fie- 191.1-) Feb. 15. J9I5 Brachysm 389 pressed, as in Plate LV, the stipule on the same side of the leaf shows a corresponding enlargement.' That brachysm in cotton is confined to the fruiting branches is doubt- less connected witli the fact that the fruiting branches have more direct relations with the floral organs. On account of the definite dimorphism of the branches of the cotton plant, no floral buds are produced on the main stalks or the vegetative branches. A study of brachysm in cotton seems to indicate that such variations represent intermediate stages or combinations of floral and vegetative characters. Brachytic varieties have leaves that are more like involucral bracts, and bracts that are more like leaves than those of normal long-jointed varieties. From this point of view it is easy to understand that leaves of the fruiting branches would be more likely to show abnormal anticipations of the characters of the floral bracts than the leaves of the vegetative branches or the main stalks, for these have no floral buds and represent earlier stages in the development of the plant. INDEPENDENT ORIGINS OF BRACHYTIC VARIATIONS It has been supposed that all of the "cluster" varieties of cotton belong to the same botanical series, but in reality the possession of short- jointed fruiting branches is not a reason for supposing that two varieties are related. Brachytic variations are very frequent and have been found in so many different species and varieties of cotton that the idea of derivation by crossing with a brachytic ancestral type is unwarranted. Brachysm is not known as a normal character of any wild species of cotton, but seems to follow as one of the results of domestication and selection. The same is apparently true in other families of plants. Brachytic variations have arisen independently in several different types of peas, beans, squashes, melons, and other climbing and creeping plants. It is in such plants that the abnormal nature of brachytic variations is most obvious. Under natural conditions, short-jointed variations of climbing plants would be placed at a still greater disadvantage than those of shrubby plants like cotton or coff'ec. If brachytic variations be supposed to represent the formation of a new character, as assumed in the mutation theory of De Vries, it is diffi- cult to imderstand why the same character should arise suddenly in so many different and unrelated types of plants. But if brachysm be con- sidered as a defect or failure of normal heredity, it becomes easier to understand that many kinds of plants might be subject to similar de- rangements. That the brachytic tendencies should appear independently in so many different genera and families of plants makes it reasonable to look for a general interpretation of this class of abnormal variations. ' A further account of these abnormalities, with iJRiires of some of the iotenncdtate forms of bracts and leaves, may he found iu an earlier bulletin. (Cook.O. F. Heredity and cotton brecdinc. U.S. Dept. Agr. Bur. Plant Indus. Bui. 356. p. 69-:^, tiR. 3-ij. pi. 4H^. 1913.) 390 Journal of Agricultural Research \u\. iii. no. s DIFFERENT DEGREES OF BRACHYSM The interest of brachytic variations as a phenomenon of heredity is increased by the fact that the tendency is manifested in many different degrees. The most extreme form of brachysra is represented b\' a com- plete abortion of the fruiting branches, but there is an apparently com- plete series of stages from these completely sterile monstrosities to the normal long-jointed forms. The strictly cluster or limbless forms have the fruiting branches reduced to a single joint or a few very short joints, as shown in Plate LVI. From this extreme condition the intermediate stages run through the ordinary cluster and semicluster types to those that show no reduction of the joints of the fruiting branches. Even in the same field of cotton it is often possible to find a wide range of varia- tions in the lengths of the internodes of the fruiting branches, especiallv in varieties like the King, that produce many brachytic variations. In some varieties, such as the Triumph cotton of Texas, the lower fruiting branches often show a tendency to brachysm not shared by branches farther up. Other varieties show the opposite tendency to form long-jointed fruiting branches on the lower part of the stalk and short- jointed branches on the upper part. Some varieties that usually have long-jointed fruiting branches show the cluster tendency when the growth of the plants is restricted by unfavorable conditions, while other varieties are always short-jointed. There may also be pronounced irregu- larities in the lengths of the joints, even on the same fruiting branch. (See Pis. L\1I and LVHI.) In addition to the fundamental difference between the vegetative and fruiting branches, the basal intemode of the fruiting branches is generally longer than the others and often maintains its length when the others are shortened. In a peculiar variety of Egyptian cotton grown in experi- mental plantings at Bard, Cal., under the name of Dale, most of the fruiting branches develop only this long basal intemode, and produce only a single boll, the other internodes being aborted. (See PI. LXII.) When the leaves, buds, and joints of such branches are suppressed, the boll appears to be borne on a greatly elongated pedicel. The extent to which these abnormalities are often carried may be more easily understood by reference to the photographs of the plant shown in Plate LXI, figure i. The main stalk was rendered completely sterile by the abortion of all of the fruiting branches, though the vegetative branches of the same plant produced a few fruiting branches and ripened a few bolls. Finally the growth of the main stalk was stopped by the abortion of the terminal bud. (See PI. LXI, fig. 2.) In more normal plants of this variety the single- jointed fruiting branches are often accom- panied by one or two other branches of similar form arising from the axillary bud. (See PI. LXII.) In an extreme case like that shown in Plate LXI, the axillary huds abort, as well as the extra-axillarv buds Feb. ij. I9I5 Brachysm 391 that normally produce the fruiting branches. The regular abortion of these axillary buds renders it the more probable that the abortion of the terminal bud was not accidental. The case shown on the left-hand figure of Plate LXII, the transformation of the terminal bud into a flower bud subtended by an abnormal leaf, affords still more definite evidence of abnormality. SHORTENING OF INTERNODES BY DROUGHT The intemodes are always longer under conditions that favor luxuriant growth of the plants than where growth is restricted by drought. Though this relation is general, some varieties shorten their intemodes much more than others. The susceptibility to shortening is sometimes so great that the same variety may be short-jointed like a cluster cotton in some places, while under other conditions it behaves as a normal long-jointed variety. Attention was called some years ago to a case of this kind in a variety called the Parker that had been grown in Texas for several seasons as a long- jointed variety, but behaved as a short-jointed or semi- cluster variety at Del Rio in the season of 1907.' The difference between the true brachysm and this false brachysm induced by changes of external conditions is that the latter is not inher- ited. The false brachysm represents an adaptive change or accommoda- tion to the conditions instead of a definite alteration of the expression relations of the characters. When whole stocks of plants or animals respond in the same way to a change of external conditions, the changes are usually in the nature of accommodations and are readily reversible. But the possibility that an increase in the number of heritable varia- tions toward brachysm might be induced by environmental shortening of the fruiting branches would be worthy of investigation. This possi- bility is suggested by the fact that individual variations in the direction of the cluster habit appear more frequently in some locahties than in others in the same variety of cotton. Thus, it was noticed in the season of 1 9 13 that plants of the cluster fonn were of frequent occurrence in many fields of Durango cotton in the Imperial Valley of California, whereas in a field of Durango cotton at Deep Creek, Va., no cluster plants could be found. Other illustrations of the influence of external conditions are afforded by irregularities in the lengths of the intemodes of the same plant, or even of the same branch. (See PI. LVII and LVIII.) In such cases the * " The Parker variety of cotton showed a pronouuced semiclustcr habit or shortcuiug of the internodes of the fruit branches, a notable departure from the previous behavior of our stock of lliis variety, which had been under observation in several different localities in the preceding years. Apart from the fact tliat every precaution is taken to avoid mistakes in labeling and planting the seed, the possibility of error in tliis case seems to be entirely eliminated by tlie fact that six different selections of Parker were grown at Del Rio and that all of them behaved in the same manner with reference to this change in the lengths of the fruiting branches. Plantings of the same strains from the same lots of seeds in several other localities in Texas in the same season produced no such results." Cook, O. F. Suppressed and intensified cbarac- ersin cotton hybrids. U. S. Dept. Agr. Bur. Plant Indus. Rul. 147. p. 30. 1909. 392 Journal of Agricultural Research voi. m. no. 5 environmental shortening or false brachysm is often accompanied by other abnormalities, like the inherited form of brachysm. On a short intemode intercalated between two of normal length the pedicel is likely to have a sloping or decurrent base, with the end of the intemode pro- longed beyond the insertion of the pedicel, though usually not so much as in cluster cottons. This irregularity is less difficult to understand when it is remembered that the floral bud of each intemode develops in advance of the growth of the next intemode. On account of this sequence of development, two structural elements may be recognized in the intemodes, the tissues that are developed early to support the pedicel and those that develop somewhat later in connection with the next internode. If a change to more favorable external conditions occurred during the growth of an intemode, its effect must be greater upon the part of the intemode that is the last to reach its full development. Thus, in connection with the development of a longer internode next to a short one, a slight elongation of the supporting part of the short internode would be induced, and a resulting tension between the two sides of the intemode. This would account for the tearing of the tissues at the base of the pedicel, which often occurs. Hence, we see that a part of the abnormality of the inter- nodes that accompanies brachysm may be capable of a simple mechanical explanation, as arising from changes in external conditions during the period of development of the affected intemodes. RETENTION OF BLASTED BUDS IN BRACHYTIC VARIETIES The facts of brachysm in cotton seem to indicate that the shortening of the intemodes of the fruiting branches is in the nature of a premature or accelerated expression of the floral character. This view seems preferable to the idea that a new character has been substituted for an old one in the mechanism of transmission, as usually assumed in theories of mutation and MendeHsm, That the shortened intemodes partake of the nature of the pedicels of the flowers is indicated not only by their reduced length but by the fact that they are often completely fused with the pedicels. At the base of a noraial pedicel is a joint or layer of spe- cialized tissue indicated externally by the absence of hairs and oil glands from the surface, but in cluster cottons the formation of this specialized layer is irregular. It is one of the recognized peculiarities of cluster cottons that abortive buds and those that are infested by boll-weevil larvae often remain attached to the plant, whereas in varieties with nor- mal fruiting branches the blasted buds are soon shed. The lack of definite differentiation between the pedicels and the joints of the branches is responsible for the more frequent retention of the buds in cluster varieties. Normal shedding of the buds takes place by the formation of a circular fissure at the base of the pedicel, the subsequent wilting of the bud, and Feb. .5. 1915 Brachysm 393 the breaking away of the central pith of the joint. The circular fissure is formed just above the shght ridge or rim that connects the base of the stipules, the position being marked in advance, as already stated, by a narrow zone of smooth skin without any of the hairs and oil glands that are scattered over all of the neighboring surfaces. The less definite diS'erentiation of internodes and pedicels in cluster varieties often interferes with the formation of a normal circular fissure for the shedding of the abortive buds, which then remain hanging on the plant. The lower side of the base of the pedicel, the side that faces the main stalk of the plant, is found to be more or less confluent with the surface of the supporting internode. The zone of smooth tissue that indicates the position of the fissure, instead of being circular, may extend far down on the internode ; or all indication of a fissure zone may be lack- ing, so that the pedicel appears as a direct continuation of the internode. The result of such malformations is that the blasted buds, instead of promptly falling off, turn brown and shrivel while still attached to the plant. The casual observer is likely to suppose that the shriveled buds have been stricken by a blight, and there is often a strip of dead tissue running down on the internode from the base of the withered bud as though a bacterial or fungous disease were extending along the branch. Though often mistaken for a diseased condition, such injuries are merely the mechanical consequences of the abnormal structure of the internode and the failure to form a properly specialized joint between the internode and the pedicel of the floral bud. The formation of the joint is usually indicated, for the death of the epidermis commonly follows a definite line, even when the bud does not drop off. A complete separation of the underlying tissues allows the shriveled bud to fall away eventually, leaving a long scar extending down the internode, instead of a normal circular scar at the end of the internode. (See PI. LVIII.) Decurrent pedicels are not confined to brachytic varieties, but are often found in abnonnal individuals of long-jointed varieties, though usually the internodes that have the decurrent pedicels are shorter, and other indications of abnormality may be present. Thus, in connection with the examples of decurrent pedicels shown in Plate LVIII there is an abnormal inequality in the lengths of the internodes, one being about five times as long as the others.' MORPHOLOGY OF DECURRENT PEDICELS As already noted, extreme cases are sometimes found in which the pedicels are not only decurrent upon the internodes but seem to lose their * A somewhat different interpretation of the decurrent pedicels of tlie cotton plant is presented by Prof. Francis I;. Lloyd. (Lloyd, F. U. Abscission, hi Ottawa Nat., v. ;8, no. ,1/4, p. 41-51, 3 fic-; no. S/6, p. 61-75. 1914) 394 Journal oj Agricultural Research voi. ui. no. s terminal positions and to arise from intermediate points. Examination of such cases at first suggested the idea that the floral bud might belong morphologically to the next internode below. In this view the pedicel would be merely coalesced wth the supporting internode, instead of being produced from it. The branch morphology of the cotton plant is unusually complicated on account of the dimorphism of the branches and the extra-axillary position of the floral buds and of the buds that give rise to the fruiting branches.' The assignment of the floral bud to the internode below seems to be forbidden by the fact that when the base of the pedicel becomes decurrent upon the supporting internode the stipule that subtends the pedicel and the stipular rim that incloses the base of the pedicel also become elongated and decurrent along the side of the internode. Thus, instead of merely a lower insertion of the flower bud, the whole internode is modified, and the nature of the modification makes it clear that the floral bud is borne normally above the stipular rim. Though other families of plants aflford instances where flower stalks or floral branches remain united with the basal portion of the next internode, it is very difficult to believe that an adnate or coalesced pedicel would be able to surmount the stipular rim and climb, so to speak, into the axil of the leaf of the internode with which it had become coalesced. In rare cases the pedicel of a cotton boll is joined to the internode above, but the result is clearly abnormal, and lends no support to the theory of coales- cence of pedicels and internodes as a normal condition. (See PI. LX.) A pedicel that is united with an internode is usually longer than the normal pedicels, while the internode is shorter than the others and is turned from its normal position to follow the direction of the pedicel to which it is attached. There is no tendency in these adherent pedicels to form an angle or a joint at the end of the internode, as might be expected if the theory of coalescence were correct. The presence of a flower bud on the basal internodeof fruiting branches offers another difficulty under the theory of coalescence. If it were assumed that each floral bud belongs, not to the supporting internode, but to the one lower down, it would necessarily follow that the floral bud of the basal joint of a fruiting branch could not belong morphologically to the fruiting branch at all, but must be assigned to the main stalk or the vegetative branch from which the fruiting branch is producd. This supposition would add new elements of complexity to the structural morphology of the cotton plant, already sufficiently complicated. BRACHYSM ACCOMPANIED BY FASCIATION AND ADHESION Further reasons for looking upon brachysm as a failure of normal differentiation of parts are found in the other abnormalities of the ' Cook. O. V. Morphology d cotton branches. In U. S. Dept. .A.gr. Bur. Plant Indus. Circ. loo. p. 1 1-16. 19U Feb. IS, 191S Brachysm 395 branches that often accompany brachytic variations and are of very frequent occurrence in cluster varieties. Fasciation, or duplication of parts, appears in many different stages, ranging from simple forking of intemodes or pedicels to inclusion of two flowers in the same involucre, or two bolls in the same calyx, or to an abnormal increase of the number of carpels.' Sometimes the fasciated intemodes remain united for their whole length and bear two leaves at the end, like the intemodes of opposite- leaved plants. In rare cases double intemodes bear two flowers or bolls, though usually the floral buds of abdominal intemodes are aborted. (See PI. LIX.) Union between the pedicel of a boll and the next intemode of the fruit- ing branch is another abnormality to which reference has already been made. Adhesion occurs in connection with brachysm, though it also appears occasionally in connection with short joints in noncluster varieties. It is more likely to be noticed in such cases because of the contrast with the normal joints of the branch. Doubtless as a result of the fact that the pedicel of the floral bud develops normally in advance of the next joint of the branch, these adherent intemodes are bent upward in the direction of the pedicels to which they are joined and form a distinct angle or elbow with the preceding intemode of the branch. (See PI. LX.) It is evident in such cases that the abnormality involves something more than a mere adhesion of the epidermal tissues, for the affected intemodes are much shorter than the others, and even shorter than the pedicel to which they are united. In most cases the short joints lose their floral buds or young bolls by abortion. ANALOGY BETWEEN BRACHYTIC VARIATIONS .\ND HYBRIDS The fact that the shortening of the intemodes is so often accompanied by abnormal leaves and involucres suggested the view here advanced, that brachysm represents a failure to maintain the normal specializa- tions of the parts. Considered as examples of intermediate expression of characters, the shortened intemodes and abnormal involucres of the cluster cottons afford a suggestive analogy with the abnormal interme- diate forms of branches and involucres that often appear in hybrids between diverse species of cotton. This analogy may be supported by the fact that sterility, or blasting of the buds or the young bolls, is very frequent in brachytic varieties of cotton, as well as in hybrids, and is especially common in involucres that liavc the abnormal intermediate forms of bracts. The idea that short-jointed variations differ from the parent stocks in only this one character, as often assumed by writers on Mendclism, is 1 Reasons for looking upon fasciation as a s>'raptom of defeneration have been given by Thomas Mcehau. (Science, v. 3. no. 70, p. 694. 18S4.) Meehan concluded that "a fasciated branch is an imperfect and pre- cocious attempt to enter on the flowerine or reproductive staRC." 396 Journal of Agricultural Research voi. m. No. 5 evidently not true of brachytic variations of cotton. While it may be that the mutations of other plants do not show changes in so many char- acters in connection with brachysm, the general fact seems to be that mutative changes affect several characters at once, instead of one charac- ter alone. It is possible, of course, to disregard the other differences and thus simplify the discussion of the Mendelian phenomena by giving exclusive attention to single differences, but many of the statements that have been made give a very misleading idea of the nature of mutative variations.' In normal cotton plants there are sharp contrasts in length between the intemodes that form the joints of the fruiting branches and those that form the pedicels of the flowers, just as there are definite differences in form and structure between the foliage leaves and the specialized leaves that constitute the involucre. In brachytic varieties these contrasts become less marked. The changes of characters are all in the direction of reduced specialization of intemodes. Instead of the normally con- trasted expression of the characters of the different kinds of intemodes represented in the normal branches and floral parts, there is a reduced contrast or intermediate expression of the characters, resulting in the formation of abnormal intemodes. Variations of this kind, resulting from intermediate expressions of characters that are normallv distinct or separate in expression, may be described as metaphanic variations. These would form a general class of abnormalities to include brachysm and other similar aberrations of heredity. BKACHYSM AND HOMOEOSIS The nearest approach to a recognition of metaphanic variations as representing intermediate expression of characters, as a general factor in heredity, is probably to be found in the theory of translocation of char- acters, or homoeosis, brought forward a few years ago by Dr. R. G. Leavitt. The theory of homoeosis is that "a character or a system of organization which has been evolved in one part of the body is transferred, ready-made, tc another part." - The theory of homoeosis might be applied to the phenomena of brachysm in cotton by considering that a partial translocation or homoeo- sis had taken place, for the characters of the intemodes, leaves, and involucral bracts are intermediate. The leaves become more bractlikc, the bracts more leaf-like. But to represent typical cases of homoeosis the leaves would need to be replaced by bracts or bracts by leaves, the ' " The difference between a tall and s dwarf |)ea is not the same as the difference between a tall and a dwarf man. In a human dwarf everyUiins is on a smaller scale than in the normal man. But a dwarf pea is not simply a miniature edition, as it were, of a tall one — it differs from a tall pea in one single clur- acteristic, the length of the intemodes, i. e., the sections of the stem between two nodes, or joints, where the leaves are given off." Darbishirc, A. D. Breeding and the Mendelian Discover>'. p. 13-ij. London, New York, 1911. ' Leavitt, R. G. A vecrtativc mutant, and the principle of homeosis in plants. In Bot. Gar., v. 4;, no. r, p. 64. 1909. Feb. 15. I9IS Brachysm 397 primary idea being that of transference of characters from one place to another, which is not the same as abnormal or intermediate expression of characters. Leavitt's idea of homoeosis may appear to be in better accord with the Mendelian theory of heredity, which assumes that varia- tions arise from differences in the transmission of characters, but meta- phanic variations seem to represent differences in the expression of the characters rather than differences in transmission. The present interpretation of metaphanic variations is based on the recognition of a definite distinction between transmission and expression as two essentially different processes, though usually described together under the general term "heredity." Characters are often transmitted without being expressed in visible form, as Darwin pointed out. Gallon made a formal distinction between patent and latent characters — that is, between characters that are brought into expression and those that are transmitted without being expressed. Latency occurs frequently in connection with Mendelian characters and has received much attention in recent years, but the importance of the distinction between trans- mission and expression is often overlooked and theories of transmission are often applied to phenomena that belong to the field of expression. With this distinction in mind, the idea suggested by metaphanic variations like brachysm is not that the characters of one part have been trans- mitted in some irregular manner to another part of the plant, but that the normal sequence of changes or contrasts in the expression of the characters is no longer observed. It is reasonable to believe that all the characters are transmitted to all the parts, including those that usually do not serve the purposes of propagation. Each of the intemodes of the cotton plant is capable of reproducing all the other parts. From this point of view the idea of translocation or transfer of characters no longer seems adequate to account for abnormal intermediate characters like brachysm. It seems more reasonable to think of metaphanic variations as arising because the characters are being confused or combined in expression. In connection with the theory of homoeosis Leavitt suggested that the translocation of characters from one part of the body of a plant or animal to another part opened the way to evolutionary changes, and many examples were given in support of this interpretation. Intermediate expression of the characters, involving a loss of specialization or differen- tiation of parts, may also be considered as one of the ways in which plants may vary and thus initiate evolutionary changes, but reasons have been given already for looking upon metaphanic variations as negative or degenerative changes of characters rather than as having progressive evolutionary value. Instead of being taken as new characters that are being added to the mechanism of transmission, metaphanic variations may mean that the mechanism of expression has become deranged so ^that the old characters are not normally developed. 398 Journal oj Agricultural Research voi. iii. No. s The prevalence of abortion in cluster cottons is one of the most striking evidences of the degenerative nature of such variations. In strongly clustered varieties completely sterile individuals are often found, the result of abortion of all the buds or young bolls or of all the buds which produce fruiting branches. Yet these completely sterile plants are less harmful to the stock than the other less abnormal degenerates that are able to reproduce themselves from seed and also to contaminate their neighbors with inferior pollen. That cluster cottons should seem more unstable than other varieties and more inclined to the production of abnormalities is easier to understand when they are considered as meta- phanic variations, representing intermediate expressions of characters that are normally distinct. AGRICULTURAL DEFECTS OF "CLUSTER" COTTONS Having considered the general nature of the cluster character of cotton and the accompanying variations, we are in better position to understand the physiological status and practical value of such forms. The first impression of cluster varieties is that they are more fruitful, for they are able to set their buds and bolls more rapidly than varieties with normal fruiting branches. But when we have learned that brachysm is in the nature of a malformation and is frequently accompanied by malforma- tions of leaves and bracts and abortion of both the floral and the vegeta- tive buds, it becomes apparent that the brachytic variants are to be avoided by the breeder, especially when they bear other marks of degeneration. The limited size and more upright habits of growth make it possible for more of the short-branched plants to stand in the same area, and very large yields may be obtained with favorable conditions of soil and season. But if the conditions prove unfavorable, cluster varieties are likely to suffer worse than others and to produce smaller crops, so that it is very doubtful whether the cluster habit is a practical advantage. The extreme forms of clustering are certainly undesirable, for in such varieties many of the plants are likely to become sterile through the blasting of all of the buds, the tendency to abortion being greatest when vmfavorable conditions are encountered during the crop season. Another agricultural consideration is that the crowding of the bolls together makes picking more difficult. This is especially true, of course, when clustering is accompanied by fasciation and other abnormalities, so that the bolls are malformed or misshapen. The opening of the bolls is also likely to be irregular, because cluster cottons often produce many late bolls on short branches developed from the axils of the leaves of the fruiting branches. Such bolls are usually undersized, as well as late in opening. A further objection to the cluster character, especially in long-staple cottons, is that the lint of cluster varieties or of individual variations ii> Feb. 15. 11)15 Brachysm 399 the direction of clustering is generally inferior to that of adjacent non- cluster plants. This may be due to the fact already noted, the greater susceptibility of cluster plants to unfavorable conditions, owing to their restricted vegetative development. Occasional exceptions have been noticed, where cluster plants produced lint as good or better than that of nonchister neighbors, but there can he no doubt that the general tendency is in the direction of inferiority of lint. CONCLUSIONS Brachysm is a term proposed to designate the shortening of the vege- tative intemodes of plants. It is a hereditary abnormality, indicating degeneracy, that has appeared in independent mutative variations in many distinct families of plants, including many cultivated forms. Brachytic variations are of frequent occurrence in cotton, giving rise to the so-called "cluster" and "limbless" varieties, and afford unusually favorable opportunities for learning the nature and physiological sig- nificance of such variations. The shortening of the intemodes of the cotton plant is usually con- fined to the fruiting branches without affecting the main stalk or the vegetative branches. JBrachytic variations occur independently in different species and varieties of cotton and do not constitute a natural group with a common origin. Brachytic varieties of cotton usually show other abnormaUties of the intemodes, leaves, and involucral bracts. There is also an increased tendency to abortion of the floral buds, and the blasted buds often remain attached to the plant, because of the absence of well-differentiated absciss-layer at the base of the pedicel. Though brachytic variations arise by mutative changes in the expression of the characters and show alternative Mendelian forms of inheritance, they afford no additional support to the general theories of mutation and Mendelism as explaining evolution. Such variations represent reduced specialization or intermediate expression of characters and are degen- erative in nature. They are not to be considered as examples of normal heredity or of the evolution of new characters. The abnormalities of brachytic variations are analogous to those found among hybrids and are likewise accompanied by tendencies to sterility or abortion of buds. Brachysm is to be associated with other fonns of intermediate expres- sion of characters, representing a general class of metaphanic variations. A more defmite recognition of this class of variations is desirable in con- nection witli the investigation of general problems of heredity and evolution. The agricultural value of brachytic varieties of cotton is impaired bv the tendency to abnormal variations and sterility and also by the fact that the cluster cottons arc more severely affected by unfavorable con- ditions. Hence, brachysm is to be avoided in the breeding of superior varieties of cotton. PLATE LI 1 1 Abnormal simple leaf on fruiting branch of Egyptian cotton, accompanied by abnormal leaf -like bract, remainder of involucre and floral bud removed. Natural size. Photographed by Mr. G. B. Gilbert. (400) Brachysm Plate LIII Journal of Agricultural Research ViJI. Ill, No. 5 Brachysm Plate LIV Journal of Agricultural Research Vol. Ill, No. 5 PLATH 1,1 V Nonnal globed le;if of Irtiiting branch of Egyptian cotton, accompanied by normal involucral bract for comparison with Plate LIII. Natural size. Photographed b) Mr. G. B. Gilbert. 75012°— 15 4 PLATE LV Abnormal leaf of fruiting branch of Egyptian cotton with one stipule enlarged and the lobe of the same side wanting. Natural size. Photographed by Mr. G. B. Gilbert. Specimen from Bard, Cal., on October 2, 1913. Brachysm Plate LV Journal of Agricultural Research Vol. Ill, No. 5 Brachysrn Plate LVI Journal of Agricultural Reseaf. h v.. I, III, N... 5 I'LATK LVI Brachytic fruiting bnmches of "cluster" cotton (Willcts Red Leaf) shortened to a single intemode b)- abortion of terminal bud. Natural size. Photographed by Mr. C. B. Doyle. Fig. I. — The boll at the right is borne by a ver>- short branch from an axillar\' bud. Fig. 2. — The boll at the right is borne by the shortened fruiting branch. The left- hand boll represents a shortened branch in the axil of the leaf that subtends tlie fruit- ing branch. PLATE LVII Normal and braLh)lic joints on same fruiting branch of Upland cotton. Natural size. Photographed by Mr. C. B. Doyle vSptcimen from Bard, Cal., on October 13, 1912. Brachys'" Plate LVII Journal of Agricultural Research Vol. Ill, No. 5 Brachysm Plate LVIII Journal of Agricultural Research Vol. Ill, No. 5 PLATE I.MII Brunches of abnormal variation of Upland cotton, witli abortive buds remaining attached to branches by deciirrent pedicels and elongated bud scars. The left-hand branch shows abnormal inequality in the lengths of the intemodes. PLATE LIX Portion of brach>-tic fruiting branch of Simpkins cotton producing twin fasciated branches from an axillary bud. The boll of the next inteniode has an elongated, somewhat decurrcnt base, while that of tlie third iutcmode of the branch is aborted and shri\-eled, only the pedicel being shown, at the left of the plate. Natural size. Photographed by Mr. C. B. Doyle. Brachysm Plate LIX Journal of Agricultural Research III, No. 5 Brachysm Plate LX Journal of Agricultural Research Vol. Ill, No. 5 PLATE LX Portion of fruiting branch of Columbia cotton, with one intemode adnate to the pedicel of the boll of the preceding intcrnode. Specimen from Easley, S. C. Natural size. Photographed by Mr. C. B. Doyle. PLATE LXI Fig. I.- — Plant of Dale Egyptian cotton, showing comjilete abortion of fruiting Ijranches on the main stalk, while the vegetative branches of the same plant produced a few fruiting branches and ripened a few bolls. Plant grown at Bard, Cal. Photo- graphed by Mr. C. B. Doyle. Fig. 2. — End of main stalk of plant sho\\"n in figure i, showing abortion of terminal l5ud and compensatory thickening of the petioles. Natural size. Photographed by Mr. C. B. Dovlc. Brachysm Plate LXI Journal of Agricultural Research Vol. Ill, No. 5 Brachys Plate LXll ^^ M ^ ^ f ^H^^^^^^^^^^^^^ v^ A Jl Y fW^ ^ V Journal of Agricultural Research Vol. Ill, No. 5 PLATE LXII Ends of main stalks of two plants of Dale Eg>'ptran cotton, showing simple fruiting branches and closely similar axillary fruiting branches. In one case the terminal bud was aborted, while in the other it was transformed into a boll subtended by an abnor- mal bract-like leaf with the stipules enlarged and the blade entire, instead of 5-lobed. Plants grown at Bard, Cal. Photographed by Mr. C. B. Doyle. ABILITY OF COLON BACILLI TO SURVIVE PASTEURIZATION By S. Henry Ayers, Bacteriologist, and W. T. Johnson, Jr., Scientific Assistant, Dairy Division, Bureau of Animal Industry INTRODUCTION The presence of colon bacilli in pasteurized milk is generally inter- preted as meaning either that the milk was not properly heated or that it was reinfected after pasteurization by careless handling. This inter- pretation is based on the low thermal death point of cultures of Bacillus coli. Van Geuns (4)' in 1899 found that the Bacillus neapolitanus of Emme- rich, the same organism as the B. coli communis of Escherich, was de- stroyed by heating for five minutes at 59° C. (138.2° F.) and one minute at 62.5° C. (144.5° F.). Based on the work of Van Geuns, Ringeling (6) examined 75 samples of pasteurized milk from 24 dairies in Amsterdam for the presence of colon bacilli. In 16 per cent of the samples exam- ined he found B. coli present. Since colon bacilli were found in pas- teurized milk from 10 of the 24 dairies, Ringeling concluded that this proportion (41 per cent) of the dairies in Amsterdam did not pasteurize or handle the milk properly. During recent years numerous investigators have studied cultures of B. coli and found that the organisms were easily destroyed at tempera- tures below 60° C. (140° F.), which is the lowest pasteurizing tempera- ture. In a previous study by the writers (i ) of the bacteria which survive pas- teurization 1 9 samples of raw milk in sterile flasks were heated for 30 minutes at 62.8° C. (145° F.). Each sample after pasteurization was examined carefully for the presence of colon bacilli, but none were found. All these results naturally strengthened the opinion that the presence of B. coli in pasteurized milk might be a valuable index as to the efficiency of the process and the care observed in handling the milk after the heat- ing process. At times, however, it has been found that high temperatures were re- quired to destroy cultures of B. coli. Gage and Stoughton (3) in a study of the resistance of B. coli to heat found that sometimes cultures were destroyed only by heating to temperatures much higher than the usual thermal death point. Heat-resistant strains have been also found by De Jong and De Graaff (5). Certain strains with which they worked re- ' Reference to "Literature cited" is made by number, p. 409. Journal of Agricultural Research, Vol. Ill, No. 5 Dcpt. of Agriculture. Wnshjngton. D. C. Feb. is. 1915 (401) A-13 402 Journal of Agricultural Research voi.iu.no. s quired 30 minutes heating at 7o°-72° C. (158° to 161.6° F.), in order to destroy them. Zelenski (9) also found strains of B. co/i which resisted high temperatures. In view of this uncertainty regarding the thermal death point of B. colt and its relation to pasteurizing temperatures, the following experi- mental work concerning the ability of colon bacilli to survive pasteuriza- tion was undertaken. The number of cultures used in the experiments was 174. METHOD OF DETERMINING THE THERMAL DEATH POINT The colon cultures were grown first in plain, neutral extract broth for 18 hours and then inoculated by means of a small-bore pipette into litmus- milk tubes. Four drops constituted an inoculation in each milk tube. In making the inoculations care was taken not to have any of the culture touch or any of the inoculated milk wash upon the sides of the tube, either during the handling or during the subsequent heating. The inoculated milk tubes, with the exception of the control tubes, were heated in a water bath in which the water was agitated, and the tem- perature of the milk was recorded in a control tube by a thermometer placed in the milk. The temperature in the tubes was not allowed to vary more than half a degree in either direction. In all experiments the heating period was 30 minutes at a given temperature. After the heat- ing, the tubes of milk were quickly cooled to about 10° C. (50° F.), incu- bated at 37° C. (98.6° F.), and the reactions recorded after 24, 48, 72, and 96 hours. Growth in the tube indicated that the organism was not destroyed at the particular temperature to which the milk had been sub- jected. In every case the tubes were run in duplicate, and in general both tubes had to show growth before the test was considered positive. The only exceptions to this were cases in which only one of the tubes showed growth after the highest heating temperature; in such cases one tube was considered a positive reaction and the organism was recorded as surviving the process. This method of determining the thermal death point was used, in order to render the conditions of heating similar to pasteurization. THE THERMAL DEATH POINT OF THE CULTURES AS A WHOLE Studies were made of the thermal death point of 174 cultures of colon bacilli isolated from the following sources: 154 from cow feces, 16 from milk and cream, 2 from flies, 1 from human feces, and i from cheese. (The cultures were supplied through the courtesy of Mr. L. A. Rogers, in charge of the research laboratory of the Dairy Division.) All of the organisms would be classified as colon bacilli according to the usual cultural tests for Bacillus coli. These cultures, with the exception of two not studied and' three noted below, were typical colon bacilli of the Feb. IS, X915 Colon Bacilli and Pasteurization 403 B. coli coimmmis or communior type according to the studies of Rogers, Clark, and Davis (7), also Rogers, Clark, and Evans (8). The three cultures not typical were probably of the B. aerogcties type. The cultures were heated in milk as previously described to temperatures ranging from 51.7° C. (125° F.) to 68.3° C. (155° F.). The results given in Table I show the number and percentage of cultures which withstood the differ- ent temperatures. Table I. — Effect of heat on colon bacilli — all cultures Exposed for 30 minutes at^ Cultures surviving. 51.7" c. dtSh 57.2° c. (133° F.). 60° C. (.40° F.). 62.8° c. (14s'' F.). 63.6' c. 68.3° C. (.55* F.). Number 174 100. 00 173 99-43 158 90. 80 95 54.59 12 6.89 I 0-57 0 It is seen from the table that 95, or 54.59 per cent of all the cultures, survived at 60° C. (140° F.). This is particularly interesting, since this temperature is the lowest used in commercial pasteurization. When heated to 62.8° C. (145° F.), 12, or 6.89 per cent, of the cultures survived. This temperature of 62.8° C. (145° F.) maintained for 30 minutes is the temperature generally used in the process of pasteurization. Only one culture sur\'ived a temperature of 65.6° C. (150° F.), and this culture when heated again failed to sun.'ive at this temperature. These results are shown more clearly in figure i, where they have been plotted. One of the cultures was destroyed at a temperature as low as 54.5° C. (130° F.). It is interesting to note that at 60° C. (140° F.) 95 of the cultures survived, while at 62.8° C. (145° F.) a difference of only 2.8° C. or 5° F., only 12 sur\'ived. In other words, 87.3 per cent of the cultures which survived at 60^ C. (140° F.) were destroyed at 62.8° C. (145° F-)- It is very evident from these results that colon bacilli may survive the process of pasteurization when a temperature of 62.8° C. (145° F.) is used. VARIATION IN THE THERMAL DEATH POINT OF THE CULTURES In order to determine whether the colon bacilli which survive at 62.8° C. (145° F.) would exhibit the same alnlity in repeated heatings, the same cultures were reheated to that temperature six times, with the results shown in Table II. 404 Journal of Agricultural Research Vol. m, No. s /oo \ ^ so I 5s €0 % 1 o "^^^^ y \ \ \ • \ ^--.^ /«5-v=: /30' /ss' Z'^o" /v^" -/so" /ss° s/.y'c. s-^s' s^s" €0" e^.e' ess" ee.3* Fig. I . — Curve showing results of heating cultures of colon bacilli for 30 minutes at various temperatures. Table II. — Variation in the ability of colon bacilli to survive pasteurization for 30 minutes at 62.8° C. (145° F.) Number of heating. GV HO HP IF IJ IL IN IS lY MF MT NV Total cultures . . . Positive First. 'Second. + + ! +- + - + ■ + • + - + - + - + - + - + - + - + - Third. Fourth. + - + - + - + - + + + - + + + + + + + - + - + + - I +- Fifth. • 12 I 12 1 12 4 + - t + + + - 1 -- + + + - + - + - + - + + + - + - + - Summao'. Both tubes Both tubes 12 6 12 9 12 O One tube + and one tube Feb. 15. 1915 Colon Bacilli and Pasteurization 405 In the above experiments two litmus-milk tubes were inoculated in the usual way with each of the 12 cultures and were heated for 30 minutes at 62.8° C. (145° F.). The first experiment shows the record of the 12 cultures which survived in the determination of the thermal death point of the cultures as a whole. When these 12 cultures were again heated, only 4 survived, on the third trial 8, on the fourth trial 6, on the fifth trial 9, and on the sixth trial none survived. These results are important since they show that at 62.8° C. (145° F.) colon bacilli may or may not survive a process of pasteurization. It is evident that this is a critical temperature and that occasionally colon bacilli may survive, owing in all probability to the resistance of a few cells in the culture. This explanation is supported by the figures in the summary of Table II, in which is seen the number of times that both the litmus-milk tubes showed growth; also when both tubes were nega- tive and when one tube was positive and one negative. It may be seen, also, that the same culture on repeated heating does not give the same results. It is evident that certain strains of colon bacilli have a thermal death point which is close to 62.8° C. (145° F.), and although they represent only a small percentage of the cultures we stud- ied, the fact that such cultures exist complicates the colon test for effi- ciency of pasteurization. The apparent scarcity of these resistant colon bacilli and the fact that 62.8° C. (145° F.) is near their thermal death point explains our failure to find them in the samples pasteurized by us under laboratory conditions, as stated previously in this paper. HEAT RESISTANCE OF COLON BACILLI From these results it seems that the colon bacilli as a rule have a low majority thermal death point and the cultures survive the higher tem- peratures only by reason of the resistance of a few cells. Gage and Stoughton (3) found this to be true of a few cultures which they studied, and although they tried to breed a race with a high majority thermal death point, their efforts were not successful. We have also tried to breed a resistant type, but thus far without success. In our experiments the thermal death point determinations were made in duplicate tubes of litmus milk, and the appearance of growth was recorded after an incubation period of 24, 48, 72, and 96 hours. In every case control tubes not heated showed a marked acid reaction in 24 hours, but in the heated tubes with the same inoculation the growth was often delayed, so that sometimes no reaction was noticed until after 96 hours, incubation. When the reaction was delayed, it showed that a portion of the bacteria in the milk were destroyed by the heating. When the heat- ing has little or no effect, the heated tubes should show a positive reaction 4o6 Journal of Agricultural Research Vol. Ul, No. 5 in 24 hours the same as the control tubes. In Table III we have recorded the percentage of cultures which gave a positive reaction in both of the litmus-milk tubes and in only one milk tube after different periods of incubation and when heated at different temperatures. Table III. — Effect of heat in relation to the time required for cultures to show grouth Hours of incubation. 34- 48. 72. 96. Total. Tube reaction. [2 tubes-)- . 1 1 tube-f . . [i tube—. . [2 tubes-|- . H tube-|-. . [i tube — . . [2 tubes-|- . [ I tube-(- . . [i tube— . . (2 tubes-l- . 1 1 tube-f. . 1 1 tube— . . S4.S°C. Per cent. 92-75 5.80 1-45 .«.2°C. (135° F.)- Per cent. 28.34 14. 17 24.41 12-59 5-52 11.03 o 3-94 60° c. (.40° F.). Per cent. II. II 26.98 17. 46 20.63 3- 17 19. 06 o 1-59 6;.8° C. (-45° F.). Per cent. 12.50 62. 50 25. 00 From the table it will be seen that 92.75 per cent of the cultures showed a positive reaction after 24 hours' incubation in both of the duplicate tubes when heated at 54.5° C. (130° F.). After 48 hours' incubation 5.8 per cent more of the cultures showed a positive reac- tion in both tubes, while 1.45 per cent showed a positive reaction in only one of the two tubes. At 62.8° C. (145° F.), however, only a small percentage of the cultures were positive after 24 hours. The majority required from 48 to 72 hours' incubation to show growth. This shows that a large proportion of the cells were destroyed so that a longer incu- bation was necessary to allow bacterial increases sufficient to cause a positive reaction. These facts are further supported by the differences in the number of cultures in which both duplicate tubes showed a positive reaction. At 62.8° C. (145° F.) only a small percentage of the cultures showed a positive reaction in both tubes, showing that in many cases all the bacteria in one tube were destroyed, while in the duplicate tube a few cells only survived. It is of interest to note that the colon bacilli are less heat-resistant than the streptococci, as is shown in a previous paper (2). Feb. IS, 1915 Colon Bacilli and Pasteurization 407 THE PRESENCE OF COLON BACILLI AS A TEST OF THE EFFICIENCY OF PASTEURIZATION The growth of colon bacilli which survive pasteurization is a matter of considerable importance, particularly when the presence of B. coli in pasteurized milk is considered as an index of the efficiency of the process. We have therefore studied the effect of pasteurization at 62.8° C. (145° F.) on two cultures of colon bacilli which were known to be able to sur\ave heating at that temperature. Flasks of sterile skim milk were inoculated with several cubic centimeters of an i8-hour-old broth culture of a colon bacillus. The number of bacteria in the milk was determined before heating and again after the pasteurized milk had been allowed to stand for 24, 48, and 72 hours at room temperature. The bacteria at the end of the 24-hour period were determined by placing as high as 3 cubic centimeters in large petri plates. Table IV shows the results of an experiment with the two colon cultures, GV and HO. Three flasks of milk were inoculated from each culture with the same amount of broth, but a bacterial count was made on only one of the three flasks before heating. Table IV. — Growth of Bacillus coli in milk liealedfor jo minutes at 62.8° C. {145° F.) and held at room temperature Flask. Number of bacteria per cubic centimeter. Culture No. Before pasteur- ization. After pasteurization. 24 hours. 48 hours. 72 hours. GV [ I 2 I 3 I 2 3 5, 000, 000 fo in 3 c. c. I in 3 c. c. [0 in 3 c. c. fo in 3 c. c. I, 050, 000 20, 000 210, 000, 000 61, 000,000 172, 000, 000 160, 000, 000 20, 000, 000 I, 140, 000, 000 1,450,000,000 1,750,000,000 I, 250, 000, 000 800, 000, 000 HO [i in 3 c. c. It may be seen from the table that, although the cultures survived the pasteurization at 62.8° C. (145° F.), there was a very great cell destruc- tion, as the bacterial count after the flasks had stood for 24 hours at room temperature was very low. However, in 48 hours' time there was a very large bacterial increase and even more after 72 hours. A similar experiment was repeated with the same cultures, except that the milk after heating was held in a refrigerator at 8 C. (46.4° F.). The results in Table V show again the great destruction of bacterial cells which takes place during the heating process. A few bacteria survived, but very little increase took place at the low temperature. 75012°— 15 5 4o8 Journal of Agricultural Research Vol. UI, No. s Table V. — Growth of Bacillus coli in milk heated for jo minutes at 62.8° C. {145° F.) and held at 8° C. (46.4° F.) Flask. Number of bacteria per cubic centimeter. Culture No. Before pas- teurization. After pasteurization. I day. 2 days. 3 days. 4tli-9thday. 3 weeks. 4 weeks. 6 weeks. GV HO I I 2 4,000,000 \ 7,000. oocj 0 in 3 c c 1 in 3 c. c I in 3 c. c 0 in 2 c. c 0 in 2 c c. 0 in 3 c. c. 0 in 3 c. c 0 in 3 c. c. 0 in a c c 0 in 3 c. c. 0 in 3 c. c. 0 in 3 c. c so 1,000 Soo 10 60 70 0 in 3 c. c. 0 in 3 c. c. 0 in 3 c c. The relation of these results to commercial pasteurization can be plainly seen. Milk is pasteurized usually at 62.8° C. (145° F.) for 30 minutes and in subsequent handling is kept at various temperatures from low to high, depending on conditions, until it is consumed. In view, therefore, of the results of our experiments, it is possible to explain the presence of colon bacilli in pasteurized milk on the ground of their ability to survive the process. These results, however, indicate that colon bacilli survive pasteuriza- tion on account of the resistance of a few cells and not because the cultures have a high majority thermal death point, in which case a large number of cells would survive. Since it is apparent that colon bacilli have a low majority thermal death point, we should not expect to find large numbers of these bacteria Ln pasteurized milk immediately after the heating process. If this condition is found we should believe from our results that the presence of the bacilli would indicate inefficient heating or a heavy reinfection. We must call attention to the fact that these opinions are based on a study of 174 cultures of colon bacilli, and consequently, while they represent a considerable number of strains of Bacillus coli, it is possible that a study of still more cultures might yield different results. It is not improbable that colon bacilli with a high majority thermal death point do exist, and if such is the case large numbers might be found immediately after pasteurization. SUMMARY AND CONCLUSIONS (i) The thermal death point of 174 cultures of colon bacilli isolated from cow feces, milk and cream, human feces, flies, and cheese showed considerable variation when the cultures were heated in milk for 30 minutes under conditions similar to pasteurization. At 60° C. (140° F.), the lowest pasteurizing temperature, 95 cultures, or 54.59 per cent, survived ; at 62.8° C. (145° F.), the usual temperature for pasteurizing, 12, or 6.89 per cent, survived. One culture was not destroyed at 65.6° C. (150° F.) on the first heating, but in repeated experi- ments it was always destroyed. Feb. 15, 191S Colon Bacilli and Pasteurization 409 (2) There is a marked difference in the effect of heating at 60° C. (140° F.) and at 62.8° C. (145° F.). Although there is only a difference of 2.8° C.,or 5° F., 87.3 per cent of the cultures which survived at 60° C. (140° F.) were destroyed at 62.8° C. (145° F.). (3) Considerable variation was found in the thermal death point of the colon bacilli which survived 62.8° C. (145° F.). When the 12 cul- tures which survived were heated again at the same temperature, it was found that many did not survive and in each repeated heating different results were obtained. It seems evident that 62.8° C. (145° F.) maintained for 30 minutes is a critical temperature for colon bacilli. (4) Among the 174 cultures studied all were found to have a low majority thermal death point, but were able to sur\'ive pasteurizing tem- peratures on account of the survival of a few cells. (5) The colon test as an index of the efficiency of the process of pas- teurization is complicated by the ability of certain strains to survive a temperature of 62.8° C. (145° F.) for 30 minutes and to develop rapidly when the pasteurized milk is held under temperature conditions which might be met during storage and delivery. The presence of a large number of colon bacilli immediately after the heating process may indicate improper treatment of the milk. (6) If milk is pasteurized at a temperature of 65.6° C. (150° F.) or above for 30 minutes, we should not expect, from our results, that any colon bacilli would survive. Consequently under such conditions the colon test for the efficiency of pasteurization may be of value. It must be remembered, however, that a study of more cultures may reveal strains of colon bacilli that are able to sur\'ive this and even higher tem- peratures. LITERATURE CITED (i) Avers, S. H., and Johnson, W. T., Jr. 1913. A study of the bacteria which survive pasteurization. U. S. Dept. Agr. Bur. Aniin. Indus. Bui. i6i, 66 p., 3° fig. (2) 1914. Ability of streptococci to survive pastetuization. In Jour. Agr. Research, V. 2, no. 4, p. 321-330, 3 fig. (3) Gage, S. de M., and Stoughton, Grace van E. 1906. A study of the laws governing the resistance of Bacillus coli to heat, fn Technol. Qiuu't., v. 19, no. i, p. 41-54. (4) Geuns, Jb. van. i88g. Uebcr das " Pasteurisiren " von Bacterien. Ein Beitrag ztir Biologic der Mikroorganismcn. In Arch. Hyg., Bd. 9, p. 369-402. (5) Jong, D. A. de, and Graaff, W. C. de. 1906. Die Coli-KontroUc der pasteurisierten Milch. Jn Ncdcrland. Weekbl. Zuivclber. Veeteelt, deel 12, afd. 37-38. Abstract in Milch w. Zcntralbl., Jahrg. 3, Heft 6, p. 265-268. 1907. Original not seen. 4IO Journal of Agricultural Research voi. iu.no. 5 (6) RlNGELING, H. G. 1903. Beitrag zur bakteriologischen KontroUe sog. krankheitskeimfreier pas- teurisierter und hochpasteurisierter Milch. In Milch Ztg., Jahrg. 32, no. 52, p. 818-819. Tr. from the Dutch. (7) Rogers, L. A., Clark, W. M., and D.wis, B. J. 1914. The colon group of bacteria. In Jour. Infect. Diseases, v. 14, no. 3, p. 4II-475- (8) and Evans, Auce C. 1914. The characteristics of bacteria of the colon type found in bovine feces. I71 Jout. Infect. Diseases, v. 15, no. i, p. 99-123 (9) Zelenski, Thadd-Eus. 1906. Zur Frageder Pasteurisation der Sauglingsmilch. /« Jahrb. Kinderheilk., Bd. 63, p. 288-307. Abstract in Centralbl. Bakt. [etc.], Abt. 2, Bd. 18, no. 4/6, p. 175. 1907. Original not seen. PRELIMINARY AND MINOR PAPERS FITTING LOGARITHMIC CURVES BY THE METHOD OF MOMENTS ' By John Rtce Miner, Computer, Maine Agricultural Experiment Station WITH AN INTRODUCTORY STATEMENT ON THE USE OF LOGARITHMIC CURVES IN BIO- LOGICAL AND AGRICULTIRAL INVESTIGATIONS BV RAYMOND PEARL, BIOLOGIST, MAINE AGRICULTURAL EXPERIMENT STATION INTRODUCTORY STATEMENT The use of logarithmic curves in the analysis of various kinds of biological and agricultural data is rapidly becoming widespread and general. It was first shown by Lewenz and Pearson (13, 22) ^ that the growth of children followed a logarithmic curve. The present writer (17) demonstrated that the phenomena of growth and differentiation in Ceratophyllum also followed a logarithmic curve. Donaldson (2, 3, 4, 5, 6) and Hatai (8, 9, 10) in a series of papers dealing with the growth and quantitative relations of the whole organism and its various parts in the white rat and the frog have shown that the same law holds for growth in those forms. Other biological phenomena than growth follow a logarithmic law. Pearl (14), in a case of regulation of the shape of abnormal eggs, and later Curtis (i) for normal eggs, have shown that the changes in size and shape of successively laid eggs are graduated with a logarithmic curve. Work now in progress in the Biological Laboratory, Maine Experiment Station, of which only a preliminary notice has yet been published (15), shows that generally the change in milk flow with age in dairy cattle is logarithmic. !3everal years ago Holtsmark (12) pointed out that the relation between the number of food units required and the milk yields of different animals was logarithmic. From this incomplete review of the literature recording the use of logarithmic curves in biological and agricultural investigations it is clear that the workers in these fields will, as time goes on, have increasing need to be able to handle these curves easily and critically. Up to the present time the only available method of fitting logarithmic curves was that of least scjuares. Several years ago Pearl and Mc- Pheters (16) published a set of tables intended to lighten materially the labor of fitting such curves by the least-squares method. For a long time, however, the writer has felt that it would be highly desirable to bring this class of curves into the general system of curve fitting worked out by Pearson {18, 19, 20, 21, 23), and known as the "method of moments." The theory of the method is extremely simple, involving as ^ Papers from the Biological laboratory of the Maine Agricultural Experiinent Station. No. 78. > Reference is made by number to " Literature cited," p. 4aa. Journal of Aericultural Research, Vol. Ill, No. s Dcpt. of Agriculture. Washincton. I). C Feb. ij, rgis Maine — 2 (411) 412 Journal of Agricultural Research vu. iii.no. 5 it does only the assumption that if we equate the area and moments of a theoretical cur\-e to the area and moments of a series of observations we shall get a reasonable fit of the curve to the obser\^ations. Experience with the method in the hands of different workers in England and Amer- ica has abundantly demonstrated that this assumption is entirely justi- fied in the fact. In the papers cited, and in others also, Pearson has given the equations for the calculation of the constants from the moments in the case of (a) skew frequency cur\-es in general, (b) sine curves, (c) parabolas of all orders, (d) the point binomial, (c) hypergeometrical series, etc. There has been lacking, however, the determination of the equations connect- ing moments and constants for the general family of logarithmic curves of the type }' = a + bx + ex- + d\og(x+ a) and its modifications. I suggested some time ago to Mr. Miner that he attack the problem, which, while theoretically simple and straightfor- ward, proved rather laborious in the actual carrying out. This he has done, with the results set forth in this paper. Raymond Peari, MOMENTS OF A LOGARITHMIC CURVE GENERAL CASE Let y,, v,, J3, . . . , y^, . . . , yjn he a. series of ordinates with cor- responding abscissae x,, x^, x^, . . . , Xf;, . . . x,„ to which is to be fitted the curve y = a + bx + c logmX. Let the unit of calculation = x^ — a;, and the origin be placed at 2Xi — %2. The abscissae expressed in units of calculation and taken from the new origin will then be i, 2, x\, . . . , x'^., . . . .v'„. In calcu- lating the moments each ordinate must first be multiplied by the base, li,r.^, of the rectangle of which it is the mid line. Otherwise the mo- ments will represent not the whole area, but strips of unit base of which the ordinates are the mid lines. For the first three ordinates Ik is I, I, 2X3—5, respectively; for higher ordinates it is found from the equation fe,,^= 2%',; -4j;\._,-f4A:',,_- ... - (- i)* 43;',-^ (- i)t 5. The upper limit of the area is given by 2x'^-2x'm-j+ . . . -(-i)* 2x'3 + (_i)* -~ = i-{-(j^ where / is the integral and q the fractional portion of the number. Let M„ represent the Hth moment about the origin as above chosen. Then Mn = j^^\a + bx + e log.„x) :,Mx=-^[(/ + ?)''«-0y^'] + c !og,„p Feb. IS, I9IS Fitting Logarithmic Curves 413 Putting w = o, I, 2, successively, we obtain the three equations which being solved give us the constants of the curve: (iv) r I I I 1 ^") + clog,„e|^(/+?) log, (/ + {{l+ i)M,-AU)- (P+3;+ i.5)M„] (xviii) 6=^[2A/,-(/+i)A/„-yjC (xix) I ,, '+ I , a=jMo-^-b-]3C (XX) where /, = 2[(/+i)(^/ + 0jlog.„i-log,„(/ + 0j + -^/= + 6/+3)log,„eJ' y2=|[(' + 2)[log.„^-log,„(/+i)}+/(/+ I) log,„e] 73 = 7J_('^ + ^ )logio(^ + 0 - ; log,o ; - nogi„e J For the curve y=a + bx + cx' + d log,oX.- d=u[2oM,~ 30(1+ i)M, + j,M,-j^M„] (xxi) c= y [6{A/,- (/+ i)M,}+ iP + 3l+i.5)M„] + ]^ (xxii) b=^[2M,-i!+ i)M„]- (/+ i)c-j^ (xxiii) a = jMo--j-b-j^.-j,d (xxiv) where ,>I2(p + f/ + 0 ;X/+i)(/= + 5/ + 0 h=j\}(l+ i)(/+;)|log.oj -log,„(/+ j)[ + g'(/^ + 6/+3)log,„ f] Feb. 15, 191S Fitting Logarithmic Curves 417 From the foregoing it is evident thai since the ;'s involve only /, and with equal intervals for the ordinates / will always be some integer, the values of the j's for a series of values of / can be tabled once for all, and in this way a great deal of labor saved in the ordinary fitting of logarithmic curves. Accordingly tables of the ;'sand of P + sl+1.5, -15' and-^have been prepared and are given in an appendix at the end of the paper. For the curve y = a + bx + c log,o (*+«): 6[M,-(l+ i)M,]+ (P+3l+i.5)M, (xxv) a'(l+2a'){/+a'){log,,{l+a')~]og,„a'}—^{P+i2a'l+i2a'%g^,e b = ^{jM, - (/ + I ) M„ - c{«' (/ + a') (log.oa' - log.„/+^') -(j + ^'ipogjj + ( a= jM^--^b-c-j-[(l+o!')logi„{l+n')-a'\og^„a'-nog^„e] (xxvii) where 'i'' = n' + K- /" ILLUSTRATIONS OF THE USE OF LOGARITHMIC EQUATIONS In order to make clear the use of the above equations, some numerical illustrations will be given. Let us first consider the data contained in Table I. These give the mean milk production in pounds over a 7-day period of Holstein-Friesian cattle at different ages. The data are taken from the official 7-day A. R. O. record of the Holstein-Friesian Association (11). The laborious task of extracting and tabulating these records and calculating the means was carried through by Mr. John W. Gowen, with the assistance of Mr. S. W. Patterson and Miss Anna B. Perkins, all of the Maine Experiment Station. In future publications from this laboratory these figures will be further dealt with, but here they are used solely for purposes of illustrating the method. Table I. — Mean y-day milk production of Holstein-Friesian cows a ( differen ( ages Age. Number of cows. Mean pro- duction. Age. Number of cows. Mean pro- duction. I yr. 6 mo. to i yr. ii mo I.09S 3.693 »,330 3,041 1,950 1,627 1,565 1. 195 1,142 882 850 68s 597 Pounds. 290. 6 316.7 347.0 376.0 407.9 428. 2 447. I 457-2 464. 2 466.3 468.0 466. 6 466.6 434 433 249 343 149 137 72 67 37 3S 30 23 10 Pounds. 467.0 466.6 8 yr. 6 mo. to S yr. ii mo ayr. 6 mo. to a yr. itmo 464.6 462.5 460.7 460. 3 9 yr. 6 mo. to 9 yr. 11 mo 10 yr. to 10 yr. s mo loyr. 6 mo. to 10 yr. 11 mo.. . 3 yr. 6 mo. to 3 yr. 11 mo 4yr. 6 mo. to 4 yr. 11 mo 455-3 1 1 yr. 6 mo. to 1 1 y r. 1 1 mo . . . S yr. 6 mo. to 5 yr. 11 mo 449. 1 12 yr. 6 mo- to 12 yr. 11 mo 444.0 6yr. 6 mo. to 6 yr. 11 mo 443- J 448.7 13 yr. 6 mo. to 13 yr. 11 mo. . . 7 y r. 6 mo. to 7 yr. 11 mo The problem now is to fit by the method of moments a logarithmic curve of the form >■ = a -f bx -I- cx^ ■\- d log * to these milk production means. 4i8 Journal of Agricultural Research Vol. Ill, No. 5 The calculations to obtain the moments are given in Table II. Table II. — Calculation of moments for data on Holstein-Friesian cows in original form Age. 1 yr. 6 mo. to I yr. 1 1 mo . . 2 yr. to 2 yr. 5 mo 3 yr. 6 mo. tojyr. 11 mo. - 3 yr.t0 3yr. s mo 3 yr. 6 mo. to 3 yr. 11 mo. . 4 yr. to 4 yr. s mo 4yr. 6 mo. t0 4yr. ixmo. . syr. to 5 yr. 5 mo syr. 6rao. t0 5 yr. 11 mo. . 6 yr. to 6 yr. s mo 6yr. 6 mo. tobyr. 11 mo. 7 yr. to 7 yr. s mo 7 yr.6 mo. to 7 yr. ji mo 8 yr. to 8 yr. 5 mo 8 yr. 6 mo. to 8 yr. 11 mo . 9yT. togyr. 5 mo 9yr. 6mo. t09yr. 11 mo, , 10 yr. to 10 yr. s mo 10 yr. 6 mo. to 10 yr. 1 1 mo 11 yr. to II yr. s mo iiyr.6mo.toiiyr.iimo la yr. to 12 yr. s mo 1 a yr. 6 mo. to 1 2 yr. 11 mo 13 yr. to 13 yr. 5 mo 13 yr. 6 mo. to 13 yr. 11 mo 14 yr. to 14 yr. 5 mo 290. 6 3'6-7 347.0 376.0 407.9 428.2 447. I 457-2 464.2 466.3 468.0 466.6 466.6 467.0 466.6 464. 6 462.5 460. 7 460. 2 455-2 453-7 449.1 444.0 443-2 448.7 440.0 Total iii3is- 45261 326. 06732 240.32913 401. 76094 361.44305 407.9 428.2 447.1 457-2 464.2 466.3 468.0 466.6 466.6 467-0 466.6 464.6 462. s 460.7 460- 2 455-2 453-7 449.1 426.81041 513. 14250 340. 49788 493- 70138 326.06732 480. 65826 1,205. 28282 1,445- 77220 2.039. 5 2,569. 2 3.129-7 3,657-6 4.177-8 4,663.0 5,148.0 5,599- 2 6,065-8 6, 538- o 6| 999. o 7.433-6 7,862.5 8,292.6 8,743-8 9, 104- o 9. 527- 7 9, 880. 2 9,816.63943 12,315.42000 8,512. 44700 12,836.23588 326. 961. 3.61s. 5.783. 10.197. 15.415- 21,907. 29, 260. 37,600. 46, 630. 56,628. 67.190. 78.855- 91.532- 104.985. 118.937- 133.662. 149, 266. 166, 132. 182,080. 200,081. 217.364- 225,782. 295,570- 212,8ll- 333. 742- 06732 31652 84846 08880 5 70689 08000 17500 13288 23 SO 92 153- 234' 338. 466. 622 806, 1,025, 1,281, J. 574 1,903, 2,272 2,686, 3,156 3,641. 4,201 4,782 5,193 7,093 5. 320 8,677 326.06732 922.63304 847- 54538 132-35520 987.5 491-2 3SS-3 086.4 401. 8 300. o 908.0 284.8 120. 2 448.0 775-0 001- 6 262. 5 802.4 511.8 600. o 715-7 016.8 002. 25847 681. 92000 279- 37SOO 295. 45488 158.369. 72291 2.806,320.01587 55.610.556. 60939 For reasons which need not be considered here it is usually desirable to use corrected rather than raw values of the moments. Here we have used one of Elderton's (7) correction methods. The column headed y' is obtained from the y column by Elderton's formula V' — i. e., by mul- tiplying the first and last ordinates by 1. 1220486, the second and last but one by 0.7588542, the third and last but two by 1.1578125, and the fourth and last but three by 0.9612847. From this table we have at once M, = 1 1,315.45261 ; Mi= 158,369.72291 ; Mj = 2,806,320.01 587; .1/3=55,610,556.60929; /=26. Substituting these values in the equations xxi to xxiv for the curve y = a + bx + cx^ + d log x, the following values for the constants are found: d= 0.060096940 (20M3— 810M2 + 8907M,— 2i829.5A/o) = 259.833i7. 0 = 0.000025250 (6M2— i62yVf, + 755.5Mo) + o.oo24io2d= —0.0533. & = 0.000341 37 (2Mj— 27M„) — 27c — 0.044239^= —6.225. a=o.03846i5M„- 13.56- 238.583333c- 1.022111^=266.38. The final equation then becomes y= 266.38- 6.225.1;- 0.0533.1-= + 259.833 logioX. The ordinates calculated from this equation in comparison with the observations are given in Table IV. Before proceeding to any discussion of the fit, let us consider a second example, where the ordinates are at irregular intervals. The data here taken for illustration are the same as those of the preceding example, except that certain of the observations have been arbitrarily combined and the new values so obtained taken as ordinates. Table III shows 'It is to be noted that the coefficieats as given by Elderton (7, p. 27) are incorrect in the last figtire. Feb. IS, 1915 Fitting Logarithmic Curves 419 how this combination has been carried out, and also the calculation of the moments from ordinates at irregular intervals. Since in this instance the four ordinates at each end are regularly spaced, Elderton's formula V may be used. W-hen the end ordinates are not regularly spaced, this is not applicable, and the ordinates may without serious loss of accuracy be left unmodified. Table III. -Calculation of moment! for data on HoUtein-Friesian cows, with grouping of certain ordinates Age, y y "-. *-/' x' *-/''' *../V' h^^y'x" I yr. 6 mo. to 1 yr. 11 mo. . . 290.6 316.7 347.0 376.0 407.9 428. 2 452. 1 466.2 466.6 466. 1 462.5 460. 7 456.4 449. 1 444.0 443.2 448.7 440.0 326.07 240. 33 401. 76 361.44 407.9 428. 2 452. I 466.2 466.6 466. I 462. s 460.7 456.4 449. I 426. 81 513- 14 340. 50 493- 70 326.07 240.33 401. 76 361.44 407.9 428.2 904.3 1.398.5 933-2 1.398.2 462.5 460. 7 1.369. I 449-1 426- 81 SI3. 14 340. SO 493. 70 I 2 3 4 S 6 7-S 10 12.5 IS 1? 20 32 23 24 25 26 326. 07 480.66 1,205. 28 1.445-77 2,039-5 2,569-2 6,783-25 13.985-0 11,665.0 20,973-0 7,862-5 8,292-6 27,382.0 9, 880- 2 9, 816- 64 12,315-42 8,512-45 12,836- 23 326.07 961.32 3,615-85 5,783-09 10, 197- 5 15.415- 2 50. 866- 88 139. 850- 0 145,812- 5 314,595-0 133,662-5 149,266.8 547,640.0 217,364.4 225,782. 70 295, 570. 08 212,811. 17 333,742-13 326.07 1,922.63 10, 847- 55 23,133-35 50,987-5 92,491-1 381,501- 56 1,398,500.0 1,822,656.25 4.718,925-0 2,272,262.5 3,686,802.4 10.952.800.0 4,782,016.8 5,193,002. 26 7,093,681.93 5. 320, 379. 37 8,677.395.45 2 yr. 6 mo. to 2 yr. 11 mo. . . 3 yr. 6 mo. to 3 yr. 11 mo. . . 4yi. 6 mo. to 5 yr. 5 mo S yr. 6 mo. to 6 yr. 11 mo... 9 yr. 6 mo. to 9 yr. 11 mo. . . 10 yr. to 10 yr. 5 mo 10 yr. 6 mo. to 11 yr. 11 mo. 12 yr. to 12 yr. 5 mo 12 yr. 6 mo. to 12 yr. 11 mo. 13 yr. to 13 yr. 5 mo 13 yr. 6 mo. to 13 yr. n mo. 14 yr. io 14 yr. 5 mo Total II. 315 45 158.369-77 2.803,263.19 55. 479. 430. 81 From Table III we have the following values: Mo= 11,315.45; M,= 158,369-77; M2^2,8o3,263.i9; M3 = 55,479,430.81 ; / + 7=26.5; /=26; ?=o.5- Substituting these values in the equations xi to xiv, we obtain: d = 0.000096940 (20M3—810A/2 + 8907M,— 2 1,829. 5M„) = 245.67899. 0 = 0.000025250 (6A'/,— 162M, + 755.5M|,) + o.oo24io2(f= —0.1338. 6 = 0.00034137 (2M, -27M„)- 27c- o.044239(i= -3.425. a = 0.03846 1 5 A/„— I3.5&— 238.583333c— 1.0221 1 id= 262.26. The final equation then becomes — j/= 262. 26- 3.425X- 0.1338*- + 245.679 logio^;. Table IV compares the fit of the two curves calculated by the method of moments with that of the curve fitted by the method of least squares. It is apparent that the two methods give results of substantially the same accuracy. By the method of moments the root mean-square error is greater, and the mean error less than by that of least squares. Neither difference is, however, large. 420 Journal of Agricultural Research Vol. III. No. s T.4BLE IV. — Comparison of the graduation of data on H ohlein-F riesian cows by different methods 13. 14- IS- 16. 17- 18. 19- 23- 24- 25- 26. Total. By least squares. By moments. y from data. 290. 6 316.7 347-0 376.0 407.9 428.2 447-1 457-2 464. 2 466-3 468-0 466-6 466.6 467-0 466-6 464. 6 462.5 460.7 460. 2 455- 2 453-7 449- I 444- o 443- 2 448.7 440. o y from curve. 272. 1 332.2 366.9 391. 1 409. 2 423-4 434-7 443-8 451-2 457-7 461- 6 46s- I 467-5 469-0 469-6 469-4 468.5 466-9 464- S 461-5 457-9 453-7 448-9 443-5 437-5 431.0 •fl8.5 -■5-5 — 19.9 — 15- 1 — 1-3 + 4.8 -fj2.4 + 13-4 + 13.0 -f 8.6 + 6.4 + t-S — -9 — 2. o — 3-0 -4-8 -6.0 -6.2 — 4-3 — 6-3 — 4.2 -4-6 -4.9 — -3 4-II.2 -^ 9. o 198. 1 342- 25 240. 25 396. 01 228.01 1.69 23-04 IS3- 76 179- 56 169- 00 73-96 40-96 2. 25 .81 4.00 9.00 23.04 36.00 38.44 18.49 39.69 17.64 21. 16 24.01 •09 125.44 81.00 Ordinates at criual inten'als. Ordinates at unequal intervals. J- from curve. 2,289.5s 260- 1 331-9 371-2 397- I 415-5 429. 3 439-8 447-8 454.0 458.6 462. o 464. 4 465. 8 466.6 466.6 466. o 464.9 463. 2 461. 1 458.6 455-7 452-4 448.8 444.9 440. 7 436. 2 +30.5 -15.2 —24.2 — 21. I -7.6 — 1. 1 + 7-3 + 9-4 + 10. 2 + 7-7 -*- 6-0 -4- 2.2 + .8 + -4 o -1-4 — 2.4 -2-5 — -9 -3-4 — 2.0 — 3-3 -4-8 — 1-7 + 8.0 + 3-8 930- 2S 231-04 585. 64 445-21 57-76 I- 21 53-29 88.36 104. 04 59-29 36.00 4-84 .64 .16 O 1.96 5-76 6. 25 .81 11.56 4.00 10. 89 23.04 2.89 64.00 14.44 177-9 2f743-33 curve. 258.7 +31-9 328.8 — 12-1 368.0 — 21-0 394-3 -■8.3 413-5 - 5-6 428.1 + .1 439- 3 + 7-8 448-2 + 9-0 455- 0 + 9.2 460-3 + 6.0 464- 2 + 3-8 467-0 — -4 468-8 -2.2 469-7 - 2.7 469- 7 — 3-1 469-0 - 4-4 467- 7 - 5.2 465. 7 - 5-0 463.0 - 2.8 459.9 - 4-7 456.2 - 2.5 452.0 - 2.9 447-3 - 3-3 442- I + I.I 436- 4 + .2.3 430- 4 + 9.6 187.0 1,017. 61 146. 41 441-00 334-89 31-36 .01 60-84 81.00 84.64 36.00 14.44 .16 4.84 7.29 9. 61 19.36 27.04 25.00 7.84 22. 09 6- 25 8.41 10.89 I- 21 151.29 92. 16 2,641.64 Least squares: Root mean-square error=9. 4; mean error= 7.6. Moments, equal intervals: Root mean-square error= 10.3; mean error=6.8. Moments, unequal intervals; Root mean-square error= 10. 1 ; mean eiTor= 7.2. In Table V are given the values of the fs and certain other constants involving only / for values of / from 3 to 40. This range will include practically all cases likely to occur in ordinary statistical work. The author wishes to acknowledge his indebtedness to Dr. Pearl for his aid and suggestions throughout the work. Feb. 15. 191S Fitting Logarithmic Curves 421 I rfl »0 C < r*i m Q "i n c lun-O r-o-wO O "Of 00 O f^-D 0<*0 ^eO r« uip* ^r^^O lAOCO MOO 0-« O O fOfOr-M &00 « ^W" i O-O Kjuni-.for^t-* 0.00 >« ^ Ov lO w 10 r^*0 t-M M u^O r» OiOim roi^VM «t ^00 CO • rri u^ 0^ m t^ •-• CC r--f»ir--N f*iO O n O ^ r* ^ Q ec ec r* u^O r-« >« >/^0 00 O fO<0 O > f»lO00 00000 't'O f^ O M rO-OM"© NCOVOM o r-VO'tN 1- OlOO »~\0 in"tv»oo< c^t oooooooooooooooooooo0o5ooooooo5oooo( 6 • uiO 10 0 »n 0 .n 0 ^ 0 xn 0 in 0 in 0 ^0 m 0 to 0 >oO »n 0 in 0 m 0 m 0 Ul 0 IT 0 u, w m Ov fi CO V* r* " ? -■ r^q •^ t^ •^ 0 r, r-. c^^o N ■T t> M 06 m t~ r. •^ - »;- 9 y I- frj 0 ^ fO -^vOOO O C'' f^ tf) t^ in t^ t^ t*i t^ u _ _, , MOO *n ^« O t^-O «^ *■ •no.^OvfO'^O'O « Oi>nM Oir - t>>00 oOO'OO'-f*!^^-*' m^ r*oO CTi O O t^ - e* Oi *-• Pt M O mO O 00 O t" rri rrj r wr-QMOOno'nt-oOOO-Of H Mio r-osQ « irtfO"* t--co f- M oc -_ > m'-'O O r^^i^f^O-O "J'OOO'O '^ron ■ O O O O < '"tmO "Iff irir~i-*i-i TTOO ^ >o "1 Pi 1 O O 1 T8Sii 1001 300 W^fOQ ^•^*l« > - I- 00 o t "1 p ^,-. lOHOO^OWinu SoO^-i O O POOOf^O w t- O^-O 0^0 V f^ - "^'OP^'S'^M^wjiOWf-'^'OWOiQOOOf*'- I ocor^-om^'j-r^m'ow pt « pj m • ■ •"QQQQQQOQQOQOOOt S >« o " >) tn >o 10 u 1 (7" d> •:* m m «A m >H O, C> M 8 ■2 10O MvO Q p^N M ^fOt~oo O ffiO M i»)vr^» -p< r~oo»0 OiO ^^m<»)M r- p^^o m co 0> » -0^0 rr-.IXt O PC l-O M 'T'O r- O ^^ '-' O f^ *■ fM ^f^oo O ■- 0>u-)0 to r^t-r-iofoOO t I r- M moo pt >nr*Q »o«nt-osM wimr-oo ' t mo 00 i^r^t^ooooooeooo 9>0>0?tC r^mOiMOOOoO t^ Q M wioO m t~» \0 O mo i^w hOoo r-r--oo mw ■-1 « 0\0<^00vt>n pooo O »0 O M r* M o o^ i*i moo O w f^ "1 mo t* fi w) >nO r^ a O " **> t mo t--00 W 000000'*«-''-'-'"««"M O O 00 •* t^O too P< r^m^oOO Oi p*CO OiQ p» f^O tM p< O &50 ^oO*^^-'0■-' nmtttt"*tt'^''l''l''l'')'')f'i'Ot'lc>w 00000000000000000000000000000 O M m r- fOO r-OOMV-iOtiv h.o.p« O'f'iM too tma« mMOOO moo o> t •- to 0>N N f^too r->00O00 POi^O ^mO n O (OOOv -ICO wjoo on»Of" O J?* r-o inio TttfOPOP^f^ 422 Journal of Agricultural Research voi. iii. no. s LITERATURE CITED (i) Curtis, M. R. 1914. A biometrical study of egg production in the domestic fowl. IV. Factors influencing the size, shape, and physical constitution of eggs. In Arch. Entwicklungsmech. Organ., Bd. 39, Heft 2/3, p. 217-327, pi. 6-10. (2) Donaldson, H. H. 1908. A comparison of the albino rat with man in respect to the growth of the brain and of the spinal cord. In Jour. Compar. Neurol, and Psych., V. 18, no. 4, p. 345-389- (3)- (4) 1909. On the relation of the body length to the body weight and to the weight of the brain and of the spinal cord in the albino rat (Mus norv-egicus var. albus). In Jour. Compar. Neurol, and Psych., v. 19, no. 2, p. 155-167. 1910. On the percentage of water in the brain and in the spinal cord of the albino rat. In Jour. Compar. Neurol, and Psych., v. 20, no. 3, p. 119- 144. (5) — 191 1. On the regular seasonal changes in the relative weight of the central nervous system of the leopard frog. In Jour. Morph. , v. 22, p. 663-694. (6) and HaTai, Shinkishi. igii. A comparison of the Norway rat with the albino rat in respect to body length, brain weight, spinal cord weight, and the percentage of water in both the brain and the spinal cord. In Jour. Compar. Neurol, and Psych., v. 21, p. 417-438. (7) Elderton, W. p. 1907. Frequency-curves and correlation. 172 p. London. (8) Hatai, Shinkishi. 1909. Note on the formulas used for calculating the weight of the brain in the albino rats, /m Jour. Compar. Neurol, and Psych., v. 19, no. 2, 169-173. (9) 191 1 . A formula for determining the total length of the leopard frog (R. pipiens) for a given body weight. In Anat. Rec, v. 5, no. 6, p. 309-312. 1911. An interpretation of growth curves from a dynamical standpoint. In Anat. Rec v. 5, no. 8, p. 373-382. (11) Holstein-Friesian Association of America. 1913. The advanced registry year book. v. 24, 1004 p. (12) Holtsmark, G. 1904. [On the relation of milk yield and feed consumption.] In Arch. Math, og Naturvidensk, v. 26, no. 2. (Cited from Exp. Sta. Rec, v. 16, P- 59I-) (13) Lewenz, M. a., and Pearson, Karl. 1904. On the measurement of internal capacity from cranial circumference. In Biometrika, v. 3, pt. 4, p. 366-397, pi. 2-3. (14) Pearl, Raymond. 1909. Studies on the physiology of reproduction in the domestic fowl. I. Regu- lation in the morphogenetic activity of the oviduct. In Jour. Exp. Zool., V. 6, no. 3, p. 339-359, 2 fig. (15) 1914. On the law relating milk flow to age in dairy cattle. In Proc. Soc. Exp. Biol, and Med., v. 12, no. i, p. 18-19. (16) and McPheters, Lottie E. 1911. A note on certain biometrical computations. In Amer. Nat., v. 45, no. 540, p. 756-760. (17) • with the assistance of Pepper, Olive M., and H.\GLE, Florence J. 1907. Variation and differentiation in Ceratophyllum. 136 p. Washington, D. C. (Carnegie Inst. Washington, Pub. 58.) (18) Pearson, Karl. 1895. Mathematical contribution to the theory of evolution. II. Skew vari- ations in homogeneous material. In Phil. Trans. Roy. Soc. London, s. A., v. 186, p. 343-414, 10 pi. (19) 1899. On certain properties of the hypergeometrical series, and on the fitting of such series to observation polygons in the theory of chance. In Phil. Mag. and Jour. Sci., s. 5, v. 47, no. 285, p. 236-246. Feb. 15, I9'5 Fitting Logarithmic Curves 423 (^°^ ^".^.orMaflT/^'atica. contributions to the theory of evoludon^ X^ Supple- ment to a memoir on skew vanation. In Phil. Trans. Koy. soc. London, s. A., v. 197, p. 443-459- ^"^ "^^ On the systematic fitting of curves to observations and measurements. II. In Biometrika, v. 2, pt. i, p. 1-23. 1° "S- ^"^ ZZ On the laws of inheritance in man. II. On the inheritance of mental '^*- '''LS moral charact^r^in man, and its comparison with the '"hen^ce of the physical characters. In Biometrika. v. 3, pt. 2/3, p. 131-190 ^^^^ "1^7 On the influence of past experience on future expectation. In Phil. Mag. and Joiu-. Sci., s. 6, v. 13. no. 75, p. 305-378- 75012°— 15— 6 ORGANIC PHOSPHORIC ACID OF RICE By Alice R. Thompson', Assislttnt Chemist, Hawaii Agricultural Experiimnl Station That phosphoric acid occurs in organic combinations with inosite in the seeds of many plants has been shown by several investigators. Postemak (9),^ Patten and Hart (8), Hart and Andrews (6), Hart and Tottingham (7), Anderson (i, 2, 3, 5), Rather (10, 11), and others have isolated this organic substance from pumpkin seed, beans, wheat, com, oats, and cotton seed. Although they have been unable to obtain it by synthesis and still disagree as to the formula and composition of the acid and its salts, it is generally known as phytin or phytic acid. Suzuki et al. (12, 13) and Anderson believe phytin to be a hexaphosphoric acid ester of inosite, and Anderson has shown that the organic phos- phoric acid in wheat bran differs materially from the acid he has obtained from a number of seeds. Phytin is completely hydrolyzed into free phosphoric acid and inosite only with difficulty, as Anderson (4) showed by boiling phytin with con- centrated nitric acid for several hours. In the previous determination of phosphoric acid in foliage and grain of rice (Oryza saliva) at the Hawaii Experiment Station (14) several methods were used in oxidizing the organic matter. On boiling the grain with a mixture of nitric and hydrochloric acids (aqua regia) the author noticed that, although the solution soon became colorless, giving the appearance of complete oxidation of organic matter, if boiled to dryness a charred mass remained in the flask. Determinations of phos- phoric acid in the solution (not boiled to dryness) in the case of the rice grain showed about one-third of the total phosphoric acid as found by the Neumann method. The determination of phosphoric acid in the foliage, on the other hand, by either method was about the same. It was thought that the reason for this resistance to the action of aqua regia is probably the fact that phosphoric acid occurs in the rice grain as phytin and is therefore not completely hydrolyzed. It was decided, therefore, to give some study to the organic phosphoric acid of rice. Suzuki et al. (12, 13) obtained an impure salt of phytic acid from rice by extracting the rice bran with a 0.2 per cent hydrochloric-acid solu- tion and precipitating with alcohol. As Anderson (i, 2, 3, 4, 5) has shown that phytin so prepared would contain the inorganic phosphoric acid of the seed, as well as other impurities, the phosphorus content as shown by the resulting analysis is not that of pure phytin. It is therefore of interest to obtain the pure salt of phytic acid from rice. In following the methods of Anderson (1, 2, 3, 4, 5) of purifying the acid by repeated solution in hydrochloric acid and precipitation with barium hydroxid, the author hoped to isolate the pure tribarium salt, but the ease of partial hydrolysis of the substance and the difficulty of eliminating all impurities which may be present, such as other phos- * Reference made by number to " Literature cited.'' p. 430. Journal of Agricultural Researcll. Vol. Ifl. No. 5 Dept. of Agriculture. WashiuBton, D. C. Feb. 15. 1915 B 4 (42s) 426 Journal of Agricultural Research voi. iii, No. s phoric esters of inoske, stand in the way of obtaining the pure substance. Special attention was paid to methods for the determination of the barium and phosphoric acid in the salt, and observations of some interest were made. The total phosphorus was determined in samples of rice bran and unpoUshed and polished rice. The following determinations were dupH- cated to within 0.02 per cent: Phosphorus in rice bran, 2.291 per cent; in unpolished rice, 0.321 per cent; in polished rice, 0.140 per cent. It is apparent that the rice bran contains a comparatively high per- centage of phosphorus. Phytin was determined in the bran by extracting 100 gm. of the sample with a 0.2 percent hydrochloric-acid solution and precipitating with alcohol. The precipitate stood over night, when it was filtered, first washed with 50 per cent alcohol, then with ether, and was dried at 105° C. It weighed 8.22 gm., amounting, therefore, to 8.22 per cent of the bran. As phytin contains considerable organic phos- phorus, it is apparent that rice bran contains much of its phosphorus in the organic form. The writer was unable to obtain phytin from the polished rice. The 0.2 per cent hvdrochloric-acid extract from several kilograms of finely ground pohshed rice yielded but a slight precipitate with alcohol or barium chlorid. With barium hydroxid added to alkalinity a considerable pre- cipitate occurred which did not behave like phytin. The phytin obtained from the unpolished rice was doubtless contained in the outer layer, which is removed in polishing. Two i)reparations of barium phytate were made according to the method of Anderson; one from unpolished rice, the other from rice bran. Six kg. of finely ground unpolished rice were treated with a 0.2 per cent hydrochloric-acid solution for several hours. This was filtered through cheesecloth and filter paper the same day and the filtrate pre- cipitated with a barium-chlorid solution. The precipitate was collected on a filter, washed with 50 per cent alco- hol, dissolved in a i.o per cent hydrochloric-acid solution, which was then filtered and precipitated with barium hydroxid. It was reprecipi- tated three additional times with barium hydroxid, once with alcohol, again with barium hydroxid, and again three times with alcohol. The filtrate from the alcohol precipitate soon ceased to give an immediate precipitate with an ammonium-molybdic solution, showing the elimi- nation of inorganic phosphoric acid from the precipitate. The precipi- tates formed by barium hydroxid were well washed with water. Crystals were then obtained by adding barium hydroxid to the acid solution until a slight precipitate resulted and then filtering and allowing to stand a couple of days. The crystals thus formed were washed, recrys- tallized in the same way, and then crystallized from the acid solution by the addition of alcohol. It was found that on adding an equal volume of alcohol, according to Anderson, only an amorphous precipitate was obtained. It was only by adding just sufficient alcohol to make the solution turbid that crystallization took place. Under the microscope the precipitate appeared to be composed entirely of crystals in the form of globules of needles with occasional needle- shaped crystals arranged in stellar form. These were probably the same substance. The crystals were washed with alcohol and ether, dried in vacuum over sulphuric acid at laboratory temperature, and finally at 105° to 110° C. Feb. IS, 191S Organic Phosphoric Acid of Rice 427 over calcium chloric! without vacuum. The moisture was then determined in the salt at 120° to 130° C. and amounted to 0.76 per cent. The phytin from the bran was purified and obtained in the same way as that from unpolished rice, but the barium salt was dried in vacuum over calcium chlorid at 105° C. The moisture determined in vacuum at 105° and at 120° C. was 1.33 per cent in both cases. The crystals resembled those of the first preparation. The salts thus obtained were practically free from chlorids and inor- ganic phosphates. Nitrogen was also absent. All the material of the first preparation was used in making repeated determinations of barium, phosphorus, carbon, and hydrogen, but the phytin obtained from the bran was analyzed also for ash constituents other than barium. In 0.60 gm. of this material an unweighable trace of calcium was found, but no iron, manganese, magnesium, or potash. The residue on precipitating out the barium and igniting the phytic acid thus left amounted to a few milligrams and was composed mostly of unvolatilized phosphoric acid. No nitrogen was found in the salt. The barium in the salts was determined by various methods. A pre- cipitate of barium sulphate was obtained in the determination of phos- phorus by the Neumann method. After boiling the salts for three to four hours with sulphuric acid to which ammonium nitrate was added at intervals, 200 c. c. of water were added to the solution, and the precipitate of barium sulphate resulting from this dilution was boiled and allowed to stand overnight on the water bath. The precipitate was collected on a filter and was washed well, dried, and ignited. This precipitate was not pure barium sulphate, but contained a large amount of silica, aluminum, etc., dissolved from the Kjeldahl flask by the concentrated solution of very hot ammonium nitrate in sulphuric acid. On heating the weighed precipitate with hydrofluoric and sulphuric acids, silica was eliminated and the weight of the residue considerably reduced. In one case 0.6853 gm. of precipitate lost o.oi 1 1 gm. by this treatment. A pure barium sulphate was also obtained by fusing the residue with sodium carbonate. The fusion was digested in boiling water, then filtered and washed. The filtrate was tested for barium and the insoluble residue dissolved in a few drops of dilute hydrochloric acid, washed through the filter paper, and the barium precipitated in boiling solution by the slow addition of 0.4 N sulphuric acid. A number of determinations were thus made in which the correction by fusion caused a decrease in calculated barium from i to 2 per cent. The results from three samples given in Table I are taken from the analysis of the phytin from bran (not calculated to moisture-free basis). Tablij I. — Percentage of barium in barium phylatc oxidized by the Neumann tnelhod Barium. Sample Xo. Uncorrected. Corrected by fusion. 1 Per crnt. 3S. 293 38.256 38. 191 Per cent. 37- 291 37.358 37- 305 2 428 Journal of Agricultural Research Vol. III. No. s Two of the corrected residues were treated with hydrofluoric acid, but no decrease in weight resulted. Barium was also determined by dissolving the barium phytate in about 200 c. c. of water with a few cubic centimeters of dilute hydrochloric acid. Twenty c.c. of 0.4 N sulphuric acid were added to the boiling solution, the whole was boiled about half an hour, and then set aside on a hot-water bath for several hours. The precipitate, after standing overnight, was filtered, washed, and ignited. On weighing, then fusing and reprecipitat- ing the barium, as was done above, a slight increase in weight was ob- served. The filtrates from the barium precipitates after fusion were united, evaporated to small bulk, acidified with nitric acid, and ammonium molyb- date added. A yellow precipitate was found after standing, and the phosphoric acid then obtained was weighed as magnesium pyrophos- phate. By fusing the impure residues amounting to from 0.2754 to 0.351 1 gm., the magnesium pyrophosphate obtained was found to be 0.0181, 0.0092, 0.0095, arid 0.0074 gm. The weight of the residues of barium sulphate before and after correc- tion for phosphate is given in Table II. Table II. — Quantity of barium sulphatt "■ and viagtiesium pyrof>hosphale precipitated from a solution of barium phytate Samples No. Uncorrected. Fused and cor- rected. Phosphoric acid as magne- siuin pyrophos- phate. Gtn. 0. 2754 ■3"7 .4672 Gm. 0. 2785 .3128 •4703 Gm. 0. 0181 . ooge " Not calculated to water-free basis. The filtrate containing the phytic acid from which the barium was precipitated by the foregoing method still contained a few tenths of a per cent of barium. On evaporating to small bulk or igniting, a small precipitate of barium sulphate was obtained. It would appear that phytic acid has some solvent action on barium sulphate. The fact that phosphoric acid was precipitated along with the barium sulphate by very dilute sulphuric acid suggests that the composition of phytic acid as determined by Anderson may have been affected. The phosphorus, if carried down with the barium sulphate, would cause a low phosphorus content in the remaining solution. It was attempted to determine barium by first igniting the salt, but a white ash could not be obtained, and the residue was extremely difficult to dissolve after ignition. Phosphoric acid was best determined by the Neumann method, filtering off the barium sulphate formed on dilution. The results were about o.i per cent lower when determined by precipitat- ing the barium with dilute sulphuric acid and evaporating the phytic acid thus obtained vrith magnesium nitrate and igniting and adding the phosphorus anhydrid found in the fused barium residue. Carbon and hydrogen were determined by the regular combustion method, passing oxygen through the apparatus during the burning. In each case the black ash remaining was ground up with potassium dichromate and rebumed. Feb. IS, 1915 Organic Phosphoric Acid of Rice 429 The analyses of the barium phytate are given in Table III and are calculated to the water-free basis. The barium and phosphorus content is lower than that reported by Anderson for tribarium-inosite-hexaphos- phate. Whether the barium phytate obtained was composed of a single salt of inosite is not absolutely certain. Table III. — Analyses of barium phytate calculated to the water-free basis Source of barium phytate. Constituent. Unpolished rice. Rice bran. Sample No. I. Sample No. 2. Sample No. I. Sample No. 2. c Percent. 6.63 I- 75 16.43 36-93 Percent. 6-97 1.84 16.38 36.84 Per cent. 6.62 1.82 16. 06 37-79 Per cent. 6- SI 1.87 16. 05 37-84 H - ... . P. Ba UNPOLISHED RICE « Sample No. Quantity used. H2O COj MgsPjOT BaSOi Gm. 0. 3729 •3551 -4752 -2365 -4514 .5042 Gm. 0. 0607 - 0605 Gm. 0. 0900 .0901 Gm. Gm. I 0. 2779 •1379 I 0. 281 1 ■3132 KICB BRAN i> Sample No. Quantity used. HiO CO2 MgjPiOj BaS04 I Gm. 0- 2559 -3271 -5514 -4765 •5514 .6730 Gm. 0.044s •0583 Gm. 0. 0613 .0771 Gm. Gm. 2 I o- 3137 .2709 2 I 0. 3496 -4251 2 J 0.5107 gm. lost 0.0039 gm. H:0=o.76 per cent of moisture. ^ 0.9717 gm. lost 0.0139 gm. H20=i.33 per cent of moisture. Inosite was prepared from the barium phytate of rice bran by heating in sealed tubes to 150° C. about 2 gm. of the salt with 20 c. c. of 30 per cent sulphuric acid for five hours. The sulphuric acid was precipitated with barium hydroxid, the excess of barium removed by carbon dioxid, and the filtrate evaporated to dryness. The residue was extracted with hot water and filtered. The inosite was precipitated by ether and alcohol and rccrystallized three times as minute needles. These gave the Scherer reaction and melted at 223° C, uncorrected. Thanks are due Dr. W. P. Kelley, of the Hawaii Experiment Station, who suggested this work on phylin in rice and gave helpful advice throughout. 430 Journal of Agricultural Research voi. in. no.s LITERATURE CITED (i) Anderson, R. J. 1912. The organic-phosphoric acid compound of wheat bran. N. Y. State Agr. Exp. Sta. Tech. Bui. 22, 16 p. (2) (3) (4) (5) 1912. The organic phosphoric acid of cottonseed meal. N. Y. State Agr. Exp. Sta. Tech. Bui. 25, 12 p. 1912. Phytin and phosphoric acid esters of inosite. N. Y. State Agr. Exp. Sta. Tech. Bui. 19, 17 p. 1912. Phytin and pyrophosphoric acid esters of Inosite. — II. N. Y. State Agr. Exp. Sta. Tech. Bui. 21, 16 p. 1914. A contribution to the chemistry of phytin. N. Y. State Agr. Exp. Sta. Tech. Bui. 32, 44 p. (6) Hart, E. B., and Andrews, W. H. 1903. The status of phosphorus in certain food materials and animal by-products, Avith special reference to the presence of inorganic forms. In Amer. Chem. Jour., v. 30, no. 6, p. 470-485. (7) and TorriNGHAM, W. E. 1910. The nature of the acid-soluble phosphorus compounds of some important feeding materials. Wis. Agr. Exp. Sta. Research Bui. 9, p. 95-106. (8) Patten, A. J., and Hart, E. B. 1904. The nature of the principal phosphorus compoimds in wheat bran. In Amer. Chem. Joiu'., v. 31, no. 5, p. 564-572. (9) POSTERNAK, S. 1900. Contribution a I'^tude chimique de 1 'assimilation chlorophyllienne. Sur le premier produit d 'organization de I'acide phosphorique dans les plantes a chlorophylle avec quelques remarques sur le r6Ie physio- logique de I'inosite. In Rev. Gen. Bot., t. 12, no. 133, p. 5-24. (10) Rather, J. B. 1912. The forms of phosphorus in cottonseed meal. Texas Agr. Exp. Sta. Bui. 146, 16 p. (II) 1913. The phosphorus compoimds of cottonseed meal and wheat bran. Texas Agr. Exp. Sta. Bui. 156, 18 p. (12) Suzuki, Umetaro, and Yoshimura, K. 1907. Ueber die Verbreitung von "anhydro-oxy-methylen-diphosphorsauren Salzen" oder "Phytin" in Pflanzen. In Bui. Col. Agr. "Tokyo Imp. Univ. Japan, v. 7, no. 4, p. 495-502. (is) , , and Takaishi, M. 1907. Ueber ein Enzym "Phytase," das " anhydro-oxy-methylen diphos- phorsaure" spaltet. In Bui. Col. Agr. Tokyo Imp. Univ. Japan, v. 7, no. 4, p. 503-512. (14) Thompson, Alice R. 1913. The determination of sulphur and chlorin in the rice plant. In Hawaii Agr. Exp. Sta. Ann. Rpt., 1912, p. 64-73. TWO CLOVER APHIDS' By Edith M. Patch, \ Entomologist, Maine Agricultural Experiment Station It recently came to my attention that two distinct species of clover aphids are rather generally confused in collections under the name "Aphis bakeri." As the range of both species extends nearly, if not quite, across the continent, it is a matter of more than local interest that this confusion should be straightened out. Specimens have been received from Messrs. J. J. Davis, A. Maxson, and H. F. Wilson, and I am in- debted to Prof. C. P. Gillette for reading the manuscript and for the determination as to which species should properly be known as Aphis bakeri. I have been unable to secure a type specimen of Aphis brevis for examination, although through the kindness of Dr. L. O. Howard, Chief of the Bureau of Entomology, and Mr. H. Hayward, Director of the Delaware Experiment Station, a thorough search was made both at Wash- ington and at Newark. I feel confident, however, that Prof. Sanderson's careful description and figures ^ are sufficient to enable us to refer the long-beaked dover aphid to that species with safety. Aphis brevis Sanderson (Long-beaked clover aphid). In the vicinity of Orono, Me., the leaves of the hawthorn {Crataegus spp.) in June are commonly twisted into dark-purple swollen curls and are inhabited by an aphid the fall migrants of which were described by Prof. Sanderson as Aphis brevis.^ This insect takes flight from haw- thorn during June and early July and returns late in the season before producing the sexual generation. I have taken the fall migrants on cul- tivated plum (Primus spp.), but as yet have made no spring collections from that host. In June and July, 1906, I collected apparently the same species from the twigs and terminal leaf curls of the Japan quince {Cydonia japonica). Not being able to find characters to separate these collections from certain specimens labeled "Aphis bakeri" received from the Middle West, I undertook some transfer tests during the summer of 191 2, and found that my A phis brevis accepted both alsike and other clover {Trifoliumspp.) . Migrants placed on alsike and white clover produced nymphs that fed with apparent satisfaction on the test plants. The potted white clover was, however, more easily managed in the laboratory, so it was selected for the main rearings. The transfer was made on June 14. The migrants fed on the clover, and their abdomens became distended. At this time the head, thorax, and cornicles were black, and abdomens olive green, with distinct black lateral dots. By June 21 their abundant progeny were established on both stem and runner. The nymphs at first were pale and pellucid, with rosy head and prothorax. By June 24 this generation had matured, but did not begin to reproduce for a day or two. By June * Papers from the Maine Agricultural Experiment Station: Entomolojn^ No. 76. » Sanderson. E. Dwight. Report of the entomologist. In Del. Agr. Exp. Sta. ijtli Ann. Rpt. [1900) /or, p. 157-158. 1901. Journal of Agricultural Research, Vol. Ill, No. 5 Dept. of Agriculture, Washington, D. C. Feb. 15. 1915 Maine — 3 (431) 432 Journal of Agricultural Research voi. m. No. s 27 these adults had lost the rosy hue they had as nymphs and had become creamy white or grayish white with black antennae, dusky legs, and deep rusty spots at base of cornicles, but with no rusty line connecting them. On August 5, 1912, my attention was called to infested sweet-pea {Lathyrus odoratus) vines, which had vigorous colonies of red aphids on the stems at the surface of the ground extending for an inch up the plants. These proved to be the long-beaked clover aphids, and the source of the infestation doubtless was the hawthorn tree a few rods distant which had been heavily attacked by the spring generations of this species earlier in the season. The spring forms on the hawthorn include two types of apterous females. One, possibly the stem mother, has the head, prothorax, and thorax soft coral pink. Joints I, II, and III of the antenna are coral pink or pellucid, while the other joints are black. The abdomen is olive green mottled with brown and pink, with a slight bloom only. The very short cornicle is light pellucid with the merest dusky tip, and the very short cauda is dark brown. The apterous female of the second (or third?) generation has the head, prothorax, and thorax crimson, overcast with a slight bloom. Joints I, II, and III of the antenna are pellucid, while IV, V, and VI are black. The abdomen is crimson mottled with olive green and 6^- Fig. I. — Aphts brei'is: Antenna of fall alate female collected from hawthorn. covered with slight bloom. There is a pale olive space about the base of the short cornicles, which are light olive green. The spring migrant before flight has the dorsal surface of the head shining black, the ventral reddish, with a black beak and the anteimae black; the prothorax is black with red membrane; the dorsal lobes of the thorax are shining black, the breast reddish black; the dorsum of the abdomen is red in form of a heavy cross, the part about the cor- nicles being pale olive green; the venter is red. These same migrants after feeding on clover juice lose with age their reddish cast, the abdo- mens then becoming olive green. The nymph, which is to become the spring migrant, in the pre-imago stage has its head and thorax coral pink, its abdomen red and more or less mottled, with a pale olive green space about the cornicle extend- ing over several segments. The fall alate form does not differ essentially from the spring migrant. The sensoria of the antenna (fig. i) are distributed over joints III, IV, and V, in practically the same numbers as the spring migrant, although some individuals of the spring form have fewer to none on joint V. V and VI are nearly subequal, and they are both sometimes longer than those shown in figure i. Joint V is sometimes a little shorter than IV. Joints IV, V, and VI in the male are relatively longer than those in the alate female and there are frequently more sensoria on IV and V, as well as sometimes from one to several on basal part of joint VI (fig. 2). Feb. js, 1915 Two Clover Aphids 433 The lateral tubercles of the prothorax and abdomen, are distinct and blunt in both the alate and the apterous generations. The beak in the apterous forms ordinarily reaches easily to the second coxa, and in the alate forms well beyond the second coxa, sometimes reaching the third. The most distinctive character of the wing is the short, broad stigma with a blunt distal end. Aphis bakeri Cowen (Short-beaked clover aphid). About the middle of August, 1914, large numbers of an aphid from- Trijolium pratense were taken by Mr. George Newman at Orono, Maine. This species was distinct from the one just discussed, and yet I found that it was commonly listed in collections as Aphis bakeri. In the f o^^w;;^ogoi2serf«8!toJ--:ffl^i,18 PUBLISHED BY AUTHORITY OF THE SECRETARY OF AGRICULTURE, WITH THE COOPERATION OF THE ASSOCIATION OF AMERICAN AGRICUL- TURAL COLLEGES AND EXPERIMENT STATIONS EDITORIAL COMMITTEE FOR THE DEPARTMENT FOR THE ASSOCIATIOH KARL F. KELLERMAN, Chairman RAYMOND PEARL Physiologist and Assistant Chief, Bureau of Plant Indusiry EDWIN W. ALLEN Assisia*it Director, Office of ExPerimefU Slaiions CHARLES L. MARLATT Assistant Chief, Bureau of Entimology Bioloffislt Maine Agricultural Experiment Station H. P. ARMSBY Director, Institute of A?iimal Nutrition, The Pennsylvania State College E. M. FREEMAN Botanist, Plant Pathologist, and Assistant Dean, Agricultural Experiment Station of the University of Minnesota All correspondence regarding articles from the Department of Agriculttire should be addressed to Karl F. Kellerman, Journal of Agricultural Research, Washington, D. C. All correspondence regarding articles from Experiment Stations should be addressed to Raymond Pearl, Journal of Agricultural Research, Orouo, Maine. JOM£ OF AGRICETIAL RESEARCH DEPARTMENT OF AGRICUETURE Vol. Ill Washington, D. C, March 25, 191 5 No. 6 NET ENERGY VALUES OF FEEDING STUFFS FOR CATTLE By Henry Prentiss Armsby, Director, and J. August Fries, Assistant Director, histitute of Animal Nutrition of The Pennsylvania State College COOPERATIVE INVESTIGATIONS BETWEEN THE BUREAU OF ANIMAL INDUSTRY OF THE UNITED STATES DEPARTMENT OF AGRICULTURE AND THE INSTITUTE OF ANIMAL NUTRITION OF THE PENNSYLVANIA STATE COLLEGE INTRODUCTION Besides supplying certain specific forms of matter (ash ingredients, proteins, lipoids, carbohydrates, vitamines, etc.) essential to the normal course of metabolism, the feed of an animal is, so far as we know, the sole source of the energy whose transformations constitute the essential phenomena of physical life. This energy is contained in the feed as chemical energy, and the maximum quantity which any substance can furnish for the vital activities by its oxidation in the body is measured by its heat of combustion. It rarely, if ever, happens, however, that this maximum effect is realized. In practically every case a larger or smaller proportion of the chemical energy of the feed escapes unutilized. These losses of energy are of two general classes. First, a portion of the chemical energy of the feed fails to be trans- formed at all, leaving the body as chemical energy in the visible excreta and in the combustible gases arising from gastric and intestinal fermentations. Second, another portion of the chemical energy of the feed is indeed transformed, but at ordinary temperatures virtually results merely in a superfluous heat production. It is true that the metabolism consequent upon feed consumption is not only unavoidable but may be regarded as a necessary expenditure of energy for the support of the activities con- nected with digestion and assimilation. Nevertheless, from the stand- point of the net gain or loss by the organism this portion of the feed energy, which ultimately takes the form of heat and escapes from the body, must be regarded as a loss. The remainder of the chemical energy in the feed, after deducting these two classes of losses, has been designated as its net energy value and Journal of Agricultural Research. Vol. Ill, No. 6 Dept. of Agriculture, Washington, D. C. Mar. 35, 1915 A-13 (43 s) 436 Journal of Agricultural Research voi. in, no. e expresses the net effect of the feed either in causing a storage of chemical energy in the form of fat, protein, etc., in the body, or, in the case of a submaintenance ration, in diminishing the amount of energy which must be supplied by the katabolism of body tissue. The investigations of the last 30 years have shown that both the general problems of nutrition and the economic questions relating to the feeding of domestic animals may be advantageously studied from the standpoint of energetics. From this standpoint, it is of importance to determine as accurately as may be the losses of energy which feed substances undergo in the two ways just mentioned and the resulting net energy values. In the following pages are reported the results of a considerable number of experiments on cattle carried out at this Institute during the years 1902 to 1912, inclusive, in which these losses have been determined for certain feeding stuffs.' These experiments up to the end of 1907 have been already reported in full (7, 8, 9, 10) ^ and it is hoped to discuss the details of the later ones in subsequent papers. Here it will be con- venient, following a general description of the experiments, to consider: I. The losses of chemical energy. II. The expenditure of energy consequent upon feed consump- tion and its factors. III. Net energy values and their computation. GENERAL DESCRIPTION OF THE EXPERIMENTS The experiments were made with the aid of a respiration calorimeter of the Atwater-Rosa type, the essential features of which have already been described (3, 4, 7). The apparatus permits a determination of the water vapor and carbon dioxid excreted, of the carbon and hydrogen in the combustible gases produced, and of the heat given off, but not of the oxygen consumed. In addition to the ordinary feeding stuffs analyses of feed and excreta, the quantitative collection of the feces and urine and the determination of the amounts of carbon, hydrogen, and energy con- tained in them were also necessarily involved. The experiments com- prised, in all, 76 single feeding periods. Each period covered at least 3 weeks, of which 11 days or more constituted a preliminary period, while the visible excreta were collected for the last 10 days, during which, on the seventh and eighth days, the complete balance of matter and energy was determined for 48 consecutive hours in the respiration calorimeter.^ The accuracy of this instrument was tested by means of numerous alcohol checks. The results of 18 such checks (10, p. 217-222) showed that the • In all. over 30 persons have taken a more or less direct part in the respiration trials and in the large amount of analytical, clerical, and miscellaneous work involved in the experiments. For obvious reasons, it is impossible even to attempt any statement of the exact part taken by individuals or to make acknowl- edgments for the specific work done by each person. This is all the more true because the most important factor in whatever success the investigation has attained, and one which by its nature is incapable of such partition, is the loyalty and zeal which all concerned have shown in the execution of the plan of the investigation and in securing the greatest attainable accuracy of details. ^ Reference is made by number to " Literature cited." p. 4S9-491. 3 For details regarding the methods employed compare the bulletins of the Bureau of ,\nimal Industry already cited (7, 8, 9, 10), especially Bulletin 128 (10), p, 200-216, as well as the detailed descriptions of the single experiments contained in that bulletin. Mar. 25, 191S Energy Values of Feeding Stuffs for Cattle 437 values obtained in a single experiment maj' be regarded as accurate to within the following percentages of the amounts determined: Carbon dioxid, 0.5; water, 6.0; heat, i.o. A further test of the accuracy of the work is found in a comparison of the observed heat production with that computed in the ordinary way from the balance of carbon and nitrogen. Comparisons of this sort for 57 feeding periods up to the end of 1909 (5, 6) showed an average differ- ence of 0.4 per cent. The results of the later comparisons reduce this difference to 0.3 per cent, or, if the unsatisfactory results of the year 1905 be omitted, to 0.04 per cent. The total amounts of heat involved are as follows : :6 periods. 68 periods. Computed heat production Calories. . i, 338, 887 i, 231, 711 Observed heat production do. .. . 1,343,071 1,231,251 Difference do ... . 4, 184 460 Percentage difference o. 31 o. 04 While the basis for the computation of the heat production is not alto- gether satisfactory, especially in the absence of determinations of the gain or loss of glycogen by the animal, nevertheless the general agreement is such as apparently to preclude the existence of gross errors. FEEDING STITFFS The dry matter of the several feeding stuffs used had the following average composition, as shown in each case by concordant analyses of two or more separate samples taken at the beginning of each experiment (Table I). The more important determinations were also repeated upon samples taken when the feed was weighed out for each period, and these period results form the basis of the computations on subsequent pages. Table I. — Composition 0/ the dry matter of the feeding stuffs Feeding stuff and experi- ment No. Timothy hay: 174 190 200 207 Red clover hay : 179 186 Mixed hay : 211 Alfalfa hay : 208 209 212 Alfalfa meal : 212 Maize stover; 210, total 210, portion eaten Ash. Protein. Per cent. 4.64 4.87 5.86 Per cent. 5.08 6.72 5.01 6. 90 6. 40 6.57 12. 90 II. 18 7.01 9. 62 9.40 10. 76 9. 06 11.86 12. 09 12.39 9.24 11-75 6. 14 6.67 4.09 4.46 Non- protein. Per cent. 0. 24 •39 ■83 .24 1. 61 .89 I. 29 2.52 I. 67 2.86 2.87 .86 .98 Per cent. 38.92 38.15 33- 02 31- IS 31.61 28.78 33-78 31.96 31- 70 30. 10 31. 12 36.15 35-23 Nitrogen- free extract. Per cent. 49.01 49-51 51. CI 54-55 44.81 49-65 46. 00 42- 75 41.86 43-63 43-17 51.46 Ether extract. Per cent. 2.08 2. 01 2. 56 2-15 2. 67 2-93 2.30 1-51 I. 92 I. 96 1.85 1.30 1. 41 Heat of combus- tion per kilogram. Caiories* 4-554 4,431 4,516 4,505 4,457 4,492 4,396 4,403 4,330 4,368 4,374 4,337 4,332 438 Journal of Agricultural Research Vol. HI. No. 6 Table I. — Composition of the dry matter of the feeding stuffs — Continued Feeding stuff and experi- ment No. Maize meal: 179 211 Wheat bran : 190 Grain mixture No. i 'fl 200 207 Grain mixtiue No. 2 ;6 208 209 Hominy chop: 211 Per cent. 1-37 I. 62 7-52 4. 10 4. 14 2.78 2-54 2-75 Protein. Per cent. 9-94 9.29 14. 50 16.97 17-95 12. 07 13.21 9-33 Non- protein. Per cent. .48 .24 -49 3-43 2. 22 .78 ■54 I. 29 Grade fiber. Per cent. 2. 60 3. 16 11.49 6. 09 6.16 5-63 5.81 5-13 Nitrogen- free extract. Per cent. 81.38 80.64 61.88 62.89 64.38 73-65 72-77 72.65 Ether extract. Per cent. 4-23 05 8.85 Heat of combus- tion per Idlogram. Calories. 4,431 4,517 4,532 4, 690 4,670 4,604 4,617 4,709 oWheat bran, 14.28 per cent; maize meal, 42.86 per cent; old process linseed meal. 42.86 per cent. bMaize meal, 60 per cent; crushed oats, 30 per cent; old process linseed meal, 10 per cent. ANIMALS Nine different steers have been used, varying in age from 1 1 months to approximately 60 months at the beginning of the several experiments.' They were either full bloods or high grades of recognized beef breeds, with one exception, steer B, which was distinctly of the dairy type and of mixed breeding (scrub), Jersey blood apparently predominating. All were docile animals and were thoroughly accustomed to the necessary handling, to wearing the apparatus for the collection of excreta, and to their surroundings in the digestion stall and the calorimeter. Further particulars concerning them are contained in Table II. Table II. — Description of the animals used in the experiments Animal No. Breed of animal. Nos. of experiments in which used. Age at begin- ning of each experiment. Average live weight in each experiment. I Grade Shorthorn Aberdeen Angus Scrub f 174 179 [ 186 f 19° 200 207 190 200 I 207 208 f 208 210 I 211 208 209 211 212 Months. 36 48 60 II 23 35 13 25 37 9 9 21 33 9 21 28 20 Kilograms. 408 528 572 274 408 A B 51° 19s 3°3 380 27s 167 33^ 449 208 C Grade Hereford ... do D E do F do. . 300 379 345 G Full-blood Hereford. .. Full-blood Shorthorn. . H * The word "experiment" is here used to designate the work of an entire season, including several feed- ing periods. Mar 2s, 191S Energy Values of Feeding Stuffs for Cattle 439 I. LOSSES OF CHEMICAL ENERGY— METABOLIZABLE ENERGY The losses of chemical energy occur substantially in the feces and urine and in the combustible gases. The energy content of the dried feces and urine is readily determined. In investigations at this Institute, Braman (16) has shown that the loss of energy in the drying of urine may be estimated with a good degree of accuracy, the error being insignificant in comparison with the total energy of the feed. The possible loss of energy in the drying of the feces has not yet been investigated directly, although Fingerling, Kohler, and Reinhardt (18) have obser\'ed a loss of carbon, the amount of which they do not state. The energy content of the combustible gases is not susceptible of direct determination, but must be estimated from their chemical com- position. The combustible gases which have been actually identified as excreted by cattle are methane and hydrogen. All investigations are in accord in showing that the former is the chief product of the normal fermentations occurring in the digestive tract, but results differ regarding the extent to which hydrogen is formed. In our experiments the gases were analyzed by passing them over platinized kaolin at a red heat. By this method in almost every instance a ratio of C to H slightly greater than that in CH^ (2-976 to i) has been found, the average of 57 experi- ments reported by the junior author elsewhere (19) being 3.167 to i, with considerable variations in individual cases. We are inclined to think that this high figure is due to failure to oxidize the last traces of hydrogen in the combustion tube. On the other hand, Markoff (37, 38), in his extensive investigations of paunch fermentation in cattle, found in nearly every instance a small amount of hydrogen, and Von der Heide, Klein, and Zuntz (20), in respiration experiments upon an ox, observed a small excretion of hydrogen in two cases out of four. In computing the energy losses in the following experiments it has been assumed that the combustible gases consisted of CH^ (methane), and the computa- tions have been based upon the observed quantity of carbon. The difference between the chemical energy of the feed and that lost in the excreta shows how much of the former is capable of being con- verted into other forms in the body, either during the changes which the feed undergoes in the digestive tract or in the course of metabolism in the tissues. This convertible portion of the feed energy has been given various names by different investigators, such as "physiological heat value," "fuel value," "available energy," etc. Without entering here into a discussion of the propriety of these names, we have preferred for our present purpose to follow the suggestion made earlier by the senior author (2, p. 270) and to designate it as " metabolizable energy." By this term is meant simply the energy capable of transformation in the body, with no implications as to the proportion of the energy thus trans- 440 Journal of Agricultural Research voi. in, no 6 formed which can be utilized by the organism. The heat evolved during the methane fermentation, for example, constitutes part of the metab- olizable energy as thus defined, although it does not enter into the tissue metabolism. The determination of the losses of chemical energy from a single feed- ing stuff or from a mixed ration is relatively simple, as is illustrated by the following example taken from the results on steer B in experiment 207. Computation of losses oj chemical energy from a ration Energy of feed: Period 1. Period 3. Timothyhay i2,477Cals. 12, 618 Cals. Grain mixture No. i 12, 549 Cals. Total 25, 026 Cals. 13, 618 Cals. Ener^ of excreta: Feces 7, 371 Cals. 5, 247 Cals. Urine I, 536 Cals. 627 Cals. Methane 2, 098 Cals. 1,057 Cals. Total 11,005 Cals. 6,931 Cals. Metabolizable energy 14, 02 1 Cals. 5, 687 Cals. Since concentrates can not be fed alone, the losses of chemical energy which they suffer, like their digestibility, must be obtained by means of a calculation by difference, which in period 2 of the foregoing example is as follows : Computation of losses of chemical energy by a concentrate chemical Chemical energy of excreta. Jletabo* energy ' " lizable of feed. Feces. Urine. Methane. energy. Calories. Calories. Calories. Calories. Calories. Total ration 25,026 7i37i 1.536 2,098 14,021 Computed for hay .12,477 5.254 59^ i.°°3 5.629 Grain mixture by difference 12,549 2,117 945 i. °95 8,392 Computed in the manner just illustrated, the losses of chemical energy per kilogram of dry matter consumed in these experiments and the metabolizable energy remaining are shown. in Table III, which includes also the percentage distribution of the feed energy between the various excreta, on the one hand, and the metabolizable energy, on the other.* For convenience, the average results for the metabolizable energy per kilogram of dry matter and per kilogram of digestible organic matter are brought together in Table IV. ' In all cases the obser\'ed energy' of the urine has been corrected to nitrogen equilibrium of the animal by adding 7.5 Calories for each gram of nitrogen retained by the animal or subtracting the same amount for each gram of body nitrogen lost, the correction being regarded as representing energy of excretory material temporarily retained in the body. Mar. as. 191S Energy Values of Feeding Stuffs for Cattle 441 Table III. — Losses of energy and their percentage distribution Feeding stuff and experiment No. Timothy hay: 174'^ 190 igo 190 190 Average. 300. 200. 300. 300. Average. 307. 307. 307. 307. Average. Red-clover hay: 179 . 179 Average. 186. 186. 186. Average. Mixed hay: 311 311 3ZI SIX 3X1 3XX Average. Alfalfa hay: 208 208 308 308 308 308 308 308 D-A 3 Average 1 4.407 Dry matter eaten per day and head. 4.483 2,001 3.493 1.774 2,610 2,470 2,647 4.42s 2,470 3.80s 2,974 4,892 2,79s 4.630 4.439 3.144 2.933 5.02s 4.139 Gm 3S6 6, 204 3.498 1,786 6.092 3.149' 1,608 2,155 1.300 4.170 2.408 1. 413 4.744 3.099 3, 119 Energy per kilogram of dry matter. Cats 4.554 4.502 4.488 4.503 4.483 4.494 4.S10 4.509 4.510 4.509 4.509 4.509 4.521 4.509 4.521 4.513 4.450 4,426 4.438 4,490 4,489 4,478 4.486 4,400 4.391 4.390 4.393 4.391 4.39° 4.393 4.407 4.407 4,410 4.407 4,406 4.403 4.407 4,406 losses. Cats. 2.439 2. 209 2.306 2. 209 2. 176 2,22s 2.17s 2.250 2.092 2. 164 2, 170 1. 841 1,875 1. 918 1,940 1. 8s 1,842 1. 817 1.852 1,837 2.077 1.837 1,956 1,906 1,878 1. 917 Cats 156 164 128 141 125 139 ISS 153 168 167 204 199 224 208 306 326 301 290 306 209 234 255 208 220 238 1.929 3.X34 2, X86 2.043 2, 105 2,037 2. I 2,066 2,013 344 266 23 X 249 248 233 243 250 Cats. 28s 309 303 342 304 315 "301 '326 b 368 357 378 355 364 309 324 303 243 254 267 285 332 345 300 319 3S8 363 273 341 Cats 1.674 X,820 I. 751 x.Sii 1.878 la 4* O go J*ercentage losses. -- Cats. 3.483 3.429 3.480 3.399 3.484 53.56 49.06 51-38 49.07 48.55 1,815 3.448 49.51 1.866 1. 80s 1.924 1,866 1, 86s 3,096 2.077 2.032 2.040 2,061 1.895 X.937 x,926 2,019 2, X29 3,082 2,076 1,829 1,990 1834 3.59 3.566 3.565 3.565 49-91 46-39 48.00 3.573 48.13 3,457 40.82 3.4821 41-77 3.378 41. S8 3.456. 42 3.443 41.66 3.373 43-61 3.452 41-95 3.460 3.637 3.578 41.03 40.49 41.36 3,5S8r 40-95 3,442 47.20 3,406 41.83 3,294 44. S7 1.979I 3.463 43-39 1.974 3.420^ 42.76 1.877 3.313 43.68 1,9x4 1,777 x,682 .895 3.390 43.92 3,529 48. 2X 3.437 49. 6x 3,697 46-33 276 1,777' 3.5x8 47.78 281 x,84o[ 3,528 46. 24 2x8 1.768 3.655 49-61 270 x,828 3.594 46-90 280 X.863 3.576 45.69 3-44 3- 64 2.86 3-14 2.79 3- 09 3-44 3-40 3-72 3.70 3.57 4-52 4.40 4-97 4-6x 4- 61 6.8: 6- so 6.69 7-25 6- 72 6.48 6. 25 6.87 6.75 7.60 6.78 7. 01 6-97 6.68 7.24 6- 92 0-94 8.X7 7.90 8.38 7-84 8.07 6.94 7-33 6.75 5-41 5.68 6.49 7' 7.85 6.82 36.7s 40.43 39. ox 40. 20 41-88 40.39 41.37 40-01 42- 6s 41-38 41.36 46.49 45.93 45.07 45-13 45-66 42.58 44-22 44-97 47-38 46.48 46.28 41.56 45- 30 41-78 43-04 7.27, 44.97 8. xsl 42. 76 7.3s 43.56 5-54 6.03 5.24 5.64 5.63 5.29 5.51 5-68 5-93 6.X7 3.46 6.26 40.32 38. X9 42.97 40.33 6.37 41.76 4.95 6. X2 6.37 40.15 41.47 42. 26 2.09s 246 262 1,804' 3.567 47.54! 5.58! 5.94 40.94 and therefore differ somewhat from the averages of o From analyses of individual samples for each period, Table I. t> Corrected to .V equilibrium, using Rubner's factor, 7-45 Calories per gram A', c Computed by difference. 442 Journal of Agricultural Research Vol. III. No. 6 Table III. — Losses of energy and their percentage distribtition — Continued Feeding stuff and experiment No. Alfalfa hay— Con. 309 309 2og Average. 212. 212. 313. Average. Alfalfa meal: 213. 212. 212. Average. Maize stover: 3IO 3IO 310 Average. Maize meal: 179 179 ■ Average. Average. "Wheat bran: 190 190 190 190 Average. Grain mixture No. i: 200 300 300 300 Average. 307. 207. 207. 307. Average. Dry matter eaten per day and head. Gm. 6,174 3,562 2, 226 6,638 S.320 3.032 6,671 5. 408 3,ISS 4.335 3.548 2.563 3.163 3,186 790 2.383 1,978 1.996 I, 712 1.779 3,608 2.631 2.433 2.4S4 2.935 2.949 2. 761 Gm. 735 3.451 I. 542 4.644 1.370 2.583 999 1.803 J. 792 4.173 1. 195 2, 129 1.996 4.759 1,398 2.677 Energy per kilogram of dry matter. Cab. 4.359 4.32' 4.328 4.338 4.354 4,338 4.413 4.368 4.364 4.379 4.379 4,333 4,327 4.337 4.360 4.366 4.363 4.526 4,508 4.S4S 4.558 4,563 4,461 4.532 4,698 4.695 4.697 4,695 4.696 4.658 4.' 4.658 4,688 Losses. Cah. 2,046 1,961 2.078 3,028 1,796 1.805 1,64a Cab. 318 344 346 336 275 370 390 1,748 1, 726 1,838 1,860 1,837 1,867 1, 855 633 401 383 948 665 1.484 1.440 1,500 1.336 611 1. 051 639 1,064 841 951 1.019 1. 104 791 353 350 370 185 195 184 441 167 175 125 330 321 254 271 358 388 385 322 338 316 294 363 353 Cab 248 264 281 264 277 283 299 250 259 293 267 335 335 354 406 S09 373 365 311 332 340 a 390 "358 "381 "343 368 397 342 439 409 01 o 2 aJ So Cab. Cab. 1.847 3.652 1.859 3, 1.733 3.470 1,810 3,006 1,980 2,181 2.056 1,973 1.973 2, 091 1,966 1,970 1.921 2.798 3.392 3.460 3.063 3.261 2,466 2.586 2.477 2.SI4 2,511 3.339 2.998 3.294 3,966 3.149 2,994 3,°33 2.752 3,13s 3.S70 3.550 3.525 3.729 3.601 3,671 3.646 3.727 3.681 3.488 3.493 3,369 3,186 3,716 3,82s 3,928 3.877 4,003 3.982 3.95 3,872 3-954 3.946 3.974 3.923 3.998 3.960 3.876 3.991 3.6S3 3., Percentage losses. 46.95 45-31 48.01 46.75 41. 26 41- 60 37-21 43-29 43-31 39-41 42-93 42-47 43-06 42-82 14-52 9-18 11.85 8-45 21.04 32.67 31-59 32-87 29-96 5-00 5- 63 5-67 6-31 6- 23 6- 58 5-69 6. 6-49 6-09 6-37 6-53 6, 6-36 5-78 3- 96 4-27 4- SO 10. 12 3-83 6-97 3-87 3-78 5-06 4-8s S-56 6-07 31-77 5-38 13-00 32-38 13-60 32- 66 17-91 20-42 21-73 23-71 16-87 7-62 6.13 8-19 6-86 7-30 6.78 6.38 7-79 7-53 6-55 5-74 5-93 6-66 7-74 7-75 8- IS II- 20 9-31 11- 25 8-25 8.04 6.82 7.28 7. 62 7-44 8-30 7-63 8-11 7-30 7.84 8.52 7.29 9.42 8-72 4.673 966 332 397 3,978 3,860 20.68 7.09 a Computed from digestible carbohydrates. 42.36 42.96 39-83 46. 06 45-64 49-43 45-19 45-04 47-75 45-37 45-51 44-29 45-06 64- 16 77-68 76-43 67.93 54-24 56-74 54-29 56-35 55-41 71-08 63-86 70. 10 63.18 67-05 64-38 64-70 59- 08 66-88 8-48 63.75 Mar. 25, 1915 Energy Values of Feeding Stuffs for Cattle 443 Table III. — Losses of energy and their percentage distribution — Continued i .§ a 6 12; T3 0 Dry matter eaten per day and head. Energy per kilogram of dry matter. n ^ II i go Percentage losses. j3 "a 2 Feeding stuff and experiment No. 1 2 s a i Losses. 3 0 a •§ 3 3 M 0 3 1 •Si a 6 3 a u OS 1 Grain mixture No. 2: 208 F E E C C I 2 3 2 3 Gm- j,oS6 S90 387 731 521 Gw. 2,122 1,163 764 1.453 1,018 Cats. 4.607 4,611 4,595 4,611 4.594 Cats. 1,258 840 901 1.246 826 Cats. 173 202 236 192 242 Cats- 349 442 346 463 486 Cats. 2,827 3.127 3,112 2,710 3,038 Cats. 3,887 3.837 3,890 3.706 3,711 27.32 18.22 19.61 27-03 17.98 3-75 4-38 5-14 4-16 5-28 7-s8 9-57 7-54 10-04 10-58 61.36 67-83 208 208 58-77 208 66.16 4,604 4,629 4,605 4,604 1,014 1,312 1,032 891 209 170 184 223 417 283 327 376 2,964 2,864 3,062 3. "4 3.806 3.979 3.980 3.896 22.03 28.35 22.40 19.36 4-54 3-68 3-99 4.85 9.06 6.11 7.10 8.16 64.37 F F F I 2 3 1,462 894 536 3.040 1,862 1,112 61. 86 66.51 67-63 4,613 4,720 4,698 1,078 542 603 193 195 167 329 496 371 3,013 3,487 3. 557 3.952 3.993 4,156 23-37 11.48 12.83 4.17 4-13 3-56 7.12 10.50 7.90 65.34 D D 2 3 l>747 3.910 1,764 3.949 73-89 75-71 4i709 4,531 4. 535 4.532 4,543 4.536 573 1.526 1.247 1.291 1.536 1.264 181 195 214 237 208 242 433 318 380 316 390 406 3,522 2,492 2,694 2,688 2.409 2,634 4,075 3,810 3,777 3.824 3.687 3.696 12.15 33-69 27-50 28.49 33-81 27-86 3-84 4-30 4-72 5-23 4.58 5-33 9. 20 7.01 8.37 6.98 8.61 8.95 74-81 Alfalfa hay and grain mixture No. 2: 208 E E E C C I 2 3 2 3 1,086 S90 387 731 521 2, 122 I, i6j 764 1.453 1.018 55-00 208 59-41 208 59-30 208 . 53-00 208 57-86 Average.. 4,535 4.527 4.S16 4.51S 1.373 1.540 1.352 1.259 219 189 199 22s 362 27s 30s 338 2,581 2.523 2,660 2,693 3,759 3,872 3.888 3.826 30-27 34-67 29.94 27.88 4-83 4- 26 4-40 4-99 7-98 6-18 6-74 7-48 56.92 F F F I 2 3 1,46a 894 536 3,040 1,862 1,112 54-89 58-92 59.65 4,519 4.483 4.468 1.384 897 1.271 204 189 156 306 443 352 2,625 2.954 2,689 3,862 3. 701 3.763 30.62 20.00 28.45 4- SI 4.22 3-49 6-77 9-88 7.88 58.09 Mixed hay and maize meal: G G 2 3 790 2,383 1.542 4,644 65.90 60.18 4.476 4. .560 4, 553 1,084 1,261 1.293 173 209 196 398 403 340 2,821 2,687 2,724 3.732 3.775 3.879 24.22 27.6s 28.40 3-87 4-58 4-30 8.89 8.84 7-47 63.0a Mixed hay and hominy chop: D D 2 3 1,747 3,910 1,764 3,949 58.93 59-83 4.557 203 37' 2,706 3,827 28.03 4-44 8.15 59-39 444 Journal of Agricultural Research Vol. Ill, No. 6 Table IV. — Average losses of chemical energy — metabolizable energy Kind of feed and experiment No. Timothy hay: 174 190 200 207 Average . Red clover hay: 179 186 Average . Mixed hay: Alfalfa hay : 208 209. 212. Average . Alfalfa meal: 212 Average of alfalfa hay and meal . Maize stover: 210 Maize meal: 179 o . . , 211. ... Average . Wlieat bran: 190 Grain mixture No. i : 200 207. Average . Grain mixture No. 2 : 208 209 Average . Hominy chop : Gross energy per kilogram of dry matter. Calories. 4,554 4,494 4,509 4,513 4.518 4,438 4,486 Losses of chemical energy- per kilogram of dry matter. 4,462 4,393 4,407 4,338 4,368 4,371 4,374 4,372 4,332 4,366 4,517 4,442 4,532 4, 696 4,673 4,685 4,604 4,613 4,609 4,709 Calories. 2,880 2,679 2,644 2,452 2, 664 2,512 2, 410 2, 461 2,479 2,603 2, 528 2,312 2,481 2,362 2,451 2,380 974 1,256 1,115 2, 021 1,547 1,695 Metabolizable energy. Per kilo- gram of dry matter. Calories. 1,674 1,815 1,865 2, 061 I, 621 I, 640 I, 600 I, 620 1,187 1,854 1,926 2, 076 1,914 I, 804 1,810 2,056 1, 890 2, 012 I, 921 1.952 3,392 3,261 3,327 2,511 3,149 2,978 3,064 2,964 3,013 2,989 3,522 Per kilo- gram of di- gestible or- ganic matter. Calories. 3,483 3,448 3,573 3,443 3,487 3,413 3,558 3,486 3,390 3,567 3,570 3,601 3,579 3.681 3,60s 3.450 3,716 3.877 3,797 3,954 3,960 3,860 ,910 3,806 3.952 3.879 4,07s o Period 4. Mar. 2s. 191S Energy Values of Feeding Stuffs for Cattle 445 The results recorded in Table III illustrate the familiar fact that the greatest loss of chemical energy, especially in the case of coarse feeds, is that in the undigested feed residues of the feces and in the relatively small amounts of excretory products which, in the case of cattle, accom- pany them. The relative proportions lost in the urine and in the methane naturally vary with the composition of the feed, one rich in protein tend- ing especially to increase the energy content of the urine, while carbohy- drates tend to increase the excretion of methane. Of greater interest, however, is the variability of the losses suffered by the same feeding stuff in different periods. INFLUENCE OF QUANTITY OF FEED CONSUMED ON LOSSES OF CHEMICAL ENERGY In considering this question it should be borne in mind that the com- parisons here reported are in every instance between different amounts of the same ration — i. e., they deal with the influence of quantity only and do not touch the question of the influence of heavy grain feeding. Fur- thermore, they relate to comparatively light feeding, many of the periods having been upon submaintenance rations, while the total dry matter of the feed seldom reached 18 pounds per 1,000 pounds of live weight. The experiments recorded in Table III include 31 cases in which different amounts either of a single feeding stuff or of an identical mixed ration were consumed by the same animal in two different periods of the same experiment, under conditions as nearly imiform as it was possible to make them. The results may be most conveniently compared on the basis of the percentage distribution of the energy, as shown in the last four columns of the table. In the following comparisons the results computed by difference for the concentrated feeds are not included. LOSSES IN METHANE In a single instance (alfalfa hay and grain mixture No. 2 in experiment 208, steer E, periods 2 and 3) the percentage loss in the methane was greater on the heavier of the two rations and in another case (corn stover in experiment 210, steer D, periods i and 2) the difference was only o.oi per cent. In two cases the determinations of methane are believed to have been inaccurate. In the remaining 29 cases the percentage loss in methane was distinctly greater on the lighter ration, the differ- ence ranging from o. 11 to 2 per cent and tending, on the whole, to be somewhat greater on the mixed rations, with their larger proportion of readily soluble carbohydrates, than on those consisting exclusively of coarse fodder. In other words, as would be anticipated, the bacterial fermentation of the carbohydrates in the digestive tract of cattle pro- ceeds to a distinctly greater extent on light than on heavy rations. LOSSES IN URINE The percentage of the feed energy excreted in the urine was also greater on the lighter ration in 28 cases out of 33, the exceptions being two experiments on alfalfa hay (experiment 208, steer E, periods 5 and 6, 446 Journal of Agricultural Research voiiii, no6 and experiment 212, steer H, periods i and 3), two on clover hay (experi- ment 179, steer I, periods i and 2, and experiment 186, steer I, periods 2a and 3a), and one on alfalfa meal (experiment 212, steer H, periods 2 and 4). In the remaining 28 cases there is one in which the difference amounts to only 0.02 per cent and two in which it is 0.04 per cent. In the remaining 25 it ranges from 0.12 to 0.78 per cent. This greater relative loss in the urine on the lighter ration can not be attributed to the presence of nitrogenous substances derived from an increased katab- olism of body protein, since the energy of the urine has been at least approximately corrected to nitrogen equihbrium (p. 440) . Since it is well established that the urine of cattle contains a considerable quantity of nonnitrogenous substances (2, p. 312-314, 320-322), it seems not impos- sible that the more extensive fermentation on the hghter ration may have resulted in an increase of these unknown constituents. LOSSES IN FECES The results regarding the losses of chemical energy in the feces are by no means so uniform as in the case of the methane and of the urine. In 22 out of 33 cases there is a distinctly smaller relative loss of energy in the feces with the lighter ration — i. e., a greater apparent digestibility — the difference in the percentages ranging from 0.28 to 8.45. In the other third of the cases, however, the difference is in the opposite direc- tion, ranging from 0.37 to 2.74, with the exception of one case of prac- tical equality, so that it appears that other factors besides the extent of the methane fermentation affected the percentage digestibility. Two rather marked cases of a greater loss of energy in the feces on the lighter ration are found in experiment 211, periods 4 and 5, with relatively very small rations. Whether the relative loss in the feces increases or de- creases with an increase of the ration seems to bear no relation to the total quantity of feed consumed either per head or per 500 kg. of live weight. The 10 instances in which a greater percentage loss in the feces was observ^ed on the lighter of the two rations include, it is true, the more extreme rations as regards the total quantity, but the averages for the two groups are not widely different (4,376 and 3,952 gm. of dry matter). PERCENTAGE OF FEED ENERGY METABOLIZABLE The bearing of the foregoing facts upon the percentage of the feed energy which is metabolizable is obvious. Clearly the fermentation which plays so large a role in the digestive processes of ruminants was relatively more intense on the lighter rations, resulting in the breaking down of a larger proportion of the carbohydrates and in a greater loss of chemical energy in the methane, accompanied in most instances by an increased loss in the urine also. On the other hand, however, the organic acids resulting from the fermentation are resorbed and oxidized in the body, and their energy, together with the heat evolved in the fermentation, constitutes part of the metabolizable energy of the feed as defined on page 439. Whether the proportion of the total energy of the feed which is metabolizable be greater or less on the lighter ration will depend, therefore, upon the nature of the additional carbohydrates fer- mented. If the increased fermentation attacks the more insoluble carbohydrates, which would otherwise escape digestion entirely and Mar. 25, 191S Energy Values of Feeding Stuffs for Cattle 447 reappear in the feces, the result will be a relative increase in the metab- olizable energy. On the other hand, if the greater fermentation on the lighter rations is at the expense of the more soluble carbohydrates which would otherwise be digested by the enzyms of the intestines, the metabolizable energy will be diminished by the quantity of chemical energy escaping in the methane (and urine). It is perhaps not sur- prising, therefore, to find that the effect of the quantity of feed upon the percentage of energy metabolizable was somewhat variable. In 19 out of 22 cases mentioned in the previous paragraph in which the percentage loss of energy in the feces decreased as the amount of feed consumed was diminished, the percentage of energy metaboUzed did in fact increase, while in the remaining 3 cases the greater losses in methane and urine overbalanced the effect of the increased digestibiUty. In each of the 10 instances in which the percentage loss of energy in the feces was greater on the lighter rations, the percentage metabohzable shows a corresponding decrease, so that of the entire 33 comparisons 14 show a greater and 19 a less percentage metabolizable on Ughter as compared with heavier rations. On the whole, then, the differences in amount of feed consumed, within the limits of these experiments, failed to show any unmistakable effect upon the quantity of energy actually liberated in the body from a unit weight of feed. Morever, it must be borne in mind that a considerable part of the additional energy secured by the more extensive fermenta- tion of the lighter ration is liberated in the digestive tract as heat of fermentation and does not enter into the energy exchange of the body tissues, so that the difference in the net nutritive effect is hkely to be less than that in the metabolizable energy as ordinarily defined. INFLUENCE OF INDIVIDUALITY ON LOSSES OF CHEMICAL ENERGY In five of the experiments the same feeding stuff or mixture of feeding stuffs was fed to more than one animal, although unfortimately the amoimts consumed were not the same either per head or in proportion to the live weight, so that exact comparisons are not possible. In experiments 190, 200, and 207 a pure-bred Shorthorn steer was com- pared with a so-called scrub. A comparison of the averages for the Ughter and the heavier rations of timothy hay, respectively, for the three successive years gives the averages shown in Table V, which fail to show any distinct individual difference between the two animals. The results computed by difference for the wheat bran in experiment 190 and for the grain mixture No. i in experiments 200 and 207 show somewhat larger numerical differences, but when the errors incident to such calculations by difference are considered they agree with those upon hay in showing no material difference between these two animals. This point has been discussed from a slightly different standpoint else- where (10). 448 Journal of Agricultural Research Vol. III. No. 6 T.\BLE V. — Influence of individuality of cattle on losses of chemical energy Loss of energy. Metabo- In feces. In urine. In methane. lizable. Timothy hay (experiments 190, 200, and 207): Average of light rations — Steer A Per cent. 46.03 46.03 Per cent. 3-87 3-96 Per cent. 7-34 7-79 Per cent. 42.76 42. 22 Steer B 0 - .09 - -45 + -54 Average of heavy rations — Steer A 47.69 46.32 3-55 3.70 7- " 7.18 41.65 42. 80 Steer B Difference + 1-37 - -15 - -07 - I- IS Average of all rations — Steer A 46.86 46. 18 3-71 3-83 7.22 7.49 42. 21 42.50 Steer B "OifFereTice -f .68 — . 12 - -27 - .29 Alfalfa hay (experiment 20S) : Average of medium and light ra- tions— .Steer D (periods i and 2) Steer E (periods 5 and 6) Steer C (periods 5 and 6) 48.91 47.01 46.30 5-79 5-64 5- 60 6. 05 6.31 6.24 39.25 41.04 41.86 Difference (D— C) -f 2.61 + .19 - -19 Heavy rations^ Steer E (period 4) 46.33 49.61 5. 24 5-29 5- 46 4.95 42.97 40. IS - 3-28 - -05 + -51 -f 2.82 Alfalfa hay and grain (experiment 208) : Average of medium and light ra- tions— Steer E (periods 2 and 3) Steer C (periods 2 and 3) 28. 00 30.83 4.98 4.96 7.67 8.78 59-35 55-43 Difference - 2.83 + .02 — I. II + 3-92 Mixed hay (experiment 211): Steer D (periods i, 4, and 5). .. . Steer G (periods i, 4, and s). .. . 44.53 43.28 5-29 5-05 7-30 7.41 42.88 44. 26 Difference + 1.25 + .24 — . II - 1.38 The three animals used in experiment 208 had been subjected to different previous treatment, steer D having received almost from birth a restricted quantity of feed, steer E a ration ample to support normal growth, and steer C as heavy feeding as practicable. In the periods on the medium and light rations of alfalfa hay these animals showed slight differences in their digestive powers in the order named, the average Mar. 25, 1915 Energy Values of Feeding Stuffs for Cattle 449 results for the two quantities being as shown in Table V. On the other hand, upon the heavy hay ration and also upon the mixed ration of hay and grain steer C was distinctly inferior to steer E, losing more chemical energy in both methane and feces. This was the most distinct individual difference in these experiments. It should perhaps be noted that steer C showed a tendency to bloat and to get off feed on heavy rations and possi- bly did not have full normal digestive power. In experiment 211 steer D, then a year older, on the average of three periods on mixed hay again showed a slight inferiority to another animal which presumably had received better feeding during growth. Clearly individual differences between the animals had no very mate- rial influence on the losses of chemical energy in these experiments. In most instances the differences are well within the limits of error for such determinations, and even in those cases where there seem to be distinct individual differences they are comparatively slight, being of about the same magnitude as those observed by G. Kiihn (28) and rather smaller than those found by Armsby (i) in experiments on three steers. VARIABILITY OF METABOLIZABLE ENERGY The results recorded in Tables III and IV show clearly that the metab- olizable energy of a feeding stuff is by no means a constant. Not only do the averages for feeding stuffs of the same name differ more or less, but the raetabolizable energy of the same sample is more or less variable in the different periods. The losses of chemical energy which a feeding stuff suffers are sub- stantially determined by the nature and extent of the digestive pro- cesses. Digestibility, however, especially in ruminants, is a very com- plex affair, depending on many factors. Broadly speaking, it may be characterized as a series of fermentations, effected in part by a variety of organized ferments and in part by enzyms secreted by the digestive organs or contained in the feed itself. Changes in the composition of the contents of the digestive tract or in the rapidity with which they move forward through it can hardly fail to influence in a ^'ariety of ways the course of these fermentations, and it seems on the whole rather surprising that they go forward as uniformly as they do. In these experiments they appear to have been affected chiefly by the variations in the amount of feed consumed. Recently Zuntz and his associates (20, 53) have reported striking instances in which the extent of the methane fermentation in particular has been markedly affected by the make-up of the rations and especially by the order in which the feeds were consumed, while Voltz and his associates (47, 48) have laid much stress on the practical importance of these results. No such marked differences occurred in our experiments, but, on the other hand, the range of feeds was not so wide. It is perhaps too early 450 Journal of Agricultural Research Vol. Ill, No. 6 to judge the full significance of Zuntz's results, but they should at least serve to correct the notion, more or less subconsciously held by not a few, of digestion as a perfectly definite process and of a digestion coeffi- cient as a sort of chemical constant. On the other hand, however, it is easy to overestimate the importance of these variations in the digestive process in their bearing upon estimates of the values of feeding stuffs. On the whole, they appear to be of far less significance than other factors to be considered later. ESTIMATION OF METABOUZABLE ENERGY FROM METABOLISM EXPERIMENTS The losses of chemical energy in feces and urine are readily determined by means of the ordinary metabolism experiment, but the determination of the losses in the combustible gases requires special and somewhat costly apparatus. A number of experimenters have therefore attempted to estimate the amounts of these gases produced, usually from the amounts of carbohydrates (crude fiber and nitrogen-free extract) digested, using, as a rule, the average factor derived from Kellner's investigations (26, p. 420) — viz, 4.2 parts of CH4 per 100 parts of digested carbohydrates. Our experiments have yielded somewhat higher figures, as shown in Table VI, giving the maximum, minimum, and average results for each feeding stuff or mixture. T.\BLE VI. — Quantity of methane per 100 gm. of digestible carbohydrates Feeding stuff. Num- ber of experi- ments. Quantity of methane. Ma.^i- nium. Mini- mum. Aver- age. Timothy hay Clover hay Mixed hay Alfalfa hay Maize stover Average Maize meal and clover hay Wheat bran and timothy hay Grain mixture No. i and timothy hay. . Grain mixture No. 2 and alfalfa hay ... Maize meal and mixed hay Hominy chop and mixed hay Average Average of author's experiments. Average of Kellner 's experiments 5 6 17 3 43 65 44 4-3 4.8 4.7 3-8 4.2 4-4 3-S 2.9 5- I 5-2 5-8 5-3 4.8 5- 2 5-2 5-3 5-5 4.7 5-0 5-5 5-5 4.6 4.6 5- I 4.8 4-7 4.8 4.8 4.9 5-0 4-S 4-S 4-7 4-7 4.8 4.2 Mar. as. is's Energy Values of Feeding Stuffs for Cattle 45 1 While there is a considerable range in the results of individual experi- ments, nevertheless an estimate of 4.5 grams of CH^ per 100 gm. of digested carbohydrates affords a fair basis for an approximate estimate of the losses of chemical energy in the combustible gases and for com- puting the metabolizable energy of a feeding stuff by means of the ordi- nary digestion experiment combined with the collection of the urine and a determination of the heats of combustion of the visible excreta. The additional labor required for this purpose is so small that it is to be hoped that in future digestion experiments it may be undertaken whenever possible. FROM DIGESTIBLE ORGANIC MATTER It is possible also to estimate the average metabolizable energy of a feeding stuff from its content of total digestible organic matter, as shown by the ordinary feeding tables. The differences shown in Tables III and IV between the metaboliz- able energy per kilogram of dry matter of the different feeding stuffs are due chiefly, as already pointed out, to differences in the proportion of the chemical energy carried off in the feces, so that the metaboliz- able energy per kilogram of digestible organic matter shows a striking degree of uniformity as between different coarse feeds, on one hand, and as between different concentrates, on the other, a fact quite in harmony with earlier results reported by Kellner (26). Expressing the results in therms per kilogram and using for the apparent metabolizable energy of Kellner and Kohler's feeding stuffs the figures computed by Armsby {2, p. 301), we obtain the follo\\'ing averages: Metabolizable energy per kilogram of digestible organic matter COARSE FEEDS Armsby and Fries: Therms. Timothy hay 3-49 Red clover hay 3-49 Mixed hay 3-39 Alfalfa hay and meal 3. 61 Maize stover 3. 43 Average 3. 48 Kellner and Kohler: Meadow hay 3-5° Oat straw 3. 74 Wheat straw 3- 31 Extracted straw 3. 64 Average 3-55 78745°— 15 2 452. Journal of Agricultural Research voi. in, no. 6 CONCENTRATES Armsby and Fries: Therms. Maize meal 3- 80 Wheat bran 3-99 Grain mixture No. 2 3-88 Average 3. 89 Grain mixture No. i 3.91 Hominy chop 4. 08 Average 4. 00 Kellner and Kohler: Beet molasses 3-47 Starch 3- 60 Wheat gluten 4. 79 Tangl (44), Tangl and Weiser (45), and Zaitschek (50) have also deter- mined the metabolizable energy of a number of feeding stuffs for cattle, the methane being estimated from the amount of digestible carbohydrates, with the following results, which are very similar to those just reported : Metabolizable energy per kilogram of digestible organic matter: Tangl's experiments Therms. Meadow hay 3-44 Ensiled meadow hay 3. 70 Hay from irrigated meadows 3. 60 Broom-corn millet meal 3. 68 Pumpkins 4- 29 On the other hand, four experiments upon a bull by Voltz et al. (48), in which the production of methane was likewise computed, gave notably higher figures, viz: Metabolizable energy per kilogram of digestible organic matter: Voltz's experiments Therms. Mixed ration (hay, straw, malt sprouts, dried brewers' grains, and potato flakes) . 3. 95 Dried distillery residue (from potatoes) 4. 84 Palm-nut meal 4- 85 Beet molasses 4. 36 The most important factor influencing the metabolizable energy of the digestible organic matter of concentrates seems to be the percentage of fat in the feeding stuff, as appears from a comparison of the data con- tained in Table I, while feeding stuffs exceptionally high in protein have also a high content of metabolizable energy in their digestible matter, as in the case of Kellner and Kohler's wheat gluten. There seems no ob%-ious explanation of the exceptionally high results obtained by Voltz, but it would seem that for the present, with the ordinary dry feeding stuffs or mixtures, the following factors may safely be made the basis for computing approximately the metabolizable energy of feeding stuffs for cattle when their content of digestible organic matter is known or can be estimated. Mar. 25. 1915 Energy Values of Feeding Stuffs for Cattle 453 Metabolizable energy of feeding stuffs per kilogram of digestible organic matter Therms. Coarse feeds 3. 5 Concentrated feeds: With less than 5 per cent of digestible fat 3. 9 With more than 5 per cent of digestible fat 4. o No similar results have been reported on succulent feeds, with the exception of Zaitschek's figure for pumpkins. II. EXPENDITURE OF ENERGY CONSEQUENT UPON FEED CONSUMP- TION AND ITS FACTORS That the consumption of feed tends to increase the metabolism of an animal has become a commonplace of physiology. The magnitude of the effect varies within rather wide limits according to the chemical and physical properties of the feed, while there is still more or less difference of opinion as to its causes. Zuntz and his associates have called it "work of digestion" and have attributed it largely to increased muscular and glandular activity of the digestive and excretory organs. Most physiological investigations in this field have been made on camivora or on man, in which the increase of the metabolism is not usually very large, except when much protein is consumed. More recent investiga- tions on these species appear to have shown that the mechanical work of the digestive organs is but a small factor and that the term "work of digestion" is not a fortunate one. With herbivora and especially with ruminants, however, the total effect on the metabolism is quantitatively very marked, while the mechanical factor is of much greater significance. This was early shown by Zuntz and Hagemann (52) in their investiga- tions upon the horse. With ruminants the most extensive investiga- tions are those made at the Mockern (Germany) Experiment Station by G. Klihn (29) and by Kellner (24, 25, 26, 27). These investigations, especially those of Kellner, were directed primarily to the determination of the relative values of nutrients and feeding stuffs, but from another point of view they constituted also determinations of the energy expendi- ture caused by the feed. The main purpose of the experiments here reported was to determine the proportion of the feed energy expended in the increased metabolism by means of direct measurements of the heat evolved, checked by determinations of the respiratory products, the relative values of the feeding stuffs being obtained by difference. DIFFERENCES IN MUSCULAR .\CTIVITY INFLUENCE OP STANDING OR LYING UPON METABOLISM We have repeatedly called attention to the very striking increase in the heat elimination of cattle when standing as compared with that when lying and have shown (11), in reply to the criticism of Zuntz, that it is 454 Journal of Agricultural Research voi. in. No. 6 accompanied by a substantially corresponding increase in the gaseous excretion. Since conclusions regarding the influence of feed consumption on the heat production must be drawn from comparisons of two or more periods on different amounts of feed, it is obvious that the periods must be made as nearly identical in other respects as practicable. It being impossible to control the standing and lying of the subject, it became necessary, therefore, to attempt a quantitative determination of the influence of standing upon the metabolism of the animal as measured by its heat production and on this basis to correct the results of each period to some uniform ratio of standing to lying. It was natural to suppose that the increment of the metabolism during standing was due to the work of supporting the body in an upright position, but it soon became evident that it was also to a very considerable extent a function of the feed con- sumption. The apparatus used in these experiments is a flow calorimeter, the temperature difference between the ingoing and the outcoming water being read every four minutes. Since the hydrothermal equivalent of the absorber system and the contained water is only 6 kg., it is possible to follow very closely the rate of elimination of heat by radiation and con- duction. The results found by the authors (ii) show that the heat car- ried off as latent heat of water vapor is substantially proportional to that eliminated by radiation and conduction. On this basis the average heat elimination per minute and for 24 hours during standing and lying has been computed for each 48-hour period (or, in experiments 174, 190, 200, and 207, from selected uniform intervals). The addition of the necessary correcdon for the gain or loss of matter by the body gives the average heat production when standing or Ipng, respectively, while the difference between the two represents the increment due to standing for 24 hours. The following example, taken from experiment 179, period i, in which 76.65 per cent of the total heat emitted was given off by radiation and conduction and 23.35 P^r ^^^^ ^^ latent heat of water vapor, and in which the correction for loss of matter by the body was - 1 27 Calories, may serve to illustrate the method. Method of computing heat production of cattle when standing or lying Standing. Lying. Time standing or lying minutes. . i, 767 i, 113 Heat emitted by radiation and conduction: Total Calories. . 12,652 5, 128 Per minute Calories. . 7. 160 4- 607 Total heat production computed per 24 hours: • Standing 7. 160X1, 440^0- 7665-127 = 13,324 Calories. Lying 4.607X1,440-^0.7665 — 127= 8,528 Calories. Difference 4. 796 Calories. Mar. : Energy V allies of Feeding Stuffs for Cattle 455 The differences thus found between the heat production of the animals per 24 hours when standing and when lying were as shown in Table VII, in which are included for comparison the total dry matter consumed and the average live weight of the animal. In order to facilitate a general comparison of the results of the several experiments, the influence of the differences in live weight has also been eliminated to a degree by comput- ing the data of Table V to a uniform body weight of 500 kg., \vith the results shown graphically in figure i . Fig. 2000 3000 -^ooo £000 eooo 7000 3000 acxfo /oooo Dffy M/ITT£ff COfi/SUMCO GffAMS -Graph showing the dry matter eaten and the increments of heat production due to standing; computed per 500 kg. live weight per 24 hours. It should be clearly understood that both Table VII and figure i show the relation of feed consumption to the heal increment during standing, computed in the manner just illustrated, and not to the total heat pro- duction. 456 Journal of Agricultural Research Vol. III. No. 6 Table VII. — Increments of heat production of cattle in standing Feeding stuff and experiment. Timothy hay; Experiment 174 — Steer I (period A) Steer I (period B) Steer I (period C) Steer I (period D) Steer I (periods D-A) . Experiment 190 — Steer A (period 3) . . . . Steer A (period 4) . . . . Steer B (period 3) . . . . Steer B (period 4) . . . . Experiment 200 — Steer A (period 3) . . . . vSteer A (period 4) . . . . Steer B (period 3) . . . . Steer B (period 4) . . . . Experiment 207 — Steer A (period 3) . . . . Steer A (period 4) . . . . Steer B (period 3). . . . Steer B (period 4) . . . . Red clover hay: Experiment 179 — Steer i (period i) Steer i (period 2) Experiment 186 (series a)- Steer i (period i) Steer i (period 3) Steer i (period 2) Steer i (periods 2-1). . Experiment 186 (series b)- Steer i (period 3) Steer i (period 2) Mixed hay; Experiment 211 — Steer D (period i). .. . Steer D (period 4). .. . Steer D (period 5). .. . Steer D (periods 1-5). Steer G (period i) . . . . Steer G (period 4) . . . . Steer G (period 5) . . . . Steer G (periods 1-5). Alfalfa hay: Experiment 208 — Steer D (period i). .. . Steer D (period 2). .. . Steer E (period 4) . . . . Steer E (period 5) . . . . Steer E (period 6) Steer E (periods 4-6) . Steer C (period 4) . . . . Steer C (period 5) . . . . Steer C (period 6) . . . . Steer C (periods 4-6) . Average live weight. Kilograms. 387 403 416 424 269 278 194 190 399 407 296 309 507 514 374 385 545 520 571 580 587 566 576 460 455 428 389 387 364 171 162 224 215 197 2S9 284 27s Dry matter eaten per head. Grams. 3-237 4,373 5, 480 6, 440 2, 001 3>493 1,774 2, 6io 2,647 4.425 2,470 3,805 2,974 4,892 2,798 4,630 4, 4S9 3, 144 2,933 4,139 5,025 4,139 5,025 6, 204 3,498 1,786 6, 092 3,149 1,608 2,155 1,300 4,170 2, 408 1,413 4,744 3,099 2, 119 Increment of heat pro- duction per 24 hours. Per head. Calories. 2,258 2,935 2,913 2,928 2, 206 2,076 1,774 1,899 2, 052 2,838 2,233 2,695 2,293 2,994 2,467 3,618 4,794 3,679 3,247 4, 018 4,438 3,080 3,452 3,51s 3,147 3, 107 3,764 4, 066 2,077 1,314 2, 145 I, 683 I, 5'2 2, 690 2, 659 2, 380 Per kilogram of increment of dry matter eaten. Calories. 1 598 } 16 210 148 442 346 364 628 848 637 474 570 420 \ - 136 24 92 102 764 196 894 264 170 230 18 286 Mar. as. 1915 Energy Values of Feeding Stuffs for Cattle 457 Table VII. — Increments of heat production of cattle in standing — Continued Feeding stuff and experiment. Kilograms. 321 308 292 349 354 337 349 349 329 34S 331 316 Alfalfa hay — Continued. Experiment 209 — Steer F (period 4) Steer F ^period 5) Steer F (period 6) Steer F (periods 4-6) Experiment 212 — Steer H ^period i) Steer H (period 3) Steer H (period 5) Steer H (periods 1-5) Alfalfa meal: Experiment 212 — Steer H (period 2) Steer H (period 4) Steer H (period 6) Steer H (periods 2-6) Maize stover: Experiment 210 — Steer D ^period i). .. Steer D (period 2 ) Steer D ^period 3) Steer D (periods 1-3) Maize meal added to clover hay: Experiment 179 — Steer I (period 2) Steer I (period 3) Steer I (period 4) Steer I (periods 4-2) Wheat bran added to timothy hay: Experiment igo — Steer A fperiod 3) Steer A (period i) Steer A (period 2) Steer A (periods 2-3) Steer B (period 3) Steer B (period i) Steer B (period 2) Steer B (periods 2-3) Grain mixture No. i added to timo- thy hay: Experiment 200 — Steer A (periods 3) Steer A (period i^ Steer A (period 2) Steer A (periods 2-3) Steer B (period 3^ Steer B (period i ) Steer B (period 2) Steer B (periods 2-3) o Basal ration of coarse lodder only Average live weight. 520 532 269 271 279 194 199 197 399 404 421 296 298 310 Dry matter eaten per head. Grams. 6,174 3.562 2, 226 6,638 5. 320 3.052 6,671 5,408 3,15s 4,335 3.548 2,563 "3.144 3,898 6,637 02,001 3,348 4,579 °i,774 2,803 3.582 1 2, 647 4,400 6,804 a 2, 470 3.628 4.583 Increment of heat pro- duction per 24 hours. Per head. Calories. 2,973 1,997 1,768 3,167 3,224 2,481 3.246 2,942 2,377 2,919 2,486 2,617 3,679 4,659 6,296 2, 206 2,995 2,639 1.774 1,741 2, 135 2, 052 2, 708 4,547 2,233 2,887 3.591 Per kilogram of increment of dry matter eaten. Calories. 372 172 304 -44 328 193 242 252 248 550 -134 170 1,300 598 748 586 -288 168 -32 540 374 765 600 56s 736 643 458 Journal of Agricultural Research Vol. Ill, No. 6 Table VII. — Increments of heat production of cattle in standing — Continued Feeding stuff and experiment. ©R.-ilN MIXTURE No. I ADDED TO TIMO- THY HAY — Continued. Experiment 207 — Steer A (period 3) Steer A (period i) Steer A (period 2) Steer A (periods 2-3) Steer B (period 3) Steer B (period i) Steer B (period 2) Steer B (periods 2-3) Alfalfa hay and gr-un mixture No. 2: Experiment 208 — Steer E (period i) Steer E (period 2) Steer E (period 3) Steer E (periods 1-3) Steer C (period 2) Steer C (period 3) Experiment 209 — Steer F (period i^ Steer F (period 2^ Steer F (period 3) Steer F (periods 1-3) Mixed hay axd maize meal: Experiment 211 — • Steer G (period 2J Steer G (period 3I Mixed hay and hominy chop: Experiment 211 — Steer D (period 2). Steer D (period 3). Average live weight. Kilograms. 5°7 499 5'^9 374 373 .^86 210 206 197 269 259 301 293 283 358 398 432 470 Dry matter eaten per head. Grains. "2,974 4>93i 7,708 a 2, 798 4,159 5,451 3,208 1,753 1,151 2,184 1,539 4,502 2,756 1,648 2,332 7,027 3,511 7.859 Increment of heat pro- duction per 24 hours. Per head. Calories. 2,293 2,704 5,lS2 2,467 3,165 4,659 1,944 1,682 1,718 2,083 2,154 2,442 1,834 2, 029 3,158 5,840 3,533 3,905 Per kilogram of increment of dry matter eaten. Calories, •) 210 I 892 610 1 513 } I, 158 828 180 -60 no -no ■} - 348 -88 144 572 86 a Basal ration of coarse fodder only. A simple inspection of Table VII suffices to show that the increment of heat production in standing can not be due to any large extent to the muscular work of supporting the body, since, in the light of Zuntz and Hagemann's (52) experiments on the horse, it must be assumed that this would be at least approximately proportional to the weight of the animal, while, in fact, in a large majority of cases the difference between the periods is very much greater than the corresponding difference in Uve weight. Even on the extreme assumption that in the periods on mini- mum rations the heat increment in standing was due exclusively to the increased muscular effort, the differences in live weight do not even remotely account for the greater increments in the other periods. Mar. as. 191S Energy Values of Feeding Stuffs for Cattle 459 RELATION TO AMOUNT OF FEED In spite of the considerable variations in the individual results and of some negative values,. it appears clear, both from Table VII and from fig- ure I, especially if the comparisons be made between the smallest and the greatest rations, that the effect of the feed in increasing the metaboUsm tended to be distinctly greater during standing than during lying, and that, on the whole, the differences tended to be greater with concentrates or with mixed rations than with coarse feeds. In other words, the differ- ence between the metabolism of the animal when standing and when lying was relatively greater on the heavier than on the lighter rations. This difference can hardly be ascribed to a greater direct stimulus of cell metabolism by the products of digestion. One explanation for it might be sought in the fact that the feed was consumed by the animal while standing. Experiments by Paechtner (39) and by Dahm (17) on cattle and by Ustjanzew (46) on sheep, in which the respiratory exchange was determined in short periods, showed that the mastication of i kg. of hay increased the metabolism of the animal by approximately 60 Calories. In our experiments this would be equivalent to the production during perhaps half an hour after feeding — i. e., at 6 a. m. and 6 p. m. — of from 12 to as much as 200 Calories of heat, or twice this amount in 24 hours, which amounts would be added to the standing metabolism. On the other hand, however, according to the same experimenters, the rumination of the feed would increase the heat production by, roughly, two-thirds as much, and this would constitute to a considerable extent an addition to the metabolism of the animal when lying, thus partially but not wholly compensating for the addition to the metabolism when standing consequent on mastication. It would appear, then, that the mastication of the heavier rations would tend to increase the ratio of heat production of the animal when standing to that when lying. The heat elimination, however, in our experiments showed no distinct evidence of such an increase. The rate of heat emis- sion per minute while the feed was being eaten showed infrequently a slight rise, which was seldom sharp and which was far less than would correspond to the presumable increase in the gaseous exchange. In many instances no effect upon the heat elimination was observed, but rather frequently there was a distinct fall. Sometimes, although not in the majority of cases, a rise was observed after the animal had finished eating. The animal was watered after the 6 a. m. feeding. Sometimes no perceptible change in the rate of heat emission resulted, but not infre- quently a fall was observed, which was occasionally considerable. It can hardly be doubted that there must have been an increased production of heat during mastication, but apparently this heat was not given off promptly. Part of it at least, it may be conjectured, was applied to warm 460 Journal of Agricultural Research voiiii. no. 6 the ingesta, especially the water, being eliminated only gradually during the succeeding 12 hours and in part during the periods of lying as well as of standing. It does not seem probable, therefore, that the heat pro- duced in the mastication of the feed was an important factor in causing the differences between the heat elimination of the animals when standing or lying which were observed in these experiments. Another plausible suggestion seems to be a tendency of the animals to greater restlessness and muscular activity in the standing position when consuming the heavier rations. We are not able to submit direct records to prove this, but indirect evidence is afforded by the fact that the animals as a rule (31 cases out of 40) changed from the lying to the standing position, and vice versa, more frequently on the heavier rations. No such distinct effect was noticeable on the percentage of time spent standing, it being greater in 22 cases and less m 19 on the heavier as compared with the lighter rations and showing a considerable degree of constancy in the individual animal. INDIVIDUAL DIFFERENCES In those cases where comparisons between different animals are possible the average results, especially those obtained by comparing the maximum and minimum rations, seem to indicate the existence of distinct individual differences between different animals in this respect. The most noticeable case of this sort is that of animals A and B, in experiments 190, 200, and 207, steer A being a typical beef animal, while B, although of mixed blood (scrub), was quite distinctly of the dairy type and of a more nervous temperament. In five cases out of six the heat increment due to standing was greater with steer B than with steer A, the average of all the results being 39 per cent higher for the former, as the following summary shows: Ht'at increifienis in Calorics per kilogram of dry matter due to standing Feed. Steer A. Steer B. f-88 148 Timothy hay I 442 346 I 364 628 Timothy hay and wheat bran 168 200 Timothy hay and grain mixture No. i l ,°° „'Jq Average 335 465 A similar instance is afforded in experiments 208 and 209 by the ani- mals C, E, and F. Steer C had received almost from birth as heavy feed- ing as practicable, while E and F had received the same feeds in quantities sufficient to insure normal growth but not to cause any material fattening. The heat increments per kilogram of dry matter of feed were as follows : Mar. 25. I9IS Energy Values of Feeding Stuffs for Cattle 461 Heal increme>its in Calorics per kilogram of dry matter due to standing Feed. Steer C. Steer E. Steer F. Alfalfa hay 118 230 304 Alfalfa hay and grain mixture No. 2 —no no 144 Average 4 170 224 INFLUENCE OP BITLK OP RATION It seems worth while to call attention also to three cases which suggest that the bulk of the ration may be a factor in determining the difference between the metabolism of the animals when standing and lying. In period 3 of experiments 208 and 209 steers C, E, and F received a light ration, two-thirds of which consisted of grain — i. e., a ration of small bulk — and on this light ration they showed a distinctly greater incre- ment of metabolism during standing than upon one considerably heavier. It seems at least a plausible suggestion that the deficient bulk may have caused a greater degree of restlessness in period 3 and consequently an increased metabolism. CORRECTION TO UNIFORM STANDING Since standing or lying exerts such a marked influence on the heat pro- duction of animals, it is evident that a correction for this influence must be made before the results of the heat determinations can be regarded as comparable, since, notwithstanding the uniformity of external conditions striven after, the proportion of time spent standing or lying, respectively, varied more or less. As already pointed out this does not seem to have been related to the quantity of feed consumed, the difference having been practically as often in one direction as in the other. The stimulating effect, if such there were, seems to have expressed itself in more frequent changes of posture and a greater intensity of metabolism while standing rather than in more prolonged standing. The percentage of time spent standing appears to be largely a matter of individuaUty, whatever that convenient term may really signify. If we compare the results in this respect in the several periods, irrespective of the amount and kind of feed, it appears that,, with a few exceptions, they show, on the whole, a rather marked degree of uniformity in the individual animal. This is especially true if experiment 190 be excepted, in which the animals were only about 12 months old. The data are contained in Table VIII, the averages in each case being computed, excepting the bracketed numbers. 462 Journal of Agricultural Research Vol III, Xo. 5 Table VIII. — Percentage of time spent standing Animal and experiment No. Steer I: 174. 179- Period No. 186. Average . 186. Average . Steer A: 190. Average. 207 . Steer B : igo. Average . 207 . Averse . A B C D I 2 3 4 la 2a 3a lb 2b 3b Percent- age of time spent standing. 57 68 57 60 61 63 56 [76] [87] 68 62 65 73 85 79 79 75 79 52 58 57 51 5° 64 46 49 53 I 65 2 46 3 58 4 28 I 47 2 45 3 57 4 33 I 48 2 46 3 43 4 39 46 Animal and experiment No. Steer C: 208. Average . Steer D: 208 f I I 2 I 210 . 2 3 I 2 3 4 I 5 Average Steer E; 208. Average . Steer F: 209 . Average . Steer G: Average . Steer H: Average . Period No. Percent- age of time spent standing. [60] 42 47 41 42 43 32 30 [S3] 38 36 3C 40 24 36 40 34 I 30 2 27 3 39 4 40 S 46 6 [56] 36 27 33 32 34 33 46 34 40 42 [28] 33 30 37 I 42 2 40 .3 I29I 4 38 5 34 0 37 38 Mar. 25. 191 5 Energy Values of Feeding Stuffs for Cattle 463 In view of this general uniformity, only a comparatively small cor- rection would be necessary in most instances to make the results on an individual animal comparable as regards standing and lying. When, however, it is desired to compare the results obtained with different individuals so as to get a general average, it is clear that the animal which tends to stand most is at a disadvantage and will show a lower net energy value for the same ration not because his feed is any poorer but because the animal is a less efficient converter and that the greater the stimulus to metabolism exerted by standing the greater will this difference become. It is necessary, therefore, to correct the results upon heat production as well as possible to a uniform proportion of standing and lying, so as to render the results applicable to an (assumed) average animal. According to Table VIII the average percentage of time spent standing was 43. In correcting the results, however, we have used 50 per cent as the standard for convenience in calculation — that is, we have taken as the corrected heat production the average of that standing and lying, computed as shown on page 454 for the entire 48 hours.' OTHER FORMS OF MUSCULAR ACTIVITY According to our interpretation of our results, the material increase in the heat production of an animal when standing as compared with that when lying is simply an instance of the well-known influence of muscular exer- tion upon metabolism. Recent investigators, notably Schlossmann and Murschhauser (42, 43) and Benedict and Talbot (14, 15) in experiments upon infants, and Benedict and Homans (12, 13) in similar trials with dogs, have emphasized the disturbing influence of this factor upon com- parisons of different periods. Benedict and his associates, in particular, have devised ingenious methods for determining the degree of muscular activity of a subject and have insisted that only periods of minimum activity can be safely compared. We have not yet had the courage to attempt to apply to an animal weighing 1,000 or 1,200 pounds methods like those which have been used so successfully for infants and small dogs, either for the indication of minor movements or for the determination of the pulse rate. It seems likely that the latter, in particular, might be of considerable aid in the interpretation of the results if it should prove possible to devise a form of apparatus which would not be injured by the movements of a heavy animal. 1 This method of correcting the rcsuUs assumes that the relative intensity of the metabolism of the ani- mal when standing or lying is not afTccted by the proportion of time spent standing. It might be imagined, however, that in a comparatively long period of standing the original stimulus to incidental muscular movements might gradually fade out. so that the average difference in the rate of metabolism in the two positions would be less for long than for short periods. In this case our method of correcting the results would be more or less erroneous. This possibility can be tested to a certain extent by comparing witli each other the two single days of each period. Out of the 64 possible comparisons, that oneof the two days in which the lesser percentage of time was spent standing showed a greater iucrenient of standing over Ildng in 26 cases and a less increment in 37 cases. On the average of the 64 days of maximum standing the percentage of time spent standing was 51.9 and oti the 64 days of minimum standing 43.5. while the corresponding average percentage increases in metabolism during standing were, respectively. 41.98 and 41.67. It does not appear, therefore, that within the range of thesocxperiments the ratio of the metabolism when standing to that when lying was materially affected by ttie proportion of time spent standing. 464 Journal of Agriciihural Research voi. iii, no. 6 In considering, however, the weight to be attached to the influence of variations in the muscular activity of the animal, and the degree of refinement necessary in experimental methods, it is important not to forget the main purpose of the experiments. This purpose was, as already explained, substantially an economic one — viz, to determine how much of the energy supplied in metabolizable form by the feed is, under ordinary conditions, dissipated through the heat production caused directly or indirectly by the ingestion of the feed and what proportion of it remains available for the physiological uses of the body. From this point of view it is immaterial whether the increased heat production is caused by "work of digestion" in the narrower sense, by the stimulating effect of the resorbed products of digestion upon the cell metabolism shown by the investigations of L,usk (31, 32, 33, 34, 35, 36, 49), or indi- rectly by giving rise to increased acti\'ity of the voluntary muscles. While it is of interest and value to learn as much as possible of the rela- tive importance of these factors, nevertheless they are "all in the day's work," and from the economic \'iewpoint their aggregate constitutes the increased energy expenditure consequent upon feed consumption. Even if it were practicable to base comparisons upon periods of minimum activity the results, however interesting physiologically, would include only a part of the effects which the feed actually exerts upon the metab- olism. If the feed causes greater restlessness in the animal while stand- ing or causes it to get up and to lie down more frequently, this gives rise, under the conditions of practice, to just as real losses of energy as does the increase of the general cell metabolism when in the lying position and from the economic point of view must be taken into account. What is needed is a comparison of periods of average rather than of minimum muscular actixaty and the correction to 12 hours of standing aims to reduce conditions to such an average (assumed) as regards this very important factor. Of course, however, the possibility of variations in other forms of muscular activity, arising from differences in external conditions other than the feed, has to be reckoned with. Naturally the endeavor has been to make those conditions as nearly uniform as possible. The feed- ing was identical from day to day during the three weeks of each period and was given at the same hours. The surroundings during the days spent outside the calorimeter were uniform, and the animals were handled by the same attendants. During the days in the calorimeter even greater uniformity of condi- tions existed. The temperature varied only a few hundredths of a degree, the triple walls of the apparatus practically shut off all external sounds except the slight monotonous click of the meter pump, while in the comparatively dim interior the change outside from daylight to arti- ficial light could not have been very noticeable. Visitors were not ad- mitted during the runs. As already stated, all the animals were docile Mar. 25, 191S Energy Valtces of Feeding Stuffs for Cattle 465 and accustomed to being handled, to the wearing of the apparatus for the collection of excreta, and to the presence of the observers. Some idea as to the extent to which these precautions were successful may be formed from a comparison of the quantities of heat produced after correction to 12 hours standing on the two successive days of each 48-hour calorimeter run. The first half of Table IX shows the corrected heat pro- duction on the first and second days of each run and likewise the mean, computed in the manner illustrated on page 454 for the entire 48 hours." Table IX. — Heat production per day and per head corrected to 12 hours standing Animal Period Corrected heat produc- tion. Analysis of heat production. Feeding stuff and experi- ment No. No. No. Stand- Rising CH, Re- First Second 4S-hour and fer- day. day. mean. ing la hours. lying down. menta- tion. main- der. Timothy hay: Cols. Cats. Cats. Cats. Cats. Cats. Cats. f I A 8,788 9.316 9.049 I, 112 17 448 7,472 I7J. I B 9,899 9.565 9.769 1,450 18 570 7.731 ■^ /*T 1 I C 10, 309 10, 723 10, 562 1,440 17 626 8,479 I D 11,013 11,130 11,149 1.437 27 864 8,821 ' A 3 5. 709 5. 532 5.636 1,091 12 282 4,251 mo A 4 ?7. 535 6.545 6,709 1,023 15 481 5.190 ■ly^ 1 B 3 4. 985 4.904 5.081 875 12 276 3.918 B 4 6,095 5.886 5.852 937 12 361 4.542 A 3 7,186 7.275 7.245 1,001 25 "378 5. 841 200 A B 4 3 8,057 6,773 8,205 8,186 1,379 1.093 40 24 f>6o6 6367 6, 161 6,707 6, 77» 5.294 B 4 7.750 7,783 7.837 I. 315 33 6540 5. 949 A 3 7.713 7,812 7,791 1,107 40 498 6,146 207 A 4 9,267 9,649 9.523 1,438 59 794 7.232 B 3 7.996 8,070 8,064 1, 190 43 481 6,350 . B 4 9, 621 9.637 9,812 1.758 51 747 7.256 Red clover hay : 170 {■: I 2 10, 939 9.327 10, gio'io, 926 2. 354 1,812 43 28 626 464 7.903 * / V 9.944 9.621 7.317 la 9.844 9.586 9.627 1,604 20 413 7.590 3a 9.835 10,618 10, 176 1.958 51 464 7,703 186 I 2a 10. 335 10, 945 10,574 2,183 36 555 7.800 3b 10, 178 10, 197 10, 206 1,492 48 464 8,202 . I 2b 10,274 10,976 10, 501 1,688 38 555 8,220 Mixed hay: f D I II, 775112,995 12,359 1,679 79 805 9I796 211 D 4 5 9.481 8,291 9,680 8,259 9.625 8,258 1,498 1.497 76 57 527 282 7.524 6,422 D G I 12, 211 11,971 12, 098 1,802 80 830 9,386 211 G 4 5 9.843 7,811 9,242 7.178 9,568 7,477 1,958 1,382 75 62 457 261 7,078 5,772 G a The 48-hour means differ from the means of the 24-hour periods for two reasons: First, the perccntaee increase of the metabolism when standing over that when lying varied in the two days, as did also the percentaKe of the total heat carried off as latent heat of water vapor. On account especially of the latter difference, the mcau of the two days taken siiifily differs from that computed from the average heat pro- duction per minute standing and lying for the whole 48 hours. Second, in experiments 190, 200. and 307 the corrected results for the 48 hours are computed from selected portions of the runs in the manner described in an earlier publication (10, p. 43), and fhenf ore differ from the mean of the results for the single days. ^ Methane computed from digested carbohydrates. 466 Journal of Agricultural Research Vol. Ill, No. 6 Table IX. — Heat production per day and per head corrected to 12 hours standing — Con. Animal Period Corrected heat produc- tion. Analysis of heat production. Feeding stuff and experi- ment No. No. No. First Second 48-hour Stand- Rising and CH( fer- Re- day. day. mean. ing 12 hours. lying down. menta- tion. main- der. Alfalfa hay: Cals. Cals. Cats. Cals. Cals. Cals. Cals. f '^ I 5,385 5,442 5,411 I, 000 39 256 4, n6 D 2 4,243 4,290 4,266 612 45 161 3,448 E 4 7,931 7,823 7,879 1,014 58 457 6,350 208 ■ s S 5,759 5,764 5,764 796 45 302 4,621 E 6 4,529 4,590 4,566 732 24 180 3,630 C 4 8,96s 9, 100 9,052 1,297 48 470 7,237 C 5 7,340 7,391 7,357 I, 292 38 380 5,647 c 6 6,136 6, 122 6, 124 I, 156 34 271 4,663 F 4 11,228 11,432 11,474 1,400 86 697 9,291 209 F 5 8,031 8,007 948 43 429 6,587 F 6 6,819 6,794 6,781 830 54 285 5, 612 H I 11,424 11,186 II, 272 1,522 62 837 8,851 212 H 3 10, 189 10, 616 10, 388 1,549 63 685 8,091 H 5 7,836 7,582 7,754 I, 196 45 415 6,098 Alfalfa meal : H 2 11,025 ", 505 11,252 1,570 53 760 8,869 212 H 4 9,889 10, 270 10, 066 1,425 46 638 7,957 H 6 7,027 7,132 7,069 1, 146 42 419 5,462 Maize stover: D I 9,550 9,203 9,363 1,428 31 495 7,409 210 D 2 8,605 8,404 8,495 1,199 44 405 6,847 D 3 7,527 7,429 7,476 1,250 59 308 5,859 Maize meal added to clover hay : f I 02 9,327 9,944 9, 621 1,812 28 464 7,317 179 I 3 4 10, 483 12,723 9,876 13,354 lo, 198 12,947 2,284 3,124 45 21 630 I, no 7, 2.39 8,689 1 I Wheat bran added to timothy hay: f A «3 5,709 5,532 5,636 I, 091 12 282 4,251 A I 7,819 7,328 7, 739 1,486 12 503 5,738 190 A B 2 «3 4,985 8,400 4,904 8,38s 5,081 1,307 875 13 12 643 276 6,422 3,918 B I 6,493 6,299 6,363 865 6 400 5,092 B 2 6,814 7,117 7,409 1,057 II 537 5-804 Grain mixture No. i added to timothy hay: f A «3 7,186 7,275 7, 245 I, 001 25 a 378 5,841 A I 8, 955 9,236 9,274 1,322 32 0679 7,241 200 A 2 12, 140 12, 667 12,514 2, 226 48 "1,043 9,197 B "3 6,773 6,707 6,778 1,093 24 0367 5, 294 B I 8,916 8,952 9,016 1,427 17 ° 556 7,016 B 2 9, Sio 9,710 9,909 1,771 25 a 684 7,429 A «3 7,713 7,812 7,791 I, 107 40 49S 6, 146 A I 9,786 lo, 181 10, 164 1,294 58 841 7,971 A 2 13, 534 12,738 13,375 2,518 73 1,222 9,562 207 B "3 7,996 8,070 8,064 1,190 43 481 6,350 / B I 9,434 9,568 9, 600 1,558 24 733 7,285 . B 2 11,658 II, 640 II, 720 2,285 45 954 8,436 <» Basal ration of coarse fodder only. Mar. 25, 191S Energy Values of Feeding Stuffs for Cattle 467 Table IX. — Heat production per day and per head corrected to 12 hours standing — Con. Feeding stuff and experi- ment No. Alfalfa hay and grain mixture No. 2; 208. 209. Mixed hay and maize meal: Mixed hay and hominy chop: Animal No. E E E C C F F F {e Period No. Corrected heat produc- tion. First day. Cats. 7.434 5.940 5. 032 6,802 6, 096 9.679 7.645 6,791 8,678 14. 510 9,782 14,877 Second day. Cats. 7.532 5.9" 5.138 6,927 6, 149 10, 099 7,809 6,659 8,362 14, 467 10, 126 15.040 48-hoiu' Cals. 7.483 5.924 5.084 6,858 6,123 9,888 7. "5 6,734 8,470 14. 561 9.947 14,936 Analysis of heat production. Stand- ing 12 hours. Cals. 916 788 824 I, oil 1,043 1. 132 844 948 1.505 2,852 1,695 1,850 Rising and lying down. Cals. 56 53 35 31 34 89 73 68 74 68 71 102 CHj fer- menta- tion. Re- main- der. Cats. 464 303 165 388 284 562 382 253 470 I, 126 64 I, 216 Cats. 6,047 4,780 4, 060 5.428 4,762 8,10s 5,816 5.465 6.421 10.515 7,538 11,768 As already stated, we estimate the experimental error in the deter- mination of the heat emitted by the animal to be approximately i per cent. In the 73 cases in which a comparison of the two days can be made, 40 show a deviation from the mean of the two 24-hour results of less than i.i per cent — i. e., the results for the two days practically agree within the limits of experimental error. Of the 53 experiments^ made since 1905 — i. e., experiments 200 to 212 — which, in our judg- ment, are, on the whole, more accurate than the earlier ones, 36 fall within this limit or error. On the other hand, however, deviations as great as 2 per cent are not uncommon, while occasionally they rise to as much as 5 per cent or even 7 per cent. The mean of the percentage deviations is for the entire series 1.45 per cent and for experiments 200 to 212, inclusive, 1.13 per cent. Moreover, the deviations of the single days from the mean are, on the whole, fully as great after reduction to 12 hours standing as before. It is clear, therefore, that despite the apparent uniformity of experimental conditions the metabolism of the animals was affected by influences other than the feed or the propor- tion of time spent standing. Mr. H. H. Mitchell, of the Illinois Experiment Station, has had the kindness to submit these data to mathematical study and finds that they present clear evidence of the existence of individual differences between the animals as regards the agreement between the two days. He writes as follows: It seems very evident to me lx)th from inspection of the data and from statistical calculations that tlie percentage deviation of duplicate determinations of the heat 78745°— 1; 468 Journal of Agricultural Research voi in. xo. 6 production of animals is variable, depending undoubtedly upon the particular animal under investigation and possibly, also, upon the nature of the ration. Thus the average percentage de\'iations for the four animals for which 10 or more obser- vations are recorded, with their probable errors, are: Animal 1 2- io±o. 23 Animal A , '^2. 25±o.35 Animal A 6 i. 77 ±0. 18 Animal B o. 7S±o. 13 Animal D i. 32 ±0. 28 While, according to these figures, animals I and A can not be differentiated from each other, they are both clearly differentiated from animal B, tlie differences, mth their probable errors, being i.35zto.26 between animals I and B and 1.02 ±0.22 between animals A and B. The difference between animals I and D was 0.78 ±0.36, the significance of which may be questioned. Using another metliod of compari- son, it may be shown tliat the odds are 124 to i that animals C and H are definitely distinct as regards the percentage deviation under discussion, I should not hesi- tate to conclude, therefore, that this percentage deviation is affected by the indi- viduality of the experimental animals. Furthermore, there is a slight suggestion, especially in the data of animals A and B, that the nature of the ration may aSect the percentage deviation of ^-our determinations. It is of interest to note in this connection that the results of Kellner's respiration experiments (uncorrected for standing or lying) likewise show variations of much the same order of magnitude between indi- vidual (not consecutive) days. When, therefore, comparisons are based upon the average results for 48 hours, it is impossible to assert that these results represent, as they should, periods of average muscular activity, although it would appear that the error thus introduced is usuallv not large. In Kellner's experiments it is still further reduced by the fact that in most cases the results of four or five single runs are averaged. ANALYSIS OF HE.\T PRODUCTION In Table VII were shown the increments of heat production per 24 hours in standing animals as compared with those lying. It is evident that of the total corrected heat production recorded in Table IX an amount equal to one-half of the corresponding increment shown by Table VII is to be regarded as the effect of t-he 12 hours' standing, while the remainder represents the metabolism of the animal per 24 hours lying. On the basis of Zuntz's recent results it is possible to carry this analysis of the heat production a little farther, at least approximately. The expenditure of energy caused by standing obviously includes that required for the muscular effort of rising and lying down. Von der Heide, Klein, and Zuntz (20, p. 823) estimate this on the basis of experiments by Klein at 9.7 Calories per 550 kg. of live weight for once rising and lying down again. The same investigators (20, p. 795) compute from Markoff's experiments (37, 38) that the methane fer- mentation in cattle gives rise to the evolution of 4.374 Calories of heat per cubic centimeter of methane, equivalent to 6.07 Calories per gram. While both the foregoing figures are confessedly but approximations, nevertheless they permit a partial analysis of the heat production with o Including questionable observation. t> Not including questionable observation. Mar. 25, 191S Energy Values of Feeding Stuffs for Cattle 469 the results shown in the second part of Table IX. The "remainder" shown in the last column includes the so-called basal or fasting meta- bolism, together with the effect of the feed in increasing the muscular activity of the organs of digestion and of the voluntary muscles in the lying position, as well as in directly stimulating the cell metabolism. No sufficient data are available for further analysis of this "remainder." PROPORTION OF FEED ENERGY EXPENDED IN HEAT PRODUCTION METHOD OF DETERMINATION The total metabolism of an animal upon any particular ration, as illustrated by the figures of Table IX, is made up of numerous factors, and a single experiment affords no means of determining the proportion due to the consumption of feed. This can be determined only by a comparison of two periods, otherwise identical, in which different quan- tities of the same feed are consinned, the additional heat production on the heavier ration constituting the measure of the additional energy expended. With camivora and with man the comparison may be made with the fasting state — i. e., the amount of feed in one of the periods may be zero. With cattle this is impracticable for obvious reasons, and it is necessary to make the comparison with a period upon a so-called basal ration. Kiihn and Kellner added the feeding stuffs to be tested to a mixed basal ration that was more than sufficient for maintenance. In our earlier tests, up to experiment 207, inclusive, the same general plan was followed, except that the basal ration consisted of coarse feeds only and was in most cases below the maintenance requirement. In the later experiments the method was modified by feeding different quan- tities of the same feed or mixture of feeds both above and below the maintenance requirement. The method of comparison for a ration con- sisting of a single feeding stuff is very simple. Thus, in experiment 207 the following results were obtained on timothy hay with steer A in periods 3 and 4. The same method of comparison may obviously be applied also to different amounts of a mixed ration of grain and hay, Computation of energy expenditure by steer A per kilogram of timothy hay eaten Quantity of dry matter eaten. Total heat produc- tion. Bistribution of heat production. Item. Standing. Rising and lying down. Fermen- tation. Remain- der. Gm. 4,892 2.974 Calories. 9.523 7.791 Calories. 1,438 1,107 Calories. 59 40 Calories. 794 498 Calories. 7.232 6,146 Period 3 1,918 1.732 903 331 173 19 9 296 154 1,086 Difference per kilogram 567 470 Journal of Agricultural Research Vol. in. No. 6 The comparison of these two periods shows that each additional kilo- gram of dry matter consumed increased the total heat production by 903 Calories, 173 of which represent energy expended in standing, 9 that expended in rising and lying down, and 154 the additional heat due to the methane fermentation, while the remainder, 567 Calories, represents the increased mechanical work of digestion plus any stimulus which the digested nutrients exerted upon the cell metabolism. Obviously the calculation by difference eliminates the basal metabolism. For concentrated feeds, which can not be fed alone, two methods have been used, as already noted. In the earlier experiments, the con- centrate was added to the basal ration of coarse fodder. Thus, in period 2 of the experiment just used as an illustration a mixture of grains (grain mixture No. 2) was added to the basal ration of period 3 with the follow- ing results: Computation of energy expenditure by steer A per kilogram of grain eaten Quantity of dry matter eaten. Total heat produc- tion. Distribution of heat production. Item. Hay. Grain. Standing. Rising and lying down. Fermen- tation. Remain- der. Period 2 Gm,. 2.949 2.974 Gm. 4,759 Calories. 13.375 7.791 Calories. 2,518 I, 107 Calories. 73 40 Calories. I, 222 498 Calories. 9.562 6, 146 Period 3 Difference Difference per kilogram of dry matter -25 4,759 5,584 I. 179 I, 41 1 298 2i 7 724 153 3,416 721 Each kilogram of dry matter of the grain increased the heat pro- duction by 1,179 Calories, which can be subdivided as before in the proportions shown.' The greater expenditure of energy per kilogram in the case of grain as compared with hay is seen to be due in part to a greater increase of the metabolism of the animal when standing and in part either to increased mechanical work in digestion or more likely to a greater stimulus of the cell metabolism. In later experiments (Nos. 208 to 212, inclusive), in place of adding grain to a ration of coarse fodder, the animals received varying quantities of a uniform mixture of coarse fodder and grain, the energy expenditure caused by the total ration being determined substantially in the manner already illustrated. The portion of the increase due to the grain alone was computed by subtracting from the total increase that due to the hay as determined in two or more separate periods on exclusive hay rations.^' The method may be illustrated by the results obtained with steer E in periods i and 3 of experiment 208. Logically the results of the comparison should be corrected for the slight difference (2s gm.) in the amount of dry matter of hay consumed. As a matter of fact, however, this correction is insigni6cant in all the experiments, araountinn in the present instance to about i Calorie. ' When more than two periods of hay fcedina were used the increased heat production per kilogram of hay was computed by comparing the periods on the heaviest and the lightest rations. Mar. 25, 191S Energy Values of Feeding Stuffs for Cattle 471 Compuiaiion of energy expenditure of steer E per kilogram of grain eaten Quantity of dry matter eaten. Total heat produc- tion. Distribution of heat production. Item. Hay. Grain. Standing. Rising and lying down. Fermen- tation. Remain- der. Gin. 1,086 387 Gm. 2, 122 764 Calories. 7.483 5.084 Calorics. 916 824 Calories. 56 35 Calorics. 464 165 Calorics. 6,047 4,060 Period 3 Difference Difference due to 699 gm . of hav "■ 699 1.358 2,399 840 92 71 21 9 299 70 1,987 690 Difference due to 1,358 gm. 1.559 I, 14S 21 IS 12 9 229 169 1,297 95S Difference per kilogram of a Computed from a comparison of periods 4 and 6. DIFFERENCES IN LIVE WEIGHT As the figures of Table VII show, the live weight of the animals varied more or less in the different periods. To what extent do these varia- tions affect the conclusions drawn from comparisons like those just illustrated ? Two effects other than those due directly to the amount of feed might be anticipated from an increase in the live weight: First, an increase in the basal metabolism due to a greater mass of tissue, and, second, an increase in the muscular work of supporting the body in the standing position. As regards the first of these, it is to be remarked that the experimental periods were short (three or, in a few cases, four weeks only) while the changes in the amount and kind of feed consumed were con- siderable. It seems altogether probable that the larger part of the variation in weight must be ascribed to "fill" — i. e., to variations in the contents of the digestive tract rather than to any considerable change in the make-up of the body proper — and that the actual basal metabolism was not greatly affected. As regards the effect upon the muscular work of standing, it has been already pointed out that this appears to be a relatively small factor in the total increase of heat elimination in stand- ing. In view of these considerations, it is to be anticipated that a cor- rection of the heat production in proportion to either the weight or the surface of the animals would materially exaggerate the effect upon the metabolism, and, on the whole, we have regarded it as safer to disregard the variations in live weight rather than to attempt a more or less con- jectural correction. 472 Journal of Agricultural Research Vol. Ill, No. 6 HEAT INCREMEXTS PER KILOGRAM OF DRY MATTER The results of the comparisons between periods made by the methods just illustrated are contained in Table X. In those cases in which more than two periods upon the same ration can be compared, the total heat increments per kilogram of dry matter are recorded for each successive ^ft/c ^-cj^ 300O ^ooo 5000 sooo 7000 SOOO 9txO JOOOO Dfy MATTER CONSUMED. Sff^lMS Fig. 3. — Graph showing the relation of heat production to dry matter consumed; computed per soo kg. live weight. pair of periods, beginning with the lighter ration. Thus, in the case of steer I in experiment 174 four different amounts of timothy hay ^ were eaten. Table X shows the total heat increments computed by comparing periods A with B; B with C; C with D; and finally A with D. The * About 350 gm. of linseed meal in each period were also fed. Mar. 25. 191S Energy Values of Feeding Stuffs for Cattle 473 final comparison, between the smallest and greatest rations, is regarded as the average result, and on it is based the computation of the distribu- tion of this energy between standing, rising, and lying down, fermenta- tion, and the "remainder," as well as, in the case of mixed rations, the computation of the energy increment due to the hay. The same results, computed per 500 kg. of live weight, are also represented graphically in figure 2, the abscissae representing the amount of dry matter consumed and the ordinates the corresponding corrected heat production, the heat production per kilogram of dry matter corresponding to the tangent of the angle between the graph and the horizontal axis. Table X. — Increments of heat production per kilogram of dry matter Animal Xo. ll Ah- a 2 1! J- Analysis of heat incre- ments. Feeding stuff and experiment No. Successive amoimts of feed. J3 a '■B in i 0 g .3 2 i ■a S 1 K 0 1 Timothy hay: 17 J. { i { t Cals. 634 Cats. 716 Cats. 612 Cals. 656 719 922 793 9°3 954 Cals. 102 -46 74 213 166 173 310 Cals. 3 2 0 8 7 9 4 Cals. 130 133 102 128 130 154 145 Cals. 421 630 100 746 180 200 490 567 207 . ... 495 782 992 453 141 412 277 S II 8 132 123 68 504 446 Red clover hay: 170 I I, series a 186 455 449 723 333 928 1. 031 344 220 41 94 10 — 10 5 4 96 103 118 137 273 186 I, series 6 { i Mixed hay: 799 1.357 I, 010 860 764 806 Average 1,078 935 q8o fi-i 5 - 7 13 5 8 5 123 III 100 76 104 118 78s 781 Alfalfa hay: D E [ c F H 1.339 I, 202 I, ii6 1,189 981 454 102 54 145 91 208 I, 204 I, 260 918 I, 161 1, 200 1,030 1.327 671 . 987 981 932 767 200 Average 1,147 1.05' 1, 165 169 S 102 889 474 Journal of Agricultural Research Vol. Ill, No. 6 Table X. — Increments oj heal production per kilogram of dry mailer — Continiied Feeding stuff and experiment No. Alfalfa meal : 212 Average of alfal- fa hay and meal Maize stover: 210 Alfalfa hay and grain mixture No. 2 : 2oS 209 Average . Alfalfa hay and grain mixture No. 2 (periods I and 2): 208 209 Average . Mixed hay and maize meal: 211... Mixed hay and hominy chop: 211 Maize meal added to clo- ver hay : 179 Maize meal (computed): 211 Hominy chop puted ) : 211 (com- Wheat bran added to timothy hay: 190. Average . Animal No. H D E C F E C F D I G D A B Successive amounts of feed. CaU 1-33° I, 184 1,034 1.395 886 1,146 765 I. 561 1,246 1,404 Cats, 939 1,028 Cats. I, 072 1,244'. I, 160 I, 004 1 7 343 934 0 2 SB •- « Cals I, igo I, 169 1,065 I, 166 1,139 I, 105 I- 139 1,072 1,248 I, 160 1,297 I, 147 952 1,434 1,365 I, 066 i,2S8 1,177 Analysis of heat incre- ments. Cals. 121 161 45 -50 64 169 127 287 36 375 386 30 84 lOI 93 13 0 Cals. 3 5 -16 Cals. 97 105 10 145 - 5 161 7 108 138 103 107 140 132 185 146 146 Cats. 969 902 875 966 1,033 926 977 871 971 921 871 972 393 906 1,180 140; 842 144 1,044 142 1 943 Mar. 25. 191S Energy Valiies of Feeding Stuffs for Cattle 475 Table X. — Increments of heat production per kilogram of dry matter — Continued Animal No. Successive amounts of feed. •a a u So ss P 0 Analysis of heat incre- ments. • Feeding sttaff and experiment No. 1 1 1 bo 1 ■0 a 5 == 1 i g B a 0 •a Grain mixture No. i added to timothy hay : r I { A B A B Cats. 1,157 1,933 1,213 I, 129 Cats. 1.348 935 1,156 1,640 Cats. Cols. 1,267 1,482 1,179 1,378 Cats. 294 298 413 Cals. 6 0 7 Cats. 160 150 153 178 Cals. 807 207 I, on 721 786 1,358 1,494 1,270 1,004 1,327 1,148 1,152 1,077 331 -99 27 4 9 -9 7 161 169 202 110 831 955 Grain mixture No. 2 (com- puted): 208. { E C F 1,058 933 200 759 1,277 1,127 1,141 1,125 1,004 1,277 -19 80 175 2 -3 9 160 116 i°3 982 811 Grain mixture No. 2 (computed from pe- riods i and 2): 208 E F 200 990 Average . . . . 1,141 128 3 no 900 CRITIC.\I, TEMPERATURE In order that comparisons like the foregoing shall be valid, the experi- ments must, of course, be made above the so-called ''critical tempera- ture" for the animal experimented with and for the minimum quantity of feed consumed, since below this temperature part of the heat produced is utilized to maintain the body temperature and thus to reduce the amount of heat liberated by the katabolism of body substance (2, p. 347-359, 407-410). Our experiments have been made at about 17° to 18° C, and we have not attempted to determine the critical temperature for cattle, but the fact clearly shown in figure 2 that the heat production per kilogram of feed consumed showed no tendency to increase as the rations were made heavier leads us to believe that even on the lightest rations the temperature was safely above the point at which the so-called "chemical" regulation of body temperature begins. Kellner's experi- ments were made at somewhat lower temperatures, mostly between 12° and 15° C, but on heavier rations. 476 Journal of Agricultural Research voi. iii. no. 6 DISCREPANCIES IN RESULTS It is apparent from both Table X and figure 2 that the single results show a considerable range for the same or similar feeds not only with different animals but also, in some instances, between different periods with the same animal. For example, in experiment 209, on steer F with alfalfa hay an increase of the ration from 2,226 to 3,562 gm. of dry matter caused an increase in the (corrected) heat production at the rate of 963 Calories per kilogram, while a further increase to 6,174 gni- re- sulted in a relatively greater increase of the heat production — viz, i ,301 Calories per kilogram. In the instance just cited one might be inclined to interpret the differ- ence as an effect of the greater feed consumption. The next line in Table X, however, sho%vs an even greater difference in the opposite direction, while it is evident from figure 2 that the data as a whole show about as many differences in one direction as the other and, as pointed out in the previous paragraph, fail to give any distinct evidence of a greater relative increase of heat production on heavy as compared with light feed or on supermaintenance as compared with submaintenance rations, the averages tending, if anything, to be a trifle lower on the heavier rations. Unavoidable differences in the muscular activities of the animal, other than those connected with standing and lying, and in other conditions have also to be considered. As already pointed out, the existence of such differences, in spite of the uniformity of the controllable experimental conditions, is indicated by the occasionally considerable divergence of the heat production upon the two days of the calorimeter runs. It is not improbable, therefore, that they may be responsible, at least in part, for the observed discrepancies, so that it is obvious that the average re- sults must be accepted with some reser\'e. On the other hand, however, it must be remembered that these are calculations by difference and that in such a calculation the experimental errors tend to accumulate in the final result. Obviously the greater we make the difference in the factor whose effect is to be determined, the less will be the relative error of the final result.^ We believe, therefore, that the results obtained by a comparison of the extreme rations, as recorded in column 4 of Table X, are decidedly more trustworthy than those computed from the inter- mediate rations, and, notwithstanding the discrepancies just mentioned, are inclined to regard them as expressing the total effect of the feed in increasing the metaboUsm, when variations in the time of standing are eUminated, with a sufficient degree of accuracy to warrant general com- parisons of the average results. These average results, both as to the total heat increment and its factors, are summarized in Table XI. 1 Out of 13 cases in which the results appear abnormally high or low. there were 6 in which the difference in dry matter consumed was less than i kg., although there were 5 other cases in which, with a similar small difference in the dry matter consumed, apparently normal results were obtained. Mar. 23, 191S Energy Values of Feeding Stuffs for Cattle 477 Table XI. — Average increments of heat production per kilogram of dry matter Feeding stuff. Coarse fodders: Timothy hay Red clover hay: Average Experiment 179 Mixed hay Alfalfa hay Alfalfa meal • Average of alfalfa hay and meal Maize stover Mixed rations: Alfalfa hay and grain mixture No. 2 Average of all Average of periods i and 2 only Mixed hay and maize meal Mixed hay and hominy chop Concentrates: Maize meal added to clover hay. . . . Maize meal computed from mixed ration Hominy chop computed from mixed ration Wheat bran added to timothy hay. Grain mixture No. i added to timo- thy hay Grain mixture No. 2 computed from mixed ration The same from periods i and 2 only Total incre- ment. Cats. 782 723 992 980 1,165 I, 190 1,169 1.065 1,139 I, 160 1,297 I, 147 1,434 1,363 1,177 1,327 I, 125 I, 141 Analysis of heat increment. Standing 12 hours. Cats. 141 344 412 68 169 121 161 lOI 20 127 287 36 375 386 30 93 331 -19 128 Rising and lying down. Cats. 10 II 4 5 3 S -16 CHi fermenta- tion. Cats. 132 96 123 123 102 97 lOI 105 138 107 140 132 i8S 146 146 142 161 160 no Remain- der. Cats. 504 273 446 785 889 969 902 87s 977 921 871 972 393 906 1,180 943 831 9S2 900 comparison op coarse feeds and concentrates The average results recorded in Table XI for the total increase in metabolism resulting from the consumption of i kg. of dry matter of the several rations — i. e., for the so-called "work of digestion" in the widest sense — are far from being in accord with common conceptions. Unconsciously misled by an unfortunate terminology, we have been accustomed to think of the more coarse and woody feeds, like hay, straw, stover, etc., as requiring a greater expenditure of energy in their digestion and assimilation than the more concentrated and highly digestible grains, for example. It may be somewhat surprising, there- fore, to note the relatively small differences found in this respect between different classes of feeding stuffs, as shown by the averages of Table XI and by figure 2. For example, the expenditure of energy caused by maize meal in experiment 179 was almost as great as that caused by 478 Journal of Agricultural Research Vol. Ill, Xo. 6 the clover hay \nth which it was fed, while in experiment 21 1 it was apparently distinctly greater than that due to the mixed hay con- sumed with it. Grain mixture No. i decidedly exceeded timothv hay in this respect, and grain mixture No. 2 was nearly equal to alfalfa hay. As a matter of fact, however, these results are in general harmonv with those of other investigators, particularly Kellner. The senior writer (2, p. 492) pointed out some 12 years ago that the total expendi- ture of energy consequent upon feed consumption, as computed from Kellner's published experiments, is strikingly uniform for the several materials experimented upon with the exception of wheat gluten, the average results computed per kilogram of dry matter being quite of the same order as those here reported, viz: Average energy expenditure per kilogram of dry malter Calories. Meadow hay i, 254 Oat straw 1,014 Wheat straw i, 138 Extracted straw i, 160 Starch; Kiihn's experiments i, 508 Kellner's experiments — Moderate rations i, 248 Heavj' rations 903 Kellner's later experiments (24, ed. 6, p. 160-168) have not yet been published in full, so that it is not possible to make an exact computarion of the energy expenditure. In certain cases, however, the percentages of digestible nutrients are reported. If the corresponding amount of metabolizable energy be computed, using the factors given on page 453, and from this the amount of energ)' gained by the animal subtracted, the difference will represent approximately the energy spent in digestion, etc. The results of such computations are as follows: Energy expetiditure per kilogram of dry matter, computed from Kellner's experiments Calories. Peanut oil " i, 727 Wlieat gluten : Ktihn's experiments 2,558 Kellner's experiments 2,096 Beet molasses ■ 988 Feeding stuff. Cottonseed meal Peanut meal . . . Palm-nut meal. . Linseed meal. .. Barley straw . . . Clover hay "Grass hay". .. . Rowen Digestible Computed metaboliza- Cm. Calories. 647 2,58s 672 2,688 624 2,496 690 2,760 464 1,624 498 iw43 5^8 1,848 487 1.705 Gain ia energy by animal. Calories. 1,869 1,798 1.739 1,828 747 811 803 747 Energy expended in feed con- sumption. Calories. 719 890 757 932 877 932 1.045 958 The approximate results thus computed for the coarse fodders are comparable in a general way with ours upon similar feeds, although a One veri' high result was rejected. Mar. 25, 191S Energy Values of Feeding Stuffs for Cattle 479 somewhat lower than Kellner's direct results just cited. Those on the oil meals appear relatively lower than ours, although even then they are not much lower than those for the coarse feeds, but it ma}^ be ques- tioned whether the estimates of the metabolizable energy of these feeds are not too low. FACTORS OP INCREASED METABOLISM Even the very approximate and partial analysis of the total heat production which is attempted in the second part of Tables IX, X, and XI serv'es to show that the degree of uniformity noted in the preceding paragraph, so far from being surprising, was rather to be expected. The notion of a greater expenditure of energy on coarse feeds is based on the idea that this expenditure is largely for mechanical work. The analysis of the heat production attempted on preceding pages, however, even though only approximate, clearly shows that a considerable por- tion of the increase in heat production is due to other causes. Roughly, from 9 to 1 7 per cent of the increase is computed to have had its source in the methane fermentation, while from 3 per cent to as much as 30 or 40 per cent appears to have been due to increased muscular activity while standing. The " remainder" may be regarded as consisting of the mechanical work of digestion plus the stimulus which the feed exerted upon the general metabolism of the animal. How large the latter fac- tor is we have no means of determining, but apparently it is not incon- siderable. It would seem that the energy expended in peristalsis can not be widely different per kilogram for the different classes of feeding stuffs. On the other hand, the work of mastication and rumination has been shown to be distinctly greater for the coarse feeds. On the basis of Paechtner's (39) and of Dahm's (17) experiments on cattle it may be roughly estimated at 100 Calories per kilogram for hay. Zuntz and Hagemann (52) found the work of masticating oats by the horse to be 28 per cent of that required for hay. On this basis an expenditure by cattle of approximately 28 Calories per kilogram of concentrated feeds, may be estimated. If these amounts are subtracted from those shown in the last column of Table XI, the following approximate figures are obtained per kilogram of dry matter consumed for the work of peristalsis plus the food stimulus to the general metabolism: COARSE FEEDS Calories. Timothy hay 404 Clover hay (experiment 179). .. . 346 Mixed hay 6S5 Alfalfa hay 802 Maize stover 775 CONCENTRATES Calories. Maize meal 878 Hominy chop i, 152 Wluat bran 915 Grain mixture No. i 803 Grain mixture No. 2 872 Whether the expenditure of energy in peristalsis in cattle is as small as it appears from recent investigations to be in man and in the car- nivora it is impossible to say, but one can hardly avoid the impression that the considerable differences shown b>^ the foregoing figures, and 480 Journal of Agricultural Research voi. in. No. 6 especially the generally higher results for the concentrates, indicate that the direct stimulation of metabolism is a large factor. It appears, then, that while the mechanical work required for the dio-estion of concentrates is somewhat less than that necessary in case of coarse fodders, this difference is more than compensated for by other factors, so that on the whole fully as great an increase in the heat pro- duction is caused by the consumption of the concentrates. As a class, concentrates are superior to coarse fodders, not because their consump- tion involves a less expenditure of energy, but because they contain more metabolizable energy, so that more remains available for body use after that expenditure has been met. DIFFERENCES BETWEEN FEEDING STUFFS But while our results do not show the existence of as great differences between the two great classes of feeding stuffs in their effects on the energy expenditure of the body as seems to have been at times assumed, they nevertheless reveal distinct differences even between feeding stuffs of the same class. Thus, among the hays (if the results of experiment 179 for clover hay are accepted) a regular increase is found in the total energy expenditure from timothy hay with an average of 782 Calories through mixed hay and clover hay up to alfalfa with an average of 1,169 Calories. Apparently the legumes cause a distinctly greater increase in the metabolism than the Poaceae (Gramineae). In the case of red clover, the difference, according to the meager results obtained, appears to result chiefly from a stimulation of the metabolism due to standing. With alfalfa, on the contrary, the increase in the standing metabolism is not materially greater than in the case of timothy hay, while that due to fermentation is somewhat less. The chief difference between the two seems to He either in their effect upon the work of peristalsis or in the degree to which they stimulate the general metab- olism. One can hardly doubt that the latter is the chief cause and is naturally inclined to associate it with the higher percentage of protein in the legumes. That other causes may also be operative, however, is indicated by the result on maize stover, which is nearly as high as in the case of alfalfa and shows a similar distribution among the several factors. Among the concentrates there may be noted particularly the marked effect of maize in both the two not very satisfactory experiments in noticeably increasing the standing metabolism. This result is of special interest in view of Zuntz and Hagemann's observations (52, p. 259) on the stimulating effect of maize upon the metabolism of the horse, which were also made on the standing animal, although no increase in the minor muscular activity is reported. Grain mixture No. i, containing 43 per cent of maize meal, likewise showed a similar effect, although with grain mixture No. 2, containing 60 per cent of maize, it was much less marked, possibly on account of the lower content of protein (12.5 as compared with 17.5 per cent). The increases caused by wheat bran Mar. 25, 191S Energy Values of Feeding Stuffs for Cattle 481 axid by hominy chop, on the other hand, appear to have affected chiefly the metabolism of the animal when lying. INDmDU.'VL DIFFERENCES Attention was called on pages 460-461 to the existence of individual differences in the effect of the feed on the ratio of the standing to the lying metabolism. These differences seem in some instances to extend also to other factors of the total heat increment. While the single results are more or less variable, this fact seems to be brought out clearly in the averages. The most striking example is afforded by the animals A and B in experiments 190, 200, and 207, for which the following averages may be computed, showdng the heat increments per kilogram of feed to have been distinctly greater with steer B than with steer A. This is, of course, the converse of the conclusion recorded in an- earher publica- tion (10). Average heat increments of steers A and B per kilogram of dry matter Animal. No. Total in- crement. Distribution. Feeding stuff. Standing 12 hours. Rising and lying down. Methane fermenta- tion. Remain- der. Xiiiiotliy hay { S Calories. 717 890 I, 066 1,288 1,223 1,43° Calories. "3 183 84 lOI 296 367 Calories. 6 4 0 — I 7 I Calories. 139 126 140 144 156 164 Calories. 459 577 842 1,044 764 898 Timothy hay and wheat bran. . Timothy hay and grain mix- ture No. 1 . No such distinct differences were observed between the other animals, which, however, were all of similar type. While steers C, E, and F showed an increased effect upon the standing metabohsm in the order named the difference in the metabolism of the animals when lying shows on the average an approximately equivalent decrease, so that no material difference in the total effect resulted. SUMMARY Tables X and XI include the results of all of our experiments which have been so far computed as to permit their discussion. In seeking to derive from the recorded results for the increased energy expenditure consequent upon the consumption of certain feeding stuffs general averages which may, with the reservations made on previous pages, afford a basis for estimating the energy values of classes of feeding stuffs and of mixed rations, a certain degree of freedom of choice and the exercise of the judgment of the experimenters seems warranted. Of our results, those on clover hay in experiment 186 appear to us particularly ques- tionable. In one period the animal did not lie down during the entire 48 hours, while in two other periods the time spent in lying was nuich less than normal. Furthermore, there was a considerable difference 482 Journal of Agricultural Research Vol. Ill, No. 6 between the observed and the computed heat production in four out of six periods, although it is true that this difference was not relatively greater than in experiment 190. Whether these facts are in any degree responsible for what seem abnormally low results and for the very large proportion of the heat increment apparently due to stimulation of the standing metabolism, it is hard to say, but the results differ so widely from all the others that we feel justified in rejecting them, pending other experiments, particularly since the total increment observed in experi- ment 179 agrees very well with that computed on page 478 from Kellner's experiments. Experiment 212 fails to show any significant difference between alfalfa hay and alfalfa meal, and the two have been averaged together. In the case of maize meal it is difficult to decide which, if either, of the discordant results is worthy of most credit. The figure of only 393 Calories per kilogram for the increase of the metabolism of the animal when lying, however, seems so low that we are inclined to attach greater weight to the later experiment. For grain mixture No. 2 we have used the results computed from periods i and 2 in the belief that the heat production in period 3 was rendered abnormally high by the restlessness of the animals, owing to the small bulk of their ration (compare p. 461), although the difference is scarcely significant. In the Mockern experiments Kellner's results on heavy rations of starch appear to be abnonnal in that the methane production was not increased, while much starch escaped digestion. Kiiim's results were obtained on rations of coarse fodder and starch alone with a nutritive ratio of about I : 20, or even wider — i. e., under conditions seldom or never realized in practice. Kellner's average for medium rations, therefore, would appear to correspond most nearly to normal conditions. The results on peanut oil were irregular in several respects, but the rejection of the very high result with ox D seems justified. Of the computed results of Kell- ner's experiments, as given on page 478, those for the oil meals seem un- questionably too low and have been rejected. On the foregoing assumptions we have formulated the following averages for the total energy expenditure resulting from the consumption of I kg. of the dry matter of the feeds named. It may not be superfluous to call attention again to the fact that these figures are simply general averages, derived in some instances from quite discordant single results, and that, as both our own and the Mockern experiments show, they are subject to very considerable variations in individual cases. Average energy expenditure per kilogram oj dry matter eaten COARSE FODDERS Calories. Timothy hay "^ 782 Red clover hay 962 Mixed hay 980 Alfalfa hay i, 169 "Grass hay" i, 045 Rowen 958 Meadow hay i, 254 Maize stover i, 065 Barley straw 877 Oat straw i, 014 Wheat straw i, 138 Extracted straw i, 160 Clover hay 932 CONCENTRATES Calories. Maize meal i, 434 Hominy chop 1.36S UTieat bran i, 177 Grain mixture No. i i, 327 Grain mixture No. 2 i, 141 Beet molasses 988 Starch 1,248 Peanut oil i> 727 Wieat gluten 2, 294 Mar. 25. 191S Energy Values of Feeding Stuffs for Cattle 483 III. NET ENERGY VALUES AND THEIR COMPUTATION The method of estimating the nutritive values of the feeding stuffs consumed by farm animals which has been current for many years may from one point of \aew be characterized in a broad way as a chemical method. On the basis of the fundamental investigations of Henneberg and Stohmann (21, 22) in the early sixties, it sought to deteiTnine the amounts of protein, carbohydrates, and fat contained in feeding stuffs in a digestible form, assuming that the groups thus determined had the same physiological values in the nutrition of herbivora as had the corresponding substances in the food of man and carnivora. It is a well-recognized fact, however, that our information regarding both the qualitative and quantitative composition of feeding stuffs is even yet very meager. Moreover, our knowledge of the physiological functions of their ingredients is even more defective, so that, as Kellner (24, p. 15) points out, the advances in our knowledge of the chemistry of plants have not led to a corresponding increase in our knowledge of their nutritive values and have left the methods for the analysis of feeding stuffs largely untouched. Kellner appears to have been the first to attempt any practical appli- cation of the conception of the feed as a source of energy to the body. In 1880, in his investigations upon the relations between muscular acti\dty and metabolism in the horse (23), he determined the additional amount of work which the animal was able to perform as a result of the addition to his rations of starch and of fat. He expressed his results in terms of the percentage of the energy of the starch or fat which was recovered as useful work and called attention to the desirability of determinations of the heats of combustion of nutrients and feeding stuffs. Sixteen years later, after Rubner (40, 41) had published his fundamental work on the replacement values of nutrients and Zuntz and his associates (30, 54) had begun their investigations on the metabolism of the horse from the stand- point of energy, Kellner was able to return to the subject and undertake those extensive investigations with cattle (cited on previous pages) upon which he based his well-known method of comparing feeding stuffs on the basis of their so-called starch values. These are in reality energy values, and, so far as they are the results of direct determinations, they were obtained by substantially the same general experimental methods used in our own investigations, although direct determinations of the heat production were not included. VALUKS DIRECTLY DETERMINED The net energy value of a feeding stuff, as slated in the introductory paragraphs, is the energy which remains after deducting from its total chemical energy the two classes of losses which have been discussed in the first two sections of this article — viz, the losses of chemical energy 484 Journal of Agricultural Research Vol. Ill, No. 6 in the excreta and the increased heat production consequent upon the consumption of the feed. For example, the alfalfa hay consumed by steer E in experiment 208 contained per kilogram of dry matter 4,408 Calories of chemical energy. From the results reported in Tables III and X its net energy value, computed from the average results of periods 4, 5, and 6, is as follows: Net energy value of alfalfa hay per kilogram of dry matter Calories. Calories. Calories. Total chemical energy 4, 408 Losses of chemical energy : In feces 2, 062 In urine 243 In methane 266 Total 2, 571 Increased heat production i, 202 Total losses 3, 773 Net energy value 635 Computed in practically this way, by subtracting from the gross energy the average losses of chemical energy recorded in Table IV and the average energy expenditure consequent upon the consumption of the feeding stuff as given on page 482, the average net energy values of the feeding stuffs used in these experiments are as follows : Net energy values of feeding stuffs per kilogram of dry mailer Feeding stuff. Timothy hay Red clover hay .... Mixed hay Alfalfa hay " Maize stover Maize meal Wheat bran Grain mixture No. i Grain mixture No. 2 Hominy chop Gross energy. Calorics. 4,518 4,462 4.393 4,372 4,332 4,442 4, 532 4,685 4, 609 4, 7°9 Losses of chemical Calories. 2, 664 2,461 2,479 2,45.1 2,380 1,115 2, 021 I, 621 I, 620 I, 187 Energy- expended in feed consump- tion. Calories. 782 962 980 I, 169 1,065 1,434 1,177 1,327 I, 141 1,365 Net energy values. Calories. 1,072 1,039 934 752 887 1,893 1,334 1,737 1,848 2,157 a Includes alfalfa meal. Kellner's results when put into the same form are as follows: Mar. 35, 1915 Energy Values of Feeding Stuffs for Cattle 485 Net energy values of feeding stuffs per kilogram of dry mailer: Kelltur's restilts Feeding stuff. Gross energy. Losses of chemical energy'. Energi- expended in feed consunjp- tion. Net energy values. Meadow hay .... Oat straw Wheat straw . . . , Extracted straw. "Grass hay" " . . Rowen " Barley straw " . . Clover hay o , . . Starch Peanut oil Wheat gluten . . . Beet molasses . . . Calories. 4,433 4,436 4,444 4, 147 Calories. 2, 260 2,848 3,062 1,013 4, 152 9,457 5,579 3,743 I, lOI 4,165 1,974 945 Calorics. 1,254 1,014 1,138 1, 160 1,045 958 877 932 1,248 1,727 2, 096 Calories. 919 574 244 1,974 803 747 747 811 1,803 3,56s 1,509 1,810 a As estimated on page 478. Very striking is the relatively low value for alfalfa hay, due in part to somewhat large losses in the excreta but chiefly to its marked effect in stimulating the metabolism. It is needless to add that this loss does not affect its special value as a source of protein, but as a source of energy it appears to have been distinctly inferior to timothy hay or even to maize stover. APPLICATION OF RESULTS TO OTHER FEEDING STUFFS It is obviously impracticable to apply the laborious methods of respi- ration and calorimeter experiments to all the vast number of feeding stuffs now in use. It is necessary to select a few typical representatives of different groups and to endeavor to apply the results obtained as well as possible to other similar materials. This Kellner sought to do in his later and as yet unpublished experiments. In the practical appli- cation of his results, however, Kellner failed to free himself from the older point of view. Aside from what seems to us the unfortunate and unnecessary concession to established usage involved in expressing energy values in tenns of matter, he approached the whole problem, as was quite natural, along the lines of the prevailing chemical methods. Deter- mining first the net energy values of the simple nutrients, he applied these values to the digestible nutrients of feeding stuffs and found that in most cases the resulting energy values were materially higher than those obtained by direct experiments on animals. In the case of coarse fodders this deficit in the obser\'ed energy values was found to be approxi- mately proportional to the total content of crude fiber, and by subtracting from the computed energy value 1.36 Calories per gram of total crude fiber results were obtained corresponding fairly well to those directly observed. 486 Journal of Agricultural Research voi. iii, no. e For finer materials like chaff, presumably requiring a less expenditure for mastication, 0.70 Calorie per gram of total crude fiber is deducted. For green forage containing 16 per cent or more of crude fiber the same deduction is made as for dry forage and for that containing 4 per cent or less of cnide fiber, the same as for chaff, while between these limits a sliding scale is used (24, 1905, p. 593-594). For concentrates a factor (Wertigkeit) is estimated from the direct results on similar feeds by which the energy value computed from the digestible nutrients is multi- plied to obtain the actual value. The method of computation just outlined is not only somewhat com- plicated but is essentially based on the older view which regarded the feed in the light of a source of matter to the body. The digestible pro- tein, carbohydrates, and fat are still the basis of the calculation, although certain more or less empirical corrections are applied to their computed effects. The energy content of a feeding stuff, however, is just as definite a quantity as its content of protein, carbohydrates, or fats, and it is entirely possible to trace the distribution of that energy in the body quite independently of any knowledge of the chemical composition of the materials. Not only so, but we believe that in discussing energy values there are distinct advantages as regards simplicity, and perhaps also as regards accuracy, in cutting loose entirely from the conventional data regarding chemical composition and digestion coefficients, as has been done in reporting our experiments on preceding pages, and in dealing directjy with quantities of energy. In making this statement we would by no means be understood to stigmatize comparisons based on chemical methods as either valueless or superfluous. The problems of nutrition are too complex and tgo difficult for us to refuse any light that can be thrown on them by any method, and the energy relations touch only one phase of them. The point is that in whatever degree their energetic aspects can be separated from their chemical aspects, to that extent we possess two independent methods of approach to them. COMPUTATION OF NET ENERGY VALUES The computation from the results of metabolism experiments or from the data of ordinary feeding tables in the manner just indicated of the net energy value of a feeding stuff which has not been the subject of direct experimental investigation with the respiration apparatus or calorimeter may be made a comparatively simple matter. The net energy value is equal to the metabolizable energy minus the energy lost as heat. It was shown on pages 450-45 1 that the metabolizable energy may be deter- mined experimentally without special difficulty and with a good degree of accuracy bj' means of the ordinary metabolism experiment in which the energy of the feed, feces, and urine is directly determined and that Mar. 2s. 191S Energy Values of Feeding Stuffs for Cattle 487 of the methane estimated from the amount of carbohydrates digested. When this is not practicable, it was further shown that the metabolizable energy may be estimated from the total digestible organic matter by the use of the factors given on pages 451-453. In one or other of these ways it is not difficult to compute approximately the metabolizable energy of the more common feeding stuffs, while the subtraction from this of the average energy expenditure due to feed consumption will give the net energy value. To illustrate, E. W. Allen, ^ gives the following data for average alfalfa hay, oat straw, and wheat bran: Percentage of dry matter and digestible food ingredients of feeding stuffs .\lfalfa Oat Wheat hay. straw. bran. Total dry matter 91.6 90. 8 88. 5 Digestible: Protein 10.58 1.20 12.01 Carbohydrates 37-33 38-64 41-23 Fats 1.38 0.76 2.87 Total digestible 49- 29 40. 60 56. 11 The sum of the digestible protein, carbohydrates, and fat equals, of course, the total digestible organic matter, irrespective of its chemical composition. Each gram of digestible organic matter, according to the averages on pages 451-453, would contain 3.5 Calories of metabolizable energy in the coarse fodders and 3.9 Calories in the bran. The average losses of energy in heat production per kilogram of feed would be the amounts shown on page 482 reduced to the average water content of the feed, as follows: Alfalfa hay . . 1,169X0.916=1,071 Calories. Oat straw 1,014X0.908= 921 Calories. Wheat bran . 1,138X0.885=1,007 Calories. The computation of the net energy values is therefore as follows: Alfalfa hay (3.5 Calories X492. 9) — 1,071 Calories=654 Calories per kilogram=29.7 T. per 100 pounds. Oat straw (3.5 CaloriesX4o6.o)— 921 Calories=5oo Calories per kJlogram=22.7 T. per 100 pounds. UTieat bran (3.9 CaloriesXs6i.i) — 1,007 Calories=i,i8i Calories per kilogram=53.6T. per 100 pounds. The methods of comptitation just illustrated are perhaps open to the charge of being to a degree summary and empirical. The idea of basing such computations on the energy values of the single ingredients may be fundamentally more scientific, but unfortunately at present it is an impracticable ideal on account of our deficient knowledge of the chemistry of feeding stuffs and of the physiological values of their in- gredients. While investigation along both these lines is highly inii)ortant and desirable, yet for a long time to come the data on which to base the practice of stock feeding will have to be obtained by more direct even ' Allen, E. W.. The feeding of farm animals. U. S. Dept. Agr. Farmers' Bui. as (rev.), p. 8-9. 1901. 488 Journal of Agricultural Research voi in. no. e if less fundamental methods. Kellner's scheme recognizes this fact and his deduction for crude fiber and his factors for relative values (Wertig- keit) are at bottom simply a method of applying the aggregate net results on typical feeding stuffs to other materials. The method here proposed seeks to do exactly the same thing more directly and simply, relating the energy content and the necessary deductions to the total dry matter or total digestible matter of the feeding stuff, independently of its chemical composition. It is true that the data for so doing are somewhat meager, but, except as Kellner has utilized unpublished data in the fonnulation of his tables, they are just as abundant in the one case as in the other. It is greatly to be regretted that Kellner's results have not yet been published in full. When they become available they will doubtless greatly broaden the basis for such computations. SUMMARY There are reported the results of 76 experiments with the respiration calorimeter upon nine steers in which the balance of matter and of energy was determined. The losses of feed energy from the animal are of two classes: (i) Losses of unused chemical energy in the feces, urine, and methane; and (2) losses in the form of heat due to the increased metabolism consequent upon the ingestion of feed. (i) Losses of chemical energy. — The losses of energy in methane and urine were relatively greater on light than on moderately heavy rations. Neither the losses of energy in the feces nor the total losses showed a distinct relation to the amount of feed consumed. Individual differences between animals had no very material influence on the losses of chemical energy. The losses of energy in methane may be computed approximately from the amount of total carbohydrates digested. The metabolizable energy per kilogram of digested organic matter showed but slight variations within the same class of feeding stuffs. (2) Losses of heat consequent upon feed consumption. — The heat production is notably greater during standing than during lying, and the difference is greater on heavy than on light rations. The increment of heat production during standing is affected by the individuality of the animal and by the kind of feed consumed. An approximate partial analysis of the heat production of the animal into its principal factors is attempted. The average energy expenditure consequent upon the consumption of I kg. of dry matter is reported for 1 1 different feeding stuffs. The expenditure of energy arising from the consumption of the coarse feeds is not on the whole materially greater than in the case of the con- centrates. The increased muscular work of the digestive organs appears to be a relatively small factor of the increased heat production. A scrub steer showed a somewhat greater increment of metabolism consequent upon feed consumption than did a pure-bred beef animal. Mar. =5, 19. s Energy V allies of Feeding Stuffs for Cattle 489 (3) Net energy values. — A summary of the average net energy values obtained in these experiments for 11 different feeding stuflts is A simple method is outlined for computing net energy values, in the absence of direct determinations, from metabolism experiments or from the data of ordinary feeding tables. LITERATXXRE CITED (l) ARMSBY, H. p. . t- C-. r. I 09 1898. The maintenance ration of cattle. Penn. Agr. Exp. Sta. Bui. 42. i8» p., 18 pi. (2) (3) (4) (5) 1913. Food as body fuel. Penn. Agr. Exp. Sta. Bui. 126, p. 59-68. (v) and Fries, J. A. „ ^ . n a • 1903. The available energy of timothy hay. U. S. Dept. Agr. Bur. Anim. Indus. Bui. 51, 77 p., 2 pi. 1905. Energy values of red clover hay and maize meal. U. S. Dept. Agr. Bur. Anim. Indus. Bui. 74, 64 p. 1903. Principles of Animal Nutrition... 614 p. New York. 1904. The respiration calorimeter at the Penns>-lvania experiment station. In Exp. Sta. Rec, v. 15, no. 11, p. 1037-1050, fig. 11-12, pi. 4-7- 1908. Investigations in animal nutrition. U. S. Dept. Agr. Bur. Anim. Indus. 23d Ann. Rpt., 1906, p. 263-285, fig. 2-3, pi. 12-14. 1913. A comparison of the observed and computed heat production of cattle. In Jour. Amer. Chem. Soc, v. 35, no. 11, p. 1794-1800. (8) (9) (10) (11) 1908. The available energy of red clover hay. U. S. Dept. Agr. Bur. Anim. Indus. Bui. loi, 64 p. iqio. The influence of type and of age upon the utilization of feed by cattle. Penn. Agr. Exp. Sta. Ann. Rpt. 1909/10, p. 324-606, i hg., 20 pi. See also V. S. Dept. Agr. Bur. Anim. Indus. Bui. 128, 245 p., 17 ng., 3 pi. :9ii. 1913. The influence of standing or lying upon the metabolism of catUe. In Amer. Jour. Physiol., v. 31, no. 4, p. 245-253. ri2) Benedict, F. G., and Romans, John. . ,. -j iQii A respiration apparatus for the determmation of the carbon dioxide ' produced by small animals. In Amer. Jour. Physiol., v. 28, no. i, p. 29-48, illus. ^'^^ 1912. The metabolism of the hypophysectomized dog. In Jour. Med. Re- search, V. 25 (n. s. v. 20), no. 3, p. 409-5°'. 3 "£•■ ' P'- (14) and T.\LBOT, F. B. . ^ , ,• r a 1912. Somefundaraental principles in studying infant metabolism. /» Amer. Jour. Diseases of Children, v. 4, p. 129-136. ^'^^ 1914. The Gaseous Metabolism of Infants. 168 p., illus. Washington. (Car- negie Inst. Washington, Pub. 201.) (j6) BRA.MAN, W. W. ..... r t T)- I (^U_™ 1914. A study in drying urine for chemical analysis. In Jour. Biol. Chem., V. 19, no. I, p. 105-113. (,7) Dahm, j^^^'-jj^j^^j^^ jgg mechanischcn Teils der Verdauungsarbeit fiir den Stoffwechsel des Rindes. /« Biochem. Ztschr., Bd. 28, Heft 5/6, p. 456-50^, 2 fig. _, , fi8) Fingerling, G., KoHLER, A., Reinhardt, FR.ctal. ,, j , IQ14 Untersuchungcii iibcr den Stoff- und Energieumsatz wachsender Schweine. In I.andw, Vers. Stat., Bd. 84, Heft 3/4, p- 149-230- 490 Journal of Agricultural Research voi. iii.no. e (19) Fries, J. A. [1912.] The combustible gases excreted by cattle. In Orig. Commun. Sth Intemat. Cong. Appl. Chem., v. 15, p. 109-119. (20) Heide, R. von der. Klein, , and Zuntz, N. 1913. Respirations- und Stoffwechselversuche am Rinde iiber den Nahrwert der Kartoffelschlempe und ihrer Ausgangsmaterialien. In Landw. Jahrb., Bd. 44, Heft 5, p. 765-844, pi. 2-5. (21) HennebErg, T- W. J. 1870. Neue Beitrage zur Begriindung einer rationellen Fiitterung der Wieder- kauer ... 457 p., i pi. Gottingen. (22) and Stohmann, F. 1860-64. Beitrage zur Begriindung einer rationellen Fiitterung der Wieder- kauer ... Hefte 1-2. Braunschweig (23) Kellner, Oskar. 1880. Untersuchungen iiber den Zusamraenhang zwischen Muskelthatigkeit und Stoffzerfall im thierischen Organismus. In Landw. Jahrb., Bd. 9, p. 651-688. (24) 1905. Die Emahrung der landwirtschaftlichen Nutztiere ... 594 p. Berlin. 1912. Aufl. 6, 640 p., front. Berlin. (25) and Kohler, A. 1898. Untersuchungen iiber den Nahrungs- und Energie-Bedarf voUjahriger gemasteter Ochsen. In Landw. Vers. Stat., Bd. 50, Heft 3/4. p. 245- 296. (26) 1900. Untersuchungen iiber den Stoff- und Energie-Umsatz des erwachsenen Rindes bei Erhaltungs- und Produktionsfutter. In Landw. Vers. Stat., Bd. 53, 474 p. (27) and others. 1S96. Untersuchungen iiber den Stoff- und Energie-Umsatz voUjahriger Ochsen bei Erhaltungsfutter. In Landw. Vers. Stat., Bd. 47, Heft 4/5, p. 275-331. (28) KuHN, GusTAV, Gerver, Franz, et al. 1883. Versuche iiber die Verdaulichkeit der Weizenkleie und deren Verander- ung durch verschiedene Arten der Zubereitung und Verabreichung sowie iiber die \'erdaulichkeit des Wiesenheu's im trockenen und angefeuchteten zustande. In Landw. Vers. Stat., Bd. 29, p. 1-214. (29) Thom.\s, A., and others. 1894. Fiitterungs- und Respirations-Versuche mit volljahrigeu Ochsen. In Landw. Vers. Stat.. Bd. 44, p. 257-581, 2 fig., i pi. (30) Lehman.n", Franz, Hagemann, O., and Zuntz, N. 1894. Zur Kenntniss des Stoffwechsels beim Herd. In Landw. Jahrb., Bd. 23, p. 125-165. (31) LusK, Graham. 1913. The cause of the specific dynamic action of protein. In Arch. Intemat. Med., v. 12, no. 5, p. 485-487. (32) (33) 1 913. The influence of foodstuffs and their cleavage products upon heat pro- duction. In Trans. Intemat. Cong. Hyg. and Demog., 1912, v. 2, pt. 2, p. 400-406. Discussion, p. 406-409. 1914. The specific djTiamic action of the foodstuffs. In Jour. Amer. Med. Assoc, V. 63, no. 10, p. 824-827. (34) and Riche, J. A. 1912. The influence of mixtures of foodstuffs upon metabolism. In Jour. Biol. Chem., v. 13, no. 2, p. 185-207. (35) (36) 1912. The influence of tlie ingestion of amino-acids upon metabolism. In Jour. Biol. Chem., v. 13, no. 2, p. 155-183. IQ12. Metabolism after the ingestion of dextrose and fat, including the behavior of water, urea and sodium chloride solutions. In Jour. Biol. Chem., V. 13, no. I, p. 27-47. (37) M.A.RKOFP, J. 1911. Untersuchungen iiber die Giirungsprozesse bei der Verdauung der Wiederkiiuer. In Biochem. Ztschr., Bd. 34, Heft 3/4, p. 211-232, 2 Mar. 2s. i9's Energy Valties of Feeding Stuffs for Cattle 491 (3S) Markoff, J. ,. • J 1- 1913. Fortgesetzte Untersuchungen iiber die Garungsprozesse bei der Ver- dauung der Wiederkauer und des Schweines. In Biochem. Ztschr., Bd. 57, Heft 1/2, p. 1-69. (30) PaEchtnER, T- . „ , -.. ,. 1909 Respiratorische Stoffwechselforschung und ihre Bedeutung fur die Nutztierhaltung und Tierheilkunde mit einem Beitrag zur Kenntnis vom Lungengaswechsel des Rindes. 64 p., 4 pi. Berlin. (40) RuBNER, Max. • 188? Die Vertrelungswerthe der hauptsachhchsten orgamschen JMahrungs- stoffe im Thierkorper. In Ztschr. Biol., Bd. 19 (n. F. Bd. i), P- 313-396- 1886. Bestimmung isod;>'namer Mengen von Eiweiss und Fett. In Ztschr. Biol. Bd. 22 (n. F. Bd. 4), p. 4o-55- (42) Schlossmann, Arthur, and Murschhai-ser, Hans. .. tt . iQio Der Grundumsatz und Nahrungsbedart des Sauglmgs gemass Unter- suchungen des Gasstoffwechsels. In Biochem. Ztschr., Bd. 26, Heft 1/2, p. 14-40. inii tjber den Einfluss massiger Temperaturschwankungen der umgebenden Ltift auf den Respiratorischen Stoffwechsel des Sauglmgs. In Bio- chem. Ztschr., Bd. 37, Heft ih, p. 1-22. (44) Tangi., Franz. .„„ ,,, ■ , ■ j i^ igos Beitrage zur Futtermittellehre und Stoffwechselphysiologie der land- wirtschaftlichen Nutztiere. In Landw. Jahrb., Bd. 34, Hett i, p. 1-92. (4O and Weiser, Stephan. . r t j t 1, k 1906. Zur Kenntnis des Nahrwertes emiger Heuarten. In Landw. Jahrb., Bd. 35, Heft 1/2, p. 159-223- STjANZEtt , ^^^ j^^j^gj^ Aquivalente der Yerdauungsarbeit bei den Wieder- kauem (Schafe). In Biochem. Ztschr., Bd. 37, Heft 5/6, p. 457-470- fd?) VoLTz WiLHELM, Paechtner, JOHANNES, and Baudrexei,, August. 101-, tJber den Futterwert der Kartoffelschlempe, ihres Ausgangsmaterials und iiber sog. spezifische Wirkungen der FutterstoSe. In Landw. Jahrb., Bd. 44, Heft 5, p. 6S5-764. ion Vergleichcnde Untersuchungen iiber den Nahrstoffbedarf bei der Mast des Rindes und des Schafes im spateren Verlauf des Wachstums . . . In Landw. Jahrb., Bd. 45. Heft 3, p. 325-437. i %•• P^- 4-"- (49) Williams, H. B., Riche, J. A., and Lusk Graham. ,-,,„» „„=,„ 1912 Metabolism of the dog following tlie ingestion of meat m large quan- tity. In Jour. Biol. Chem., v. 12, no. 3, p. 349-376- fi;o') Zaitschek, Arthur. ,„..,. , t 1 1906. iiber die Zusammensetzung und den Nahrwert des Kurbis. /k Landw. Jahrb., Bd. 36, Heft 1/2, p. 245-258. (51) ZunTZ, N^_..^^ ^^ j^^ Abhandlung "Zum Studium der Respiration und des Stoffwechscls der Wiederkauer." In Landw. \ ers. Stat., Bd. 83, Heft 3/4, p. 283-284. (1:2) ZuNTz, N., and Hagemann, Oscar. , , j T>t . u ■ -o u^ ..-nH ^^ 1898. Untersuchungen iiber den Stoffwechsel des Pferdes bei Ruhe und Arbeit. In Landw. Jahrb., Bd. 27, Erganzungsbd. 3, 438 P-. i fig--7Pl- ,, .. i^-,-) , Heide, R. von der, Klein, . et al. , , , ,,- j " iqn Zum Studium der Respiration und des Stoffwechsels der W.eder- kauer. In Landw. Vers. Stat., Bd. 79/80, p. 781-814, fig. 11. ^''^ ^7slo UnteTs;!chunTen"^ber den Stoffwechsel des Pferdes bei Ruhe und Arbeit. In Landw. Jahrb., Bd. 18, p. 1-156, pl. 1-3- AIR AND WIND DISSEMINATION OF ASCOPORES OF THE CHESTNUT-BLIGHT FUNGUS By F. D. Heald, M. VV. Gardner, and R. A. Studhalter, Agents, Investigations in Forest Pathology, Bureau of Plant Industry^ HISTORICAL INTRODUCTION Wind dissemination of the chestnut-blight fungus {Endothia parasitica (Murr.) And.) was first suggested by Murrill (13) ^ in 1906, although he apparently had only the pycnospores in mind", as is shown by the follow- ing quotation: Later the fruiting pustules push up through the lenticels and give the bark a rough, warty appearance; and from these numerous yeIlowish-bro\vn pustules millions of minute summer spores emerge from day to day in elongated reddish-brown masses to be disseminated by the wind and other agencies, such as insects, birds, squirrels, etc. A few years later, in a discussion of the means of spreading the disease, Hodson (9) says: Wind is probably the principal agency, but the spores are no doubt carried by ani- mals, birds, insects, and by shipment of infected material. He also cited some observations to substantiate the wind-dissemination theory, but it was not brought out clearly whether he had in mind the ascopores or the pycnospores only. A similar opinion is expressed by Mickleborough (12) a little later. After speaking of both the ascopores and the conidal, or summer, spores, he states: The minute spores are carried by the wind, on the feathers of birds, and the fur of squirrels. Referring to the spore horns, Mickleborough writes: These threads are dissolved and washed away by the rain and the spores are blown about by the wind. ° There are two possible ways in which pycnospores might be dissem- inated by the wind: First, by the direct transport of spore horns or small fragments of these structures; second, by the transport of dust particles bearing spores previously washed down by rains. Fulton (4) reports experiments which indicate that the former method of transport of pycnospores is of little importance in the spread of the disease. He concludes his discussion of this topic with the following statement : It seems likely the detachment was largely of small bits of the tendrils made up of large numbers of spores, and that these are too heavy to be carried great distances; ' The writers received valuable assistance in this work from Mr. R. C. Walton, also an agent. Investi- gations in Forest Pathology. 3 Reference is made by number to "Literature cited." p. 525-526. Journal of Agricultural Research, Vol. III. No. 6 Dept. of Agriculture, Washington. D. C. Mar. 25. 1915 G— 41 (493) 494 Journal of Agricultural Research voi. iii.no. e and suggests that under natural conditions infection may be spread short distances by the wind. The second possibility is brought out by j\letcalf and Collins (ii), as may be noted in the following quotation: As both kinds of spores appear to be sticky, there is no evidence that they are trans- mitted by wind except where they may be washed down into the dust and so blown about with the dust. While it has not yet been demonstrated that pycnospores are carried in this way, the tests of Heald and Gardner (7) on the longevity of pycno- spores in soil give added plausibility to the theory, since these spores were found to persist in the soil between periods of rain and were able to withstand complete desiccation in the laboratory for months. Attention was first directed to the strong probabihty of wind dissem- ination of ascopores by Rankin (14), who reported their forcible ex- pulsion. In a later report the same writer (15) makes the following statement : Under moist conditions the ascospores are shot forcibly out in tlie air where they can be caught up by the wind and carried for a considerable distance. The speaker found ascopores being shot from mature pustules during every rainy period last sum- mer. * * * The question at once arises, Wliy could not these ascospores once shot into the air be carried long distances and, owing to their abundance, cause a large majority of the infection? After carrying out field experiments during the summer of 191 2, Rankin (16), referring to ascospores, says: They are shot out in vast numbers with every rain during the summer and are carried by the mnd. Detailed field work on dissemination was carried out by Anderson (i) and his assistants for the Pennsylvania Chestnut Tree Blight Commission (2). These publications confirm the statement of Rankin that expulsion of spores takes place only when the pustules are moist. The seasonal duration of shooting under natural conditions was not determined, as the field tests were confined to the month of August. Under artificial con- ditions in the laboratory, the time required for moistened bark bearing perithecia to begin the expulsion of spores was determined, the shortest time recorded being three minutes. The duration of the shooting period following a rain was determined by artificial tests in either the field or laboratory, performed by soaking the specimens or drenching cankers with water. The maximum dura- tion recorded was five hours and two miniutes. While these tests under artificial conditions gave suggestive results, they were not necessarily a reliable indication of what would happen under natural conditions. It was also determined that bark kept constantly moistened continued to expel spores for a maximum period of 25 days, and the point was emphasized that no continuous rainy weather would be longer. The fact that ascospores expelled during a rain would be washed down to Mar. 2s. 1915 Dissemination of Chestnut-Blight Fungus 495 the ground without being carried any appreciable distance is not men- tioned. Since they germinate at once in rain water, the great bulk of such spores would be lost for anything but ver}' local infections. The really important point would appear to be the length of time shooting continues after a rain ceases, for at that time the conditions of the atmos- phere would be such as to favor a \\'ider dissemination. This question does not seem to have been satisfactorily answered. The data given on height and horizontal distance of projection, as well as the rate of expul- sion, certainly indicate the importance of wind transport of spores following rainy periods. The spore content of the air was studied b}' means of aspirator tests and exposure plates. In this work, carried out during drj' weather, Anderson and his assistants failed to get positive results under natural conditions in the field. They report the use of over 100 exposure plates and tests of 500 liters of air without finding a single spore of the chestnut- blight fungus. Tests made of aspirated air and by exposure plates gave positive results, however, when the cankers were artificially drenched ■with water. For the aspirator tests the horizontal distances of the aspirator opening from the canker varied from 2 inches to 5 feet (?) and the maximum vertical distance was 22 feet. The tests made by exposing agar plates under artificial conditions in the field again pointed to the probability of wind dissemination, but one is forced to admit that they were not conclusive, since the conditions were so different from the natural in that the cankers were drenched with water artificially instead of waiting for a rain. The results with exposure plates may be summed up as follows: No spores of the chestnut- blight fungus were obtained under natural conditions in the field during dry weather; by the use of artificially drenched cankers spores were obtained at distances varying from i inch to 51 feet, with very few at the maximum distance. The final and most conclusive argument in favor of wind dissemination in the minds of the authors cited was afforded by inoculations made by offering an opportunity for wind-borne spores to be introduced into wounds. There is little doubt in the minds of the writers of this paper that infection did take place in the way claimed, but it should be pointed out that a covering of cotton would not prevent spores from being washed into the wounds by rains (6). A fairly compact mass of cotton has been shown to retain but few of the pycnospores present in water passing through it. It must therefore be admitted that, under the conditions of the experiments reported, infection by spores washed down by rains was one of the possibilities. It is interesting to note in this connection lliat Kittredge (10), as a result of field observations on the spread of the disease around a center of infection, arrives at the following conclusion: The location of infected trees in partially infected groups of sprouts shows that wind is not the prime factor in the distribution of the spores. 496 Journal of Agricultural Research voi. m. No. 6 The author admitted, however, that the observations reported were rather meager in support of this conclusion. PURPOSE AND SCOPE OF PRESENT WORK Since most of the previous work on wind dissemination of the chest- nut-blight fungus which yielded positive results was done under artificial conditions, it was the aim of the present writers to studj' the problem under absolutely natural conditions. Briefly stated, the purpose of these tests was to determine whether or not, and if so, to what extent, wind ' acts as an agent in dissemination of the spores of this fungus. It was also the object of the work herein recorded to ascertain at what particu- lar times under natural conditions spores of Endothia parasitica are preva- lent in the air, the possible distances transported by the wind, and the kind of spores (whether ascospores or pycnospores). The locality chosen in which to conduct our tests was a 4-acre plot of native chestnut {Castanea dcntata) coppice near West Chester, Pa. The trees in this plot ranged from 4 to 8 inches in diameter and all were badly infected with the chestnut blight, many having already succumbed. In these tests, which covered a period of 36 consecutive days during August and September, 191 3, four methods were employed in studying the points in question. To determine the prevalence of spores of En- dothia parasitica in the air at particular times and places a series of 756 exposure plates was made. The occurrence of ascospore expulsion was detected and its exact period of duration ascertained by the examina- tion of ascospore traps in the shape of object slides supported over peri- thecial pustules on the trees. The number of spores present in the air was determined quantitatn^ely by the aspirator method. Rather pro- longed exposures of water spore traps, consisting of sterile water in dishes, were made to secure additional information as to the kind of spores in the air, the periods of occurrence, and the distance transported. EXPOSURE-PLATE TESTS In testing the spore content of the air among diseased trees in the field for the presence of spores of Eiidolhia parasitica the exposure of sterile poured plates of chestnut-bark agar proved to be the most satis- factory method. The use of chestnut-bark agar ^ was found advan- tageous, since this medium inhibits the development of bacterial colo- nies and retards the growth of rapid-growing fungi, spores of which are ' Falck has pointed out the importance of convection currents in the dissemination of ascospores. (Falck, Richard. Uber die Luftinfektion dcs Muttcrkomcs (Claviceps purpurea Tul. ) und die Verbrcitimg pflanzlicher Infektionskrankheiten durch Temperaturstrumungen. /n Ztschr. Forst- u. Jagdw., Jahrg. 43. No. 3. p. 202-227. 4 fig., 1911.) For this reason we have used the word "air" in the title of the present paper. 2 Chestnut-bark agar was made according to the following formula: Add 50 gm. of finely chopped or ground air-dry chestnut bark to j.ooo c. c. of distilled water and boil for 15 minutes. Filter through cheesecloth or absorbent cotton and add water to make up to 1,000 c. c. Add 15 gm. of agar and boil until the agar has melted; then cool to 60° C. or under, clear with the whites of two eggs, filter, and sterilize in the autoclave. Mar. 25. 1915 Dissemination of Chestnut-Blight Fungus 497 present in the air. At the same time the growth of E. parasitica on this medium is vigorous and characteristic. As supports or stations on which to expose the plates, it was found con- venient and satisfactory to make use of the numerous large flat-topped stumps scattered throughout the coppice stand of diseased trees. To facilitate the recording of data, all of the stumps used were numbered with crayon and carefully described and located with regard to surrounding trees (fig. i). Here it may be mentioned, however, that other supports, such as the top rail of a fence or the top of a stake driven into the ground, were used in case of emergency attendant upon certain weather conditions. Fig. -Map of chestnut coppice growth at West Chester. Pa., in and near which the experiments on wind dissemination of the chestnut-bhght fungus were carried out. The stumps, rails, and stakes used for this purpose were all of such an age or nature that they were entirely free from lesions of the chestnut blight. Under conditions of ordinary fair weather the routine followed in making the exposures was similar throughout the tests. Plates were ex- posed at the rate of one about every half hour during the day, and the average length of exposure was about 5 minutes for each plate during the first 18 days. Then it was found advisable to lengthen the time of expo- sure, and thereafter 10 minutes, more or less, was the usual time allowed. Wind direction determined what stations were utilized each day, since an 498 Journal of Agricultural Research Vol. m, No. 6 effort was usually made to expose plates at stations where there were many diseased trees to the windward. Fig. -Map showing the location of some of the important outlying exposur^plate stations. Station si is at the comer of tne plot represented in figure i. During wet weather the routine was often varied considerably, espe- cially just after the cessation of a rain. At such times plates were often exposed in more rapid succession, even to the extent of exposing several in different locations at about the same time. Mar. 25, 1915 Dissemination of Chestnut-Blight Fungus 499 An anemometer was erected in the experimental plot, and from the successive readings of this instrument the wind velocities were computed. Continuous records of temperature were secured by means of a thermo- graph located in a standard instrument shelter near the plot, and by use of a rain gauge the exact rainfall in inches was determined. As com- plete data as possible were also secured relative to the exact duration of all rains. In describing and locating the stations used for exposure plates, meas- urements were made to the nearest diseased trees and to the nearest lesions, the horizontal distance being recorded. To supplement the de- scription, detailed topographic maps were made, showing the location of each station (figs, i and 2). The exposed plates were incubated at room temperature, and two rec- ords were usually taken. First, at the end of three days after exposure all fungous and bacterial colonies visible were marked and counted, and those suspected of being Endolhia parasitica were especially noted. After six or seven days of incubation the final record on each plate was taken. This included the total number of fungi, the number of bacterial and yeast colonies, and the number of colonies of E. parasitica, if any were present. In case of doubt as to the identity of the latter, owing to crowding by other colonies, transfers were made to 3 per cent dextrose agar, on which medium the growth of this fungus is even more characteristic than on chestnut-bark agar. The results obtained in the exposure-plate tests are presented in a somewhat summarized form in Tables I and II. Table I. — Summary of exposure-plate tests at West Chester, Pa., in IQI3, giving number 0} fungous colonies caught Num- I Number of fungous Total ber Length of Total colonies in any plate. Total nimiber Date of of plates exposure for cacfi time repre- Rain- fall. number of of culttires. fungous colonies of ex- plate. sented. Maxi- Mini- colonies. Undothia posed. mum. mum. parasitica. Minutes. //. m. Inches. Aug. ig 18 5 to 6 I 39 0. 40 Si 3 3i^ I 20 22 6 to 10 2 I9>4 0 24 I 227 0 21 19 5 to 7 I 48>^ 0 26 0 168 0 22 23 20 2S i>^to rA 4 to 8 1 44 2 19 }o-25 1 35 1 28 2 I 340 256 0 0 24 17 aH to 6 I 28X 0 47 I 206 0 25 19 4 to 6 I 34 0 75 5 419 0 26 20 4 to 7 I 48K 0 95 8 574 0 27 28 4 to 93.^ = 44>^ o- 175 70 4 +46 94 28 18 4 to 6K I 32 0 20 2 196 0 29 30 19 22 4 to s'A 5 to 6K I SVA •I. 10 { dl I 0 309 363 0 0 31 20 4 to 7 I 48 0 32 I 261 I Sept. I 19 5 to ty. I 44 0 12 I 122 0 2 19 4 to 6 I 37 0 39 3 213 0 3 19 s to ty^ I 37M 0 10 I 68 2 4 18 5 I 30 0 48 5 245 I S 18 5 to 6 I 31 0 40 0 •79 I 78745°— 15 5 500 Journal of Agricultural Research Vol. III. No. 6 Table I. — Summary of exposure-plate tests at West Chester, Pa., in ipij, giving number of fungous colonies caught — Continued ■NJi]Tn. Number of fungous Total ber Length of Total colonies in any plate. Total number Date of of exposure time Rain- number of of cultures. plates for each repre- fall. fungous colonies of ex- plate. sented. Maxi- Mini- colonies. Endothia posed. mum. mum. parasitica. Minutes. H. m. Inches. Sept. 6.. 18 5 to io>^ 2 I4K 0 6 0 39 0 7-- 19 7 to iiK 3 4H |o.3, 1 22 I 15 I 186 0 8.. 27 7X to 12 4 16 I 195 9 g.. 18 6 to iiK 2 23 0 31 2 194 2 lo. . 18 8 to iij^ 2 5°K 0 31 0 138 0 II . . 18 9K to 14K 3 19X 0 . ^5 0 118 0 12. . 13- ■ 18 19 5 to 14 7 to 13 3 12K 3 5^ }o- °95 41 I 73 I 2 121 332 0 0 14.. iS 9 to iiX 3 9K 0 20 2 126 0 15- ■ 18 9J-i to II 3 6 0 20 6 202 0 16.. 19 9>^ to 27 3 34K 0 28 0 251 0 17.. 17 9K to 15K 3 13K jo. 26 lo. 68 1 '' 67 1 400 1 160 ^ 60 2 221 "2 18.. 28 9K to 13K 5 ISX I 539 55 19.. 26 loX to 17 S 24X jo. 09 jo- 43 jo. 73 I 814 7 20. . 28 9X to 18 6 13K I 194 90 21. . 25 3K to 20 s S2K 5 1,187 160 22. . 36 6>^ to 30 9 5° 0 494 2 23 •■ 24 12X to 23 6 36 0 30 0 261 0 o From wind-blown bark fragments. Table II. — Detailed record of all exposure plates in which spores of Endothia parasitica were caught at West Chester, Pa., in igi3 Plate No. 4008 4376 4382 4383 4384 4385 4386 4460 4509 4529 4553 4599 4600 4602 4606 4609 4610 4617 4619 4626 4627 Date of exposure. Rainfall. Aug. 19. Aug. 27 . ....do.. ....do.. . . . .do. . ....do.. . . . .do. . Aug. 31 . Sept. 3.. Sept. 4. . Sept. 5. . Sept. 8. . do.. do.. do. . do.. do.. do.. do.. Sept. 9. , do.. Inches. 0.4 12 12 05s 05s 05s 055 10 10 10 10 37 37 37 37 37 37 37 37 37 37 Time elapsed since cessation of rain. h. 6+ 10+ o o o 5+ 5 5 7 8 9 13 14 8 8 o 42 44 18 20 22 35 o o 30 Length of exposure. Minutes. hi 6 6i^ 6 8 5 5 5 5 5 loK "K 12 12 10 8 9 6 8K Horizontal distance to nearest blight lesion. Feet. "iK 15X 15K II 25 4K a 2 a 2 a 2 a 2 II 14 a I «2 "iK a 2% a 2 "2 a 2 ''4'/2 Number of spores of Endothia parasitica caught. .1 I 3 16 21 33 20 I 2 I I 2 I I I o stumps more or less overhung by diseased sprouts. Mar. 15. 191S Dissemination of Chestnut-Blight Fungus 501 Table II. — Detailed record of all exposure plates in which spores of Endothia parasitica were caught at West Chester, Pa., in IQ13 — Continued Plate No. 4772 4784 4787 4788 4789 4790 4791 4792 4793 4795 4796 4797 5006 5010 5018 5020 5027 5°29 5033 5°37 5°4i 5042 5043 5044 5°45 5046 5047 5048 5049 5050 5°5i 5052 5053 5069 5070 5071 5072 5073 5074 5075 5087 5098 5102 Date of exposure. Sept. 17. ....do... Sept. 18. ....do... do... ....do... ....do... do... do... do... do.. do... do.. do.. vSept. 19. do.. do... do.. do.., do... Sept. 20. do.. do... do.. do.. do.. ....do... ....do... ....do... ...do... ....do... ...do... ....do... Sept. 21. do. . do... do.. do.. do.. do.. do.. Sept. 22. do.. Rainfall Inches. 0.095 09s 26 26 26 26 26 26 26 26 26 26 26 26 68 68 68 68 68 68 09 09 09 09 09 09 09 09 09 09 09 09 09 35 35 08 08 08 08 08 07 73 73 Time elapsed since cessation of rain. O o I I 2 2 2 2 2 3 3 3 9 II I 2 6 7 10 12 I 2 2 2 2 2 2 2 3 3 3 4 5 SI 54 o 19 35 48 54 25 34 58 14 19 23 7 4 8 5 7 55 o I 8 19 23 25 45 12 37 45 29 14 o 5 IS 17 44 56 I 20 O 30 Length of exposure. Minutes, 10 loK 9J< 10 10 13 II II 12 II II loX 13K II iiX I2K 11;^ I2K 13K 16 13 14 I2K I2X 15 15 u'A iiX 13 ibK 10 13K WA loK 20 19 13K 15X 12 14K 16 IS Horizontal distance to nearest blight lesion. Feel. "2^ a 2 WA 16 4K II 7 ISK "aA ^sA ^lA 16 Oi 02 4K 27 «2 "5 "S 27 217 195 no 27 85 180 237 7 II II 86 4K 5 19 17K i7>^ ^9 . 17K 14 7 19 77 iK Niunber of spores of Endothia parasitica caught. II 16 9 4 5 2 2 2 I 10 7 4 22 II 7 10 6 7 3 I I I 12 20 62 24 19 I 2 20 I I "> stumps more or less overhung by diseased sprouts. >> From fragments of bark. <■ Approximately. To supplement the tables, it may be well to give in chronological order a more detailed record of the actual routine pursued. A rain occurred the night previous to August 19, and examination of the ascospore traps (see "Ascospore-trap tests") showed that abundant expulsion of ascospores had occurred. But when the first plates were exposed too long a time had evidently intervened since the rain, as no positive results were obtained, except that one colony of Endothia para- sitica developed in a plate exposed in the early afternoon. 502 Journal of Agricultural Research Vol. m. No. 6 The next two days were fair, and as was expected for these weather con- ditions, no spores of this Endothia were caught. In the evening of August 22 there was a rain of 0.25 inch; 6 plates, therefore, were exposed early the next morning before the sun had dried the vegetation. Although the ascospore traps gave evidence that expulsion had occurred, no posi- tive results were obtained, which is explained by the fact that again too long a time had elapsed after the rain ceased. Dry, hot weather now continued until the afternoon of August 27, and the exposure plates yielded no evidence of the presence of spores of Endo- thia parasitica in the air. In the afternoon of August 27, however, two thunder storms occurred, in consequence of which the regular routine was departed from. Tables II and III show the outcome of the tests of this date. After the first storm two sets of plates were exposed in the course of an hour and a half. Two out of the second set yielded colonies of E. parasitica. Since the ascospore-trap tests (Tables X and XVII) did not give evidences of expulsion occurring when the first five plates were exposed, negative results were to be expected in those plates, and it is not surprising that only two out of the second set of seven plates yielded colonies of E. parasitica when the 19 ascospore traps examined for this particular period showed evidence of expulsion from only one perithecium. The meagerness of these results is partially accounted for by the small amount of rain, rapid drying, and the fact that the perithecia had hardly been wet a sufficient length of time. Table III. — Record of exposure plates made on August sy, IpiJ, at West Chester, Pa. BEFORE RAIN. Plate No. Time. Length of exposure. Station No. wind. Number of bacteria and yeasts. Total number of fungi. Number of colonies of Endothia parasitica. Direction. Miles per hour. 4359 4360 4361 4362 4363 4364 4365 4366 4367 4368 4369 4370 8.35 a. m 9.09 a. m 9.32 a. tn 10.03 a. m. . . . 10.29 a. m. , . . 10.57 a. m. . . . 11.34 a. m 11.57 a. m. . . . 1. 16 p. m 1-39 P- m 2 .09 p . ra 2-45 P- m Minutes. 4 5 5 5 6 ^K S 5 I 3 I 13 14 3 I I I 13 13 3 sw. sw. sw. sw. sw. w. w. w. w. w. w. w. 1.6 1.6 1.6 1.6 1.6 1.6 1-7 1-7 1-7 1-7 1-7 1-7 2 I I 0 I 3 2 0 0 0 I I 7 8 6 6 7 9 19 IS 22 33 13 7 0 0 0 0 0 0 0 0 0 0 0 0 RAIN NO. I (0.12 INCH, 3.15 TO 3.28 P. M.) 4371 3- 4372 3- 4373 3- 4374 3- 4375 4376 3- 4- 4377 4378 4- 4- 4379 4380 4381 4382 4- 4- 4- 5- 48 p. m 50 p. Ill 49 p. m 48 p. m 51 p. m 10 p 12 p 14 p ISP 17 p 40 p 12 p 7 3 w. 2.4 I 12 6K IS W. 2.4 0 10 6| 13 W. 2-4 0 17 6K I w. 2.4 0 32 7 8 w. 2.4 0 9 9K 6 w. Trace. 0 7 854 11 w. Trace. 0 8 8i 3 w. Trace . 2 15 8 16 w. Trace. 0 II sl IS w. Trace . I 9 4 3 w. Trace. 0 4 6 3 w. Trace. r 13 Mar. 35, 191S Disse initiation of Chestnut-Blight Fungus 503 Table III. — Record of exposure plates made on August 27, /p/j, at West Chester, Pa. — Continued RAIN NO. 2 Co.os.'i INCH. 5.35 TO 5.50 P. M.) Plate No. Time. Length of exposure. Wind. Number of bacteria and yeasts. Total number of fimffi. Number of No. Direction. Miles per hour. colonies of Endothia parasitica. 4383 4384 438s 4386 6.08 p. m 6.10 p. m 6.12 p. m 6.23 p. m Minutes. 6K 6 8 5 3 ! NW. I 1 NW. 16 1 NW. 6 i N. I. 2 1. 2 I. 2 ■ 7 0 0 2 0 30 25 70 25 16 21 33 20 The second shower on August 27 took place late in the afternoon, and though the precipitation was Hght, the cumulative effect of this rain upon that of the preceding one caused abundant expulsion of ascospores. The four plates exposed within about half an hour after this shower yielded colonies of E. parasitica in such numbers as to prove bejond doubt that the ascospores were at that time very prevalent in the air. The ascospore- trap tests for this period (Tables X and XVII) showed that, although out of the 14 examined only i bore any evidence of spore expulsion during the first 15 minutes after the cessation of the rain, 12 out of 14 showed expulsion of ascospores during the time in which the plate exposures were made. The sun had gone down, and the weather conditions following this storm were not conducive to the rapid drying of the bark. The results of this date were the first evidence secured which indicated beyond doubt that ascospores of Endothia parasitica are disseminated by wind under natural conditions. During the dry, hot weather of August 28 evidently no sporesof Endothia parasitica were present in the air, nor were any detected on August 29. On this date the humidity was high and cloudiness prevailed, accompanied by traces of rain insufficient to cause ascospore expulsion. As there was a rainfall of i.io inches in the evening of August 29, several plates were exposed in rapid succession the next morning, but no spores were caught. This failure is attributed to the fact that once more too long a time had passed since the rain ceased, and spore expulsion, though probably abun- dant in the night, had no doubt ceased long before the first exposures were made. Throughout the following week there was no rain, and no ascopore expulsion occurred at any time. For the night previous to September 8 a rainfall of 0.37 inch was recorded, the time of cessation being prior to 1.30 a. m. The ascopore traps gave evidence of plentiful spore expulsion. Between 6.27 and 8 a. m. eight exposures were made before the sun had dried the vegetation and while the bark was still wet in places. Three of these plates yielded colonies of Endothia parasitica, and as two of them were exposed at sta- tions more or less in the open, it would seem that ascospores were at that time prevalent in the air to some extent. The third plate and also five others exposed at later intervals during the day each yielded one colony of E. parasitica. Of the plates exposed on September 9, a dry, hot day, two in the morn- ing also yielded one colony each of Endothia parasitica. During the dry weather of ,Septembcr 10 and 1 1 negative restiUs were obtained. 504 Journal of Agricultural Research Vol. III. No. 6 Cloudiness prevailed on September 12, with traces of rain insufficient to cause spore expulsion. In the night less than one-tenth of an inch of rain fell, and subsequent examination of the ascospore traps showed that very light ascospore expulsion had occurred. Five plates were exposed before 8 o'clock the next morning, but the bark was dry at the time and no spores were obtained in any of the plates exposed that day. Three days of clear, hot weather followed, and no spores were caught. Of the plates exposed on September 17, two yielded one colony each of the fungus. In the evening of September 17 a series of rains began, occurring usually in the night. Our most important positive results were obtained from tests made following these rains. Tables II and IV give the results of the plates exposed on September 18. Although the rain ceased before 4 a. m., a heavy fog prevailed in the early morning, there was only a trace of wind, and it was more or less cloudy all day. Because of these condi- tions the bark of the trees was slow in drying, and examination of the ascospore traps (Tables XI and XVII) showed that abundant spore expulsion had occurred in the night and was still in progress while the first four exposure plates were made. A few of the traps gave evidences of the continuation of expulsion during the time in which the next 1 1 plates were exposed. Six of these yielded colonies of Endothia parasitica in varying numbers, and two exposed much later in the day also showed one colony each. Table IV. — Record of exposure plates made on September 18, iQl3,at West Chester, PaA Plate No. 4787 4788 4789 479° 4791 4792 4793 4794 4795 4796 4797 4798 4799 5000 5001 5002 5003 5004 5°oS 5006 S007 5008 5009 5010 5011 5012 5013 S014 Time. 5.51 a. m. . . 5.54a. m.. . 6.00 a. m. . . 6.19 a. m. . . 6.35a. m.. . 6.48 a. m. . . 6.54 a. m. . . 7.02 a. m. . . 7.25 a. m. . . 7.34a. m... 7.58a. m... 8.04 a. m. . . 8.29 a. m. . . 9.20 a. m. . . 9.49 a. m. . . 10.21 a. m. . 10.54 a. m. . 11.25 3- ■"■ ■ II. 51 a. m. , 1. 14 p. m. ., 1.44 p. m. . 2.15 p. m.. 2.46 p. in. . 3.19 p. m. . 3.45 p. m.. 4.16 p. m. . 4.44 p. m.. 5.23 p. m.. Length of exposure. Minutes. 10 10 13 II II 12 loK II II loX loK 10 II loX 12 II iiK r3K loX loX 15 II II 13K iiX loK' Station No. 3 15 6 I II 3 4 16 3 13 IS 9 17 I 3 6 15 33 3 26 21 22 6 33 21 2S 23 I Direction. NW. NW. NW. NW. NW. NW. NW. NW. NW. NW. W. W. W. W. W. w. w. w. w. w. w. w. w. w. w. w. w. w. Miles per hour. Trace. Trace. Trace. Trace. Trace. Trace. Trace. Trace. Trace. 0.4 •7 ■7 •7 1. I 1.8 2.8 2.8 3-7 3-7 2. 2 2.8 1.8 I. 2 I. 2 •9 •9 . 2 . I Number of bacteria and yeasts. 6) 6) Total number of fungi. Niunberof colonies of Endothia parasitica. 20 58 39 59 67 16 32 23 3° 8 II 2 I 8 25 10 19 12 13 17 7 10 12 12 16 I 21 3 16 9 4 5 2 2 o 2 I I o o o o o o o I o o o I o o o o o a Rainfall, night previous, 0.26 inch. Time of cessation, prior to 4 a. m. ^ Nimicrous. Mar. 3s. 191S Dissemination of Chestnut-Blight Fungus 505 In the night of September 18 a rain of 0.6S inch was recorded, the time of cessation being the next morning before 3. 45. Fog again pre- vailed all day September 19 and a noticeable spray fell until 5.38 a. m. and began again after 2.41 p. m. Ascospore-trap tests (Tables XII and XVII) showed that in 10 out of the 17 traps examined there was spore expulsion after 7.35 a. m. Of the eight plates exposed prior to this time but two yielded colonies of this Endothia. The first three plates were exposed in an open field at considerable distances from the trees (fig. 2, stations 41, 42, and 43) in the same direction toward which the wind was blowing, but no spores of E. parasitica were caught. These negative results may be accounted for by the action of the falling mist. Later in the day four exposures at various inter\'als yielded one colony each of the chestnut-blight fungus. Table V. —Record of exposure plates made on September 20, igi3, at West Chester, Pafl Plate No. Time. 5041 5042 5043 5044 SO45 5046 SO47 5048 SO49 5050 S051 5052 5°53 5054 5°5S 5056 5057 5058 S°59 5060 5061 5062 S063 5064 5065 5066 5067 5068 Length of exposure. 5.503. m.... 5.55 a. m.... 5.56 a. m 6.03 a. m. . . . 6.14 a. m. . . . 5.18 a. m. . . . 6.20 a. m. . . . 6.40 a. m. . . . 7.07 a. tn. . . . 7.32 a. ra.... 7.40 a. m. . . . 8.24 a. m. . . . g.09 a. m. . . . 9.37 a. m 10.01 a. m. . . 10.32 a. m. . . II. og a. m. . . 11.44 a. m. . . 12.08 p. m. .. 1.05 p. m. . . . 1.35 p. m.... 2.08 p. m. . . . 2.36 p. m. . . . 3.09 p. m. . . . 3.42 p. m.... 4.06 p. m. . . . 4.41 p. m. . . , 5.25 p. m.... Minutes. 16 13 14 I2j< I2X % II>^ ^3 , 16K 10 I3K I4K I4K 10 II>< 10 WA 13K 14K 15K ^5% 14 18 12X 12 Station No. 44 43 46 SI 47 48 49 II I 9 5° 6 8 27 6 10 29 26 10 30 II 10 8 9 29 23 12 12 Wind. Direction. Miles per hour. NE. 2.6 NE. 2.6 NE. 2.6 NE. 2.6 NE. 2.6 NE. 2.6 NE. 2.6 ENE. 2.6 ENE. 2.6 ENE. 2.6 ENE. 2.6 ENE. 2-5 ENE. 2-5 ENE. 30 ENE. 2.4 E. 2.4 E. 2.6 E. 2.6 E. 2.7 E. 2.7 E. 2-3 E. 2- 5 E. 2-7 E. 2. 7 E. 2.7 E. 2-3 ESE. 2-7 SE. 2.7 Xumber of Total number of fungi. N^umberof bacteria and yeasts. colonies of Endothia parasitica. 0 12 10 0 9 7 0 10 4 0 23 22 0 14 II 0 9 7 0 II 10 0 6 6 0 9 7 0 6 3 0 6 I 0 I I 0 I I 5 II 0 0 2 0 0 3 0 0 12 0 0 6 0 0 2 0 0 2 0 0 I 0 0 19 0 2 4 0 0 I 0 ' 0 2 0 0 4 0 0 4 0 3 4 0 " Rainfall, night previous. 0.09 inch. Time of cessation, 3.25 to 3-55 a- ni. While there was but 0.09 inch of rain in the night of September 19, very important results were obtained on September 20. Tables II and V give the results secured. Fog prevailed during the entire day, and the bark on the trees dried very slowly. Examination of the ascospore traps to determine the duration of spore expulsion (Tables XIII and X\'II) showed that in 8 out of 19 the perithecia were active after 9.18 a. m. 5o6 Journal of Agricultural Research Vol. ni. No. 6 All of the 13 plates exposed previous to this time yielded colonies of Endothia parasitica. Eight of these exposures were made in the open field at varying distances south and west of the plot of diseased trees (fig. 2), the wind being from the northeast. The distance relations brought out by these tests are discussed later. Although five ascospore traps showed evidences of the occurrence of spore expulsion after 10.07 a. m., no colonies of this fungus appeared in any of the plates exposed after 9.23 a. m. This indicates that spores were evidently not sufficiently numerous in the air after that time to be detected by the exposure- plate method. The results obtained on September 21, as shown in Tables II and VI, bring out again the direct relation of rain to wind dissemination. Two plates exposed during a i6-minute interval between showers in the early morning yielded colonies of Endothia parasitica in such numbers as to prove without doubt that ascospores were very prevalent in the air at that time. After the second rain, ending at 8.20 a. m., only the five plates exposed within an hour after its cessation yielded colonies of E. parasitica, even though 14 out of the 21 ascospore traps examined showed that considerable spore expulsion had taken place after 10 a. m. (Table XI). However, a south wind of increasing velocity prevailed, and at 9.2 1 the sun appeared, causing a marked rise in temperature, so that the bark dried very rapidly after that time. Furthermore, the higher wind may also have dispersed and scattered the fewer spores expelled thereafter to such an extent that none happened to fall into the exposed plates. Table VI, — Record of exposure plates viadc on September zi, IQIJ, at West Chester, Pa. RAIX XO. I (about 0.35 INCH. CEASED 6.24 A. M.) Plate No. Time. Length of exposure. Station No. Wind. Number of bacteria and yeasts. Total number of fungi. Number of colonies of Endothia parasitica. Direction. Miles per hour. 5069 5070 6.23 a. tn 6.30 a. m Minutes. WA loK 12 13 SSE. SSE. 2.6 2.6 (0) (") 30 50 12 20 RAIN NO. 2 (ABOUT o.oS INCH, 6.40 TO 8.20 A. M.) 5071 5072 5073 5074 5075 5076 5077 5078 5079 5080 5081 5082 5083 5084 8.35 a.m. 8.37 a. m . 9.04 a. m . 9.16 a. m . 9.21 a m. 9.40 a. m. 9.43 a. m. 10.25 a. tn 10.28 a. in 10.46 a m 11.00 a. tn 11.07 ^- ™ 11.48 a m 1.19 p. m . 20 13 SSE. 2.6 (*) 70 19 12 SSE. 2.6 2 28 13^ 13 SSE. 3- I I 21 I5K 19 S. 3- I I 5 12 II s. 3- I 0 9 15 13 s. 3- I 3 37 I3K 37 s. 3- I I 18 15 13 ssw. 5-8 3 30 16 19 ssw. S-8 12 80 10 37 ssw. 5-8 0 21 12 14 ssw. 5-8 6 44 I3K 37 ssw. 5-8 I 28 16K 12 ssw. 5-8 S 98 isK 13 s. 5-6 0 26 62 24 19 I o Numerous. 6 Few. Mar. as. 1915 Disseifiination of Chestnut-Blight Fungus 507 Table VI. — Record of exposure plaUs made on September 21, IQ13, at West Chester, Pa . — Continued RAIN NO. 3 (.\BOUT 0.03 INCH, I.4J TO 2. II P. M.) Plate No. Time. Length of exposure. Station No. Wind. Number of bacteria and yeasts. Total number of fungi. Number of colonies of Endothia parasitica. Direction. Miles per hour. 5085 2.20 p. m Minutes. 3K 13 s. 6.2 8 160 0 RAIN NO. 4 (about O. 04 INCH, 2. !0 TO 2.50 P. M.) 13 12 s. s. 6.2 6.2 0 0 144 3S 5086 5087 2.53 p.m. 3.10 p.m. o 20 RAIN NO. s (ABOUT 0.03 INCH, 3.26 TO 4.07 P. M.) 5086A S087A 5088 5089 SO9O 509 1 4.08 p. m 4.17 p. m 4-35 P- m 4-SSP- m 5.16 p. m 5-33 P- m I9K 13 S. 8.0 10 58 14 37 S. 8.0 3 5° 10 II s. 8.0 4 45 15 12 s. 8.0 6 42 I7K II s. 8.0 3 38 II 10 s. 8.0 5 20 In the afternoon of the same day three light showers occurred, and one plate exposed in the interval after the second of these caught 20 ascospores. After the first of these showers the exposure was cut too short by the recurrence of rain to give a reliable test. It will be seen that none of the six plates exposed during the i hour and 36 minutes after the last shower yielded colonies of Endothia parasitica, despite the fact that 6 out of 1 1 ascospore traps examined (Tables XIV and XVII) gave evidence that expulsion had occurred during that period. In explanation it may be stated that the wind had attained a higher velocity at this time and was blowing quite briskly in the open. It is readily conceivable that with such a wind the spores as they were ex- pelled might have been transported with such speed and their numbers dissipated so rapidly that none chanced to fall on the rather small area represented by the exposure plates. A rather heavy rainfall was recorded on the night of September 21, but it ceased before 12.45 3- ™- Examination of the ascospore traps showed that there was abundant spore expulsion during the night, and 5 out of 21 traps gave evidences of the occurrence of expulsion after 7.30 a. m., on September 22 (Tables XV and XVII). Of the 13 plates exposed between 5.56 a. m. and 7.35 a. m. but 2 yielded positive results (Table II). No spores were caught in any of the plates exposed there- after, even though two ascospore traps bore evidences of the occurrence of light expulsion after 1 1 .24 a. m. The meager results obtained on this date are no doubt due to the long period of time intervening since the cessation of rain the night before. Clear, hot weather prevailed during September 22 and 23, and no spores were caught. 5o8 Journal of Agricultural Research Vol. in. No. 6 The relation of the time elapsed since the cessation of rain to the prevalence of ascospores in the air among diseased trees is shown in Table VII. Table VII. — Relation of the time elapsed since the cessation of rain to the number of spores failing on an area of I square foot per minute in igi^ at West Chester, Pa. PLATES EXPOSED ON SEPTEMBER 18. I9I3 Number of spores of No. of Time elapsed Number of colonies of Endothia parasitica plate. tion of rain. Endothia parasitica. falling on an area of I square foot per minute. H. m. 4787 I 51 II 15-74 478S I 54 16 22.32 4789 2 9 12-55 4790 2 19 4 4-29 4791 2 35 5 6.34 4792 2 48 2 2-53 4793 2 54 2 2.32 4795 3 n 2 2-53 4796 3 34 I 1.27 4797 3 58 I 1.36 4798 4 4 0 0 4799 4 29 0 0 PLATES EXPOSED ON SEPTEMBER 21. I9I3 5°7i 15 62 43-24 5072 17 24 17. 62 5073 44 19 19.64 5074 56 I 0. 91 5075 I I 2 2.32 5076 I 20 0 0 5077 I 23 0 0 5°78 2 5 0 0 An examination of these tables shows that on September i8 the spore content of the air decreased more or less gradually during the third and fourth hours after the rain, while on September 21 the spore content decreased very abruptly and no spores were obtained after the first hour following the cessation of the rain. The duration and the abundance of the ascospore expulsion on these dates (Table XI) are seen to have differed likewise, and a comparison of the weather conditions gives the probable explanation, since it was calm and foggy on the i8th and hot and sunny with a brisk wind just following the rain of the 21st. Condi- tions following the rain on the i8th were such as to prevent rapid drying of the bark, so that spore expulsion continued during a much longer time than on the 21st, when the bark dried rapidly. Furthermore, the brisk wind of September 21 would tend to disperse the spores very rapidly, whereas the comparative calm of September 18 would be favorable to a more prolonged prevalence in the air near their source. In this regard Mar. 25, 1915 Dissemination of Chestnut-Blight Fungus 509 it should also be noted that on September 20 (Table II), when foggy weather followed the rain, spores were prevalent in the air during at least five hours after the rain had ceased. A glance at the figures representing the number of spores falling each minute on a surface equal to i square foot shows that during periods of one to four or more hours after a rain — in other words, during such time as expulsion continues — healthy trees among diseased ones would be sub- ject to infection, since some of the ascospores would find lodgment upon exposed parts of trunks and branches. The results obtained in the early morning of September 20 by making exposures in an open field at varying distances from the principal source of spores (figs. 2 and 3) are presented in Table VIII. Table "VIII. — Relation of distance from source of spores to number of spores falling on an area of I square foot per minute in IQI3 at West Chester, Pa.<* Number of Distance spores falling Plate No. Time. from source of on an area of spores. I square foot per minute. Feel. 5°44 6.03 a. m. 27 24.07 5°4S 6.14 a. m. 85 12. 52 5046 6.18 a. m. 180 6-51 5042 5-55 a. m. 266 7-51 5047 6.20 a. m. 409 9- 30 S°4i 5.50 a. m. 414 8. 71 a Plates exposed on Sept. 20, 1913. These exposures, all made within about half an hour and in the same general direction from the plot of diseased trees — i. e., the direction toward which the wind was blowing — show that in a general way the number of spores falling upon equal surfaces in equal intervals of time decreases as the distance from the source of spores is increased. The fact alone that in an open field at rather long distances from diseased trees ascospores were prevalent in the air to such an extent that every minute from 6 to 24 spores were settling upon a surface equal to i square foot (PI. LXIII, figs. I and 2) indicates that at such a time many opportunities would be offered for exposed parts of undiseased trees at considera- ble distances from diseased ones to become infected by wind-borne ascospores. Furthermore, these results show that the maximum distance over which ascospores might be transported by the wind was by no means obtained, and the large numbers found at the longest distances in this experiment, given in Table VIII, when a light wind prevailed, indicate that even with a relatively light wind ascospores are probably conveyed distances far greater than these. In a consideration of the exposure plates yielding the high numbers of colonies of Endothia parasitica it is interesting to note the relatively large proportion of the spore content of the air fomied by ascospores of this fungus at certain times (Table IX). 5IO Journal of Agricultural Research Vol. m. No. 6 Table IX. — Percentage of the tiumber of spores of Endothia parasitica to the total spore content of the air, as shown by exposure-plate tests on chestnut-bark agar in igi3 at West Chester, Pa. Plate No. Date. Total num- ber of fun- gous colo- nies. Num- ber of colo- nies of Endo- thia para- sitica. Per- centage of colo- nies of Endo- tJhia para- sitica to total spore content of air. Plate No. Date. Total num- ber of fun- gous colo- nies. Num- ber of colo- nies of Endo- thia para- sitica. Per- centage of colo- nies of Endo- thia para- sitica to total spore content of air. 4383 4384 438s 4386 4787 5041 5042 5043 5044 5045 Aug. 27 ...do ...do ...do Sept. 18.... Sept. 20. . . . ...do ...do . . .do ...do 30 25 70 25 20 12 9 10 23 14 16 21 33 20 11 10 7 4 22 II 53 84 47 80 55 83 77 40 95 78 5046 5047 5048 5°49 5°5° 5069 5070 5°7i 5072 5°73 Sept. 20. .. . ...do ...do ...do ...do Sept. 21 ... . .. .do ...do ...do ...do 9 II 6 9 6 30 5° 70 28 21 7 10 6 7 3 12 20 62 24 19 77 91 100 77 50 40 40 88 85 go In connection with these figures it should be borne in mind that the fungi represented are such as will grow only on chestnut-bark agar. Taking into consideration, however, the relatively large numbers of other fungi ordinarily developing in the exposure plates (Table I), it is a note- worthy fact that at certain periods when ascospore expulsion was in progress the spores of this one species should constitute from 40 to 100 per cent of the total spore content of the air. Since these plates were exposed not long after a rain, a possible expla- nation suggested is that spores of other fungi were washed from the air by the rain and the supply had not yet been replenished, whereas condi- tions were very favorable to the abundant expulsion of ascospores of Endothia parasitica. It has also been suspected that certain types other than this fungus which were often found in plates exposed at such times represented other ascomycetous fungi the spores of which had just been expelled. SUMMARY OF EXPOSURE-PLATE TESTS In all of the exposure plates yielding colonies of Endothia parasitica it was determined from the time of appearance of these colonies that all originated from ascospores. Therefore we may safely state at the outset that under the conditions of the tests little or no wind dissemination of pycnospores occurred. By comparison with ascospore-trap tests it is evident that ascospores of Endothia parasitica were caught in the exposure plates in numbers and at some distances from trees only during certain periods following rains when ascospore expulsion was in progress. The possible exception occurred on the morning of September 8, when no series of observations was made on the ascopore traps. As the occurrence of ascospores in the air in considerable numbers is the prime requisite for wind dissemination and as ascospore expulsion Mar. 25. I9IS Dissemination of Chestnut-Blight Fungus 511 occurs only when the perithecia-bearing bark has been wet by rains, the following facts are presented to show that wind dissemination is directly dependent upon weather conditions causing spore expulsion. Of the total number of 756 plates exposed during these tests 95 were exposed while ascospore expulsion was known to have been in progress, and of these, 41 yielded colonies of Endolhia parasitica. Of the remaining 661 plates exposed at other times than those noted above, but 23 yielded colonies of E. parasitica, and 14 of these were exposed within 12 hours after expulsion was known to have occurred. To bring out in a more striking manner the relation of rain to wind dissemination, it is worthy of note that out of a total of 427 ascospores of Endothia parasitica caught in the exposed plates 402, or 94 per cent, were caught in plates exposed while spore expulsion was known to have been in progress, and of the remaining 25 spores 3 were caught within 5 hours after the cessation of a rain (Sept. 8) and 12 more were caught within 12 hours after ascospore expulsion was known to have occurred. This leaves but 10 out of 427 spores, or 2.3 per cent, seeming to be stray ascospores bearing no relation to a rain. As to the origin of the 22 colonies of Endothia parasitica appearing in plates exposed when spore expulsion was known not to be in progress (see Table II), the following points are cited to prove that they originated from stray ascospores which, after expulsion, lodged on near-by or, per- haps, distant trunks, limbs, or leaves and were subsequently loosened by the mechanical action of some agency. 1. All but one of the 21 plates containing these colonies yielded only a single colony of Endothia parasitica each. 2. In one colony a fragment of bark was visible at its center. 3. All except one of these spores were caught at stations more or less overhung by branches of diseased trees, and all except three were caught on stumps surrounded by sprouts. 4. Only I out of 192 plates exposed at unsheltered stations when expulsion was not in progress yielded a colony of Endothia parasitica. If these had been stray spores that were still floating in the air since expulsion, they would have fallen just as frequently into plates exposed out in the open at unsheltered stations. During a period of ascospore expulsion following a rain it seems probable that the spores would not all be swept away by air currents but that some few would find lodgment upon near-by leaves and branches. Such lodgment is especially hkely to take place Lf there is no noticeable wind when expulsion is in progress. Thus, it seems quite probable that the colonies obtained when perithccia were not active originated from spores dislodged from either healthy or diseased parts of trees more or less overhanging the plates. Unless attached to a bark fragment, the path of these spores in falling would not necessarily approach the vertical, and such spores might be transported by the wind just as readily as though they were freshly expelled. This explains, perhaps, why one spore was caught in plate No. 5037, exposed 27 feet from the nearest chestnut tree. The probable reason, then, why, with this exception, such stray spores were caught only under trees is that the rareness of their occurrence in the air pre- vented their detection elsewhere than in very close proximity to their place of temporary lodgment, since with the exposure-plate method the chance of detecting these spores decreases very rapidly as the distance from their source is increased. 512 Journal of Agricultural Research voi. iii, no. s Obviously no exposures could be made during a rain, but ascospore- trap examinations have shown that abundant spore expulsion may occur during the actual fall of the rain. It is evident, however, that at such a time wind dissemination would be reduced to a minimum, because the spores upon expulsion would soon be washed to the ground or to near-by bark or fohage. Therefore, under the conditions of our tests it can be said that, with the exception of the few stray ascospores loosened from temporary lodgment, wind dissemination of Endothia parasitica occurs only during certain periods after rains, when ascospore expulsion is in progress. ASCOSPORE-TRAP TESTS In order to detect ascospore expulsion whenever it occurred, use was made of what we have termed "ascospore traps." An ascospore trap consisted of a glass object slide held in place over perithecial pustules on the bark of a diseased tree by means of a wooden bracket either above or below the slide (PI. LXIV, figs, i and 2). The slide was wedged firmly into a slot in the bracket so as to be suspended about one-eighth of an inch or less from the papillae underneath. These traps were placed on lesions of various ages on trees more or less scattered throughout the experimental plot (fig. i). As the ascospores of Endothia parasitica are expelled they adhere to the glass, and the spores expelled from each ostiole usually form a definite "spot," so that the number of spots on the slide represents the number of perithecia in the area underneath which have exjjelled spores. During the progress of the work on wind dissemination, it was found possible by means of these traps not only to detect the occurrence of ascospore expulsion but to determine even with some degree of accuracy the exact duration of perithecial activity. As has been brought out in the discussion of the exposure-plate tests, the occurrence of ascospores in the air in numbers is directly dependent upon the continuation of their expulsion after a rain has ceased. The duration of expulsion becomes, therefore, an essential factor in deter- mining the period during which wind dissemination may occur. In making this determination the method of procedure was as follows: Out of the total number of 69 ascospore traps usually about 20 were selected, representing areas of vigorous perithecia where previous expe- rience indicated that abundant expulsion was most likely to occur. The sUdes from these traps were collected as soon as possible after the rain and were replaced with clean shdes. Then, after a convenient interval, this second set of slides was collected and replaced with clean ones. This operation was repeated at intervals of several minutes to several hours until none of the slides bore spots of expelled ascospores. Several series of trap collections were usually made after each rain, and a subsequent examination of each slide revealed whether or not any expulsion had occurred under that trap in the period during which that particular slide had been in place on the tree. Although usually visible to the unaided eye, an examination with a hand lens was often necessary to detect very faint or very diffuse spots of ascospores on the slides. The detailed results for September 20 are given to show the behavior of individual traps (Table X). The results given in the summary for the other dates were obtained in a similar manner and the individual records will therefore be omitted. Mar. 2s. 191S Dissemination of Chestnut-Blight Fungus 513 Table X.- -Record of ascospore-trap collections on September 20, 1913, Chester, Pafi at West Time from cessation of rain to — Number of perithecia expelling ascospores between times stated. Repladng of slides. Collection of slides. 6 t H 44 23 6 0 0 6 0. s 22 7 4 I 0 i 6 Z 2 118 0 vO 6 Z 0. 2 73 66 17 8 5 3 0 s 6 2; a m H 79 43 IS I 0 0 0 6 6 1 i i s 6 Z 1 d a CO «^ d ^: a ca 1 2 I 0 0 ■■ ;; 6 % 7 3 2 I I 0 0 0 d a 13 2 0 0 0 0 0 0 d 2; a c8 8 J I 0 1 i 6 Z 9 H 5 4 0 0 0 6 2 a 0 0 0 0 0 0 0 4 d a 0 0 0 0 0 0 0 0 d 0. a 2'»48° t0 3''io™ .. .. 53 5 I I 0 40+ 23 9 I 0 14 0 0 0 238+ lOI 22 11 4 3 0 0 82 53 30 4 0 0 4 0 0 6^12" to b^ii°^ 10034™ to iil'3i°^ lahi-m tQ jihjgm 0 II 9 2 0 0 0 0 ° lo"34'» to li''3l'° I4''4o" o Rainfall, night previous. 0.09 inch. Time of cessation, between 3.25 and 3.55 a. m. Table XI. — Summary of records of ascospore-trap collections in IQ13 at West Chester, Pa. Date. Rainfall. Number of traps Time from cessation of rain to — Replacing of slides. Collection of slides. Number of perithecia expelling ascospores between times stated. Aug. 27.. Do... Sept. 18.. Sept. 19.. Sept. ao.. Sept. 21.. Do... Sept. 22.. Oct. 20. . . ^Do... Inches. 0-12 o" to i8"» o™ to 15™ 45™ to !*» 2^tP^ to 2''17™.. 31119" to 3''S4'». 4'20» to 4''38". S''47" to6''37'». l''48" to 2'>4",. jlijjm to s'jo". 2''48= to jliio". 4*'5™ to 4^27™., 51*23™ to 5l>4l'°. 6h,2m to 6''33". 61153™ to jlill™, jh.jm to 7l'47">. 101134™ to 11*131' is'm" 1I14O™ to 211. . . . 3116" t0 3ll22".. 33" to 41" 53" to ill sliio" to sli30". 6l'3l°i to6li5l". 8116= to 81125"., 10I139" to 111*2" ,lij" to jliii".. 12" to 27" 20" to 31" 37" to 43" 45" to 54" 57" to lli9". . . , l'l2" to 1I126". 1I128" to 11146". llijo" to 2I19".. o" to 18" 1I122" to ll*27".. o" to 15" 45" to ill 1411 (Aug. 28)... 2I16" to 2I137". . . 3I119" to 3°54". . 4I120" t0 4li38".. 51147" to 7I11Q". . j2h25" to 13^10" 11148" to 2114". . . 31135°' 1051130".. jhjjn. t0 5li30».. 12I110" to 13I1 . . . 2I148" t0 3lil0". . 4115" to 4"27". . . 5I123" to 51141". . 6lii2" to 6I133". . 6I153" to 7I111". . 71135" to 71147'°.. 101134" to 111131" 13I114" to I3l'38" I3'i49'» 1I14O" to 211 3116" to 3I122". . . 51128" to 51136".. 33" to 41" 53" to I 111" l''3l'° 5I110" to 51*30". . 6I131" to 61151". . 8116" to8li2j"°... 9I13S" to ll^Il". 14I112" to 14I116" 1I13" to iliii". . . 1I116" to 1I121". . 13" to 27" 22" to ii^ 37" to 43" 45" to 54™ 57" to 1I19" lhi2" to ill 36". . ll'2S'° to ll'46«>.. 1I155" to 2119" 5I121" to 51*54". . 360 227 1.047+ Si 9 35 I 236 219 IS 46 875+ 433 130 28 3,494+ 211 1.199+ S77 + 6 618 82 24S 303 387 379 390 262 •52 37 d This summary includes the records of traps Nos. 58, 59, 72, and 74 only. SH Journal of Agricultural Research Vol. Ill, No. 6 The data secured relative to the duration of ascospore expulsion after certain rains are given in a summarized form in Table XI. As will be seen, all of these, except the results obtained on October 20, bear refer- ence to rains occurring during the progress of exposure-plate tests. Although the exact time of cessation of rain is a very important point, in the cases of September 18, 19, 20, and 22 it could not be more accu- rately determined, because the rains all ceased in the night. A com- parison of the figures presented in these tables with the results obtained in the exposure-plate tests will show a close interrelation. On October 20 another opportunity was offered to obtain data relative to duration of spore expulsion. By selecting only a few traps and chang- ing the slides at much shorter intervals more in detail was learned in regard to the activity of the perithecia (Table XI). Since the point had been often suggested that these ascospore-trap tests might not yield results typical of natural conditions because of the protection from drying afforded by the glass suspended over the bark and that because of this spore expulsion was greatly prolonged, occasion was taken on October 20 to determine the validity of this contention. About two hours after the cessation of rain, when expulsion had appar- ently ceased under most of the ascospore traps that were being tested (Table XI), a number of clean slides were placed at random over other areas of perithecia-bearing bark which appeared still to be damp, to determine whether or not perithecia on bark unprotected by glass slides were expelling spores at this time. Owing to the high south wind and occasional sunshine, such promising areas of bark were found only on the north side of trunks, either where loosened bark about a bad lesion had become soaked or in locations more or less protected by sprouts or by stumps from the drying action of the wind. The slides were held in place by a cord tied around the trunk, and care was taken to prevent contact with the papillas. The results obtained from these 22 test traps are given in Table XII. Table XII. — Record of test traps on October 20, 1913, at West Chester, Pafi Trap No. Time of placing slide. I 2 3 4 5 6 7 8 10 II 12 14 IS 16 17 18 19 20 21 22 11.20 a. ra 11.25 a. m 11.27 a. m 11.27 a. m 11.32 a. m 11-352. ni 11.35 a. m 11.39 3- ™ 11.34 a. m 11.47 ^- T" 11.47 3- rn 11.55 a. m 12.00 to 12.04 p. m 12.00 to i2.o4p. m 12.00 to i2.o4p. m 12.00 to i2.o4p. m 12.00 to i2.o4p. m 12.00 to 12.04 p. m 12.00 to i2.o4p. m 12.17 P- ni Time of colJectiou. 1.30 to 1.30 to 1.30 to 1.30 to 1.30 to 1.30 to 1.30 to 1.30 to 1.30 to 1.30 to 1.30 to 1.30 to 1.30 to 1.30 to 1.30 to 1.30 to 1.30 to 1.30 to 1.30 to 1.30 to 2.00 p. ra 2.00 p. m 2.00 p. m 2.00 p. m 2.00 p. m 2.00 p. in 2.00 p. m 2.00 p. m 2.00 p. m 2.00 p 2.00 p 2.00 p 2.00 p 2.00 p 2.00 p 2.00 p 2.00 p 2.00 p 2.00 p 2.00 p Results of examination. I faint spot. o. I light spot. o. o. o. o. o. I faint spot o. o. o. o. 3 spots; I rather heavy. o. 1 light spot. o Kainfall. 0.86 inch. Time of cessation, 9-09 a. m. Wind, high SSW. Mar. 2s. 191S Dissetnination of Chestnut-Blight Fungus 515 SUMMARY OF ASCOSPORE-TRAP TESTS From the standpoint of \\dnd dissemination, the all-important feature proved beyond doubt by these tests is that in every case where ascospore expulsion occurred at all it continued for a time after the cessation of the rain, thus insuring a supply of spores in the air. A glance over these results shows that in a general way the volume of ascospore expulsion, as measured by the character and number of spots on the slides, is greatest during or shortly after the rain and decreases more or less uniformly as the bark dries. On August 27 the rains were of the thunderstorm type, being of very short duration, and consequently the perithecia had hardly been wet for a sufficient length of time when the rain ceased. The greatest volume of expulsion occurred, therefore, a little later, evidently between 15 minutes and i hour after the rain had ceased. With the exception of the rain in the afternoon of September 21, the tests of September 18, 19, 20, 21, and 22 are rather unsatisfactory, since no records could be obtained until some time after the rain had ceased. The summary for these dates (Table XI) shows that the maximum vol- ume of spore expulsion had occurred before the first collections were made, and whether the climax occurred during the rain or shortly after- wards can not be stated. In the case of the afternoon rain of September 21 very evidently the maximum volume of spore expulsion took place before 33 minutes had elapsed after the cessation of the rain. On October 20, after the bark had been thoroughly saturated by a rain in the night, the greatest volume of expulsion occurred within one hour after the rain. Two hours and nine minutes later a light rain of 58 min- utes' duration began, and the results secured after this shower (Table XI) show that in three traps tested the greatest volume of expulsion occurred, not during the rain, but after 22 to 43 minutes had elapsed since its ces- sation. As to the rate of subsidence of ascospore expulsion after the rains, Table XI shows a marked contrast between the results obtained on differ- ent dates. This has been mentioned in the discussion of the exposure plates and the relation of the subsidence of ascospore expulsion to weather conditions. In the cases of September 18, 19, 20, and 22 the duration of expulsion is seen to have been prolonged after the rains, and in all cases except September 1 9 the data show that the rate of subsidence was very gradual. Except for the last three hours of the duration of expulsion on September 22, fog or cloudiness and low wind prevailed, and the weather conditions were not favorable to rapid drying of the bark. After the rains of September 2 1 and October 20 the rate of subsidence of ascospore expulsion was relatively abrupt and rapid, and its duration was comparatively short, especially after the second rain on September 2 1 . Here, again, the relation of duration of expulsion to rapidity of drying of the bark is shown, since the rains on these dates were followed by brisk winds, and, except for the second rain of September 21, by rapid clearing and sunshine. Such weather conditions were, of course, very conducive to the rapid drying of the bark. The maximum duration of ascospore expulsion as determined by these tests after each of these rains is shown in Table XIII. In considering these data the weather conditions just described should be borne in mind. 5i6 Journal of Agricultural Research Vol. III. No. 6 Table XIII. — Maximum duration of ascospore expulsion after the cessation of rain, as determined by the examination of slides in ascospore traps at West Chester, Pa., in igi3 Date. RaJnfall. Maximum duration of spore expulsion after rain. Date. Rainfall. Maximimi duration of spore expulsion after rain. AUP" 2 7 Inches. 0-I75 .26 .68 .09 H. m. 45 6 15 5 27 13 14 Sept. 21. . .' Inches. 0-43 . 10 •73 .86 H. m. I 58 40 II 2 Sept iS 21 10 22 . . Oct. 20 3 8 It should be mentioned in this connection that the figures given in the Table XIII were secured in all cases, except that of October 20, from bark that had been protected continuously by the trap slide from the drying action of the wind, and it is possible that under such conditions the duration of expulsion may be slightly prolonged. But the data relative to the maximum duration of expulsion on October 20 were secured from bark previously unprotected by slides, since seven perithecia in five exposed areas were found to be expelling spores after expulsion had ceased in all but one area protected by the ascospore traps (Table XII). These tests prove beyond doubt that under natural conditions certain exposed areas of diseased bark do remain wet enough to cause spore expulsion fully as long as the particular areas protected by the ascospore traps. Of course, such areas would usually be in locations more or less protected from the wind or sun; but, nevertheless, they would continue to act as a source of spores for wind dissemination as long as any expulsion was in progress. The direct bearing of the results of these ascospore-trap tests upon the results obtained in the exposure plates has been brought out in the discussion of the latter topic. ASPIRATOR TESTS It has already been brought out in the historical introduction that previous analyses of air by the aspirator method under natural conditions in the field during dry weather failed to show the presence of spores of the chestnut-blight fungus (2). Positive results were obtained, however, under artificial conditions in the field, and it seems probable that failure to detect spores under natural conditions was due to the fact that most of the analyses were made during dry weather. If positive results were obtained following periods of rain, that fact was not brought out in the discussion (2). In order to obtain definite information on this point, the aspirator tests reported in the following pages were made so as to include the filtration of air immediately following periods of rain, as well as during the intervening drv weather. METHOD OF MAKING THE ANALYSIS The apparatus used in this series of tests consisted of a 4-liter aspirator bottle set on a level stump near the center of the field (fig. i). The nearest trees were 15 feet north, 19 feet east, and 33 feet west, and the Mar. 25, 1915 Dissemination of Chestnut-Blight Fungus 517 nearest lesion was on a branch 13 feet to the north. The standard sugar- tube method of making a quantitative bacteriological analysis of air was employed. The bottle was refilled with 4 liters of water at intervals of 20 or 30 minutes, thus making the aspiration practically continuous. One sugar tube was generally used each day, and the quantity of air drawn through each tube averaged 58 liters, with a maximum of 96 liters. The medium employed was a 3 per cent dextrose agar, with a reaction of +10. Ten plates were poured for each test and were incubated and the colonies counted in the same way as those in the experiments with the water spore traps (p. 520). Table XIV. — Summary of results of aspirator tests in IQI3 at West Chester, Pa. Sugar tube No. 2 3 4 5 6 7 8 9 10 II 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 Date of aspiration. Aug. 19 Aug. 20 Aug. 21 Aug. 22 Aug. 23 Aug. 24 Aug. 25 Aug. 26 Aug. 27 ....do Aug. 28 Aug. 29 Aug. 30 Aug. 31 Sept. I Sept. 2 Sept. 3 Sept. 4 Sept. s ....do Sept. 6 Sept. 7 Sept. 7 and 8. vSept. 8 Sept. 9 Sept. 10 Sept. II Sept. 12 Sept. 13 Sept. 14 Sept. 15 vSept. 16 Sept. 17 Sept. 18 do Sept. 19 Sept. 20 Sept. 21 do Sept. 22. .... . Sept. 23 Rainfall." Inches. O. 40 O O O •25 175 37 ■095 .26 Average. .68 .09 •43 . 10 •73 Quantity of air repre- sented. Number of bacteria and yeasts per liter. Liters. 12 28 36 48 68 56 56 56 47 12 68 40 60 56 92 72 72 56 68 4 76 64 12 76 60 76 72 80 68 72 72 72 80 52 64 96 76 40 28 ^2 80 26. 42 3-21 •83 2. 29 1.47 2.86 3-96 4.64 2. 12 3^75 1-35 2-5 I. 2S I. 16 1-25 I. 18 1.32 1.78 I. 69 16. 78 1. 12 4. 61 6.23 •39 .16 2. 17 •59 ■ 14 .62 .41 1. 00 7^59 • 15 1.04 •59 1.50 1.32 2.30 •87s Total number of fungi per liter. Number of spores of Endothia parasitica per liter. 57.9 2.91 4. 16 4.28 3-05 5- 21 17-65 II. 07 3-96 8-93 8-3 5-0 4.04 5-875 3^9i 7.67 38. 37 1.87 7.91 36. 60 20. 00 6.2s 8.09 3-90 8-33 4.86 I. 41 6. 90 1.94 2.58 1.91 1. II I. 18 '•25 I. 12 14-23 5-94 4.27 '•3i 4- 50 6.25 1.82 I- 31 Number of species of fungi. •35 •42 .080 •03 4 9 9 7 3 5 7 10 II 9 12 9 7 10 II 8 12 ir II 7 12 13 10 lO 5 10 7 9 6 )0 6 8 8 14 8.8 « Al! rains occurred durinc the night previous to the date of aspiration, except on Aue. a? 5i8 Journal of Agricultural Research Vol. III. No. 6 DISCUSSION OF RESULTS OF ASPIRATOR TESTS The results obtained from these tests are presented in Table XIV. The average number of bacteria per liter of air was 2.91, while the number of fungi per liter averaged 7.03. The number of fungus species repre- sented in the cultures ranged from 3 to 14. In only five instances did any colonies of the chestnut-blight fungus appear in culture, and the number of spores per liter was never large. It is not impossible that the small numbers of spores of Endoihia parasitica obtained may be due to the effect of sunlight, for in those instances where tlie rains were followed by fair weather the aspirator was exposed to the direct rays of the sun for a part of the day. This may also be the expla- nation of the fact that no spores of E. parasitica were obtained after some of the rains when ascospore-trap collections made it certain that expulsion was taking place, notably those of September 7 and 8 and September 21 and 22 (Table XV). Unfortunately there are no published investigations which give any information on the effect of sunlight on ascospores of the chestnut-blight fungus. Table XV. — Relation of aspiration tests to rainfall in 1013 at West Chester, Pa. Date of rain. Aug. 27. . . 29-30 Sept. 7-8 . . 12-13 17-18 18-19 19-20 20-21 21 . . . 21-22 Rainfall. Date of aspiration. Iiuhes. 0-175 10 37 09s 26 68 09 43 10 73 Aug. 27 30 Sept. 7-1 13 18 19 20 21 21 22 Quantity of air tested. Liters. 12 60 12 68 52 96 76 40 28 52 Number of spores of Endotliia parasitica to 10 liters of air. 4.2 O o o I. 92 o o l'2S Results with exposure plates. + + + + + + + The chief explanation of the small number of spores of Endotliia para- sitica to the liter is to be found in the small amount of air drawn through each tube. While this averaged 38 liters for those tubes yielding positive results, only a few liters were drawn through the tube in the several hours during which copious expulsion of ascospores took place. The figures given in the tables are therefore smaller than the actual number of spores per liter during the period of copious expulsion. In view of these facts. Tables XIV and XV do not represent the true number of ascospores present in the air during the time of their actual prevalence, since the period of aspiration included many hours when the}' were not prevalent, as shown by the exposure-plate tests. The rate of development of the colonies of the chestnut-blight fungus showed that they all originated from ascospores and none from pycno- spores (5). The spores obtained from sugar tube No. 7 two days after a rain may have been stray spores similar to those obtained in several exposure plates. Mar. 25. I9IS Dissemination of Chestnut-Blight Fungus 519 The aspirator tests do not appear to have given as reliable results as the exposure-plate method, since it may be noted from Table XV that negative results were obtained on certain days when the exposure plates showed that ascospores were prevalent. The importance, however, of the aspirator tests lies in the fact that ascospores were obtained under perfectly natural conditions in the field at a distance of 13 feet from the nearest lesion and that they were obtained at times when ascospore expulsion was taking place. WATER SPORE-TRAP TESTS The use of water spore traps for testing the transport of spores of the chestnut-blight fungus by the wind was the outcome of our attempts to use the method of Burrill and Barrett (3) in their study of the wind dissemination of Diplodia zeae. First, substituting a funnel for the glass plates employed by the writers just cited, an attempt was made to find some mixture which could be applied to the inner surface of the funnel and which would fulfill the necessary requirements, as follows: 1. The mixture must contain no substances toxic to spores of the chestnut-blight fungus. 2. It must spread readily and adhere to a glass surface. 3. It must be sticky, so as to retain the spores which lodge upon the surface, which is coated with it. 4. It must retain its sticky character at least 24 hours under field con- ditions. 5. It must be readily soluble in water. Glycerin of various percentages was tried alone, as well as in combina- tion with various quantities of gum arable or gelatin, but in all cases the mixtures either dried too soon or did not spread well on a glass surface. The fact that pycnospores do not germinate in water (4) suggested the substitution of dishes of sterile water for the funnels. The first idea was that analyses of the water from these dishes exposed in the field under natural conditions could be made at intervals of some days and would reveal the presence of pycnospores if they had been carried by the wind. Experience in the field, however, proved that the method was also well adapted to the study of ascospore dissemination. DESCRIPTION OF THE WATER SPORE TRAPS A water spore trap consisted of a crystallizing dish 5 cm. deep and 10 to 12 cm. in diameter, into which sterile water was introduced. The dishes were wrapped in paper and sterilized in the laboratory for transport to the field. Each dish was supported about 2 feet above the ground by a tripod of three small stakes driven into the ground. Ten-penny nails were driven into the ends of the stakes, whose ends were converged to make a support for the dish. The nails were held in proper position by a heavy cord attached to them and encircling the dish. By this means they were so firmly secured that they were never in danger of being blown out by the wind (PI. LXIV, fig. 3). After placing a dish in its proper field location, 100 to 150 c. c. of sterile water were introduced. Water for this purpose was kept in stock in small Ivrlenmeyer llasks. The dishes of water were exposed in the field in various selected loca- tions and analyses made at certain intervals (fig. 3; also PI. LXV, figs. I and 2). 520 Journal of Agricultural Research Vol. nx. No. 6 METHOD OF MAKING A TEST At the end of an exposure period the contents of each dish were emptied into sterile flasks provided for the purpose and transported to the labo- ratory at the University of Pennsylvania, where the work of making an analysis was completed. They were then replaced with other sterile dishes and sterile water introduced as before. For each water spore trap 15 to 20 plate cultures were employed, and these were made by introducing o.i to 0.5 c. c. of the water by means of a graduated i c. c. pipette into each Petri dish. In this way only 4 to 5 c. c. of the total water returned to the laboratory v/ere used in each Fig. 3. — Map showing the location of water spore-trap stations Nos. I to VT. Stations I and II are in the chestnut coppice, the detailed composition of which is shown in figure -i ; Stations III to V are at various distances from the same coppice; Station VI is to the north of a mixed chestnut and oak woodland. test, but special pains were taken to secure a uniform suspension before the removal of the quantities used. Chestnut-bark agar was used for all of these analyses (see p. 496 for formula) since experience had proved that it was a poor medium for the growth of bacteria, which were always present in some quantity. In fact, the medium is so unfavorable for the development of ordinary bacteria that in most cases the colonies remained as minute specks during the period the plates were under obser^'ation and with proper dilution offered no hindrance to the development of colonies of Endothia parasitica and other fungi. All cultures were incu- bated as nearly as possible at 25° C, and the colonies of fungi suspected of being the chestnut-blight fungus were marked at the end of three days. The count was completed on the fifth day, and any uncertain colonies Mar. 25, 191S Dissetnination of Chestnut-Blight Fuiigus 521 were transferred to 3 per cent dextrose agar for further study. In gen- eral, it may be said that transfers were not necessary, for the colonies of E. parasitica are very characteristic on chestnut-bark agar at the end of five days, if they have had sufficient room in which to develop. It was only in the case of plates badly crowded \vith other fungi that such trans- fers were necessarv. RESULTS AND DISCUSSION OF TESTS The water spore traps were exposed in or near the same plot of badly diseased chestnut trees at West Chester, Pa., which was employed for the exposure plates previously reported. Six different stations were selected for the location of water spore traps at distances varying from 15 to 389 feet from the nearest blight lesions, although Station V was 404 feet from the nearest probable source of spores. More detailed in- formation in regard to these stations is given in Table XVI. Table XVI. -Relation of water spore-trap stations to diseased chestnut trees in IOI3 at West Chester, Pa. Trap station No. Distance and direction from station to nearest lesions. Position of station with reference to diseased chest- nut trees. I II Ill IV 25 feet west and east 15 feet southwest; 19 feet north. 50 feet north; 50 feet west. . 38^ feet northwest .... Surrounded by 15- to 18-year-old coppice. Do. In cornfield 50 feet from coppice. Across cornfield from coppice. Across open field from a single tall tree. Across pasture from a single tall tree. Do. 398 feet north V 26^5 feet south Across cornfield from a single tall tree. Across cornfield from coppice. Across havfield from older forest. VI 389 feet south Table XVII. — Summary of the tests with -water spore traps in IQ13, at West Chester, Pa. Test No. Trap sta- tion No. Culture Nos. Period of exposure. Date of cultures. Total number of fungous spores. Number of spores of Endothla parasitica. I I II III III I I II III I VI IV V I IV V VI I V VI 4227-4234 4235-4242 4243-4250 4821-4835 4906-4920 4951-4960 4967-4976 4983-4992 5303-5312 5313-5322 5345-5354 5355-5364 5371-5380 5381-5390 5391-5400 54OI-5410 5411-5420 5421-5430 5431-5440 Aug. 30 to Sept 4.. Sept 5.. . . .do ". 215 13, 868 23. 239 1.236 6,375 (?) 10, 666 10. 583 7,886 0 3 4 S 6 7 8 Aug. 3 1 to Sept. 4 . . Sept. 4 to SepL 8. . . Sept. 4 to Sept 13. . Sept. 13 to Sept. 18. . . do ...do Sept 10. . Sept 16. . ...do . . do. . . . 0 0 0 1.749 3-507 2. "3 0 do ...do 9 10 Oct 13 to Oct 20. . . Oct 20, a. m.-p. m. Oct 19 to Oct 22 . . do Oct 21. .. ...do Oct. 23 . . 1,666 30 0 . . .do 13 14 IS 16 17 18 Oct. 22 to Oct. 27. . do Oct 28. . . ...do 22. 139 13. 148 4,591 113. 203 Numerous. Numerous. Numerous. 378 380 do ...do do Oct 27 to Nov. 10. . do ...do Nov. II.. . . do. . 431 0 0 19 ....do ...do 0 ;22 Journal of Agricultural Research Vol. in. No. 6 The map (fig. 3) shows the location of the coppice growth and other chestnut trees than those used in the test, with the position of the exposure stations. The character of the diseased coppice growth is shown in Plate LXV, fig. i , which shows a view taken from Station V. The older forest, which was the source of the ascospores for the traps exposed at Station VI, is shown in Plate LXV, fig. 2. The period from August 30 to November 11, 1913, was covered bv the tests presented in Table XVII. The time and amount of rainfall, and in some cases the wind direction, are necessary in interpreting the results. Table XVIII gives the rainfall for the time covered by the water spore-trap tests. Table XVIII. — Rainfall record for period covered by the water spore-trap tests in igi3 at West Chester, Pa. Date of rain. RainfaU. Bate of rain. RainfaU. Date of rain. Rainfall. Aug. 29 and 30 Sept 7 and 8 12 and 13 17 and 18 Inches. I. 10 •37 .095 .26 .68 .09 •43 Sept 21 and 22 ^ 30 Oct. I Inches. 0.83 . 02 1. 12 . 12 .06 • 73 .86 Oct. 20 24 25 26 Nov. 8 and 9. . Inches. 0.08 1.44 18 and 19 19 and 20 20 and 21 Tests Nos. 1,2, and 3 were started in the field after the rain of August 29 and 30, late in the day, and the traps were taken to the laboratory for analysis before the next rain. Judging from the results obtained from our exposure plates, no ascospores should have been present, and our failure to get any colonies of the chestnut-blight fungus in the test cultures suggests that during that period there was no wind dissemina- tion of either pycnospores or ascospores. There was a small amount of rain during the period that traps 4 and 5 were exposed, but the analyses were not made until two and three days later. Considering the fact that ascospores germinate at once in water, the failure to get any colonies of the Endothia parasitica in these tests is not surprising and again points to the absence of pycnospores. Traps 6, 7, and 8 were removed from the field a few hours after the heavy rain of September 18 and 19, and the analyses gave a large number of colonies of the chestnut-blight fungus. It appears probable that the spores were caught during the few hours following the rain, since the cultures indicated the origin of the colonies from ascopores only (5). It should be noted from Tables XVII and XVI 1 1 that traps 9 to 12 were removed from the field just following periods of rain. The vnnA was blowing from the infected trees toward trap 10 only, and this was the only one in the series which yielded the blight fungus. Traps 13 to 16 were removed from the field shortly after the rainy period of October 24 to 26, and all yielded positive results, trap 16, located 389 feet from the nearest chestnut tree, giving 431 spores. The length of time after the rain when the tests were made and the direction of the wind are the possible explanation for the negative re- sults for traps 17 to 19. Unfortunately no traps were exposed during the rainy periods of October i to 3 and October 1 1 . It is probable that the figures recorded for tests Nos. 6 to 8 and 13 to 16 represent the number of spores blown into the traps during the few hours Mar. 2s. 1915 Dissemination of Chestnut-Blight Fungus 523 following the rain. This appears to be substantiated by negative re- sults obtained during dry periods and by positive results obtained with exposure plates and slide traps just following a period of rain. The results are briefly summarized in Table XIX. It is interesting to note the number of viable ascospores of the chestnut-blight fungus that must have fallen on each square inch of water surface for the time represented. This information is presented in Table XX. Table XIX. -Summary 0/ positive results obtained from water spore traps in igij at West Chester, Pa. Station. Number of tests represented. Distance of station from nearest lesion. Total number of spores of Endotbia parasitica caught, as determined by cultures. I 2 I I I I 2 Feet. 2S IS 50 383 389 2,136 3,S07 2,113 380 820 II Ill IV V VI 461 Table XX. — Number of ascospores of Endothia parasitica falling on each square inch of water surface at various distances in igij at West Chester, Pa. Test No. Surface area of water trap. Number of spores of En- dothia parasit- ica falling on each square inch of water surface. Distance to nearest lesion. Test No. Surface area of water trap. Niunber of spores of En- dothia parasit- ica falling on each square inch of water surface. Distance to nearest lesion. 6 1::::::; 13 Sq. inches. 12-5 12. 5 16.5 139 280 169 23 Feet. 25 50 25 14 15 16 Sq. inches. 16.5 16.5 16. s 23 50 26 Feet. 383 'M The large number of spores of Endothia parasitica falling on each square inch of surface for a single rainy period certainly emphasizes the fact that healthy trees in the vicinity of badly diseased ones have innu- merable opportunities to become infected by wind-borne spores. It should be mentioned in this discussion of the results obtained by the water spore traps that there are some possibilities of error. It might be claimed that the spores found in the water traps were carried by birds or insects. This, however, appears exceedingly improbable. The cultures alwavs indicated ascospores and tests have shown that birds are carriers of pycnosporcs only (8). The position of the traps was such as to reduce the insect visitors to a minimum. Insects tested as carriers of the chestnut-blight fungus yielded both pycnosporcs and ascospores, but the former were very much more abundant (17). Besides, it was rare that any insects were found in the exposure dishes. Furthermore, spores were present in the traps only at periods following rains when other tests had indicated their prevalence. 524 Journal of Agricultural Research voi. iii.no. e CONCLUSIONS (i) As a result of 756 exposure plates made in or near the badly diseased chestnut coppice at West Chester, Pa., it can be definitely stated that ascospores of Endothia parasitica (Murr.) And. are prevalent in the air and after expidsion are carried for varying distances from their source. (2) As shown by the same exposure plates, the period of prevalence of ascospores varies with the conditions following the cessation of rains; when there is a rapid drying of the bark, this period is short, but when drying is retarded, this period is correspondingly extended. The tests indicate a general prevalence of ascospores within the first 5 hours fol- lowing the cessation of rains, with less abundance during later hours. The longest period for our entire series was 14 hours. (3) During periods of dry weather ascospores, although not generally prevalent, may occasionally be detected by the exposure-plate method. These are apparently stray ascospores expelled during some previous period of rain and now loosened from lodgment on some near-by objects. (4) In and near badly diseased chestnut groves or forests the number of ascospores falling on each square foot of exposed surface following a period of rain, as indicated by exposure plates, is very large and is suffi- cient to offer abundant opportunity for new infections. (5) Ascospores are forcibly expelled in large numbers from the peri- thecia during and after each warm rain in case the amount is sufficient to soak up the pustules. Following a dry period a rain of 0.18 to 0.25- inch has been observed to cause copious expulsion of ascospores, while rains of o.oi to o.io inch, if immediately preceded by a copious rainfall, have been sufficient to cause the resumption of spore expulsion. (6) As determined by the ascospore traps, the duration of expulsion depends on the rapidity with which the bark dries and only continues when the stromata are moist. Under natural conditions in the field the period of expulsion for eight rains varied from 45 minutes to 13 hours and 14 minutes. (7) In some cases at least the maximum of ascospore expulsion occurs after the cessation of rain. (8) The fact that the period of ascospore expulsion as determined by the ascospore traps coincides in general with the period during which spores were obtained by exposure plates points to these forcibly expelled spores as the ones prevalent following periods of rain. This is definitely substantiated by the development of colonies in the exposure plates from ascospores only. (9) It is possible to determine the presence of ascospores of the chest- nut-blight fungus in the air under natural conditions in the field by the standard aspirator method of bacteriological analysis. By this method positive results were obtained following four different rainy periods, but only when the period of aspiration included a period of copious ascospore expulsion. (10) By the use of water spore traps stationed at varying distances from diseased trees it was possible to determine that ascospores are prevalent in the air and fall upon exposed surfaces in considerable numbers, the number diminishing with the distance from the source of supply. Mar. 25. '915 Dissemination of Chestnut-Blight Fungus 525 (11) By making possible long exposures the water spore traps offered some advantages over the exposure-plate and aspirator methods. The presence of spores of the chestnut-blight fungus, however, was never shown by this method unless the period of exposure included a period of ascospore expulsion. (12) The failure to obtain colonies of the Endoihia parasitica from the water spore traps exposed during dry periods, as well as the fact that only ascospore colonies were indicated in the aspirator and exposure- plate tests, points to the conclusion that pycnospores are not generally prevalent in the air at any time. If present they certainly would be detected by the prolonged exposure of water spore traps. (13) The time immediately following a rain, when the bark is still moist, would appear to be a favorable one for new infections, since the supply of moisture would offer opportunity for germination of spores. It is a noteworthy fact that it is only during this favorable period for germination that the dissemination of ascospores takes place. (14) All of these experiments point to air and wind transport of the ascospores of the chestnut-blight fungus as one of the very important methods of dissemination and substantiate the conclusions of Rankin (15, 16) and Anderson (i, 2). It can now be said with absolute certainty that following each warm rain of any amount ascospores are carried away from diseased trees in large numbers. Since they have been obtained in large numbers at distances of 300 to 400 feet from the source of supply, the conclusion of the authors that they may be carried much greater distances is justified. During dry periods wind dissemination of ascospores does not occur at all or sinks to a very insignificant minimum. LITERATURE CITED (i) Anderson, P. J. 1913. Wind dissemination of tlie chestnut blight organism. In Phylopathol- ogy- ^- 3> no. I, p. 68. (2) and Babcock, D. C. 1913. Field studies on the dissemination and growth of tlie chestnut blight ftuigus. Penn. Chestnut Tree Blight Cora., Bui. 3, 1912, 45 p., 14 pi. (3) BuRRiLL, T. J., and Barrett, J. T. 1909. Ear rots of corn. 111. Agr. Exp. Sta. Bui. 133, p. 65-109, 11 pi. (4) Fulton, H. R. 1912. Recent notes on the chestnut bark disease. Penn. Chestnut Blight Conf., Rpt. of Proc., 1912, p. 48-56. (5) Heald, F. D. 1913. A method of determining in analytic work whether colonies of tlie chest- nut blight fungus originate from pycnospores or ascospores. In Mycologia, v. 5, no. 5, p. 274-277, pi. 98-101. (6) and Gardner, M. W. 1913. The relative prevalence of pycnospores and ascospores of the chestnut blight fungus during the winter. In Phytopathology, v. 3, no. 6, p. 296-305, pi. 26-28. Preliminary note in Science, n. s., v. 37, no. 963, p. 916-917. 1913. (7) 1914. The longevity of pycnospores of the chestnut-blight fungus in soil. In Jour. Agr. Research, v. 2, no. i, p. 67-75. (8) and Studh.vlter, R. A. 1914. Birds as carriers of the chestnut-blight fungus, /k Jour. Agr. Research, V. 2, no. 6, p. 405-422, 2 fig., pi. 38-39. Preliminary note in Science, n. s., V. 38, no. 973, p. 278-280. 1913. (9) HoDSON, E. R. 1908. Extent and importance of the chestnut bark disease. U. S. Forest Scrv., [Misc. Publ.], 8 p. 526 Journal of Agricultural Research voi. in. nos (10) KiTTREDGE, J., Jr. 1913. Notes on the chestnut bark disease (Diaporthe parasitica, Murrill) in Petersham, Mass. In Bui. Harvard Forestry Club, v. 2, p. 13-22. (11) Metcalf, Haven, and Collins, J. F. 191 1. The control of the chestnut bark disease. U. S. Dept. Agr. Farmers' Bui. 467, 24 p., 4 fig. (12) MiCKLEBOROUGH, JOHN. 1909. A report on the chestnut tree blight, the fungus, Diaporthe parasitica, Murrill. 16 p., 2 pi. (i col.). Pub. by Penn. Dept. Forestry. (13) Murrill, W. A. 1906. A serious chestnut disease. In Jour. N. Y. Bot. Gard., v. 7, no. 78, P- 143-153. fig- 13-19- (14) Rankin, W. H. 1912. The chestnut tree canker disease . In Ph)rtopathology, v. 2, no. 2, p. 99. (15) 1912. How further research may increase the efficiency of the control of the chestnut bark disease. Penn. Chestnut Blight Conf., Rpt. of Proc, 1912, p. 46-48. (16) 1913. Some field experiments with the chestnut canker fungus. In Phytopa- thology, V. 3, no. I, p. 73. (17) Studhalter, R. A. 1914. Insects as carriers of the chestnut blight fungus. In Phytopathology, V. 4, no. I, p. 52. PLATE LXIII Fig. I. — Petri-dish culture 5044 from 12 minutes' exposure of chestnut-bark agar, made on September 20, 1913, 2 hours and 8 minutes after the cessation of a rain, at station 51, located 27 feet from the nearest lesion. Fig. 2. — Petri-dish culture 5041 from 16 minutes' exposure of chestnut-bark agar, made on September 20, 1913, i hour and 55 minutes after the cessation of a rain, at station 49, located 414 feet from the source of the spores. Ten of the twelve colonies are those of Endothia paiasiiica. Dissemination of Chestnut-Blight Fungus Plate LXIll Journal of Agricultural Research Vol. Ill, No. 6 Dissemination of Chestnut-Blieht Fungus Plate LXIV Journal of Agricultural Research Vol. Ill, No. 6 PLATE LXIV Fig. I. — Ascospore trap 51. This consists of a wooden bracket which supports an object slide over perithecial pustules. Fig. 2. — Ascospore trap 52. Fig. 3. — Water spore trap located at Station V. The trap consists of a crystallizing dish containing sterile water and is supported on a tripod. PLATE LXV Fig. I. — View looking towards the coppice growth from water spore-trap Station V. The trees in the background at the right are at the end of the plot shown in text figure I. Fig. 2. — View of a mixed chestnut and oak grove taken from water spore-trap Sta- tion VI. This grove was the source of the spores of the blight fungus caught at Sta- tion VI. Dissemination of Chestnut-Blight Fungus Plate LXV Journal of Agricultural Research Vol. Ill, No.6 INDEX Page. Ability of Colon Bacilli to Survive Pasteuriza- tion (paper) 401-410 Achillea lanulosa — forage value of 97 viability of seed of, in range lands 106 Acid, phosphoric, in Oryza saliva 425-430 A oastache urtici folia — forage value of 97 viability of seed of, in range lands 106 Affrilus — bilmeatus — control of 292-293 ecology of 284 enemies of 292 life history of 283-294 spp., habits of 184 vittaticollis — control of 184-185 description of burrows of 181 enemies of , 184 habits of 183 host plants of 180 life history of 181-183 Agoseris glauca — forage value of 97 viabihty of seeds of, in range lands 106 A gropyron — flexuosum, forage value of 97 spicalutn, viability of seed of, in range lands 106 spp.— characters of lemma of 279 characters of palea of 280 characters of rachilla of 278-279 identification of seeds of 275-282 seed characters of 276-277 shape of seed of 278 water requirement of 44,46,52,60-62 violaceutn, forage value of 97 A grosiis rossae — forage value of 97 viability of seed of. in range lands 106 Air and Wind Dissemination of Ascospores of the Chestnut- Blight Fungus (paper) 493-526 Alfalfa. See Medicago. Alfalfa hopper, three-cornered 343-363 Allard, H. A. (paper), Effect of Dilution upon the'lnfectivity of the Virus of the Mosaic Disease of Tobacco 295-299 Allard et al. (paper). Oil Content of Seeds as Affected by the Nutrition of the Plant 227-249 Allium validum — forage value of 97 viability of seed of, in range lands 106 AUernaria spp., growth in butter. . 302-303,3051307 Amaranthus — graecizans, water requirement of 60,62 retroflexus, water requirement of 47.49,53162 Ambrosia artemisi folia, water requirement of . 47-4S, 53,60.62 Amygdalus persica, food plant of Ceraiitis caPitata 3 14 78745''-i5 7 Page. AndropoQon sorghum — water requirement of 20-25, Si> 53, 55-56,58,61-62 aethiopicus, water requirement of ao, 22-23,36-37, 39, 51, 53, 6j Aphid — clover 431-433 long-beaked clover. See Aphis brevis. short-beaked clover. See Aphis bakeri. A phis — bakeri — description of 433 host plants of 433 brevis — description of 43 1-433 host plants of 431-433 Apple. See Malus. Apple Root Borer (paper) 179-186 See also Agrilus vitlaiicollis. Arachis hypogaea, oil content of seed of 244-245 ArgioPe transversa, enemy of Stictocephala festina 359 Arm.illaria tnellea, relation to injury by Ag- rilus bilineatus 284-285 Armsby, H. P.. and Fries, J. A. (paper), Net Energy Values of Feeding Stuffs for Cattle 435-493 Artemisia frigida, water requirement of.. 43,50,60,62 Ash analyses of leaves and twigs of Carya illinoensis 165-166, 169 Aspen. See Populus tremuloides. Aspergillus repens, growth in butter 307 Assimilation of Colloidal Iron by Rice (paper) 205-210 Atanycolus sp., parasite of Agrilus bilineaius . 292 Avena saliva — host plant of Stictocephala festina 346 water requirement of 11-12, 5o-53»5S-56»59.6i-*2 Ayers, S. H., and Johnson, W. T.. jr. (paper), Ability of Colon Bacilli to Survive Pasteuri- zation 401-410 Bacillus colt — ability to survive pasteurization 401-410 effect of heat on 403-40S presence of, index of efficiency of pasteuri- zation 403-408 thermal death point of 402-406 Back. E. A., and Pcmbcrton, C. E.— Life History of the Mediterranean Fruit Fly from the Standpoint of Parasite Intro* ductiou (paper) 363-374 Life History of the Melon Fly (paper) . . . 269-274 Susceptibility of Citrous Fruits to the Attack of the Mediterranean Fruit Fly (paper) 3"-33o Bactroccra cucurbitae — comparison with Ceraiitis caPitata 370-37r host plants of 269 life history of 269-374 Banana. Sec Musa sapientum. 5»7 528 Journal of Agricultural Research Vol. Ill Page. Barley- See Hordeum. Bean — horse. See Viciafaba. Mexican. See Phascolus vulgaris. navy. See Phascolus vulgaris. soy. See Glycine hispida. wild soy. See Glycine soja. Bearce. H. W. (paper). Studies in the Expan- sion of Milk and Cream 251-268 Beardtongue. blue. See Pentstemon procerus. Beet, sugar. See Beta vulgaris. Beta vulgaris, water requirement of- 16, 50, 55, S9. 61 Bird enemy of StictocePkala festina 360 Blight fungus, chestnut. See Endothia para- sitica. Bluegrass — little. See Poa sandbergit. mountain. See Mclica spp. Boebera pa pposa. water requirement of 47-"4S. 52,60,62 Borer — apple root 179-1S6 oak. See Agrilus bilineatus. two-Uned chestnut. See Agrilus bilineatus. Bolryosphaeria vtarconii, n. comb., causal or- ganism of fungous disease of Cannabis saliva . 83 BouUloua gracilis, water requirement of 44-45. 53,60,62 Brachysm— a Hereditary Deformity of Cotton and Other Plants (paper) 3S7-400 comparison with nanism 387 description of 3S7-395 origin of 389 relation to homoeosis 396~398 relation to hybrids 395~396 Brassica spp., water requirement of 4i~42. 52-53. 59. 61 Briggs, L. J., and Shantz. H. L- (paper). Rela- tive Water Requirement of Plants 1-64 Brome-grass. See Broraus. Bromus — hordeaceus, viability of seed of, in range lands 106 inermis, water requirement of . 44-46,51,60-61 marginatus — forage value of 97 viability of seed of, in range lauds 106 Brooks, F. E. (paper). Apple Root Borer. . 179-1S6 Buckwheat. See FagoPyrum fagopyrum. wild. See Polygmiuvi phytolaccaefolium. Budding, effect of, on pecan rosette . . . 158-159, 167 Buffalo grass. See Bulbilts daclyloides. Bulbilis dactyloides, water requirement of ... . 44-45. 53,60.62 Butter— moldiness in 301-310 moldy, analyses of 301 types of mold in 302-303 Butterweed. See Senecio triangularis. Cabbage. See Brassica. Cala magroslis — canadensis, viability of seed of, in range lands 106 rubesccns, viabihty of seed of, in range lands. 106 Calophyllum inophyllum, host plant of Cera- litis captiaia 3i3-3i4'3i6.364 Page. Cannabis sativa. fungous disease of 81-84 Cantaloupe. See Czuumis vielo. Capriola dactylon, food plant of Stictocepftala festina 346 Capsella bursa- Pastor is, host plant of Aphis bakeri 433 Carbohydrate, transformation of, in Ipomoea batatas during storage 336-339 Carbonate, soil, decomposition of 79-80 Carex — exsiccata — forage value of 97 viability of seed of. in range lands 106 geycri — forage value of 97 viabihty of seed of. in range lands 106 hoodii — forage value of 97 viabihty of seed of. in range lands 106 illota— forage value of 97 viability of seed of, in range lands 106 Carica papaya, host plant of Ccraiitis capi- lata 316 Carophyllus jambos, host plant of Ceratiiis caPitata 3^4 Carrero. J. O.. and Gile, P. L. (paper). Assimi- lation of Colloidal Iron by Rice 305-210 Carya illinocnsis. ash analyses of leaves and twigs of 165-166,169 Castanea deiitaia, host plant of Agrilus bili- neatus 283, 284 Cattle, net energy values of feeding stuffs for 435-49* Celery, wild. See Ligusticum oreganum. Ceratatis caPitata — comparison with Bactrocera cucurbitae. . . 370-371 control of 328 effect of oil from rind of Citrus spp. on 317-319 habits of 314-31S host-fruit preference of 3^4 host plants of 312-314 injury to citrous fruits by 323-328 life history of 3^3-374 mortality among eggs of 315-319 mortaUty among lar\'ae of 315-317-319-322 susceptibihty of citrous fruits to the attack of 311-330 Chaetochloa iialica, water requirement of 26-27, 38-39- Si> S3. 55-56- 58,61-62 Changes in Composition of Peel and Pulp of Ripening Bananas (paper) 187-203 Chapman, R. N. (paper). Observations on the Life History of Agrilus Bihneatus — 283-294 Chamacnerion angustifolium — forage value of 97 viability of seed of. in range lands 106 Charles. V. K., and Jenkins, A. E. (paper), A Fungous Disease of Hemp 81-84 Chenopodium album, water requirement of. . 47-481 52*63 Chestnut-bhght fungus. See Endothia para- siilbtu Chestnut borer, two-lined. See Agrilus bilineatus. Oct., i9x4-Mar., 1915 Index 529 Page. Chrysophyllum cainito, food plant of CeratUis capHata 314-31S Chick-pea. See Cicer arietmum. Cicer arietinum, water requirement of 28.30, 50-51.59 Cinna UUifolia, viability of seed of, in range lands 106 Ciirullus vulgaris — host plant of Bacirocera cucurbitae 369 water requirement of 40-41)52-53,59,61 Citrus spp. — effect of oil from rind of, on eggs of Ceraii- tis capitata 317-319 host plants of Ceratitis caPitata 311-330 Climate- effect on development of Ceralitis capitata. 324-32S effect on oil content of seed 245 infience on development of forage plants. 109-115 Clover- bur. See Medicago denliculata. crimson. See Trifoltum incarnatum. red. See Tnfolium pratense. sweet. See Melilotus alba. yellow sweet. See Melilotus officinalis. Clover aphid. See Aphis. "Cluster," cotton, agricultural defects of 39S Cocklebur. See Xantfiitim commune. Collins, G. N. (paper). A More Accurate Method of Comparing First-Generation Maize Hybrids with Their Parents 85-91 Colon bacilli, ability to survive pasteuriza- tion 401-410 Coloring Matter of Raw and Cooked Salted Meats (paper) 211-226 Coneflower. See Rudbeckia occidentalis. Cook, O. F. (paper), Brachysm, a Hereditary Deformity of Cotton and Other Plants. . ^87-400 Com. See Zea mays. Cotton— "cluster," agricultural defects of 398 See also Gossypium. Cowpea. See Vigna sinensis. Crataegus spp., host plant of Aphis brevis. . 431-432 Cream — effect of temperature upon density of 254-262 effect of temperature upon volume of . . 262,265-268 expansion of 251-268 Cnidfer, water requirement ol 41-42, 59»6i Cucumber. See Cucumis sativus. Cucumis — melo— host plant of Bactrocera cucurbitae 269 water requirement of 40-41, 52-53* 59» 61 sativus — host plant of Bactrocera cucurbitae 269 water requirement of AO,s^53tS9>6i Cucurbit, water requirement of 40-41.61 Cucurbita — maxima, water requirement of 40« 52- 59- 61 pepo— host plant of Bactrocera cucurbitae 269 water requirement of 40, 52, 59, 61 Cunntnghamietla sp.. growth in butter 307 Curtis. M. R. (paper). Relation of Simultane- ous Ovulation to the Production of Doublc- yolkcd Eers 375-386 Page. Curve, logarithmic — fitting by the method of moments 4x1-423 moments of 412-417 use of, in agricultural and biological inves- tigations 411-412 Cydonia japonica, host plant of Aphis brevis. . 431 Dahlberg, R. C. (paper). Identification of the Seeds of Species of Agropyron 275-282 DandeUon, mountain. See Agoseris glauca. Decomposition of Soil Carbonates (paper). . . . 79-80 Deschampsia — caespitosa. viabihty of seed of, in range lands 106 eltmgata, viability of seed of. in range lands. 106 spp., forage value of 97 Dirkinus giffardii, parasite of Ceratitis capitata. 363 Disease, mosaic, of tobacco, effect of dilution upon the incfectivity of the virus of 295-299 Durra. See AndroPogon sorghum. Dwarfing, true, of plants, comparison with brachysm 387 Effect of Dilution upon the Infeclivity of the Virus of the Mosaic Disease of Tobacco (paper) 295-299 Egg— double-yolked — origin of 376-582 ovarian relation of the two follicles fur- nishing the yolks for 380-382 relation between rate of fecimdity and type of doubling of 379-380 relauon of simultaneous ovulation to tt^ producrion of 375-386 relation of the nature of the doubling to the functional divisions of the ovi- duct 376-37S types of 375-376 ovarian doubling of yolk of 382-384 Elymus glaucus, viability of seed of, in range lands 106 Emmer. See Triticum. Endothia Parasitica — detection of ascospores of, by ascosiwre traps 512-516 detection of ascospores of, by aspirator tests 516-519 detection of ascospores of, by exposure plates 496-51 J detection of ascospores of. by water spore traps 519^5*3 dissemination of ascospores of 493-526 Energy— dicmical — influence of quantity of feed consumed on losses of 445-447 loss in feed 439-449 expenditure consequent upon feed con- sumption 453-483 mctabolizab le — determination of. in feed 450-453 in feed 439^453 variability in feed 449~4So Environment, effect on oil content of seed. 239^341 Eriobotrya jaPonica, host plant of Ceratitis capUala 313 Errata IV 530 Journal of Agricultural Research Vol. in Page. Erythraeus sp., enemy of SliUocephala festina . 359 Euchlaena mexicana.-watev requirement of. . 17-20, S3.s8,6i FaQoPyrumfagoPyrum. water requirement of. 59,61 Feces, loss of chemical energy of feed in — 440-449 Feed- composition of dry matter of 43T-438 computation of net energy value of 4S3-488 influence of. on heat production in cattle, 456-458 influence of individuaUty on losses of chemi- cal energy' of 447~449 influence of quantity consumed on losses of chemical energy of 445-447 loss of chemical energy of 439-453 metabolizable energy of 439-453 net energy values of, for cattle 435-492 relation of. to heat production of cattle in various positions 453-482 Fertilizer- effect on oil content of seed : 245-247 effect on pecan rosette 159-162, 168 Festvca viridula — forage value of 97-127 viability of seed of 106-107 Feterita. See Andropogon sorghum. Fireweed. See Chatnaenerion angustifolium. Fitting Logarithmic Cur\'es by the Method of Moments (paper) 411-423 Flax. See Litium usilalissimum. Fly, Mediterranean fruit. ^^eeCeratitis capitata. Fly, melon. See Bactrocera cucurbitae. Forage plant- development under yearlong protection. , 121-125 factors influencing establishment of repro- duction on range lands 109-115 influence of physical conditions on devel- opment of 109-115,143 loss of, in range lands 142 production of, under deferred grazing 125-143 reproduction of. imder yearlong grazing. . 11S-119 Foubert, C. L.. et al. (paper). Oil Content of Seeds as Affected by the Nutrition of the Plant 228-249 Foxtail, white. See Sitanion veluiinuni. Fries. J. A., and Armsby. H. P. (paper), Net Energy Values of Feeding Stuffs for Cattle 435-492 Fruit fly. Mediterranean. See Ceratilis capi- tala. Fungus, chestnut-blight. See Endolkia para-, silica. Fungous Disease of Hemp, A (paper) 81-84 Fusarium sp.. growth in butter 307 Galesus sihestrii, parasite of Ceraiitis capilala. 363 Gardner, M. W.. et al. (paper). Air and Wind Dissemination of Ascospores of the Chest- nut-Blight Funi^is 493-526 Gamer, W. W., Allard, H. A., and Foubert. C. L. (paper). Oil Content of Seeds as Af- fected by the Nutrition of the Plant 227-249 Geranium viscosissiinutn, viabiUty of seed of, in range lands 106 Gile. P. L., andCarrero, J. O. (paper). Assimi- lation of Colloidal Iron by Rice 205-210 Page. Glycine — hispida- host plant of Stictocephala festina 346 oil content of seed of 230-241 water requirement of 30~3 1- 34- 5^7 S9> 61 soja, water requirement of . . 3c>-3i'34-52-53r59»6i Gore, H. C. (paper), Changes in Composition of Peel and Pulp of Ripening Bananas. . . 187-203 Gossypium spp. — hereditary deformity of 387-400 oil content of seed of 230-^32, 239, 242, 246 water requirement of ifr-i7) 50t 52-561 59, 61 Grafting, effect on pecan rosette is8r-i59, 167 Grain, water requirement of 5-8-27,37-39,50-^2 Grama grass. See Bouteloua gracilis. Grapefruit. See Citrus. Grass — Bermuda. See Capriola dactylon. brome. See Bromus inermis. brome, short-awned. See Bromus mar- gin atus. buffalo. See Bulbilis dactyloides. bimch. See Agropyron; Festuca. elk. See Carex geyeri. grama. See Boutel-oua gracilis. hair. See Deschampsia. Johnson, See Sorghuvi halepense. Uttle needle. See Siipa minor. mountain bunch. See Festuca viridula. mountain wheat. See A gropyron violaceitnt. needle. See Stipa occidcntalis. onion. See Melica. red bxmch. See Agropyron fiexuosum. reed. See Ciuna latifolia. Sudan. See Andropogon sorghum aetki- bpicus. tall meadow. See Panicularia nervata. tall swamp. See Carex exsiccata. water requirement of 43-46,52-53,60-62 western porcupine. See Slipa occidenialis. wheat. See Agropyron. Grazing^ deferred- advantages of. in range lands 143-14S influence on reproduction of forage plants in range lands. 125-146 selection of lands for 14S relation of. to growth of forage plants 115-146 relation of, to revegetation 115-146 season-long. See Grazing, yearlong. yearlong — influence on reproduction of forage plants 11&-135 on range lands "6 Grindclia squarrosa, water requirement of . . 43 1 So> 60 Guava. See Psidium. Gumweed. See Grindelia squarrosa. Hasselbring, H., and Hawkins, L- A. (paper),- Physiological Changes in Sweet Potatoes during Storage 33 i-34a Hawkins, L. A., and Hasselbring, H. (paper). Physiological Changes in Sweet Potatoes during Storage 331-342 Hawthorn. See Crataegus. Oct., Z9i4-Mar., 1915 Index 531 Page. Heald. F. D., Gardner, M. W., and Studlial- ter. R. A. (paper). Air and Wind Dissemi- nation of Ascospores of the Chestnut- Blight Fungus 493-526 Heart-Rot of Oaks and Poplars Caused by Polyporus Dryophilus (paper) 65-78 Heat- effect of, on Bacillus colt 403-408 influence of various feeds on production of, in cattle 453-482 loss of, in cattle, consequent on feed con- sumption 453-482 Hedgcock, George G., and Long. W. H. (paper), Heart-Rot of Oaks and Poplars Caused by Polyporus Dryophilus 65-78 Heliantkus spp. — host plant of Stictocepkata festma 346 water requirement of 47-48, 52, 53, 60, 62 Hellebore, false. See Veratrum viride. Hemochromogen. nitric-oxid. coloring matter of cooked salted meats 221-224 Hemoglobin, nitric-o.^id, coloring matter of raw salted meat 212-221 Hemp. See Cannabis sativa. Hieracium cynoglossoides — forage value of - 97 viability of seed of, in range lands 106 Hoagland. Ralph (paper), Coloring Matter of Raw and Cooked Salted Meats 211-226 Hordeum spp. — host plants of Stictocephala festina 346 water requirement of 13-14,50-51.55,59,61 Horse bean. See Viciafaba. Horsemint. See Agasiache urtici folia. Humidity, effect on mold growth in but- ter 304-306, 308-309 Identification of the Seeds of Species of Agro- pyron (paper) 275-282 iPomoea batatas — carbohydrate transformations in, during storage 336-339 physiological changes in, during storage. . 331-342 Iron, colloidal, assimilation of, by rice 205-210 Jacob's- ladder. See Polemoniutn humile. Jenkins, A. E., and Charles. V. K. (paper). Fungous Disease of Hemp 81-84 Johnson, W. T.. jr., and Ayers. S. H. (paper), AbiUty of Colon Bacilli to Survive Pasteuri- zation 401-41C Juncoides glabratum — forage value of 97 viability of seed of, in range lands 106 Juncus spp., forage value of 97 Kafir. See AtniroPogon sorf/htim. Slamani — ball. See Calopkyllum inophyllum. winged. See Tcrminalia cataPpa. Kaoliang. See Andropogon sorghum. Koeleria cristata, viability of seed of, in range lands 106 Lamb's-quarters. See Chenopodium album. Latkyrus odoratus. host plant of Aphisbrevts. . 432 Legumes, water requirement of 27-39,55-56,61 Page. Life History of the Mediterranean Fruit Fly from the Standpoint of Parasite Introduc- tion (paper) 363-374 Life History of the Melon Fly (paper) 269-274 Ligusiicum oreganunt, viabiUty of seed of. in range lands 106 Linum usitatissimum, water requirement of. 15-16, 52-53,59.61 Logarithmic curves, use of 417-421 Long. W. H., and Hedgcock, G. G. (paper). Heart-Rot of Oaks and Poplars Caused by Polyporus Dryophilus 65-78 Loquat. See Eriobotrya jaPonica. LyccPersicon esculcntum. food plant of Sticto- cephala festina 345 Maclntire, W. H. (paper), Decomposition of Soil Carbonates 79-80 Maize, comparison of first-generation hybrids with parents 85-91 Malus spp., host plant of Aphis bakeri 433 Mangi/cra indica, host plant of Ceratitis caPi- '"^'a 313-314.364 Mango. See Mangifcra indica. Marigold, fetid. See Boebera paPposa, Meat, salted, coloring matter of 211-229 Medicago — dcnticiilata, host plant of SlictocePhala festina 346 saliva — host vX^ntoi Stictocephala festina 344-346, 350,357-359 injury to, hy Stictocepkata festina 357-359 spp., water requirement of 27-39, SO,52,S5-s6,6oHS2 Mediterranean fruit fly. See Ceratitis capitata. Melica — bella— forage value of 97 viability of seed of. in range lands 106 spp. , forage value of 97 Melilotus — alba, water requirement of 27-30,50,55,59,61 oJ0icinalis, food plant of Stictocephala fes~ tina 346 Melon — musk. See Cucumis tnelo. water. See Cilrullus vulgaris. Melon fly. See Bactrocera cucurbitae. Membracis festina, syn. Stictocephala festina. Mesquite. See Prosopis juliflora. Metabolism, in cattle, influence of position on 453-482 Methane — loss of chemical energy of feed in 44(^449 cjuantity in various feeds 450 Mexican bean. See Phaseolus iitlgaris. Milk— and cream, expansion of 251-268 effect of temperature upon density of 254-263 effect of temperature upon volume of . 262,265-268 Millet. Sec Chaetochloa italica 26-2 7 Milo. See Andropogon sorghum. Moisture, soil, relation to loss of forage plants in range lands 142 532 Journal of Agriculiural Research VoI.IU Mold— Page. black, growth in butter 303,305,307 green, growth in butter 303» 305. 307-308 red, growth in butter 303» 307-308 relation of humidity to growth in butter. 304-306, 308-309 smudged, growth, in butter 302-303.305 Moldiness in Butter (paper) 101-310 Moments, fitting logarithmic curves by method of 411-423 More Accurate Method of Comparing First- Generation Maize Hybrids with Their Parents, A (paper) 85-91 Mosaic disease. See Disease, mosaic. MucoT sp., growth in butter 303>30S'307 Mumsops elangi. host plant of Ceratitis cap- itata 364 Musa sapientuni — changes in composition of 187-203 changes in composition of, in ripening . . . 187-203 composition of igor-igS respiration of i99 Muskmelon. See Cucumis melo. Nanism, comparison with brachysm 387 Natural Revegetation of Range Lands Based upon Growth Requirements and Liie His- tory of the Vegetation (paper) 93-14S Net Energy Values of Feeding Stuffs for Cattle (paper) 435-492 Nitrogenous Soil Constituent, A: Tetracar- bonimid (paper) 175-178 Oak. See Quercus. Oak borer. See Agrilus btUneatus. Oats. See Avena sativa. Observations on the Life History of Agrillus Bilineatus (paper) 283-294 Oidium laciis. growth in butter 303*305.307-308 Oil Content of Seeds as Affected by the Nutri- tion of the Plant (paper) 227-249 Orange. See Citrus. Organic Phosphoric Acid of Rice (paper). . 425-430 Orton, W. A., and Rand, F. V. (paper). Pecan Rosette 149-174 Oryza sativa — growth of. with dialyzed iron 206-209 growth of, with ferric chlorid 206-209 organic phosphoric acid of 425-430 water requirement of 15-50, 52-56>59.6i Ovulation, siniultaneous, relation of, to the production of double-yolked eggs 375-386 Onion — mountain. See Allium validum. wild. See Allium. Opius humilis, parasite of Ceratitis capitaia. . . 363 Panicvlaria nerxata — forage value of 97 viability of seed of. in range lands 106 Panicum miltaceum, water requirement of.. 26-37, 3&-37. 39. 51. 58. 61-62 Papaya. See Carica papaya. Pasteurization, ability of colon bacilli to sur- vive 401-410 Patch, E. M. (paper). Two Clover Aphids. . 431-433 Pea — Page. Canada field. See Pisum sativum. sweet. See Lathyrus odoratus. Peach. See A mygdalus perstca. Peanut. See Arachis kyPogaea. Pearl, R., on use of logarithmic curves in biological and agricultural investigations. 411-412 Pecan. See Carya illinoensis. Pecan Rosette (paper) 149-174 Pemberton, C. E., and Back, E. A.— Life Histor>' of the Mediterranean Fruit Fly from tlie Standpoint of Parasite In- troduction (paper) 363-374 Life History of the Melon Fly (paper) . . . 269-374 Susceptibility of Citrous Fruits to the At- tack of the Mediterranean Fruit Fly (paper) 311-330 Penicilliumspp., growth in butter. . 303,305,307-308 Pentsteman procerus — forage value of 97 viabihty of seed of. in range lands 106 Phaseolus vulgaris, water requirement of ... . 30, 34-35.52-53.59.61 Pkleum alpinum, viability of seed of, in range lands 106 Physiological Changes in Sweet Potatoes during Storage (paper) 331-342 Pigweed. See Amaranthus retrofiexus. Pisum sati'imm, water requirement of 30, 34-35. 52-53. S9» 61 Plants- effect of nutrition on oil content of seeds . . 227-349 effect of varietal differences on oil content of seeds 237-^39 relative water requirement of x-64 Poa sandbergii — forage value of 97 viability of seed of, in range lands 106 Polemonium kumile — forage value of 97 viability of seed of, in range lands 106 Polygonutn phylolaccacfolium — forage value of 97 viability of seed of, in range lands io9 PolyPorus dryoPkilus — distribution of 72-75 causal organism of heart-rot of oaks and poplars 65^8 description of sporophores of 70-72 Poplar. See Populus. Popuius — spp. , heart-rot of 65-78 tremuloides, host plant of Polyporus dry- opkilus 66,68-71,73-75 Portulaca oleracca. water requirement of 47. 49»S3»6o,62 Potato- Irish. See Solatium tuberosum. sweet. See Ipomoea batatas. Prosopis glandttlosa, food plant ol SiictocePhala festitta 346 Proso. See Panicum miliateutn. Pruning, effect of, on pecan rosette 152, 168 Prunus spp.. host plant of Aphis brevis 431 Psidium spp., host plant of Ceratitis capitata. 324,364 Pumpkin. See Cucurbita Pepo. Purslane. See Portulaca oleracea. Oct., 1914-Mar., 1915 Index 533 Quercus spp. — Page. heart-rot of 65-78 host plants of Agrilus bdinealus 28^-284 Quince, Japan. See Cydonia jaPontca. Ragweed, western. See Ambrosia ariemisi- folia. Rand, F. V.. and Orton, W. A. (paper). Pecan Rosette 149-174 Range land — character and distribution of vegetation. . . 95-100 dissemination of seed crop of 108-109 factors influencing establishment of repro- duction of forage plants of 109-115 flower-stalk production in 102-104 life history of forage plants of 101-115 management of, during revegetation 146 period of seed maturity of 104-105 revegetation of 93-148 selection of, for deferred grazing 145 systems of grazing on 115-146 viability of seed crop on 105-108 Rape. See Brassica. Rcdtop. alpine. See Agrostis rossae. Relation of Simultaneous Ovulation to the Production of Double-YolkedEggs(pa per). 375-386 Relative Water Requirement of Plants (paper) 1-64 Revegetation of range lands 93-148 Rhizopus nigricans, growth in butter 306-307 Rice — assimilation of coUoidal iron by 205-210 organic phosphoric acid of 425-430 See also Oryza saliva. Root borer, apple 179-1S6 Rose-apple. See Carophyllus Jambos. Rosette, pecan — comparison with other diseases 169-171 control of 172-173 distribution of 149-150 effect of budding on 158-159, 167 effect of fertilizer on iso-162,168 effect of grafting on- - 158-159, 167 effect of priming on 152, 16S effect of spraying with Bordeaux mixture on 162 effect of transplanting on 152-155, 167-16S effect on germination of nuts 155-156 inoculation experiments with 156-157, 166-167 isolation of microorganisms of 156-157* ^67 nature of 171-172 orchard observations of 163-165, 168 parasitism of 166-167 symptoms of 150-151 virulence of 150-151 Rot- heart, of oaks and poplars caused by Poty- porus dryopkilus 65-78 piped — control of 76 in Populus spp 65-78 in Pofmliis tremuloides, characters of 68-70 in Qucrcus spp 65-78 in Qjiercus spp., characters of 66-68 Rudhcckia occidcntalis — forage value of 97 viability of seed of. in range lands 106 Rush. See Juncus spp. Page. Rush, wood. See Juncoides glabratum. Rye. See Secale cerealc. Rye, smooth wild. See Elymus glaucus. Sage, mountain. See Artemisia frigida. Salix nuttallii. forage plant of 97 Salsola pestifer, water reciuirement of 60, 62 Salt, effect of, on mold growth in butter. . 304-309 Sampson, A. W. (paper), Natural Revegeta- tion of Range Lands Based upon Growth Requirements and Life History of the Vege- tation 93-148 Saperda Candida, host plants of 180 Secale cereale, water requirement of 14, 50-51. 55. 59. 61 Seed- effect of climate on oil content of 345 effect of environment on oil content of . . . 239-241 effect of fertilizer on oil content of 245-247 effect of length of growing period on oil content of 236-237 effect of nutrition of plant on oil content of 227-249 effect of partial defoliation on oil content of 232-234 effect of partial removal of seed pods on oil content of 234-235 effect of size on oil content of 235 effect of soil on oil content of 241-245 oil content of, at successive stages of devel- opment 230-232 varietal differences in oil content of 237-239 Senecio triangularis, forage value of 97 Shantz, H. L.. and Briggs. L. J. (paper). Relative Water Requirement of Plants 1-64 Shaw. R. H.. and Thom. C. (paper), Moldi- ness in Butter 301-310 Shepherd "s-purse. See Capsella buTsa~Pas~ toris. Shorey. E. C, and Walters. E. H. (paper), A Nitrogenous Soil Constituent: Tetracar- bonimid 175-178 Sitanion velutinum, viability of seed of. in range lands ro6 Skunkwecd. See Polemoniutn kumilc. Soil carbonate — decomposition of 79-So effect on oil content of seed 241-345 Solatium tuberosum, water requirement of. 4i**43 Sa-53>s6»S9.6i Sorghum. See AndroPogon sorghum. Sorghum kaUpense, food plant of Slictocephala festina 346 Soy bean. See Glycine hispida. Soy bean. wild. See Glycine soja. Spraying, effect on pecan rosette 162 Squash. Hubbard. See Cucurbiia tnaxima. Star-apple. See Ckrysophyllum cainito. Siictoccphala festina — control of 361 description of 346-34S distribution of 344-345 effect of altitude on distribution of 345 eneniies of 3Ssr-36o habits of 348-357 host plants of 345-346 life history' of 34ft-3S7 534 Journal of Agricultural Research Vol. Ill Stipa— minor — Page. forage value of 97 viability of seed of, in range lands io6 occidentalis — forage value of 97 viability of seed of. in range lands io6 Studhaltcr. R. A., et al. (paper). Air and Wind Dissemination of Ascospores of the Chestnut-Blight Fungus 493-526 Studies in the Expansion of Milk and Cream (paper) 251-268 Sudan grass. See A ndroPogon sorghum aethi- opicus. Sedge. See Carex spp. Sugar beet. See Beta vulgaris. Sunflower. See Helianthus spp. Susceptibility of Citrous Fruits to the Attack of the Mediterranean Fruit Fly (paper). . 311-330 Tansy, wild. See Achillea lanutosa. Teosinte. See Euchlaena mexicana. Terminalia catappa, host plant of Ceratitis capitata 316, 364 Tettacarbonimid, a nitrogenous soil constit- uent 175-178 Tetrasiichus giffardii, probable parasite of Ceraitits capitata - 3^3 Thistle, Russian. See Salsola peslifer. Tbom, C, and Shaw. R. H. (paper), Moldi- ness in Butter 301-310 Thompson, A. R. (paper). Organic Phos- phoric Acid of Rice 425-430 Three-Cornered Alfalfa Hopper (paper) 343-362 Timothy, alpine. See Pldeum alpinum. Tobacco, effect of dilution upon the infectiv- ity of the virus of the mosaic disease of 295-299 Topography, relation of, to loss of forage plants in range lands 142 Transplanting, effect of. on pecan rosette. . 152-155, 167-168 Trickoderma sp., growth in butter 307-308 Trichogrammidae, parasites of A grilus bilinea- tus 29-J Tri/olium — incar7iatum, water requirement of 30. 33>35.52»59j6i praiense — host plant of Aphts bakeri 433 host plant of Slictocephala festina 346 spp.. host plant of Aphis brevis 431 Trisetum spicalum, forage value of 97 Triticum spp. — effect of fertilizer on water requirement of. .Sj 50-51 effect of screened inclosure on water re- quirement of 3 host plant of St iciocephala festina 346 water requirement of 8-10, 50-56, 58, 60-62 Tumbleweed. See Amaranthus graecizans. Turnip. See Brassica. Two Clover Aphids (paper) 431-433 Urine, loss of chemical energy of feed in . . .^ . 440-449 Page. Use of logarithmic curves in biological and agricultural investigations 41 1-413 Wallowa National Forest, topography and soil 94-95 Walters, E. H.. and Shorey. E. C. (paper), A Nitrogenous Soil Constituent: Tetracarbon- imid 175-178 Watermelon. See Citrullus vulgaris. Water requirement — definition of a of plants, relative 1-64 Weed, water requirement of 46-49 Weed, woolly. See Hieracium cytwglossoides. Wheat. See Triticum. Wheat-grass. See Agropyron. WUdennuth. V. L. (paper), Three-Cornered Alfalfa Hopper 343-36a Willow. Nuttall. See Salix nuttalln. Wind dissemination of ascosp>ores of Endothia parasitica 493-526 Valeriana sitchensis, forage value of 97 Vcratrum viride — forage value of 97 viability of seed of, in range lands 106 Vetch. See Vicia. Vicia spp.. water requirement of 3o~3^, 33-36, S2> 59.61-62 Vigna sinensis — host Yt^a-ntoiStictocephala festina. . 34S.35o-35i»359 injury to, by Slictocephala fesitna 359 water requirement of 30-31, 35>52-53»S9.6i ^anthitnn commune, water requirement of. . 47, 52,60,62 XyloPhruridca agrili, parasite of Agrilus vitta- ticollis 184 Yarrow. See Achillea lanidosa. Zea viays, water requirement of 17-20, 50,53-55-56,58,61 Zone- Arctic- Alpine, character of vegetation of . . 97-9S Canadian — character of vegetation of 96 climate of 98-100 inception of growth of forage plants in. . 101-102 Hudsonian — character of vegetation of 96-97 climate of 98-100 development of forage plants under year- long protection in 121-1 24 flower-stalk production in 102-104 inception of growth of forage plants in. loi-ioa Transition — character of vegetation of 95-96 climate of 98-100 development of forage plants under year- long protection in 124-125 inception of growth of forage plants in. . 101-102 ' if. 00263 3996