f Historic, archived document Do not assume content reflects current scientific knowledge, policies, or practices, Bui. 185, Bureau of Plant Industry, U. S, Dept. of Agriculture, Plate I. Fig. 1— Experimental Plats of Alfalfa in Hills and Drill Rows at Dickinson, N. Dak., October 1, 1908. Each row is seeded with a different strain, 68 kinds in ail. % '^ Fig. 2.— The Same Experimental Plats Illustrated in Figure 1 on June 10, 1909, Showing the Great Diversity in the Winter Resistance of Alfalfa from Dif- ferent Parts of the World. U. S. DEPARTMENT OF AGRICULTURE, BUREAU OF PLANT INDUSTRY— BULLETIN NO. 185. B. T. GALLOWAY, Chief of Bureau. COLD RESISTANCE OF ALFALFA AND SOME FACTORS INFLUENCING IT. CHARLES J. BRAND, Physiologist, Crop Physiology and Breeding Investigations, AND L. R. ^ALDRON, Superintendent of the Dickinson Substation, North Dakota Agricultural Experiment Station. Issued September 16, 1910. WASHINGTON: government printing office. 1910. BUREAU OF PLANT INDUSTRY. Chief of Bureau, Beverly T. Galloway. Assistant Chief of Bureau, G. Harold Powell. Editor, J. E. Rockwell. Chief Clerk, James E. Jones. Crop Physiology and Breeding Investigations, sclentific staff. Walter T. Swingle, Physiologist in Charge. S. C. Mason, ArhoricuUurist. G. P.Rixford, Expert. Bruce Drummond, W. L. Flanery, E. W. Hudson, and E. M. Savage, Assistants 185 2 LETTER OF TRANSMITTAL U. S. Department of Agriculture, Bureau of Plant Industry, Office of the Chief, Washington, D. O., May 6, 1910. Sir: I have the honor to transmit herewith a paper entitled ''Cold Resistance of Alfalfa and Some Factors Influencing It." This report was prepared by Mr. Charles J. Brand while Physiologist in Charge of Clover and Alfalfa Investigations in the Office of Crop Physiology and Breeding Investigations, and by Mr. L. R. Waldron, superin- tendent of the subexperiment station conducted by the State of North Dakota at Dickinson. I recommend that it be published as Bulletin No. 185 of the special series of this Bureau. One of the important lines of investigation in the Bureau of Plant Industry is the determination of the factors that make it possible for plants to thrive in the extremely cold sections of the West and North- west. It is a broad, complicated question, and all of the factors can be determined only by studies that extend through many winters of extreme cold. One such season, the winter of 1908-9, occurred during the progress of the studies of the writers and forms the basis of the principal part of the discussion. The investigations reported, however, have been in progress for five years in different parts of the Northwest. They include 80 dift'erent kinds of alfalfa, many of which are introductions of this Bureau. It would have been desirable for the observations of the writers to have extended over a greater number of severe winters. The need of information on the part of those in the cold regions is so pressing, however, that it seems advisable to present the facts contained in the paper as a contribution to the comprehensive series of similar studies in progress in the Bureau. The deductions of the writers are especially helpful in suggesting lines of investigation that are likely to be developed further by investigators in the cold regions, as well as by those wdthin the Bureau. Another phase of alfalfa hardi- ness has already been discussed in Bulletin No. 169 of this Bureau, entitled ''Variegated Alfalfa," by Mr. J. M. Westgate. Respectfully, B. T. Galloway, Chief of Bureau. Hon. James Wilson, Secretary of Agriculture, 185 CONTENTS. Page. Introduction 9 Purpose of the investigations 9 Experiments at Dickinson, N. Dak., begun in 1906 11 Rainfall 13 Temperature 13 Reasons why alfalfas sui'vived the winter of 1906-7 15 Termination of experiments begun in 1906 18 Field plat experiments at Dickinson, N. Dak., begun in 1907 18 Diversity in previous treatment of land used 18 Soil preparation, weather conditions, and other observations during 1907 and 1908 20 Winterkilling (1908-9) and yields (1909) of the 1907 field plats 22 Field plat experiments at Dickinson, N. Dak., begun in 1908 23 Winterkilling and yields 23 Observations on diversity of conditions 23 Differences in original stand 24 Differences in amount of winterkilling 24 Differences in cold resistance of two samples of Grimm alfalfa 25 Differences in color, vigor, and amount of growth of plants 27 Differences in yield 27 Importance of inoculation 29 Drill-row and hill experiments at Dickinson, N. Dak., begun in 1908 30 Soil of experimental area, previous cropping, and preparation 31 Origin of strains used in experiment 31 Weather conditions during the summer of 1908 32 Autumn moisture and its effect on dormancy 33 Amount of soil moisture in alfalfa stands of varying thickness 34 Weather conditions during the winter of 1908-9 37 . Effect of winter of 1908-9 on alfalfa varieties 39 Winter resistance of alfalfas in drill rows and hills' in the 1908-9 ex- periments 43 Alfalfas from South America 43 Alfalfas from Africa 43 Alfalfas from Europe 44 Alfalfas from Asiatic sources 45 Arabia 45 Turkestan 45 Mongolia 46 Alfalfas from North American sources 47 Canada 47 Mexico ^ 48 5 6 CONTENTS. Drill-row and hill experiments at Dickinson, N. Dak., begun in 1908 — Cont'd. Weather conditions during the winter of 1908-9 — Continued. Winter resistance of alfalfas in drill rows and hills in the 1908-9 experiments — Continued . Alfalfas from North American sources — Continued. Page. United States 48 Ordinary American alfalfa 48 Progeny of Old World introductions 50 Grimm alfalfa 50 Acclimatized Turkestan alfalfa 51 South Dakota No. 167 55 Comparison of conditions in field plats and in rows and hills 56 Further observations on the effect of autumn moisture 60 Diversity in alfalfa 63 Methods by which alfalfas become hardy 64 Conclusion 67 Description of plates 72 Index 73 185 LLUSTRATIONS. PLATES. Page. Plate I. Fig. 1. — Experimental plats of alfalfa in hills and drill rows at Dick- inson, N. Dak., October 1, 1908. Fig. 2. — The same experimen- tal plats illustrated in figure 1 on June 10, 1909, showing the great diversity in the winter resistance of alfalfa from different parts of the world Frontispiece. II. Fig. 1. — Experimental plats at Dickinson, N. Dak., June 12, 1909, showing the importance of a snow covering. Fig. 2. — Experi- mental plats at Stockton, Kans., showing the winterkilling of North African and Arabian alfalfas during the mild winter of 1908-9 72 III. Fig. 1. — Experimental plats of European alfalfas after the winter of 1908-9 at Dickinson, N. Dak. Fig. 2.— Experimental plats at Dickinson, N. Dak., August, 1909, showing three rows of Mon- golian alfalfa, the hardiest foreign-grown strain 72 IV. Fig. 1. — Experimental plats at Dickinson, N. Dak., showing the winterkilling of North American strains of alfalfa. Fig. 2. — A field of Grimm alfalfa in Carver County, Minnesota, showing the winterkillmg in 1905-6 in the "draws" or natural drainage courses of the field 72 TEXT FIGURE. Fig. 1. Ground plan of the experimental field plats of alfalfa in 1907 and 1908, showing their location and estimated percentage of stand in June and August, 1909, and, in connection with Table V, the previous cropping and treatment of the various areas composing the field 19 185 7 B. P. I.— 579. COLD RESISTANCE OF ALFALFA AND SOME FACTORS INFLUENCING IT. INTRODUCTION. Interest in the culture of alfalfa continues to increase. As a wealth producer this crop is winning for itself a predominant place in the agriculture of the world. While in some parts of Europe it can no longer be grown with the old-time success, and serious people are discussing measures for bringing its culture back to a paying basis, the United States has as yet developed only a small fraction of its alfalfa possibilities. So far as the northern part of the Great Plains area is concerned the investigations reported upon in the present paper show that weather, tillage, and the use of a suitable strain of seed are the con- trolling factors in successful alfalfa production when soils of average fertility are employed. Snow covering and the presence or absence of excess moisture in the soil were found to be fully as important in winterkilling as minimum temperatures; differences in tillage and in the previous treatment of the land had profound effects, and finally the greatest possible diversity was found in the endurance shown by strains of seed from different parts of the world. PURPOSE OF THE INVESTIGATIONS. In the autumn of 1904 the Bureau of Plant Industry began a comprehensive study of the alfalfa plant. During the preceding years many new strains had been introduced'^ with a view to finding a During the period that elapsed from May, 1898, until November, 1904, the Office of Foreign Seed and Plant Introduction of the Department of Agriculture introduced a total of forty-two strains of alfalfa from different parts of the world. Included among these earlier importations were the following: Fifteen numbers secured by Professor Hansen in Turkestan which constituted what was, so far as we know, the original introduction into North America of the form of alfalfa grown in that country; four numbers sent in by Mr. Swingle, including the well-known Provence and Poitou strains, also what was probably the Department's first introduction of the so-called commercial sand lucern, from France, and the highly prized Oasis alfalfa of the Sahara from the Oued Rirh ; three interesting Algerian strains grown in the province of Con- stantino, secured by Messrs. Fairchild and Scofield, including the interesting form 185 9 10 COLD KESISTA^TCE OF ALFALFA. better kinds for our older alfalfa-producing areas and also sorts that would succeed in areas where the ordinary strain had failed to establish itself. Little was kno^Ti as to how these new plant immigrants would behave in their new surroundings, and as their number grew it became very apparent that there was acute need for information as to what factors governed their successful growth. Without this knowledge the most effective distribution in this country was impossible ; hence the investigations begun m 1904 w^ere planned with a view to satis- fying this need. The many forms of alfalfa secured by the Bureau of Plant Industry have now been grown or are being grown in groups of from 2 to more than 200 in several hundred places scattered over the country. These different localities represent in a very full measure the great diversity of climate and soil present in the United States and furnish an opportunity for studying and comparing the behavior of the various kinds of alfalfa under a great range of conditions. Throughout the southwestern portion of the United States and in practically all areas where irrigation is practiced farmers have had very little difficulty in securing and maintaining highly profitable stands of alfalfa, but up to the present time Kansas and Nebraska are the only States that have made a notable success in growing this crop without irrigation. The alfalfa, or lucern,^ grown in these States grown in rows by Mr. G. Ryf, of Setif (see Fairchild, Bulletin 72, pt. 1, Bureau of Plant Industry, U. S. Dept. of Agriculture), a sample of the so-called sand lucern, and the variety known to botanists as Medicago sativa gaetula; two numbers from Peru, two from Argentina, and two from Arabia, seciu-ed by Messrs. Lathrop and Fairchild; one from Egypt, from Mr. Fairchild; two numbers constituting a new introduction of the Oasis alfalfa, sent by Mr. Kearney from the Oasis of Tuggm-t in Algeria; seven new lots from various parts of Turkestan, secui'ed by Mr. Bessey ; a commercial importation from Turkestan ; two strains from Mexico, and the interesting Peruvian alfalfa described in Bulletin 118 of this Bm'eau (see Brand, "Peruvian Alfalfa," Bulletin 118, Bureau of Plant Industry, U. S. Dept. of Agriculture). Since the autumn of 1904 nearly two hundred numbers have been introduced from various parts of the world; in addition a large number of more or less distinct strains have been secured from various parts of the United States, and a still larger number have been secured by propagation of seed from previous introductions. All of these domestic and foreign races and their progeny constitute the richest material for the study and breeding of alfalfa that has ever been brought together. o Alfalfa and lucern are wholly synonymous common names for Medica sativa (L.) Mill. [See Scofield, Carl S., "The Botanical History and Classification of Alfalfa," Bulletin 131, pt. 2, Bureau of Plant Industry, U. S. Dept. of Agriculture.] The for- mer is used in practically all Spanish-speaking countries and countries that drew their original supplies of seed directly or indirectly from Spain. This includes nearly all of the western continent. The name ' 'lucern, ' ' on the other hand, is the name generally applied in Germany, France, and contiguous countries. It is also the name commonly used by the Mormon farmers of Utah. The common name used in Italy is " herba medica." 185 EXPERIMENTS AT DICKINSON, BEGUN IN 1906. 11 came from California and Utah and is probably descended from the Chilean alfalfa introduced into this country about 1855. In a little more than a half century, so readily adaptable is alfalfa, this strain has spread over the agricultural areas of the greater portion of the western half of the United States, where it constitutes the backbone of agriculture. This strain, in the form to which it was unconsciously modified as it moved eastward, has been found hardy in Colorado, Kansas, Wyoming, south of the middle line of Nebraska, and by con- tinued growth has been extended throughout Nebraska and in Mon- tana as far north as the valley of the Milk River in the northern part of the State. In Kansas, farmers who have been growing alfalfa for from twenty to twenty-five years state that the sort they have used has never winterkilled to a noticeable extent. Speaking broadly, there are three great areas for which no well- adapted alfalfa has been available in quantity: (1) The humid Eastern and Southern States; (2) the semiarid regions where irriga- tion water is not available, including especially the drier parts of the Great Plains and Intermountain areas, and (3) the cold northern part of the country. In the region north of the Valley of the Platte in Nebraska and extending eastward to the Atlantic Ocean none of the strains of alfalfa, either of Chilean or European origin, obtainable in commercial quantities have proved hardy in the sense that they can be abso- lutely depended upon. Hence, one of the great needs of our northern agriculture and that of Canada has been an alfalfa that is perfectly hardy and reliable in this area. It is the purpose of the present paper to discuss rather fully some of the factors that appear to be important in the wintering over and winterkilling of alfalfa. The greater part of this work has been done in North Dakota at the Dickinson subexperiment station, but advan- tage has also been taken of using, for collateral observations, numer- ous other experiments located in various parts of the West, North, and Northwest. Dickinson is located in the western part of North Dakota at an altitude of 2,453 feet. The range of temperature thus far recorded is from —47° to 110° F. The winters are severe on plant life. The soil is similar to that of much of the Great Plains area. An alfalfa that does well at Dickinson is likely to succeed in a large part of the surrounding region. EXPERIMENTS AT DICKINSON, N. DAK., BEGUN IN 1906. With a view to gaining some knowledge as to their comparative cold resistance and also to determine the factors affecting cold resist- ance in various strains under moderately light rainfall, a series of 185 12 COLD RESISTANCE OF ALFALFA- 22 strains of alfalfa (see Table I) was sown in duplicate on plats one- fortieth of an acre each in the spring of 1906. One complete set was sown on the best soil -on the experiraent farm, a rich sandy loam, while the duplicate series was placed on a light sandy piece of land on another part of the farm. The series on sandy soil was sown May 9, 1906; that on the loam May 10. The following table shows the serial number, the region of origin, and the estimated percentage of stand on both the sand and the loam during the first or seeding year and the percentage of stand remaining after the severe winter of 1906-7. Table I. — Results of experiments with different strains of alfalfa grown on sandy soil and on loam at Dickinson, N. Dak., during 1906 and 1907, showing the j^ercentage of a perfect stand which survived the icinters. Origin. Sandy soil. Loam. No. 1906. 1907. 1906. 1907. Stand. Date. Stand. Date. stand. Date. stand. Date. Perct. Per ct. Perct. - Perct. oD. E. S. 223 Fargo, N. Dak. (Grimm, crop ol 1904). 80 Aug. 15 80 June 3 95 Aug. 1 95 June 3 D. E. S. 241 Gladstone, N Dak. (crop 95 ...do.... 65 Do. of 1905 ^.b D. E. S. 282 Commercial al- fal!a.& 70 Aug. 15 10 June 3 o S.P.I. 9303 Peru 10 ...do.... 1 ...do.... 40 Aug. 1 r; June 3 S. P. I.] 2694 Provenoe,Franoe ...do.... 4 ...do.... 85 ..do.... 40 Do. S.P.I.n211 Commercial Tur- kestan. 80 ...do.... 75 ...do.... 85 ...do.... 85 Do. S. P. 1. 12398 Colorado 80 ...do.... 15 ...do.... 90 ...do.... 60 Do S. P. 1. 17449 Commercial American. 75 ...do..-. 12 ...do.... 93 ...do.... 50 Do. S. P. 1. 12747 Yellowstone Valley, Mont. 75 ...do.... 20 ...do.... 90 ...do.... 65 Do. S. P. 1. 13291 Fayette ville, NY. 75 ...do.-.. 50 ...do.... 90 ...do.... 65 Do. S. P. 1. 12784 Utah (irrigated). 70 ...do.... 5 ...do.... 90 ...do.... 50 Do. S. P. 1. 12816 Milk Eiver Val- ley, Mont. 75 ...do.... 15 ...do.... 80 ...do.... 55 Do. S. P. 1. 12801 Texas Panhan- dle. 75 ...do.... 5 ...do.... 80 ...do.... 40 Do. S. P. 1. 13259 Southern Ne- braska. 60 ...do.... 25 ...do.... 80 ...do.... 65 Do. S. P. L 13237 Chinook, Mont.. 75 ...do.... 4 ...do.... 85 ...do.... 55 Do. S.P.I. 12671 Kansas 75 ...do.... 40 ...do.... 85 ...do.... 60 Do. S. P. 1. 12820 North em Ne- braska. 70 ...do.... 45 ...do.... 90 ...do.... 65 Do. S. P. 1. 12748 Rhine Province, Germany. 75 ...do.... 45 ...do.... 85 ...do.... 35 Do. S. P. 1. 14786 Tashkend, Tur- kestan. 85 ...do.... 85 ...dO...- 85 ...do.... 85 Do. S. P. 1. 12409 Utah (not irri- gated). Kharkof, Russia. 75 ...do.... 20 ...do.... 85 ...do.... 20 Do. S. P. 1. 13858 90 ...do.... 25 ...do.... 95 ...do.... 80 Do. S. P. 1. 13857 Simbirsk, Russia Average... 60 72.7 ...do.... 55 ...do.... 70 ...do.... 60 Do. 29 84.6 56.1 a These abbreviations refer to the serial numbers of the Dickinson subexperiment station and of the Office of Foreign Seed and Plant Introduction of the Bureau of Plant Industry. b These strains were not in duplicate. It will be noted that the stand on the loam is in nearly every case better than that on the sand and that the loss in stand observed in 185 EXPERIMENTS AT DICKINfeON, BEGUN IN 1906. 13 1907 as compared with 1906 is practically always greater on the sandy soil than on the loam. Although this loss in stand is undoubtedly due in large part to winterkilling, it is no doubt true that a part of it may be accounted for by the fact that the sandy soil is much less retentive of moisture than the loam, so that the deterioration of the stand on the sandy soil would be due to a combination of drought and cold. The likelihood of this will be more apparent after reading the discussion of the precipitation of the autumn of 1906 given below. RAINFALL. The rainfall during 1906 was 19.99 inches, while the yearly average based on a 16-year record is 15.11 inches. Despite the fact that the precipitation exceeded the average considerably, there were times during 1906 and 1907, as may be judged from the following table (Table II), when a lack of moisture interfered notably with the growth of the crop. This was particularly true during July, 1906. The dry spell that occurred in this month was in part responsible for the small growth made on all plats during the season of 1906. The most important factor in contributing to the lack of vigorous growth was the failure to secure inoculation. The effect of inoculation on cold resistance will be more fully discussed later. Table II.— Rainfall in inches at Dickinson, N. Dak., for 1906, 1907, and 1908. a Average for 16 years. 1906. 1907. January February . . March April May June July August September. October November. December. . .48 .99 1.18 2.43 2.92 2.12 1.82 1.04 .70 .56 0.25 .32 .98 1.10 7.11 5.40 .16 2.64 .25 .14 .87 .77 0.80 .14 .39 .30 1.36 2.52 4.82 1.89 1.11 .10 .02 .22 0.28 .73 1.42 1.27 3.50 4.30 1.41 1.43 1.67 2.47 .78 .24 Annual. 15.11 19.99 13.67 19.50 oThis table is quoted from Waklron, L. R., First Annual Report of the Dickinson Sub-Experiment Station for the year 1908, Fargo, N. Dak., 1909. TEMPERATURE. The winter of 1906-7 was one of extraordinary severity. The minimum temperature for October was 11° F.; .November, —1° F.; December, -24° F.; January, -35° F.; February, -31° F.; March, — 8° F.; and April, 5° F. During January there were only two days when the minimum temperature registered was above zero (3° F. on January 1 and 5° F. on January 11) and only one day when the maximum rose above freezing (35° F. on January 23). 185 14 COLD RESISTANCE OF ALFALFA. In order to show more adequately the nature of the weather which the experimental alfalfas shown in Table I were called upon to endure the report of the Weather Bureau for the month of January, 1907, is presented in its entirety m Table III. Table III. — Daily record of temperature and precipitation for the month of January, 1907, at Dickinson, N. DaJc; latitude, 46° 50^ N.; longitude, 102° 50' W.; altitude, 2,453 feet. Temperature. Precipitation. Date. Maxi- mum. Mini- mum. Range. Record ate p. m. Snow- fall. Depth of snow on ground at time of ob- servation. 1 ° F. 5 - 1 - 1 6 - 2 5 6 - 8 14 9 15 5 - 6 -23 -13 - 9 11 23 24 0 24 12 35 30 - 1 4 - 4 - 7 - 6 7 30 ° F. 3 - 8 -20 - 4 - 5 - 4 - 7 -14 -23 -22 5 -13 -14 -33 -35 -16 -22 -16 -14 -28 - 3 -13 - 3 - 6 -15 -13 -20 -21 -25 -17 - 2 ° F. 2 7 19 10 3 9 13 6 35 13 10 18 8 10 22 7 33 39 38 28 27 25 38 36 14 17 16 14 19 24 32 ° F. 4 - 8 - 3 2 - 3 3 - 3 -14 -12 15 5 -11 - 9 -29 -13 -14 - 6 14 -14 0 - 3 11 24 - 2 - 1 - 5 -20 -12 -12 3 22 Inches. 1 1 Inches. 7.5 8.5^ MONTHLY SUMMARY, 2 Temperature. 3 4 Mean maximimi, 5.93° F. 5 6 Mpfin minirrmm l.*^ 74° Tf 7 oT. 8 Mean —3 9° F 9 10 Maximvun 35° F ■ date 23d 11 3 aT. 1 11.5 12 Minimum, -35° F.; date, 15th. Greatest daily range, 38° F. 13 14 12.5 15 16 1 13.5 17 18 Snow. 19 20 Total fall, 8 inches. 21 22 23 24 REMARKS. 25 26 Temperature, 15.1° F. below nor- 27 28 1 14.5 29 30 , 31 Sum 184 -428 Mean 5.9 -13.8 a T.= trace. An examination of the columns of Table I showing the stand for 1907 reveals the fact that some plants on every plat in both series survived the winter. In several cases the strains came through with no injury whatever, and even the plats of the Peruvian variety, which next to the Arabian is perhaps the least hardy alfalfa known, had a sprinkling of living plants amounting on the sandy soil to about 1 per cent of a stand and on the loam to about 5 per cent. 185 EXPEEIMENTS AT DICKINSON, BEGUN IN 1906. 15 REASONS WHY ALFALFAS SURVIVED THE WINTER OF 1906-7. It is surprising, in view of the severity of the winter of 1906-7, which was almost unprecedented in North Dakota, that so large a percentage of alfalfa plants survived. From a careful study of the weather records for Dickinson from September, 1906, to May, 1907, inclusive, it seems that two factors were largely concerned in the result, namely, autumn weather of a "kind that induced com- parative dormancy before the real winter weather came, and a heavy blanket of snow throughout the coldest winter months. Beginning with September 10 there was on the whole a gradual downward movement of the mean daily temperature, which cul- minated in killing frosts on the 26tli and again on the 29th of the month. This and the dryness of the soil, the precipitation for the whole month being only 0.25 of an inch, served to harden the tissues of the plant and inhibit the production of new and tender growth. In October a very similar condition of affairs obtained. A pre- cipitation of only 0.14 of an inch for the whole month was recorded. The temperature went down rather gradually, with killmg frosts on the 5th, 9th, and 10th, while from the 18th to the 31st of October the daily minimum temperature ranged from 37° to 11° F., only two days (37° F. on the 25th and 35° F. on the 26th) having minima above the freezing point. The result of this combination of dryness and gradually increasing cold was first to brmg about a gradual cessation of the physiological processes involved in growth, and finally to bring the plants unharmed into a comparatively dormant condition. Several years of observation have led to the conclusion that a capacity to react with reasonable promptness to stimuli producing dormancy is one of the most important qualities, if not the most important, in producing hardmess in alfalfas and clovers. Similarly during dormancy, resistance to growth-producing stimuli is important. Elsewhere ^ an attempt has been made to explain the failure of one variety to become dormant while another becomes so, by conceiving varieties and even individuals as having a more or less constant minimum temperature, below which growth ceases in the autumn and above which it starts in the spring. Breeders desiring to estab- lish races for particular purposes should seek wherever possible to correlate one physiological character mth another rather than morphological characters with physiological, unless there is a fairly obvious relation between the physical character used as a basis of judgment and the physiological character toward which selection is a Brand, Charles J. Peruvian Alfalfa: A New Long-Season Variety for the South- west. Bulletin 118, Bureau of Plant Industry, U. S. Dept. of Agriculture, pp. 8-14. 185 16 COLD EESISTANCE OF ALFALFA. being made. That dormancy is of paramount importance in deter- mining whether alfalfas will winterkill or not appears to be shown by the results of 1906 and 1907 as compared mth those of 1908 and 1909, given later. No two conditions are more potent than drought and cold in retard- ing plant growth or bringing it to a complete standstill by hardening the tissues. Both temperature and moisture conditions were ideal in western North Dakota during the autumn of 1906 for bringing about this result. Those for September and October have already been described. November weather was generally favorable for inducing dormancy, and a snow covering was present during a portion of the month. The total precipitation for the month amounted to 0.87 inch, of which 0.22 inch fell before the middle of the month and had a good chance to become absorbed by the soil. Only a portion of this soaked in. The remainder of the precipitation was in the form of snow, which remained until the end of the month. The precipitation for December was above the normal, but nearly all of this came in the form of snow, which remained in place. A short period of warm weather at the beginning of the month removed the snow, but on the 9th and 13th more snow fell, forming a covering over 7 inches deep. The temperature for December was 4.1° F. below the normal. The minimum temperatures ranged from 28° to — 24° F., while the maxima ranged from 47° to - 1° F. The snow afforded the plants a high degree of protection. The weather conditions during January, 1907, are shown in Table III and discussed in that connection. Attention here is directed only to the snowfall. The month opened with 7.5 inches (see Table III) on the ground; on the 15th there were 12.5 inches, and at the end of the month there were 14.5 inches. Without this blanket of snow and with the persistently low temperatures that occurred throughout the month it is probable that there would have been 100 per cent of winterkilling on the majority of the experimental plats, and this despite the extremely favorable conditions under which the alfalfa entered the winter. As it was, red clover, winter wheat, and some alfalfa on every one of the twenty-two plats survived the winter. As is always the case in this northern region, the weather during February, 1907, was also severe, the mean of the minimum tempera- ture being 4° F. and the absolute minimum —31° F. The month began with 14.5 inches of snow on the ground, 6 inches remaining on the 15th and 4 inches at the end. In order to make possible a comparison of mean temperatures endured by the alfalfas under experiment during the winter of 1906-7 185 EXPEEIMENTS AT DICKINSOX, BEGUX IX 1906. 17 vV'ith those endured by the strains under experiment during 1908-9, the mean temperatures for these years are shown in the following table: Table IV. — Mean temperatures for DicJcinson,^ N. Dak., during 1906, 1907, 1908, and 1909, the years covered by the alfalfa experiments reported upon in this bulletin. Month. Aver- 1906. Mean Mean for ^^ of the of the max- min- ima, ima. years Mean. 1907. Mean Mean of the of the max- min- ima, ima. 1908. Mean. Mean Mean of the of the ima. mia, 1909. Mean' Mean of the of the January February March April May June July August September October November December Annual. . o jr 11.0 11.9 23.5 42.6 52.6 60.9 68.0 67.2 56.8 44.0 27.3 18.6 40.3 ° F. 28.4 29.5 31.9 63.0 61.6 73.2 82.9 79.1 77.8 81.8 37.2 25.4 54.3 o p 5.8 5.2 8.1 33.0 38.8 49.2 50.9 53.6 44.6 31.0 15.4 3.6 o p 16.5 17.2 19.8 48.3 50.3 61.2 66.9 66.3 61.2 46.4 26.3 14.5 28.0 O J^ 5.9 27.3 41.6 47.5 59.5 72.7 77.6 81.2 67.2 62.6 46.4 36.9 O J^ -3.9 15.6 28.5 34.2 44.5 60.2 65.3 65.3 52.8 46.3 31.8 23.5 o p 35.5 24.7 34.1 53.7 62.0 71.4 84.9 80.0 76.4 53.5 44.6 27.7 o p 6.9 6.-9 13.6 30.0 37.0 48.6 55.0 48.5 47.0 31.3 20.4 6.6 41. 52.2 24.5 38. i 54. 0 29. 5 o p 22.1 15.7 23.8 44.3 49.5 60.0 64.2 61.7 42.4 32.5 17.1 41.5 o p 18.5 27.2 37.1 48.6 65.8 74.1 78.3 82.7 74.7 58.1 44.8 17.4 -2.6 6.2 18.2 22.2 38.0 51.2 55.5 54.6 44.6 30.9 18.0 0.0 7.85 16.70 27.70 44.30 49.50 62. 60 66.90 68.60 59.60 44.50 31.40 8.70 52. 28.0 40.70 a The weather records for Dickinson have been kept at the experiment farm since March, 1906, and thus the data given are almost identical with those obtaining at the plats, as the two points are less than a quarter of a mile apart. ^ These figures are based on records beginning in 1893, when the first observations were taken. The most striking thing to be noted in view of the fact that a con- siderable percentage of the alfalfa plants lived through the winter, despite their lack of vigor due to imperfect inoculation and drought, is that the mean for January, 1906 ( — 3.9° F.), was 14.8 degrees colder than the normal temperature that had prevailed up to that time for that month, based on a 13-year record. March, 1907, was slightly warmer than usual, but the snow cover- ing remained until practically the 15th of the month. Though a slight growth began to show in manjT plants in April, decidedly inhibitive minimuxn. temperatures continued throughout the month. There were killmg frosts May 1-10, 13-15, 18-19, 26-27, and a final one on June 5. The weather was such that the spring awakening was almost as gradual as the enforced dormancy of autumn. It was on account of this unusual combination of autumn drought, the gradual com- ing of winter, the protective covering of snow, and the gradual return of spring that some of even the tenderest alfalfa varieties came through this winter, which was one of almost unparalleled severity. 44281°— No. 185—10 2 18 COLD EESISTAXCE OF ALFALFA. TERMINATIOX OF EXPEKOIENTS BEGUN IX 1906. Although a moderately good stand of alfalfa was obtained, the average for all plats on the sandy soil being 72.7 per cent and for those on the loam 84.6 per cent, the condition of the 1906 seedings was never satisfactory. The following statement upon this experi- ment is found in the records for 1906: ''The alfalfa on these plats is a failure, due almost entirely to acute nitrogen starvation. The plants were not inoculated ^ * ^ ^nd they were of weak, sickly appearance." This lack of inoculation primarily, together with the dry autumn of 1906 and the severe winter that succeeded it, kept the plants m an unthrifty condition. Following this, the summer of 1907 was one of the driest of recent years. A total of only 13.67 inches of rain fell, while in 1906 and 1908 there were 19.09 inches and 19.50 inches, respectively. This combination of trying conditions kept the plats in such an unpromising state that in the late autumn of 1907 they were plowed up to make room for other crops. FIELD PLAT EXPERIMENTS AT DICKINSON, N. DAK., BEGUN IN 1907. In the spring of 1907 a second experiment was started. Only four regional strains were seeded. Two samples of American alfalfa, one from Colorado and one from Utah, and two strains that had shown some degree of hardiness in the 1906 experiments, one from Turkestan and one from Simbirsk, Russia, were selected. At this time no seed of the Grimm alfalfa of Minnesota was available for use. DIVERSITY IX PREVIOUS TREATMEXT OF LAXD USED. The different portions of the field on which the experimental plats were located had not had the same cultural history. The virgin sod was broken in different years in some cases; some parts in autumn, others in spring; one part was left without a crop the first year, in other years the different parts bore different crops; one part received a dressing of rotted manure in a year when none of the others were fertilized. These diversities of cultural history furnished an unusually good opportunity to study their effect on cold resistance. As interesting differences occurred in the amount of winter loss on different areas of the field, Table Y and text figure 1 are inserted to show the different treatments these areas had received. As the 1908 field plats to be 185 FIELD PLAT EXPEEIMEKTS AT DICKINSON. BEGUN IN 1907. 19 discussed later were seeded on an adjacent part of the same field, they also are shown in the figure." Table V presents the history of Area I AreaM Area M Area IV ^/£ Fig 1.— Ground plan of the experimental field plats of alfalfa in 1907 (5) and 1908 {A), showing their location and estimated percentage of stand in June and August, 1909, and, in connection with Table V, the previous cropping and treatment of the various areas composing the field. (Scale, linch=100feet.) a The 1908 field plats are fully discussed on pages 23 to 30, inclusive, but the writer's desire to call especial attention to the fact that the great differences in stand observed in 1909 on the various portions of these plats are due in large part to tillage effects and to differences in previous cropping and not to winterkilling. The extent to which winterkilling was effective in thick as compared with thin stands may be seen best by a study of Table VIII. page 23. A comparison of the two ''Total" and two "Loss" columns in this table shows that the percentage of loss was greater in every case on the square yard having the thinner stand; that is, the smaller total number of plants. 185 20 COLD EESISTAKCE OF ALFALFA. each area of the plats shown in figure 1. Estunations of the stand were made early in June, 1909, and again late in August. Table Y.— Record of previous crops groivn and of treatment given the land used for the experimental field plats of alfalfa in 1907 and 1908. [See also fig. 1.] Year. Area I. Area II. Area III. Area IV. 1904 Virgin prairie Virgin prairie. broken some years before, cropped, and then allowed to re- tm-n to sod. 1905 Broken in spring, not seeded, and grew up ^ irgin prairie Rebroken in spring and sown to oats. Virgin prairie. to Russian thistles. Fall-plowed. 1906 Small - grain plats. Plowed and disked in autumn. Broken up in October. Grain. Fall-plowed.. B, Manured on stub- ble in autumn. Broken up in October. 1907 B, AHalfa (1907 field plats). A, Oats, stubble ma- B, Alfalfa (1907 field plats). A, Oats, stubble ma- B, Alfalfa (1907 field plats). A, Oats, stubble ma- B, Alfalfa (1907 field plats). A, Oats, stubble ma- nured. nured. nured. nured. 1908 B, Alfalfa (1907 field B, Alfalfa (1907 field B, Alfalfa (1907 field B, Alfalfa (1907 field plats). plats). plats). plats). A, Alfalfa (1908 field A, Alfalfa (1908 field A, Alfalfa (1908 field A, Alfalfa (1908 field plats). plats). plats). plats). 1909 ^ and 5, Alfalfa .•1 and 5, Alfalfa A and £, Alfalfa ^ and 5, Alfalfa. SOIL PREPAKATIOX, WEATHER CONDITIONS , AND OTHER OBSERVATIONS DURING 1907 AND 1908. In the spring the land for the 1907 plats was double-disked and harrowed on April 25 and double -disked crosswise and harrowed on June 1. Seeding took place on June 5, after a good rain. The rainfall of the summer was fairly adequate, though almost hah of the 4.82 inches shov\'n in Table II for July, 1907, occurred in two heavy storms only three days apart. Over half of the August rain fell in one storm. The autumn was dry. September had about a normal precipitation, while October had only 0.10 inch and November only 0.02 inch. The first killing frost did not occur until September 29. Before this time growth had stopped and the plants had become practically dormant because of the dry weather. Russian thistles, more drought resistant than alfalfa, came up thickly and overtopped the latter. An ideal plant covering to hold snow was created. On October 11, 1907, the condition of the plats was noted. The estimated per- centage of stand on this date for each plat was as follows: Per cent. Turkestan (S. P. I. No. 14786) 80 Simbirsk, Russia (S. P. I. No. 13857). 70 Utah, irrigated (S. P. I. No. 12784) 90 Colorado, irrigated (S. P. I. No. 12398) 90 The winter of 1907-8 was an unusually mild one; the average tem- perature for the winter months was 20.4° F., this being almost 7 degrees above the 16-year normal. 185 FIELD PLAT EXPEEIMEXTS AT DICKINSON, BEGUK IX 1907. 21 The lowest temperature of the whiter occurrecl on February 1 and was preceded by two da3^s of snow. This was the only really cold spell experienced during that season, and the alfalfa was amply pro- tected by snow throughout the period. Mild as this winter was, some winterkilling occurred in all the strains. The amount was negligible in the Turkestan and Russian strains, the former showing the smaller loss. The Colorado and Utah plats suffered some loss, though not enough to reduce them to unprof- itableness. All stands were sufficient to justify leaving the plats over during the winter of 1908-9. They will receive further discussion in connection with the field plats seeded in the spring of 1908. A fact of some interest was noted regarding the comparative hardi- ness of alfalfa and red clover during this dry open winter. Six plats of the latter on land adjacent to the Simbirsk and Turkestan plats showed losses in stand ranging from 55 to 95 per cent. In order to show the temperature conditions endured by the 1907 seeding as compared with the 1906 and 1908 experiments. Table VI is presented. Table YI. — Monthly mean temperatures for the winters of 1906, 1907, and 1908, at Dickinson, N. Dale. Monthly mean. Month. Average for 16 years. 1906-7. 1907-8. 1908-9. °F. 21. Z 18.6 11.0 11.9 23.5 °F. 26.3 14.5 -3.9 15.6 28.5 °F. 31.8 23.5 22.1 15.7 23.8 °F. 32.5 December 17.1 January 7.9 16.7 March 27.7 18.5 16.2 23.4 20.4 The Turkestan and Russian strains of alfalfa not only did not suffer serious loss, but by the summer of 1909 showed an improved condition of stand. This change was one of appearance and not an actual thickening of the stand. It varied in the different areas shown in figure 1 and appeared to be due in large part to a progressive inoculation of the soil with nitrif}dng bacteria as the plants became better established. The most important fact gleaned from the small experiment begun in the spring of 1907 was that the common form of alfalfa, which in a little more than half a century has been extended so generally over the western half of the United States and which has for many years proved sufficiently hardy throughout Kansas and over a large part 185 22 COLD EESISTANCE OF ALFALFA. of Nebraska, is not cold resistant enough to endure perfectly even a mild winter in the coldest parts of the Great Plains. The autumn of 1907 furnished another illustration of the importance of dry weather during the fall in producing dormancy. The winter, as a whole, was an open one, but there was a snow covering of utility during the only critical period. Although the winter was an abnor- mally mild one, these two factors were probably dominant in the comparatively successful wintering over of the two American strains. These same factors and a greater inherent hardiness appear to have produced the better results observed in the Turkestan and Simbirsk varieties. WINTERKILLING (1908-9) AND YIELDS (1909) OF THE 1907 FIELD PLATS. The winter of 1908-9 is discussed fully later in connection with the 1908 row experiments; hence it need not be taken up in this connec- tion. Its effect on the 1907 field plats, as determined in the spring of 1909, is shown in Table VII. The figures presented in this table were secured in May, 1909, by counting the living plants and the winterkilled crowns on selected square yards which showed the highest and lowest amount of loss. Table VII. — Results of experiments showing the relative proportion of ivinterMUing (1908-9) of the alfalfas in the experimental field plats of 1907 and their yield in 1909. S.P.I. Strain. Plants on square yard showing the least winterkilling. Plants on square yard showing the greatest winterkilling. Yield of whole plat. No. Living plants. Dead plants. Total. Loss. Living plants. Dead plants. Total. Loss. 13857 12398 14786 12784 Simbirsk, Russia Colorado Turkestan... Eniery,Utah. Number. 69 236 120 20 Number. 23 70 10 91 Number. 92 306 130 111 Per cent. 25.00 22.87 7.69 81.98 Number. 45 32 31 4 Number. 33 191 30 127 Number. 78 223 61 131 Per cent. 42.30 86.09 49.18 96.94 Pounds. 2,065 1,131 2,502 1,175 A comparison of the figures in the percentage of loss columns in Table \T!I will show how great the differences in winter loss were in the same varieties, but on different parts of the same plat. Some discus- sion of the particular factors that may have produced these differences will be taken up in connection with the 1908 field plats. Attention here is directed only to Colorado alfalfa, S. P. I. No. 12398, which originally had an unusually thick stand over the whole plat. On the square yard where the least winterkilling had taken place only 70 plants (23 per cent) out of 306 were killed, while on the other 191 (86 per cent) out of 223 succumbed. 185 FIELD PLAT EXPERIMENTS AT DICKINSON^ BEGUN IN 1908. 23 FIELD PLAT EXPERIMENTS AT DICKINSON, N. DAK., BEGUN IN 1908. WINTERKILLING AND YIELDS. In the spring of 1908 a second series of field plats was seeded by the subexperiment station at Dickinson, N. Dak. These are in charge of the junior writer cooperating with ^Ir. J. M. Westgate, Agronomist in the Office of Forage-Crop Investigations. As may be seen by refer- ence to figure 1, the new seedings are situated immediately north of the 1907 field plats and share with these the diversities in the previous cropping and treatment shown in Table V. Counts of all the living and dead plants on selected square yards were made, as in the case of the 1907 seedings. The results are shown in Table VIII. Table VIII. — Results of experiments showing the relative jprojportion of winterkilling {1908-9) of the alfalfas in the experimental field plats of 1908 and their yield in 1909. 0' S.P.I. No. Variety or strain. Plants on square yard show- ing the least winterkilling. Plants on square yard show- ing the greatest winterkilling. Calcu- lated Liv- ing plants. Dead plants. Total. Loss. Liv- ing plants. Dead plants. Total. Loss. yield per acre. 20988 Tiorkestan Num- ber. 99 87 76 88 78 100 45 65 Num- ber. 8 1 0 0 9 41 0 36 Num- ber. 107 88 76 88 87 141 45 101 Per cent. 7.47 1.13 0.00 0.00 10.34 29.07 0.00 35.65 Num- ber. 60 42 51 40 8 22 18 16 Num- ber. 30 5 0 9 6 11 13 23 Num- ber. 90 47 51 49 14 33 31 39 Per cent. 33.33 10.63 0.00 18.35 42.85 33.33 41.93 59.00 Pounds. 1,156 21735 Grimm (Nebraska) 1,474 21827 (rriTTiTn (Afontana) 1,487 2,136 21247 912 21269 21217 Darmstadt, Germany do 1, 603 1,960 19508 1,044 a Mr. Charles H. Clark, Special Agent, Crop Physiology and Breeding Investigations, Bureau of Plant Industry, rendered painstaking and efficient help in making many of the determinations presented in this paper. Mr. O. J. Grace, Acting Superintendent, also rendered much assistance. 6 This plat is located between two plats sown in 1907. The seed was imported as sand lucem from Europe for the substation. On an examination of the plants it proves to be neither the true sand lucem nor the so-called commercial or false sand lucern. It is almost wholly lacking in variegated flower color, and by its general habit of growth, appearance, leafage, etc., suggests" the possibility that it is a Turkestan alfalfa adulterated with a small quantity of ordinary middle European alfalfa, the plat shown in the upper left-hand corner of figure 1 is froni the same seed. Its yield was not taken. c This plat is also located between two 1907 plats (see figure 1). According to Mr. J. M. Westgate it was grown from seed a part of which at least was from a field planted with variegated-flowered alfalfa seed, presumably procured originally from central Europe. OBSERVATIONS ON DIVERSITY OF CONDITIONS. Several interesting conditions were observed on both the 1907 and 1908 plats. These are shown in a measure in Tables VII and VIII and in figure 1. The following deserve especial consideration: (1) Differences in the original stand on different parts of the same plat. This becomes evident on comparing the columns marked ^' Total" for each plat shown m Table VIII. (2) Differences in the amount of winterkilling on different parts of the same plat. In these cases the greater percentage of loss appears to have taken place on the part that had the poorer stand originally. 185 24 COLD RESISTANCE OF ALEALFA. (3) Differences in the percentage of loss in the two Grimm samples, despite the fact that these were both grown from the same original stock. (4) Differences in the color and vigor of the plants and finally in the amount of growth on the different areas. The latter finds partial expression in the diversities in yield. (5) Differences in yield which show practically no correlation to any of the conditions mentioned above except the last. Differences in Original Stand. That some of the differences in the original stand are due to differ- ences in the preparation of the soil and the previous croppmg is probable, as they follow^ exactly the division lines of the '^ lands," sho^\Ti by the broken lines in figure 1, page 19. How^ sensitively alfalfa reacts to even slight variations in preparing the soil is show^n on both Areas II and III, but especially on the latter. Here the uphill or east furrows have a much better stand and higher growth than the downhill or w^est furrows. To show more concretely the diversities in stand on the differently treated strips of the field after the ^vinter of 1908-9, the estimated percentage on each for all plats is stated in figure 1. In the case of plat 21217 the difference in estimated stand between the east and w^est ends amounts to fully 70 per cent. The soil w^as very dry in the fall of 1906 when the sod was broken up on Area II, so dry, in fact, that the plowman had great diffi- culty in getting his plow to ^^bite" when going down hill. Differences in Amount of Winterkilling. The differences in the amount of winterkilling are traced with greater difficulty and less certainty than those producing differences in the original stand. However, there are certain interrelations. The square yard of each plat showing the thinner stand suffered the greater loss in each case. Hence, a lack of thickness of stand itself probably has an important bearing. An inferior snow-holding capacity results and a greater consequent exposure to the weather. Furthermore, thick stands, through greater transpiration, more quickly exhaust excess moisture from the soil. The probable impor- tance of this factor is more fully discussed later. Another factor both directly and indirectly of importance in winter- killing on the 1907 and 1908 field plats is inoculation. Its direct importance rests on the robustness, vigor, and generally greater hardihood that results from perfect nutrition. Inoculation conduces to a better root development, higher growth, and more leafage, the two latter making a more nearly perfect covering to bind the winter's snow. 185 FIELD PLAT EXPERIMENTS AT DICKINSON^ BEGUN IN 1908. 25 Differences in Cold Resistance op Two Samples of Grimm Alfalfa. The difference in the amount of winterkilHng in the two Grimm samples raises the important and interesting question as to what cause or causes this is due. Both the Montana and Nebraska seed were grown from S. P. I. No. 12991, the former at Chinook, in northern Montana, and the lat- ter at Alma, in southern Nebraska, near the Kansas line. Grimm No. 12991 was grown in Minnesota in 1904. The weakness of the Nebraska race as compared with that from the Montana seed was only slightly apparent wdiere the stand was thickest, but amounted to over 10 per cent on the square yard with the thinner stand. ^ « The difference in the winter hardiness of the two Grimm samples is not so great as to be outside the limit of possible error in field plat work. This is especially true of the present plats, which were put in only with ordinary care on land shomng a wide range of diversity in tillage effects. Nevertheless, several possibilities that suggest themselves as being concerned in this difference in cold resistance deserve notice in this connection: (1) Free hybridization between the Grimm field and neighboring fields of the far less hardy ordinary alfalfa. Common alfalfa is grown in both sections; hence there would be some tendency toward equalization as to the factor. However, natm-al selection toward cold resistance is operative every winter in northern Montana and very rarely operative in southern Nebraska. The result of this should be a hardier race of common alfalfa in Montana than exists in Kansas or in Nebraska south of the Platte River. There is little doubt that this is actually the case, both from experi- mental evidence and from the fact that there is a general demand for Montana seed in preference to Kansas, Nebraska, or Utah seed for use in cold regions. Hence, the weakness that probably results from hybridization would be greater in a measm'e in the southern-grown race. (2) Further selection toward cold resistance in Montana and a slight reversion, if anything, in southern Nebraska. Alma is approximately 300 miles farther south than the Grimm section of Minnesota and 600 miles south and almost as much east from Chinook. The selective power of the Montana as compared with the Nebraska climate is shown in a measure by the monthly minima for Alma and Chinook presented in the following table. The winters shown are those that intervened between the seeding, which took place in the spring of 1905, and the harvesting in the fall of 1907 of the seed used at Dickinson on the 1908 field plats. Minimum monthly temperatures for the cold months of 1905, 1906, and 1907 at ChinooJc, Mont., and Alma, Nehr. Locality. Winter of 1905-6. Winter of 1906-7. Oct. Nov. Dec. Jan. Feb. Mar. Oct. Nov. Dec. Jan. Feb. Mar. Chinook a °F. - 8 17 "F. -28 5 "F. -18 6 °F. -21 - 3 "F. -16 _ 2 =i4 op 15 16 ^F. - 7 7 °F. -15 2 "F. -45 0 "F. -43 — 6 "F. - 8 a The records for both Chinook and Havre, Mont., were used in securing these minima on account of gaps in the record of the former. 185 26 COLD EESISTANCE OF ALFALFA. Another point to be noted in Table VIII with reference to the Grimm samples is that the one whose loss from winterkilling was greater, nevertheless gave practically the same yield as the one that suffered no loss. This is due to the fact that the Nebraska Grimm alfalfa had fully 5 per cent better stand on the plat as a whole than the Montana Grimm. The following comparison of their germination gives a reasonable explanation of this difference in stand, and at the same time brings to light another difference in the progeny of S. P. I. No. 12991 as grown in Montana and Nebraska quite as profound as the diversity in hardiness, namely, in the percentage of hard seed. S.P.I No. strain. Germina- tion. Hard seed. 21735 21827 12991 Grimm (Nebraska) . Grimm (Montana).. Grimm (Minnesota) Per cent. 77.5 45.5 60.0 Per cent. 16.00 42.75 36.50 The percentages of germination and hard seed of the Nebraska- grown stock are the average of five tests, those of the Montana of two tests. These tests were made in the winter of 1908. In the spring of 1909 both strains were again tested, with the following- results: S.P.I. No. Strain. Germina- tion. Hard seed. Length of test. 21735 Grimm (Nebraska) Per cent. 83.5 53.5 Per cent. 10. 0 30.5 Days. 4 21827 The subject of hard seed is little understood, and attention here is directed to it only because it accompanies an observable difference in hardiness in the same race grown under unlike climatic conditions. The difference in winter weather shown by these minima is a great one, and might of itself be a suflScient explanation of the difference in the winter loss of the two Grimm plats. (3) The third factor of probable influence is new-place effect, due to a disturbance of local adjustment in the variety brought about by transporting it from one set of conditions to another. The warm climate at Alma would be especially likely to call forth new-place diversities. It is of com'se impossible to determine the amount of change any one of the factors discussed may have produced. Their potency, however, is scarcely open to question. The important phenomenon of local adjustment in cotton varieties has been dis- cussed recently by Mr. 0. F. Cook in "Local Adjustment of Cotton Varieties," Bulletin 159, Bureau of Plant Industry, September, 1909. Observations made the past season by the senior writer of the present paper indicate that this phenomenon also occurs in alfalfa when subjected to a change of place. 185 PIELD PLAT EXPEEIMEKTS AT DICKINSOK^ BEGUN IN 1908. 27 Differences in Color, Vigor, and Amount of Growth op Plants. Tlie differences in the color, vigor, and amount of growth of alfalfa plants find partial expression in Table VIII in the figures giving yields, which are discussed below. They are so evident on the plats themselves as to suggest immediate inquiry. On September 1, 1909, the plants of the 1908 seedings on Area IV (fig. 1, p. 19) averaged full}^ 16 inches high, those on Area III about 8 inches, while those on Areas I and II averaged about 7 and 8 inches, respectively. Furthermore, the plants of the thick stand on Area IV had the healthy dark-green color of well-inoculated alfalfa, while Areas I and II had unthrifty plants of a pale j^ellowish green. On Area III even the plants making up the thin stand on both up and down furrows showed more growth and better color than did the plants on Area II. Differences in Yield. a In view of the differences in the extent of winterkilling and in the carefully estimated percentages of present stand, shown in figure 1, the diversities in yield for the different plats are decidedly interesting. Take, for example, S. P. I. No. 20988, which lost only about 7^ per cent of the plants on a selected square A^ard where the stand was thickest and 33^ per cent w^here thinnest. Compare this with S. P. I. No. 21269, which lost over 29 per cent on the better part of the plat and 33^ per cent where the stand was poor. Although there is no great difference in the average number of living plants remaining on the better portions of both plats, and the number of living plants on the poorer part of No. 20988 is full}" three times as great as that on the corresponding part of No. 21269, the latter outyielded the former b}^ 450 pounds per acre. About September 1, 1909, careful estimates were made and the following comparison shows the stand on the different areas of both plats : S.P.I. No. Strain. Good. Medium. Poor. 20988 Turkestan . Per cent. 80 90 Per cent. 80 33 Per cent. 75 21269 Coniniercial sand lucern 20 A similar disagreement in yield and stand exists between S. P. I. Nos. 21217, commercial sand lucern, and 21827, Montana-grow^n Grimm alfalfa. Although the former has a thin stand over two- thirds of the plat and showed some winterkilling where the stand was a See Table VIII, p. 23, in connection with this discussion. 185 28 COLD EESISTANCE OF ALFALFA. thin, while the latter was wholly uninjured and has on the whole a far thicker stand, still No. 21217 gave a yield of 2,000 pounds per acre, while No. 21827 produced only 1,487 pounds. An even more striking result is apparent on comparing Turkestan, S. P. I. No. 20988, in Table VIII, with Utah, S. P. I. No. 12784, in Table VII. The average winterkilling for the two parts of the Turkestan plat was about 20 per cent, with 80 per cent of a perfect stand remain- ing. The Utah plat winterkilled almost 90 per cent and still, accord- ing to the scales, produced as much hay as the Turkestan. A part of this discrepancy may be explained by the general inoculation of the surviving plants of the Utah strain, but the controlling factor was the weeds harvested and weighed with the alfalfa. Weeds constituted a large proportion of the product of all the plats or parts of plats which had thin stands originally or which suffered severe winter losses. In the 1907 section of the field the latter con- dition applies particularly to the Utah, Kansas, and Colorado strains, and the former in a measure to all parts of the plats included in Areas II and III. On the western two-thirds of the east tier of the 1908 plats the stand of alfalfa was very light, and weeds made up a large part of the hay yield. The three plats on the west side of the field (see fig. 1, p. 19) had sufficiently thick stands of alfalfa to hold the weeds under complete check. The Canadian plat's product, small as it is, was still more largely made up of weeds. The original seeding on this plat was very thin. These differences are traceable almost wholly to inoculation. Almost one-third of plat 21217 (all of Area IV) is perfectly inocu- lated, while only scattering plants on 21827 show signs of inoculation. As previously stated, on September 1 the growth on Area IV aver- aged about 16 inches high, while that on Areas I and II was less than 8 inches. The yields given are for the first crop. Those for the second crop vary even more strikingly, but are not given, as rains made it necessary to scatter and rebunch the hay several times. In doing this some confusion arose, making the figures as a whole unreliable, hence their omission. That the diversities in yield commented upon are really due to a difference in the extent of inoculation is readily apparent upon even a casual examination. A comparison between the 1908 plats (Table VIII) and the 1907 plats (Table VII) furnishes additional evidence that is of interest. The latter are uniformly inoculated, while the former are not. The average stand on Turkestan plat No. 14786, seeded in 1907, is practically identical with that of Turkestan No. 20988, seeded in 185 FIELD PLAT EXPERIMENTS AT DICKINSON^ BEGUN IN 1908. 29 1908. The yield of the former, 2,502 pounds per acre, was more than twice that of the latter, which gave only 1,156 pounds. Another factor with reference to the 1908 field plats that deserves mention is a difference in the amount of seed applied to the different plats. The following shows the rate per acre at which the seeding of the various strains of alfalfa was done : Pounds. Turkestan (S. P. I. No. 20988) 15.4 Sand lucern (S. P. I. No. 21217) 13.7 Sand lucern (S. P. I. No. 21269) 9.2 Grimm (S. P. I. No. 21827) 13. 9 Grimm (S. P. I. No. 21735) 10.2 Canadian (S. P. I. No. 21247) 4.4 Kansas (S. P. I. No. 19508) 14. 9 IMPORTANCE OF INOCULATION. As the foregoing discussion shows, many factors, some of them probably unrecognized, interacting in a complex and imperfectly understood manner, must be taken into consideration in comparing the yields given of the different plats, and too great reliance must not be placed on them. Another plat of Grimm alfalfa seeded in 1908 with Minnesota seed gives an opportunity for comparing the same variety under uniform field conditions with the Grimm as grown under the very dissimilar conditions prevailing on the 1908 experimental seedings. The small field of Grimm which is referred to as the ^' valley plat" is located in a slight depression farther up the slope and about 600 feet northwest of the 1908 field plats. It was seeded the same day as the latter and has a perfect stand except at one point where evidently some of the drill hose became clogged. The valley plat contains a net area of 1.22 acres and produced 4,098 pounds of cured hay per acre for the first cutting. This yield should be compared with the Nebraska- grown Grimm which had only a slightly poorer stand, but produced less than 1,500 pounds of hay per acre. The former had general inoculation, while the latter had practically none. It might be mentioned that a pure culture was used to inoculate the seed sown on the Nebraska Grimm plat, while the valley Grimm plat was inocu- lated both with a pure culture and with alfalfa dirt. In this case the soil inoculation appears to have been the more efficacious, though diversity in conditions makes satisfactory comparisons impossible. The valley plat soil is somewhat better, though the two soils do not differ materially. The valley plat, originally virgin prairie, was broken and disked July 20, 1907. It was well disked and harrowed 185 30 COLD EESISTAXCE OF ALFALFA. in the spring of 1908 and seeded to alfalfa May 27, 1908. It is coni- monty believed that alfalfa does not do well on new groimd. That inoculation has even more direct relations to cold endurance than those already discussed appears to be sho^^^l b\^ observations made during the autumn of 1909. On the night of September 25 a hglit frost occurred. During the following day Mr. Clark made the interesting observation that the well-inoculated parts of all plats remained uninjured", while the parts that had only shght inoculation or none at all were markedly affected. During the night of October 8 a sharp frost (24° F.) occurred, which affected all areas ahke except Area IV. The latter had the heavy growth, dark-green color, and general robustness that accompany perfect inoculation. The same observations were recorded for the vaUey plat of Grimm alfalfa, which Jias like evidences of inocula- tion but is located farther up the slope and northwest from the 1907 and 1908 field plats; however, in this case a slightly higher altitude must also be taken into consideration. Similar observations as to the frost resistance of inoculated plants were also recorded in the case of well-inoculated row experiments located alongside the field plats. DRILL-ROW AND HILL EXPERIMENTS AT DICKINSON, N. DAK., BEGUN IN 1908. In the early spring of 1908 arrangements had been concluded for a more comprehensive test of the hardiness of alfalfa than any pre- viousty undertaken in the United States. In accordance with the plan that was made, sixt3-eight strains or numbers were sown on May 15 in duphcate series. One complete set was sown in drill rows 3 feet apart with one row of each kind. These rows are 50 feet long, and as the summer advanced the plants in them w^ere thinned out to one every 8 to 12 inches, leaving only the most vigorous plants. The second series was in hills 2^ feet apart. There were 25 hills of each strain which durmg the summer were thinned out to one individual in each hill. The hills of each strain were planted at the end of the drill rows, so that both methods of seeding were combined in one continuous row. The experimental plat, the soil of which is a rich, black, sand}' loam, lies near the base of a ridge that extends in a general east and west direction and furnishes some degree of protection from the prevailing winter winds which come from the Northwest. The plat slopes slightly toward the middle on accouiit of a slight depression running diagonally across it. This affects nearly all rows equally. 185 experiments at dickinson, begun in 1908. Soil of Experimental Area, Previous Cropping, and Preparation. 31 The area used for the drill-row and hill experiments was virgin prairie which was first broken in the spring of 1905 and planted to vines. The land was plowed in the fall of 1905 and then disked. It was disked and harrowed in the spring of 1906 and sown broadcast to small clover plats. The land was plowed in the fall of 1907 about 6 inches deep and worked down. The ground was well worked in the spring of 1907 with a disk and harrow before starting the alfalfa experiment. The soil was inoculated by sowing earth, taken from an experimental alfalfa plat on the station farm, broadcast at the rate of about 400 pounds per acre. The soil would probably be described as a fine black, sandy loam — the Morton loam as described by the Bureau of Soils. The following mechanical analysis is furnished by the Bureau of Soils of the United States Department of Agriculture: No. of sample. Description of soil analyzed. Fine gravel, 2 to 1 mm. Coarse sand, 1 to 0.5 mm. Medium sand, 0.5 to 0.25 mm. Fine sand, 0.25 to 0.1 mm. Very fine sand, 0.1 to 0.05 mm. Silt, 0.05 to 0.005 mm. Clay, 0.005 to 0 mm. 13634 0.4 .3 0.9 .3 1.4 .9 10.6 9.7 19.3 20.4 47.3 42.0 20.1 13635 Second, 8 inches 26.5 Origin of Strains used in Experiment. The strains used in the experiment were chosen with a \iew to assembling a fairly complete collection of the cultivated alfalfas of the world; but in the case of American-grown seed no attempt was made to represent a climate milder than that of central Utah, pre- vious observations having shown that no degree of hardiness worthy of note could be found in seed from the milder States of the West and Southwest. Irrigated and unirrigated or dry-land strains were included in the plan in order to secure accurate data as to whether the characters producing drought resistance were in any way corre- lated with those producing cold resistance. To give a fuller conception of the wide range of climatic and other conditions represented b}" the sixty-eight strains, Table IX, showing their geographical origin, is presented. ISo 32 COLD EESISTANCE OF ALFALFA. Table IX. — Record of the geographical origin of the sixty-eight strains of alfalfa grown in the drill-roiv and hill experiments at DicJdnson, N. Dak., in 1908-9. (^ North America (26): Canada (2). 42, 43. Mexico (2). 44, 45. United States (22) : Colorado (1) irrigated. 47. Kansas (1). 48. Minnesota (1). 51. Montana (3) irrigated. 12, 41, 49. Nebraska (3). 26, 50, 53 (dryland). North Dakota (2). 52, 68. South Dakota (2). 50, 54. Europe (15): France (5). 23, 24, 25, 57, 60. Germany (5). 17, 58, 59, 62, 63. Italy (1). '61. Russia (2). 27, 29. Spain (1). 21. Commercial (1). 46. Asia (19): Arabia (4). 6, 7, 8, 9. Mongolia (3). 2, 3, 4. Turkestan (12). 28, 30, 31, 32, 33, 34, 35, 36, 37 Utah (9). 38, 39, 67. Irrigated (4). Africa (3): 1, 5, 10, 64. Algeria (2). Not irrigated (5). 18, 66. 19, 22, 55, 56, 65. Tunis (1). South America (5): 20. Aro:entina (3). 14, 15, 16. Ecuador (1). 13. Peru (1). 11. a The figures in parentheses indicate the number of strains from the country or State named, while those without parentheses refer to the numbers of the rows in which the strain was grown, as shown in Table XIII. Numerous strains known to be tender, especially from foreign sources, were included for comparison. The least hardy of these were the Arabian and Algerian strains, and the Andean variety repre- sented by the strains from Peru and Ecuador. On account of the great interest that has been taken in Turkestan alfalfa, especially in the Northwest, where much has been claimed for it in the wa}^ of hardiness, a selection of twelve strains from that country was included in the test. Weather Conditions During the Summer of 1908. The spring and summer of 1908 were exceedingly favorable for getting a catch of alfalfa. Seeding was completed on May 15 and was followed by prompt germination. A good stand was secured in prac- tically all of the rows, such deficiencies as occurred being due chiefly to poor germination. As may be seen from Table II, the rainfall was ample. Indeed, all the conditions except a frost that occurred 185 EXPEEIMENTS AT DICKINSON, BEGUN IN 1908. 33 on August 22, wliich did not do permanent injury, were more favor- able than had prevailed in any pre^^ous year since experimental work began at tliis place. Growing weather continued ^^dthout interrup- tion until the 27th and 28th of September, on which dates killing frosts occurred. Thereafter the weather grew gradually cooler. KilHng frosts came again on October 7, 10, and 11, and beginning with the 17th continued nightly with one exception to the end of the month. Nearly all growth had ceased in the strains in the nursery by October 23. The minimum temperatures fell gradually during the first hah of November, culminating in zero weather on the night of the 14th. By the end of September, when growth-inhibiting frosts began, all of the alfalfas had become fully estabhshed. They had been clipped once during the summer (on August 5), and the second growth had reached an average height of 12 inches or more. The crowns had stooled out well and the plants were vigorous and of a healthy color, having become well inoculated. In discussing the experiments begun in the spring of 1906, emphasis was laid on the behef that the dryness that prevailed during the fall of that year was one of the very important favorable factors that enabled a part of all of the 22 alfalfas under test to pass unharmed through the unprecedentedly severe winter of 1906-7. This has been discussed in some detail earlier in these pages. The following statement of the rainfall for September and October of 1906 and 1908 shows an important disparity: Average for, -.^r,. 16 years. •^^"^• 190& September. October. .. es. Inches. 1.04 0.25 .70 .14 Inches. 1.67 2.47 AUTUIMN MOISTUEE AND ItS EfFECT ON DORMANCY. In respect to the factor of autumn moisture and its effect on the dormancy of alfalfa plants, conditions in 1908 were vastly different from those of 1906. The figures given above show the wide differ- ence in the quantity of rain that fell. In the humid portions of the United States a precipitation of approximately 1^ inches in Septem- ber and 2^ in October occasions no comment, but in the dry parts of the Great Plains it is A^er}^ unusual. Only one other October (that of 1898) during the 16-year period covered by the Dickinson records had a greater precipitation than that of 1908. The rains of these months fell in such a manner that they were largely taken up by the soil. Furthermore, the amount of moisture available to 44281°— No. 185—10 3 " 34 COLD RESISTANCE OF ALFALFA. each plant is far greater in stands sown in rows and hills than in thick stands secured by the ordinary methods of seeding. The appearance of the experiment on October 1, 1908, is shown in Plate I, figure 1 (frontispiece). As a result of this moist condition, the plants went into the cold weather that came toward the end of October (16° F. on the 23d) in a comparatively unhardened condition. Temperatures inhibitory to growth occurred late in September and at intervals during October, but these did not occur frequently enough nor were they long enough sustained to produce the gradual and complete dormancy observed in 1906. By the middle of November, zero weather having arrived shortly before, it was no longer possible to find growing stems either in the drills or hills. On the other hand, in the field plats mentioned in Tables VII and VIII, which were sown thickly in the usual way, it was possible as late as the 20th of the month to find plants with stems still in a fresh and growing condition. Inasmuch as the winter as a whole dealt much less severely with the field experiments than with the rows, a careful canvass has been made of all possible explanations. It is rarely, if ever, possible to tell positively in such cases as this what particular factors brought about the results observed, but it seems probable that the field plats escaped the injury suffered by the hills and rows for some or all of the following causes: (1) The greater protection due to thick stand, the thicker stand serving to hold the snow; (2) the warmer location on a pronounced slope (the rows are located near the base of this slope); (3) the only cutting of the 1907 and 1908 field plats being done much earlier than the cutting of the rows, the tissues in the plants of the field plats had a chance to become hardened, and finally (4) on account of the dry condition of the soil in the plats as compared with the rows. The last factor might have acted directly upon the roots of the plants through the temperature and indirectly by bringing about a hardening of the tissues of the plants in the field plats. The dryness was brought about by the far greater run-off due to hard, uncultivated topsoil and sidehill location. Amount of Soil Moisture in Alfalfa Stands of Varying Thickness. Unfortunately, determinations of the percentage of soil moisture in the field plats, hills, and drill rows were not made in the autumn of 1908. In the spring of 1909, experiments were undertaken to determine the drying effect on the soil of alfalfa stands of varying thickness. In these experiments Grimm alfalfa plants from a 1908 185 EXPEKIMENTS AT DICKINSON, BEGUN IN 1908. 35 seeding were transplanted into plats with definite distances between the plants. Early in November, 1909, the moisture in the soil in these plats was carefully determined to a depth of 4 feet. The results of these determinations and several others are presented in the following table: Table X. — Results of determinations showing the percentage of soil moisture in the plats of alfalfa broadcasted and planted in drill rows and in kills having stands of varying thickness, Dickinson, N. Dak., November, 1909. Depth. Broadcast. Drills. 3 feet apart. Hills. 16 by 16 inches. Hills. 24 by 24 inches. Hills. 18 by 36 inches. Hills. 36 by 36 inches. First foot Per cent. 12.6 9.7 9.9 8.4 Per cent. 13.9 11.9 10.7 14.5 Per cent. 13.2 13.7 11.6 12.4 Per cent. 14.5 14.7 12.2 16.1 Per cent. 16.6 15.5 17.9 18.8 Per cent. 17 6 Second, foot 16 8 Third foot 14.7 Fourth foot . 18.7 It will be noted that in general the amount of moisture in the soil varies directly with the thickness of the plant covering, the thickly broadcasted plats having the least and the hills 18 by 36 and 36 by 36 inches apart the most moisture. That a similar con- dition prevailed in the autumn of 1908 to that shown in Table X for 1909 can not be doubted." Moreover, it seems likely that the percentage of moisture in 1908 was much greater than in 1909, as the following comparison of the rainfall for the two years would indicate : 1909. September . October Inches. 1.67 2.47 Indies. 0.83 1.08 The distribution of the rainfall in 1908 was such that a greater degree of saturation probably resulted than in 1909. September, a The soil moisture figures for both 1908 and 1909 are available for certain corn and wheat plats. The following percentages are the average for the top 3 feet of soil: Corn. Wheal. Date. A. B. Date. A. B. Oct 21 1908 Per cent. 21 14.8 Per cent. 21.4 14.6 Oct. 21, 1908. Per cent. 15.5 14.9 Per cent. 16.4 Nov. 11, 1909 Nov. 2?,, 1909 14.8 These plats received the same treatment both years. 185 36 COLD RESISTANCE OF ALFALFA. 1908, had eight clays with rain; September, 1909, nine da3^s. Octo- ber, 1908, had six days with rain; October, 1909, had three. The moisture determinations shown in Table X all relate to stands less than a year old. To determine whether similar conditions prevail in stands of greater age, with a varying thickness of the plant covering, 4-foot borings were also made in various places in the 1908 seedings. The results, which are corroborative, are pre- sented in Table XI. Table XI. — Results of determinations shoiuing the j^ercentage of soil moisture in the 1908 experimental field plats and drill rows of alfalfa at Dickinson, A^. Dak., in November, 1909. a Depth. Thick stand. Plat 21217. Thin stand. Plat 12784. 206 plants on 1,200 square feet. 171 plants on 1,200 square feet. 64 plants on 1,200 square feet. First foot Per cent. 10.7 9.7 9.5 8.3 Per cent. 11.9 11.7 11.5 11.4 Per cent. 17.7 13.2 10.9 11.1 Per cent. 16.5 14.0 15.1 15.6 Per cent. 18.2 Second foot 18 5 Third foot 20.1 17.0 9.5 11. G 13.2 15.4 18.4 a The moisture determinations presented in Tables X and XI were made bj' Mr. John C. Thysell, through the courtesy of Prof. E. C. Chilcott, Agriculturist in Charge, Dry-Land Agriculture Investigations, Bureau of Plant Industry. As in the case of the 1909 seedings shown in Table X, the quantity of moisture varies directl}^ with the thickness of the stand, the area with the thinnest plant covering having on an average practically twice the amount of moisture of the thickest stand. The thick and thin stands showm in the first and second columns of this table are located on Area IV, as shown in figure 1, and ar^ adjacent. The Utah strain (S. P. I. No. 12784) was seeded in 1907 and winterkilled badly during the winter of 1908-9, so that only about 15 per cent of a stand remains on this, the best portion of the plat. The commercial sand lucern plat was seeded in the spring of 1908 and the portion on which the moisture determination was made has an estimated stand of 95 per cent. The determinations shown in the third column of Table XI are froixi borings made between drill rows 39 and 40 (commercial Turkestan and Wheeler's acclimatized Turkestan), as shown in Table XIII; those in column 4, between rows 51 and 52 (Minnesota and North Dakota Grimm) ; those in column 5, between rows 58 and 59 (two samples of commercial sand lucern from Germany). It is evident from these differences in the amount of soil moisture that just comparisons in regard to the winterkilling of alfalfa can not be made in the future without taking the differences into account. If, as appears likely, a higher moisture content produced a severer test 185 BEGUN IN 1908. 37 condition, an accurate comparison of winterkilling can not be made between the rows in question, as the ones with a thinner stand will have been subjected to more trying conditions than the others. The result of these moisture determinations also indicates that the pro- gressive thinning of a stand, which goes on more or less rapidly accord- ing to the hardiness of the strain used, is in itself a menace to the success of any field. Weather Conditions ditking the Winter of 1908-9. On November 25 and 26, 1908, no less than 7 inches of snow fell, but a severe blizzard on November 30 blew this into banks, leaving the alfalfa nursery, which is located on the north side of a shelter belt of young, deciduous trees, practically bare of snow. On December 1 a minimum of — 10° F. was recorded. There was nothing particu- larly worthy of note in the weather of December, but as compared with the same month in 1906 very little snow fell. In 1906 the total fall for the month was 7.5 inches, with a like amount on the ground at the end of the month, while in December, 1908, the total fall was 2.4 inches, with only 2 inches on the ground at the end of the month. In the alfalfa nursery, on account of the lack of a sufficiently dense plant covering to hold the snow, there was even less than this. This resulted in the rows and hills having practically no snow covering to protect the plants. Plate II, figure 1 , illustrates more effectively than can be described the great importance of even a light covering of snow. The small patch of red clover seen in the middle of the foreground is all that remained on a series of plats immediately adjacent to the 1907 field plats of alfalfa. These plants have lived over without doubt as a result of their being protected by 2 or 3 inches of snow. Extending along the right edge of the clover is the rather high ridge of another season's back furrow, which held a sufficient crust of snow to prevent the extermination of this group of plants, while the loss on the remainder of the plats was 100 per cent. A similar phenomenon explains the living over of the alfalfa plants, but in this case the growth made during the previous autumn which was not removed served to hold the snow. This point will be more fully discussed later. Table XII shows the complete weather record, except the wind direction and the character of the day, for January, 1909, at Dickin- son. It is presented here to make possible a careful comparison of the weather of this month with that of Januarv, 1907, which is shown in Table III. 185 38 COLD EESISTANCE OF ALFALFA. Table XII. — Daily record of temperature and precipitation for the month of January, 1909, at Dickinson, N. Dak.; latitude, 46° 50^ N.; longitude, 102° 50' W.; altitude, 2,453 feet. Temperature. Precipitation. Date. Maxi- mum. Mini- mum. Range. Record at 6 p. m. Snow- fall. Depth of snow on ground at time of ob- servation. - 1 ° F. 23 39 31 20 -16 -15 - 8 -13 -15 -13 - 6 6 - 3 - 6 0 41 32 41 34 41 40 34 34 34 37 36 44 37 15 12 38 ° F. -12 2 18 -17 -25 -31 -19 -25 -25 -21 -28 -20 -18 -20 -17 - 3 - 5 21 14 28 22 31 15 18 9 19 20 3 - 5 -15 4 •^35 37 18 37 11 12 10 8 22 26 15 14 17 44 37 20 20 13 18 3 19 16 28 17 24 34 20 27 34 ° F. 14 Inches. MONTHLY SUMMARY. 2 Jemperature. 3 19 -16 -20 -17 -16 -20 -19 -18 -14 - 5 -10 -13 - 1 6 28 26 30 40 31 33 25 27 29 29 38 12 - 5 5 26 4 0.2 .1 Mean maximum, 18.5° F. 5 . .. 6 Mean minimum, —2.6° F. 7 •7 .1 .2 .2 8 Mean, 7.85° F. 9 10 Maximum, 44° F.; date, 27th. 11 12 Mmimum, -31° F.; date, 6th. 13 14 Greatest daily range, 44° F. 15 .7 .5 16 Snow. 17 18 a T. Total fall, 2.7 inches; on ground 19 20 15th, — inches; at end of month, 21 22 very little, and that in spots. 23 24 REMARKS. 25 26 Chinook on the 16th. 27 28 High winds on the 4th, 5th, 9th, 29 30 14th, 15th, 16th, 17th, and 18th 31 of the month. Sum .... 574 -82 2.7 18.5 -2.6 a T.=trace. January, 1909, was the coldest month of a winter of nornial tem- perature (13,8° F. for December, January, and February), during which all strains of alfalfa, including the very hardiest, winterkilled, from a few per cent to 100 per cent. January, 1907, on the other hand, was the coldest month of a winter of extraordinary severity, through which some plants of all strains, including the very tenderest, passed uninjured. An inspection of Table XII shows that for a period of two weeks, extending from the 4th to the 17th of the month, the minimum temperature fell from —3° to — 31°F. daily, while during one week of this period the maximum did not rise above —6° F. The snow- fall for the month was light, amounting to 2.7 inches. There were seven or eight days with high winds, so that the small amount of snow that fell did not remain in place except where there was some- thing adequate in the way of a plant covering to hold it. As a 185 BEGUN IN 1908. 39 result of these winds the alfalfas had practically no snow to protect them. A chinook, a climatic factor that is sometimes very trying to per- ennial crops and winter grains in this section, occurred on the 16th of the month. On this date the mercury rose rapidly from —3° to 41° F., a range of 44 degrees for the day. The mean temperature for January, 1909, 7.8° F., was almost 12 degrees higher than that for January, 1907, —3.9° F., and a trifle more than 3 degrees lower than the normal for sixteen years. Hence, while this month was considerably warmer than the same month in 1907, it was nevertheless colder than the average, and, considering other climatic conditions, furnished a splendid test of the hardiness of alfalfa. The mean temperatures of February and March, 1909, were higher than usual, as may be seen by an inspection of Table IV. From February 7 to 15 the minimum temperatures ranged from -9° F. on the 7th to -25° F. on the 14th. Fortunately for the plants a snowfall of 5 inches occurred between these dates. No severe winds to blow the snow into drifts were noted. March weather was much less trying than that of 1907 and 1908. In 1907 the range for this month was from -8° to 69° F.; in 1908, from - 17° to 54° F. ; while in 1909 the lowest temperature was 7° F. and the highest 58° F. EFFECT OF WINTER OF 1908-9 ON ALFALFA VARIETIES. The spring awakening in those strains that endured the winter with any marked degree of success came about the usual time, beginning with the middle of April. In many strains whose mortality was high those plants that survived did not begin growing until a month later. In order to make possible an accurate comparison of the resistance to cold of the sixty-eight strains used in the 1908-9 experiments, Table XIII is presented. 185 40 COLD RESISTANCE OF ALEALFA. S o^ ?; ^^ ;S oo ?^ CO 2^ txj ?«. C ^ ;::^ ^ '^ a, I ft:! X h^ft 5ft u )(NOO OOiOOiOOOJ 005000 to O O (MOC^Otr^OcD t^r lO O lO O O 05 O 00 05 00 CO 00 t^ OJOOOOOt^OJ OOOOCDiOOSt^ K^OOOOOOO OOOOOOO OOOOOOOO OOOOOOOilM ■^oooo ooinoiooo o o lo o r- 02 (M 1 I 'Ji (M !>. ■* 005 lOO O O ^ O 00 o -^ iMOiOOO C^OOO--l.-iiO OOOt^cOi-HCO I CO 00 O O O O 00 O 00 < ■ T-l • • (M • • • C3 C3 CS^-* — o ill 'go - 'ss - © O O O >i =3 S§§§a2 : dog :e^ o a w pi O o C 3 C= S 0^ 3 ce 3-5 « c3 I- o , T3 3 oj ii , " "i s- --'C3 o c3 ^ ^ 3 . St; ?^ 3 s-i w ra O c3 • I CO (N CJ < -J 05 rt --I 1 i-H ^ C<1 (N (M I Cq T-l ,-1 (M .-< i-( (N i-( T h^hh^^W^_;hh>-; Mt-St-nW^ p^ (i; hJ H^ pL^ PlJ p^ pL^ PlJ (ij iJ fC p^ fLj Ph CL| 1-5 Ph Ph Ph Ph Ph Ph Ph Ph (l^ Ph Ph Ph Ph Ph Ph P^ Ph ai «i Ph Ph CQ M 02 02 oj oi Ph oq od 02 02 02 Ph oi 02 03 o5 mxAmmmmm znmmmxAm 1-1 C^ CO ■* lO CO t^ 41 M->*'COOi-l(Mt^ lM-0^0;-^.-i?occc4 -^ot^o-io^cj -^oo'-oc^jo OOOOOiOC^CO 0^0000030000 00 t^ OOCOO 00 r- 00 O O C3 O OO o o o o> o O'^C^'^iOiOtO iC*^Cii^ooio>o -^ -^ ac ci c6 -^ tA t^t^t^ccoJc5o6 ^-llOl-^■-^l-^co OCO'OOOC^ CO (N'^'OOt^t^OOOO coo t^COOO OOiOiOOCOCOOi OOOOClt^COOO CO ^ TT CO t^ O O -^OGOOOOI 'OOt^OCO-^ t^OCMOO^t^ t^l^LOO-^t^ ir-^COiO-^iO-^ •OCO-^CO'T'-^"^ lOlO-^-^-^'OiO iQiO'O'O'OiOiO iO'O'5'iOiOCO iC3O00-^OCl OdOO-^t^O 0500C30COIOO Ol 1-1 CO -^ CO -'I' (M 1-1 rt T-l -7' -^ i-( CO 1-1 (M iM I UO lO »0 »0 -^ ■* -^ i-t CO £ Si 3 C3 o 0~ ■ o^ ' ^ c3' Q . O c3 ! >• .Q'3 ■ M.2 . M c3 • O 5 5 ill 2 o""' • o jj-^ — ;sfe ••sis i^^ i'S -5^-5 2 OOCmoPh •o^oor^-iMiMr^ or^i-ic^ioocr^ (Nioococoio-^ co-^oior~c;t^ OOOOOIO^COIM CO -^ rt rt 1-1 !M CM i-l.-(i-l(N(N i-H 1-1 CM 1-1 i-(i-( 1-1 T-l 1-5 1-5 1-5 1-5 h^ " 1-5 ("HMt li— II— IS-l"^ CQ OQ CQ CQ OQ Ph cd OQ 02 «2 CQ 02 CO 02 (ilPH'p^^a;'-^!^ i^p;fL;p;fi;pHPH p^phPhPh'si^i-] 02 02 02 P^ 02 Pi P^ P-i CO Oi cd 02 OJ 02 CO02O2CQ0QPh •^ lO O lO lO lO iC 185 42 COLD RESISTANCE OF ALFALFA. The thirty-four varieties named in the first half of this table, comprising the east section of the nursery, are shown in Plate I, figure 2 (frontispiece), just back of the middle-ground. The photo- graph was taken June 10, 1909. Row No. 1 is designated by the white stake to the left of the figure 1 in the plate, and extends from this stake to the one next to the figure 35. The latter designates the row of this number. The stake labels stand at the head of the rows, or east end, while the photograph was taken from the foot of the rows of the west section, which comprises the varieties shown in the second part of Table XIII. The stakes scattered about in both the east and west sections indicate plants that were marked in the autumn of 1908 for work in selection and studies in correlation. The stakes appearing beyond the row from 1 to 34 are in another section of the nursery started in 1909, which is not a part of the present discussion. In September, 1908, the number of living plants in both the hill and drill rows was determined by actual count. This count was made with very great care, and no row was left until the person counting was certain of the correctness of the count. On June 15 another count was made of the living and the dead plants in all rows and this was compared with the count of the previous autumn. In this way the percentage of winterkilling for the various strains was secured, as shown in Table XIII. The loss of the sixty-eight strains may be summarized as follows: 12 strains killed out, 100 per cent. 16 strains killed out, 90 to 99 per cent. 14 strains killed out, 80 to 90 per cent, 7 strains killed out, 70 to 80 per cent. 5 strains killed out, 60 to 70 per cent. 4 strains killed out, 50 to 60 per cent. 2 strains killed out, 40 to 50 per cent. 4 strains killed out, 30 to 40 per cent. 1 strain killed out, 20 to 30 per cent. 2 strains killed out, 5 to 10 per cent. 1 strain killed out, 1 to 5 per cent. For the sake of convenience the winter resistance of the strains under experiment will be taken up on the basis of their immediate geographical origin, as shown in Table IX, although this may lead in some cases to combining in one group several things that differ materially, both physiologically and botanically. 185 EXPERIMENTS AT DICKINSON, BEGUN IN 1908. 43 WINTER RESISTANCE OF ALFALFAS IN DRILL ROWS AND HILLS IN THE 1908-9 EXPERIMENTS. Alfalfas from South America. Five alfalfas from South America showed an average mortality of 99.6 per cent. These strains were known to be nonhardy but were included in the experiment for purposes of comparison. The Peru- vian, P. L. H. No. 3260, which lost 100 per cent, was grown at Yuma, Ariz., from the original S. P. I. No. 9303, described in Bulletin 118 of this Bureau. By reference to Table I it will be seen that 1 and 5 per cent of No. 9303 endured the severe winter of 1906-7 at Dickin- son. This same strain suffered a loss of only 12.8 per cent of its stand at Stockton, Kans., during the winter of 1908-9. The strain from Ecuador is the famous Guaranda alfalfa, a short account of which also occurs in the bulletin cited. It is practically identical with the Peruvian and was completely killed out at Dickin- son, N. Dak., while at Stockton, Kans., the average loss of two Guaranda samples was 10.6 per cent. The Argentine samples resemble the foregoing in some respects, but on the whole their affinities are more strongly toward the Chilean type represented by our own Utah and California sort. Alfalfas from Africa. Rows 18 and 20 (S. P. I. Nos. 9322 and 12846, respectively), which killed out 100 per cent, are the highly prized Oasis alfalfa of the Sahara. The tenderness to cold of the Oasis variety will confine its usefulness to the warmer parts of the Southwest, where it is of promise because of the fine quality of its hay. It winterkills severely as far south as Kansas, as may be seen from Plate II, figure 2, which is more fully described later in discussing the strain from Arabia. The photograph used for this illustration was taken on the farm of Mr. E. Bartholomew, near Stockton, Kans. The experiments there, as well as those at Dickinson, showed that merely living through the winter is not a sufficient recommendation for the use of a certain strain of seed. The variety must be so hardy that it will emerge from winter fit to take advantage of the first growing temperatures of spring. Nonhardy individuals of Oasis alfalfa No. 12846, the row shown in the middle of the picture, severely injured by the mild winter, did not recover sufficiently to begin a visible growth until a month later than the more resistant individuals. Of this strain 51.5 per cent killed out at Stockton and 100 per cent at Dickinson. No. 9322 did not suffer as severely in comparison, losing 35.2 per cent at Stockton, while at the northern station its loss was also 100 per cent. 185 44 COLD EESISTAXCE OF ALFALFA. The third Xorth African strain (S. P. I. Xo. 12S03, row 66) is the one gTO^\Ti in rows at Setif. Algeria, and referred to in a preceding footnote (pp. 9-10) of this paper. Seventy-nine per cent of the plants of this variety, which appears very similar to the Poitou alfalfa of France. T\dnterldlled. The average number of plants v-interkihed for the three African strains at Dickinson was 93 per cent. Alfalfas from Europe. There were inchuled in the experiment lo samples from European sources. The average loss of phmts for all strains was ST. 9 per cent. Per cent. 2 Russian strains lost an average of 83. 9 5 German strains lost an average of 83. 1 5 French strains lost an average of : 89. 6 1 Italian strain lost 98. 7 1 Spanish strain lost 100 1 commercial strain lost 79. 5 One of the German strains (row 17, P. L. H. No. 3321, wliich occurs in the inventories of the Office of Foreign Seed and Plant Introduc- tion under Xos. 22467 '^ and 24767 ^) is of sufficient interest to deserve a preliminary note in the present connection. This strain is known as ''alt-deutsche frankische Luzerne/" and all evidence thus far secured points to its being the progenitor of the highly esteemed Grimm alfalfa, which is mentioned later in these pages. It takes its name from the circles of Franconia in northern Bavaria, and the present sample was secured only a few miles from the ancestral home of TVendelin Grimm, who introduced alfalfa into Minnesota in 1857. It is interesting to note that tliis strain has the same chstinctive habit of growth as the Grimm. The appearance of a majority of the rows of European alfalfa, or lucern, after the ^vinter of 190S-9 is shown in Plate III, figure 1. The old German variety preA'iously mentioned is row 17 and can be seen best in Plate I. figure 2 (frontispiece). It forms the continuation of the right-hand row of Grimm alfalfa near the center of the picture. Row 46 (S. P. I. Xo. 11275), a commercial sample purchased in the open market, appeared from its botanical characteristics to be of European origin. The seedsmen who furnished it state that it was purchased fi'om an Italian source. Tliis number suffered to the extent of 79.5 per cent. It resembles strongly S. P. I. Xo. 12S03 from Setif; Algeria (row 66). a See "Seeds and Plants Imported during the Period from Januar\'l to March 31, 1908," Bulletin 137. Bureau of Plant Industiy, U. S. Dept. of Agriculture, 1909, pp. 56-57. & See '"Seeds and Plants Imported during the Period from January 1 to March 1, 1909," Bulletin 162, Bureau of Plant Industry. U. S. Dept. of Agricidture, 1909, p. 42. 1S5 EXPEKIMENTS AT DICKINSON, BEGUN IN 1908. 45 Alfalfas from Asiatic Sources. The discussion of the strains from Asia falls naturally into three groups. Taking up first the Arabian strains, rows 6, 7, S, and 9, their loss was 100 per cent for each of the four numbers. As far south as Stockton, Kans., these same numbers killed out 100 per cent, though the ^dnter there was an extremely mild one. The condition of these. rows at Stockton on June 1 is sho^\Ti in Plate II, figure 2. The two rows in the middle of the picture shomng a certain percentage of hving plants are African strains, the one having the better stand being identical ^dth Dickinson row 20, Kebilli Oasis alfalfa (S. P. I. No. 12846), wliile the other is from Tripoh (S. P. I. No. 12847). The pairs of rows on either side are the Arabian numbers. TURKESTAN. Among the twelve strains listed as from Turkestan three were pro- duced in the United States from Turkestan seed of known origin. S. P. I. No. 13999 (row 33) was growTL on unirrigated land in Utah in 1904 from S. P. I. No. 991 (see row 28), which was so^^m in 1900; S. P. I. No. 19968 (row 37) was grown in Kansas in 1906 from S. P. I. No. 1295, from Samarkand, so^vn in 1899. S. P. I. No. 20437 (row 38) was produced in 1906 b}^ the same field as No. 13999, noted above. The average loss in stand of these three numbers was 84.3 per cent. Inasmuch as practically all of the Turkestan seed used b}^ American farmers comes from a great range of unknown conditions of soil and climate under the commercial name '' Turkestan," four samples pur- chased in the open market were included in the experiment. These samples are rows 35, 36, 39, and 67. The average winterkilhng of the strains was fairly low, 58.6 per cent, but the range was from almost 36 per cent (row 39) to almost 93 per cent (row 35). When one con- siders that Turkestan has an area almost as great as that part of the United States h"ing west of the Mississippi and a climate that ranges from as cold as Montana to as hot as Arizona, this range in the hardi- ness of commercial samples of seed is not surprising. It suggests also the danger confronting the northern grower who purchases seed whose only recommendation is that it came from Turkestan. The last class of Turkestan strains includes ^Ye secured from known sources. These came from a considerable range of territor}', extend- ing from Askhabad, in the southern part of the Transcaspian district, northeastward through Bokhara and Tashkend and northwest to Karabulak, north of the Aral Sea. It is a matter of regret that no sample from Yerni, which is said to be the coldest alfalfa-producing section of Turkestan, was included in this experiment. In the new 185 46 COLD EZSISTAXCE OF .ILFALFA. experiments begun during tlie spring of 1909 a sample of A'erni or Tscliilik alfalfa has been included. The average winter injury to five strains of known origin amounted to 76.2 per cent, the range being from a little more than 56 per cent (row .32 I to almost 91 per cent (row 2S\ The average winterkilling of Turkestan strains may be summarized as foUows: Per cent. Commercial strains {-i) .58. 6 Sn-ains from known sources (5) 76. 2 American-grown strains (3 ) S4. 2 Average for all strains 72. 3 Average for imported strains (^commercial and known sources com- bined ■ t3S. 3 It is apparent from the foregoing and from Table XIII that wliile none of the Turkestan strains in their present condition are hardy enough for the cold Xorthwest, several of them are promising for accHmatization by selective breeding methods. One American-grown strain of Turkestan alfaha. South Dakota Xo. 240 (P*. L. H. Xo. 32.52, row lO). has been assigned to South Dakota and has not been discussed here. This variety was secured through Prof. TT. A. TVheeler and is the progeny of S. P. I. Xo. 991 (see row 2S, Table Xni\. the original imponation of Turkestan alfalfa introduced in 1S9S by Prof. X. E. Hansen. A plat of Xo. 991 was seeded at the Highniore branch of the South Dakota station in the spring of 1S99. The seed used at Dickinson was produced on this plat in 1906. the seventh year after seeding. A comparison of rows 2S and 40. Table XIII, shows that the parent strain. Xo. 991, lost 5S of 64 plants, wliile its progeny grown in South Dakota lost only 7 out of a total of 75 plants. Rows 33 (S. P. I. Xo. 13999^ and 3S (S. P. I. Xo. 20437' are also the progeny of Xo. 991. Some further chscussion of these strains -^^dll be given later. MONGOLIA. The three rows of Mongohan alfalfa were all secured from the same importation. P. L. H. Xos. 2124. row 3, and 2125, row 4. were selected from the bulk lot. S. P. I. Xo. 21232, row 2. on the basis of color and size of seed. Xo. 2124 comprised the largest seed contained in the original lot. It was of rose-purple color, unusually large size, and fairly good germination. So. 5 per cent. Xo. 2125 was somewhat smaller than the former, but considerably larger than the remainder of the bulk lot and of the typical greenish color of alfalfa seed. Xo. 21232 was the original sample, noted above, ^^4th the large seed removed. 1S5 EXPERIMENTS AT DICKINSON, BEGUN IN 1908. 47 The average winterkilling of these three samples was 33.5 per cent. The average for the two seed selections was 39 per cent, the loss being greatest for the larger seed. It would not be safe to assume from this that size of seed is correlated ^-ith hardiness. Nevertheless, the facts are of interest, especially as the Arabian and Peru^^an varieties which are the tenderest alfalfas known also have the largest seeds. The bulk lot containing the smaller-sized seed suffered onty to the extent of 23.5 per cent, showing the lowest percentage of loss among all of the many imported samples included in the experiment. This interesting and valuable alfalfa was secured in 1907 by the Office of Foreio^n Seed and Plant Introduction through Hon. W. W. Rockliill, American Minister to the Chinese Empire, on the basis of a memorandum presented by Mr. Walter T. Swingle, Physiologist in Charge of Crop Physiology and Breeding Investigations. Mr. Swingle, in the course of his studies of Chinese agriculture, found an account of a French missionary,*^ written over forty years ago, of the growing of alfalfa in Mongolia. This led to further investigation and to the ultimate securing of the seed from the Belgian missionaries stationed in the region. The seed came from the high and relatively cold Ordos country, where it has probably been grown for centuries. Rows 2, 3, and 4 in Plate III, figure 2, show the excellent condition of this variety in the summer of 1909. Alfalfas from North American Sources. Of the alfalfas from North America, two came from Canada, two from Mexico, and twentv-two from the United States. Two Canadian strains, purchased in the open market, whose exact origin is unknown to the writers, showed a promising degree of hardi- ness. Their average loss was 45.4 per cent. The hardier of the two, which suffered only to the extent of 39.24 per cent, was secured in Canada in 1904 by what was then the Office of Grass and Forage- Plant Investigations. It is apparent from the botanical characters of these strains (S. P. I. Nos. 13436, row 42, and 21247, row 43) that they are of European origin, a fact to which attention was first directed by Mr. J. M. West- gate, of the Bureau of Plant Industry. An e:S:amination of rows 42 and 43 in Plate IV, figure 1, shows the appearance of the Canadian alfalfa after the wdnter, as compared ^vith. other American strains surrounding it. a David, Pere A., in Flore des Serres et des Jardins de I'Europe, vol. 18, pp. 124 and 126, 1869-70 (Extracts from a letter to M. Decaisne). 185 48 COLD RESISTANCE OF ALFALFA. MEXICO. The average winterkilling of the Mexican rows 44 and 45 (S. P. I. Nos. 11651 and 11652) was 85 per cent. The original source of these strains is unknown, but their immediate geographical origin would have led one to expect a higher percentage of loss. Rows 44 and 45 are shown in Plate IV, figure 1, next to the Canadian numbers. In habit of growth, leafage, etc., they resembled considerably some of the Turkestan strains. UNITED STATES. Among the strains attributed to this country two classes must be recognized: (1) Those that are without doubt ordinary American alfalfas of the Chilean type, of which there are eighteen numbers, and (2) those that are clearly the progeny of introductions from the Old World. Of the latter there are only four (rows 40, 51, 52, and 54), but among them are the alfalfas of the greatest present importance to agriculture in the cold parts of the United States. Ordinary American Alfalfa. — Ordinary American alfalfa, which came to the United States from Chile, has often proved not sufficiently hardy for our severe northern climate. The average winterkilling of the eighteen strains included under this head was 83.3 per cent. The greatest mortality is found in three of the Utah dry-land strains, as follows: Per cent. Row 19, S. P..I. No. 12409, Diamond Fork, Utah .96. 6 Row 22, S. P. I. No. 21867, Nephi, Utah 95. 2 Row 65, S. P. I. No. 22559, Gunnison, Utah 94. 3 Average 95. 4 The hardiest strain of ordinary American alfalfa was a Montana strain (row 41, S. P. I. No. 12747). This seed was produced in the Yellowstone Valley near Billings, where it had been grown for a number of years. This sample suffered to the extent of 43.7 per cent. Next to the above in hardiness was the sample from Clearwater in north-central Nebraska, row 50, S. P. I. No. 12820. This number suffered to the extent of 66.7 per cent. These and several other American strains are shown in Plate IV, figure 1 (rows 41, 47, 48, 49, and 50). Only two other strains of this class showed a degree of hardiness that entitles them to mention. These strains and the proportion of the plants winterkilled are as follows: Per cent. Row 49, S. P. I. No. 12816, Chinook, Mont 70. 9 Row 53, S. P. I. No. 21945, Sextorp, Nebr 75. 7 The latter is a highly esteemed dry-land form grown by Mr. Lewis Brott on the high plains of western Nebraska. Unlike the Utah dry- 185 EXPERIMENTS AT DICKINSON, BEGUX IX 1908. 49 land strains this one is grown in a section that is subject to decidedly low temperatures at times, with little or no snow covering to protect plant growth. The normal autumn rainfall in this section is very light. The question whether American alfalfa of the ordinary sort has undergone any change in hardiness since its introduction, about 1855, has been discussed recently by alfalfa breeders and growers. No experimental evidence bearing on the point has been published heretofore. Hence, the results obtained in the present experiment while by no means conclusive are of more than ordinary interest. The eighteen strains of ordinary American alfalfa showed a great difference in hardiness, as winterkilling under the uniform condi- tions of the experiment ranged from about 44 per cent (row 41) to almost 97 per cent (row^ 19). The mere presence of so high a degree of diversity with, reference to one physiological character is extremely suggestive. Botanically, alfalfa or lucern {Medicago sativa) is an exceedingly diverse species composed of numerous strains, varieties, or possibly subspecies. Even the most uniform of our cultivated sorts, such as the Peruvian and Arabian, comprise numerous strains that are easily recognizable on inspection. Ordinary alfalfa, of course, shares in the diversity common to the species as a whole. That some of these component lines or individuals are more susceptible to cold than others seems axiomatic. Resistance to cold would then depend largely at least on the severity of the climate to which a strain has been subjected; that is, to the amount of ''weeding out" of unfit lines and individuals that has taken place. The selective value of the Montana climate with reference to hardiness is imquestionably greater than that of Utah or most of Nebraska; in fact, it does not seem mifair to suppose that that of northern Nebraska is superior to that of Kansas. Different ^\dnters have different selective or acclimatizational values with reference to various test factors. Very severe winters with heav}^ snows that remain in place have small selective effect; mild ^vinters mth even a moderate snowfall, such as that of 1907-8 at Dickinson, have practically no effect on strains having only the customary degree of hardiness. Tender sorts, like the Arabian, Peruvian, Tunisian, and Tripolitan, are susceptible of selection with even comparatively mild winters, such as that of 1908-9 at Stock- ton, Kans. 44281°— No. 185—10 4 50 COLD EESISTAKCE OF ALFALFA. The following statement shows by States the average winterkilling of common alfalfa in the row and hill experiments at Dickinson: strains. Percent. Utah a 10 90.4 Colorado 1 86.1 Kansas 1 84.8 Nebraska 3 76. 4 Montana 3 65.4 The number of strains from the different States is too small to use as a basis for generalization, but it is hoped that the data here given will lead other investigators to make a careful record of results of similar experiments. Progeny of Old World Introductions. — It is among the progeny of alfalfas introduced from the Old World that we find the highest degree of hardiness, a degree of hardiness that in at least one case might almost be called wonderful. Each of the four numbers included in this class emerged from the winter of 1908-9 mth a notably high percentage of living plants. Their combined average loss is 13.5 per cent, while the percentages of loss for the individual strains were as follows : Table XIV. — Results of experiments showing the percentage of winterkilled alfalfa plants of four strains introduced from the Old World. Row No. Accession No. Source. Propor- tion of winter- killed plants. 52 P. L. H. 3235.. S. P. I. 21938.. P. L. H. 3252.. P. L. H. 3251.. Fargo, N. Dak. (Grimm) (originally from Baden, Germany) Per cent. 2.8 51 40 54 Clearwater, Minn. (Grimm) (originally from Baden, Germany) Highmore, S. Dak. (originally from Tashkend, Turkestan) (S. Dak. No. 240). Brookings, S. Dak. (S. Dak. No. 167. Original source not known, but probably from Europe). Average 7.0 9.2 34.6 13 4 Grimm alfalfa.^ — Grimm alfalfa, which is named after Wendelin Grimm, the German immigrant by whom it was introduced, is the hardiest form of the cultivated plant known at the present time. Its progenitor was brought to Minnesota in 1857 from Ktilsheim, between Wertheim and Tauberbischofsheim, in the valley of the Tauber of Baden, Germany. a This includes row 18, common alfalfa which endured the mild winter of 1907-8 in North Dakota. h A short account of the history and probable method of development of Grimm alfalfa by the senior writer of this bulletin has been published, entitled "The Accli- matization of an Alfalfa Variety in Minnesota," Science, n. s., vol. 28, December, 1908, pp. 891-892. 185 51 The drainage basin of this httle stream, which empties into the Main at Wertheim, known as the Taubergrund, is at the present time the chief seat of the production of seed of ^' alt-den tsche frankische Luzerne" (row 17), already mentioned in these pages. This strain of alfalfa is highly prized in Germany and appears to have been under cultivation in the old Franconian circles of Baden, Wlirttemberg, and Bavaria for several hundred years. The two Grimm strains included in the present experiment proved to be the hardiest of the sixty-eight under observation. The hardier one, P. L. H. No. 3235 (row 52), was produced in 1904 on a plat about 4 years old on the farm of the North Dakota experiment station. This seed was presented to the Department of Agriculture in October, 1904, by Prof. J. H. Shepperd, vice-director and agriculturist. This strain lost less than 3 per cent of the plants in hill and drill rows com- bined. Only one plant of twenty-five in the hills, which are 2| feet apart in the row with the rows themselves 3 feet apart, succumbed during the winter. In the 50-foot drill row only one plant out of forty-seven was killed. Second in hardiness was a Minnesota-grown strain of Grimm alfalfa, S. P. I. No. 21938 (row 51). This seed was secured from Mr. A. B. Lyman, of Excelsior, Minn. It was produced near Clearwater, Stearns County, from Carver County seed. In the hill row one plant in twenty-five winterkilled, in the drill row four out of forty-six, the average loss for both being 7 per cent. As an indication of the vigor that the Grimm alfalfa had after passing through the winter of 1908-9, which was so destructive to ordinary alfalfas, mention may be made of the yield of seed from a tenth-acre plat of Grimm grown just adjacent to the experiment under consideration. This plat was sown in the spring of 1908 in rows 3 feet apart and kept cultivated. The plants were quite tliick in the row. A clipping was made of this plat in 1909 early in the season, when it had attained a growth of about 8 inches. It was not cut again until September 28, when a cutting was made for seed. The plat yielded 33.6 pounds of clean seed, which is at the rate of 336 pounds, or 5.6 bushels, per acre. This is certainly an admirable showing. If, as appears likely, the old German Franconian lucern (row 17) is the progenitor of the Grimm, we have here a further suggestion as to the possibilities of habituating alfalfa to a new or foreign climate. The loss of the former was over 75 per cent, the average for two strains of Grimm less than 5 per cent. Acclimatized Turkestan alfalfa. — Plant Life History No. 3252, row 40 (S. Dak. No. 240), was produced at Highmore, S. Dak., by Prof. W. A. Wheeler, from Prof. N. E. Hansen's original introduction of 185 52 COLD RESISTANCE OF ALFALFA. Turkestan alfalfa from Tashkencl (see row 28), S. P. I. No. 991. In Bulletin 101 ^ of the South Dakota Agricultural Experiment Station Professor Wheeler has given a short account of this strain, a part of which is as follows : No. 240. Seed was obtained from the United States Department of Agriculture under S. P. I. No. 991. This was collected (obtained) by Prof. N. E. Hansen at Tashkend; Turkestan, in 1898. It has been grown in plots at Highmore since 1899 and apparently has not killed out. This is a severe test of its resistance to both cold and drought. A small plot has been cut for seed in both 1905 and 1906 and some seed has been obtained. So far, it does not appear to be nearly as good a seed producer as any of Nos. 162, & 164, c or 167,^ but it can not be compared with these until it has gone through as many generations from seed in this State as have the other three numbers cited. This alfalfa is thought to be of the same introduction as No. 164. If this is correct it will probably become a better seed producer after it has been grown from several seed-producing generations in this State. The quality of this number is about the same as No. 164. This ranks as one of the four best alfalfas which have been tried at Highmore. Numbers 162, 164, 167, and 240 are hardy and are possessed of high forage qualities, so that they can be recommended for the conditions under which they have been tried. Numbers 162 and 167 start earlier in the spring and produce a stronger and more vigorous first crop than numbers 164 and 240. With the second crop this differ- ence is not so noticeable. These four numbers have been tested from eight to ten years in this State and with the exception of number 240 have gone through several generations from seed. The fact that they have reproduced from seed through several generations seems to have more to do with acclimating alfalfa than has the mere fact that they have grown so many years under these conditions from a single seeding. This is where No. 164 leads over No. 240. * * * For the present the work will be restricted entirely to selected stocks from Nos. 162, 164, and 167. Work with No. 240 progresses more slowly because of its poorer seed production. An examination of the published station records shows that although No. 240 has been grown at Highmore since 1899, even approximately accurate data concerning its behavior have been taken only since 1903. The seed used in the Dickinson experiments was produced in 1905, the seventh 3^ear from seeding. This acclimatized Turkestan strain ranked third in hardiness among the sixty-eight strains tested, and appears to show that the parent strain had become profoundly changed in cold endurance by the seven winters spent in South Dakota. This change probably resulted from the successive eliminations of the unfit lines of descent. « Wheeler, W. A., and Balz, Sylvester. Forage Plants at the Highmore Substation, 1906. Bulletin 101, South Dakota Agricultural Experiment Station. March, 1907. b This number is the Grimm alfalfa of Minnesota, grown for several years at Brook- ings and Highmore. c No. 164 is supposed to be the progeny of the same number (S. P. I. No. 991) as the present sample, P. L. H. No. 3252. No. 164 was produced from No. 991 sown at Brook- ings in 1898, while No. 240 was produced from No. 991 sown at Highmore in 1899. d No. 167 is the parent of P. L. H. No. 3251 (row 54 at. Dickinson). 185 EXPERIMENTS AT DICKINSON, BEGUN IN 1908. 53 This strain of Turkestan alfalfa which underwent seven years nat- ural selection in South Dakota presents an opportunity for several important comparisons. These are not presented as final in any sense, but rather to stimulate further investigation. Row 28 was seeded with a small quantity of seed remaining from the original lot of S. P. I. No. 991, the first importation from Turkes- tan. Row 40 is the South Dakota progeny of No. 991. Rows 33 and 38 are the Utah progeny of the same number. Both were har- vested from the same field which was seeded down in 1900. Row 33 was harvested in 1904, row 38 in 1906. Table XV shows the comparative winter resistance at Dickinson of No. 991 and its Utah and South Dakota progeny. Table XV. — Comparison of Turkestan alfalfa, S. P. I. No. 991, ivith certain strains grown from it, showing the percentage of plants winterkilled. Row No. Accession No. Source. Propor- tion of plants winter- killed. 28 S.P.I. 991.... P. L. H.3252.. S. P. I. 20437.. S.P.I. 13999.. Turkestan (imported 1898).. . Per cent. 90.6 40 Highmore, S. Dak. (1899-1906) 9.2 38 Wellsville, Utah (1900-1906) . . 89.6 33 Wellsville, Utah (1900-1904) 98.7 The distance between row 28, the original §tock, and row 33, the 1904 progeny, is only 15 feet, while that between rows 38 and 40 is only 6 feet. All conditions were uniform so far as known, and still this great diversity appears. The greater age of the original seed might be responsible for some of the difference in hardiness between this and the South Dakota-grown strain. But here we are confronted with the fact that the 1904 progeny from Utah — comparatively fresh seed — suffered even more than the foundation stock, while the 1906 crop suffered an almost equal loss. It is not possible in the limits of the present paper to discuss, nor is there sufficient data to warrant discussing, the numerous biologic factors that may be concerned in these diversities. The differences in loss in the 1904 and 1906 Utah strains suggest that the two additional winters endured by the latter may have had greater selective value than those from 1900 to 1904. It appears from the records that this was the case. The lowest temperature that occurred during these years at Logan, Utah, the nearest point to Wells- ville having a temperature record, was — 17° F. on February 14, 1903; but on the 11th and 12th of the month more than 2 inches of snow had fallen, and the observer makes the comment that ^'fields were gener- ally well covered with snow throughout the month, and it is therefore 185 54 COLD EESISTANCE OF ALFALFA. not probable that any damage was done to fall grain by low temper- atures." The minimum temperature during the winters of 1904-5 and 1905-6 was — 15° F.^ which occurred in both mnters. The minimum of the former fell on February 12. Between January 1 and February 12 almost 2^ inches of rain had fallen. This no doubt resulted in a satu- rated condition of the soil that led to considerable injury to the plants when 15° F. below zero weather occurred on the 12th. January had been a warm, balmy month that tended to stimulate new growth in the plants. The minimum temperature of the winter of 1905-6 (-15° F.) fell on January 8. Two inches of snow had fallen during the last days of December. This probably gaye enough protection from the cold. The seeding capacity of this strain, P. L. H. No. 3252, is not satis- factory according to Professor Wheeler; howeyer, it is scarcely more than axiomatic to state that the seed production may be improyed by the selection of those lines within the strain that produce seed better than the ayerage of the strain. If we assume that the yarious lines within the strain remain constant from generation to generation, then we are not modifying the alfalfa in any other way when we improye its seed production. We are merely isolating those lines of high seed production. The strain is merely in the process of puri- fication. On the other hand, it is not at all unlikely that mutations, of greater or less moment, are more or less constantly taking place within the strain. In such a case it would be quite proper to speak of the alfalfa as being modified, though such modifications would be almost purely spontaneous. In other words, we haye not progressed enough in the science of breeding to be able to bring about desired mutations at will. The function of open poUination in alfalfa breed- ing is something about which we haye almost no exact knowledge. One difficulty in interpreting the results from the different sowings from No. 991 is that we are not well enough posted as to the preyious history of the imported seed. The following statement concerning it is found in Inyentory No. 1 of the Section of Seed and Plant Intro- duction : 991. a Me Die A GO SATivA. Alfalfa. From Tashkend. Received through Prof. N. E. Hansen, June 4, 1898. (200 bushels.) Variety "turkestanica.'" This subspecies of alfalfa was obtained from eight different sources varying widely in climatic conditions. It endures droughts which kill European alfalfa. Deemed very promising for trial in droughty regions. See No. 469. ^Foreign Seeds and Plants Imported by the Section of Seed and Plant Introduction. Inventory No. 1, 1898, p. 81. 185 EXPEEIMEKTS AT DICKINSON, BEGUN IN 1908. 55 469,0 Medicago sativa. Alfalfa. From Turkestan Agricultural Society, Turkestan. Received through Prof. N. E. Hansen, March, 1898. Native alfalfa. Endures drought better than European alfalfa. Eows 28, 38, and 33 run reasonably close together in hardiness. The deviation is perhaps no greater than would be expected from any run of commercial seed. Whether the seed from the eight original sources was mixed or not, a reasonable hypothesis would be that some of the hardier seed was planted in South Dakota. Even this seed was probably a mixture representing more or less hardy strains. By the result of natural selection the hardiest lines or strains were brought to the seeding stage in South Dakota and then transferred to Dickinson. It is an interesting fact that No. 991 was evidently imported more for its drought resistance than for any potential hardiness. South Dakota No. 167. — Although the source from which the strain of seed, South Dakota No. 167, originally came is unknown, it is obvious from its botanical characteristics that it is of European origin.^ The name Baltic was first applied to it in 1906 by Prof. W. A. ^Yheeler, formerly botanist to the South Dakota Agricultural College and Experiment Station, from the fact that it had been grown for about ten years near the little village of Baltic in ^linne- haha County, S. Dak., and not as has been supposed in the Baltic Sea region of Europe. The appearance of tliis strain, which is P. L. H. No. 3251, in the spring of 1909 is shown in Plate III, figure 1, row 54. Although shoAving a promising degree of hardiness, at Dickinson this strain was the tenderest of the four hardy American races derived from the Old World introductions. In the drill rows, where it is possible that the thicker stand by binding more snow may have been afforded sUghtly more protection, the loss was 34 per cent, while in the hill a Op. cit., p. 42. b Mr. J. M. Westgate, of the Office of Forage-Crop Investigations, first directed atten- tion to the probable European origin of this alfalfa. Mr. Westgate noted the very inter- esting fact that several European alfalfas show greater diversity in flower color than do our ordinary American or the Turkestan types. In the Dickinson experiments the diversity is about equal in the Grimm samples (rows 51 and 52), South Dakota No. 167, the so-called Baltic alfalfa (row 54); " alt-deutsche frankische Luzerne" (row 17); commercial or false sand lucern. No. 20896 (row 60); and first-quality com- mercial alfalfa, probably from Italy, No. 11275 (row 46). It occurs to a less degree in Simbirsk alfalfa, No. 13857 (row 27); in Poitou alfalfa. No. 12695 (row 24); Italian alfalfa, No. 22416 (row 61); and Algerian alfalfa. No. 12803, from Setif (row 66). At Stockton, Kans., a very similar degree of diversity in color was found in Italian alfalfas, Nos. 12696 and 13481. A much greater variegation in color was found to exist in commercial sand lucern, Nos. 21187, 21217, 21269, and 22418 (rows 57, 58, 59, and 63), though the last number showed less diversity than the other three. 185 56' COLD RESISTANCE OF ALFALFA. rows, where the plants were subjected to the severest test, the loss was 36 per cent. It may be of interest to state that the entire nursery killed out 2.3 per cent niore than did the plants in the drill rows. Climatic conditions at Highmore, S. Dak. (altitude 1,860 feet), where the seed generation used in the present experiments was produced, are no doubt less severe than at Dickinson (altitude 2,453 feet), which is located about 300 miles farther northwest on the Great Plains. Professor T\Tieeler found the Baltic alfalfa very promising at both Highmore and Brookings and pubhshed liis observations in Bulletin 101 of the South Dakota station, alread}^ cited. His notes are of sufficient interest to be quoted in full, as follows : No. 167. The original source from which the seed of this number was introduced into this country is unknown. Mr. "W'. F. Kelly, of Renner, S. Dak., purchased the seed about ten years ago from a seed dealer at Hartford, S. Dak. He and Mr. E. C. Evans of the same place have grown it for about ten years and consider it better than other alfalfas tried in their vicinity. Mr. Kelly furnished several pounds of seed to the Agricultural College in 1905. A two years' trial of the variety is rather short upon which to draw conclusions. In all our tests, however, both at Highmore and Brook- ings, this number has shown itself equal to any in quality, hardiness, and seed pro- duction. It has not been sown in the selection rows and so has not been put to quite so severe a test of hardiness as No. 162 (Grimm alfalfa), and as there is no record of its ha^'ing been tested under northern conditions in the past, as has No. 162, it can not be compared in hardiness to the latter number. In seed production it is equal if not superior to No. 162. In quality and the vigor of early spring growth it appears to be equal. The test in the nursery rows at Dickinson was probably more severe than any this strain had previously endured. Hence, although not the equal of a few other strains it is decidedly valuable. It ranked fifth among the sixty-eight strains at Dickinson. Of the new intro- ductions only one, MongoHan No. 21232, excelled it in ^\'inter resist- ance. It ranked high as a seed producer at Dickinson in 1909. COMPARISON OF CONDITIONS IN FIELD PLATS AND IN ROWS AND HILLS. The hill and row experiment just discussed represents the severest possible test of the winter resistance of alfalfa for the weather con- ditions that prevailed while the work was in progress. Fortunately, the two small series of field plats already discussed, which are on a rather pronounced southern slope above the alfalfa nursery, offer in comparison what might be termed decidedly more favorable condi- tions than alfalfa is likely to be gro^\m under in most years in this region. Except for January the ^\dnter of 1908-9 was no more severe than may be expected any year. One set of the field plats containing four different strains was seeded in the spring of 1907. These plats should have been cut in ordinary farm practice twice during their second year, as it would 185 CONDITIONS IN FIELD PLATS AND IN ROWS AND HILLS. 57 not pay to leave such a heavy stand of alfalfa on the field at the advent of winter. They were cut onl}^ once (on July 20) and the second growth was left on the ground to furnish winter protection. Precipitation to the amount of 0.65 inch occurred during three cloudy days preceding the cutting and a copious rain fell a few days after the cutting. As a result, recovery was rapid and growth con- tinued until well into the autumn. Indeed, the growth of over two months and a half remained on the ground. The favorable conchtions under which these plats entered the winter could scarcel}^ be duphcated; they were in their second year, had enjoyed ample moisture for continued growth through the season, had been cut only once, and that in midseason, leaving a good stand to bind the snow. In spite of this the four varieties lost from 15 to 90 per cent of their stands, the loss varying greatly as we have seen on different parts of the plats. On June 10, 1909, a Turkestan strain from Tashkend, S. P. I. No. 14786, stood best of the four, foUowed closely by S. P. I. No. 13857 from Simbirsk, Russia. The loss in these strains based on careful estimates was 15 and 18 per cent, respectively. An irrigated strain from Utah (S. P. I. No. 12784) and one from Colorado (S. P. I. No. 12398) lost about 80 and 90 per cent, respectively. A slight portion of the loss in these two cases occurred in the mild mnter of 1907-8, but most of it was due to the mnter of 1908-9. The snow protection during the former was such that only the really tender kinds suffered even slightly. The other two strains, which were known to have a certain amount of hardiness, maintained full stands through the winter of 1907-8. The nine other field plats were sown in May, 1908. This was an unusually favorable year for getting a stand of alfalfa. The nursery rows as well as the field plats shared in the beneficial conditions that prevailed, the latter in a higher degree than the former, as in dry or even ordinary A^ears fields sown by the usual methods suffer consider- able hindrance in growth because of a lack of moisture. This results in the autumn in an insufficient covering of vegetation to hold the snow that is necessary to furnish a wdnter protection to the plants. April, 1908, had a Kttle more than the normal rainfall. May, the month in which seeding was done, had a precipitation of 3.50 inches. This is more than an inch above normal and insured rapid germination and a good stand of young plants. June rainfall amounted to 4.30 inches, about Ih inches above normal. This enabled the plants to become firmly estabHshed and to make a fine growth throughout July, when f inch less than the normal amount of rain fell. These plats were clipped a few days after a heavy rain that feU July 24 and about ten days earlier than the experimental rows in the nursery. They began a prompt recovery which was no doubt gradu- 185 58 COLD EESISTANCE OF ALFALFA. ally retarded by a hot dry spell that lasted until August 12. The hill and drill rows, which together contained an average of about seventj-fiye plants of each variety, were cut on August 5 in the midst of this dry time. As no rain fell until August 12 they were at some disadvantage in making a recovery. The field plats, on the other hand, had already begun growth and were forced to harden their tissues to a considerable extent. The presence of this hardened tissue was a factor of undoubted value in wintering over. During the second half of August the rainfall was ample ; the short- age for the whole month amounting to less than J inch as compared ^\ath the 16-3^ear normal. September's rainfall (1.67 in.) exceeded the normal more than J inch, while October's precipitation (2.47 in.) was over IJ inches more than this month usualty receives. The rainfall of these two months had opposite effects on the row and field experiments, affecting the former unfavorabl}^ and the latter favor- ably. This seems paradoxical, but really is not. In the first place because of their sidehill location and hard uncultivated topsoil, as compared with, the rows, the run-off was greater from the field plats than from the nursery. The moist autumn benefited the field plats by enabling them to make a late growth through the autumn. This increased their snow- holding capacity. At the same time on account of the run-off and the far greater transpiration from a thick stand, the dangers attendant upon an excess of moisture in the soil during the mnter were avoided. In the nursery rows the opposite was true. There was a sufficient growth of plants in height, but the stand was too thin to hold much snow. The bare ground, the distance between the plants, and the cultivation resulted in a dangerous saturation of the soil. In the final analysis it was found that the presence of excess moisture in the soil and the absence of snow covering killed the plants in the rows, and the presence of snow covering and the absence of excess moisture enabled the plants in the field plats to five. It was dry enough in the plats to cause the formation of hard tissue without intercepting growth and wet enough in the rows to cause a sappy growth to con- tinue until hard frosts stopped growth altogether. The season and treatment were such that when winter came the plants in the field plats had made ample growth to hold the snow. Thickness of stand combined with height increased the snow-binding capacity. The thick stand, conducive to a greater transpiration, and the topography of the field, which promoted the run-off, pre- vented an excess of moisture in the soil when cold weather came. The opposite was true in the case of the rows. Furthermore, the field plats have a warmer, more protected situation, due to a southern slope and the location below the brow of a ridge. The rows are 185 CONDITIONS IN FIELD PLATS AND IN ROWS AND HILLS. 59 located near the base of this ridge in such a way that they are rather more exposed to the prevaihng winter winds which come from the northwest. The row test is the severest to which alfalf.a can be put in the area which Dickinson typifies, but as this general region, which includes large parts of North Dakota, South Dakota, and Montana, has a normal rainfall of less than 18 inches, row cultivation of alfalfa may prove to be the best farm practice. If this proves to be the case, the alfalfa suitable for this area must be able to bear this rigorous test successfully. The rainfall for September and October, 1908 (4.14 inches), while unusual for Dickinson and all of the region west of the one hundredth meridian in the States just named, is normal in Mnnesota, Wis- consin, and ^Michigan, and not infrequently occurs in eastern North and South Dakota and, as was the case in 1898 and 1908, in the western part of the Dakotas. The alfalfa for this region must be inured to this amount of autumn moisture followed by low tempera- tures with or without snow. The winterkilling of varieties on the 1907 and 1908 field plats has already been discussed. It was found that there was a great diver- sity in loss on the thick and thin parts of the same plat. Where the stand was thin or inoculation poor, greater loss occurred. Ten of the varieties on the field plats were duplicated in the rows. In no case was the loss as great in the former as in the latter, but the com- parative behavior in both cases was similar; that is to say, when ranged in series on the basis of winterkilling, the duplicate strains occupy a very similar relative position. The six strains that were shown to be the hardiest on the field plats rank in the following order: Montana Grimm, S. P. I. No. 21827; Nebraska Grimm, S. P. I. No. 21735; Turkestan, S. P. I. No. 20988; commercial sand lucern, S. P. I. No. 21217; Turkestan, S. P. I. No. 14786; Canadian, S. P. I. No. 21247. In the rows where the test was severer, the ranking was as follows : North Dakota Grimm (row 52); Minnesota Grimm (row 51); Turke- stan, S. P. I. No. 20988; Canadian, S. P. I. No. 21247; commercial sand lucern, S. P. I. No. 21217; Turkestan, S. P. I. No. 14786. Only the Canadian strain in the field plats fails to show the same relative strength. This is probably due to the unsubdued condition of the soil on this plat, which was new breaking, as well as to the original thinness of the stand, and not to any greater inherent weakness in the strain. Among the four weakest strains, in wliich the loss in the rows ranges from about 80 per cent up, the agreement is not so close. 185 60 COLD KESISTANCE OF ALFALFA. From this it is apparent that we are deaUng not so much with a difference in the hardiness of the varieties under comparison, but merely with differences in the amount of loss under tests of varying severity. Some of the factors that contributed to the greater loss in the nursery rows as compared with the thickly seeded plats may be summarized as follows: (1) The 1907 field plats were cut between two rains, two weeks earlier than the rows; the 1908 plats were cut about ten days earlier than the rows and a few days after a good rain; the rows were cut August 5 in the midst of a hot dry spell that lasted over two weeks, the most protracted of an otherwise unusually favorable season for growing alfalfa. (2) The field plats had a thick growth which by transpiration aided in exhausting excess moisture from the soil before winter came and later, together with weeds, furnished lodgment for a protective covering of snow that acted like a blanket in shielding the plants from cold. (3) The field plats have a warmer, more protected situation, due to their location on a pronounced southern slope and below the brow of a ridge. The rows are located near the base of this ridge in such a way that they are more exposed to the prevailing winter winds which come from the northwest. This slope, by promoting the run-off, further aided in preventing the saturated condition of the soil, which undoubtedly acted disastrously on the nursery rows. FURTHER OBSERVATIONS ON THE EFFECT OF AUTUMN MOISTURE. In the preceding pages much emphasis has been placed on the injurious influence of autumn moisture on the wintering over of alfalfa. In order to determine whether similar losses have occurred in other experiments the alfalfa work of the North Dakota, South Dakota, and Minnesota stations has been studied especially. The winter of 1898-99 was one noted for its extremely low temperatures in the Great Plains region. At Brookings, S. Dak., the experiment station had a plat of the original importation of Turkestan alfalfa, S. P. I. No. 991, that lived through the winter successfully. Inas- much as this same stock of seed suffered quite severely in the present row experiments at Dickinson the records were investigated to learn what caused the difference. So far as temperature was concerned conditions were certainly more trying in the winter of 1898-99 than during the winter of 1908-9. According to Prof. E. C. Chilcott's records, during December, 1898, there were at Brookings thirteen days with temperatures below zero, the range being from - 2° to — 22° F. There was no snow during the month. January had fourteen days with minima below zero, the 185 OBSERVATIONS ON THE EFFECT OF AUTUMN MOISTURE. 61 absolute minimuin temperature for the month being —27° F., which occurred on the 30th. The total snowfall for this month was 1.4 inches, which would have made a very light covering in any case, but because of winds was not very evenly distributed. A period of low temperatures began on January 26 and extended until February 13, culminating on February 9 in — 41° F. On the 11th the mercury again fell to — 38°F. Only 2.1 inches of snow fell during February ; hence, the plat of Turkestan alfalfa had only such snow protection as was held by its own plant covering. Between February 5 and 12 the highest minimum temperature experienced was —25° F., while the liighest of the maxima was 5° F. The spring weather was in all respects favorable, there being no alternate thawing and freezing worthy of mention. The severe temperatures endured by S. P. I. No. 991 at Brookings in February, 1899, as compared with the same strain at Dickinson in January, 1909, makes it evident that there must have been some other difference in weather than temperature between the two win- ters to explain the extensive killing in one case and the living over in the other. An investigation of the precipitation record at Brookings shows that during September, 1898, 1.31 inches of rain fell between the 1st and the 15th; that only a little over 1 inch fell in October, all before the 15th, with snowfall on the 18th. It seems evident from this that the difference in autumn moisture and the consequent saturation of the soil was probably the determining factor at Brook- ings in the wintering over of No. 991. That autumn moisture inter- feres with plants becoming dormant agrees in a way with orchard practices in other sections, especially in irrigated areas. Orchardists desiring to harden their trees and force them into dormancy stop irrigating to produce the desired effect. The experiences of the North Dakota Agricultural Experiment Station at Fargo tend to bear out this hypothesis. In 1901 a plat of Turkestan alfalfa was seeded on the station farm. This endured suc- cessfully the winter of 1901-2, but, according to Professor Shepperd, ''failed to pass the winter of 1902-3 successfully, having been badly winterkilled. It was given the most trying conditions by having all of its stem growth mowed close to the ground just before cold weather began in the fall. A second plat of Turkestan seeded in the spring of 1902 which was left with a growth of stems a few inches high killed but Httle during the same winter." The record of rainfall at Fargo during the autumn of 1902 shows that over 5 inches of rain fell in October. This probably accounts largely for the greater amount of winterkilling in the plat from which no moisture was removed through transpiration than in the one left with some growth. Furthermore, the latter had the advantage of a plant covering to hold the snow. 185 62 COLD RESIST AXCE OF ALFALFA. Still other observations have been made which leave Httle doubt in the matter of winterkilling due to an excess of autumn moisture." Prof. W. A. TTlieeler in a paper presented before the Dry Farming Congress, at Billmgs, Mont., in October, 1909, called attention to the fact that the whiter of 1905-6 resulted very disastrously to the nursery row experiments at the South Dakota substation at Highmore. Professor THieeler commented especially on a northern-gro^sTi strain of common alfalfa which was almost wholly killed out. Only the Grimm alfalfa, acclimatized Turkestan alfalfa, and South Dakota No. 167 — the so-caUed Baltic alfalfa — endured thisTSTnter successfully, and there was some killing in the two last named of these strains. The precipitation record for Highmore shows that almost 2 inches of rain fell in October, 1908, while the precipitation for November amounted to 1.29 inches. The alfalfas under experiment that suffered the worst were grown ui rows, as at Dickinson; hence, the cUsastrous effect of this amount of rainfall was no doubt much greater than in thick stands. This same winter (1905-6) is the only one in recent years which has injured seriously the stands of Grimm alfalfa in Carver County, ^lin- nesota. Farmers have generally beheved that the kiUing which took place that winter was due to the alternate tha-vx^ing and freezing dur- ing February. A careful study of Minnesota winters for a period of years shows that many have had more severe conditions in this respect, hence it did not seem hkely that this condition furnished a sufficient explanation. Again, the precipitation records were con- sulted and it was found that the September rainfall amounted to 4.15 niches, the October rainfall to 2.49 inches, while the precipita- tion for November which was practically all rain was almost 2 J inches. It was noted that the killing was especially bad in the '^ draws," or depressions, which furnished drainage to the fields. Even where these ''draws" or natural drainage ditches were on side- hills, as shown in Plate IV, figure 2, where water could not remain standing, the killing was severe. From this it seems that the condi- tion was not one of an ice covering, but of saturation. These examples have been selected to give further emphasis to the importance of autumn moisture. It has frequently been the case in irrigated sections that late irrigation has caused the kiUing out of alfalfa fields. However, it has generally been beheved that the killing occurred only because the winters were open. The results at Brook- ings after the winter of 1898-99 indicate that open winters preceded by o A brief but interesting discussion of the winterkilling of alfalfa under irrigation has been published recently by Prof. Samuel Fortier. See Farmers' BuUetin 373, U. S. Dept. of Agriculture, entitled "Irrigation of Alfalfa, '' 1909, pp. 41-42. 185 DIVEKSITY IN ALFALFA. 63 dry autumns, even when the temperatures are very severe, are not to be compared in injuriousness to winters preceded by moist autumns. In the Milk River Valley of Montana ordinary alfalfa maintains its stands from eight to fifteen years, despite the fact that temperatures considerably lower than — 40° F. occur rather frequently. Progressive thinning of the stand of course makes it necessary to reseed sooner or later. This is a region of scant autumn rainfall, which probably accounts for the success obtained in growing ordinary alfalfa. This same strain in Minnesota, North and South Dakota, Wisconsin, and many other Northern States succumbs in from two to five years, some- times even sooner, though the temperatures generally are not nearly so low as those of northern Montana. All of these facts seem to indicate that moisture in the soil is one of the most important factors in deter- mining whether or not alfalfa will winter over successfully. DIVERSITY IN ALFALFA. Alfalfa appears to be quite as diverse physiologically as it is botan- ically, and in the majority of cases botanical or morphological lines of descent seem to have no correlation with the physiological. That this should be the case is not surprising in a crop where the greatest freedom of interbreeding promoted by special adaptations for making insect visits effective is the rule. As a result of this condition strains that have by natural selection been rendered uniform with reference to physiological characters, such as cold resistance or drought resist- ance, may contain still the whole range of botanical diversity, so far as this is expressed in the external appearance of the plants. Strains of Grimm alfalfa, for instance, contain lines of great mor phological diversity. Some are prostrate, some upright; some leaf profusely, others very sparingly; some incline to coarse woody stems, others to fine tender ones; some have small leaves, others large; some show only one of a number of flower colors, others are variegated, and so on through the whole range of possible differences. In the experi- ments thus far conducted none of these diverse elements have shown any observable differences in cold resistance. This suggests the pos- sibility that the hardiness may be due to some liistological character common to aU the lines of descent. Investigations of the physiology and histology of cold resistance have not yet produced much definite information, but it does not appear necessary to assume any one character as the basis of hardi- ness. Some lines may live through severe test conditions from one cause and some from another; in some it may be due to the faculty of becoming dormant at the proper time, in others to the habit of growing from deep-seated crowns, in still others to the power of 185 64 COLD KESISTANCE OF ALFALFA. reestablishing a root system quickly when heaving or freezing has broken off the root. As our knowledge of individual alfalfa plants becomes more inti- mate, some true correlations between the morphological and physio- logical characters may be found. At present we have few data along this line. The belief has grown up recently that diversity or variegation in flower color is correlated with hardiness, and that drought resistance and cold resistance are also correlated. The evidence of the present investigations has been negative on both of these points. On the field plats the Kansas strain No. 19508, which undoubtedly came from Europe originally, has practically the same range of flower color as the Grimm strains. While one of the latter suffered no winterkilling and the other but little, the Kansas strain killed so badly that only about 20 per cent of a stand remained. The Grimm strains, as we have already seen, differed from one another in hardiness, though not in diversity of flower color. The two strains of commercial sand lucern S. P. I. Nos. 21217 and 21269 are far more variegated in color than any Grimm plants, still they proved less hardy on both the field plats and in the rows and also in experiments elsewhere in the Northwest. They themselves showed considerable difference in hardiness. On the other hand, strains like Wheeler's acclimatized Turkestan (S. Dak. No. 240), which has no variegation, proved hardier than any variety except the Grimm. This indicates that in the strains mentioned, at least, the factors producing cold endurance are independent of variegation or the factors producing it. Similarly, it has been believed that drought resistance is correlated with hardiness. Some of the best drought-resistant varieties, however, proved in the row experiments to be the very weakest strains from a cold-resistance standpoint. The only dry-land form that showed any cold resistance worthy of comment came from Sextorp, Nebr., where in addition to a lack of sufficient rainfall low temperatures are often experienced. Ordinary American alfalfa, which probably traces its history back to a single lot of seed taken from Spain to South America several cen- turies ago, also showed a range of more than 50 per cent in hardiness. METHODS BY WHICH ALFALFAS BECOME HARDY. A study of the whole group of alfalfas reveals the fact that the total range in variation in hardiness is much greater than in our com- mon American strains. The best strain in the Dickinson experi- ment, a Grimm, lost less than 3 per cent, while twelve kinds killed out 100 per cent. Among the latter were four Arabian stocks that winter- killed completely in row and hill tests as far south as Stockton, Kans., 185 METHODS BY T^'HICH ALFALFAS BECOME HAEDY. 65 where the winter of 1908-9 was milder than normal. This variation in hardiness of the different stocks is correlated positively with an in- definite minimum temperature or complex of climatic conditions of the country in which they have long been gro^m. In other words, the tenderest of the alfalfas come from the warmest regions and the hardier ones come from the colder regions. The first method of securing a hardy alfalfa is illustrated in the case of Grimm alfalfa, South Dakota Nos. 240 (originally from Turkestan) and 167, and by those strains of ordinary American alfalfa that show more than average hardiness. The evidence is good and fairly conclusive that the alfalfa winch Wendehn Grimm brought over from Germany in 1857 was neither better nor worse than the average central European alfalfa. It also contained many lines, some of which were hardier than others. ^ ^Ir. Grimm did not start a nursery and propagate one plant that showed itseh superior to the others. He sowed the whole lot of seed and harvested what he could. This was clone repeatedly and as the years passed by nature weeded out the tender lines, while the hardier ones persisted. This process no doubt gradually brought about a profoimd change in the stock as a whole. It is, however, not necessary to suppose that the lines which have persisted in the Grimm alfalfa are any dif- ferent than they were in the origmal seed brought over fift}' years ago. They are simply more abundantly represented in the popula- tion making up the strain. On the other hand, it may be that the Grimm stocks used hi our experiments do possess lines hardier than any contained in the foundation stock. These could have arisen readily by mutation, a recombination of characters,^ or otherwise smce the date of the introduction of the strain. This assumption is not necessary, as all the central European alfalfas that underwent the severest test of the row and hill experiment possessed some hardy lines which could be made the basis of further propagation. Another method may be used, but the ultimate principles mvolved are the same. As already pointed out, any stock of alfalfa seed, even when grown in a restricted area for a long time, is made up of a num- ber of lines or families which vary both morphologically and physio- logically and which breed more or less true. In our study of hardiness the majority of strains or regional varieties possessed some plants which lived through the northwestern winter. The plants that lived through were representatives of the hardier lines. Xow evidently a « Methods by which, recombinations produce a change in cross-fertilized plants are discussed by Prof. W. J. Spillman. See ''Application of Some of the Principles of Heredity to Plant Breeding," Bulletin 165, Bm-eau of Plant Industry, U. S. Dept. of Agriculture, 1909, pp. 36^3. 44281°— No. 185—10 5 66 COLD EESISTAXCE OF ALFALFA. simple method of seeming cold endurance is to select any representa- tive hard J plant and increase it. By this method of breeding we are led to the establishment of a pure or nearly pure strain. The Minnesota Agricultural Experiment Station and Prof. W. A. Wheeler, of South Dakota, have followed this method, basing their pure strains largely on selected Grimm plants. During the past season Professor Wheeler produced as much as 200 pounds of seed that traces its ancestry to a single plant. This method is also being followed by Mr. A. C.^ Dillman, of the Office of Alkah and Drought Resistant Plant Breeding Investigations, in his work at Belle- fourche, S. Dak. In this method the singling out of the representative plant is the important thing. It naturally presupposes some antecedent method of getting rid of the weaklings and hence applies not so much to the establishment of a hardy race as to the isolation of a pure or uniform race from a mixture of predetermined hardiness. The advantages and disadvantages of this method are little known. Observations on other crops indicate that there is considerable danger in a too narrow restriction of lines of descent .° The length of time required to produce from a mixed stock by un- conscious mass selection ^ an approximately pure race, so far as cold resistance is concerned, is uncertain. It will evidently depend ahnost wholly on two factors, namely, the abundance of hardy in- dividuals in the original stock and the frequency with which test winters occur. South Dakota Xo. 240 (see pp. 51-55) is an example of the estab- lishment of a very hardy and nearly pure strain from a mixed strain having a lesser degree of hardiness. The new hardy strain was devel- oped immediately, without the necessity of intercrossing or long- continued selection. The experience with this strain and observations on others indicate that the time required need not be long. The original seeding in the case of Xo. 240 was made in 1S99; the seed used in the experiment was harvested in 1906. A study of the records for Highmore, S. Dak., and neighboring points shows that test weather conditions prevailed several times during the interim. "^ a Our knowledge of the amount of cross-pollination that takes place when alfalfa plants are left e .posed is very deficient. If we wish to select a line and propagate it pure, the prevention of cross-pollination is necessary. If the plant selected is an F^ cross with regard to any character, then it is obvious that the selection and self-pol- lination of several individual plants may be necessary in the next generation before a really pure strain is obtained. b Cook, O. F. The Superiority of Line Breeding over Narrow Breeding. Bulletin 146, Bureau of Plant Industry, U. S. Dept. of Agricultm-e, 1909, pp. 28-29. cThe following minimum temperatures at Highmore, S. Dak., indicate the nature of the test that alfalfa endures in this locality: 1904, -27° F.; 1905, -36° F.; 1906, -31° F.; 1907, -27° F. 185 COXCLUSIOX. 67 The seed production of Xo. 240 has been uniformly poor. Most stocks of Turkestan alfalfa share in this weakness. In some years the plats appear not to have been cut at all; in 1904 only the first crop was removed. In consequence of this, the surviving plants undoubtedly produced seed which was not harvested, but fell on the ground, thus bringing about natural reseeding. In this case seven years of unconscious mass selection toward cold resistance resulted in an unusually hardy strain. The Grimm variety is the product of fifty years of similar selection in the Northwest. The term "selective acclimatization" maybe applied to this process of change whereby hardy lines or races in a crop plant are accen- tuated in later generations by the elimination from year to year of unfit lines. The versatility of a race produced in this way in meet- ing new conditions in other areas than the one m which it was devel- oped will probably hinge largely on the range of selective factors to which it has been subjected in its previous history. Hybridization between alfalfa and its wild relatives has been sug- gested as an immediate method of securing hardiness. ° Its use, however, would not dispose of the necessity of selection both before and after the making of the crosses. Antecedent selection will be required for determining the hardiness of the chosen parents of both the wild and cultivated stocks, wMe the subsequent selection will be essential for eliminating the nonliardy by-products of crossing and isolating the lines of high forage and seed value. CONCLUSION. The investigations that form the basis of the present paper have been under way at Dickinson, X. Dak., for nearly four years. At other places in the X^orthwest they have continued a year longer. The observations recorded show that the employment of good tillage, the use of a suitable strain of seed, and the character of the winter weather are the controlling factors in successful alfalfa production in the Xorthwest when soils of average fertility are used. With reference to tillage it may be said that the previous cropping and treatment of the soil and its thorough preparation are of pri- mary importance. Inoculation, irrespective of the manner in which it is secured, determines success or failure, even though aU other conditions are satisfied. It was found that bv varvins; the time of cuttino- in order to leave considerable growth on the field in the autumn, even the strains of inferior hardiness endured, though the conditions were severe. This result appeared to be due to the exhaustion of moisture from the o For an extended discussion of this point of view, see Westgate, J. M., •'Variegated Alfalfa," BulJetin 169, Bureau of Plant Industrj^ U. S. Dept. of Agriculture, 1910. 185 68 COLD EESISTAXCE OF ALFALFA. soil in the autumn and to a plant covering which held a protective blanket of snow over the alfalfa cro^\Tis during severe weather. Thinness of stand resulted in a high percentage of winterkilling in all except the hardiest varieties. Farmers who are unable to secure sufficient seed of hardv strains for acreage plantings may find it profitable to use less hardy strains, planning to leave the late sum- mer and early fall gro^^Hh standing to hold the snow. Thoroughness in preparing the soil promotes inoculation, which is essential to satisfactory yields. Lack of attention to this point has led in thousands of cases to the plowmg up of alfalfa stands in the Northwest that would otherwise have been successful. Inoculation is just as necessary for hard}' as for nonhardy strains of seed. Present experiments show that undoubtedh' the most important cause of failure in actual practice has been the use of seed from regions that m the nature of things can not produce adapted seed. In practically all of the present alfalfa-growing areas of the L^nited States, ordinary American alfalfa and commercial importations of seed from abroad have proved to be entirely satisfactory or else so easily adaptable to conditions as to become satisfactory after one or two generations of seed have been produced. This is particularly true where alfalfa is grown under irrigation. Even as far north as the Yellowstone and Milk River valleys of Montana ordinary Utah seed has given satisfactory results, except in unusual years. In the more humid parts of the cold Xorth and Northwest this has not proved the case. After a severe winter in these sections fields of nonhardy forms of alfalfa show only the dry crowns of dead plants, with here and there a growing plant. The profusion of the latter depends wholly upon the degree of winter resistance of the particular strain used, that is. upon the percentage of cold-resistant individuals included in its population. Our knowledge is still meager as to what quality or qualities pro- duce hardiness in different races of alfalfa. The importance of ability to become dormant in spite of weather calculated to force continued growth has been pointed out. At least two other characteristics of the plants themselves have been observed that are believed to be of importance. The Mongolian alfalfa which proved to be the hardiest under the severe test at Dickinson of any of the newly imported strains was found both there and at Stockton, Kans., to have crowns more deeply set in the soil than other varieties. It seems likely that this adaptation, if it may be called an adaptation, is of importance in giving the tenderest part of the plant the needed protection. Another condition has been noted, especially in Grimm alfalfa. On taking up the crowns in old stands it has been observed that at some previous time the taproot has been completely broken off from 185 CONCLUSION. . 69 4 to 8 inches below the surface.^ Living plants in such cases have developed new roots laterally to take the place of the broken ones. This condition has suggested the possibility that the long endurance of stands of the Grimm strain may be due in some measure to a capacity for putting out new roots and reestablishing itself after the taproot has been broken. The data collected on this point do not yet justify anything like positive statements, but it is mentioned in this connection to stimulate growers and breeders to make note of similar conditions. With reference to weather conditions, which because of their impor- tance have been discussed rather fully in the present paper, the follow- ing appeared to be the most destructive to alfalfa stands under the conditions studied : (1) A lack of sufficient snowfall or the exposure of plants, due to drifting. The latter condition can often be remedied by making the last cutting of hay early enough so that the field will have 8 or 10 inches of plant growth to bind the snow. (2) Successive thawing and freezing, which may kill the plants outright by ruptur- ing their tissues or till them indirectly by heaving the soil, thus breaking the main roots and stripping off the lateral roots. The crown of a plant killed in this way can be drawn from the ground without effort, and in the spring it usually dries out before it can reestablish itself. (3) An excess of moisture in the autumn which by preventing the plants from hardening up their tissues results in their being in a more or less tender and succulent condition when winter weather sets in. Alternate thawing and freezing are especially deadly when a wet autumn or a late irrigation leaves a high percentage of moisture in the soil. Dry winters have been supposed to be highly injurious to alfalfa, but the writers' observations indicate that it is only the dry winters which are preceded by moist autumns that do the greatest injury. A dry fall furnishes more favorable conditions for wintering over by producing more perfect dormancy. In the well-recognized alfalfa- growing areas of the United States winterkilling has never occasioned excessive losses. Observations elsewhere than at Dickinson have shown at least two other important external conditions that conduce to winterkilling, namely, sleet or a coat of ice which is usually an accompaniment of alternate thawing and freezing, and severe pasturing, whether by cattle, sheep, or hogs, especially if it is carried late into the autumn. In the case of ice formation it seems probable that the plants are injured both by a rupture of the tissues and through being deprived of the necessary aeration. a Mr. T. H. Kearney calls attention to the fact that ordinary alfalfa in Arizona does this when the taproot has been destroyed by Ozonium. This disease has not been found in Carver County, Minnesota. 185 70 COLD EESISTAXCE OF ALFALFA. Xo doubt many other factors in addition to those discussed in this bulletin are concerned in the rapid depreciation of alfalfa stands in cold climates. It is hoped that the present investigations may con- tribute to a fuller knowledge of the subject and assist ui the estab- hshnient of an alfalfa industry ui the Northwest comparable to that which has grown up farther west and in the Southwest within the past fifty years. 155 PLATES. 185 71 DESCRIPTION OF PLATES. Plate I. {Frontispiece.) Fig. 1. — Experimental plats of alfalfa in hills and drill rows at Dickinson, N. Dak., on October 1, 1908. Each row was sown to a different strain in the spring of 1908, 68 kinds in all, and on the date the photograph used for this illustration was taken practically every one had a full stand. Fig. 2. — The same experimental plats at Dickinson, N. Dak., on June 10, 1909, showing the great diversity in the winter resistance of alfalfa from different parts of the world. The statistics of winterkilling for the section in the immediate foreground, rows 35 to 68, counting from right to left along the white stakes in the middle of the picture, are shown in the second part of Table XIII. Those for rows lto34, comprising the section between the first and second complete lines of stakes, are presented in the first half of Table XIII. Plate II. Fig 1. — Experimental plats showing the importance of a snow covering. The small patch of red clover in the foreground stands in the meager shelter of an old back furrow. This furrow, which is only slightly higher than the soil around it, held a few inches of drifting snow over the plants and preserved them from injury, while the remainder of the plat, which is only a few hundred feet from the alfalfa nursery, killed out completely. Fig. 2. — Experimental plats at Stockton, Kans., showing the winterkilling of North African and Arabian alfalfa in the mild winter of 1908-9. The middle row is Kebilli Oasis alfalfa, No. 12846, secured by Mr. T. H. Kearney from Tunis. At Dickinson, N. Dak., this strain killed out 100 per cent. (See row 20, Table XIII.) The row at the right of this, showing a few plants, is from Tripoli. Two rows to the right of the Tripolitan strain and at the left of the Tunisian are Arabian alfalfa, the tenderest variety known. Not a living plant at Stockton remained in any of the Arabian rows in the spring. These rows are identical with Nos. 6, 7, 8, and 9, Table XIII, Dickinson experiments. Plate III. Fig. 1. — Experimental plats showing seven rows of European alfalfa after the winter of 1908-9 at Dickinson, N. Dak. The four rows to the left of the arrow nearest the middle of the picture. Table XIII, rows 57 to 60, inclusive, are the so-called commercial or false sand lucern from France and Germany. Row 61 is Italian alfalfa and row 62 Provence alfalfa grown in Germany. Row 63, imme- diately to the right of the arrow nearest the margin, is another sample of com- mercial sand lucern. The two rows to the right of the middle arrow, rows 56 and 55, Table XIII, are Utah dry-land alfalfa. Next is South Dakota No. 167, row 54, the so-called Baltic alfalfa, while between this and the two rows of Grimm alfalfa, at the right, is a row of Brott's Nebraska dry-land alfalfa. Fig. 2.- — Experimental plats showing three rows of Mongolian, the hardiest newly imported alfalfa included in the experiment, Dickinson, N. Dak., August, 1909. (See rows 2, 3, and 4, Table XIII.) Plate IV. Fig. 1. — Experimental plats showing the winterkilling of North American alfalfa and some other strains at Dickinson, N. Dak. Photographed June 10, 1909. The row indicated by the arrow, row 47, Table XIII, is Colorado alfalfa, which was included in the field, plat, and row experiments. In order, the rows to the left of the arrow are as follows: No. 48, Kansas; 49, Chinook, Mont.; 50, Clear- water, Nebr.; 51, Minnesota Grimm; 52, North Dakota Grimm. To the left of the arrow are the following: No. 46, first quality commercial, of European origin, with variegated flower color; 45 and 44, Mexico; 43 and 42, Canada; 41, Billings, Mont.; 40, Wheeler's acclimatized Turkestan. Fig. 2. — A field of Grimm alfalfa in Carver County, Minnesota, showing the nature of the severe winterkilling of 1905-6. Although this field is located on a sidehill, so that no water could remain standing on it, considerable winterkilling took place in the "draws" or natural drainage courses of the field where the soil had become oversaturated by autumn moisture. Photographed in September, 1909. 185 72 Bui. 1 85, Bureau of Plant Industry, U. S. Dept. of Agriculture. Plate II. ^'^m^aiii^ Fig. 1.— Experimental Plats at Dickinson, N. Dak., June 12, 1909, Showing the Importance of a Snow Covering. The patch of red clover in the foreground was preserved by a light covering; the remainder of the field was killed out. Fig. 2.— Experimental Plats at Stockton, Kans., Showing the Winterkilling of North African and Arabian Alfalfas During the Mild Winter of 1908-9. Bui. 185, Bureau of Plant Industry, U. S. Dept. of Agriculture. Plate III. Fig. 1.— Experimental Plats of European Alfalfas after the Winter of 1908-9 AT Dickinson, N. Dak. The seven rows between the arrows (Nos. 57 to 63, counting from right to left) are from Europe. Fig. 2.— Experimental Plats at Dickinson, N. Dak., August, 1909, Showing Three Rows OF Mongolian Alfalfa, the Hardiest Foreign-grown Strain. Bui. 1 85, Bureau of Plant Industry, U. S. Dept. of Agriculture. Plate IV. Fig. 1 .—Experimental Plats at Dickinson, N. Dak., Showing the Winterkilling OF North American Strains of Alfalfa. The rows are numbered from right to left, the arrow indicating No. 47. Fig. 2.— a Field of Grimm Alfalfa in Carver County, Minnesota, Showing the Winterkilling in 1905-6 in the "Draws" or Natural Drainage Courses of the Field. INDEX. Page. Acclimatization, selection, definition 67 African alfalfa. See Alfalfa, African. Alfalfa, acclimatization of variety in Minnesota 50 African, winter resistance, 1908-9 43-44 Algerian strains, introduction into United States 9-10 winter resistance, 1908-9 44 American, hardiness, diversity under uniform conditions 49 origin, winter resistance, 1908-9, etc 48 Arabian, introduction into United States 10 winter resistance, 1908-9 43, 45 Argentina, introduction 10 winter resistance, affinities, etc 43 Canadian, origin, winter resistance, 1908-9, etc 47 Chilean, introduction 11 parent strain hardy in certain areas 11 cold resistance, experiments at Dickinson, X. Dak 11-60 purpose of investigations 9-11 Colorado, hardiness at Dickinson, N. Dak., comparison with other varieties 20, 22 common, lack of hardiness in coldest parts of Great Plains 21 or Chilean, adaptability to new conditions 11 variation in hardiness 49-50 winterkilling by States 50 differences in color, vigor, and growth of plants 23, 27 yield, causes, etc 27-29 diversity in strains 49, 63-64 morphological and physiological, not correlated 63 dormancy, cold a factor in production 16 drought a factor in production 16 effect of autumn moisture 33-34 importance in determining hardiness 16 production, importance of drj^ weather 22 drought resistance, not correlated with hardiness 64 dry-land, Brett's Nebraska, winterkilling 48 Utah, winterkilling 48 Ecuador, winter resistance, 1908-9 43 effect of autumn moisture, comparison of plats 60-63 weather conditions, winter of 1905-6 62 1908-9 on 68 varieties 39-42 Egyptian, introduction into United States 10 European, winter resistance, 1908-9 44 field plats, experiments, 1907-8, record of previous crops and treat- ment 20 flowers, color, diversity, comparison of strains 55 relation to hardiness 64 185 , 73 74 COLD EESISTANCE OF ALFALFA. Page. Alfalfa, French, winter resistance, 1908-9 44 See also Lucern, sand. German, winter resistance, 1908-9 44 Grimm, cold resistance, difference between two samples, causes, etc.. 25-27 diversity of strain 63 experiments at Chinook, Mont., and at Alma, Nebr., difference in results 25-27 hardiness, method of producing 67 origin, etc 50-51 growing, controlling factors 9 drill-row and hill experiments at Dickinson, N. Dak., begun in 1908 30-56 experiments, Dickinson, N. Dak., begun in 1906 11-18 1907 18-22 winterkilling and yields.. 22 1908 23-30 comparison of field plats and rows and hills 56-60 diversities, etc 23-30 field plats sown in 1907, management 56-57 field plats sown in 1908, management 57-59 pre\'ious treatment of land . . 18-20 soil preparation, weather conditions, etc., 1907 and 1908 20-22 with 22 varieties, table of results, etc 11-18 Kansas, without irrigation 10-11 Nebraska, without irrigation : 10-11 Guaranda, winter resistance, 1908-9 4,3 hardiness, comparison with red clover 21 dormancy as factor 15 effect of autumn weather conditions 15 hybridization 25 not correlated with diversity in flower color 64 range, etc 10-11 size of seed, apparent correlation , 47 variations, causes 65 hardy, methods of production 64-67 hay, yield of different plats, effect of inoculation 29-30 hybridization, means of securing hardy strain 67 inoculation and thickness of stand, relation to cold resistance 25, 59 of plants, importance 18, 25, 29-30, 67, 68 introduction into United States of Algerian strains 9-10 irrigation, autumn, effect 62 Italian, winter resistance, 1908-9 44 Kansas, growing without irrigation 10-11 winter resistance, 1908-9 • 50 mass selection, production of pure race 66 185 INDEX. 75 Page. Alfalfa, Mexican, introduction into United States 10 winter resistance, 1908-9 48 Mongolian, hardiest foreign-grown strain 47 hardiness, variation with size of seed 46-47 introduction 47 winter resistance, 1908-9 46-47 Montana, winter resistance, 1908-9 48, 50 mowing close, effect on winter resistance 61 method, effect on winter resistance 67-68 Nebraska, growing without irrigation 10-11 winter resistance, 1908-9 48. 50 need of suitable strain for humid Eastern States 11 Southern States 11 Northern States 11 semiarid regions 11 North American, winter resistance, 1908-9 47-56 Oasis, introduction into United States 9, 10 value in Southwest for hay production 43 winter resistance, 1908-9 43 old German Franconian, winter loss 51 Peruvian, introduction into United States 10 winter resistance, 1908-9 43 plants, protection by snow 37 Poitou, introduction into United States 9 production, Northwest, important factors 67 of a pure strain 66 Provence, introduction into United States 9 range, hardiness, etc 10-11 Russian, winter resistance, 1908-9 44 Sahara, winter resistance, 1908-9 43 seed, selection, importance of care 68 size, relation to tenderness of plants 46-47 seeding, experimental plats, Dickinson N. Dak., 1906 and 1907 12, 20 selective climates 49 Setif , winter resistance, 1908-9 • 44 Simbirsk, hardiness at Dickinson, N. Dak., comparison with other varieties 20, 21, 22 South American, winter resistance, 1908-9 43 Dakota, origin, winter resistance, etc 55-56 Spanish, winter resistance, 1908-9 44 stand, effect of different soil treatments 24 original, cause of differences 24 stands of varying thickness, amount of soil moisture present 34-37 strains used in drill-row, and hill experiments, origin 31-32 survival of winter of 1906-7, causes 17 taproot, breaking, results 68-69 tillage, important considerations 68 Turkestan, acclimatized, history, experiments, etc 51-56 seeding capacity 54 winter resistance, comparison with other strains 52-55 hardiness at Dickinson, N. Dak., comparison with other varieties 20, 21, 22 185 76 COLD EESISTANCE OF ALFALFA. Page, Alfalfa, Turkestan, introduction into United States 9-10 seed, production 67 purchase by northern growers, caution 45-46 winter resistance, 1908-9 45-46 use of name 10 Utah, hardiness at Dickinson, N. Dak., comparison with other varieties. 20, 22 winter resistance, 1908-9 48, 50 winter resistance, importance of snowfall 69 winterkilling, cause of differences in amount 24-25 caused by coat of ice 69 excess moisture 69 sleet , 69 successive freezing and thawing 69 prevention, snow a factor 15 results of experiments at Dickinson, N. Dak. 23 yields, results of experiments at Dickinson, N. Dak 23 See also Lucern and Medicago. Alfalfas, local adjustment 26 progeny of Old World introductions, hardiness, results of experiments . 50 winter resistance in drill rows and hills in 1908-9 experiments 43-56 Algerian alfalfa. See Alfalfa, Algerian. Alma, Nebr., temperature, minimum monthly, for cold months 25 "Alt-deutsche frankische Luzerne," origin of name, relation to Grimm alfalfa, etc 44 winter resistance, 1908-9 44 American alfalfa. See Alfalfa, American. Arabian alfalfa. See Alfalfa, Arabian. Argentina alfalfa. See Alfalfa, Argentina. Bessey, E. A., introduction of Turkestan alfalfa into United States 10 Blizzard, November 30, 1908, Dickinson, N. Dak 37 Canada, need of suitable strain of alfalfa 11 Canadian alfalfa. See Alfalfa, Canadian. Chilean alfalfa. See Alfalfa, Chilean. Chinook, Dickinson, N . Dak., January 16, 1909 39 Mont., temperature, minimum monthly, for cold months 25 Climates, different, selective values 49 Clover, hardiness, dormancy as factor 15 red, hardiness 37 comparison with alfalfa 21 Cold, factor in producing dormancy of alfalfa 16 resistance of alfalfa, Grimm, difference between two samples 25-27 purpose of investigations 9-11 Colorado alfalfa. See Alfalfa, Colorado. Conclusion of bulletin 67-70 Dickinson, N. Dak., blizzard, November 30, 1908 37 Chinook wind, January 16, 1909 39 frosts, killing, first, fall of 1906 15 occurrence, fall of 1908 33 location, elevation, etc 11 precipitation, January, 1909, daily record 38 rainfall, 1906-8 13, 14 185 ■ INDEX. 77 Page. Dickinson, N. Dak., rainfall, September and October, 1908 59 snowfall, January, 1907 14, 16 1909 39 November 25-26, 1908 37 temperature, January, 1909, daily record 38-39 winter 1906-7 13-14 temperatures, mean, 1906-1909 17 weather conditions and other observations, 1907 and 1908. 20, 22 January, 1909 38-39 summer of 1908 32-33 winter of 1908-9 37-39 daily record, January, 1907 14 Drill-rows, experiments in alfalfa growing, at Dickinson, N. Dak., begun in 1908 30-56 Drought, factor in producing dormancy of alfalfa 16 resistance of alfalfa, relation to hardiness ". 64 Ecuador alfalfa. See Alfalfa, Ecuador. Egyptian alfalfa. See Alfalfa, Egyptian. European alfalfa. See Alfalfa, European. Experiments, cold resistance of alfalfa 11-63 Fairchild, D. G., and Scofield, C. S., introduction of Algerian alfalfa into United States 9-10 Flowers, alfalfa, color, diversity, comparison of strains 55 relation to hardiness 64 Freezing and thawing, successive, cause of winterkilling of alfalfa 69 French alfalfa. See Alfalfa, French, and Lucern, sand. Frost, occurrence August 22, 1908, Dickinson, N. Dak 33 Frosts, killing, first, Dickinson, N. Dak., fall of 1906 15 1907 20 late, Dickinson, N. Dak., spring of 1907 17 occurrence, Dickinson, N. Dak., fall of 1908 33 German alfalfa. See Alfalfa, German. Germany, Tauber River valley, production of Franconian alfalfa 51 Great Plains, coldest parts, hardiness of common alfalfa 21 Grimm alfalfa. See Alfalfa, Grimm. Wendelin, introduction of Grimm alfalfa 50 Guaranda alfalfa. See Alfalfa, Guaranda. Hansen, N. A., introduction of alfalfa into North America 9 Hardiness, alfalfa, comparison with red clover 21 determining, importance of dormancy 16 dormancy as factor 15 effect of autumn weather conditions 15 hybridization 25 Grimm, method of producing 67 variations, causes 65 clover, dormancy as factor 15 red 37 comparison with alfalfa 21 See aZso Winterkilling. Hay, alfalfa, value of Oasis variety in Southwest 43 yield of different plats, effect of inoculation 29-30 Herba medica, Italian name for alfalfa 10 185 78 COLD EESISTAXCE OF ALFALFA. Page. Hills, experiments in alfalfa growing at Dickinson, X. Dak., begun in 1908 30-56 Hybridization, effect on hardiness of alfalfa 25 method of producing hardy strain of alfalfa 67 Ice, coat, cause of \dnterkilling of alfalfa 69 Inoculation, alfalfa, effect on cold and frost resistance 30 importance 18, 25. 29-30, 67, 68 Introduction to bulletin 9 Irrigation, autumn, effect on alfalfa 62 Italian alfalfa. See Alfalfa, Italian. Kansas alfalfa. See Alfalfa, Kansas. Kearney, T. H., introduction of Oasis alfalfa into United States 10 Kelly, W. F.. production of South Dakota alfalfa 56 Lathrop, Barbour, and Fairchild, D. G.. introduction of certain strains of alfalfa. 10 Loam. See Soil, loam. Lucern, sand, introduction 9 "winterkilling 64 use of name 10 See also Alfalfa, "Alt-deutsche frankische Luzerne, "" and Medicago. Medica sativa, common names 10 Medicago sativa gaetula, introduction 10 turkestanica, introduction, nature, etc 54-55 See also Alfalfa and Lucern. Mexican alfalfa. See Alfalfa. Mexican. Moisture, autumn, effect on alfalfa 60-63 dormancy of alfalfa plants 33-34 excess, cause of winterkilling of alfalfa 69 soil, amount in alfalfa stands of varying thickness 34-37 Mongolian alfalfa. See Alfalfa, Mongolian. Montana alfalfa. See Alfalfa, Montana. Chinook, experiments with Grimm alfalfa 25-27 Milk River Valley, winter resistance of alfalfa, causes 63 Mowing, alfalfa, close, effect on winter resistance '61 method, effect on winter resistance 67-68 Nebraska alfalfa. See Alfalfa, Nebraska. Alma, experiments with Grimm alfalfa 25-27 North American alfalfa. See Alfalfa, North American. Dakota, Dickinson, alfalfa growing, experiments 11-56 Oasis alfalfa. See Alfalfa, Oasis. Peruvian alfalfa. See Alfalfa, Peruvian. Plates, description 72 Poitou alfalfa. See Alfalfa, Poitou. Precipitation, Dickinson, N . Dak., daily record, January, 1909 38 See also Rainfall. Provence alfalfa. See Alfalfa, Provence. Rainfall, Dickinson, N. Dak., 1906-1908 13,14 September and October, 1908 59 See also Precipitation. Red clover. See Clover, red. Rockhill, ^Y. ^Y., introduction of Mongolian alfalfa 47 Rows, tests of alfalfa growing, compared with other tests 59 Russian alfalfa. See Alfalfa, Russian, thistle. See Thistle, Russian, 185 INDEX. 79 Page. Ryf, G., reference 10 Sahara alfalfa. See Alfalfa, Sahara. Sandy soil. See Soil, sandy. Scofield, C.S., and Fairchild, D.G., introduction of Algerian alfalfa into United States 9-10 Seed, alfalfa, selection, importance of care 68 size, relation to tenderness of plants 46-47 Turkestan, production 67 purchase by northern groM^ers, caution 45 Seeding, alfalfa, experimental plats, Dickinson, N. Dak., 1906 12 1907 20 Setif alfalfa. See Alfalfa, Setif . Simbirsk alfalfa. See Alfalfa, Simbirsk. Sleet, cause of winterkilling of alfalfa 69 Snow, factor in preventing winterkilling of alfalfa 15 protection to alfalfa plants 37 Snowfall, Dickinson, N. Dak., January, 1907 14, 16 1909 ■. 39 November 25-26, 1908 37 importance in winter resistance of alfalfa. 69 Soil, alfalfa-growing experiments, Dickinson, N. Dak., preparation, 1907 20 previous treatment. . . . 18-20 effect of different treatments on stand of alfalfa 24 loam, experiments with 22 varieties of alfalfa, table of results 12 moisture. See Moisture, soil. nature and treatment in drill-row and hill experiments with alfalfa at Dickinson, N. Dak 31 sandy, experiments with 22 varieties of alfalfa, table of results 12 Soils, alfalfa growing at Dickinson, N. Dak., experiments 12-13 South American alfalfa. See Alfalfa, South American. Dakota alfalfa. See Alfalfa, South Dakota. Spanish alfalfa. See Alfalfa, Spanish. States, Eastern, humid, need of ^suitable strain of alfalfa. 11 Northern, need of suitable strain of alfalfa 11 Southern, humid, need of suitable strain of alfalfa. 11 Swingle, W. T., alfalfa, Mongolian, introduction ' 47 Taproot, alfalfa, breaking, results ." 68-69 Tauber River valley, Germany, production of Franconian alfalfa 51 Temperature, Chinook, Mont., and Alma, Nebr., minimum monthly for cold months 25 Dickinson, N. Dak., daily record, January, 1909 38-39 winter of 1906-7 13-14 winters of 1906, 1907, 1908, 1909 21 Temperatures, Dickinson, N. Dak., mean, 1906-1909 17 Thistle, Russian, use as plant cover for holding snow 20 Tillage, alfalfa, important considerations 67 Turkestan alfalfa. See Alfalfa, Turkestan. Utah alfalfa. See Alfalfa, Utah. Weather conditions, Dickinson, N. Dak., autumn, effect on hardiness of alfalfa 15 January, 1907, daily record 14 1909 38-39 summer of 1908 32-33 185 80 COLD KESISTANCE OF ALFALFA. Page. Weather conditions, Dickinson, N. Dak., winter of 1908-9 37-39 winter of 1905-6, effect on alfalfa 62 Weeds, occurrence in alfalfa plats 28 effect on yields 28 Westgate, J. M. , observations on South Dakota alfalfa 55 Winterkilling, alfalfa, cause of differences in amount 24-25 caused by coat of ice 69 excess moisture 69 sleet 69 successive freezing and thawing 69 experiments at Dickinson, N. Dak 22 prevention, snow a factor 15 results of experiments at Dickinson, N. Dak 23 See also Hardiness. Yield, alfalfa, differences, causes, etc 27-29 alfalfa, experiments at Dickinson, N. Dak 22, 23 185 O