SB 
 
 UC-NRLF 
 
 
 THE LARGE SEED FACTOR 
 
 I H 
 
 CROP PRODUCTION 
 
 A THESIS 
 
 BY 
 
 Lowrie Baird Nevin 
 
 Presented in partial fulfillment 
 of the requirements for the degree 
 
 of 
 Bachelor of Soienoe 
 
 University of California 
 College of Agriculture 
 Division of Agronomy 
 May 1916 
 
 UMVEKSHY CFCA) :: 
 
 LIBRARY 
 
 COLLEGE OF AGRICULTURE 
 DAVIS 
 
THE LARGE SEED FACTOR 
 
 I B 
 CROP PRODUCTION 
 
 A THESIS 
 
 BY 
 
 Lowrie Baird Nevin 
 
 Presented in partial fulfillment 
 of the requirements for the degree 
 
 of 
 Bachelor of Scienoe 
 
 University of California 
 College of Agriculture 
 Division of Agronomy 
 May 1916 
 
 UNIVERSITY OF TA;,;! 
 
 LIBRARY 
 
 COLLEGE OF AGRICULTURE 
 DAVIS 
 
CONTESTS 
 
 Introduction 2 
 
 
 
 The Influence of Large Seed upon the Yield 3 
 
 Effect of Large Seed upon Germination 14 
 
 The Effect of the Specific Gravity upon the Yield ... 15 
 
 Relation of Size and weight of Kernel to Composition. . 18 
 
 Mechanical Aids to the Selection of Seed 22 
 
 Fanning Mills 25 
 
 PART II. 
 
 Experimental 28 
 
 Conclusions 33 
 
 References 35 
 
 Figures 37 
 
 -1- 
 

THE LARGE SEED FACTOR IB CHOP PRODUCTION. 
 Introduction 
 
 The question of larger yields is a very vital one with the 
 farmer of to-day. In many oases more than is realized "by the 
 farmers themselves do the profits depend upon the last refine- 
 ments in efficient production. Therefore any factor that gives 
 larger returns for practically the same effort merits attention. 
 Such a factor is the influence of large heavy seed in crop pro- 
 duction. 
 
 In opening his discussion of experiments with wheat in which 
 some of the advantages of large plump seed were demonstrated, Dr. 
 1. A. Cobb (1) remarks that any effort to prove the lower value 
 of small seed will be regarded by many as an effort to kill a 
 dead horse; but as long as there exist among us advocates of 
 small seed, and the bulk of wheat growers use poor seed, it will 
 be necessary to harp on the superior value of large seed. 
 
 It is the intent of this paper to deal principally with 
 the large seed factor as applying to cereals, especially wheat; 
 but from time to time reference is made to experiments with 
 other plants to show that the principle applies not alone to 
 the cereals. 
 
 There has been no attempt to make an exhaustive survey 
 of the field, either as regards the literature or as a dis- 
 cussion of the subject. 
 
 The one definite objective of the experimental part of 
 the task has been to secure some experimental ground as a 
 basis for the assumption that the extra development of the 
 plants from large seed is due to the additional supply of 
 
 endosperm available to the germinating plantlet. 
 
 -2- 
 
; 
 
 ' 
 
 36160 
 rf , 
 
 
There is no lack of literature showing that this is the 
 explanation favored by many tut it has not been the good 
 fortune of the writer to find any experiments in support 
 of the theory, logical as it certainly seems. 
 
 The subject has been taken up from the standpoint of 
 the influence of large seed upon the yield, of large seed 
 upon germination, of the specific gravity or weight upon 
 the yield, and the relation of the size and weight of the 
 kernel to composition. There has been included also a 
 short discussion of some of the available mechanical aids 
 to the selection of large, heavy seed. 
 
 The Influence of Large Seed upon the Yield. 
 
 Sanborn(2), at the Utah Experiment Station, reports ex- 
 periments to ascertain the effect of using seed separated 
 for size. He used five grades: large, medium, small, or- 
 dinary, and shrivelled. His results are given as follows :- 
 
 : pounds per acre yields: Average 
 Kind of seed. : : 4 years 
 
 :1890 : 1891 : 1892 : 1893 : bu.per A. 
 
 Large 88.5 72.5 111.0 >63.0 18.72 
 
 Mediua 70.0 87.0 67.0 16.60 
 
 Small 94.0 105.0 64.0 74.0 18.72 
 
 Ordinary 84.0 95.0 87.0 29.5 16.42 
 
 Shrivelled 42.0 78.0 31.0 11.25 
 
 His conclusion that very little if any advantage is to 
 be gained by separating seed wheat and planting the large 
 kernels is of questionable value on account of the irregu- 
 larity of the results. 
 
 Latta, at the Indiana Experiment Station, (3) conducted 
 experiments in which wheat was cleaned of chaffy seed and 
 impurities and separated into light and heavy kernels by 
 
means of a fanning mill. The experiments were continued 
 for three years, each year taking seed not grown from seed 
 so selected. The results showed an average gain for large 
 seed of two and one half bushels per acre. 
 
 Georgeson, at the Kansas Experiment Station, has done con- 
 siderable work on the subject and we find the results of his 
 work in several of the publications of that station. In 
 Bulletin No. 13 the following results are presented: 
 
 Grade of seed Yield of grain per A. 
 
 Light 21.6 bu. 
 
 Common 24.0 " 
 
 He avy SO . " 
 
 These tests, he says, prove "only what is already 
 well known". 
 
 He later seeded plats with seed of different weights 
 per bushel and one of seed from selected heads. The grades 
 were as follows: 
 
 Light seed 56 Ibs. per bu. 
 Common seed 62.5 62.5 " 
 Heavy seed 63 " " n 
 The seed from selected heads. 
 
 Each year seed was selected from wheat not grown from selec- 
 ted seed. The average results from three years trials were 
 as follows: 
 
 Grade of seed Bu. of grain per acre. 
 
 Light 25.19 
 
 Common 26.57 
 
 Heavy 27.07 
 
 Select (Average for two years) 25.82 
 
 In Bulletin So. 54 of the same station the average for 
 six years shows as follows: 
 
 Light 8.48 
 
 Medium 29.85 
 
 Heavy 30.76 
 
The results of eight years' experiments recorded in Bulletin 
 No. 54 of the same station are similar. 
 
 Of oats he says in Bulletin 63, n light inferior seed 
 is certain to produce less than seed of a fair quality, "but 
 "between a fair quality of seed oats and heavy sifted seed 
 
 there is not much difference." n the average for seven 
 
 years is, however, in favor of heavy seed." 
 
 In 1892-3 Eesprez at the Experiment Station of Capello 
 (4) experimented with seed of different sizes and from dif- 
 ferent locations in the spike, with results as follow: 
 
 Kilograms 
 1892 From From Dif. 
 
 large small favor Ige . 
 Var. Ho. 1. 5,726 4,799 927 
 
 Early ears, middle 
 
 Late 6,172 4,235 1,937 
 
 Yar. So . 2 . 
 
 Early ears, middle 5,231 3,123 2,108 
 
 Late 4,680 2,456 2,224 
 
 Var. Ho. 3. 
 
 Early ears, middle 5,879 3,543 2,336 
 
 1893 
 Yar. Ho. 1. 
 
 Early ears, middle 5,835 5,796 66 
 
 extremity 5,492 4,425 1,067 
 
 Late middle 5,869 4,347 1,522 
 
 " extremity 5,291 4,491 800 
 
 Var. Ho. 2. 
 
 Barly ears, middle 5,i4E 5,035 107 
 
 extremity 5,587 5,242 345 
 
 Late middle 6,330 4,543 1,787 
 
 extremity 4,897 4,393 504 
 
 Var. Ho. 3. 
 
 Early ears, middle 6,365 6,161 204 
 
 The seed was sown in drills eight inches apart and eight 
 inches in the rows. They were sown with large-seed drills 
 alternating with small-seed drills. The author draws no 
 conclusions. 
 
 Two years later (5) he reports experiments with large 
 kernels selected from a crop grown from large seed for three 
 years and with small seed grown from small seed for several 
 years. Five varieties of wheat were used. The average re- 
 

 
 - 
 
 ' 
 
 
 . 
 
 
 
 
 ' 
 
 
 , 
 
 . 
 
 > ' f 
 
 . 
 
 
 . 
 
 
 
 
 
 . 
 
 . 
 
 
 
 1 
 
6 
 
 suits for three years were a difference of from 1,067 to 
 1,828 kilograms of grain per hectare in favor of the large 
 seed. But the difference was in general greater the first 
 year than later. The use of large seed gave a crop with 
 kernels larger than those grown from small seed. 
 
 Middleton (6) reporting experiments with wheat, oats, 
 and beans, says the large seed yielded almost twice as much 
 as small seed, in the case of wheat; the difference was less 
 marked with oats and with beans there was scarcely any. 
 
 Bolley (7), of north Dakota!^, after experiments lasting 
 for four years in which plump kernels of large size and plump 
 kernels of small size were selected for seed, concludes from 
 the results that "perfect grains of large size and greatest 
 weight produce "better plants than perfect grains of smell 
 size and light weight, even when the grains come from the 
 same head". 
 
 After presenting data of fourteen years' experiments 
 with large, mediuiu and small seeds of oats, barley, field 
 peas, spring and winter wheat, mangels, sugar beets, swedes, 
 fall turnips, field carrots, rape, and potatoes C. A. Zavitz(9) 
 of the Ontario Agricultural College concludes that "it 
 seems very evident that large seeds will give a greater 
 yield than will an equal number of small seeds, in the case 
 of at least twelve different classes of farm crops." 
 
 *? i 
 
 In a report of the same college (8) he submits the follow- 
 ing table: 
 

 
 
 - 
 
 
 
 ' 
 
Class 
 
 Years 
 
 seleo- : 
 
 bu. 
 
 Average 
 
 yield per A. 
 
 
 test 
 
 tion : 
 
 wt. : 
 
 tons 
 straw 
 
 : bus. 
 : grain 
 
 Spring wheat 
 
 8 
 
 L. plump 
 S. n 
 Shrunken 
 
 59.1 
 58.3 
 56.9 
 
 1.4 
 1.3 
 1.2 
 
 1.7 
 
 18.0 
 16.7 
 
 Winter wheat 
 
 6 
 
 L. plum: 
 S. " 
 Shrunken 
 Split 
 
 59.4 
 59. 
 59.1 
 54. 
 
 .6 
 . 
 .1 
 .6 
 
 46.9 
 40.4 
 39.1 
 9.3 
 
 Oats 
 
 7 
 
 L. plump 
 Medium 
 Small 
 
 33. 
 32. 
 31.8 
 
 1.9 
 1.8 
 1.8 
 
 6.0 
 
 54.1 
 46.6 
 
 Barley 
 
 6 
 
 L. plump 
 S. n 
 Shrunken 
 Broken 
 
 49.5 
 48.8 
 49.1 
 48.6 
 
 1.5 
 1.5 
 1.4 
 1.3 
 
 53.8 
 50.4 
 46.0 
 42. 
 
 Peas 
 
 6 
 
 Large 
 Small 
 
 56.3 
 56.3 
 
 1.3 
 
 1.1 
 
 8.1 
 3.0 
 
 tr 
 
 9 
 
 Sound 
 Split 
 
 58.1 
 59.7 
 
 1.4 
 
 .6 
 
 9. 
 10. 
 
 P. P. Deherain (10) reports a slightly "better yield of 
 wheat from large seed and he and Dupont (11) are cited as 
 reporting yields from large and small kernels of a number 
 of varieties of wheat to have been in all oases in favor 
 of the large kernels, "but a large difference in yield was 
 obtained only when there was a marked difference in the weight 
 of the kernels. 
 
 Soule and Vanatter (1) of the Tennessee Experiment 
 Station conducted experiments for three years in which 
 large and small wheat kernels were separated by means of selves, 
 after the first year each grade being selected from wheat 
 grown from a similar grade. A check plot of unseleeted 
 seed was planted. 
 
 "The average difference in yield at the end of three 
 years between large grains (689 per ounce), and small grains 
 

 
 ' 
 ' 
 
 ' 
 
 - 
 
 
 J-et ' ' 
 
 - 
 
 ' 
 ' 
 
6 
 
 (888 per ounce) with Mediterranean wheat, was 2.06 bushels 
 in favor of large grains as compared with the commercial 
 sample, aad 5.18 bushels in favor of large grains over 
 small grains. The difference in yield between the large 
 grains and the commercial sample chiefly occurred the first 
 year; but it is possible, though hardly probable, that the 
 difference was partly due to variation in the soil. The 
 experiment has been carried out in different parts of the 
 field for the last two year4 and the difference in the yield 
 is now only 0.32 in favor of the large grains." 
 
 Cobb (13) gives a summary of the comparisons of the re- 
 sults of some very carefully conducted experiments to deter- 
 mine the relative advantages of large seed and small seed 
 of wheat. He used sieves with rectangular openings of meas- 
 ured width which give a fairly accurate grading on the basis 
 of diameter of kernel. Three grades are cited in the summary: 
 very large, large, and medium which were respectively 3.25, 
 .00 and 2.50 millimeters in diameter. 
 
 If the comparison is made on the basis of yield of grain 
 and straw, the following is a general statement of the results; 
 
 The very large or 3.25 grade excelled in 66.7$ of the ttials. 
 " 3.00 89.7$ n " " 
 
 medium n 2.50 98.1$ " " " 
 
 * 
 
 This statement fails to tell the whole story for the 
 reason that the excess of yield in the majority oases is 
 greater than in the minority cases. Thus 
 
 Grade When excelling did When excelled 
 
 so by per cent was only by $ 
 
 Very large or 3.25 14.7 8.5 
 
 Large or 3.00 30.4 12.3 
 
 Medium or 2.50 44.5 3.5 
 
 In these statements the basis of the percentage cal- 
 culations is the weight of the lower of the two constants. 
 

 
 
 
 
 
9 
 
 If instead of taking the sum of the weights of grain and 
 straw as the criterion of yield, we take the weight of the 
 grain alone, we arrive at the following: 
 
 The very large or 5.25 grade excelled in 58.3 % of the trials. 
 
 large " 2.00 n " n 93.1 " " " n 
 n medium " 2,50 " " " 86. E " " " " 
 
 As before, we find, however, that the victories of the 
 large seed are more decisive than those of the small seed, 
 and this must be taken into account in estimating the superior- 
 ity of the larger grades. We find that: 
 
 Grade Excelling by Excelled by 
 3.25 12.9 < 6.6 % 
 
 3.00 26.5 % 7.0 % 
 
 2.50 40.5 % 3.3 $ 
 
 If instead of taking the sum of the weights of grain 
 and straw as the criterion of yield we take the yield of the 
 strew alone, we arrive at the following: 
 
 The very large or 3.25 grade excelled in 66.7 % of the trials. 
 large " 3.00 89.7 % " " n 
 
 medium " 2.60 " " 93.1 % * n " 
 
 But again we find that the victories of the large grades 
 are much more decisive than those of the small grades, and this 
 must not be forgotten ie estimating the superiority of the 
 larger grades. Examination of the table proves that:-- 
 
 Grade $ excelling by % excelled by 
 
 when excelling. when excelled. 
 
 3.25 19, B 10.8 
 
 3.00 29.6 19.4 
 
 2.50 40.7 4.5 
 
 The above trials include some twenty-nine varieties of 
 wheat. Also, the comparisons have been made on the basis 
 of equal numbers of seed and not between equal bulks or 
 equal weights. 
 
 Grenfell (14) selected plump and shrivelled kernels 
 from the same bulk of grain. They were sown with the plump 
 

 - : 
 
 . 
 
 . 
 . 
 . 
 
 <rf 
 
10 
 
 and shrivelled kernels in alternate rows. The germination 
 appeared much the same but the plants in the rows sowed from 
 plump grain soon began to gain on the others and kept ahead 
 for the remainder of the season. The tillering was better 
 in the plump grain. The average of the plants that grew was 
 92.7 % for the plump and 88.5$ for the shrivelled: the plump 
 produced and average of 180 heads from each row which origin- 
 ally was planted with 150 seeds, giving a tillering power of 
 1.J5E. The tillering power of the shrivelled seed was 1.23 
 and the yield per acre in bushels was 9.8 for the plump and 
 7.5 for the shrivelled. 
 
 Von Liebenberg (15) working with red clover fount that 
 the smaller the seed the less favorable the result. 
 
 P. de Caluwe (16) working with oats to ascertain the 
 effect of different sizes of seed reports that very large, 
 large, ordinary, small and very small grains of oats were 
 sown; all at the rate of 111 pounds per acre. The yields of 
 grain were slightly in favor of the small and ordinary seed. 
 The result was attributed to the greater number of grains 
 sown per plat with the smaller classes of seed. The author 
 considers it advisable to employ a greater weight of large 
 than of small seed. 
 
 In a test of whole and cut tubers of potatoes the same author 
 (17) says that the largest yield, after deduction of the seed 
 potatoes planted, was afforded by the large whole tubers, 
 and the next largest yield by the medium sized whole potatoes. 
 
 The yield of large and small seed of wheat is reported 
 by Despres (18) as being the greatest from large seed except 
 in two cases where there was a larger yield of straw for the 
 

 
 
 
 ' 
 
 
 
 . 
 
 
 
 ' 
 
 
 
 . 
 ' 
 
 
 
 
 
 . 
 
 
 
 ' 
 
 
11 
 
 small seed. 
 
 By growing selected lots of large and small radish seed., 
 Mr. B. T. Galloway (19) found that the large seeds germinated 
 more quickly and with more certainty and produced marketable 
 crops sooner and more uniformly than the small seeds. The lat- 
 ter, however, gave proportionally larger plants. He submits 
 the suggestion that the radish has "been cultivated for the root 
 and selection has been made continually with the development 
 of the root in view without attention to the seed and says 
 that if more nutrition is utilized in root development with 
 plants of equal vigor, less would probably remain for seed 
 development, resulting naturally in small seed. 
 
 The reasoning in this case, however, does not appeal to 
 the writer of this thesis as it would seem that the very fact 
 of a large uvuolus being present would result in better seed 
 if the conception of its function as a storage organ for the 
 production of seed is correct. Rather the answer must be 
 looked for in some natural tendency if it should always prove 
 to be the case that large seed gives small root development. 
 It is not at all uncommon to find that some undesirable quality 
 is emphasized Just as rapidly as the desired one when selection 
 is practiced in animal breeding, and selection for root alone 
 might result in small seed merely according to the chances 
 in selection. 
 
 In experiments with wheat, barley, oats and sugar beets 
 Lubanski (20) reports that the results show the influence 
 on yield and to some extent on quality of crop was in favor 
 of large seed. 
 
 According to the Ontario Agricultural College experi- 
 ments (21) with wheat, oats and barley during a five to eight 
 

 ' 
 
 
 
 , 
 
 
12 
 
 year period show that the average yield of grain and straw 
 and the weights per measured "bushel were in favor of large 
 plump seed as against either medium size or small seed. 
 
 7/ith oats the Aberdeen and lorth of Scotland Agricul- 
 tural College (22} found that when equal "bulks of small and 
 large seed were planted the "results were slightly in favor 
 of the small seed*. The two lots of seed were selected 
 from the same field. . 
 
 The \Vest of Scotland Agricultural College (23) reports 
 a series of experiments with oats in which the oats were di- 
 vidAd into (a) singles, or kernels that hear rudiments that 
 might have developed into "bosom oats, ("b) and firsts, or 
 oats in which the rudiments did so develop. The small seeds 
 are made up of seconds or "bosom oats broken from the firsts. 
 The firsts produced heavier yields of straw and grain than 
 either the seconds or the singles. The seconds produce^ more 
 straw and less grain per acre than the singles. The differences 
 reported range from twenty- three to thirty- six per cent in 
 terms of the weights yeilded hy the singles. 
 
 The Vermont Station (24) reports the influence of size 
 
 of seed in golden wax "beans as follows: 
 
 Season 1910 (green weights) 
 
 Size "So. T7t. UQ. Wt. Av. wt. pods 
 
 seed used Seed plants green per plant 
 
 grams pods grams. 
 
 lagge 500 312 426 5,894 12.8 
 
 Uedium 500 252 429 4,762 11.1 
 
 Small 500 169 370 3,629 9.8 
 
 Air dry weights 
 
 Size I>ry Total Tfo. ITo pods Ho. Wt. 
 
 seed wt. Ro. empty w. 1 or of "beans 
 
 pods. pods pods more "beans "beans grams 
 
 Large 1,899 2,545 1,580 965 2,503 964 
 Medium 1,389 2.032 1,390 642 1,578 624 
 Small 709 1,211 694 517 1,445 567 
 

 
 
 
13 
 
 The same station shows the following results with sweet 
 pumpkins:-- 
 
 Size lb. *o. total No. pounds 
 
 seed vines fruits wt. ripe ripe 
 
 rbs fruits fruit 
 
 Large 76 212 1,034 133 696 
 
 Small 76 256 1,247 68 386 
 
 Fo. fruitsTTt. 
 
 fruits unripe 
 
 unripe fruit 
 
 80 388 
 
 188 361 
 
 This shows that the plants from large seed (76 vines) 
 yielded sixty- five more ripe fruits and three hundred and 
 ten more pounds of editle product. 
 
 Fron the same station we have an interesting tabulation 
 of results with sweet peas. 
 
 Siae 
 
 "Kb. 
 
 Tit. seed 
 
 Total Ho. Total Ho. 
 
 Average Ho. 
 
 seed 
 
 sown 
 
 grams 
 
 plants "blossoms 
 
 "blossoms 
 
 
 
 
 
 per plant . 
 
 Duke of "'estminster 
 
 Large 
 
 50' 
 
 4.58 
 
 45 11,831 
 
 263 
 
 Lledium 
 
 50 
 
 3.70 
 
 40 10,069 
 
 E52 
 
 Small 
 
 50 
 
 2_..74 
 
 40 8,240 
 
 206 
 
 
 
 
 Earliest of All 
 
 
 Large 
 
 50 
 
 6.25 
 
 40 3,496 
 
 87 
 
 Iledium 
 
 50 
 
 4.11 
 
 33 2,932 
 
 89 
 
 Small 
 
 50 
 
 2.51 
 
 16 567 
 
 35 
 
 
 
 
 Her Llajesty 
 
 
 Lex** 
 
 50 
 
 4.50 
 
 34 12,244 
 
 360 
 
 Medina 
 
 50 
 
 3.19 
 
 35 10,585 
 
 321 
 
 Small 
 
 50 
 
 S.45 
 
 18 3,702 
 
 206 
 
 
 
 
 Cocclnia 
 
 
 Large 
 
 50 
 
 5.13 
 
 40 16,226 
 
 407 
 
 Medium 
 
 50 
 
 3.07 
 
 40 13,578 
 
 339 
 
 Small 
 
 50 
 
 4.22 
 
 46 14,673 
 
 19 
 
 
 
 
 Agnee Eckford 
 
 
 Large 
 
 30 
 
 3.13 
 
 . 25 1,701 
 
 68 
 
 LlecJiua 
 
 30 
 
 2.51 
 
 24 2,199 
 
 89 
 
 Small 
 
 30 
 
 1.87 
 
 25 2,145 
 
 32 
 
 
 
 
 Apple Blossom 
 
 
 Large 
 
 50 
 
 5.40 
 
 39 11,678 
 
 315 
 
 Medium 
 
 50 
 
 3.78 
 
 29 10,443 
 
 266 
 
 Small 
 
 50 
 
 2.65 
 
 38 9,592 
 
 273 
 
 They conclude that large seed is heavier, germinates 
 "better, seems more likely to produce strong plants, and 
 more important, yields a larger number of "blossoms. Also 
 

 
 . 
 
 I 
 
 . 
 
 , 
 - 
 
 
14 
 
 their results with. Hubbard squash indicate that while the 
 seeds do not vary as widely as those of some other plants, 
 large seed gave more pounds and more numbers of edible squash- 
 Let tuoe responded very uniformly, the larger seed giv- 
 ing heavier heads and more salable ones. 
 
 Effect of Large Seed on Germination. 
 
 Kerpelley (25) found that the growth of plants from large 
 and fully developed wheat kernels was more uniform and vig- 
 orous and produced the largest number of seeds capable of 
 germinating. 
 
 A. J.J. Vandevelde (26) reports on the germination of 
 1800 each of large, average and small seeds of peas, oats, 
 rye, wheat and barley. He states that the time required for 
 germination is longer for the large seed but only slightly 
 so. Total germinations were greater with small in every 
 case except barley. 
 
 Cobb (27) Gives as an average of experiments with 
 twenty eight varieties of wheat these figures. Out of 
 two hundred seeds each, large seed showed 12.5 failures, 
 medium seed 18.6 and small seed 34.1 failures. This gives 
 almost three times as many failures in small as in large. 
 
 Eiseninenger (28) found in experiments with spruce and 
 Scotch and Austrian pine that large seeds germinated quicker 
 and reached the period of maximum germination earlier than 
 small seeds. 
 
 In certain experiments by !.. Slegrell (29) the author 
 points out the fact that with one exception the period re- 
 quired for germination increased with the specific gravity. 
 A comparison of germination by the heavier and lighter 
 kernels in the same head is given "by the Kansas State Agri- 
 

 
 
16 
 
 cultural College: (30) 
 
 Ho. kernels Ho. kernels per cent 
 Grade tested germinated germination 
 
 Heavier than the average 4,707 4,684 98.74 
 
 Lighter 2.972 2.877 96.80 
 
 Total 7.679 7,561 97.99 
 
 Separation into five sizes by sieves showed; There is 
 a slight "but not constant difference "between the germination 
 of kernels separated "by sieves. (2) The largest kernels do 
 not always germinate the "best. (3) The smallest kernels 
 usually but not always germinate the poorest. 
 
 Separation according to density by means of a wind 
 blast gave the following results; 
 
 Grade Average % 
 
 germination 
 
 1 99.19 
 
 2 98.00 
 8 95.16 
 
 4 87.63 
 
 5 71.70 
 
 6 53.95 
 
 The Effect of the Specific Gravity 
 Upon the Yield. 
 
 A 
 
 Pommel and Stewart at the Iowa Experiment Station (31) 
 report that of thirty-seven samples examined the average 
 specific gravity of the seed of wheat was 1.469, the amount 
 varying between 1.503 and 1.407. Their method was to first 
 weigh the seed in air and again while immersed in Kerosene 
 oil. That this does not represent the range of the specific 
 gravities of the wheat kernel is readily seen from the suc- 
 ceeding reference. 
 
 "Seed wheat of four varieties was separated by Church (32) 
 by means of solutions of calcium chloride having specific 
 
 gravities of 1.247, 1.293, and 1.31. The seed was first 
 

treated with a solution of mercuric chloride to remove the 
 adherent air. Each lot of seed was planted separately. From 
 the results the following oonolusions are drawn: 
 
 1. The seed wheat of the greatest density produced the 
 densest seed. 
 
 2. The seed wheat of the greatest density yielded the 
 largest amount of dressed grain. 
 
 3. The seed of medium density generally gave the largest 
 number of ears, but the ears were poorer than those from the 
 densest seed. 
 
 4. Seed of medium density generally produced the largest 
 number of fruiting plants. 
 
 5. The seed wheat that sank in water, but floated in a 
 solution having the density 1.247, was of very low value, yield 
 ing on an average only 34.4 pounds of dressed grain for every 
 100 yielded "by the densest seed." 
 
 Working with soy beans Hicks and Pabney (33) found that 
 the plants and roots were heavier from heavy seed and the same 
 was true with Alaska peas with the additional advantage of 
 earliness. Earliness also developed with radish and Kafir 
 corn from heavy seed. Also, with barley, oats, and rye the 
 weights of the seedlings were closely proportional to the 
 weights of the seed. 
 
 Yon Llebenberg (34) says that the results of the ex- 
 periments with Hanna barley grown in five different places 
 "corroborate those of other experiments and it can he safe- 
 ly stated that with a heavy weight per "bushel go a greater 
 weight per kernel, a smaller percent of husk, and a richer 
 content of extract, and finally a greater yield." 
 
 The North Bakotah Station, (35) from a four years' 
 

 
 - 
 
 ' 
 
 . -. . , 
 
 
 
 
 ' 
 
 
 - 
 
 
 
 ;- 
 
 
heavy seed : 
 
 light seed 
 
 : riff, favor hee^ 
 
 29.4 
 27.9 
 27.1 
 38.6 
 
 24.3 
 22.8 
 25.2 
 33.7 
 
 4.6 
 5.1 
 1.9 
 4.9 
 
 17 
 
 test, concludes that perfect grains of large size and high 
 weight produce better plants than those of smaller size 
 and weight even though the grains come from the same spike. 
 
 The Burdue University Experiment Station submits a 
 summary of results taken from a number of stations thus: (36) 
 
 Yields in bu. per acre 
 
 Stations 
 
 Minnesota 
 Nebraska 
 Kansas 
 Ontario 
 
 Averages 30.7 26.6 4.1 
 
 Large seed : Small seed : riff, fav. Lge 
 Indiana 30.5 27.9 2.6 
 
 'Ohio 16.3 16.3 
 
 Ontario 46.9 40.4 6.5 
 
 Tennessee 28.6 23.4 5.2 
 
 Averages 30.6 27.0 3.6 
 
 The figures from experiments by Burnett at the Iowa (37) 
 Station tend to show that with oats the heavy seed and the 
 light seed were about equal pound for pound but not seed 
 for seed, nor measure for measure. All plots were sown at 
 the rate of three bushels per acre. 
 
 The Minnesota Station (38) state that a plump heavy 
 kernel of grain will produce a stronger plant and nourish 
 it better than will a shrunken, light $ernel. They quote 
 
 from the Nebraska station the following table: 
 
 Yield per acre in bushels 
 
 Kind of seed 1900 1901 2 yr. av. 
 
 Heavy 29.5 29.3 29.4 
 
 Light 23.0 26.7 24.8 
 
 Gain in favor of heavy seed 4.6 
 
 A similar experiment in their own station gave an in- 
 crease of 9-fc bushels per acre in favor of the heavy seed and 
 heavy wheat gave a yield of 36 per cent greater than that 
 of light wheat. 
 

 
 : 
 
 . 
 . 
 
 . 
 
 . 
 ft. 
 
 . 
 ' 
 
 . 
 
 - 
 
 
18 
 
 Lyon (39) cites Tollny as objecting to experiments by a number 
 of experimenters with various cereals in which almost without 
 exception kernels of high specific gravity produced the beet 
 yields, because no distinction was made between absolute 
 wiight and specific gravity in the kernels. He claims that 
 the value of the seeds lies in the kernels of absolute 
 heavy weight rather than in kernels of high specific gravity, 
 concluding that the specific gravity of the seed exerts no 
 influence on the yield. Lyon goes on to say that in the light 
 of the experiments that have been conducted it would seem 
 that there is a difference between seed of low specific grav- 
 ity and high specific gravity in favor of the dense seed, but 
 there is little difference between seed of high and medium 
 specific gravity. 
 
 Relation of Size and height of Kernel 
 to Composition. 
 
 It is not the purpose of this thesis to enter far 
 into this phase of the large seed factor. Therefore only 
 brief reference will be made. 
 
 Lyon (40) concludes that in general it may be said that 
 as between wheat kernels of the same variety grown in the 
 same season and upon the same soil, the specific gravity is 
 inversely proportional to the nitrogen content. This conclu- 
 sion is given after citing several experimenters who pre- 
 sent varying results. He suggests that as the ash, which 
 varies so considerably with the soil on which the wheat is 
 grown, has such a high specific gravity compared to that of 
 the other constituents this would prevent the establishment 
 of a constant relation. H further suggests that the number 
 
 and size of vacuoles would affect the specific gravity. 
 
, 
 
 T&* ' ' ' 
 
 
 - 
 
 ' ' 
 ' 
 
 , ' 
 
 r&' 
 
 - 
 
19 
 
 Comparing oata Hienrich (41) O onsidered that lie had 
 found very little difference in feeding value between oats 
 of heavy and light weight. In general the lighter grains 
 had more fibre and less carbohydrates. 
 
 Johannsen (42) found great variation in the nitrogen 
 content of fully developed ripe barley grains from different 
 heads of the same variety of barley grown under like con- 
 ditions. Ho definite law relative to this variation was 
 found but in general the nitrogen content increased with 
 the weight of the kernels. By careful selection, for four 
 years, a strain of barley was found which yielded heavy 
 
 grains low in nitrogen. A later and fuller table may be 
 found in the Experiment Station Record, Vol XII page 236. 
 
 Snyder (43) submits results to show that light-weight 
 kernels contain a somewhat larger percent of nitrogen, 
 phosphoric acid and potash than the heavy seed but the total 
 amount of these is much less. 
 
 Sperling (44) was unable to correlate weight in barley 
 seed with protein content. 
 
 Shaw (45) working with California white wheats, con- 
 cludes that the normal kernels usually carry a larger per 
 cent of nitrogen than smaller kernels of the same type. 
 
 It is important to have a good start because three 
 fourths of the total mineral matter is taken from the soil 
 in the first fifty days (46) . 
 
 The Minnesota Station presents a table of the composition 
 of the ash of heavy and light weight wheat kernels. Under 
 the head of total ash in the wheat is given the number of 
 pounds of ash in every hundred pounds of wheat. The figures 
 given for the potash, lime, etc. are the per cents of those 
 
' 
 
 
 - 
 
 . '- 
 
 ' 
 
 . 
 
 aw ; 
 
 fw 
 
 - 
 
 10 
 
 
 
 
 
 3* Jfll 
 
substances found by the analysis of each ash. This wheat 
 was all grown from one lot of seed in different parts of Min- 
 nesota. (47) 
 
 Eote that the heavy seed has more potash and phosphoric 
 acid. 
 
 Composition of the Ash of Heavy and light Weight Wheat. 
 
 No B "- 
 ujt 
 
 Jalo.1 
 ask 
 
 Pdajtl- 
 
 Joda. 
 
 Lime 
 
 
 Maguey a- Iron 
 
 "t 
 Phafi*5 silic*. k 
 
 Chlor- Carbon 
 ide.3 dioxid<L 
 
 I 65 
 
 2.00 
 
 30.44 
 
 .50 
 
 3.12 
 
 12.10 
 
 .41 
 
 50.83 
 
 .76! .11 
 
 m 
 
 a. so 
 
 65 
 
 1.91 
 
 32.46 
 
 .67 
 
 2. 55 
 
 12.39 
 
 .27 
 
 50.76 
 
 .37 .06 
 
 .11 
 
 
 3 64 
 
 2.02 
 
 31.85 
 
 .86 3.16 
 
 12.54 
 
 .38 
 
 50.95 
 
 .09 --- 
 
 .04 .11 
 
 4 
 
 63 
 
 2.07 
 
 30.45 
 
 .06 3.00 
 
 12.10 
 
 .34 
 
 50.95 
 
 .24 
 
 .10 
 
 .10 1.64 
 
 5 63 
 
 1.94 
 
 31.29 
 
 .09 3.89 
 
 IE. 54 
 
 .29 
 
 50.60 
 
 .26 .07 
 
 .10 
 
 6 
 
 62 
 
 2.04 
 
 31.16 
 
 .62 
 
 3.16 
 
 12.62 
 
 .34 
 
 48.86 
 
 .35 
 
 .06 
 
 .11 
 
 1.96 
 
 7 
 
 61 
 
 1.92 
 
 30.02 
 
 .74 3.06 
 
 13.17 
 
 .62 
 
 48.04 
 
 .39 .09 
 
 .09 2.12 
 
 8 
 
 60 
 
 2.10 
 
 30.16 
 
 .26 3.02 
 
 12.86 
 
 .44 
 
 48.10 
 
 .27 .71 
 
 .07 2.86 
 
 9 59 
 
 2,08 
 
 29.01 
 
 .89 3.91 
 
 13.48 
 
 1.16 
 
 45.18 1.12 .16 
 
 .14 2.99 
 
 10,58 
 
 2.16 
 
 28.86 
 
 1.45 3.86 
 
 14.90 
 
 1.33 
 
 44.17 1.35 
 
 .10 
 
 .13 
 
 11 
 
 57 
 
 .05 
 
 28.41 
 
 .79 4.81 
 
 15.30 
 
 .81 
 
 43.52 
 
 1.67 
 
 .11 
 
 .47 3.04 
 
 12 
 
 55 
 
 2.09 
 
 28.77 
 
 .88 
 
 4.52 
 
 14.48 
 
 .86 
 
 43.07 
 
 1.92 
 
 .12 
 
 .42 
 
 3.02 
 
 In Minnesota Snyder (48) has divided small kernels of 
 wheat into two classes; those small because shrunken, and 
 those well filled though small. As "between small, shrunken 
 kernels and those large end well filled he finds that the 
 former contain the higher per cent of nitrogen. But be- 
 tween small well filled and large well filled kernels 
 the latter contain the higher per cent of nitrogen. 
 
 A balancing of the foregoing results and opinions 
 seems to leave little room for the contention that there is 
 no advantage in large seed. Conceding this, we come to the 
 question as to the reason for the superiority of large heavy 
 seed. 
 
 Endeavoring to answer this several writers make certain 
 statements as if the reasons given were established but none 
 submit experimental proof. The reason given is that it is 
 on account of the larger supply of nutrients for the young pla.n1. 
 
. 
 
 
 
21 
 
 This is advanced by Shaw when he says (49) that the main 
 reason for the difference is physiological, and lies in the 
 greater reserve supply in the larger seed. It has been shown, 
 he goes on to say, that heavier seeds possess more of the im- 
 portant plant foods (phosphoric acid, nitrogen and potash (50) ) 
 Continuing, he says that there is abundant material immediately 
 available to the plantlet and the young plant is soon in a po- 
 sition to take a vigorous hold upon the soil* 
 
 Blanohard, (51) gives the same reason and goes on to say 
 that this is a fact well understood by all who are endeavoring 
 to improve their crops by careful and practical selection of 
 
 seed. 
 
 In connection with the sub- 
 ject of effect of inheritance in 
 the case of wheat a cut, herewith 
 reproduced, in Seed Wheat by Cobb 
 (52) shows the variation in the 
 size of kernels in the same spike. 
 
 Lyon (53) quotes BUnker who gave 
 the weights of individual grains of 
 a spike which was presented as be- 
 ing an average head; the figures 
 show kernels varying from 36 to 71 
 milligrams. This cut from Cobb's 
 work bears out the observation of 
 others that the heaviest grains in a 
 spike are found in the lower middle 
 part. 
 
 A suggestion comes from the 
 
 
 *houn r 
 
 - 
 
 d with lii 
 
 i'OllJd I* 
 
 "ii the . . 
 
 tainty nf in. : 
 ng- 
 
 r tin- I] 
 
 A_:i i, tilhir.' in tin* 
 1 dat.i as to 
 varieties of wheat 
 'I'll.- record i.r his 
 Anniiji] Ufpurt of 
 :-li"\is tli.it rln- 
 
 !< \n\],,\\ 
 
 l/White i, ; m 
 
 _>portuD 
 traveuin 
 
 I>artsof the State 
 fects, and the infor- 
 
 that the samples 
 he present in.jini v 
 the following, und 
 pie Straw, :,-2/ 
 wedel, 14; Red 
 % ; Blount'a Lam- 
 
 . at the time of col- 
 J wheat being "i\\n 
 immas, and Strin- 
 *e of Purple Straw, 
 e*e varieties either 
 cause they can get 
 give the man 
 pver c-nni.-s liandv. 
 dismiss this latter 
 on, as experivn.-f 
 d opinions, and ;iii- 
 neglect to think 
 > make a living. 
 rnatives, namely, 
 ties because thc'-v 
 an get no others, 
 he former. Other 
 " within the reach 
 ' use some otlitT 
 iimot get it, nor 
 
 This Rt.-1'lns tu tin- 
 
 the farmers must 
 sey Consider them 
 o grow. 
 
 
 Fig, 11. \ 
 
 Defiance wheat, 
 with th, 
 arranged as ei- 
 tracted from one 
 
 side of i'. 
 

 
 ' 
 
 
22 
 
 Kentucky Experiment Station (54) that this may "be accounted 
 for by the fact that the florets at the center of the spike 
 are pollinated first. 
 
 Since it is so that there ia such a variation in weight 
 batr/een kernels of the same spike there are a number of men 
 who consider that it is not to "be expected that all large seed 
 should suppass smaller seed because the small seed may possess 
 the same inheritance as the large. It would seem, then, that 
 on account of this fact the smaller seed in a selection fl>r 
 size and weight would contain a larger proportion of the im- 
 mature and poorly developed seed the first time than it would 
 the second after the weaker plants resulting from such poor 
 seed would have been eliminated. The second and succeeding 
 seasons selections was practiced from wheat grown from seed 
 so selected the small size would contain a larger proportion 
 of seeds which are naturally small and this is pointed out by 
 Lyon (55) with the observation that several of the experiments 
 cited exactly fulfil these conditions. 
 
 Mechanical Aids to Selection 
 of Seed. 
 
 It is obvious that some of the finer methods of de- 
 termining the weights of seed are utterly impractical for 
 use on the farm. Thus far there seems to be but one method 
 of selecting for weight which is at all practical for the 
 average man and at the same time is generally understodd. 
 This is the air blast applied in some form. 
 
 There are various devices making use of centrifugal 
 force or of the ability of seed to bounce from taught wires 
 but at the present time the most help would seem to lie 
 

 
 
 
 
 - 
 
 X1;' 
 
 
23 
 in some type of fanning mill. 
 
 It is the opinion of the writer of this thesis that no 
 separation of seed wheat or "barley is adequate that does not 
 select for size as well as weight. 
 
 Herewith are presented a few references to literature on 
 the subject. It will "be noticed here also that there is a 
 certain amount of difference of opinion as to the value of 
 such selection "but it may "be possible that other factors have 
 affected the results of some of the dissenters or that their 
 machines have not "been mechanically what they should have been. 
 
 Montgomery (56) found that the use of a fanning mill for 
 separating heavy and light seed wheat did not result in im- 
 provement in either the yield or the quality of the grain. 
 He concludes that; as every wheat plant contains both heavy 
 and light seed, the fanning mill will give about the same 
 kind of wheat, so far as inheritance is concerned, in the 
 light as In the heavy. 
 
 The Ohio Station (57) pronounce against the fanning mill 
 but their test cannot be very well guarded as they refer to 
 hand~sorted tests which are giving results in favor of large 
 seed. 
 
 Experiments in Missouri indicate that, while the differ- 
 ence is not great, the large seed gives the best yields. The 
 fact is also pointed out that more seed by weight should be 
 used when large seed is planted. In these tests the seed was 
 put first through a fanning mill and then through a grader, 
 separations being made .into three grades. (59) 
 
 Burlison (60) found in a test of fanned and unfanned seed 
 vjheat of three varieties that the fanned showed an advantage 
 over the unfanned of 1.455 bushels per acre in yield. 
 

 
 - 
 
 ' 
 
 
 
 ' 
 
 
 
24 
 
 Kansas Bulletin 33 gives the resuls of two years' trials with 
 fanned and unf aimed seed as follows: (61) 
 
 Yield per acre 
 
 Bu. grain Tons straw 
 
 Heavy seed SI. 90 1.92 
 
 Light seed 30.03 1.62 
 
 Common seed SI. 13 1.73 
 
 The seed was graded with a clipper fanning mill. The 
 "common" was wheat as it came from the thresher "but cleaned 
 of trash. It weighed 62^ pounds per bushel. The heavy seed 
 weighed 64fc pounds, and the light weighed 60 7/8 pounds per 
 "bushel. 
 
 The statement of Montgomery above should be taken in the 
 light of the statement of page 22 of the same bulletin that 
 the first two years the lightest seed produced least, the ordin- 
 ary was nezt, and heavy seed yielded the best. Since then the 
 results have varied from year to year and the average for the 
 eight years does not show a marked advantage for either. 
 
- 
 
25 
 
 Fanning Mills. 
 
 Since a bushel of wheat contains somewhere in the neigh- 
 borhood of a million kernels, it is obvious that hand selection 
 of seed wheat is impractical except for careful experimen- 
 tal purposes. 
 
 For commercial production of small grains it is neces- 
 sary to have some means of selection which is of rather high 
 efficiency and, according to the results noted in the preceding 
 pages, there should be good returns from its use. The type 
 of machine commonly employed for such a purpose is generally 
 called a fanning mill. 
 
 The functions of a fanning mill (62) are; (1) to clean 
 the grain; (2) to grade the grain; and (3) to separate differ- 
 ent kinds of grain. 
 
 The modern mill uses several physical differences for the 
 separation of grain; among them being difference in weight, 
 difference in size, and difference in shape. Also the rough- 
 ness of the hull, and to some extent the location of the heavy 
 part of the seed, may be used. But in this connection we are 
 concerned chiefly with the difference in size and density. 
 
 Selection for different sizes may be made by the use of 
 sieves or riddles with different sizes of openings. Since 
 denser seed offers less surface to the air, an air current 
 may be utilized to select seed for weight. 
 
 The first of the machines used for cleaning and grading 
 grain made use of the air blast only and hence the term fan- 
 ning mill has come to be applied to all machines of this type; 
 although the best of them combine the principles of sieves 
 anfl si- blast. 
 

 
 . 
 
 
 
 
 
 _ 
 
 ' 
 . 
 
 ' 
 . 
 
 . 
 
 
 , 
 
 . . 
 
26 
 
 There are tvro general tyi.es of fanning mills. In the first 
 the air is directed upon the grain as it passes over the sieves; 
 and in the second the air "blast is independent of the sieves and 
 riddles. 
 
 The first of these is the older type. It has a rather larger 
 capacity for the amount of sieve surface provided and when proper- 
 ly handled will do good work. A diagram of a mill of this sort 
 is shown in figure 2. The latter type, however, has the great- 
 er refinement and is capable of more careful selections. Diagrams 
 of machines of this type are presented in Figs. 3 and 4. 
 
 From a Minnesota Bulletin we get a diagram that is quite 
 helpful to the understanding of some of the arrangements of a 
 mill of this sort. See Fig. 5. (63) 
 
 Figure G. (64) shows an arrangement where selection is be- 
 ing made on the "basis of density alone. This would also blow 
 out a great deal of the trash and dust. 
 
 Other machines are designed to make selection merely on 
 the basis of size. Herewith are presented three cuts to illus- 
 trate one means of sorting for size which should be capable of 
 considerable refinement. In any mill of this kind the efficiency 
 of selection decreases with the increase in amount of grain 
 put through per hour. These are taken from Cobb's work (65) 
 and are presented to show first (Fig. 7) the simple principle 
 on which they work and are kept clean. Here note the brush 
 for keeping the meshes cleared of grains which are Just large 
 enough to wedge slightly without going through. It seems to the 
 writer that considerable efficiency must be lost by some of the 
 commercial machines in this manner. 
 
 Fig. 8 shows a more elaborate mill which makes selection 
 of several sizes and blows out dust and chaff. 
 
27 
 
 The apparatus shown in ?ig. 9. is interesting on account of 
 the adjustability of the meshes tut the danger of "bending the 
 wires out of shape would seem to lessen its probable durability. 
 
 * 
 
 A criticism applying to the three last machines is that they 
 make use only of size in selecting seed and do not select on 
 the basis of weight as well. If Wollny (39) is right in 
 saying that the advantage of large seed lies in absolute weight 
 rather than in density, the very way to secure this absolute high 
 weight is by combining selection for size with selection for 
 density. 
 
8 
 
 PART II. 
 EXPERIMENTAL 
 
 If there Toe much truth in the assumption that the advan- 
 tage of large seed over small is due to the extra start given 
 to the very young plantlet "by the additional endosperm, it would 
 seem that removing the extra endosperm should counteract to 
 some extent if not nullify this effect. 
 
 Acting upon this assumption, experiments were undertaken to 
 determine the effect of so removing the "balance of the endo- 
 sperm at the time when the plant becomes able to support it- 
 self. The plan was to allow the seed to germinate and, as soon 
 as the young plant became able to draw its nourishment from the 
 soil and air, remove the balance of the endosperm. 
 
 Three sets of plantings were made, the second of which was 
 lost on account of the fact that wet weather kept the seeds 
 submerged until they failed to germinate at all. The plantings 
 will be described more at length later. 
 
 The seed for these plantings was prepared with considerable care. 
 Separations were made first by size. In order to do this the seed 
 was put through a series of three sieves which are shown in Fig. 
 1. Sieve llol will pass a grain of 3-j millimeters in diameter, 
 3Jo. 2 will pass grains of 2-J- millimeters and Fo. 3 passes all 
 2-^ millimeters or less. 
 
 In this manner seed was separated into two classes, all 
 not falling in these two. being rejected from the experiment. 
 The first class, which will hereafter be designated as large, 
 consists of kernels that are retained by the 3^ millimeter 
 sieve. The seed retained by the 2-jjf millimeter sieve was re- 
 jected and the kernels passed by it but retained by the 2-j 
 
29 
 
 sieve were used for the second class, which will "be referred 
 to hereafter as small seed. 
 
 These two classes of seed were each separated further into 
 heavy and light kernels "by means of solutions of sodium nitrate. 
 Seed that sank in a solution of specific gravity 1.35 but floated 
 in a solution of 1.40 was used for the heavy, and those kernels 
 which floated in a solution of specific gravity 1.26 "but sank 
 in one of 1.20 were classed as the light. In this manner there 
 were four grades of seed obtained: large heavy, large light, 
 small heavy, and small light. 
 
 The first plantings were made in the Agronomy green house 
 at Berkeley and ordinary nursery flats were used. Soil was pre- 
 pared "by taking surface soil, a clay adobe, from the field south 
 of Agriculture Hall and mixing it with about equal parts of 
 sane? . The seed v.as planted at a depth of about two inches in 
 three of the six flats used and at a depth of one inch in the 
 others. Each was planted with all of the claseee of seed ar- 
 ranged according to the following diagram in order to secure 
 results as nearly comparable as possible. 
 
 aabcdad a- 5 large heavy 
 
 babcdbc b large light 
 
 cabcdcb c - small heavy 
 
 dabcdda d = small light 
 
 aabcdad 
 
 babcdbc 
 
 cabcdcb 
 
 dabcdda 
 
 The endosperms were removed two weeks later but by that 
 time it was found that practically ell of it had been used up 
 by the plants so there could be no results expected. 
 
 But the effect of large and small and heavy and light 
 seed as such was noted in the following data taken from the 
 experiment: 
 
so 
 
 Large Small 
 
 Heavy light heavy light 
 
 Per cent 
 
 germination 77 62.5 64.5 33 1/2 
 
 Average height 24.9 eon. 22.4 Gnu 16.2 Cm. 12.0 ttm. 
 
 The second set of plantings, which was made on the thirtieth 
 and thirty-first of December of the same year, wes the most 
 comprehensive under taken. Fifty seeds each of the four classes 
 were planted at distances of four inches in each direction. 
 They were planted ten seeds in a row and five rows of each. 
 Around the whole were two rows planted in like manner and at 
 the same intervals to serve for guard rows. Hone of this seed, 
 sprouted. The seed of this lot were placed in pockets of sand 
 in a manner similar to that described for the following Bet. 
 
 The third set of plantings, was made February twentieth and 
 twenty-first at the University Farm, Davis; Cal. having had to 
 wait on late rains until the soil was in proper condition. Ow- 
 ing to the slow method of placing the seed, which is described 
 later, all could not be planted the first day and rain the next 
 compelled a delay of about a week before the last could be plant- 
 ed. Fortunately this amounted to only the guard ro on either 
 end of the experiment, or rows <) and 14. The seed was planted 
 according to the following plan. Twelve seed were planted in 
 a row at intervals of four inches and the rows were planted 
 four inches apart. The rows constituted by the first and last 
 plants of all the rows are considered guarc! rows, leaving ten 
 plantfc in each row to be considered in the experiment. 
 
 The class of seed used in each row is shown by the following; 
 Clase of seed rows planted 
 
 Small light 1,2,2, 
 
 heavy 4,5,6, 
 
 Large light 7,8,9,10, 
 
 heavy 11,12,13. 
 
31 
 
 To facilitate the later work each seed was planted in a 
 small body of loamy sand collected from the "bed of the creek 
 where it had been washed down "by recent rains. A dibble was 
 used to make a small hole about tv/o inches deep and one and one 
 half inches across the top. A moderate amount of sand was in- 
 troduced and the hole opened again with the dibble to a depth 
 of about one and one quarter inches. Each seed was placed in 
 its hole by means of a pair of forceps so that it lay with the 
 suture down, embryo uppermost end pointing in the same direction 
 as all the others, the long diameter of the seed being about 
 horizontal. This placed all the seed in the same relative po- 
 sition so it was possible later to find the endosperm with a 
 minimum disturbance of the soil about the plant. The effectiv- 
 ness of this precaution was apparent, for in the large majority 
 of oases the surfece of the soil where the plant emerged was 
 not broken while removing the endosperm. 
 
 The unabsorbed portion of the endosperms in rows 1,3,4, 
 6 ,7, 9, 11, & 13 was removed early in the afternoon of the first 
 of March. In the case of the small light seed, the endosperm 
 vvas nearly all liquified in some cases but it was not all used 
 up in any. noticeably more was left in the small heavy. Be- 
 tween the light and heavy large seed there was less apparent dif- 
 ference in the amount of endosperm remaining but between the small 
 end large size the difference was very marked. \Sfhile it was 
 nearly used up in the small seed, apparently half of the endosperm 
 in many cases and in some about two thirds remained in the large % 
 
 Owing to force of circumstances the endosperms were not re- 
 moved for a little more than twenty- four hours after it was plan- 
 ned to do so, but at the time they were removed the plants were 
 but two and one half inches in height. 
 
32 
 
 It was about two weeks "before any difference could "be noted 
 in the appearance of the rows thus treated and those not treated. 
 Then for a time there teemed to be a checking of the growth in 
 the rows treated but this difference has disappeared or is so 
 slight that at the present time it is not possible to tell 
 
 which rows have been deprived of part of their endosperm and 
 
 which rows have not. See Figs. 10 and 11. 
 
 On the twenty-second of March the heights of the plants measured 
 
 "by the longest leaf were as follows: 
 
 Heights of plants in millimeters. 
 
 Plant s 
 
 :TowB: 1; 2: 3: 4: 5: 6: 7; 8: 9: 10: 11: 12: 13; 
 
 1 152 162 170 178 218 184 216 246 248 
 
 2 142 191 148 199 222 171 197 236 218 226 158 
 
 3 175 169 161 141 208 182 163 208 218 215 240 170 
 
 70 150 128 176 193 199 154 222 E15 23 222 188 
 
 5 176 200 90 186 199 204 223 201 219 216 211 236 
 
 6 110 174 136 128 196 152 232 198 188 230 2Z6 160 221 
 
 7 150 226 153 226 119 199 210 228 187 276 220 248 213 
 
 8 140 201 173 209 228 220 236 239 226 221 210 
 
 9 175 200 196 200 193 233 210 244 211 176 226 238 
 10 130 197 196 SIC 118 201 161 220 233 234 217 245 
 Averages 138 185 164 171 179 188 200 214 214 28 220 222 213 
 
 '.Vhile the longest leaf may not be a very reliable criterion 
 of the development of the plant, the averages represent fairly 
 well the superficial appearance of the experiments at the time. 
 It took a careful scrutiny to make sure which rows had had the 
 endosperm removed and which had not, if no reference was made 
 to the record. But on the other hand there was no difficulty 
 in deciding which rows were plantec with large seed and which 
 from small. 
 
 ?ig. 1C shows the appearance of the plat on the fourteenth 
 of April, and on the twenty-seventh a plant was taken from each 
 row, which was as nearly as possible a representative plant 
 for the row. Herewith, Fig. 11, is shown the comparison of the 
 development of the plants. There seems to be no definite oorre- 
 

33 
 
 lation between the removal of the endosperm and the comparative 
 development at this time. 
 
 The numbers run from right to left in order that they may 
 correspond with the location of the rows in Pig. 10. It was 
 necessary to take the picture of the plat from this side on 
 account of more advantageous light conditions. 
 
 COUCLUSIOES. 
 
 Though perhaps it is futile to attempt conclusions at this 
 stage of the experiments, or in the light of the various diffi- 
 
 ties in the way of making a completely satisfactory trial, 
 it seems to the writer that there is considerable doubt east 
 upon the theory of the extra food, supply. If this factor were 
 at all the most important one it would seem that the removal 
 of one half or two-thirds of the material would certainly cause 
 a profound difference in the development at least of plants 
 from the same class of seed. Moreover, the difference between 
 the effect on the small seed and on the large should be less 
 u^on the small seed, proportionally since less endosperm in pro- 
 portion was removed from it than from the large. The facts as 
 developed from an examination of the figures in regard to the 
 heights show that the reverse is the case and that the effect 
 on the small amounted to 9,2$ while the effect on the large 
 was but 4,2,1. 
 
 is certain that all the factors concerned in what we are 
 pleased to refer to as vitality and vigor, etc., are really 
 but little understood. Therefore since there is something 
 which we refer to as vigor, or response to environment, or 
 indicate by various other terms, is it not possible that it 
 

 ;-. 8rt 
 
 _ 
 
 
 
 
 
 
 
 
 
 
24 
 may be present in varying degrees in the germ itself? 
 
 Grant, for the sake of argument, that it is an inherent 
 quality of the soluble proteins in the embryo. *7e must consider 
 the embryo not as a gamete or as a pair of them but as a zygote. 
 The combination of the hereditary equipment has already been 
 made and determined. 2ven though the pollination is by the 
 same plant there is a chance for a certain amount of varia- 
 bility on account of the maturation process. Then is it not 
 entirely possible that a certain degree of the capability of the 
 plent is already determined in this "plant packed and ready 
 for shipment". There may tie some correlation between the ex- 
 tent of the development of the seed and the capability of mak- 
 ing a large plant under proper circumstances after planting. 
 A young animal loses its power of response to proper food quite 
 early in life, as is common knowledge to stock men. It is en- 
 tirely concievable to the author that some measure of loss of 
 power to develop may be due to the same conditions, whatever 
 they are, that cause the development of a smaller endosperm. 
 
 It becomes necessary to discover some other factor than 
 hereditary equipment on account of the variation of th<- weight 
 of kernels in the same spike, if their equipment be considered 
 the same, and this certainly is not shown to be extra food supply 
 
 . 
 
 by the experiments here recorded. 
 

 
 
 
 
 ' ' 
 
35 
 
 'REFERENCES. 
 
 1. Seed wheat by B.A.Cobb. I.S.Wales. Mi so, pub. Ho. 6E5 p. 2. 
 
 2. S.P.I. Bui. #78 p. 30 from Utah Ex. Stn. Rpt 1892, p. 168. 
 
 3. Same p 31 from Indiana Ex. Stn. Eul 36, p. 110-128. 
 
 4. Abstract in Expt. Stn. Record , Vol. V. p. 526. 
 
 5. Same Vol. VII. p. 679. 
 
 6. Same Vol XII. p. from Univ. College of Wales Eep't. 1899, p. 68-70 
 
 7. I. Dak. Exp. Stn. Kept. 1901, p. 30. 
 
 8. Eept. Ontario Agr. College 1911, p. 182. 
 
 9. Abstr. E.S.Eec. V. XZVI. p. 434 from Proc. of Amer. Soo. of 
 
 Agron. 1(1907-1909) pr,. 98-104. 
 10. ; 'jon. Agron. 26(1900) Ho. 1. pp. 20-23. 
 
 p. 672. 
 
 11. B. P. I. Eul. Ho 78 p. 32. Abstr. in Es. Kec v Xv, 
 
 12. Same from Tenn. Bui v. 16, Ho. 4, p. 77. 
 
 13. Same as reference 1. Page 38. 
 
 14. B.P.I. Bui 78, fr. Agr. Gazette of H.S.V:ales, 12(1901) #9, 
 pp. 1053-1062 
 
 15. E.S.S.v V. p. 404, from Centbl. Agr. Chem. p. 404 
 
 16. E.S.E.v VII. p. 209. 
 
 17. Z.S.E.v VII. p. 209 
 
 18. S.S.E.v IX. p. 553. 
 
 19. Year Book U.S.D.A. 1896, p. 92. 
 
 20. ^ S.B. v ZIV 432. 
 
 21. Cal. Ex. Stn. BuL 181, from Eept. Ont. Agr. Farm, 1901, pp 82-111 
 
 22. E.S.E.v 111. p. 37 from Aberdeen & H. of Scotland Coll. of 
 Agr. Bui. 11, 1907, p. 8. 
 
 23. E.S.E.v XXVI p. 636, From West of Scotland Agr. Col. 
 Eept. 10(1911) pp. 225-229. 
 
 24. Vt. Agr. Exp. Stn. Eept. 1913-1914, Bui. 177. M.E.Cummings. 
 
 25. E.S.E.v V. p. 226 fr. Centbl. Agr. Chern. 1892 p. 545. 
 
 26. S.S.E.v XI. 353. 
 
 27. Same as ref. 1, p. 31. 
 
 28. S.S.E.v XVIII. p. 47. 
 9. E.S.B. v X21. p. 632. 
 
 30. Giro Ho. 11, Kansas State Agr. Col. 
 
 31. Bui. 25, Iowa Sxp. Stn. 
 
 3.P.L Bui. 78, p. 38. 
 
 U.S.I'.A. 1'ear Book 1896, pp. 305-322 
 
 E.S.E.v VIII. p. 117. 
 
 Eept . Dak. Stn. 1901. pp. 30-44. 
 
 Circ. So. 23, Purdue Unlv". Ept. Stn. Ind. 
 
 Bui 128, Iowa Exp. Stn. pp. 93-127 
 
 Bui 115, Minn. 2xpt. Stn. p 
 Same as reference 1. p. 38. 
 
 368 
 
 32, 
 33, 
 ?4, 
 35, 
 36, 
 ?7, 
 38, 
 39, 
 40- 3&me, p. 40 
 
 41. E.3.E. v. VII, p. 497. 
 
 42. E.S.E. v. XI p. 633 fr. Chem. Ztg. 23(1899) So. 74, p. 275. 
 
 43. Bui 90, Minn. Expt. Stn. 
 
 44. E.3.E. v. ZS.III p. 336 fr. Illus. Landw. Ztg. 30(1910) 
 i;o. 19, pp. 175-176. 
 
 45. Bui. 181, Cal Expt. Stn. p. 394 
 
 46. Bui. 29, iiinn. Expt. Dtn. 
 
 47. same , p. 149. 
 
 48. 3ul 85, ilinn. Expt. Stn. 
 
 49. Bui 181, Cal Expt. Stn. p. 163. 
 
 50. Bui 29, iiinn Zxpt. Stn. 
 
 51. B.P.I. Eul 178 p. *3. 
 
52. Same as reference JTo. 1. page 5. 
 
 53. B.P.I.Bul. 78, p. 33 
 
 54. Bui. 113, Kentucky Expt . Stn. 
 F5. B.P.I. 78, p. 34. 
 
 56. Bui 104, Uebr, Expt. Stn. p. 34 
 
 57. Bui. 231, Ohio Expt. Stn. p. 11. 
 58. 
 
 59. Bui. 21, Mo. Sxpt. Stn. 
 
 60. Bui. 128, So. Dak. Expt. Stn. 
 
 61. Bui 33, Kan. Sxpt. Stn. p. 14. 
 
 62. Agricultural .Engineering "by Davidson, p. 281 on. 
 
 63. Bui 115, 1,'inn. Agr. Expt. Stn. p. 369. Kept. 1908-1909. 
 
 64. 7th Ann. P.ept. Amer. Soo. Agr. Engineers, p. 57. 
 
 65. Sante ae ref. So 1. pp. 50, 52, 53. 
 
inin 
 mini 1 
 
 minimum 
 
 iimimiiiiiii 
 
 iiiiiiiiiiiiiiiniii 
 
 ~ 
 
 Jig. 1, 
 Original . 
 
 Fit' ISO. A section of a fanning mill in 
 l,l;,st is directed below an-1 
 i lie sieves. 
 
 J-ig 2. 
 
 Jfrom Agricultura^ Engineering, Davidson. 
 
I.-;., ITS A section of a fanning mill in 
 which h<' blast does not strike the gram 
 until after it has passed through the sieves. 
 
 ig. 3. 
 
 Fig. 179. Another view of the type of 
 machine shown in Fig. 178. 
 
 Jig, 4. Both above from Agricultural Engineering, Davidson. 
 
FANNING MIT. I. SELECTION. 
 
 39 
 
 Shake',' Mill for Si-paratinj,' S > 
 
 fvrei-n NO. T H! I'uhl be Just coarse i!i-> ii-h i" b-i i: 
 It- is used sjni|.lv in run oft .- ; i i'-K 
 
 Ki'iiin bai-kw-aril in the null, so as to l.-i ii drop thro I n 
 
 ...... , .Tin' li^ht Uci-iifls ;M-f blown pasl tin.- >ri:,l cif screen No. ''. 
 
 lic-.icji'i' kiTnc'I>; f::ll . No, ". Roard 'No. i' may bo moved to 
 
 ..kw:ii-il I" thro mull per cent <>( u-i-Min mi screen X" 
 
 ::. ns ili'sin-il. Screen No. :; shdMlil )" coarse I-IUIM.II to ! i the small 
 Is i hrough onto ci een X". i. 
 
 It is ;Mijiis;:,i,i,- :is tn'slMiii an,d maj i> mnvi'd forward r backwurd 
 in n-j;-nlate the amount ni' ^niiii it will catch. Screen X" I is inn- 
 oninii;li t" c:i:-ry nearly nil of the KIM!:) over into market grain. An> 
 <=nli- shake mill may be tixi-il nc in this \v;iy. Si'parai inn by weight caii 
 easily !>._> made wi.th tho cnil shako mills, but tin- I:H-KI- ke in In- 
 
 separated from the small ones' in any proportion ilesiri-d. simply by 
 using coarse or fine scve.-ns in the haver part of mill. 
 
 Jig. 5. (62) 
 
 i-"i. s,'trti>i>f Grain .\<-<- ( n'(lina to Weight. 
 
 . 6. (64) 
 
40 
 
 Fig. So. 
 
 . 
 
 fall on 
 
 hitiir, but nt ' 
 Imslit'j- , 
 
 
 . t.y j.irmers. The 
 Le crank, E. A brush, A A, is 
 -ulated by 
 , ink- ihe tailings 
 i .1 50 bushels per 
 per liour as at 20 
 r it i done. The brush, A A, 
 .ne ei*e of 
 ailiiji,' <-,iii still be done by working the 
 
 -^till be <loiie by working the 
 cimteoiltpictgrardea iuuy be prepared 
 
 wit)' This nun-Inn.-, :ii ' For Australian wheat tne 
 
 
 IPig. 7. (65) 
 
 . 34. 1, li"!-r revolving 
 
 . . fot Hi.- r<'Uiv;i] .,t' COSTM !II;L 
 l,Iinm-h whii'Ii 
 In-! .nnl i-liiitl L- t'li.v. 
 
 . ted with 3: ',, *, v. to. si^m- 
 
 f nn u wlii.rh vsn'Huih '_T.t.ic-. ,il _'r;nn;ii>- 
 
 le in-u^ii 
 
 i. >!IM\MI ,H'tiii'_- mi 
 
 . 
 
 i.-. ,IL t Itn-i* piirts t-iicli riisily rrjiliir.'- 
 sililf )>y .1 >.'f.'.']i ipf ;iny 'l.'^i; 
 
 
 n.t'tii!. ivlii.-l, ii.- Iliit v. i 
 
 '.til III" lll.irllll.r-, ;i|ii| !,. I|. r 
 II!. 
 
 8. (65) 
 
41 
 
 Fig. 36. An English yrader in which the meshes are made of wire in s;n 
 13 1..- adjuxtdbU in width, without removal. The machine gives fours.-'. 
 and .lot's very irood work. At the back near the top is the brush for kr 
 meshes dear! This machine may be had with or without a fan. ' 
 
 Tig. 9. (65) 
 
42 
 
 , 10. Photograi-li "by the author. 
 
 11. Photograph. ~by the author. 
 
THIS BOOK IS DUE ON THE LAST DATE 
 STAMPED BELOW 
 
 BOOKS REQUESTED BY ANOTHER BORROWER 
 ARE SUBJECT TO RECALL AFTER ONE WEEK. 
 RENEWED BOOKS ARE SUBJECT TO 
 IMMEDIATE RECALL 
 
 LIBRARY, UNIVERSITY OF CALIFORNIA, DAVIS 
 
 D4613 (12/76)