R.ARY OF THE U N IVLR.SITY OF ILLINOIS ~" : : <*H5WS ' ,< P4* Pf 43? l3fe{ i 1 :& L v j, ^4 * Arf , UNIVERSITY OF ILLINOIS Agricultural Experiment Station, URBANA, DECEMBER, 1902. BULLETIN NO. 82. METHODS OF CORN BREEDING.* BY CYKIL GEORGE HOPKINS, PH. D., PROFESSOR OF AGRONOMY IN THE AGRICULTURAL COLLEGE AND CHIEF IN AGRONOMY AND CHEMISTRY IN THE AGRICULTURAL EXPERIMENT STATION. It is a well established fact that there now exist markets and demands for different kinds of corn. The price of corn varies, say, from > cent to 1 cent per pound. The cost of protein in the principal stock feeding- states varies from 3 to 5 cents per pound. In other words, the protein is several times more valuable per pound than corn itself. Consequently, stock feeders want more protein in corn. (Very possibly the feed- ers in the southern states want more carbohydrates to supplement their present more abundant supplies of nitrogenous food stuffs.) The price of corn starch varies from 2 or 3 cents to 5 or even 10 cents per pound, depending- upon the wholesale or retail nature of the sale. The manufacturers of starch and of g-lucose sug-ar, glucose-syrup, and other products made from starch want more starch in corn. *Read before the Section on Agriculture and Chemistry of the Associa- tion of American Agricultural Colleges and Experiment Stations, at Atlanta, Georgia, October 8, 1902; and before the Illinois Live Stock Breeders' Associa- tion, at Bloomington, Illinois, November 20, 1902. 525 526 BULLETIN NO. 82. [ December, In its own publication a large commercial concern, which uses enormous quantities of corn, makes the following 1 statements: "A bushel of ordinary corn, weighing- 56 pounds, contains about 4^2 pounds of germ, 36 pounds of dry starch, 7 pounds of gluten, and five pounds of bran or hull, the balance in weight be- ing made up of water, soluble matter, etc. The value of the germ lies in the fact that it contains over 40 per cent, of corn oil, worth, say, 5 cents per pound, while the starch is worth \% cents, the gluten 1 cent, and the hull about YZ cent per pound. "It can readily be seen that a variety of corn containing, say one pound more oil per bushel would be in large demand. "Farmers throughout the country do well to communicate with their respective agricultural experiment stations and secure tneir cooperation along these lines." These are statements and suggestions which should, and do, attract the attention of experiment station men. They are made by the Glucose Sugar Refining Company of Chicago, a company which purchases and uses, in its six factories, about fifty million bushels of corn annually. According to these statements, if the oil of corn could be increased one pound per bushel, the actual value of the corn for glucose factories would be increased 5 cents per bushel; and the President of the Glucose Sugar Refining Company has personally assured the writer that his company would be glad to pay a higher price for high oil corn whenever it can be furnished in large quantities. The increase of five cents per bushel on fifty million bushels would add $2,500,000 to the value of the corn pur- chased by this one company each year. The Glucose factories are now extracting the oil from all the corn they use and are unable to supply the market demand for corn oil. On the other hand, to these manufacturers, protein is a cheap by-product and conse- quently they want less protein in corn. Corn with a lower oil content is desired as a feed for bacon hogs, especially for our export trade, very extensive and thorough investigations conducted in Germany and Canada having proved conclusively that ordinary corn contains too much oil for the pro- duction of the hard firm bacon which is demanded in the markets of Great Britain and Continental Europe. The methods of corn breeding devised by the Illinois Experi- ment Station and now used not only by us, but also by the Illinois Seed Corn Breeders' Association, and, to some extent, by other Ex- periment Stations and other corn breeders,have for their object the improvement of corn in yield and in quality. In the main the methods are now the same as we have employed for the past six METHODS OF CORN BREEDING. 527 years and they have given results which enable us to assert with confidence that by these methods corn can be improved in a very marked degree and for many different purposes. The yield of corn can be increased, and the chemical composition of the kernel can be changed as may be desired, either to increase or to decrease the protein, the oil, or the starch. Following- is a brief description of the methods of corn breed- ing- which we practice and which we have recommended to others: PHYSICAL SELECTION OF SEED CORN. The most perfect ears obtainable of the variety of corn which it is desired to breed should be selected. These ears should con- form to the desirable standards of this variety and should possess the principal properties which belong- to perfect ears of corn, so far as they are known and as completely as it is possible to secure them. These physical characteristics and properties include the leng-th, circumference, and shape of the ear and of the cob; the number of rows of kernels and the number of kernels in the row; the weig-ht and color of the grain and of the cob; and the size and shape of the kernels. In making- this selection the breeder may have in his mind a perfect ear of corn and make the physical se- lection of seed ears by simple inspection, or he may make absolute counts and measurements and reduce the physical selection almost to an exact or mathematical basis. In this connection let me suggest that there is some dang-er of corn breeders making- too much of what might be called fancy points in selecting- seed ears. We should learn the facts which are facts and not base our selections too much upon mere ideas and opinions. For example, it is not known that ears whose tips are well filled and capped with kernels are the best seed ears. Indeed it is not improbable that the selection of such seed ears will cause the production of shorter ears and a reduced yield per acre. It is true that the percentag-e of shelled corn from a given ear is the greater, the greater the proportion of corn to the cob, but our inter- est in that percentage is very slight compared to that of yield per acre, and perhaps for the greatest possible yield of shelled corn per acre it requires that the ears shall have good sized cobs. Pos- sibly the corn which shall untimately surpass all others for yield per acre will have tapering and not cylindrical ears. These are some of the points regarding which men have some ideas and opinions but as yet we have no definite facts, and we shall need several years more to obtain absolute knowledge regarding some of these points. Let us base our selections of seed corn first upon 5 28 BULLETIN NO. 82. [December, known facts and performance records, and secondly upon what one may call his "type" of corn. CHEMICAL SELECTION BY MECHANICAL EXAMINATION. The selection of seed ears for improved chemical composition by mechanical examination of the kernels is not only of much as- sistance to the chemist in enabling- him to reduce greatly the chem- ical work involved in seed corn selection, but it is of the greatest practical value to the ordinary seed corn grower who is trying 1 to improve his seed corn with very limited service, if any, from the analytical chemist. This chemical selection of seed ears by me- chanical examination, as well as by chemical analysis (which is described below), is based upon two facts: 1. That the ear of corn is approximately uniform throughout in the chemical composition of its kernels. 2. That there is a wide variation in the chemical composition of different ears, even of the same variety of corn. These two facts are well illustrated in Table 1. TABLE I. PROTEIN IN SINGLE KERNELS. Ear A, protein, per cent. Ear B, protein, per cent. Ear C, protein, per cent. Ear D, protein, per cent. Kernel No. i 12.46 II . "iS 7-4S 8.72 Kernel No. 2 12. ;4 12.^2 7. 54 8.41 Kernel No. 3 12.44 12. 10 7.6o 8.73 Kernel No. 4 12. SO 12. Cd 7 47 8.31 Kernel No. 5 I2.3O 12 14. 7 74 8 02 Kernel No. 6 1 2. 4Q 12. OS 8.70 8.76 Kernel No. 7 12. t;o 12 84 8.46 8 80 Kernel No. 8 12. 14 * 8 60 o 02 Kernel No. 9 12 14 12 O4 8 86 8 96 Kernel No. 10 12 71 12 7S 8 10 8 89 It will be observed that, while there are, of course, small dif- ferences among- the different kernels of the same ear, yet each ear has an individuality as a whole, the difference in composition be- tween different ears being- much more marked than between differ- ent kernels of the same ear. The uniformity of the individual ear makes it possible to esti- mate or to determine the composition of the corn by the examina- tion or analysis of a few kernels. The remainder of the kernels on the ear may then be planted if desired. The wide variation in the composition between different ears furnishes a starting point for the selection of seed in any of the several different lines of de- sired improvement. The methods of making- a chemical selection of ears of seed *Determination lost by accident. 1902.] METHODS OF CORN BREEDING. 529 HIGH-PROTEIN KERNELS (little starch) LOW-PROTEIN KERNELS (much starch) PLATE i. corn by a simple mechanical examination of the. "kernels is based upon the fact that the kernel of corn is not homogenous in structure, but consists of several distinct and readily observable parts of markedly different chemical composition. (See illustra- tions.) Aside from the hull which surrounds the kernel, there are three principal parts in a grain of corn: 1. Tne darker colored and rather hard and horny layer lying next to the hull, principally in the edges and toward the tip end of the kernel, where it is about 3 millimeters, or y% of an inch, in thickness. 2. The white, starchy-appearing part occupying the crown end of the kernel and usually also immediately surrounding, or partially surrounding, the germ. 530 BULLETIN NO. 82. [ December, HIGH-OIL KERNELS (large germs) LOW-OIL KERNELS (small germs) PLATE 2. 3. The germ itself which occupies the central part of the kernel toward the tip end. These different parts of the corn kernel can be readily recog- nized by merely dissecting- a single kernel with a pocket knife, and it may be added that this is the only instrument needed by any- body in making a chemical selection of seed corn by mechanical examination. The horny layer which usually constitutes about 65 per cent, of the corn kernel contains a large proportion of the total protein in the kernel. The white, starchy part constitutes about 20 per cent, of the IQO2.] METHODS OF CORN BREEDING. 531 whole kernel, and contains a small proportion of the total protein. The germ constitutes only about 10 per cent, of the corn kernel, but, while it is rich in protein, it also contains more than 85 per cent, of the total oil content of the whole kernel, the remainder of the oil being 1 distributed in all of the other parts. By keeping 1 in mind that the horny layer is larg-e in propor- tion and also quite rich in protein and that the germ, althoug-h rather small in proportion is very rich in protein, so that these two parts contain a very larg-e proportion of the total protein in the corn kernel, it will be readily seen that by selecting ears whose kernels contain more than the average proportion of g-erm and horny layer we are really selecting ears which are above the average in their protein content. As a matter of fact, the method is even more simple than this, because the white starchy part is approximately the complement of, and varies inversely as, the sum of the other constituents; and to pick out seed corn of high protein content it is only neccessary to select those ears whose ker- nels show a relatively small proportion of the white, starchy part surrounding the germ. As more than 85 per cent, of the oil in the kernel is contained in the germ, it follows that ears of corn are relatively high or low in their oil content, according as their kernels have a larger or smaller proportion of germ. In selecting seed corn by mechanical examination for improve- ment in composition we remove from the ear a few average kernels; cut two or three of these kernels into cross sections 4 and. two or three other kernels into longitudinal sections and examine these sections as they are cut, usually simply with the naked eye. If we are selecting seed ears for high protein content we save those ears whose kernels show a small proportion of the white starch immediately adjoining or surrounding the germ. If se- lecting corn for low protein content we look for a larger propor- tion of white starch surrounding the germ. Our results have shown that the white starch in this position, that is, surrounding the germ toward the tip end of the kernel, is a better index of the protein content than the starch in the crown end. If we are selecting seed ears for high oil content we save those ears whose kernels show a large proportion of firm and solid germ; while if seed of low oil content is desired, we look for a small pro? portion of germ in the kernel. It should be emphasized that it is not the absolute, but propor- tionate, size or quantity of germ or of white starch which serves as a guide in making these selections. 532 BULLETIN NO. 82. [December, CHEMICAL SELECTION BY CHEMICAL ANALYSIS. In selecting 1 seed corn by chemical' analysis we remove from the individual ear two adjacent rows of kernels as a representative sample. This sample is ground and analyzed as completely as may be necessary to enable us to decide whether the ear is suitable for seed for the particular kind of corn which it is desired to breed. Dry matter is always determined in order to reduce all other determ- inations to the strictly uniform and comparable water-free basis. If, for example, we desire to change only the protein content, then protein is determined. If we are breeding 1 to change both the pro- tein and the oil, then determinations of both of these constituents must be made. For a satisfactory breeding 1 plot, about 20 to 40 selected seed ears are required. If the breeder desires to make only physical improvement then he should select, say, 40 of the most nearly per- fect ears which it is possible to pick out by inspection or by exact physical measurements. If it is desired to improve the composition or quality of the corn as well as the physical properties, then at least 200 physically perfect ears should be selected, and, from these 200 ears, the 40 ears which are most suitable as seed for the partic- ular kind of corn which it is desired to breed should be selected, either by mechanical examination of sections of kernels, which any- body can make, or by chemical analysis, or by a combination of these two methods. In our own work we now commonly select by physical inspection or measurement the 200 ears; then, from these 200 ears, we select by mechanical examination of sections of ker- nels the best 50 or 100 ears, and from this lot we finally select by chemical analysis the best 20 to 40 seed ears for planting 1 . This combination of methods effects a very satisfactory seed selection and requires only one-half as much chemical work as would be re- quired if the method of chemical analysis alone were employed. Table 2 shows very fairly the degree of seed improvement which may be accomplished by these different methods of selection, when breeding 1 to change only the protein content of corn. It may be stated that equally satisfactory results may be ob- tained in chemical selection by mechanical examination for secur- ing 1 seed ears of hig-h or low oil content. For example, the writer has selected by mechanical examination, from a lot of 272 ears of corn, 18 ears for high oil content which averaged 5.24 per cent, of oil; and, from the same lot of corn, 30 ears were selected for low oil content which averaged 4.13 per cent, oil, making an average difference of 1.11 per cent, of oil. 1902.] METHODS OF CORN BREEDING. 533 TABLE 2. SOME FAIR ILLUSTRATIONS OF ACTUAL RESULTS OBTAINED IN SELECTION OF SEED CORN. (Protein, average per cent.) Variety 200 average seed ears. 50 ears selected by mechanical examination. 28 ears selected by chemical analysis. 10 best seed ears. Best single seed ear. Silver Mine 10 oo 9.47 8.77 7.97 7.00 Boon County White. . Learning 10.57 ii .96 9.72 II .36 9-36 10.79 8.84 10.08 8.69 8.82 Learning i ii .96 12.44 13.33 14.03 14-63 Learning ii .27 11.84 12.43 13. 12 14.71 Yellow Dent 11.14 ii .64 12. II 12.55 13.24 Riley's Favorite ii .02 11.^8 12.41 12.99 15.78 Burr's White 12.48* 12.88 14.36 14.87 15.71 Burr's White . ,, 9.2Ot 9. 10 7-77 7.56 7.08 Learning II .26 12. 14! Learning II .26 10.67$ * Average protein content of ten field rows of Burr's White after four years' breeding for high protein. | Originally from same stock of Burr's White as preceding, but bred four years for low protein. J Two lots of 42 ears each selected from the same lot of 200 ears for two breeding plots, high protein and low protein, the seed for which is selected by physical inspection and mechanical examination but without chemical analysis of individ- ual ears. If the method of mechanical examination alone is employed in making- the chemical selection, then, if possible, there should be some chemical control of the work, at least until the breeder has become sufficiently skilled, or has had sufficient experience, to feel that he knows how to make a chemical selection of seed ears by mechanical examination of kernels. Such a chemical control does not involve a large amount of chemical work. In Illinois the Ex- periment Station offers such a chemical control to farmers who will agree to make the selection of the best possible seed, both by phys- ical inspection of ears and mechanical examination of kernels and who will further agree to secure data and breed the corn in accord- ance with our directions. This control is affected by analyzing- only two samples of corn each year; one composite sample of the rejected ears, five averag-e kernels being- taken from each ear, and one composite sample of the 20 to 40 selected seed ears, twenty averag-e kernels being- taken from each of these ears, and each of these two composite samples being- properly labeled and analyzed. One of the best selections which has yet been made by mechan- ical examination was accomplished last spring- by a farmer who is breeding- corn for higher protein content. Out of a lot of 165 ears 534 BULLETIN NO. 82. [December, of corn he selected 15 ears whose protein content averaged 1.48 per cent, higher than that of the 150 rejected ears, as was determined by the chemical analysis of a composite sample from each of the two lots. Because of the chemical control which the Station affords him, he knows each year just how much he has accom- plished. If the purpose of breeding- a kind of corn is principally to change its content of a single constituent, as to increase protein, then the selection of the best 40 ears is simple and regular by either method; but if it is desired to effect changes in the content of two constituents, as to increase the protein and to increase the oil in the same corn, then one could hardly expect to make much progress in both directions, if he relied solely upon mechanical examination of kernels for chemical selection of seed ears. Even after the chemical analyses of 100 ears have been made it requires some com- putation to determine which are really the best 40 ears. For ex- ample, an ear may be desirable for seed because of its high protein content, but it may not be sufficiently high in oil. In order to re- duce the selection to an exact basis, we have adopted simple math- ematical computations for all such cases. For high protein and high oil in the same corn, we multiply the percentage of protein by the percentage of oil and use the product as the selection coefficient, the forty highest products des- ignating the forty best ears. For low protein and low oil we multiply the percentages to- gether and use the lowest product as the selection coefficient. For high protein and low oil in the same corn, we divide the percentage of protein by the percentage of oil and use the highest quotients as our selection coefficients. TABLE 3. SELECTION OF SEED CORN FOR HIGH PROTEIN AND HIGH OIL. No. Ear Protein in corn. Oil in corn. Selection coefficient- I 11.17 6.03 67.30 2 12.66 4.90 62.00 3 13.60 4.92 66.89 4 10.85 4-55 49.89 5 I I OI >f ki > ^W r -v^ % CM /i?J? ^JBK ^^/^v ^ ' yfrt -v l vvi 9T9B ^.Jv7 .v.ciSftv /-^ .' c^ fra .. i7xV*>.r^/ ^r ^gr- ^-^r- TSXJ r^TT &> h, . ; , M - ^-