SB j;^ U. S. DEPARTMENT OF AGRICULTURE. • U^ BUREAU OF PLANT INDUSTRY— BULLETIN NO. 167. B. T. GALLOWAY, Chief of Bureau. NEW METHODS OF PLANT BREEDING GEORGE W. OLIVER, Plant Propagator. Issued Februaky 7, 1910. WASHINGTON: GOVERNMENT PRINTING OFFICE. 1910. U. S. DEPARTMENT OF AGRICULTURE. BUREAU OF PLANT INDUSTRY— BULLETIN NO. 167. B. T. GALLOWAY, Chief of Bureau. NEW METHODS OF PLANT BREEDING. GECmOE W. OLIVER, Plant Propagator. Issued Feukuakv 7, I'JIO. WASHINGTON: GOVKRNMKNT PRINTING OFFICE. 1!> 1 U . 'O'- BUREAU OF PLANT INDUSTRY. Chic/ of Bureau, Beverly T. CJalloway. Assistant Chief of Bureau, Albert F. Woods. Editor, J. E. Rockwell. Chief Clerk, James K. Jones. Foreign Seed and I'lant Introduction. SCIENTIFIC staff. David Fairchild, Agricultural Explorer in Charge. P. H. Dorsett, Albert Mann, George W. Oliver, Walter Van Fleet, and Peter Bisset, Experts. Frank N. Meyer, Agricultural Explorer. H. V. Harlan, H. C. Skeels, and R. A. Young, Assistants. Edward Gouelier and P. J. Wester, Assistant Propagators. 167 2 peg 11 1910 a^^^ 10- 2 ^C^r" LHTTHK OF TRANSMITTAL U. S. Department of Agriculture, Bureau of Plant Industry, Office of the Chief, Washington, D. C, Septemher 25, 1909. Sir: I have the honor to transmit herewith a manuscript entitled •'New Methods of Plant Bi-eeding," by Mr. Georj^e W. Oliver, Plant Propagator of this Bureau, and recommend that it be ])ublishe(l as Bulletin No. 167 of the Bureau series. Respectfully, B. T. Galloway, Chief of Bureau. Hon. James Wilson, Secretary of Agriculture. 167 3 CONTENTS' Page. Introduction 7 Origin of the new methods of plant breeding 8 Tools required for depollinatiou and emasculation 11 Devices used in depollinating flowers 12 Emasculation 13 Obstacles in emasculating the flowers of Compositsp 18 How hybrids and crosses of composite flowers may have originated in the past. 14 Preparation of seed and pollen bearing parents ]5 Condition of the stigma at the time of pollination 16 The application of water to all flowers 16 Crossing alfalfa 16 Work accomplished in the past 16 Peculiar features of the work 17 Old methods of crossing alfalfa 18 New methods devised by the writer 18 Depollinatiou by water 18 Depollinatiou by compressed air 22 The growing of the plants and their care 22 Raising large quantities of seed from a cross 23 Crossing large-flowered legumes 23 Crossing in large and small numl:)ers 25 Hybridizing species 26 Helianthus hybrids 26 Grass hybrids 27 Crossing cultivated varieties on natural species 28 Pansies 28 Dahlias 28 Cross-pollinating clovers 29 Methods of emasculating and pollinating common flowers 30 The essential organs of the flower 30 Preparation of flowers to be pollinated 30 Removal of the anthers 31 Securing pollen 32 Applying the pollen to the stigma 32 Description of plates 33 Index 35 167 r> ILLUSTRATIONS. PLATES. Page. Plate I. Fid on very easih" and the combination is an ex- ceedingly useful one. Several pairs of for- ceps which are self- closing (shown in fig. 1, B) are necessary in such lines of work as pollinating depolli- nated stigmas of al- falfas and many other small flowers. These forceps enable the operator to hold a stamen or a sexual column of a small legume while pollinating. It is difficult to perform this work satisfactorily with the fuigers. Watch glasses and small-sized moist chambers are needed to pre- serve pollen. Tags for keeping records while the seeds are ripening, ranging from one-half inch by three-c(uarters inch to those of much larger dimensions, are indispensable. Some small-sized camel 's-hair brushes, some good white blotting paper torn into small ])ieces, and a lens should also be included in the kit. 107 Fig. 1.— Tools used in depollination and emasculation: A, scissors useful in removing small organs; JS, self-closing forceps; C, forceps with attachment; D, scissors for severing large organs. 12 NEW METHODS OF PLANT BREEDING, DEVICES USED IN DEPOLLINATING FLOWERS. The contrivances required for the work of depolhnation by water are inexpensive and easily secured. When the method was first used attachments to the garden hose were employed, but these were found too clumsy and uncertain; moreover, the hose is not always available when it is desired to depollinate flowers in the field. A fairly good substitute is a modification of the common putty bulb (fig. 2, E). This answers the purpose very satisfactorily for large flowers of the Composit??, Leguminosfe,and other groups. The putty bulb will hold about a pint of water. By imscrewing the spout or ejector, the rubber bag may be quickly filled with water and the ejector replaced in a few seconds. The size of the jet of water is regulated by using a suitable piece of ])am- boo reed or other con- trivance fixed firmly in the tube. When the l)idb is full of wa- ter a slight pressure with the hand will cause a fairly strong Init fine stream of wa- ter to be emitted. It will l)e found that a very small jet of wa- ter is needed for very small and fragile flow- ers. For this purpose the various kinds of chip blowers (fig. 2, A and B) and water bulbs sold by dental supply companies suit the plant breeder's needs admirably. These devices are inexpensive, and when many flowers are to be treated at one time several of the chip blowers or water bulbs may be brought into use. As soon as the contents of one are exhausted and while the bull) is still ])ressed in the hand it may be put in a vessel of water; then when the others are emptied of their contents the first one will be ready to l)e used again. There is a still further choice in instruments of this nature in the shape of rul)l)er bulbs (fig. 2, (^and D) similar to those used on cameras. 1C7 Fig. 2.— Devices used in the depollination of flowers: A and B, chip blowers or water bulbs; C, water bulb with valve at bottom pro- vided with celluloid ejector; D, old rubl>er bulb with glass tube inserted; E, " putty bulb" with attachment to give a small jet of water. OBSTACLES IN EMASCULATING ELOWEES OF COMPOSITiE. 13 A piece of glass tubing can be pushed into one of these ; the other end of the tube can be brought to a fine point by heating it in a flame and breaking off the point, leaving an aperture about one-sixteenth inch in diameter. This device does satisfactory work, but the others are preferable. EM A SCUL ATION. A necessary preliminary to successful plant breeding by hybridiza- tion or cross-fertilization consists in preventing pollen of the plant used as the seed bearer, or pollen from other plants of the same variety or species, or pollen of any other closely allied variety or species from gaining access to the stigmatic surface of the flower intended to be pollinated instead of the pollen selected by the operator. With the greater number of plants, especially those having stamens of large size, emasculation is a simple matter, and consists of remov- ing the stamens before the anthers shed their })()lleii. Examples of plants having large stamens are found in the rose (see PI. XI, fig. 1), cherr}^, peach, etc. The l)opp>', carnation, and tobacco are also familiar examples in which emasculation is an exceedingly easy process, as the stamens can be removed Ix'foro dehiscing with a pair of forceps. OBSTACLES IN EMASCULATING THE FLOWERS OF COMPOSITiE. It is well known that the florets of the various genera of the com- posite flowers are so constructed that the anthers can not be removed without the florets being injured beyond recovery. The anthers in the disk florets unite and form a tid)e (PI. VIII, fig. 1, B, and PI. IX, fig. \,B) over and around the upper part of the immature pistil so that in the effort to remove tliis tube before the floret opens or the anthers dehisce the anthers are certain to be rup- tured and the pollen scattered over the stigma. The parts of the florets are so small that it is practically impossible to remove the anthers in a young stage without ruining the florets. There is also great danger of injuring the Yevj slender pistils during the attempted emasculation. Some genera have only female organs in the ray florets and both male and female in the disk florets, and in some important genera some species have sterile ray florets and depend on the disk florets alone for fertilization. (PI. V, figs. 1 and 2.) But to depend on the ray florets alone for careful crossing means in all instances slipshod work, as the disk florets must then all be removed, causing too much mutilation of the flower head. Depollination of the disk florets by the method herein described (PI. VIII, fig. 2, B) means that the work is absolutely certain to produce the desired results. Moreover, depollination of the disk florets followed by crossing is many times easier than the use of the ray florets as seed bearers, as this is 167 14 NEW METHODS OF PLANT BREEDING. necessarily followed by the removal of the disk florets to prevent self-pollination. In several genera of the Compositse it will be found that the staminal tubes are not drawn down into the tubular corolla until the second day after shedding the pollen. This happens with some of the species of Helianthus (PL IX, fig. 1,B), Gaillardia, etc. A short, sharp needle fijfed in the end of a piece of wood less than the diameter of a pencil and 4 inches in length can be veiy successfully used in opening the collar formed by the united anthers around the stigma, so that it may be depollinated before the pollen grains germinate. All of the crop of florets wliich come out in any single day can be opened after the manner described, the pollen washed from them, and the flower head bagged, pollination being delayed until the following day, when the circles of florets contiguous to those already depollin- ated will have come to maturity. These are in turn treated with the needle and depollinated by water. The remaining unopened florets may then be pulled out with a pair of forceps, the two lots of treated stigmas pollinated at the same time, and the flower head bagged. A method still easier is to dej^ollinate the two outer circles of stigmas and then remove them with the aid of a pair of forceps, which in the larger number of genera is easy of accomj^lishment. When the remainder of the florets expand, depollination is effected and the flowers are pollinated as soon as the flower head is cleared of water (PI. VII, figs. 1 and 2). In all composite flowers wliich attract insects the head should be bagged to prevent undesirable pollen from gain- ing access to the stigmas after being pollinated. It may be said that unless the method of depollination herein described is used with flowers of the Compositne, especially those of the division Liguliflorse, there is very little chance, if any, of abso- lutely certain results from cross-fertilization between different varie- ties of the same species or from hybridizing distinct species, HOW HYBRIDS AND CROSSES OF COMPOSITE FLOWERS MAY HAVE ORIGINATED IN THE PAST. It is quite possible that a long continuous shower will wash the pollen from the stigmas of a composite flower and that as soon as the sun shines these flowers are visited by insects whose bodies are occa- sionally covered with pollen obtained from florets of other varieties or species which open and shed their pollen, either before or subse- quent to the pollen-removing shower of rain. It is certain that cross- fertilization may take place in this way. Those species and their varieties having infertile ray florets can be crossed only when the disk florets, containing both male and female organs, are used as the seed bearers. (PL VII, figs. 1 and 2.) All of the work in breeding compos- 107 PREPARATION OF SEED AND POLLEN BEARING PARENTS. 15 ite flowers is rendered simple by the depollinating method. Especially is this the case with chrysanthemums, asters, dahlias (PL VI, fig. 2), marigolds, cineraria (PI. VI, fig. 1), cosmos, zinnias, lettuce, and with all the numerous genera having infertile ray florets, PREPARATION OF SEED AND POLLEN BEARING PARENTS. Too much care can not be bestowed on isolating from insects the plants from wliich pollen is to be selected for use in crossing. Tliis is a part of the plant "breeder's work wliich is apt to be neglected. PoUen is usually considered satisfactory irrespective of the condi- tions under wluch the flowers have been produced. It is not unna- turally supposed that the pollen is pure while the flowers are in the bud stage, and although this is the case in many instances, especially where large numbers of one variety or species are growing side by side, yet there is no doubt that the pollen of these plants is often interchanged, as in alfalfa and other plants of the legume family in -which the anthers dehisce in the flower bud and which are visited by the pollen-eating thrijis. Much experimental work in plant breeding b}^ crossing and hybridi- zation gives negative results through lack of care in the selection of pure pollen. It is just as necessary to protect the flower selected to supply the pollen, in order to prevent foreign pollen from being depositefl on or near the anthers either by insects or by wind, as it is to protect the seed-bearing flowers. In selecting pure pollen it is a good plan to have isolated plants growing in pots in the greenhouse, where they can be protected from insects during the flowering period by wire screens. If this is not feasible, the flowers should be bagged when the buds are nearing the opening stage, to prevent insects from depositing pollen on or near the flowers. Absolute success means careful attention to the very minutest detail. Omission of the neces- sary care in this respect has undoubtedly caused a vast amount of work to be unproductive of good results. Not only should the pollen bearer be grown under glass, but in every case where it is possible the seed-bearing parent should also be gro^\^l in this manner in order to have the work absolutely under control. Even then with the aid of wire screens only the large insects can be ke])t from the flowers as the pollen-eating thrips found in nearly all flowers is one of the greatest carriers of pollen from flower to flower. It is not unusual to find one of these minute insects with several grains of pollen attached to its body, rendering the work of evidently careful emasculation of no avail. The plants infested with these insects should be treated to a slight fumigation with hydrocyanic-acid gas and afterwards i)ro- tected with structures covered with some kind of fine white fabric to prevent the flowers being visited immediately before emasculation and until fertilization takes place. 107 16 NEW METHODS OF PLANT BREEDING. CONDITION OF THE STIGMA AT THE TIME OF POLLINATION. It has often been stated that the stigmatic surface of the pistil must be free from moisture when the pollen is applied. When the con- trary condition is present, the pollen is said to be less effective than it would be were it applied to the stigma when free from moisture deposited from the atmosphere. Be tliis as it may, it is absolutely certain that pollen is as effective when applied to stigmas wliich have been thoroughly treated with water and the moisture adhering to them partly removed with the aid of bibulous paper applied edgewise as it is wlien the stigma has not come in contact with Mater. THE APPLICATION OF WATER TO ALL FLOWERS. For the removal of pollen wliich may have been deposited upon the stigma of a flower previous to pollinating, water should be used in every case where the flower has been exposed through inadvertence or otherwise to the visits of insects. Its use in such a case, if it be applied within a certain period after the pollen is deposited, is desir- able, as it will render the operation of crossing with another flower more certain, because if pollen is present on the stigma of a (lower, especially if the pollen be of tlie same variety or species as the flower which it is desired to use as the seed bearer, its own pollen will in jnany cases take effect in preference to the pollen of the flower of a separate species or variety. Another case may be mentioned in wliich this adjunct to perfect emasculation can be used advantageously. It sometimes happens that an operator comes across an open flower of some Idnd which he may wish to cross, with the stamens already deliisced and the stigmas well covered with pollen. If the pollen has been deposited on the stigmas for only a short time,- in most flowers every grain can be removed effectively by the aid of the depollinating method herein described. In the case of lettuce the flowers have been depolHnated after the pollen had been in contact with the stigma for an hour and a half and a successful cross has followed. CROSSING ALFALFA. WORK ACCOMPLISHED IN THE PAST. For the purpose of demonstrating how the new method of preparing flowers for crossing can be applied to a wide range of subjects, alfalfa may be taken as an example. While it is true that alfalfa crosses and hybrids are effected by insects, the genus is an exceedingly difficult one to deal with by ordinary methods of emasculation, and there is probably no record of intentional crosses among the many 107 CROSSING ALFALFA. 17 forms of Medicago sativa^ or hybrids between it and other species, if we except those of Urban,'' who succeeded in making reciprocal crosses with Medicago sativa and M. falcata in 1S77. The work of Urban, however, was evidently done without taking the necessary precautions to preclude the possibility of error, as the pollination was a crude imitation of that effected by insects. We can not afford to relegate the crossing of the varieties of this increasingly important plant to insects or to risk pollen of unknown or undesirable forms on the plants we wish to cross. If we were to copy the insect method, we should, of course, get crosses, but we should remain as much in the dark concerning the parentage of the seedlings as we have been in the past. PECULIAR FEATURES OF THE WORK. None of the species and varieties of alfalfa sets seed from self- pollination if the flowers remain untripped (PI. II, fig. 1,^). The pollen in the untripped flower, being of a slightly adhesive nature, does not get an opportunity to move after being discharged from the anthers while still within the closed keel, and after the dehiscing period the surface of the stigma is protected from it by being close against the keel. This is the case with the flowers of most of the varieties, but there are forms the flowers of which often have the stig- mas completely hidden by the pollen (see PI. Ill, C), and yet even these do not set seed if the flowers remain untripped. When, however, the flower is tripped (PL II, fig. I, B), either by drawing the closed hand along the raceme or by snipping the indi- vidual flowers with the forceps, this action releases the column from its imprisonment within the keel and permits it to spring upward with a very rapid movement. In doing so the pollen grains are throw^n on the banner in large numbers and the stigma falls with force among them, causing a mass of the grains to be embedded in it (PL II, fig. 1, B). This action usually results in fertilization. When the flowers are visited by certain insects, the upper part of the sexual column falls with considerable force on the imder part of the insect. The stigma in this case is then partly pollinated with })ollen from the same flower, from other flowers on the same plant, or from flowers of other })lants of the same variety or from other varieties, or from pollen of other species previously deposited on the body of the insect. a The botanical history and nomenclature of this species have been discussed in pre- vious publications of the Bureau of Plant Industry; by C. S. Scofield in Bulletin 131, part 2, "The Botanical History and Classification of Alfalfa," and by Charles J. Brand in Bulletin 118, "Peruvian Alfalfa: A New Long-Season Variety for the Southwest." i> Urban, I . Verhandlungen des Botanischen Vereins der Provinz Brandenburg, vol, 19, p. 125. 1877. 11026— Bull. 167—10 3 18 NEW METHODS OF PLANT BREEDING. OLD METHODS OF CROSSING ALFALFA. Perhaps the method first used in crossing alfalfa, at least so far as can be ascertained, consisted in introducing a sharjiened piece of wood, resembling in shape the proboscis of a bumblebee, into the suture formed by the blades of the keel of the pollen bearer. The piece of wood was dusted over with pollen secured by allowing a sexual column to trip on it. It was pushed into the suture of the flower of the proposed seed bearer, and the resulting tripping of the column caused the stigma to come in contact with the pollen already secured on the piece of wood, the supposition being that, in some cases at least, foreign pollen is prepotent over that of the seed bearer on its own flowers. The second method — one which the writer tried several years ago — consists in emasculating the flowers in the bud stage, but as the anthers dehisce before the buds expand the operation must be performed when the buds are quite small, and the danger of bruising the flower enough to cause it to wither is great. This operation is easily performed with the aid of a binocular dissecting microscope, but even when the parts of the flower are left uninjured the method is clumsy and exceedingly uncertain. Another method which involves a considerable element of uncer- tainty consists in planting one or more plants of a known variety in the midst of a large field of another variety. These plants are used as seed bearers and the work of pollination is left to the insects. Thus it will be seen that when a cross of known ancestry is desired between two varieties of alfnlfa the chances of securing it by methods heretofore in vogue are very remote. NEW METHODS DEVISED BY THE WRITER. A few years ago, while investigating methods of crossing alfalfa and tr3ang to cross a hardy form of Peruvian alfalfa with a variety from Turkestan, it was found that there are at least three variations in the method of compelling this ])lant to capitulate readily to the wishes of the plant breeder. DEPOLLINATION BY WATER. The first method to be described requires close attention to details. The tools and other necessary material are as follows : Four pairs of forceps, three of them self-closing; a pair of scissors (see fig. 1, A); a few pins; a small chip blower such as dentists use (see fig. 2, A and B) ; a vessel of water; and some pieces of good blotting paper. The operations of dei)ollinating and applying pollen to the stigma can be performed satisfactorily with the unaided eye, but the operator should take his first lesson with the aid of a low-power binocular 167 CROSSING ALFALFA, 19 dissecting microscope. The magnification should not be higher than 8 diameters. Three or four flowers on a raceme shoidd be selected for crossing. The others may ])e cut off, although this is not necessary if they be tripped and the surj)lus pollen removed by wasliing with water. However, when all but three of the flowers on a raceme are removed, those intended for crossing can be reached moi-e easily. The age of the flowers used as seed bearers for crossing seems to make little difference, provided they do not show signs of withering. It is per- haps the safer j)lan to select those flowers near the center of the raceme just when the buds at the end of the raceme are about to expand. The flowers should not be mutilated in any way, and of course should be handled as little as possible because of their delicate nature. The first requisite consists in having pollen from the male ])arent at hand ready to be applied to the prepared stigmas. To do this most conveniently, as the flowers can not very well be manipulated with the fingers alone, it is desirable to have three pairs of forceps, one for each flower that is to be pollinated. Self-closing forceps are best, because they hold firmly the sexual column (PI. Ill, C and D) used in pollinating. Take a flower from. a raceme of the male })arent, bend down and secure the banner between the tips of the thumb and the forefinger, then ])ress wntli a pin or dissecting needle sidewise against the suture of the keel, beginning at the base and gradiuxlly drawing upward. If this o})eration is done carefully, the column will come out gently without disturbing the pollen from the anthers. When the flower has been trippetl or exploded in this way, the ter- nunal part comprising the stigma, with the masses of pollen surround- ing the empty anthers closely arranged around it, hangs toward the operator free from contact with anything. With the aid of the self- closing forceps, sever the column from the flower, laying the columns aside ready to be applied to the stigmas of the flowers of the seed bearer when the latter is depollinated. Now comes the inost critical part of the operation, and the amount of skill with which it is performed determines the success or failure of the work. It must be understood that in the flower to be used as the seed bearer the stamens dehisced while the flower was in the bud stage, perhaps a day or two previously, and the pollen lies in masses all around the soft stigma (PI. Ill, C and D), but still incompetent in that position to perform the acts of pollination and fertilization while the column is untripped. It should be the aim of the operator to trip the column in a manner that will cause a minimum disturbance of the arrangement of the pollen and prevent the terminal part of the column containing the sexual organs from springing with considerable force on tlie awaiting banner, thereby embedding a large number of its own pollen grains in the soft, pulpy stigma. 167 20 NEW METHODS OF PLANT BREEDING. With this end in view, grasp one of the flowers gently but securely between the tips of the thumb and forefinger, with the back of the keel resting against the tip of the index finger. Cut an ordinary small-sized pin or needle in two, take the pointed end between the thumb and index finger of the free hand, place the half pin or needle against the lower part of the suture of the keel, and press gently against the keel, bringing the pin or needle up to the central part or a little beyond it, increasing the pressure gradually. This will compel the sexual column to alter its position, or "trip," gradually as the gentle pressure of the pin retards or prevents its springing with force enough to disturb the arrangement of the pollen grains around the stigma. Allow the pin supporting the sexual column to come gently toward the awaiting banner. It will then be found that, with the pin resting on the banner, the fingers may relax their hold, the pressure of the column toward the banner keeping the pin in place; and owing to the position of the pin, for the time being the terminal part of the sexual column, consisting of stigma and an abundance of pollen around it, is prevented from pressing on the standard, as it is about one thirty-second to one-sixteenth of an inch above it. (PI. II, fig. 1, C, and PI. II, fig. 2.) This gives the opportunity for the removal of the pollen grains b}^ the use of a jet of water from the chip blower. The action of the water effectually depollinates the flower without causing the least injury (PI. II, fig. 1, Z>); in fact, the jet may be of sufficient force to remove even the empty anthers without injury to the stigma. (PI. Ill, E'and F.) However, the first few flowers oper- ated on by the beginner should be examined before proceeding with the pollination to ascertain if the treatment given has been sufficient to depollinate the flower thoroughly. After the jet of water has been applied there will be a considerable quantity of moisture covering the difi^erent parts of the flower, espe- cially the empty anthers and stigma. This is immediately removed by touching these organs and other parts of the flower with a piece of blotting paper applied edgewise. When this operation has been completed the exposed stigmas are pollinated in the following manner: Take one of the self-closing forceps, holding one of the previously prepared sexual columns from the flowers of the male parent, and with the stigma pointing upward push the end of the column containing stamens and freed pollen closely surrounding the stigma under the end of the column which has been depollinated, giving it a very slight circular movement to make certain that the large masses of pollen come in contact with the soft stigma of the depollinated flower. Wlion this has been satisfac- torily done, take hold of the supporting pin by the blunt end and gently withdraw it; the column then assumes its place on the banner 167 CROSSING ALFALFA. 21 with the stigma closely pressed against its surface (PI. II, fig. 1, F), and a goodly number of pollen grains are embedded in the soft stig- matic surface. When the supply of pollen is unlimited a number of flowers may be tripped over a watch glass and the pollen applied to the stigma with a small brush, the hairs of which are held together with a weak sirup of sugar and water. If the details described are carried out in a painstaking way, all that is done simply consists in substituting pollen from another variety for that which originally surrounded the stigma of the flower of the proposed seed bearer. The operation is performed in much less time than it takes to describe it, and the operator is rewarded by a fairly high percentage of successful crosses. The first time the writer tried this method with two distinct varieties of Medicago saliva more than two-thirds of the flowers worked set seeds. The second method of crossing alfalfa also involves depollination with the aid of a jet of water on the dehisced stamens, but it has now been discarded in favor of that just described. It is a simpler opera- tion, requiring less delicate manipulation, but the percentage of suc- cessful crosses by it is very low. Besides, it involves the mutilation of the floral envelope and the exposure of the pollinated stigma to the atmosphere instead of allowing it to assume its natural position on the banner of the flower after being pollinated. By this method all the flowers on a raceme may be used. First, by the aid of the scissors shown in figure l^A, cut off all of the unopened buds and the banner of each flower left on the rachis. The reason for removing the banner is that when the flower column is trippe