HYBWZMON’ 0F BAR‘LEY ,7 - r V HGRDEUM VULGARE L. EMEND-TLAM - z: ’ WiTH ITS WiLD RELATIVES . . . . x . ‘ b l . r O 7 Dissertation for the Diegreggf, PhD; ‘MlCHiGAN. STAT-E umveasm ROBERT PARKER; STE-1%: , . v 1‘" ‘ ’r—c’ .._,,,, , r...- >‘ln' .vr v -.., ”vi-rum. ,,,, p,-_.,,~ ,,’,pr‘ ,,..,,, ., ..., .7 (r r. 7 ‘tv: - y - o nap" N w " w W ya..- . "" '- v'"rr' . . u _ 'r'm~r>r~w~u- ”nova..- ~~I~ ~rl-rrr’ v"; 'r mv—r "WWW my- ,wwrowrrwn no», Wroc— "pm I - — r -r 3::rfi-a 1-,. or»... .9 , . _ , — #93". :w-J n' ' j-I-IBRARY ' M This is to certify that the . ;thesis entitled Hybridization of Barley (Hordeum Vulgare L. Emend. Lam.) with Its Wild Relatives. presented by Robért‘Parkef Steidl . . *" ‘ a -- hasbeen. accepted towards fulfillment of the requirements for Ph .0. degree in Wee 221% Major profesfl Date (3/0/7é 0-7639 e 3 _' g H' amomc IY . r”fumsasns Llnmnv muocns‘ It . ’1’” ll Ill/I /I MSU l LIBRARIES l RETURNING MATERIALS: Place in book drop to , remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. ABSTRACT HYBRIDIZATION OF BARLEY (HORDEUM VULGARE L. EMEND. LAM.) WITH ITS WILD RELATIVES BY Robert Parker Steidl Hybrids with Hordeum jubatum L. (4x) and the a. gubatum amphiploids g. jubatum X g. compressum Griseb. (6x) and Agropyron sericeum Hitchc. X g. jubatum (8x) were made to transfer to barley (fl. vulgare L. emend. Lam.) economically useful improvements from g. jubatum in these traits: winterhardiness, resistance to the cereal leaf beetle (Oulema melanopus L.), tolerance to salty soils and tolerance to wet soil conditions. To transfer these traits, attempts were made to produce an intermating pop- ulation containing the desired variation from the wild species, expressed in the genetic background of commercial barley. Recurrent selection and intermating could be practiced in this population until transfer of the desired variation was achieved. Robert Parker Steidl Chromosome elimination and cytoplasm from‘fi. lu- batum were both instrumental in blocking production of this intermating population by standard crossing methods. Cytoplasm from g. jubatum conferred gross lack of vigor and male sterility to g. vulgare derivative lines. Chrom- osome elimination greatly reduced the frequency of produc- tion of certain genomic constitution hybrids. Chromosome elimination also was shown to be the most logical explana— tion for the complete loss of wild chromosomes in second 5. vulgare backcross progenies of (HJ X HC)(6x) X VV. Thus, it appeared chromosome elimination partially to completely blocked production of certain genomic consti- tution hybrids and backcross progenies. Problems with the g. jubatum cytoplasm were circumvented by making the hybrids with E. vulgare as female. A different approach was undertaken, utilizing chromosome elimination in conjunction with callus culture and other techniques to produce the desired parental ma- terial for the intermating population. This would either substitute for or be a part of the more traditional back- cross method of producing the desired parental material. Robert Parker Steidl This approach involved several intervening steps with the eventual production of a population consisting of plants with different chromosome number and chromosome constitution. Then, hopefully, selection of the desired parental material could be made. If not, the most promis- ing material could again be run through the intervening steps of callusing, subculturing and regenerating large numbers of plants to eventually produce the desired parental material. All intervening steps were worked out so that pro- duction of large numbers of plants by callus culture of one hybrid have already been achieved. Thus, I am confi- dent that by taking advantage of chromosome elimination, techniques including tissue culture can eventually produce the desired parental material. HYBRIDIZATION OF HARLEY (HORDEUM VULGARE L. EMEND. LAM.) WITH ITS WILD RELATIVES BY Robert Parker Steidl A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Crop and Soil Sciences 1976 ACKNOWLEDGEMENTS I wish to express my sincere appreciation to my co-advisors Dr. J. E. Grafius and Dr. D. H. Smith for their encouragement and guidance throughout the course of this investigation and for their critical review of the manuscript. My deep appreciation is also extended to Dr. Carter M. Harrison for his critical review of the manuscript. My deepest thanks go to my fellow graduate stu- dents Gil Starks, Lynn Murry, Rye Ho Huang, and Tom Orton for their suggestions and collaboration on this research problem and to their major professors William Tai and Peter Carlson. A special thanks to a fellow graduate student, Barb Steidl (my wife), for her assistance in review and typing and for her understanding. ii TABLE OF CONTENTS LIST OF TABLES. . . . . . . . . . . . . . . . . LIST OF FIGURES O O O O O O O O O O O O O O O 0 INTRODUCTION 0 O O O O O O O O O O O O O O O O O LITEMTUE “VIEW 0 O O O O O O O O O O O O O 0 MATERIALS AND METHODS . . . . . . . . . . . . . I. II. III. IV. V. RESULTS I. II. III. PLANT MATERIALS. . . . . . . . . . . . METHODS OF HYBRIDIZATION . . . . . . . EMBRYO CULTURE MEDIA . . . . . . . . . EMBRYO CULTURE . . . . . . . . . . . . CHEMICAL TREATMENT OF PLANT MATERIAL . AND DISCUSSION. . . . . . . . . . . . . CROSSES MADE BY CHUNG LEE. . . . . . . SEQUENCE AND OBJECTIVES OF CROSSES AND COLCHICINE TREATMENT . . . . . . . . ICOLCHICINE TREATMENT OF PLANT MATERIAL CROSSING AND PROGENY PRODUCED. . . . . HYBRID SEED DEVELOPMENT. . . . . . . . iii Page vi 15 15 15 22 25 27 31 31 42 52 55 77 TABLE OF CONTENTS (cont'd.) Page VI. CHROMOSOME ELIMINATION AND EVIDENCE FOR THE E. VULGARE BACKCROSS OF (HJ X HC) (6x) X VV AND THE PENTAPLOID HYBRID FROM (HJ X HC)(6X) X VVVV HAVING THE SAME BASIC GENOMIC CONSTITUTION. . . . . . . 81 VII. BREEDING IMPLICATIONS OF THE HYPOTHESIS THAT THE Bl.2 HAS A PENTAPLOID CONSTITUTION. . . . . . . . . . . . . . _93 FUTURE DIRECTION O O O O O O O O O O O O O O O O O 98 LITERATURE CITED 0 O O O 0 O O O O O O O O O O O O 101 APPENDIX 0 O O O C O O O C O C O O O C O O O O O O 109 iv LIST OF TABLES Table Page 1. The designation and source of plants. . . . . 16 2. Crosses made by Chung Lee . . . . . . . . . . 32 3. Designation of all crosses attempted and mature progenies produced. . . . . . . . . 45 4. Order in which crosses were made. . . . . . . 47 5. Colchicine treatment. . . . . . . . . . . . . 48 6. Stability of hybrid progeny . . . . . . . . . ' 65 7. Starks' metaphase I counts of progeny from (HJ x HC) (6X) x W o o o o o o o o o o o 68 8. 'Starks' metaphase I counts of the pentaploid and B102 0 O O I O O O O O O O O O O I I O 69 9. Orton's preliminary embryo squash counts of (HJ X HC)(6x) X VVVV . . . . . . . . . . . 34 Appendix A. Norstog's agar medium II. . . . . . . . . . . 109 B. Wick embryo culture medium. . . . . . . . . . llO LIST OF FIGURES Figure 1. 32.2 spike, crossing technique, B2.2 seed and flat with plants. . . . . . . . . . . . 2. Embryo culture . . . . . . . . . . . . . . . . 3. Spike morphology . . . . . . . . . . . . . . . 4. Parthenocarpic fruit and other cross results . vi Page 21 24 35 6O INTRODUCTION Varietal improvement can only be achieved by utilizing useful genetic variability. This variability can be obtained from induced mutations or by introgres- sion within or between biological species. Induced muta- tions have provided only a relatively minor portion of varietal improvements. Variability within a crop's biological species is usually readily transferable. Transfer of a single gene into the desired genetic background is most easily achieved by the backcross method of plant breeding. Transfer oftquantitatively inherited improvements is much more effectively achieved by recurrent selection and in- termating. If the appropriate genetic background is par- tially fixed by one or two backcrosses, then selection can be mostly limited to the desired trait for transfer. The basic approach is the same for transfer from related species. That is, transfer of a single gene is usually most easily achieved by the backcross method while transfer of a quantitatively inherited improvement is usually most successfully achieved by selection in an intermating population containing the desired trait ex- pressed in the crop's genetic background. Plants with the desired chromosomes or translocations from the donor parent can be more easily recognized during the selection phase of cyclic intermating and fertility is generally high enough so the next cycle can be produced. The greatest difference in transfer of a quanti- tatively inherited improvement from a related Species as opposed to the same species involves problems in obtain- ing the desired intermating population. These problems include everything from greater sterility and lower re- combination in hybrid* and backcross generations to com- pletely unexpected phenomena operating in the material. Cytogenetics can be an extremely useful tool in coming to an understanding of these phenomena and in determining the best approach to deal with them. Ease of transfer from a related species is deter- mined by the closeness of species relationship and the *Throughout this thesis, the term "hybrid" refers to a species hybrid. ploidy level of the recipient species. Closely related species have greater chromosomal pairing and more similar functions of corresponding genes. Thus recombinant chrom— osomes will tend to produce changes in a smaller number of gene products and the changes will result in smaller devi- ations in enzymatic products. Slight deviations in pro- duction of enzymatic products for relatively few genes usually result in relatively minor imbalances in gametes and zygotes. Polyploids from closely related species have repeated copies (one copy per chromosome set) of eSSentially the same gene. Thus, in polyploids from Clo sely related parents chromosome changes such as addi- tion, substitution and recombination will result in Smaller changes in production of enzymatic products. Diploids are generally weakly buffered. The vast majority of genes producing essential enzymes are not re- peated throughout the genome. Thus each enzyme and its corresponding gene must perform its function at the cor- 1'th rate for the biological system. Translocation of a non<=orresponding chromosome piece,when fixed, results in esaentially four copies of some genes and no copies of other genes. This type of translocation has little chance i- m: u..- it. e, m. '1 J: ff ‘1. of survival because it results in biologically profound changes in the enzymatic rates of all enzymes involved. Transfer of a short and corresponding chromosome piece will hopefully be within the buffering capacity of the diploid species. Transfer of traits requires production of hybrids and subsequent generations with transfer at some point. Many hybrids between species of the Triticeae (Triticum, Aegilops, Hordeum, Secale, . . . ) have been produced. Generally these hybrids were relatively easily obtained. But production of specific combinations can not be effec- tively predicted, because phenomena such as chromosome elimination, ploidy level and inhibition of fertilization determine ease of production of certain hybrids. Thus certain combinations can be difficult to obtain. Dr. J. E. Grafius chose to use a g, jubatum-g. compressum amphiploid, which was generously supplied by Dr. A. B. Schooler, as the donating parent of g. jubatum traits to g. vulgare (Schooler, 33 al., 1966). §.'jubatum a perennial tetraploid species, is native to much of the North American continent. g. compressum is a perennial diploid species native to South America. ‘L‘. ~v ‘ I n A“ (I) The main objective of this work was to transfer useful levels of variability from g. jubatum for any of the following characters: winterhardiness, resistance to cereal leaf beetle, tolerance to salty soils, and toler- ance to wet soil conditions. g. jubatum has very high levels of useful improvements for all these traits. Al- though extensive efforts have been made to find good re- sistance to the cereal leaf beetle within g. vulgare, only moderate levels have been found. Useful levels of variability for salt tolerance do exist, but are only moderate when compared with the tolerance of g. jubatum. Secondary aspects of this study deal with various problems and techniques arising from attempts to transfer improvements from g. jubatmm. LITERATURE REVIEW Transfer of genes is the major objective of this study, therefore it is important that we understand the most important factors controlling ease of transfer. These factors are ploidy level of the recipient parent and closeness of species relationship. Their importance to gene transfer can be better understood by thinking of their effect on the buffering capacity of a Species (Har- lan, 1966). Closely related species generally have rea- sonable levels of pairing and correSponding genes on the chromosomes perform the same function at essentially the same rate. Thus, transfer in‘closely related species generally involves a very small translocation which per- forms essentially the same functions at the same rates. When the parents of a polyploid are closely re- lated, there are repeated copies (one copy per chromosome set) of essentially the same gene. Thus polyploids from closely related parents are much more buffered against biologically important changes in enzymatic rates. Hexaploid wheat is a beautiful example of the greater buffering capacity which polyploidy confers. Comparison of hexaploid wheat, Triticum aestivum, with its diploid progenitors T. monococcum (AA), Aegilops speltoides (BB), and A. squarrosa (DD) and its inter- mediate progenitor, tetraploid T. turgidum (AABB) pro- vides an Opportunity to compare the buffering capacity of polyploid levels of this polyploid series. All members of this series have only bivalent formation due to genetic suppression of homoeologous pairing. Suppression or removal of SBL (gene causing the greatest suppression of homoeologous pairing) allows considerable multivalent formation at the tetraploid and hexaploid levels (Riley and Kempanna, 1963). Species crosses at the diploid level when SBL is either not pres- ent or suppressed have considerable bivalent formation. Based on pairing data, these species are relatively closely related. Chromosomes of these three genomes (AA, BB, and DD) can be grouped into seven sets of three chromosome pairs. The degree to which one of the three homoeologous chromosome pairs can compensate for another pair of the homoeologous group also indicates the degree of chromosome relation. If considerable compensation occurs, then the homoeologous chromosomes have largely the same genetic content. Lilienfeld (1951) showed substitution did not occur at the diploid level between the A and B genomes. Substitution lines at the tetraploid level were difficult to obtain and there was good reason to assume many of the substitution chromosomes were actually recombinant chrom- osomes (Sears, 1969). At the hexaploid level there was enough buffering capacity that the compensation potential of all 42 pos- sible homoeologous substitution lines (e.g. 1A tetrasomic, 1D nullisomic) could be observed (Sears, 1966). Compen- sation was measured by comparing the substitution line with its correSponding nullisomic line. Compensation ranged from some restoration of vigor but no restoration of fertility, to near normal size, vigor and fertility of all six homoeologous substitution lines of group III. Although pairing configurations showed these species were fairly closely related, they diverged to tile "A: n'xo: 9" u‘ u a." 4:- OH. unz- . «av ‘uuL A.-. (I) the extent that considerable buffering was needed to allow reasonably easy chromosome substitution. A more dramatic expression of the substitution potential of hexaploid wheat was demonstrated by the nu- merous intergeneric substitution lines. Disomic substi- tutions of rye homoeologous groups 1, 2, 3, 5, and 6 chromosomes were made for part or all of their correspond- ing homoeologouschromosomes (Sears, 1968; O'Mara, 1947; Riley, 1965; Lee, etpal., 1969; Mettin, 23.31., 1973). Compensation was generally good. The commercial varie- ties ‘Saladin,’ 'Orlando,‘ 'St 2153/63,‘ 'Zorba' and 'Salzmunder' are lR/lB substitution lines (Zeller, 1973). 1R imparts high resistance to rust and moderate resis- tance to powdery mildew. Disomic substitution of Agrqpyron elongatum homo— eologous chromosome groups 1, 3, 4, 5, 6 and 7 have been produced (Sharma and Knott, 1966; Larson and Atkinson, 1972; and Dvorak and Sosulski, 1974). Compensation was very good. Disomic substitution lines with several other species have been produced. Monosomic and disomic addition or substitution lines are very useful in transferring traits because 10 they provide continued Opportunity for producing recombinant chromosomes. Most of the early transfers also utilized ir- radiation to break chromosomes. Transfers utilizing irradia- tion include leaf rust resistance from Aegilops umbellulata (Sears, 1956), stem rust resistance from Agropyron elongatum (Knott, 1961), leaf rust resistance from Agropyron elongatum \ (Sharma and Knott, 1966), leaf rust and powdery mildew resis- tance from rye (Driscoll and Jensen, 1963), leaf rust resis- tance from Agropyron intermedium Wienhues, 1966), and leaf rust resistance from rye (Mukade, at 31., 1970). All of these are well documented. Transfer has also been achieved from rye by natural translocations (Driscoll and Sears, 1963; and Mettin, 23 31., 1973). These translocations involve 4A/2R and lB/lR. Nat- ural breakage and transfer is the most likely explanation for both of these lines, because wheat and rye chromosomes do not appear to pair even with SBL suppressed (Riley and Kimber, 1966). Furthermore, most of the chromosomal tranST locations in the literature appear to be of this type. Larsen (1973) showed that of the tWenty reported chromo— somal translocations in wheat only three involved homoeolo- gous chromosomes. Thus stable interspecific addition and 11 substitution lines can be very useful in obtaining the desired transfers even without any appreciable pairing. Recently, transfer from species with some pairing affinity with wheat has been very successfully achieved by interfering with the genetic control over homoeologous pairing. Riley (1968) used genetic suppression of SBL to transfer yellow rust resistance of Aegilops comosa to wheat. Athwal and Kimber (1972) showed genetic suppres— sion of SBL could be used to transfer traits from Aegilops umbellulata. Sears (1973) used nullisomic SB plants to transfer leaf rust resistance from Agropyron elongatum chromosomes to their homoeologous wheat chromosomes. Many 3Ag/3D and 7Ag/7D transfer lines were produced. For both translocation types, male transmission was excellent for some lines. Pairing of 3Ag with 3D and of 7Ag with 7D was about 30% and a corresponding percentage of the progenies had translocations. These results are so spectacular that it would seem reasonable to conclude transfer from many other more closely related species could be just as easily achieved by utilizing similar methods. 12 Hexaploid wheat's highly buffered genetic system can handle considerable genetic manipulation. Thus trans- fer from species like rye, with essentially no pairing homology, is possible. The weak buffering of diploids is demonstrated by the failure of fairly extensive attempts to transfer traits from wheat to rye (Deorikar, l 9 6 3) . Attempts to transfer traits to H. vulgare is another good example of fairly extensive efforts to trans- fer traits to a diploid species. No properly verified examples of transfer to Ii. vulgare exist in the litera- ture. The only possible examples of transfer are from E. W (Hamilton, 5-33 §_1_., 1955) and from g. bulbosum (Schooler, 1964) . Transfer of improvements in any one of Se\reral traits to g. vulgare could result in substantial improvement in overall productivity of the crop. Many of the previous attempts to transfer improve- meI'lts to g. vulgare have been made with g. jubatum as the doIlor parent. The main objective of these previous WOt'kers was to transfer improved levels of winterhardi- he as from g. jubatum. g. jubatum has excellent winter- haII:'diness. Transfer of winterhardiness would result in 13 a large increase in overall productivity of this crop, because a winter hardy barley can utilize considerably more of the environmental resources than spring barley. Improvements in winterhardiness within _I_I_. vulgare have plateaued and little progress has been achieved in the last 10 years. (Personal communication, Dr. J. E. Grafius) The B. jubatum-Ii. vulgare cross was first made by Q11 incke (1940) . Since then several researchers—~Morrison, $ 3;. (1959), Vinogradova (1946, cited in Smith, 1951, and Price, 1968), and Kerber (unpublished)-—have success- fully repeated this cross, but none have gotten beyond the high levels of F1 hybrid sterility. Cytogenetic results of Kerber's cross were cited by Wagenaar (1960) . My crosses all have 3. jubatum or a g. jubatum a1TlE>hiploid as one of the parents. 5. jubatum, like wheat, is a member of a polyploid series (Rajhathy and Morrison, 19 61) . Relatively little work has been done on this poly- ploid series. When using polyploids either as the recip- ient or donor species, the possibility of genetic control over homoeologous pairing must be seriously considered. MI1:1:‘ry (1975) found essentially no pairing in the hybrid 14 from the cross diploid _Ii. vulgare X g. ju’bat’um. This could be caused by either genetic control over pairing, no homology, or physiological problems inherent in this hybrid. Some data exist on the pairing homology of the Species in the g. .jubatum amphiploids I used. Starks and Tai (1974) found It. compressum has some homology with g. jl—lbatum. _Ii. jubatum also has some homology with one of the genomes of Agropyron sericeum (Murry, 1975) . The on 13; species reported to have some pairing with the bio- l‘Dgical species 31. vulgare is _I_I_. bulbosum (Kasha and Sadasivaiah, 1971) . o—vo - ‘ _mv and 0 fr MATERIALS AND METHODS I . PLANT MATERIALS The stock of the Hordeum jubatum X _}_I_. compressum amphiploid was obatained from Dr. A. B. Schooler. The natural hybrids AgrOpyron trachycaulum X g. jubatum and A. sericeum X g. jubatum were collected in Alaska by Dr. J- E. Grafius with help from Dr. W. W. Mitchell and Dr. R- Taylor of the Alaska Experiment Station Staff (Table 1). II. minions OF HYBRIDIZATION Good growing conditions must be maintained. Hy- brids are much more successfully produced from vigorously QrOWing parental material. Frequent repotting and high fertility levels are very helpful in maintaining good Vigor, The most vigorous growth was obtained with day- time temperatures between 15C and 29C. Hybrid material in E- jubatum cytoplasm required warmer conditions (gen- erally between 22C and 29C) for most vigorous growth. 15 TABLE 1 . 16 The designation and source of plants. PLANT DE S IGNAT ION. SOURCE Hordeum lubatum X g. compressum amphiploid Wren trachycaulum X _11- jubatum (spontaneous hybrid) A. sericeum X g. jubatum (spontaneous hybrid) Hordeum j ubatum 5.9m vulgare (tetraploid) m vulgare (diPIOid) (HJ x HC)(6x) ATX HJ AS X HJ HJ Selfed seed from Dr. A. B. Schooler, North Dakota State U., Fargo, N.D. Clone from Dr. W. W. Mitchell, Palmer, Alaska Clone from Dr. W. W. Mitchell, Palmer, Alaska Seed from Dr. W. W. Mitchell and Dr. R. Taylor, Palmer, Alaska A. B. Schooler, North Dakota State U., Fargo, North Dakota Several varieties, bulks and populations from Dr. J. E. Grafius, Michi- gan State University, East Lansing, Michigan ‘1‘ (I) lie: I A A U ‘ u H '\_) I In I Dr (3 KW- V‘ t. ‘v 17 In some crosses temperatures above 26C were detrimental to good hybrid seed development. When using I_I_. m as the male parent, night temperatures of 15C or below were necessary to allow the spike to protrude from the boot before the pollen was shed. Procedure Used When Female Is Male Sterile 1. Healthy heads of the female parent with some open florets were selected. Florets were clipped, remov— ing 1/3 to 1/2 of the lemma and palea. Several heads were then grouped together and loosely wrapped with aluminum foil. If possible, the flag leaves were kept outside the aluminum foil; otherwise, the flag leaves were removed. When necessary, the head wrapped in aluminum foil was sup- PortEd with an iron rod. 2(a) . When _I-I. ‘12.];9252. was used as the male parent, Spikes with undehisced yellow anthers were selected. Flor— ets Were clipped, removing 1/3 of the lemma and palea, and crosses made by the approach method (Curtis and Croy, 1958) or the anthers were allowed to fully extrude and pollen was disseminated by gently tapping the spike of the male parent above the female . p a .3- CRY -:r "l bU sic: 18 2(b). When wild species were used as the male parent, spikes with freshly extruded anthers were gently tapped above the female. This required luck in guessing when anther extrusion would occur. Clipping of florets to cause anther extrusion met with little success. Abundant fresh pollen was obtained by cutting spikes with a few extruded anthers and storing with stems in water at 5C for two days. After storage, spikes were transferred to the greenhouse in a cold (5C) box. Rapid anther extru- sion occurred after removal of spikes from the box. 3. Greatest success was achieved when plants were P°1linated with massive amounts of pollen for two or three daYs- Good success in producing hybrid or backcross seed required much higher pollen loads than species selfing. migres Used When the Female \e Fertile 1. Healthy spikes, which were to be used as the female, with yellow-green anthers were selected. The older the spike, the greater the chance of selfing, but emas<21.11ation and drying out caused much more injury to y("inger spikes. Spikes with pollen shed during emascula- tlon or having several obvious selfed seeds were usually its: 2 A: C' . .‘.-.IE. ‘1- ‘n ll..v' . durin 19 discarded. This precaution and the obvious morphological differences between interspecific hybrid and selfed seed helped avoid mistaking a selfed seed for a hybrid seed during embryo transfer. In all the species crosses, selfed seed had plenty of starch deposition before endosperm collapse occurred in hybrid seed. At this time, hybrid seeds usually took on a shrunken or thinner appearance from reduced starch deposi- tion. Hybrid seed with H. vulgare as the female remained Plump with a watery sac-like endosperm with little or no StarCh deposition. 2. Each floret was gently opened by holding the tip Of the palea with the thumb and using tweezers to pull open the lemma. Anthers were then removed. Usually the bottom and top spikelets were too immature and were re- moved. In material with more than one floret per spike- let, e.g. (AS X HJ) (8x) , all florets past the second were remOVed. | 3. Crosses were made when the female stigma was featihery (Figure 1B) . Sometimes this was the same day as ema~'°'<=ulation. Florets either opened on their own or were gently opened before pollination. 20 FIGURE l.--BZ.2 spike, crossing technique, B2.2 seed and flat with plants. A. 32.2 plant with approximately 40% self-fertility B. Emasculated spike of (HJ X HC)(6x) with open receptive florets C. Crossed spikes wrapped in aluminum foil D. B2.2 seed E. Flat of colchicine treated 69-101 H oudmflm '1”- ill: Hr? III 22 III. EMBRYO CULTURE MEDIA Embryo culture should allow nOrmal embryonic growth and differentiation; then normal germination should occur. Normal development and germination are two separ- ate stages. At present, the Optimum media for normal de- velopment of the two stages is quite different. I had no particular success with transferring embryos to a second medium to Optimize development in both stages. Thus the best results were Obtained by using a wick medium system* for well differentiated embryos and Norstog's Medium II (Norstog, 1973) for poorly differentiated embryos (See Tables A and B Of the appendix for a complete listing of chemicals in these media.) Embryos near completion of de- velopment (having a well-developed scutellum) were very successfully grown on a wick medium (Figure ZD), while embryos with poor development could only be successfully grown to mature plants on Norstog's Medium II. Even Nor— stog's medium resulted in a low production of mature plants from poorly differentiated embryos. NO mature plants were produced by embryos lacking in differentia- -tion (proembryos). Both media were moderately successful *Wick.medium system is hereafter referred to as wick medium. 23 FIGURE 2.--Embryo culture. Chung Lee's material on wick medium with cotton stOpper wrapped in alum- inum foil. Chung Lee's transfer technique; 1/2 caryopsis placed on wick medium. Seedling Of [(HJ X HC)(6x) X HJ] X VVVV near death. 82.2 seedling on wick medium. HJ X VV seedling with tillers, auricles and roots on agar medium. - an: Ewan-1... __:...;- '_. . - _ . .___. “e“‘fir‘an -.“—_—.A_- "‘ is. .4... _._"r—."_- ' 24 N wusmwm It“ :4 7n 1 l () (J x 25 in the production of mature plants from embryos with in- termediate development . IV. EMBRYO CULTURE Embryos were dissected from the seeds after col- lapse of the endosperm occurred. This stage was deter- mined by noting when the caryopsis changed from green to white or yellow. The best success was achieved by waiting until this stage, because of better development of the embryo. The only exceptions were caused by dormancy and chromosome elimination for the second _H_. vulgare backcross of (HJ X HC) (6x) X W and W X (AS X HJ) (8x) , respectively. In the case of the second H. vulgare backcross, embryo dormancy occurred after normal starch deposition began (early dough stage) - I The procedure of excising the embryo and transfer- ring to media was as follows: 1. The transfer hood was turned on and allowed to run for several minutes. The transfer area was sponged down with 95% alcohol. Tweezers and dissecting needles were flame-sterilized before transfer of each embryo. 26 2. Seeds were surface sterilized by immersing for 2 minutes in a 20% Clorox solution. When possible, the lemma and palea were removed before sterilization. If not, they were removed in the first step of dissection Of the embryo. 3. The embryo was dissected out under a dissecting microscope with a tweezer in one hand and a dissecting needle in the other. The embryo is at the basal end of the seed on the lemma side. Thus the most effective method of dissection was to hold the distal half of the seed in the tweezers with the lemma side facing up and carefully slit the lateral side of the seed, starting in the middle and working toward the basal end. The embryo would usually then pOp out. If it did not, it could be easily found just below the seed coat at the basal tip Of the seed on the lemma side. The location of the embryo and its distinctive appearance make even very small em- bryos easy to find after a little practice. 4. The excised embryo was then transferred to the Culture medium. This required flaming the culture bottle IMNIth after removing the foam stopper and again just be— fore reinserting the stopper. To transfer onto wick 27 medium required careful orientation of the embryo with the scutellum facing the surface of the filter paper and the radicle oriented downward. Failure to orient the radicle correctly can be corrected later, but incorrect orienta— tion of the scutellum is fatal. On agar media, either the embryo was placed on the medium surface and oriented with the scutellum facing down or the embryo was submerged slightly and orientation ignored. V. CHEMICAL TREATMENT OF PLANT MATERIAL Colchicine Treatment Due to the high levels of hybrid sterility, use of chromosome doubling agents is common in interspecific crossing programs. For doubling purposes I chose to use nine different colchicine treatments. Treatments one and two were the method Wellensiek (1947) used and a modifica- tion of it, respectively. (1) Near single-tiller divi- sions were alternately immersed (dipped) in a .05% colchi- Cine solution during daytime and in water at night for four consecutive days; and (2) several consecutive daily sPrayings of leaves of vigorously growing clones were 28 done using a .05% colchicine solution with 4% DMSO (di- methyl sulfoxide, a surfactant). Treatments three through nine were all modifica- tions of POpe and Love's (1952) treatment. In these treatments washed tiller divisions were dipped in .5% colchicine. Plants were wounded by making a short slit in the leaf base near the meristem with a dissecting needle. These treatments were: TREATMENT TIME (3) nonwounded root dip 6 hours (4) wounded root dip 6 hours (5) nonwounded root dip 12 hours (6) wounded root dip 12 hours (7) horizontal treatment on 7 hours paper towels with roots protected from colchicine (8) wounded horizontal treatment 7 hours (9) inverted plant dip with roots 12 hours protected from drying with wet paper towels After colchicine treatment, divisions were planted in flats at the rate of approximately 200 per flat in ' lirree or four rows per flat (Figure 1E). Many treatments resulted in 50% mortality. A few hybrid seedlings were .K is also 'n'lC .o.. fen I ‘ t an. n .l K \n h \, nu - , § in tall 29 also colchicine treated using Pope and Love's method. The wick embryo culture medium was poured Off at the two leaf stage and .05% colchicine solution added to the test tube. Seedlings were treated for 8 hours, the colchicine solu- tion was carefully washed off, and the seedlings trans- ferred to soil. g-Amino Caprioic Acid (EACA) Treatment EACA is a relatively cheap, highly water—soluble animal-effective immunosuppressant. Bates, 23.. a_1_._., (1974) concluded EACA appeared to be the most effective of four immunosuppressants tested at blocking a hypothetically Similar immuno-system in plants. The immunosuppressants were injected daily into the leaf enclosing the spike Starting before meiosis and ending with emasculation. In a personal communication, Dr. Bates suggested a modifi- Cation for narrow leaf barley material. This modification, which I used, involved spraying spikes with 1 mg/ml EACA with .l% X-77 (a surfactant) from before meiosis to near trarleer time. The surfactant was used to insure adequate uptake of the immunosuppressant. C‘T C) (I) ("I (‘14 30 Gibberellic Acid Treatment Gibberellic acid (6A3) treatment of the maternal plants during seed development was started by Larter and Enns (1960). Their objective was to maintain more normal post—fertilization development of the seed tissues of hybrids. They treated tetraploid H. vulgare crossed with diploid H. vulgare by clipping the flag leaf tip and plac- ing the flag leaf in a shell vial with 100 ppm GA. Treat- ment was continuous from time of emasculation to embryo transfer. This treatment produced near-normal development Of the endosperm and embryo, while untreated seeds were shrunken and had much smaller embryos. Kruse (1967, 1969, 1973) found this method time consuming and not particularly effective. He placed one drop Of 75 ppm GA on the stigma of each floret for each of the two days directly following pollination. Using this technique, he had remarkable success in producing interspecific and intergeneric hybrids. I modified this technique, because I found it extremely tedious for large numbers of crosses. Spikes were immersed once daily in 75 ppm gibberellic acid for a few seconds the next one or two days after pollination. RESULTS AND DISCUSSION The enormous phenotypic plasticity of weeds (Har- lan, 1975) is very similar to the plasticity in the "weed" X cultivated barley crosses. Many of these plants vary from an extremely robust and large to a very small plant. Thus phenotypic criteria used in describing these plants must be sharply limited. I have, however, found these criteria useful: overall plant type and vigor in several environments, vigor in cooler environments, presence and relative size of auricles, relative size of lemmas and overall head type. I. CROSSES MADE BY CHUNG LEE Chung Lee, a former graduate student of Dr. J. E. Grafius, crossed the H. brachyantherum X H. depressum amphiploid with diploid H. vulgare in the spring of 1969. NO further work was done with this cross. In the spring of 1969 he also crossed the H. jubatum X H. compressum amphiploid with diploid H. vulgare (Table 2). His embryo 31 H. ‘11 I? l I". m. the .uw 1 ‘u. .h, .clu '12- 32 TABLE 2. Crosses made by Chung Lee. HYBRID DATE PED IGREE PARENTAGE DESIGNATION CROSS MADE DESCRIPTION (HJ X HC)(6x)X CI 12528 12-1* \ March 1969 69-101** 12-3 69-102 12-4 69-103 (HJ X HC)(6x)X CI 12518 11-2 March 1969 69-104 11-3 69-105 (HJ X HC)(6X)X CI 12518 14-3 March 1969 69-106 (HJ X HC)(6x)X Manchuria 16-1 March 1969 69-107 (HJ X HC)(6x)X Manchuria 18-1 March 1969 69-108 (HJ X HC)(6x)X CI 12518 10-1 March 1969 69-109 (HJ X HC)(6x)X CI 12518 26-1 March 1969 69-110 (HJ X HC)(6x)X Paragon 12-1/Bl Dec. 1969 69-101-Bl.2*** (abbrev.: Bl.2 *12-1 stands for a particular mature F 1 plant culture tube number designation Of each mature hybrid from that tube _ first generation of selfing lZ-l-l—ldfsecond generation Of selfing year Of crossing ** 69-10111; individual F1 from 100 series cross L—-first selfing generation *** Bl.2 is a simplified notation used by Mather and Jinks (1971) to denote a first backcross to parent two (BCIPZ) 311*. he < '1 «an. t . 'W' Us @471 \o l h u. #2. .‘l 33 culture techniques are shown in Figure 2A,B. At that time he only cut the seed in half and didn't dissect out the embryo. This method is not particularly reliable and only produces mature plants from very well-developed embryos. He was successful in backcrossing one Of these hy- brids, 69-101, to diploid H. vulgare. This clone, 69—101-Bl.2 (abbreviated Bl.2) , is morphologically and CYtologically similar to the "pentaploid" hybrid, 74-405, from (HJ X HC) (6x) X W (Figure 3E and Table 8), which leads to speculation that Bl.2 is the progeny from a normal 5: lulgare pollen grain and either an unreduced egg or a dOllbled sector Of 69-101. Some of the F1 hybrids produced selfed seed. In many cases, subsequent generations were essentially com- Pletely fertile. I grew plants of 69—101-1, 69-101-1-1, 69-3|.Ol-1--1-1, 69-102—1, 69-105-1, 69-106-1, 69—106-1-1, 69-107-1, 69-107-1-1, 69-109—1, and 69-109-1-1. .Only 69-101-1 had any morphological differences from the E- j'ubatum-fi. compressum amphiploid (Figure BB). This plant was also extremely different cytolog‘ically. 69-101-1 had a highly variable chromosome number at metaphase I of elosis and multipolar cell diViSions 1n melOSlSo Starks 34 FIGURE 3.--Spike morphology (left to right) A. H. compressum H. jubatum (HJ x HC)(6x) (HJ X HC)(6x) X HJ B. 69-101-1 69-101-B1.2 (Bl.2) Spike and tiller with auricles Of VV X (HJ X HC)(6x) hybrid 69-101 VV spike and tiller with clasping auricles Figure 3 A,B 36 FIGURE 3. Spike morphology (cont'd.) C. 69-101 69-lOl-Bl.2 (Bl.2) Selfed progeny of B2.2 B3.2 D. VV X (HJ X HC)(6x) Haploid H. vulgare spike from VV X HJ progeny Normal spike of same progeny from VV X HJ HJ X VV progeny D C 3 e r u g .1 F 38 FIGURE 3. Spike morphology (cont'd.) E. Polyhaploid from (HJ X HC)(6x) X VVVV Intermediate progeny from (HJ X HC)(6x) X VVVV Pentaploid from (HJ X HC(6x) X VVVV Bl.2 F. (HJ x HC)(6x) VVVV Spike and tiller with clasping auricles F E 3 e r u g .1 F 40 FIGURE 3. Spike morphology (cont'd.) G. Agropyron sericeum H. jubatum (AS X HJ) (AS X HJ) (8x) H. W vv X (AS x HJ)(8x) (AS X HJ) (8x) x VVVV H G 3 e r u g .1 F 42 (unpublished) did not find multipolar cell divisions in either 69-101 or Bl.2 although both were studied exten- sively. The other hybrid selfs could easily be contamin- ant selfed progenies from the H. jubatum-H. compressum amphiploid. This amphiploid and H. jubatum Often produce contaminant plants in greenhouse pots and field clumps. II. SEQUENCE AND OBJECTIVES OF CROSSES AND COLCHICINE TREATMENT This section contains a sequential description of the crossing and colchicine treatment program. At present, several of these crosses are essentially worthless for transferring traits to H. vulgare, but their production was useful in coming to a much better understanding of the problems involved in making these transfers. Problems such as lack of good pairing and hybrid sterility are expected barriers to transfer of traits in this type of material. I also found H. jubatum cytoplasm imparted gross lack of vigor and cytOplasmic sterility to H. vulgare derivatives. Chrom- osome elimination also made certain combinations difficult and relatively useless in a standard backcross intermating program . 43 Kasha (1974) listed several criteria as evidence for chromosome elimination in interspecific or intergeneric crosses: 1) genome balance in hybrids; 2) high seed set, but seeds were small and usually non-viable; 3) chromo- some instability of hybrid leading to mosaics, chimeral sectors and variable chromosome number in pollen mother cells; and 4) production of hybrids and haploids. In this section and preceding sections I have attempted to describe my results so that they might be of help in interpreting other crossing results. In Michigan, the greenhouse crossing season begins in September and lasts through June. Many of the potential parents (species and hybrids) do not produce good heads un- til November. By June, greenhouse conditions are too hot for most hybrid seeds to develop properly. Summer green- house use is mostly restricted to maintaining stocks in as good condition as possible. My crossing program began with making the second H. vulgare backcross of (HJ X HC)(6x) X VV. Our Objective was to produce parental material for an intermating population containing the desired genetic variation. Then cyclic in- termating and selection would be practiced with the hope 44 of achieving the desired recombination and ultimately a commercial variety with the desired improvements. Although this cross was made during the first two greenhouse seasons (1972-73 and 1973-74), it failed to produce the desired parental material (Tables 3 and 4). Figure 3A to 3H shows spike morphology of progeny and parents of all crosses. During January to March of the first crossing sea- _ son, clones of AT X HJ and AS X HJ were divided to tiller divisions and were colchicine treated (Table 5). The ob- jective was to produce higher level amphiploids as poten- tial donor parents to H. vulgare. Amphiploids of both hy— brids were Obtained. An extensive attempt to double (HJ X HC)(6x) X VV was also begun. The objective was to produce an amphiploid parent which, when crossed with diploid H. vulgare would produce progeny with two sets of H. vulgare chromosomes. It was hoped these hybrids would be more fertile and could produce the parental material for an intermating population. Attempts to double this cross were terminated in late 1974 after 10,000 tiller divisions were treated without success. In May of the first greenhouse crossing season, another cross-~VVVV X (HJ X HC)(6x)--was attempted with 45 movlvh woolen mow 0» Hovlvh Nam OD Homlmn How 0» Hmmlmh mom on Homlvo N.leaoalmo Am.am .>ounbac N.Hmuaoaume «we Ou HmHImh «NNH OU.HHHI¢> OHH Ou HOHImo NH Ha NH ea NH 0H Aeeoaemueoev >> ea .aa anemone sunfloofiuoucfl Amofloadmrmaomv =4-E :4 ed .dfl :fi .< .<<¢ . >.<¢ >> >>=4.¢¢ mmouoxoon oncoHs> =<.<¢> > :¢.¢¢ > =<.¢¢ >>>> x Axeciom x ems or x AxeCAom x are mm x Axeciom x are mm x >> >> x we >> x A>> x Hoaumec >> x HOHImo AerAom x ems x >> >> x Axmciom x ems ZOHBdZUHme mmmOHQmm BZ mm sues moans»: How oo>uomou one mmIHm dawns .Emmadouho ourmadb nmlsos sues oranges How couscmflmoo our moeuom concede moo muo>o mo meldo muonsdz mmouo umasowuumd menu scum Hm comm on» we «NH some mw3 mmouo oru use» or» we we «malwba 46 .l: o .5geam as x Ga: Ga x m5 Selma H 222? one :5 x 9: x >> ST: A >>.<<2m >>>> x :81; x 9: an- o Erma >25 x mm in e in E x semi? 111 o 2.525 33c 8m x 2: x E A.U.HGOOV m mqmda 47 TABLE 4. Order in which crosses were made. CROSS TIME . PROGENY Bl.2 X VV Nov. l972-June 1973 6 VVVV X (HJ X HC)(6x) May 1973 0 (HJ X HC)(6x) X HJ June 1973 12 (HJ X HC)(6x) X HC June 1973 0 Bl.2 X VV Nov. 1973—June 1974 6 (HJ x HC)(6x) x vvvv Feb.-April 1974 5 [(HJ x HC)(6x) x HJ) x vvvv Feb.-April 1974 f 0 (HJ X HC)(6x) X VV Feb.-April 1974 6 (AS X HJ)(8x) X HJ April 1974 0 (AS X HJ)(8x) X VVVV April 1974 1 HJ X VVVV Summer 1974 0 HJ X VV Summer & Fall 1974 9 (HJ X HC)(6x) X VV Summer & Fall 1974 6 VV X HJ 3 Oct. 1974 to Feb. 1975 ll VV X (HJ X HC)(6x) Feb.-April 1975 14 VV X (AS X HJ)(8x) March-May 1975 1 [VV X (HJ X HC)(6x)] X VV Aug. 1975 to March 1976 0 48 o ooo.oa same .>oznmema seesaw >> x AerAom x ems m oooa mama eoumznsumseme Lem x sac om OOOH meme rouezusumseme Lem x mac omosoomm omeamme mmmque moHoqumeza so mmmzoz oonme ezmSeamme omeamme anmmemz .usofiumouu oseownoaou . m mqmdfi 49 the hOpe that two sets of H. vulgare chromosomes and H. vulgare cytoplasm would greatly increase fertility of the hybrid and backcross progeny. It was hoped this would provide a better chance to form an intermating population containing the desired variation. Attempts to produce this cross were stopped after the first Six heads failed to Show Signs of floret closure, stigmatic change from a receptive appearance, or development of seeds. In June of the first crossing season another cross --(HJ X HC)(6x) X HJ--was successfully made. Progeny of this cross were used by Starks and Tai (1974) to better elucidate the genomic relationship of H. jubatum and H. compressum. At that time, six heads of the H. jubatum X H. compressum amphiploid were also backcrossed to H. compres- EBB for the same Objective. Although the amphiploid was massively pollinated, no signs of floret closure, stigmatic change, or seed development were observed. During February, March, and April of 1974, (HJ X HC) (6x) X VVVV, 73-304 X VVVV and (HJ X HC)(6x) X VV were at- tempted. The first two crosses were made with the hopetmat two sets of chromosomes from H. vulgare would greatly in— crease the fertility of the hybrid and backcross progenies. 50 True hybrids and progenies with partial and complete elim- ination of the chromosomes Of H. vulgare were obtained from (HJ X HC)(6x) X VVVV. Seedlings of 73-304 X VVVV died before transfer to soil, probably due to extreme chromosome elimination (Figure 2C). (HJ X HC)(6x) X VV was made to obtain seedlings for colchicine doubling treat- ment. None of the Six treated seedlings was doubled. Two other crosses were also attempted in April of 1974--(AS X HJ)(8x) X HJ and (AS X HJ)(8x) X VVVV. Four heads of the ngopyrog sericeum X H. jubatum amphiploid were pollinated with H. jubatum. Five days later, after no signs of floral closure, stigmatic change, or seed set, these heads were pollinated with VVVV. Four seeds rapidly developed, producing one (AS X HJ)(8x) X VVVV hybrid. The H. jubatum backcross was attempted to elucidate the genomic relationship of this cross. The H. vulgare hybrid was made for trait transfer purposes. By the summer of 1974, healthy H. vulgare plants were repeatedly obtained under growth chamber conditions. Therefore, I attempted a cross during the summer. H. 13- EEEET X VVVV gave a respectable seed set ranging from 0% to 90% on different Spikes, but only proembryos--globular, 51 undifferentiated embryos--resulted from this cross. Crosses with good seed set, but with a very low number of mature hybrid plants produced is good evidence for chromosome elim— ination (Kasha, 1974). Again, this cross was attempted to produce a hybrid with two sets of H. vulgare chromosomes with the hope of greater hybrid fertility. Two other crosses--HJ X VV and (HJ X HC)(6x) X VV-— were started during the summer and completed in the fall with the crossing taking place in the growth chamber. These crosses were made to determine the effects of EACA, an im- munosuppressant, which Bates, SE.21° (1974) claimed spec- tacular effects from. The HJ X VV cross was also made for breeding purposes and to help elucidate the genomic rela- tionship Of H. jubatum and H. vulgare. By this time, the deleterious effects on vigor and male fertility of the H. jubatum cytoplasm on H. vulgare derivatives were quite apparent. Thus reciprocal crosses were attempted. Between August and late December, spring habit, normal H. vulgare and genetic male sterile H. vulgare plants were used. Both failed to produce any hybrid seed. Starting in early January, hybrids were produced from winter genetic male sterile diploid barley X H. jubatum. (HJ X HC) 52 (6x) and (AS X HJ)(8x). During August and September of 1975 seventy heads of the VV X (HJ X HC)(6x) hybrid were back- crossed with VV. Only one lateral floret developed, and this had an endosperm with normal starch deposition but no embryo. III. COLCHICINE TREATMENT OF PLANT MATERIAL During January to March of 1973 clones of AT X HJ and AS X HJ were divided into tiller divisions. For colchi- cine treatment, I chose to use eight colchicine treatments and no attempt was made to keep track of which were effec- tive, because many factors such as vigor of material, par- ticular species, stage of growth, temperature and lighting determine the best colchicine treatment (Elliott, 1958). One thousand vigorous tillers of each spontaneous hybrid were divided into eight nearly equal groups. Each group received one of the treatments used. Colchicine treatments used were 1 and 3-9. After treatment, tillers were care- fully washed and planted in flats. Several treatments caused 50% mortality. 53 By June, the surviving plants were growing vigor- ously and beginning to head. The heads were cut Off and the plants transplanted to the field. Field conditions were good for rapid tiller production and vigorous growth, but these plants failed to produce large numbers of heads over the next two years in the field. Evidently the right conditions for floral initiation never occurred. These hybrids seemed to require very long days and a moderate temperature for floral initiation. For a few weeks after transplanting to the field a few plants continued to head. Several of these spikes produced a few seeds. Three amphiploids of AT X HJ were Obtained from the seed (Table 5). They all had good fertility under faVorable growing conditions. Only one of these plants has been main- tained. See Huang (1975) for cytological investigation of this amphiploid. Thirty doubled plants of AS X HJ were obtained from the seed. Some of these were completely sterile under fimmmh able growing conditions, while others had very good fertil- ity. Only those with 56 chromosomes at anaphase I and a good fertility level were maintained. Only one of these 54 was used in crossing and chromosome studies. See Murry (1975) for cytological investigation of this amphiploid. From the Spring of 1973 to the fall of 1974, all nine doubling techniques were extensively used to double the hybrid (HJ X HC)(6x) X VV. Ten thousand plants were treated with no fertile seeds produced. With a very Similar cross (HJ X VV) three Canadian research groups (personal communications from Dr. Eric R. Kerber, Steve Symko, and Dr. Bryan Harvey) also failed to produce fertile doubles. Extensive attempts were made by all three research groups. Both Bryan Harvey and Eric Kerber used all the standard colchicine doubling techniques for several years. There are many possible explanations for the fail- ure to produce fertile seeds or visibly doubled sectors in these crosses. Some of these are: failure to produce doubled cells; lack of competitive ability of doubled tis— sue; rapid reversion of doubled cells; and sterility of doubled tissue. Lindstrom (1966), working with colchicine— treated tetraploid rye, concluded that failure to produce octaploid seeds and/or octaploid meiosis was probably due to lack of competitive ability of the octaploid tissue. 55 IV. CROSSING AND PROGENY PRODUCED Crosses vary greatly in seed set, transfer survival, response to environmental stresses, etc. An understanding of how several crosses behave can be very useful in making further crosses and in interpreting the results of crosses made by other researchers. The crossing method and results of crosses including progeny produced and environmental in- fluences are herewith discussed. Second and Third Backcross Generations of (HJ X HC) (6x) X VV Extensive efforts were made to produce the second backcross to H. vulgare of (HJ X HC)(6x) X VV. This in- volved careful selection and backcrossing of 2000 healthy Bl.2 heads over two greenhouse seasons (1972-73 and 1973- 74). Good heads Of Bl.2 were seldom produced before late November and by late May, greenhouse conditions were too hot for further successful crosses. Crossing was performed on a regular basis, but success was very dependent on weather conditions. Depending on environmental conditions, from 5% to 50% of the Spikes would begin development of Small seed(s). Usually these would live for 6 to 10 days 56 and cease development at the proembryo (no differentiation) stage. In an attempt to provide a better environment for embryonic development, thirty spikes were treated with gibberellic acid by the Larter and Enns (1960) method. The treatment delayed collapse of some seeds until they were twenty—five days old, but only proembryos were pro- duced by this treatment. This lack of success caused me to abandon gibberellic acid treatment until I used Kruse's (1967) methods on crosses with H. vulgare as female parent. On untreated plants, chance and the combination of environmental factors, particularly temperature and light- ing, allowed continued development Of the embryo on rare occasions. Many of these continued growth and development until they were completely differentiated and similar to small mature embryos. Much Of the chance factor probably involved elimination of certain chromosomes to produce a diploid H. vulgare (see section on chromosome elimination). Twenty B2.2 seeds developed to the normal H. vulgare seed size (Figure 1D) and contained completely differenti- ated embryos comparable in size to those of H. vulgare. These were all transferred to wick type embryo culture 57 media (Figure 2D). Five developed abnormally on media and died. Three remained dormant and white with an appearance: of viability for at least six weeks. These embryos were left on the plant until early dough stage and dormancy seems to have been induced during starch deposition. The remaining 12 B2.2 embryos produced mature plants (Table 3). Only one of these had a relatively high. self—fertility of approximately 40% (Figure 1A). A few of the remaining plants set an occasional selfed seed. All selfed seed occurred in a hot summer greenhouse. Selfed seed and third backcross seed from all plants had no self- fertility when grown in cooler environments. Only seven normal H. vulgare bivalents were Observed in selfed andthird.backcross H. vulgare plants. These plants closely resembled Dr. Schooler's restorer lines (H. vulgare in H. jubatum cytoplasm) except, due to their lack of re- storer genes, they failed to set selfed seed in cooler en— vironments. Other than for cytoplasmic effects, these plants only exhibited H. vulgare characteristics. H. jubatum cytoplasmic effects were the most distic- tive features of the.restorer and.second.and third backcross H. vulgare plants. These plants were far less vigorous than 58 normal H. vulgare, even under ideal growing conditions. Under less than ideal growing conditions the plants often failed to head normally and died at any age from one to nine months. Under extremely unfavorable conditions, plants produced many abnormal tillers, probably due to failure of the development of main tillers. Figure 4D Shows a representative examplecfifweak and abnormal tiller production. Extreme lack of vigor made it very difficult to use the restorer lines as parents in crosses. Thus forma— tion of an intermating population among the restorer lines would have been nearly impossible. Therefore the H.'jubatum cytoplasm seems a dead end even with restoration genes present. (AS X HJ)(8x) X VVVV and (HJ X HC)(6x) X HJ These crosses are representative of the two dis- tinctly different seed set groups of my crosses. (HJ X HC) (6x) X HJ, like several other crosses had a very good seed set, ranging from 0 to 95% on individual Spikes. (AS X HJ) (8x)x VVVV had one of the highest seed sets of the low seed set group. It set a total of four seeds on four Spikes. 59 FIGURE 4.-—Parthenocarpic fruit and other cross results Large parthenocarpic fruits (normal seed size) in central florets of six row ‘H. vulgare Smaller parthenocarpic fruit in all florets. Stigmas still feathery Four hybrid seeds from VV X HJ. Other florets still with feathery stigmas 18 days after pollination Selfed progeny from 82.2 with yellowing of leaves (nicofume injury) and many weak and often deformed tillers Seed from [VV X (HJ X HC)(6x)] X VV Seedling from HJ X VV on Norstog's agar medium II 60 Figure 4 61 None of the species X species or species X amphiploid crosses fell into an intermediate category. Seed set of the high seed set group was comparable with that expected for a cross within a species, but some- what more susceptible to environmental influences. Crosses of the low seed set group are either extremely susceptible to environmental influences (like oat crosses in the green- house in Michigan) or they have fairly effective isolation mechanisms preventing fertilization. Seed Of the cross (HJ X HC)(6x) X HJ produced a reasonable amount of endosperm starch and could possibly have germinated normally without embryo culture. Thirty well-differentiated embryos were transferred to Norstog's agar medium. Only twelve of these developed into mature plants (Table 4). Probably many more mature plants would have developed on the wick medium from these completely differentiated embryos. Self fertility was 70%, which is very high for pentaploid hybrids. Morphologically, the plants were intermediate between their parents in spike characteristics and overall morphology (Figure 3A). For cytological verification and discussion of pairing, see Starks and Tai (1974). 62 Seed of the cross (AS X HJ)(8x) X VVVV also pro- duced well developed embryos with some starch deposition in the endosperm. Only one mature plant developed from these embryos. Morphologically, this plant more closely resembled its female parent, although the head was somewhat different and auricles were definitely more pronounced. For cytological verification and discussion of pairing, see Murry (1975). Other Crosses Resulting in Embryos with H. jubatum Cytoplasm The following crosses, (HJ X HC)(6x) X VV, HJ.x VV, HJ X VVVV, [(HJ X HC)(6x) X HJ] X VVVV, and (HJ X HC)(6X)X VVVV all had a relatively high seed set, which can be favor- ably compared with the seed set of (HJ X HC)(6x) X HJ. All crosses had a seed set ranging from 0 to 95% on different spikes. This group of crosses demonstrated the effects chromosome elimination had on the ease of obtaining mature progeny. It was relatively easy to obtain mature progeny from crosses of HJ X VV and (HJ X HC)(6x) X VV. The crosses produced fairly completely differentiated embryos. Seed of 63 the cross (HJ X HC)(6x) X VV had some endosperm starch depo- sition, while HJ X VV produced enough starch that mature plants were produced without embryo culture (Kerber, per- sonal communication). I also produced one mature HJ X VV hybrid without embryo culture while eight hybrids were produced using embryo culture. Twelve mature (HJ X HC) (6x) X VV hybrids were produced (Table 3). Transfer suc— ceSs was around 30% for both crosses on agar or wick media. Lack of complete embryo development probably restricted transfer success. Phenotypically, the hybrids Of both crosses more closely resembled their female parents. Hybrids of both crosses can be easily distinguished from their respective female parents on the basis of overall plant type, slightly more H. vulgare-type heads and presence Of auricles (Figure 3A—D). All the hybrid plants produced from these crosses were self-sterile. Approximately l/2 of the hybrids of each.cross were produced using EACA, an animal—effective immunosuppressant, to test the effects of this chemical. No morphological differences were noted between the treated and nontreated hybrids. All treated progenies were just as sterile. 64 Colxchicine treated seedlings and divisions of (HJ X HC)(6x) X VNJ seemed just as stubborn in producing seeds as divi- Sions Of 69-101 were. Thus in contrast to Bates, EE.El°'S (1974) claims, I was unable to Observe any effects of EACA. Metaphase I chromosome counts Of 69-101 were all 28, thus the genomic constitution AA'A"V seemed quite stable (Starks, unpublished). This agreed with Tom Orton's preliminary embryo squash data (Table 9). Kerber (unpub- lished) found AA'V metaphase I counts varied from 16 to 23, so it would appear AA'V was not quite as stable. While crosses of (HJ X HC)(6x) X VVVV, [(HJ X HC) (6x) X HJ] X VVVV and HJ X VVVV had just as high seed set, production of mature plants seemed to be strongly controlled by chromosome elimination. See Table 6 for genomic stabil- ity of this crossing series. Crosses of HJ X VVVV produced only globular undifferentiated embryos (proembryos) which were not successfully cultured on Norstog's medium. Crosses of [(HJ X HC)(6x) X HJ] X VVVV and (HJ X HC)(6x) X VVVV produced many reasonably well differentiated embryos. These embryos were essentially at the same stage of devel- opment as (HJ X HC)(6x) X VV embryos. Approximately 100 embryos of each cross were transferred to embryo culture 65 TABLE 6. Stability of hybrid progeny. HYBRID CROSS METAPHASE I PROGENY AA' A“ V (HJ )( HC)(6x) X VV Stable-- all counts 28 VAA' VV X HJ variable-- l6-22* AA' V HJ X VV l7-23** AA' A" VV (HJ )( HC)(6x) X VVVV stable 21 22-28 variable all hybrids 8 hybrids--3 of which produced some haploid tillers all hybrids 3 polyhaploids (AA'A") 1 intermediate progeny 1 pentaploid hybrid *Murry (1975) **Rajhathy and Morrison (1959) 66 media. If the stage of embryo develOpment were the only factor controlling production of mature plants these crosses should have produced many mature plants. The more differen- tiated embryos of both crosses were transferred to a wick medium. Only a small percentage of the embryos of both crosses had no development on culture media. Approximately 1/2 of the remaining embryos of each cross produced only abnormal growth ranging from undiffer- entiated growth or hypocotyl development only, to abnormal roots and shoots. The other 1/2 for a brief time developed nearly normal shoots and roots. [(HJ X HC)(6x) X HJ] X VVVV embryos only maintained near-normal growth for a short time before dying. Thus no mature plants were produced from this cross. Figure 2C is a good example of how long near-normal growth usually occurred. In contrast, embryos from the cross (HJ X HC)(6x) X VVVV produced five mature progeny: three “polyhaploids“ --haploids of the amphiploid; one "pentaploid“ hybrid; and one "intermediate" progeny--intermediate between the penta- ploid and the polyhaploid progenies in morphology and chromosome number. Figure 3E,F shows morphological 67 comparisons of the plants and Table 7 Of Gil Starks' unpub— lished data shows the cytogenetics of these plants. The polyhaploids had the expected morphology of haploids of the amphiploid, that is, more but smaller tillers with somewhat smaller floral parts. There were no H. vulgare characteristics in these plants. For example, H. vulgare and all true H. vulgare interspecific or inter- generic hybrids that I made had auricles. Like H. jubatum, H. compressum and (HJ x HC)(6x), the polyhaploids had no visible auricles. The pentaploid hybrid was intermediate in morphology between the parents. This was more obvious when the pentaploid was compared with hybrids from (HJ X HC) (6x) X VV. The pentaploid had fewer but more robust tillers with larger, often clasping auricles and a definitely more H. vulgare-like head. Plants from the cross (HJ X HC)(6x) x VV had small auricles that were rarely clasping. Like most hybrids from crosses with chromosome elimination, the pentaploid had a variable chromosome number at metaphase I Of meiosis (Table 7 of Starks' unpublished data). The intermediate progeny also had a variable chromo- some number. 68 TABLE 7. Starks' metaphase I counts Of progeny from (HJ X HC)(6x) X VVVV. CHROMOSOME NUMBER POLYHAPLOIDS INTERMEDIATE PENTAPLOID 19 l 20 21 80 22 23 24 25 26 27 28 24 29 30 31 32 33 34 35 01 o ~4 a: e; H H .s so ea a..> as as a..> to Total number 80 50 of cells (A U! 69 TABLE 8. Starks' metaphase I counts of the pentaploid and Bl.2. CHROMOSOME Bl.2 COUNT PENTAPLOID PENTAPLOID PENTAPLOID NUMBER IN 1974 1974 COUNT DEC. 1975 COUNT*' DEC. 1975 COUNT** 15 16 17 18 19 NUII—‘NN 20 21 22 23 24 25 ubQWUlNN 26 27 28 29 3O 0090\1‘0 mepe l-‘ O to 31 32 33 34 l 35 w 1‘ e.a1+d.b H H UIP‘ Total number 74 35 19 10 of cells *Count from.four or five florets from two spikes. **All counts from one floret. 70 Both the intermediate and pentaploid progenies changed morphologically since the 1974 data were taken. Partly due to lack of vigor, change in the intermediate progeny was difficult to describe, but this plant seemed more vigorous with some changes in overall plant type and head morphology. The pentaploid head morphology became impossible to distinguish from [(HJ X HC)(6x) X VV] X VV heads. Results of these crosses are good examples of the effects of chromosome elimination on a crossing series. Thus these results can be useful in interpreting the re- sults of other researchers. For example, Morrison, et 21' (1959) transferred 1400 embryos of g, murinum X VV. These were well developed embryos, but produced almost identical results to those of [(HJ X HC)(6x) X HJ] X VVVV. Earlier, this research group had produced a hybrid, which later broke down into E, vulgare and fl} murinum sectors (Hamilton, 1955). Taken as a whole, I consider this good evidence that chromosome elimination was operating in this g. murinum- ‘g. vulgare hybrid. 71 Crosses withgg vulgare as Female In May of 1973, the cross VVVV X (HJ X HC)(6x) was attempted on a limited scale. Although the six heads were massively pollinated, the florets remained open and receptive. Both lack of floral closure and lack of stigma collapse are good evidence of nonfertilization. Thus I assumed that this was good evidence this cross was improb- able. Later, when the extreme disadvantage of g. jubatum cytoplasm became all too obvious, other crosses with g. vulgare as the female were attempted. Preliminary attempts with crosses of VV X HJ and VV X (HJ X HC)(6x) gave similar results. All florets not obviously injured by excessive drying or poor emasculation remained open and appeared receptive for many days. Lack of stigma collapse was taken to indicate continued recep- tive condition (Figure 4B,C). Crossing on a large scale was then undertaken with the aid of'a spring type, genetic male sterile population and a winter type, male sterile population. These populations had the added advantage of genethzdiversity for many traits and thus possible varia- bility for receptivity to foreign pollen. 72 Crossed heads were also treated with gibberellic acid (6A3). Kruse (l967) was highly successful in using this chemical in achieving interspecific and intergeneric Hordeum hybrids. Crossing onto spring type male sterile plants was soon stopped after difficulties occurred with incomplete sterility and frequent development of parthenocarpic fruit. I use the term "parthenocarpic fruit“ for fruits appearing to have only maternal tissue development. Sometimes small parthenocarpic fruit were produced by spring type plants even without gibberellic acid treatment and without pollin- ation. Gibberellic acid treatment seemed to be necessary to stimulate parthenocarpic fruit development in the winter type male sterile plants. Partial genetic control of par- thenocarpic fruit development was indicated by the more frequent and more uniform development in the spring type population. The external morphology of parthenocarpic fruits was highly variable (Figure 4A,B) with only one character remaining constant. All develOping parthenocarpic fruits had noncollapsed stigmas. Lack of stigma collapse could be observed in all parthenocarpic fruit for at least ten days. 73 Size of parthenocarpic fruits ranged from l/3 nor- mal to normal seed size. On some heads only the central florets developed into parthenocarpic fruits. Most park thenocarpic fruits could easily be distinguished from nor- mal seed, but a few were very similar to normal seeds. Three hundred fruits were dissected. Internal morphology was less variable than external morphology. Fruits generally appeared to have a spongy proliferation of ovary wall cells, which made it very easy to distinguish parthenocarpic fruits from normal seeds. All three hundred dissected fruits had no signs of embryo or endosperm. Fifty winter type male sterile heads were massively pollinated with g. jubatum. Florets remained open and stigmas failed to collapse for many days. Only two spikes developed hybrid seeds. One spike had ten hybrid seeds, while the other had four hybrid seeds. Figure 4C shows the spike with four hybrid seeds and other florets with non-collapsed stigmas. The picture was taken eighteen days after pollination and still some florets had noncollapsed stigmas. Embryos were transferred immediately after this picture was taken. The hybrid embryos were nearly as big as g. vulgare selfs, and well—differentiated. Usually the 74 scutellum was deformed in that it curved back and partially covered the shoot end of the embryonic axis. I assumed this was a geotropic response, since all hybrid embryos of this cross were in water sac-like endosperms which appeared to allow considerable freedom of embryonic movement. The ten most normal appearing embryos were trans- ferred to a wick medium and all developed into mature plants. Eight of these were normal hybrids. Three of these hybrids sometimes produced haploid g. vulgare tillers (Figure 3D). The other two remaining plants were neither normal hybrids nor haploid g. vulgares. These plants re- mained very short for many months and failed to head. They lacked g. jubatum leaf hairs and had less distinct auricles than true hybrid progeny. Eventually these plants produced heads morphologically identical to normal hybrids. Prob- ably these plants were true hybrids but had phenotypic sup- pression of normal hybrid pubescence and auricle expression. I based this statement on some callus culture results of true hybrid tissue. Regeneration plantlets from callus formed from true hybrid tissue have the same failure of exPression of the g. jubatum pubescence and hybrid auricle tYpe. Metaphase I chromosome counts of one of these plants 75 were mostly 2l. Thus it seemed reasonable to assume this plant was a true hybrid with phenotypic suppression of some characters. The remaining four embryos were transferred to a callus promoting medium. Three embryos soon became dis- eased and died. The remaining embryo rapidly callused and was turned over to Dr. Tom B. Rice for experimental purposes. Seventy winter type male sterile heads were mas- sively pollinated with (HJ X HC)(6x). Four spikes set a total of 17 seeds. Seed sets were 9, 5, 2, and 1 seed per spike. Seeds in this cross were very similar to those of VV X HJ. In both crosses embryos were nearly normal g. vulgare size, well differentiated and had some overgrowth of the shoot tip of the embryonic axis by the scutellum. In both cases the endosperm remained a water sac with very little or no starch deposition. All seventeen embryos were transferred to a wick medium. Fourteen of these developed into normal hybrid plants (Table 3). The other embryos failed to develop into mature plants. Morphologically, these hybrids and VV X HJ hybrids were very similar to their respective reciprocal crosses. But these hybrids 76 were distinctly more vigorous than their reciprocal crosses with cytoplasm from E. jubatum. Vigor is dependent on growing conditions and thus is not a good character to distinguish reciprocal crosses grown under different micro- environmental conditions. The most consistent morphological difference between reciprocal crosses was that hybrids with cytoplasm from g. vulgare had a somewhat wider lemma. One hundred and eighty winter type male sterile heads were massively pollinated with (AS X HJ)(8x). Again, essentially all florets remained open and stigmas failed to collapse for many days. Six spikes set a single seed per spike. Four of these seeds had large, well-differen- tiated embryos in a water sac-like endosperm. Again some overgrowth of the embryonic axis by the scutellum occurred. All four of these embryos were transferred to a wick medium. Seedling development in all four cases soon became abnormal, resulting eventually in death. The remaining two embryos were transferred at a much earlier stage of development. Although embryos had a partially develOped scutellum, they were so small that proper orientation on a wick medium was left up to chance. One of these hybrids developed into a mature plant. 77 Chromosome elimination could possibly explain these results. The much lower mature plant production frequency of some crosses in which chromosome elimination operated seemed to indicate that embryos in the process of losing chromosomes had a much lower survival frequency. Early embryo transfer results in a higher percentage of hybrids (Kasha, 1974). In a similar cross, (AT X HJ)(8x) X VV, Huang (1975) found chromosome number in metaphase I of meiosis was variable. Thus possibly this cross was also unstable. Morphologically, plants of the cross VV X (AS X HJ)(8x) were similar to (AS X HJ), but they did have distinctly more g. vulgare-like auricles. This hybrid had a stable chromosome number of 35 in metaphase I of meiosis. So it was impossible to make any interpretations on whether chromosome elimination was operating in this cross. All the hybrids with g. vulgare cytoplasm were dis- tinctly more vigorous than their reciprocal crosses. Thus vigor of g. vulgare derivative lines should be good. V. HYBRID SEED DEVELOPMENT Normal seed develOpment involves normal development of many tissues. This process is set in motion by 78 pollination and fertilization (Maheshwari, 1950 and Nitsch, 1971). Initially, activity is most intense in these mater- nal tissues: the diploid integuments, nucellus and ovary wall and the haploid antipodal cells. In the cross Hordeum jubatum X Secale cereale, Cooper and Brink (1944) attributed failure of normal endosperm development to failure of the antipodal cells to develop normally. Thompson and Johnston (1945) attributed breakdown of hybrid seed development directly to irregular development of the endosperm, includ- ing formation of giant nuclei and slowing down of the mi- totic cycle. These authors seem to imply normal embryo and endo- sperm development are strongly interdependent. In actual fact, the embryo and endosperm follow independent and auton- omous developmental sequences and interaction between the two systems is not a prerequisite for normal formation of either tissue. Results of interspecific crosses support this statement. ' Interspecific crosses in Datura sometimes produce seeds with only endosperm or embryo. From the results of several Triticeae interspecific and intergeneric crosses Ivanovskaya (1962) concluded there was no relationship 79 between the degree of embryonic development and production of endosperm tissue in mature seeds. The results of my crosses also agree with the au- tonomous developmental idea. Completely differentiated embryos occurred in the crosses VV X HJ, VV X (HJ X HC)(6x), and VV X (AS X HJ)(8x) with no or only a watery sac-type endosperm develOpment. These embryos were at least as com- pletely differentiated as those of (HJ X HC)(6x) X HJ or HJ X VV. In these latter crosses enough endosperm starch deposition occurred that normal germination of the seed was possible. To complicate the matter even more, crosses between different ploidy levels of the same species can result in severe disturbances of the endosperm and embryo. Larter and Enns (1960) found hybrid embryos of tetraploid X diploid g. vulgare were so small that even with the aid of embryo culture only a very few normal seedlings could be produced. Whatever the cause of developmental breakdown, there is no indication of incompatibility between the em- bryo and the maternal tissues. In fact, results of crosses between different ploidy levels of the same species is very good evidence against incompatibility between these tissues. 80 Incompatibility due to maternal tissue also seems very un- likely, because embryos usually have more normal develOp- ment than the endosperm. It would seem any incompatibil- ity with maternal tissue would more severely affect tis- sues having a greater proportion of their genetic material from the paternal parent. Thus transfer should be normally delayed until collapse of the endosperm tissue occurs. This allows for production of more completely differentiated embryos, which when embryo cultured usually produce many more hybrid plants. 1 I observed two exceptions to this rule. Those ex- ceptions were due to these intervening factors: seed dor- mancy and chromosome elimination. Prevention of seed dor- mancy probably only requires transfer before much endOSperm starch deposition occurs. Transfer time in the case of chromosome elimination must be the best compromise between greater developmental problems with young embryos and the increasing proportion of older embryos mixed up with chromosome loss problems. In some crosses, e.g. HJ X VVVV, chromosome elimination was so severe that there was no optimum time for transfer. 81 VI. CHROMOSOME ELIMINATION AND EVIDENCE FOR THE H. VULGARE BACKCROSS OF (HJ X HC)(6x) X VV AND THE PENTAPLOID HY- BRID FROM (HJ X HC)(6x) X VVVV HAVING THE SAME BASIC GENOMIC CONSTITUTION This section deals with the gradual change in phenotype and metaphase I chromosome number of the penta- ploid hybrid from (HJ X HC)(6x) X VVVV. I did not find any similar examples in the literature on Hordeum. Ob- viously, rapid change from hybrid to a haploid of one of the parents is easily observed and easily documented for publication. 3. bulbosum-E. vulgare hybrids have become a classic example of this. These hybrids have also become a classic example of the cause of this rapid change, chromosome elimination. Chromosome elimination also causes many other conditions, which Kasha (1974) has listed as indirect evidence for chromosome elimination. These expressions of chromosome elimination in interspe- cific or intergeneric crosses are: l) evidence of genome balance, e.g. triploid hybrid VBB is essentially stable, but VB, VVBB, and VVB hybrids are unstable; 2)good seed set, but seeds non-viable without embryo transfer; 3) 82 chromosome instability of hybrids leading to mosaics, chimeral sectors or variable chromosome number in meta- phase I; and 4) production of hybrids and haploids. Obviously, some of these factors, taken individually, could be caused by other variables. Direct evidence for chromosome elimination can be obtained from embryo squash counts. For example, Subrah- manyam and Kasha (1973) made a very convincing argument for chromosome elimination in VB hybrids from this type of count. Their results were: 1) near hybrid counts for most very young embryos and 2) a gradual reduction in chromosome number of most counts over the next few days until most counts were near haploid level. Most counts began near a hybrid level, so production of a reasonable number of haploids was not due to male or female parthen- ogenesis. The gradual reduction in chromosome number of most counts shows chromosomes were lost gradually and not by whole genomes at a time. Embryo squash counts were also made on the essentially stable VBB hybrids. Much less variation in chromosome number occurred in these embryos. 83 My crosses exhibited the following expressions of chromosome elimination: (l) (2) (3) (4) VAA'A" from (HJ X HC)(6x)X VV or reciprocal was completely stable; VAA' from HJ X VV or recipro- cal was slightly unstable; AA'VV from HJ X VVVV produced only proembryos; AA'AFVV from (HJ X HC) (6x) X VVVV was somewhat unstable, exhibiting variable metaphase I counts and production of polyhaploids, intermediate progeny and a hybrid. Very good seed set, but very low production of mature plants through embryo transfer, while closely related crosses produced mature plants relatively easily with the aid of embryo culture. Variable metaphase I counts. Production of hybrids and haploids. Preliminary embryo squash results also agreed with the chromosome elimination hypothesis (Table 9). Orton found the mean chromosome number of (HJ X HC)(6x) X VVVV embryos was depressed from the expected zygotic chromosome 84 .h.mm ma >> x meVHum x one no moaunam UHo >mo cm>om Hon Hones: osOmQEOHao cooza m.mN o v o m o v m m m N o o N N H o N m .m.mN N e H o m H H H . m o H o o o o N N N m.NN H m m o N H H o N H m H m o o N v o ¢.Nm m e o o N H H H m H H H o N o o o m 0.0m H m H N N o m m H H o o o N H H N v 8.0m H v m N H o N H H H N N o o o H o m n.0m N H o m H N N v H m N N o H o N o N mm.mN H m o H H o N H N H H H m o H N o H mmmzsz (mm sane mm «H mm Nm Hm om mN mN NN 6N mN «N mN NN HN HN amps H@ mm“ mm ozomozommo u o 6H onsazHHHom magma muse zmmz EMU mmm mmuflumOSOmmU .>>>> x meVAum x way no mucooo noodvm omunfim humcfleflamum m.souuo .m mqmda 85 number of 35. This suppression was significant at the .05 level for all days except day five. Only a relatively small portion of this decrease in zygotic chromosome num- ber could be attributed to reduction in gametic chromosome number of (HJ X HC)(6x) (Starks and Tai, 1974) and E. X217 ga£g(4x) (Orton, unpublished). Thus by definition, chrom- osome elimination had occurred. Chromosome elimination had also caused considerable instability in chromosome number and the variance in chromosome number increased with the age of the embryos. Gupta and Gupta (1973) found a gradual transition in phenotype and chromosome number in a cross of tobacco similar to that expressed by the pentaploid. Nicotiana suaveolens (2n=32) was crossed with N, glutinosa (2n=24). In hybrid tissues, 5. glutinosa traits were dominant. For the first few weeks the plants maintained a E. glutinosa phenotype. Gradually, the phenotypic expression began to shift toward the E. suaveolens phenotype. Root tip and premeiotic anther squashes showed a high average frequency of aberrant mitotic anaphases (12% with bridges; 11% with fragments). Metaphase I counts showed a range of 24-29 chromosomes in a single flower. Mitotic metaphase, 86 examined in petals, also showed a similar variation rang- ing from 24-29. The ranges in mitotic and meiotic counts seemed to reflect prior mitotic abnormalities. Also, metaphase I counts seemed representative of the somatic variation of nearby tissues. By the use of mitotic kary- otype, the authors were able to show that the transitional change in phenotype was positively correlated with loss of pg. glutinosa chromosomes. Their results indicated: 1) metaphase I count variation corresponded to somatic varia- tion of nearby tissues, 2) loss in characters of one parent corresponded to loss in chromosome number, particu- larly loss in chromosomes of that parent, and 3) mitotic and meiotic counts reflected preceding mitotic abnormali- ties. It is only reasonable to assume that many of these findings also apply to Hordeum material. Comparision of the tobacco results with those available for the E. 231227 EEETEI vulgare material also shows evidence of actual loss of chromosomal material. Triploid VBB hybrids haveiabout 5% of their somatic cells with micronuclei. This compara- tively low frequency closely agrees with the small varia- tion in chromosomal number in embryo counts of this hybrid. 87 The frequency of micronuclei in shoots of the diploid VB hybrid was much higher and more variable, which was in close agreement with embryo squash counts. The metaphase I chromosome number range was large for both hybrids, but the variance for VB was considerably higher. This seems to reflect the preceding mitotic abnormalities of these hybrids. Thus it would appear metaphase I counts are representative of the somatic variation of surrounding tissues. The greater variability in micronuclei of the VB hybrids possibly indicates that once the chromosome num- ber deviated enough from the hybrid constitution more rapid loss of chromosomes occurred. This agrees with the cross results indicating g. bulbosum chromosomes exert a stabilizing effect for their own retention. This would also effectively explain the rapid change in phenotype of some VB tillers. Researchers working with the g. bulbosum-g. vulgare hybrids found little or no reciprocal differences in chrom- osome elimination. I found definite maternal effects in g. vulgare-g. jubatum crosses. Metaphase I counts of ma- terial with a hybrid phenotype from HJ X VV and VV X HJ 88 are essentially the same. Murry (1975) found a metaphase I range of 16—22 for my VV X HJ hybrids and Rajhathy and Morrison (1959) found a range of 17-23 for their HJ X VV hybrids. Thus it appeared that the same basic somatic variability was expressed in hybrid tissue of both crosses. None of my nine HJ X VV hybrids produced haploid g. vulgare tillers. Three of my eight VV X HJ hybrids produced hap- loid g. vulgare tillers. This would seem to indicate that in g. jubatum cytoplasm there was a strong lack of competi- tive ability of cells, tissues and tillers losing too many g. jubatum chromosomes. The extreme lack of vigor of g. vulgare derivatives in a g. jubatum cytoplasm strongly supports this conclusion. The pentaploid (HJ X HC)(6x) X VVVV hybrid was pro- duced in the spring of 1974. The first g. vulgare back- cross of (HJ X HC)(6x) X VV-B1.2- had been produced by Chung Lee in December of 1969. Thus, before I began my study the Bl.2 had nearly three years to stabilize pheno- typically and in chromosomal variability. This plant has remained basically stable in phenotypic expression since then. Its phenotype was as stable as any of these hybrids of the wild species. 89 By the fall of 1974, these plants could be compared cytologically and morphologically. Starks (unpublished) found metaphase I chromosome number varied from 15-34 for the Bl.2 and from 19 to 35 for the pentaploid progeny (Table 8). In 1974 the pentaploid had the expected pheno- typic characters to go with a larger culm diameter. The pentaploid had fewer but larger tillers, which were usually taller and had a larger head. The pentaploid also had an more g. vulgare-type head and the auricles were usually more pronounced. For all these characters the pentaploid expression was more like fl. vulgare. The pentaploid underwent a gradual shift in mor- phology until by December of 1975, head morphology of the pentaploid and Bl.2 were essentially identical. That is, heads removed from both plants and mixed could not be distinguished. Expression of the other characters was very similar for the two plants. These characters were much more affected by microenvironmental factors, particu- larly within pot competition factors. Clones of both the pentaploid and Bl.2 were divided and paired comparisons made in two environments. Again, other morphological traits appeared essentially identical. 90 The pentaploid expressed a somewhat greater variation in head type indicating it was not as yet completely stable phenotypically. From a mixture of heads of both penta- ploid clones, some pentaploid partial reversion heads could be separated out. The rest of the heads of the pentaploid were still impossible to distinguish from the heads of the Bl.2. In the winter of 1975, another set of metaphase I counts were made on the pentaploid by Gil Starks (Tablein. One count involving 19 cells was from four or five florets from two spikes. This count probably represented a chance event of finding two spikes with essentially the same chromosome variation. The remarkably small range also indicated that variation within a spike was small. The other count of ten cells came from one floret on another Spike. Again there was a remarkably narrow range in chromosome number. It would seem that selection for the pentaploid's 1974 phenotype could still be partially suc- cessful. Morphologically, the pentaploid hybrid had lost its more intense expression of g. vulgare characteristics until it could not be distinguished from the Bl.2. From 91 morphological observations it appeared that many of the cells had lost several E. vulgare chromosomes. The 1975 preliminary cytological observations were too biased (the sample was unrepresentative of the range and variation present in the whole clonal population of cells) to draw any conclusions. Development of considerable variability in chromosome number would seem reasonable for unstable hybrids having stable intermediate forms. For example, the tetraploid progenies of (HJ X HC)(6x) X VV with one set of g. vulgare chromosomes were stable. Loss of g. vulgare chromosomes in pentaploid progenies would tend toward this type of intermediate form. The variation present in embryo squash counts of the (HJ X HC)(6x) X VVVV cross strongly supported the idea that pentaploid progenies could develOp the chromosome variation found in the Bl.2 progeny. I can postulate only one viable hypothesis on how the pentaploid progeny could develop the same phenotype as the Bl.2. Fortunately this hypothesis also has a good explanation for the other observed results. I postulate the Bl.2 and the pentaploid are basically both of penta- ploid constitution. This requires that the Bl.2 was 92 produced from the fusion of a tetraploid egg from the tetraploid hybrid with a normal pollen grain from diploid H. vulgare. The tetraploid egg could be produced from a doubled sector or could be an unreduced egg. The extremely high sterility of the (HJ X HC)(6x) X VV hybrid supports the conclusion that the Bl.2 origin- ated from a tetraploid egg. In the cross Triticum durum X Aegilops longissima, Vardi and Zohary (1967) found all viable T. durum backcross progenies were pentaploids and thus originated from unreduced eggs. The frequency of viable progeny closely corresponded with the frequency of restitution nuclei (all chromosomes end up in one nucleus at the end of meiosis I). Restitution nuclei have not been observed in the tetraploid (HJ X HC)(6x) X VV hybrid, but this does not preclude the idea that this mechanism or some other mechanism could not on rare occasions pro- duce unreduced eggs. Chung Lee treated 69-101 with col— chicine, so a small doubled sector could also have been produced. Thus production of a Bl.2 pentaploid seems quite reasonable; in fact, it is probably the most likely type of progeny. 93 The pentaploid constitution hypothesis easily ex- plains why the B1.2 and the pentaploid have the same phe- notype now. It also fits nicely with the observed meta- phase I chromosome variation of the Bl.2. Furthermore, it offers a reasonable explanation for the origin of 69- 101-1. 69-101-1 and the intermediate progeny have con- siderable gross morphological similarities. 69-101-1, originating as a chromosome elimination product from a pentaploid zygote seems much more reasonable than from selfing of the extremely sterile (HJ X HC)(6x) X VV hybrid. In summary, observation of how the pentaploid hy- brid has slowly changed over time gives an example of how the Bl.2 probably originated. Since there is excellent evidence for the Bl.2 having a pentaploid constitution, the next section deals with the breeding implications of this hypothesis. VII. BREEDING IMPLICATIONS OF THE HYPOTHESIS THAT THE Bl.2 HAS A PENTAPLOID CONSTITUTION The diploid g. vulgare genomic constitution of the 12 second H. vulgare backcross progenies presents a 94 difficult obstacle to the production of the desired par- ental material for an intermating populatiOn. The paren- tal material should ideally be a series of plants with reasonable fertility and all expressing part of the de- sired improvements in the gentic background of the recip- ient species. Thus the plants could be essentially a diploid fl. vulgare with one or two g. jubatum chromosomes. These plants could be cyclically intermated with selection pressure to maintain the desired g. jubatum chromosomes and to fix desired translocations when they occur. There are two possible explanations for the failure to produce the desired parental material. These possible mechanisms must explain why the only viable mature progeny have a diploid g. vulgare constitution. The possible mech- anisms are: 1) All viable eggs have a haploid g. vulgare constitution and 2) Viable eggs have some wild chromosomes plus a haploid g. vulgare set, but the only progeny to sur- vie to maturity have all the wild chromosomes eliminated. Either one of these mechanisms would have profound implications on the chances of getting the desired parental material. Thus knowing which is operating could be ex- tremely important. Without knowing the genomic 95 constitution of the Bl.2, it is impossible to evaluate which is the more likely mechanism operating in the material. Since the only viable hypothesis on the genomic constitution of the Bl.2 is the pentaploid hypothesis, I have evaluated the re- sults in terms of the Bl.2 beingga pentaploid. Chromosome elimination is a logical explanation be- cause the most likely viable egg gametes from an AA'A"VV pro- geny are gametes with essentially two or more full genomic sets, for example, AV, A'V, A"V, and A'V. I base this state- ment on the genomic constitution overserved in backcross pro-- geny of other Triticeae crosses (Linienfeld, 1951; Kihara, 1963; Furuta and Tanaka, 1970; Vardi, 1973; and Vardi, 1974). Fusion of these types of viable eggs with a pollen grain from diploid g. vulgare would produce zygotes which could be com- pared with those produced in species crosses. Kasha (1974) showed the genomic constitution of the embryo was the most important factor in determining embryo chromosomal stability and the genomic constitution of the female parent had sur- prisingly little effect. AA‘A"V from (HJ X HC)(6x) X-VV was at the genomic balance point (Table 6). .VAAf from VV X HJ was slightly underbalanced (not enough wild chromosomes) and had 96 considerable chromosomal variation occurring at metaphase I (a range of 16-22; Murry, 1975). These hybrids also occasionally produced haploid g. vulgare tillers. Thus progeny of this constitution tended to lose wild chromo- somes. AA'A"VV progeny from (HJ X HC)(6x) X VVVV were slightly overbalanced and tended to lose g, vulgare chrom- osomes. AA'VV progeny from HJ X VVVV only developed to the proembryo stage, suggesting excessive problems with chromosome elimination. The most viable eggs (excluding haploid g. vulgare eggs) would be poorly balanced genomic- ally and chromosome elimination would be expected to be a strong force. At this point, it was important to determine the chance of obtaining haploid g. vulgare eggs from AA'A“VV hybrids. This required making some assumptions and using the binomial expansion. I chose to first estimate the upper limit for production of haploid g. vulgare eggs. The most favorable circumstances for production of haploid eggs would be to have perfect pairing of the g. vulgare chromosomes (a 7-7 Split at anaphase I) and no pairing of the E. jubatumHH, compressum (AA'A“) chromosomes. Thus the AA'A“ chromosomes would distribute randomly to the 97 poles. The population of egg gametes would all have a haploid g. vulgare constitution plus §.jubatum—§. £227 pressum chromosomes determined by the binomial expansion of (1/2 + l/2)21. The chance of any one egg having only a haploid g. vulgare constitution is (1/2)21, or approxi— mately 5x10-7. Since approximately 80,000 florets of the Bl.2 were pollinated, the chance of producing one diploid g. vulgare was 80,000 (5X10)'—7 or approximately .04. The chance of producing twelve diploid g. vulgare progenies from pollination of 80,000 florets is 5.5 X 10-26. This comes from the Poisson distribution, where u12/121eu is the chance for 12 such events in any one sample (Sokal and Rohlf, 1969). U can be found by solving u/eu=.04, where e=2.71828 and .04 is the chance for one event. U is approximately equal to .0417. Thus, even making these ridiculously ideal assumptions, the chance of producing 12 haploid g. vulgare eggs is so close to zero it can be ignored. Therefore, it would appear that chromosome elimination played a very important role in preventing production of the desired parental material by classical backcrossing methods in this cross. FUTURE DIRECTION This study has shawlthat material having a cyto- plasm from H. jubatum is a dead end and chromosome slime ination is a powerful force in determining numbers and type of mature progenies produced from many of the crosses. Both of these factors have a profound influence on the chances of producing the desired parental material for an intermating pOpulation. Some progress has been made with material having H. vulgare cytoplasm. Techniques includ— ing tissue culture are being utilized to take advantage of chromosome elimination in the production of the desired parental material for an intermating population to trans- fer traits to H. vulgare. Dr. Tom B. Rice has found it relatively easy to callus, subculture and regenerate large numbers of plant- lets of these hybrids. He has used colchicine treatment of callus cells of VV X HJ hybrids to produce amphiploid material of this cross. Root tips of roughly one half of the plantlets from treated callus material have the 98 99 amphiploid chromosome number. A single H, vulgare back- cross seed was produced from a mature plant from one of these plantlets. This seed appeared to have a completely normal endosperm with plenty of starch, but no embryo was present. These results indicate there are no major barriers to the production of callus cultures, treatment of callus cells) and regeneration of large numbers of plantlets. These plantlets can be very successfully grown to maturity. These are all necessary steps in the use of tissue culture to produce an amphiploid of VV X HJ and the subsequent use of tissue culture to help produce the desired parental ma- terial for an intermating population from this amphiploid. At present, I consider two different approaches to be the most promising prospects for the production of this parental material from this hypothetical, at present, (VV X HJ)(6x) amphiploid. The first approach is production of embryos of the backcross to H, vulgare and the subsequent production of large numbers of plants with different chrom- osome number from these embryos by utilizing tissue culture techniques to callus embryos and regenerate large numbers of plantlets. Chromosome elimination will provide the 100 directional driving force for the production of material with different chromosome numbers. Hopefully, many of the plants will be essentially back to a diploid H. vulgare constitution, but the direction chromosome elimination would operate in embryos of VVAA' genomic constitution can not be determined from the available cross data. The second approach involves use of callus culture of the amphiploid to produce large numbers of plantlets with different chromosome numbers.‘ Again, chromosome elimination will provide the directional driving force for the production of material with different chromosome num- bers. Hopefully, many of the plants will be essentially back to a diploid H. vulgare constitution. Since chromo- some elimination is operating in the VV X HJ hybrids in the direction of loss of H. jubatum chromosomes, it would seem quite reasonable to assume H. jubatum chromosomes would also be lost in its amphiploid. I am very hopeful that one of these approaches will eventually produce the desired parental material for an in- termating population, which will lead to the transfer of traits to H. vulgare. LITERATURE CITED LITERATURE CITED Athwal, R. S. and G. Kimber. 1972. The pairing of an alien chromosome with homoeologous chromosomes of wheat. Can. J. Genet. Cytol. 14: 325-333. Bates, L. S. Personal communication. 1 Bates, L. S., A. Campos V., R. Rodrigues R., and R. G. Anderson. 1974. Progress toward novel cereal grains. Cereal Science Today 19: 283-286. Cooper, D. C. and R. A. Brink. 1944. Collapse of the seed following the mating of Hordeum jubatum X Secale cereale. Genetics 29: 370-406. Curtis, B. C. and L. J. Croy. 1958. The approach method of making crosses in small grains. Agron. J. 50: 49-51 0 Deorikar, G. B. 1963. Rye: Secale cereale L. Indian Council of Agriculture Research, New Delhi. 152 pages. DrisColl, C. J. and N. F. Jensen. 1963. A genetic method for detecting induced intergeneric translocations. Genetics 48: 459-468. Driscoll, C. J. and E. R. Sears. 1963. The nature of a Spontaneous transfer of hairy neck from rye to wheat. -Proc. XI Int. Congr. Genet. 1: 123. Dvorak, J. and F. W. Sosulski. 1974. Effects of additions and substitutions of Agropyron elongatum chromo- somes on quantitative characters in wheat. Can. J. Genet. Cytol. 16: 627-637. Elliott, F. C. 1958. Plant Breeding and Cytogenetics. McGraw-Hill, Inc., New York. 395 pages. 101 102 Furuta, Y. and M. Tanaka. 1970. Experimental introgres- sion in natural tetraploid‘Aegdlops species. Japan. J. Genetics 45: 129-145. Grafius, J. E. Personal communication. Gupta, S. B. and P. Gupta. 1973. Selective somatic elhm- ination of Nicotiana glutinosa chromosomes in the F1 hybrids of N..suaveolens and N. glutinosa. Genetics 73: 605-612. Hamilton, D. G., S. Symko and J. W. Morrison. 1955. An anomalous cross between Hordeum leporinum and Hordeum vulgare. Can. J. of Agr. Sci. 35: 287- 293. Harlan, J. R. 1966. Plant introduction and biosystema- tics. In: Plant Breeding: A Symposium Held at Iowa State University. Ed. K. J. Frey. Iowa State University Press, Ames, Iowa. 430 pages. Harlan, J. R. 1975. Crops and Man. American Society of Agronomy. Crop Science Society of America, Madi- son, Wisconsin. 295 pages. Harvey, B. Personal communication. Huang, R. H. 1975. The genomes of Agropyron trachycaulum (Link) Malte and Hordeum jubatum L. and their hy- bridization with Hordeum vulgare L. MS thesis, Michigan State University. 62 pages. Ivanovskaya, E. V. 1962. The method of raising embryos on an artifical nutrient medium and its applica- tion to wide hybridization. In: Wide Hybridiza- tion of Plants. Ed. N. V. Tsitsin. Israel Pro- gram for Scientific Information, Jerusalem. 364 pages. Kasha, K. J. 1974. Haploids from somatic cells. In: Haploids in Higher Plants--Advances and Potential. Proceedings of the First International Symposium. K. J. Kasha-Ed. The University of Guelph. 421 pages. 103 Kasha, K. J. and R. S. Sadasivaiah. 1971. Genome rela- tionships between Hordeum vulgare L. and g. bul- bosum L. Chromosoma (Berl.) 35: 264-287. Kerber, E. R. Unpublished data. Kihara, H. 1963. InterSpeCific relationship in Triticum and Aegilops. 'Seiken Zik6. 15: 1-12. Knott, D. R. 1961. The inheritance of rust resistance. VI. The transfer of stem rust resistance from Agropyron elongatum to common wheat. Can. J. Plant Sci. 41: 109-123. Kruse, A. 1967. Intergeneric hybrids between Hordeum vulgare L. ssp. distichum (v. Pallas, 2n=l4) and Secale cereale L. (v. Petkus, 2n=l4). Royal Vet- erinary and Agricultural College Yearbook, Copen- hagen, Denmark. pages 82-92. Kruse, A. 1969. Observations on some interspecific and intergeneric hybrids in Gramineae. Abstract. Hereditas 63: 459. Kruse, A. 1973. Hordeum X Triticum hybrids. Hereditas 73: 157-161. Larsen, J. 1973. The role of chromosomal interchanges in the evolution of hexaploid wheat, Triticum aestivum. Proc. Fourth Int. Wheat Genet. Symp. pages 87-93. Larson, R. I. and T. G. Atkinson. 1972. Isolation of an Agropyron elongatum chromosome conferring resis- tance to the wheat curl mite on a Triticum- Agropyron hybrid. Abstract. Can. J. Genet. Cytol. 14: 731-732. Larter, E. N. and H. Enns.‘ 1960. The influence of gibber- ellic acid on the development of hybrid barley ovules i2_vivo. Can. J. Genet. Cytol. 2: 435-441. 104 Lee, Y. H., E. N. Larter and L. E. Evans. 1969. Homo- eologous relationship of rye chromosome VI with two homoeologous groups from wheat. Can. J. Genet. Cytol. 11: 803-809. Lilienfeld, F. A. 1951. H. Kihara: Genome-analysis in Triticum and Aegilops. X. Concluding review. Cytologia 16: 101-123. Lindstrom, J. 1966. Colchicine treatments of polyploids in the group Hordeae. Hereditas 54: 177-201. Maheshwari, P. 1950. An Introduction to the Embryology of Angiosperms. McGraw-Hill Book Co., Inc. New York. 453 pages. Mather, K. and J. L. Jinks. 1971. Biometrical Genetics. Cornell University Press, Ithaca, New York. 382 pages. Mettin, D., W. D. Blfithner and G. Schlegel. 1973. Addi- tional evidence of spontaneous 1B/1R wheat-rye substitutions and translocations. Proc. Fourth Int. Wheat Genet. Symp. pp. 179-184. Morrison, J. W., A. E. Hannah, R. Loiselle, and S. Symko. 1959. Cytogenetic studies in the genus Hordeum. Can. J. Plant Sci. 39: 375-383. Mukade, K., M. Kamio and K. Hosada. 1970. The transfer of leaf-rust resistance from rye to wheat by in- tergeneric addition and translocation. In: Mutagenesis in Relation to Ploidy Level. GAMMA Field Symposium No. 9. Natl. Inst. of Radiation Breeding, Ministry of Agriculture and Forestry, Ohmiya-machi, Japan. 93 pages. Murry, L. E. 1975. A cytogenetic investigation of X Agrohordeum pilosilemma. Ph.D. Thesis, Michigan State University. 96 pages. Nitsch, J. P. 1971. Perennation through seeds and other structures: Fruit deVelopment. In: Plant 105 Physiology Vol. VI A. F. C. Steward, Ed. Academic Press, New York. 541 pages. Norstog, K. 1973. New synthetic medium for the culture of premature barley embryos. In vitro 8: 307- 308. O'Mara, J. G. 1947. The substitution of a specific Secale cereale chromosome for a specific Triticum vulgare chromosome. Genetics 32: 99-100. Orton, T. Unpublished data. Pope, W. K. and R. M. Love. 1952. Comparative cytology of colchicine-induced amphiploids of interspecific hybrids: Agropyron trichophorum X Triticum durum, g. timopheevi, and g. macha. Hilgardia 21: 411- 426. Price, P. B. 1968. Interspecific and intergeneric crosses of barley. In: Barley: Origin, Botany, Culture, Winterhardiness, Genetics, Utilization, Pests. Agriculture Handbook No. 338, Agriculture Research Service, USDA. pp. 85-95. Quincke, F. L. 1940. Interspecific and intergeneric crosses with Hordeum. Can. J. Res. C 18: -372-373. Rajhathy, T. and J. W. Morrison. 1959. Cytogenetic studies in the genus Hordeum. IV. Hybrids of H. jubatum, H. brachyantherum, H. vulgare, and a hexaploid Hordeum Sp. Can. J. Genet. Cytol. 1: 124-132. Rajhathy, T. and J. W. Morrison, 1961. Cytogenetic studies in the genus Hordeum. V. H. jubatum and the new world Species. Can. J. Genet. Cytol. 3: 378-390. Riley, R. 1965. Cytogenetics and plant breeding. Proc. 11th Int. Congr. Genet. 3: 681-688. Riley, R., V. Chapman, and R. Johnson. 1968. Introduction of yellow rust resistance of Aegilops comosa into 106 wheat by genetically induced homoeologous recom- bination. Nature 217: 383-384. Riley, R. and C. Kempanna. 1963. The homoeologous nature of the non-homologous meiotic pairing in Triticum aestivum deficient for chromosome V (5B). Heredity 18: 287-306. Riley, R. and G. Kimber. 1966. The transfer of alien genetic variation to wheat. Rep. P1. Breed. Inst. (Cambridge), 1964-65. pp. 6-36. Schooler, A. B. 1964. Wild barley crosses Show disease resistance. North Dakota Farm Research 23: 13-15. Schooler, A. B., J. Nelson and A. Arsten. 1966. Hordeum jubatum L. crosses with Hordeum compressum ' ‘ Griseb. Crop Sci. 6: 187-190. Sears, E. R. 1956. The transfer of leaf-rust resistance from Aegilops umbellulata to wheat. Brookhaven - Symposium in Biology, No. 9, pp. 1-22. Sears, E. R. 1966. Nullisomic-tetrasomic combinations in hexaploid wheat. In: Chromosome Manipulations and Plant Genetics. R. Riley and K. R. Lewis, Eds., a supplement to Heredity 20: 29-45. Sears, E. R. 1968. Relationships of chromosomes 2A, 2B and 2D with their rye homoeologue. Third Int. Wheat Genet. Symp. pp. 53-61. Sears, E. R. 1969. Wheat cytogenetics. Ann. Rev. Genet. 3: 451-468. Sears, E. R. 1973. Agropyron-wheat transfers induced by homoeologous pairing. Proc. Fourth Int. Wheat Genet. Symp. pp 191-199. Sharma, D. and D. R. Knott. 1966. The transfer of leaf- rust resistance from Agropyron to Triticum by irradiation. Can. J. Genet. Cytol. 8: 137-143. 107 Smith, L. 1951. Cytology and genetics of barley. Bot. Rev. 17: 1-355. Sokal, R. R. and F. J. Rohlf. 1969. Biometry. W. H. Freeman and Co., San Francisco. 776 pages. Starks, G. D. Unpublished data. Starks, G. D. and W. Tai. 1974. Genome Analysis of Hor- deum jubatum and H. compressum. Can. J. Genet. Cytol. 16: 663-668. Subrahmanyam, N. C. and K. J. Kasha. 1973. Selective chromosome elimination during haploid formation in barley following interspecific hybridization. Chromosoma 42: 111-125. Symko, S. Personal communication. Thompson, W. P. and D. Johnston. 1945. The cause of in- compatability between barley and rye. Can. J. of Res. 23 C: 1-15. Vardi, A. 1974. Introgression from tetraploid durum wheat to diploid Aegilops longissima and Aegilops Speltoides. Heredity 32: 171-181. Vardi, A. and D. Zohary. 1967. Introgression in wheat via triploid hybrids. Heredity 22: 541-560. Vinogradova, N. M. 1946. Hybridization between cultivated barley and wild Species of Hordeum. Trudy Zonal. Inst. Zernavovo Khoziaistva Nechernozemnoi Polosy SSR. 13: 134-137. Wagenaar, E. B. 1960. The cytology of three hybrids in- volving Hordeum jubatum L.: The chiasma distribu- tions and the occurrence of pseudo-ring-bivalents in genetically induced asynapsis. Can. J. of Bot. 38: 69-85. Wellensiek, S. J. 1947. Methods for producing Triticales. J. Heredity 38: 167-173. 108 Wienhues, A. 1966. Transfer of rust resistance from Zeller, Agropyron to wheat by addition, substitution and translocation. Proc. Second Int. Wheat Genet. Symp. pp 328-340. F. J. 1973. lB/lR wheat-rye chromosome sub- stitutions and translocations. Proc. Fourth Int. Wheat Genet. Symp. pp 209-221. APPENDIX 109 TABLE A. Norstog's agar medium II. COMPONENT AMOUNT g/liter medium Major minerals KH2P04 0.91 KCl 0.75 MgSO4o7H20 0.74 CaC12°2H20 0.74 Trace elements and vitamins* mg/liter medium MnSO4'H20 3.0 H3303 ‘ 0.5 anO4°7H20 0.5 CoC12-6H20 0.025 , CuSO4°5H20 0.025 NaMoO4 0.025 Inositol (meso-) 50.0.‘ Thiamine-HC1 0.25 Ca-pantothenate 0.25 Pyridoxine-HC1 0.25 Iron source* mg/liter medium Sequestrene (sodium ferric 16.0 diethylenetriamine pentaacetate) Amino acids g/liter medium L-Glutamine 0.4 L-Alanine 0.05 L-Cysteine 0.02 L-Arginine 0.01 L-Leucine 0.01 L-Phenylalanine 0.01 L-Tyrosine 0.01 Other Malic acid** 1.0 Difco Purified Agar 6.0 Sucrose 34.2 *100X stock solution is useful **Dissolve malic acid in 30 ml H20 and adjust pH to 5.0 110 TABLE B. Wick embryo culture medium. MAJOR MINERALS‘ g/liter medium Ca (N03) '4H20 0.58 MgSO4'7H20* 0.33 KNO3 0 . 08 KCl 0.005 NaH2P04-HZO 0.038 MICRONUTRIENTS** mg/liter medium MnSO4 0,45 H3BO3 0.00375 KI 0.03 Glycine 3.0 Thiamine-HC1 0.1 Ca-pantothenate 2.5 CuSO4-5H20 0.025 NaMOO4 0.025 CoC12-6H20 0.25 IRON SOURCE** Sequestrene (sodium ferric diethylenetriamine pentaacetate) 16.0 SUGAR SOURCE q/liter medium Sucrose 20.0 *Dissolve separately. . **100X stock solution is useful. Adjust complete medium to pH 5.4. MICHIGAN STAT ll WI 11)) 32 mm“ 1