’ \ I 5| i W *1] WW I1 w i H I i ‘I W» l | € 1| \ (t r i H AQUA \I (D\H>~ {N-HEMTANCE OF REE-STANCE TO POWDER? .MlLDEW' EN THE CANTMEUPE VARIETY SEMSMQLE O ' Gs I n Thesis €03“ fhe ufirna a? M. 3. 1 x. - r~ oo‘u , .. .. . .. ‘ Q ~ 'V' 9.".IU'535545N h u}. i E“; [INA/H13? a Y ' :._ M! if“: ,3... , a _ . . .. ~ Ram-2m magma? s lawman: ,4 a I' {.1 as", it a" "a: "' “a: (masts LIBRARY Michigan State University n . ' t . I I - ' ‘ ‘ . ‘ . I M “1 .. v‘ J in?“ _.M- -~o— W "'—"""""'—‘-' ABMT JNHERITANCE CF REBTANCE TO mm MIIDEW IN THE CANTAIDUPE VARIET! SMOIE Powdery mildew WWD£J has been a season pest on cantaloupe in the United States since 1925. Control of the fungm has been accomplished in the southern areas where the disease is especially destructive by introduction of mildew re- sistant varieties. In northern areas occasich mildew outbreaks still may reduce late-season yields and quality as most varieties grown in these areas lack resistance. One of the sources of mildew resistance med in a program to incorporate resistance into Michigan-adapted varieties was the Florida variety Seminole. The objectives of the current stuiy were to determine the inheritance of resistance in Seminole and to determine its relationship to other known sources of resistance. Deliciom 51, a New York variety, was med as the susceptible parent. The populations med in the genetic study incluied both parents, the F1, F2, F3 and bachcrosses of the F1 to both parents. These populations were screened for resistance in a greenhome soil bench in the winter when air temperatures could be lovered as desired. The seeds were planted directly into the bench in rows three feet long With It inches betHeen seeds in the row and 1. inches between rows. The temperatm'e was held at 75°F until the grim leaves were £- inch in diameter, then lowered to 60°F for the reminder of the screening. Innoculation was rude by breath- blowing conidia over the plants at the time the temperature was lowered. Ratings of the degree of infection were made three weeks after ixmoculation using a 5-category rating scale. This greenhome seedling rating had good correlation with field screen- ing. It was found that resistance to powdery mildew in the canta- loupe variety Seminole was controlled by two gene pairs whose effects were uneqml and at least partially additive. Both genes exerted dominance with the major gene showing incomplete dominance and the minor gene exhibiting complete dominance. An F2 population of the cross Seminole 1 PM? 1.5 showed segregation of 3 dominant genes for mildew resistance. The genes for resistance in Seminole were thus different from and not linked to Pml, the gene for resistance in PM? 1.5. The F2 of Seminole X IBDA 5 was also screened to establish the relationship of Seminole resistance to that reported for IBDA 5. Since no segregates for susceptibility were found it was concluded that the genes are either identical, allelic or closely linked. Names for genes A and B are undesignated pending further study to their relationship to m2 in IBDA 5. INI-IERITANCE OF RESISTANCE T0 POWDERY MIIDEW IN THE CANTAIDUPE VARIETY SEMINOLE By Richard Roland Harwood A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree MASTER OF SCIDICE Department of Horticulture 1966 'I ,/ ACKNOWLEDGMENES The author wishes to express appreciation to Dr. D. Markarian for his warn.encouragement and guidance. His many extra efforts have added immeasurably to the value of the education and training received during the course of this study. I wish also to acknowledge the assistance and suggestions of Mr. PhillipflRowe. To my wife Betty, my sincere thanks for her support and encouragement as well as for help in preparation of the manuscript. ii TABLE OF CONTENTS INTRODUCTIDN. . . . . ..... . . LITERATURE REVIEM . . . . The Pathogen. . . . . . . . . mthhgc O O O O 0 Host Range. . . . Conditions Promting Epiphytoss Control. . . . . . . . . . . 01161316810 0 O O O O O 0 00:00.00 Genetic . . . . . . . . . . Resistance. . . . . . . . . . . . Inheritance....... I. O O O O O O O O O 0 Screening for Resistance. . . . WEEDS....... ..... Genetic Populations Used. . . . Pollination Technique . . . . . . Greenhowe Screening. . . . . . . Procedure . . . . . . Mildew Rating Scale . . . CORRELATION BETWEEN FIELD AND @133me SCREENING. . . . . . . . Field Screening . . . . . . . . Procedure.... Mildew Rating Scale . . . Greenhouse Correlation. . . . . MEYI‘IDDOFANAIXSIS.......... GENETIC HYIOTHEIS. . . ...... IDENTIFICATION OF THE GENE . . . . . DISCIBSION. . . . . . ........ CONCLIBIOIB............... LITERATIRE CITED ..... . . . . . . iii 0 I O O O O 00000 O O O O O O O O 000000000000 0 O ..... c O c c o c o c c O O O O 0 O 0 O O O o O O cccccccccccccc O O c O I O O O cccccccccccccc O O O 0 0 O O O o ccccc 0 c O O O 5' “8 IJST OF TABIIS Page Table 1. Correlation between greenhouse and field screening. . . 13 Table 2. Observed distribution of mildew ratings . . . . . . . . 15 Table 3. Observed and expected grouped percentages in each rating category 0 c c c O c c c c o O O O c c c c c c c 16 Table 4. Genetic hypothesis expressed as percentages in each ratingcategory............. ....... 21 Table 5. Chi-square test of observed and expected values . . . . 22 Table 6. Computation of the expected F2 frequencies. . . . . . . 23 Table 7. Expected F3 populational frequencies . . . . . ..... 25 Table 8. mpothetical populational genotypes. . . . . . . . . . . 26 Table 9. Computation of the expected Seminole X M 45 F2 . . . . 33 iv Figure 1. Figure 2. Figure 3. Figure A. LIST OF FIGURES Page Greenhome mildew rating scale. . . . . . . . . . . . . 11 Populational distributions: resistant parent, susceptible parent,F1,F2..................... 28 Populational distributions: backcross to the resistant parent, backcross to the susceptible parent, F3 #9, F3 #10, 11’ 1'4 0 O O O O O O O 0 O O O O O O O O O O O O O m Pupulational distributions: F3 #12, l3, 16, 18, 19, F3#15, F3#17, M45XSeminoleF2. . . . . . . . . . 32 INTRODUCTION Powdery mildew W W D.C.) has been a parasite of the cucurbits and especially gm 11191-2 in the United States since about 1925. Jagger and Scott (12) stated that in 1925-26 the incidence of mildew in the commercial cantaloupe growing areas of California reached such epiphy'totic proportions that yields and qmlity were drastically reduced. Since the 1920 's powdery mildew has been a constant threat to cantaloupe production through- out the United States and particularly in the southern and western states. In the northern states powdery mildew is not a constant threat, being dependent on weather conditions during the growing season. Under conditions favorable to its spread, however, it can be equally devastating in these northern areas. This happens in Michigan about once in every three or four years when the harvest period is shortened three weeks or more by nid-or-late-season out- breaks. LITERATURE REVIEW The W W m W D.C., an obligate parasite, is a member of the Pyrenomycetes or perithecial fungi (1). It is in the order Erysiphales , having a completely closed cleistothecium. The genus Erysiphe is characterized by having meloid cleistothecial append- ages resembling somatic hyphae. Morrison (16) reported the two rating types which upon rating gives rise to the cleistothecium bear- ing two or rarely three ascospores. This cleistothecial stage was observed in 1956 by Randall and Menzies (20), but so far no sexual stage has been found in the northern areas. This lack of a sexml stage in Michigan may prove of importance in keeping mutation and development of new races to a minimm. The pathogen has superficial rycelium except for the haustoria (1). It multiplies and spreads by means of asch spores called conidia which are produced by septation of the aerial mycelium into ovoid spores 6-7 In. in diameter. These windblowm spores land on the leaf epidermis . Upon germination the spore produces a short myceliun which terminates in a round, flat appressorium. From the base of this appressorium a penetration peg is produced which penetrates the cuticle and epidermal cell wall by mechanical pressure and enlarges into a branched haustorium which is enfolded by the mlruptured cell membrane. The aerial mcelium then proliferates -2. .3... and ultimately produces more conidia at its tips. mm The fungus has a rather narrow host range. . It is a. conmon past of the Cucurbitaceae (26) but has few other hosts (22). The most comon non-cucurbit hosts are zinnia and phlox (21). Wm mm Win Since the fungm is spread largely by conidial dissemination, the rate of spread is determined by the rapidity of production of conidia and their rate of germination. Morrison (11., 15) and Yarwood :00, 31), found that conidial germination is most often the limiting factor. Yarwood (31) first reported a diurnal cycle in spore germi- nation with the best germination during the daytime. Harrison (11.) observed that thespores were dormant for 8 days in the dark. Germi- nation was inhibited by free water, but good germination was seen at 95-100! humidity. The spores would germinate, although poorly, at 5% humidity. Temperature proved to be the most important factor with optimum germination at 59-61. and with little or no germination at 50 and 90°F. Knowledge of these factors is essential to an understanding of the suiden field epiplwtoses on melons as well as to the develoment of an efficient greenhouse screening method for resistance. mm W Control of powdery mildew on cantaloupe both in the field and the greenhouse has been the subject of many papers (4. 5, 6, 8, 13), but generally chemical control has been limited in its effectiveness to those conditiom favorable to limited mildew multiplication. Earlene The most effective control of powdery mildew has been genetic resistance in the established varieties. Following the severe outbreak of mildew in California in 1925-26, material from India was screened for resistance and a resistant selection was made. In 1931 M 50 was introduced followed by PM! 45 in 1934 (12). In 1938 a new race of E. W was reported in California which was virulent on HR 45 (10). A new source of resistance was found in Indian material, resulting in the release of cantaloupe #5, often called IBDA #5, in 191.3 (2'7) . Since that time mny resistant varieties have been released by several breeders ('7, 9, 18, 23, 25, 29) with the met recent being the varieties "Campo" and "Jacmnba" by Bohn et. a1. (2). Most of these later releases have resistance derived from EDA #5 or its same 8013309. W tanc Studies on the genetics of resistance were mde by Jagger et. a1. (11) and Whitaker and Pryor (28). They concluded that resistance of mm 1.5 to race 1 of the fungw was controlled by a single gene showing incomplete dominance. A recent study by Bohn and Whitaker (3) has shown that resistance to race 2 was governed by a single partially dominant gene with two recessive epistatic modifier genes. W No studies have been made on the type of resistance to mildew found in cantaloupe. (Screening.fazwflssiaianse Because of the unreliability of mildew epiphytcses and of the expense of growing melons in the field, a greenhouse screening technique had to be developed for screening melon populations for resistance. Pryor and Whitaker (19) first pre- sented correlations between field screening and greenhouse seedling screening. later Whitaker and Pryor (24) correlated resistance of leaves, cotyledons and stems of m male to powdery mildew. Bohn et. a1. (3) used this technique in their study of resistance to race 2 of the fungus. This work was done at the La Jblla, California experiment station which has ideal weather conditions for mildew development. The plants were grown in an unheated greenhouse in the winter. Little information was given on actual environmental conditions during these studies. MATER IALS In 1960 a cantaloupe evaluation program was started at the Michigan Agricultural Experiment Station in which varieties and breeding material from across the United States were evalmted for characteristics which would be of pose ible benefit to the Michigan cantaloupe industry. Included in this material was the Florida variety Seminole (29). It was noted that Seminole had very good resistance to powdery mildew as well as to other diseases. The powdery mildew resistance, moreover, seemed to be of a different quality from that of other material in the trials. Seminole was subsequently used as a source of resistance in a back- cross program to incorporate mildew resistance into several Michigan adapted varieties. A One parent of Seminole originated in a volunteer vine found in a Florida variety trial in 1946. Open pollinated seed was saved and grown with selection being made for several years. One of the powdery mildew resistant selections was crossed with Georgia 47, a mildew resistant line derived from a P.I. source and the variety Seminole was selected from an advanced generation of this cross. There is no known mildew resistant parentage comnon to both Seminole and the California resistant types . Seminole is completely resistant to powdery mildew in Florida, where Pm 1.5 shows no resistance. Thus it would seem unlikely that the genes for resis- ~6- -7. tance in Seminole are the same as those in either of the California varieties PM! 1.5 or IBDA 5. Breeder seed of Seminole was obtained from the Florida Agricultural Experiment Station and used for the resistant parent. Foundation seed of Deliciom 51, a highly susceptible variety, was supplied by the New York Foundation Seed Association and used as the sweeptible parent. The innoculum was transferred from field-infected plants to mature greenhouse plants and maintained as a source for the screen- ing. No attempt was made to make single spore isolations or to otherwise insure that only one race of the mildew was being used since greenhouse isolation from natural mildew infection was impossible with the available facilities. marinas Gasman Winds Lead 9mm Maia The genetic study included the parental lines, the F1, F2, F3 and backcrosses to both parents. The reciprocal of the susceptible parent I resistant parent (Ps X Pr) cross was also made and its F2 produced . Won Technician The original parental crosses were trade in the field in the summer of 1961.. The perfect flowers were emsculated during the morning of the day before they opened. The corolla was left on the flower and was med to hold it while the anthers were being removed. This 1 emitted handling of the flower without touching the ovary, which is easily bruised. After emsculation a metal -8— clitiwas used to hold the corolla closed. At this time the male flower was selected to open the next day and a metal clip used to hold it closed. Pbllinations were made at approximately 8 o'clock the following morning after the pollen began to dehisce and before the nectar started to flow. After pollination the corolla was closed with cellophane tape to prevent contamination by insect- borne pollen. The F1 selfs were made in the greenhouse in the winter of 1964-65, using essentially the same techniques but eliminating the precautions against insect contamination. The F3 and backcrosses were made in the field in the summer of 1965. firecnhguse cheenipg Procedure To screen for resistance a greenhouse soil bench 3 feet wide and 50 feet long was used. The seeds were planted in 36 inch rows across the bench, 4 inches apart with 1% inches between seeds. Every sixth row was planted to the susceptible parent (Ps) as a check for uniform infection. The soil was leached prior to planting and no nutrients were added during the screening. The temperature was held at 75°F until the primary leaves were «2: inch in diameter, and then dropped to 609F for the remainder of the screening. Innoculation was made by breath-blowing conidia from.the previously infected pdants over the seedlings each day for a week. This was begun at the time the temperature was lowered. Care was taken in watering to prevent water from touching the leaves once innoculation had taken place. The symptoms were read three weeks after innoculation. lfilébflffléiiflfihfisélé The scale (as depicted in.Fig. 1) used to rate the seedling populations in the greenhouse was as follows: W W192 06 no mildew on plant 16 mildew on the cotyledons only 26 mildew light or dissipating on the true leaves with little or no sporulation 36 colonies numerous with moderate sporu— lation AG colonies numerous and coalescing on true leaves with heavy sporulation This scale was chosen after several partially successful pre- liminary screening trials because it permitted rapid macroscopic evaluation of the pflants with a minimum of subjective error. CORREEATIDN’BETWEEN”FIEID AND GREENHOUSE:SCREENING Weenies Ezgcedgge Since the environment of the greenhouse screening was highly unnatural and the plants were held in a state of low vigor under optimum conditions for mildew grOWth, there was a question as to what was really being measured in this screening. Consequently, in the summer of 1965 several.varieties and their*F1 hybrids were rated for mildew resistance in the field. The readings were taken twice, on the let and 15th of September following severe mildew outbreaks. Figure 1.-— Greenhouse mildew rating scale. A. Category 06: No mildew on plant. B. Category 16: Mildew on the cotyledons only. C. Category 2G: Mildew light or dissipating on the true leaves with little or no sporulation. D. Category 36: Colonies numerous with moderate sporulation. E. Category LG: Colonies numerous and coalescing on true leaves with heavy sporulation. .mmnpml mmnmquwntuvumlmu: I ' 1' | ‘ azscsorug|1zscss1ugmsz I u i u l Mmduuhn miuulnwauz:u;:m.,' Milieu m finals The scale used to rate cantaloupe lines in the field was as follows: Megan: Dasssiatiga 0:? no mildew on plant 1F isolated colonies on the leaves with no more than 2 or 3 per leaf 2F colonies numerous on leaves but still separate 3F all leaves with numerous coalescing colonies 4F entire plant heavily infected, plant dying or dead Greenhouse Correlation In the fall of 1965, 25 of the varieties and F1 hybrids rated in the field were grown in the greenhouse and screened using the seedling technique. In this screening, as in the field screening, a single rating value was assigned to each variety. The results of the two screenings are tabulated in Table l. METHUD40F ANALESIS The normal method of genetic analysis involving partitioning of the components of variance relies on deriving the environmental portion from the non-segregating populations (parental and F1). In these populations all variance is assumed to be environmental. This estimate of environmental.variance is then superimposed upon each genotype in the genetic model in computing expected frequencies. The hypothetical genetic model is established by analysis of the parental, F1, backcross and F2 distributions. The F3 3950.5“ 38 co 32s. R Seance ma ass its me :88 poet! a a 5333 me 36: x 838 3&5...“ ma Sea a cassettes 838 me sin a R 83.58 me Set a 88o alt! ms 528:... s .3: a as: snowflsg 888 ups-.8 309:6 Ana Search N 5283: as 03.5 A may 333$ N «E cue—alum ease-£382 econ nos! :8 .38: .382 me Bee .8 Been: a Sea aim a on S 525 peat! 5883.. co eat me engage a 3m. 5 an a an .8 3e: 18 38 co E u 3:: 9a «Ea a £88 Egan ma ease has! a on B are 3% a Set 388.: me 588.: n 588:... co oer... ma 3c: u 838%.: 838 38 no 3.8.: :8 8a! 9a 35 n 880 as R Beans ma :88 est! x a 16335 me 36: n a 8333— ?e 5885.. a .85 a 9mm sea-.538: 588.; «e eers ma siege «as 388 335.... 580 a!!! :8 none: cur—8a 838.3; 538 m 38 3 on on on 8 .8355 3.58 Etonefioz as 288m 3858.6 so 5383 :83: Pica no sect-B8 .H .22. -14, distributions are then used as proof of the model. In this normal method the effects are measured quantitatively on a continuous scale. This permits computation of the means and variances as well as enabling population curves to be drawn. A major difficulty with the rating of severity of disease in this study was the discontinuity in the rating scale. It was not known what relationship existed between categories 10 and 26 or if the interval between them was greater or less than that between 26 and 36. There was no practical way to achieve continuity, as a metrical measurement of the amount of mildew present was impossible. This obviously precluded the computation of means or any other quantity which assumed continuity of function. Since variance computations as such could not be made there was no way to separate environmental from genetic variance. Con- sequently an empirical method of analysis which described each genotype in terms of its interaction with environment was used. Each genotype was thus described in terms of its phenotypic response. This description was made using the percentage of the population falling into each rating category. These percentages were then used to calculate the expected percentage values in each category for a population composed of several genotypes. It was necessary to group categories 0G and 16 as well as 36 and AG. The observed number of individuals in each category is listed in Table 2 and the observed and expected percentage in each of the grouped categories is listed in Table 3. Delicious 51 (Pa) Seminole (Pr) F1 (Pa 1 Pr) F2 (Pa I Pr) F3 " #9 F3 " #10 r, n #11 r, n #12 F3 " #13 F3 " #11. F3 " #15 F3 n #16 F3 " #17 F3 n #18 F3 " #19 Bcépemxprrr) x Pr Bcépzoxpspr) 1 Ps PM! 45 F103”! 45 X Pr) F2 (ms 45 x Pr) IBDA 5 F1 (IBM 5 x Pr) F2 (ram 5 x Pr) 145 1171 119 32 101 '75 50 65 68 104 9'7 257 20 46 as 18 1.5 322 151. 20 46 190 18 45 22 18/. 15 15 11 18 31 1'7 21 10 11 10 18 -16. Table 3. Frequency Distribution of Resistance Rat 7‘- " ings With Classes 1.'.. Resistance Rating 0-1 2 3.4 Population observed expected observed expected observed expected 1361106117113) 2.0 2.0 98.0 98.0 Seminole (Pr) 86.0 88.0 14.0 12.0 31 (Pa 1 Pr) 45.0 55.0 55.0 39.0 6.0 F2 (PB 1 Pr) 41.5 40.6 33.1 35.6 25.1 23.8 F3 " #9 39.3 40.6 43.1 35.6 17.3 23.8 F3 " #10 35.2 35.0 57.9 58.0 6.7 7.0 F3 " #11 31.2 35.0 50.0 58.0 18.7 7.0 F3 I " #12 14.0 13.8 52.0 46.0 35.0 40.3 F3 " m 10.6 13.8 37.3 46.0 52.0 40.3 F3 " #14 50.0 35.0 50.0 58.0 7.0 F3 " #15 69.1 74.8 21.5 21.5 9.4 3.8 F3 " #16 21.9 13.8 41.1 46.0 36.7 40.3 F3 " #17 54.7 49.5 33.6 31.5 11.4 19.0 F3 " #18 17.7 13.8 50.0 46.0 32.2 40.3 F3 " #19 7.1 13.8 50.5 46.0 42.2 40.3 F3 #10:n,14 38.8 35.0 52.6 58.0 8.5 7.0 F3 13.16.1833 14.3 13.8 46.2 1.6.0 39.6 40.3 BC to Pr 67.96 71.5 28.0 25.5 3.0 3.0 so to P6 6.4 16.3 45.1 35.8 48.3 48.0 PNR L5 100 100 F ins 45 x Pr) 100 100 F £18545 1 Pr) 85.5 85.1 6.1 8.9 8.2 5.9 .17.. Because of this unmual method of analysis the normal method of establishing a genetic tapothesis by using the distributions of the parents, backcross and F2 populations was of limited mefulness. 6mm HYPOTFEB The populations used to develop the genetic twpothesis were the parental populations, F1, backcrosses and several F3 populations. An anabsis of the percentages in these populations showed tint the genetic model which best fitted the distributions was one of We gene pairs, with one pair accounting for most of the resistance and the second contributing only minor resistance. Both genes were dominant with the major gene having incomplete dominance and the minor gene complete dominance. The effects of the 12:0 were additive. These gene pairs will temporarily be designated as AA for the mjor gene pair and BB for the minor. The basis for this Impothesis was as follows: If categories (m, 16, 26) and (36, 48) are grouped, one observes a 3:1 ratio in the 1'2, 8. 1:1 ratio in the bachcross to the susceptible parent CBCe), andnearlyallplantsfallintoOG,lGorminthebachcrosstothe resistant parent (801-). This would lead to the conclmion that the inheritance was mnogenic. With this mode of inheritance, one would expect the F3 populations to fall into one of three classes correspond- ingtothatofthePr (AA).F2 (1112218: laa)orfh (aa)sinceeach F3 population came from the self of a single P2 plant. The fact that the F3 populations did not, for the most part, fall into these categories disproved the plausibility of the monogenio model. Since no distributions were seen in the 11 F3 populations grown which class]? fit the parental distributions (Table 3), apparently more than a single gene is involved. -18- Assuming that two genes were involved, the F3 #10 could be fitted to a model. From the raw data (Table 2) it is seen that #10 had few individuals falling into category 00 and none in 40. It could now be assumed that there were few, if any, parental genotypes represented in this population. The only F2 that would give the (#10 F3 distribution would be a genotype of AAbb or aaBB. Since the distribution more closely resembled the Pr in terms of total re- sistance, it was believed that the homozygous gene which was present in #10 was accounting for a large portion of the resistance that was found in Pr. It therefore seemed evident that the effects of the two gene pairs were unequal. At this point gene A was arbri- trarily assigned as the "major" gene and gene B as the ”minor“ gene. AAbb was then designated as the genotype of F3 #10, and the observed percentages in each rating category of #10 were the result of the expression of genotype AAbb. Likewise, the observed percentages in each category of the parental populations were taken as the expression of genotypes AABB and aabb, and the observed F1 as AaBb (Table 3). If the phenotypic distributions of two genotypes are known, that of a third genotype can be obtained from.a populational distri- bution containing all three. The percentage of a population of a known genotype falling into each category is multiplied by the frequency of occurrence of that genotype in the composite to obtain the contribution of that genotype to the composite. If the con- tribution of two genotypes is subtracted from a composite of three, the remainder is the expected frequency distribution of the third genotype. -l9- Given the express ion of the genotypes AAbb and aabb, the distribution of an F3 from an F2 of genotype Aabb could be de- termined. The expected F3 ratio would be lAAbb : 2Aabb , laabb. Since f of the population was of genotype AAbb, construction of the expected F3 was begun by adding & of the percentages in each category of F3 #10 to at of the percentages of the Pa. When i of the values for Aabb was added to the total of these two, the expected distribution for the F3 was derived. Thus the contribution of AAbb and aabb was: 06 - 16 26 36 - AG AAbb 35.28/4 3 8.8% 57.9%/4 2‘ 14.5% 6.7%/I. : 1.7% aabb 0 2%/4 : 0.5% 98 %/4 : 24.5% Total 8.8% 15$ 26.21 Now according to our model, the genotype Aabb should be somewhat more susceptible than the F1, AaBb. It could be expected to center around the 20 rating. When the above totals were subtracted from the observed per- centages of F3 #12 which had a distribution approximting that which would be expected from Aabb, the results were: 00 - 10 2G 30 - AG 14.0% - 8.8% .2 5.2% 5% - 15% z 37% 3571'» - 26.2% x 8.8% These values, seeming that F3 #12 was from an F2 of genotype Aabb, were then 50% of the values expected for genotype Aabb. Aabb was consequently phenotypically 10.4%, 74$ and 17.6% for O-lG, 26 and 30—40 respectively. -20- These values were computed assuming that the genotypic frequency of'F3 #12 was as stated. This assumption was consistent with the proposed hypothesis in that the relative percentages in each category for each genotype conformed to what was logically expected. Now that several of the phenotypic frequencies had been estimated, values were assigned to others. Under the proposed genetic hypothesis, if the minor gene (B) had complete dominance, the phenotypes of AABB and AABb would be the same. Likewise AaBb and AaBB would be alike in phenotype. New, using the 308 the expected percentages for aaBb were calculated which were the same as for aaBB. After the percentages were calculated as above for each genotype, minor adjustments were made in the percentages in order to get a better fit of the expected and observed values for the parents, F1, backcrosses and all.F3 populations. The values arrived at are listed in Table A. The values of the Chi-square test for goodness of fit are listed in Table 5. These values were computed after converting expected percentages back to expected number of plants for each population. Now that the phenotypes were known for each genotype, an expected F2 population could be computed and the comparison with the observed population used as a test of the hypothesis. Since thein population was large, the observed frequencies could be expected t. be dependable. The computation of the expected F2 is shown in Table 6 and its Chi-square test for goodness of fit is in Table 5. .- ., tv .3K’n‘:°.'5 Resistance Rating 2 genotype 0—1 3-4 M38 88 12 AABb 88 12 M33 55 39 6 AaBb 55 39 6 AAbb 35 58 7 Aabb 10 66 24 aaBB 36 61+ aaBb 36 6!. aabb 2 98 -22.. Table 5. Chi-Sqmre Test for Comparison of Observed and Expected L 29W _ Epulation x2 Pm F1 (Pa 1: Pr) 5.75 .059 F2 (Pa 1: Pr) 2.065 .357 F3 " #9 5.05 .080 F3 " #10 0 F3 " #11 6.92 .032 at F3 " #12 1.1.1. .484 F3 " #13 4.30 .122 F3 " #11. 7.26 .025 * F3 ., #15 5.1.5 .066 F3 " #16 3.97 .136 F3 " #17 3.78 .150 F3 .. #18 2.79 .250 F3 " #19 3.54 .175 F3 " #10,11,1z. 0.992 .607 F3 ., #12,13,16,18,19 0.19 .950 BC to Pr 1.186 .559 BC to Fe 5.19 .075 F2 (Pm 45 x Pr) 3.507 .175 1‘F'rom Pearson and rhrtley (17) “um “a; 1 -23- Table 6. Computation of the Expected F2 (PB X Pr) Populational _ J‘W Genotype Frequency as % 0-1 RatingéFrequencyé_4 AABB 6.25 5. 50 .75 AABb 12.5 11.0 1.50 AaBB 12.5 6.88 4.88 .75 AaBb 25.0 13.75 9.75 1.5 AAbb 6.25 2.19 3.63 .44 Aabb 12.5 1.25 8.25 3.0 aaBB 6.25 2.25 4.0 aaBb 12.5 4.5 8.0 aabb 6.25 .125 6.13 Expected F2 Population 100 40.56 35.61. 23.82 -24.. The expected F3 distributions from all possible F2 genotypes have been computed and listed in Table 7, and the genotypes of each of the observed distributions are listed in Table 8. The observed populational frequencies grouped in Table 3 as well as the expected frequencies for each population as calculated from the model were plotted in Fig. 2-4. Those F3 populations which were assumed to be from F2 plants of identical genotype were grouped and plotted on the same graph. IDENTIFICATION OF THE GENES In order to compare the Seminole genes for resistance with those previously studied such as m1 in He 45 and sz in [EDA 5, crosses were made between Seminole and these two varieties. The F2 populations were then screened for resistance. In the M 45 I Seminole F2 the genotypes containing Pm1 were completely resistant and fell in category 06 as seen from various PM? 45 F1 '3 in the general variety screening (Table l) . Using the values computed for each genotype and the ratios expected from normal segregation of three genes, the expected F2 was computed in Table 9. This cross served not only to distinguish between the three genes but to further support the genetic hypothesis proposed for the inheritance of resistance in Seminole. The F2 of Seminole X IBDA 5 was also trade and screened, but since no smceptible types appeared (Table 2) , no model will be proposed . 130m -25. Table 7. Expected Frequencies of F3 Populations Expressed as Resistance Rating 2 F2 Genotype 0-1 3-4 AABB 88 12 AABb 75 22 04 1333 50 32 19 AaBb 41 35 24 AAbb 35 58 7 Aabb 14 46 4O aaBB 36 64 aaBb 28 73 aabb 2 98 ~26- ab e l t na n Deliciom 51 (P9) aabb Seminole (Pr) AABB F1 (Pa 1 Pr) AaBb F2 (Pa x Pr) distribution as listed in table 6 BCr MABB : lAABb : lAaBB : Mb 303 1AaBb : lAabb : laaBb : laabb Genotype of the Parental F2 Plant F 3 (P8 1 Pr) #9 11st F3 « #10 use F 3 " #11 Mbb F 3 ,. #12 Aabb F3 " #13 Aabb F 3 " #11. AAbb F 3 " #15 AABb F 3 " #16 Aabb F 3 " #17 MB 1“ 3 ,, #18 Aabb F3 " #19 Aabb -27. Figure 2.-— Percentages of the genetic populations falling into the grouped mildew rating categories. A . Seminole (Pr) Genotype: AABB. B. Delicious 51 (Is) Genotype: aabb. c. Delicious 51 x Seminole F1 (Ps x Pr F1) Genotype: AaBb. D. Delicious 51 X Seminole F2 (Ps X Pr F2) For genotype see Table 6. ,7 98 98 7//////////1///////////// 43—h... mo 5.sz O 6 O I’I O 0 4 2 1 3+4 O-I-I 3+4 b O 0 4. 2 hzmo mud RATING RATING 25 24 3+4 33 35 as .7 mV//////// W e w . w J "D RATING 0+| RATING :b e ' ,on 9f t : hay: =3 eu-io : i 39; ist-a ratio 30-1 2 3-4 <‘_ 0.2 48/64 .75 AABBcc 1/64 .014 .0019 We 2/64 .0275 .0038 AAch 1/61. .0055 .0091. .0011 AaBBcc 2/64 .017 .0122 .0019 AaBbcc 4/61. .034 .0240 .0033 Aabbcc 2/61. .003 .0210 .0075 aaBBcc 1/64 .0056 .010 aaBbcc 2/64 .0112 .020 aabbcc 1/61. .0003 .015 total .8510 .0894 .0593 * C is used here to represent the gene Pml and c for PI;- DISCUSSION The greenhouse screening method used was designed to manipulate the dynamics of the relationship between pathogen and host. In the early stages of plant growth the amount of the seedling covered by mildew is a function of the relationship between the growth of the plant and growth of the fungus. From the time of innoculation with mildew spores, about ten days is needed for a colony to be started and three weeks before the colony is well developed. If during this time the cantaloupe plant is growing rapidly, the leaf area my be increased by a factor of six or eight. This would serve to dilute the concentration of colonies. Since the resistance rating depends on the density of colonies, control of the relative growth rates was important. It is fortunate that cantaloupe seedling growth can be inhibited by temperatures below 659F and that mildew growth is restricted by temperatures above 759E. This allowed the seedlings to be grown at fairly high temperature to the size desirable for innoculation, at which time the temperature was lowered to increase mildew growth. The comparison between field resistance and greenhouee seedling resistance showed that there was generally good agreement between the relative ratings of the varieties. The greatest variation, however, was between those rated in categories 0 and l as well as between 3 and 4. It appeared that the fine differences separating these categories could not be reliably differentiated in the screening -34- -35- tests as conducted. This lends support to the argument for combining these rating categories as was done in the genetic analysis. The method of handling the populations by working with percent— ages was awkward and inferior to a method using means and standard deviations. The latter method would have permitted a much more sophisticated analysis yielding a good dea1.more information. For example, the amount of domdnance could have been computed instead of merely estimated. It was unfortmate that with a problem of this type the scale could not be made continuous. There was no practical way that this could be accomplished. The proposed genetic hypothesis itself seemed to fit the pop- ulations acceptably well as seen in the graphs in Fig. 2-4 and in the Chi-sqmre values of Table 5, but the test of its validity using the F2 should have been more fully supported. In the case of the gene- types represented in low frequency in the F2, a major change in phenotypic frequency would have had little effect on the total ex:- pected F2 distribution. It would have been desirable to have had several more F3 populations whose F2 parents represented the entire range of F2 genotypes. This is now being done in order to further support the hypothesis. The results of the PMR 45 X Seminolein supported the hypothesis that the two genes for resistance in Seminole were different from Pml in PM?! 45. This was to be expected from the differences in phenotype seen in the field evaluation. In comparing the Seminole resistance with that of USDA 5, the results were inconclusive. No F2 segregation was seen for resistance. . ~36- There were several possible reasons for this. It is possible that with the several genes involved, the population screened was not large enough to have a high probability of a susceptible type occurring. Another possibility is that the major gene A is either closely linked with m2 or is an allele. Further studies are being made on this question. CONCLUSIONS A greenhouse cantaloupe seedling screening test for re- sistance to powdery mildew was developed and correlated with a field rating. The results of this stuiy indicated that resistance to powdery mildew in the cantaloupe variety Seminole was controlled by two gene pairs whose effect was at least partially additive. Both genes exerted dominance, with A, the mjor gene, showing incomplete dominance and B, the minor gene, showing complete dominance. The genes A and B were different from and not linked to Pml, the gene for resistance in PMR 45. Names for genes A and B are undesignated pending further study on their relationship to P212 in [EDA 5. 10. 11. 12. 13. LITERATURE CITED Alempoulos, C.J. 1962. Introductory )tvcolog. John Wiley and Sons, Inc. New York. 292-300. Bohn, G.W., G.N. Ihvis, R.E. Foster and T.W. Whitaker. 1965. Campo and Jacumba, new cantaloupe varieties for the south- "881'” 09.1. [Ago 8.100 and T.W. Whitaker. 1964. Genetics of resistance to powdery mildew race 2 in mmkmelon. Phytopath. 54: 587-91. Conover, R.A. 1958. Control of powdery mildew on melons. Ann. Rept. Fla. Ag. Exp. Sta. :6: 806. Dekker, J. and A. J. P. Oort. 1964. Lbde of action of 6-azauracil against powdery mildew. Phytopath. .54: 815-818. . 1963. Effect of kinetin on powdery mildew. Nature. 12:: 1027-28. Godfrey, S.H. 1953. New Gold from the Rio Grande. Southern Seedsmn 16: (12) . . 1950. Cantaloupe powdery mildew control with dinitro capryl phenyl crotonate. Phytopath. 43,: 335-37. Ivanoff, 8.8. 1957. The Homegarden cantaloupe, a variety with combined resistance to downy mildew euio ) powdery mildew W c choracea and W m1).Phytopath. 52: 552-5 . Jagger, I.C., TM. Whitaker and D.R. Porter. 1938‘. A new biotic form of powdery mildew on muskmelons in the Imperial Valley of California. Plant Dis. Reptr. _2_2_: 275-76. . raw. Whitaker and D.R. Porter. 1938b. Inheritance in gm m of resistance to powdery mildew . Phytopath. 2,8: 671. and G.W. Scott. 1937. Developnent of powdery mildew resistant cantaloupe no. 45: U.S.D.A. Circ. “:13 1-5. Middleton, J .T. and C.E. Yarwood. 1946. Fungicidal control of cantaloupe powdery mildew. Cal. Ag. Exp. Sta. Bull. 622. -38.. 15. l6. 17. 18. 19. 20. 21. 22. 23. 25. 26. 27. 28. -39- Harrison, R.W. 1964. Germination of conidia of m . 1960a. Studies of clonal isolates of gighgmcearum on leaf disc culture. Wool. 2: 388-93. 1960‘”. Compatibility of several clonal lines of Mimic 21211229222113- }tvcol- games-91.. Pearson, EM. and H.0. Hartley. 1958. Biometrica Tables for Statisticians. University Press, Cambridge. 122-29. Pryor, D.E., T.W. Whitaker and S.N. mvis. 1946. The develop- ment of powdery mildew resistant cantaloupe. A.S.H.S. 441: 347-56 0 and T.W. Whitaker. 1942. The reaction of cantaloupe strains to powdery mildew. Phytorath. 32: 995-1004. Randall, T.E. and J.D. Menzies. 1956. The perithecial state of the cucurbit powdery mildew. Plant Dis. Reptr. 49: 225. Schmitt, J.A. 1955. The host specialization of W W from zinnia, phlox and cucurbits. Mycol. 41: 488-701. Stone, 0.14. 1962. Alternate hosts of cucumber powdery mildew. Ann. Appl. Biol. .59: 203-10. Van mltern, F. 1951. Georgia 47, mildew 0. Southern Seeds- mn M: (11). 31, 450 Whitaker, T.W. and D.E. Pryor. 1947. Correlated resistance of leaves, cotyledons and stems of Que-min male 1.. to cantaloupe powdery mildew W gichgraceam D.C.) . Hiytopath. 31: 865-67. and G.N. Davis. 1946- Cantaloupes 6 and 7. Southern Seedsman 3: (2). 14, 64. and D.E. Pryor. 1945. The reaction of 21 species in the cucurbitaceae to artificial infection with cantaloupe powdery mildew We meanness-am) . Phytopath. 25: 533-34. . D.E. Pryor and 6.11. Davis. 1943. Cantaloupe #5. Western Grower and Shipper. M: (9) . 8, 22-23. and D. E. Pryor. 1942. Genes for resistance to powdery mildew in Emmi; gen. A.S.H.S. Al: 270-72. '40- 29. Whither, BQF. 1960. Seminole, a high-yielding good quality, downy and powdery mildew-resistant cantaloupe. Fla. Ag. up. Sta. 3-122. 30. Yarwood, C.E. 1957. Powdery mildew. Bot. Rev. 22: 235-300. 31. . 1936. The diurnal cycle of the powdery mildew Mg m. Jour. Agr. Res. 52: 645-57. "'II'I‘IIIII‘I'III’IIIIIIIIIII'IIIIII“ 3120