mmnnmnou Ann lNHERITANCE or FLOWER MGMENTS m nmom ‘MEDICAGOC specms Thais fer tho We cf Ph. D. MlCHiGAN STAT! UNIVERSWY Richard L.‘ Coop“ 1961 IHES|5 ) , This is to certify that the i 5/ p ‘ thesis entitled » V‘. I, / Identification and Inheritance of Flower Pigments in Diploid Alfalfa (Medlcago) presented by Richard L. Cooper has been accepted towards fulfillment of the requirements for Ph.D . degree in Agriculture @066. 52% Major professor Date December 15, l36l 0469 L I I I A R Y Michigan Sue University i '2 g i g IDENTIFICATION AND INHERITANCE OF FLOWER PIGMENTS IN DIPLOID mmcgao SPECIES BY Richard L? Cooper AN ABSTRACT Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Farm Crops Year 1961 Approved .. .9 .-. _. - rag- . ' .49.“?! ABSTRACT IDENTIFICATION AND INHERITANCE OF FLOWER PIGMENTS IN DIPLOID MEDICAGO 51336:st ' by Richard L. Cooper Flower color has been used as a convenient tool for the study of tetraploid inheritance in alfalfa. Such stu- dies have been only partially successful because of the complexities of tetraploid inheritance and the phenotypic interaction of flower pigments. A technique whereby the complicating factors of tetraploid segregation and pheno- typic interactions are removed would be useful in determin- ing the inheritance pattern of flower color in alfalfa. From the F2 intercross progeny of a tri-species diploid hybrid (M. gaetula x E. falcata) x g. sativa, 42 flower color types were identified. These plants were used as source material in an intensive crossing and selfing pro- gram of the various color types. Flower pigments were separ- ated chromatographically from 700 plants within 23 segrega- ting families, identified, and their inheritance patterns determined. Three anthocyanin pigments were found in every flower containing anthocyanin. The production of these pigments was under control of a single dominant gene. Richard L. Cooper Nine flavonol pigments consisting of six quercetin glycosidee and.1flaree kaempferol glycosides were also observed. Two quercetin glycosides and one kaempferol glycoside were pre- sent in every plant examined. Although a definite segrega- tion pattern was not established, two hypotheses were pro- posed for the inheritance of quercetin pigments. Two kaempferol glycosides exhibited independent segre- gation in a 3:1 ratio indicating each was controlled by a single dominant gene. The carotenoid pigment in alfalfa flowers was identi- fied as xanthophyll ester and exhibited a l:h:6:h:l segrega- tion in some families, indicating control by quantitative factors at two loci. Phenotypic correlation of pigments indicated that blue and purple colors were due to mixtures of three anthocyanin pigments. Evidence obtained indicated that kaempferol gly- cosides produce a phenotypic effect by copigmentation with anthocyanin pigments. Certain intensity levels of quercetin glycosides produced a yellowing effect, but yellow flower color intensity was most closely associated with intensity of xanthophyll pigments. From the inheritance of flower pigments and their pheno- typic correlation, an inheritance chart for flower color in diploid alfalfa is proposed. IDENTIFICATION AND INHERITANCE OF FLOWER PIGMENTS IN DIPLOID MEDICAGO SPECIES By Richard LgyCooper A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Farm Crops Year 1961 ACKNOWLEDGEMENT The author wishes to express his sincere gratitude to Dr. Fred C. Elliott for his guidance in this study and the pre- paration of the manuscript. The author is grateful to Dr. Charles R. Olien for his helpful suggestions on chromatographic procedures and to Dr. Carter M. Harrison for assistance in preparation of the manuscript. Also, the author is very grateful to his wife Norma for her help and encouragement in completing this study and typing of the manuscript. ii LIST OF TABLES LIST OF FIGURES REVIEW OF LITERATURE MATERIALS AND METHODS EXPERIMENTAL RESULTS LITERATURE CITED O TABLE OF Page iv 10 21 46 66 67 Table l. 2. 3. 7. 8. 9. 10. ll. 12. 17. 18. 19. LIST OF TABLES Solvent systems used for the chromotograpny of flavonoid pigments . . . . . . . . . . . . . . . Authentic compounds and their source . . . . . . Rf values of anthocyanins in SAW and HAc-HCI . . Rf values of anthocyanin glycones in Forestal solvent and i; ‘31:; o o e o o e o o o o e o e e o o Rf values of anthoxanthin in BAU’and HAO . . . . Rf values of anthoxanthin aglycones in Forestal Solvent 0 I O I O O O O O O C O I O I O I O O I Observed Observed Observed Observed Observed segregation for anthocyanin production. segregation of kaempferol glycosioe F . segregation of kaempferol glycoside G . segregation of xanthophyll . . . . . . segregation of xanthopnyll into combined classes . . Joint segregation of P Joint segregation of anthocyanin P . . . Joint segregation of Joint segregation of and anthocyanin P . Joint segregation of Joint segregation of glycoside F . . . . Joint segregation of glycoside G . . . . kaempferol F and anthocyanin kaempferol glycoside G and kaempferol glycoside F and e kaempferol glycoside F and G xanthophyll and anthocyanin. xanthophyll and kaempferol xanthophyll and kaempferol Phenotypes and genotypes proposed for parental clones . . . . . . . iv . Page 11+ 18 2L; 25 26 28 31 32 33 3'4 35 36 37 37 38 38 39 39 60 Figure l. 2. 3. LIST OF FIGURES Page Flower color types in diploid alfalfa . . . . . 22 Chromatograms of flower pigments in diploid alfalfa 0 O O O O O 0 v. 0 O I I 0 I O C O O O O 23 Flower color inheritance in diploid alfalfa (Chart) 0 O I O O O O O O O O O O O O O O O I O 65 INTRODUCTION Flower color has been used as a convenient tool for the study of inheritance in tetraploid alfalfa. These studies have been only partially successful because of the phenotypic interactions of the blue and yellow pigments, and the compli- cations resulting from tetraploid segregation. The availability of natural diploid alfalfa provides a tool whereby the inheritance of flower color can be studied without the complication of tetraploid inheritance. Once an inheritance pattern is established at the diploid level, it may serve as a useful guide in the interpretation of tetra- ploid inheritance. Separation of the component pigments and study of their inheritance individually removes the phenotypic interactions encountered in a study of flower color inheritance. From the inheritance of these pIgments, an explanation of flower color inheritance in diploid alfalfa can be obtained. The purpose of this study was to (1) identify the flower pigments in alfalfa (2) study the inheritance of these pig- ments (3) use this information to explain flower color in- heritance in diploid alfalfa. REVIEW OF LITERATURE Flower Color Inheritange in Alfalfa Flower color has been used as a convenient tool for studying inheritance in tetraploid and diploid alfalfa. Earlier investigations were with tetraploid alfalfa. In these studies, flower colors were grouped into three main classes "purple", which included purple and variegated flow- ers, "yellow", which included all gradations of yellow, and "white". Using this system of classification, 3:1, 15:1 and 63:1 F2 ratios of purple to non-purple were observed from crosses between tetraploid fl. sativa (purple flowered) and E. falcata (yellow flowered) and also in crosses between purple and white flowered E. sativa plants. Most investiga- tors explained these ratios on the assumption of disomic in- heritance and segregation for one to three supplementary (duplicative) factors for anthocyanin production (Lepper and Odland 1939; Armstrong and Gibson l9hl; Weihing 1948). Stan- ford (1951) obtained the same ratios but pointed out that they could be explained either on the basis of tetrasomic or disomic inheritance. The occurrence of purple flowered progeny from white x white or white x yellow crosses has been cited as evidence of complementary gene action (Lepper and Odland 1939; Weihing 19h8). Complementary gene action in diploid alfalfa was re- ported by Twamley (1955) who obtained 9:7 ratios in some F2 2 3 families of a sativa-falcata diploid cross. Yellow color inheritance has not been as extensively studied because of the epistatic effect of purple pigment and the quantitative nature of yellow color segregation. Grouping all gradations of yellow into a single class, Odland and Lepper 1939, proposed a single gene for yellow pigment production. Weihing 19h8, in a study of F2 segre- gating families observed 3:1, 15:1 and 63:1 segregation of yellow to white. From these observations he proposed at least 3 factors for yellow pigment production. A more intensive study of yellow pigment inheritance was made by Twamley (1955) in both diploid and tetraploid plants. From the distribution into intensity classes of 85 yellow F2 plants of a diploid sativa-falcata cross, he pro- posed yellow pigment production was controlled by four quan- titative genes, or more probably by four factors, some of which are quantitative and some which are qualitative. Flower Piggents in Alfalfa Flower pigments in alfalfa can be divided into three main classes, anthocyanin (blue and red pigments), anthoxan- thins (yellow sap soluble pigments) and carotenoids (yellow plastid pigments). These three classes of pigments were first reported by Twamley (1955) who identified the antho- cyanin as primarily malvidin. Lesins (1956), however, using paper chromatography was able to separate four anthocyanin v A v. f v . \ bl. .. -1 ._ (.1. . - t 41.- :3; - u .. .1 . g , .. i: E- \ \fi l .‘_ g, . \‘LL _1 L _ .' u .4 ~1..4 ‘ \I,u _ .4 i , _ l , .7 . A J . . v . .L .1 .-... a J . ‘_._ . . ~ a». k -D __.. . > , _ v. A4... k. , '. nil. Z , A ‘ . | -L‘ J no ,_.._’. ’ It. '.'\ C L l .a I... _ -s. 1 .(J -1 ,. I. 1' 1 ‘ .Lu. .1 . -._ v 3 hav». J ("-11 ..:7.'~-:L.‘ .1 .I‘... _. "CLAN/ . . ~‘..1_J_:\' 9...", . tan-{.21. ., :..-I')111!i'.1 L, .2 1'. file ..‘_1 s ...x- .;.ii . I pigments from alfalfa, and identified their aglycones as delphinidin, petunidin and malvidin. Davies (1958) identi— fied delphinidin, cyanidin and malvidin, as well as the antho- xanthins, quercetin, kaempferol, myricetin, and tricin from extracts of alfalfa flowers. In a histological study of alfalfa flowers, Lesins (1956), observed that the coloring pigments were located in the epidermal layers of the petals. In variegated flowers, the coloring matter consisted of yellow carotenoid plastids, anthocyanins and anthoxanthins. Green color was explained as a result of a side by side purple-and-yellow cell mosaic occurring when not all cells had both purple and yellow pig- ments. When numerous cells contained both yellow and purple pigments, a color subtraction phenomenon was produced, giving dark color. Egpetics of Flower Pigments Anthocyanins and anthoxanthin§.-Much of the knowledge about inheritance of anthocyanin and anthoxanthin pigments stems from the research of Scott-Monerieff 1936, Beale 1939, and Lawrence 1935. These investigators combined the infor- mation on chemical structure of flower pigments (Robinson 1933) with the accumulated knowledge of flower color inheri- tance, and established the first examples of the relationship between single genes and a simple biochemical difference in higher plants. 5 All flavonoid pigments arise from a common 015 precursor and are elaborated into the different classes by processes of hydroxylation, methylation, glycosidation and acylation (Harborne 1958). Specific examples of these types of action have been reviewed by Scott-Moncrieff (1936) and Lawrence 33 El (l9h0). More recent examples have been found by the use of paper chromatography to separate complex mixtures of pig- ments (Geissman £3 31 195k; Dodds and Long 1955; Nordstrom 1956; Alston and Hagen 1958; Harborne and Sherratt 1958). Of particular pertinence to the study of flower pigment inheritance in alfalfa are the genetic studies of flower color in sweet pea, Latnyrus odoratus. Beale (1939) reported that some varieties of sweet peas contained a mixture of delphin— idin, petunidin, and malvidin glycosides. He also observed the presence of an anthoxanthin, which in the presence of anthocyanin formed weak additive complexes, making the antho— cyanin bluer than it would be in absence of anthoxanthin (co- pigmentation). This presence of anthocyanidin mixtures and copigmentation effects of anthoxanthins was confirmed by the recent work of Pecket (1960). Based on observations of a number of different Lathyrus species he concluded that the bluing of the wings within a species or variety and the blu- ish red to blue flower color of other species is due primari- ly to the copigmentation action of the anthoxanthins. 6 Carotenoids.-One of the first genetic studies on caro- tenoid production was reported by Mangelsdorf and Fraps (1931), in a study of yellow endosperm color inheritance in corn. They found that zeaxanthin, the principle pigment in yellow endosperm was controlled by a single quantiative fac- tor Y. Because of the triploid (3N) condition in the endosperm, four dosage levels of Y, (yyy, yyY, yYY, and YYY) were possi- ble with the four corresponding phenotypic classes (white, pale yellow, dilute yellow, and deep yellow endosperm color). LeRosen 33 31 (19h1) in an inheritance study of lyco- pene, the red carotenoid pigment in tomato, found production of this pigment controlled by a single dominant gene. In an intensitive review of the literature on flower color inheritance, Clark 35 51 (1960) compiled inheritance charts for each of 75 different species. Of these 75 species, only 4 species had genetic factors identified as affecting carotenoid production. These were (1) Mimulus cardinalis with a single quantitative factor for carotenoid production, segregating 1:2:1 (Brozek 1932; Vickery and Olsen 1956) (2) Tropaeolum majus with two complementary genes for production of carotenoid (Moffett 1936; Sutton 1939) (3) Eschscholtzia californica, in which carotenoid production is controlled by a single dominant gene (Douwes 1943) (h) Primula ER. with a single dominant gene for carotenoid production (Frimmel 1932). 7/ Chromatographic Iggntification of Pigments Anthocyanins.-The earlier color and distribution tests developed by Robinson (1931) for the identification of antho- cyanins, have largely been replaced by paper chromatography. Bate-Smith (l9h8) was the first to use paper chromatography for separation and identification of flavonoid pigments. Since that time, paper chromatography has been used by many investigators for separation of complex mixtures of flower pigments (Bate-Smith 1950, 1954, 1956; Bate-Smith and Westall 1950; Geissman £3 31 195A, 1955; Nordstrom gt 3} 1953, 1956; Hayashi 195h, 1957; Harborne 1958; Alston and Hagen 1958; and many others. A very excellent review of the paper chro- matographic procedures used in identification of anthocyanin pigments was published by Harborne (1958). Extensive lists of Rf values for numerous anthocyanins and anthocyanidins, in various solvent systems, are avail- able in the literature, serving as useful guides for tenta- tive identification of flower pigments (Bate-Smith 1950; Geissman 195h, 1955; harborne 1958). For more positive identification, these authors suggest that unknown pigments should be compared directly with an authentic sample on the same chromatogram (cochromatography), preferably in two or more solvent systems. Anthoxanthins.-In general the chromatographic proce- dures used for the separation and identification of antho- cyanins are equally applicable to anthoxanthins. Techniques- I“ 9 fcu? extraction of anthoxanthin and their hydrolysis for obtaining the aglycones have been reported by several in- vestigators (Geissman 1954, 1955; Nordstrom 1953, 1956; Bate-Smith 1954; Roberts 1956). Several published lists of Rf values for the anthoxan- thins are available for use in tentative identification of extracted anthoxanthin pigments (Bate-Smith 1950, Gage 23 31 1951; Geissman 1954, 1955; Roberts 1956). Carotenoids.-Separation of carotenoid pigments has been primarily by use of adsorption columns. Procedures have been worked out and summarized by Strain (1943) and Goodwin (1955). An excellent method for preparation of columns by a wet pack- ing method was described by Williams (1948). a 1‘1? y£flf was MATERIALS AND METHODS In the F2 intercross population of a trispecies hybrid between diploid fl. sativa and the F1 (fl. gaetula X 5. falcata), forty different color types were obtained, ranging from white to yellow, through green (variegated) to purple. Color types were identified by comparison with freshly picked flowers from standard color plants. Whenever a new color was found, it was given a new number and was made a standard plant for that phenotype. An effective color chart was made by placing freshly picked flowers from each of the standard plants into water-filled glass vials carried in a tray. Two plants, representative of each flower color pheno— type, were moved into the greenhouse for use in a crossing program. Special emphasis was placed on crosses of purple and variegated flowered plants to yellow and white flowered plants. Yellow by white crosses were also made as well as crosses within the major color classes. Several plants were selfed, with particular emphasis on the variegated flower types. Since there were no plants with the orange yellow flower color of i. falcata, three different sources of E. falcata were also included in this crossing program. All flowers were emasculated using Lesins (1955) tech- nique of dipping the tripped flowers into 57% ethanol and rinsing quickly in fresh water. When nearly dry, the adher- ing anthers and pollen were removed by air suction. As a 10 11 check, 100 flowers of a self-fertile clone were emasculated and not pollinated. Only 2 pods and one seed developed. Pollination was accomplished by transferring pollen from plant to plant with a tooth pick. Racemes were then tagged as to date and parentage. After maturity, seed was harvested from the various crosses, germinated on moist filter paper in petri dishes, and transplanted to the greenhouse into 2"X2" peat pots. The pots were placed on top a 4" layer of sand. A strip of mesh wire was stretched about 5" over the top of the bench to support developing seedlings. In this manner, 1000 seed- lings were grown to the blooming stage in one 3'X20' green- house bench, without transplanting. Using the fresh flower color chart, the progeny were classified as to phenotype. Special emphasis was made to group phenotypes within progenies so it would not be necessary to analyze chromatographically all plants. A maximum of five plants within each phenotype were analyzed, but if segregation for pigments appeared in these five plants, additional plants were examined. For chromatographic analysis, a minimum of 20 flowers per plant were needed, with a maximum of 30 flowers. For many plants it was necessary to make three or four collec- tions to obtain the minimum number of flowers. To insure that the same plants were used in suosequent collections, each seedling was tagged as to cross and plant number. 12 thn1 flowers were collected the petals were separated from the other flower parts by clipping off the standard, trip- ping the flower and pulling off the wings and keel in a unit. The petals were put in small coin envelopes labeled as to cross, plant number and color type, and placed back in the envelope box. by storing the box in a cool dry place the small petals dried readily and could be held for 2 or 3 months with no apparent breakdown of pigments. Extraction Procedure.-Using the technique developed by Vickery and Olsen (1956), the plastid borne, carotenoid pig- ments were extracted by placing 3 mg of dried alfalfa petals in a mortar and grinding in petroleum ether. A few grains of sand were added for abrasive action. The ether extract was deCanted off and centrifuged to remove any suspended material. As soon as the residue in the mortar had dried, 3 ml of .1N H01 was added for extraction of the sap soluble anthocyanin and anthoxanthin pigments. The tissue was ground further in this acid solution and then, while the residue was still in suspension, poured into the test tube contain- ing the dried precipitate from the petroleum ether extract. The solution was stirred, centrifuged and the supernatant extract poured off. In some flowers, particularly the bright yellow types, the residue after acid extraction still con- tained some yellow plastid pigment. This was removed by adding 3 m1 of isopropyl alcohol to the residue, stirring it up and letting it set 10 to 15 minutes. Centrifuging and 13 deceuating off the yellow isopropyl alcohol extract left a colorless white residue. For convenience in centrifuging, extractions were done on two petal samples at a time. Chromatographic Analysis.-The acid soluble pigments, anthocyanins and anthoxanthins, were separated and identi- fied by use of paper chromatography. Inital comparisons be- tween ascending and descending chromatography indicated that the ascending technique described by Vickery and Olsen (1956), with some modification, would more aptly meet the require- ments needed. Whatman No. 1 filter paper was cut into 14 1/2 X 1 1/2 inch strips and the pigment to be tested spotted 3 cm from the bottom. The paper was attached to the pro- per size cork and then inserted into a large test tube (15" X 2") such that the bottom of the paper was 0.5 cm in the appropriate developing solvent. The cork sealed the tube, permitting the formation of a saturated atmosphere. Since most of the acid extracts were rather dilute sol- utions, it was necessary to concentrate the pigment by re- peated application of the extract to the paper. To speed up this procedure, a mimeoscope was utilized for light as well as heat to increase rate of evaporation. The standard pro- cedure was to apply twenty drops from a glass rod which had been rounded on the end by heating. The same glass rod could be used from sample to sample by rinsing it in running water and wiping it with a clean cheese cloth. Although this is not as precise as the use of micropipettes, it proved both satisfactory and expedient. L . Lw p \ (“pl‘i‘m u‘ at .‘lj‘lfiwl‘iml‘ f... . T .‘ ‘ ,v, 4111.4);4 wind 11+ The prepared strips were placed in twenty test tubes, supported in wooden racks and placed inside a chromatocab. The insulated chromatocab decreased temperature fluctuation on the tube walls. Under these conditions, a very uniform solvent front was obtained with readily repeatable separa- tion of pigments. Solvent Systems.-Four solvent systems were used in this T- study and are listed with abbreviation, composition, and run- to ing time in Table 1. Table 1. Solvent systems used for the chromatography of flavonoid pigments. Abbrev- Composition Ratio Running iation V/V Time hrs. 1 % BAW n-butanol-acetic acid-water 4:135 20 HAc-HCl2 water-acetic acid-cone. HCl 82:15:3 8 15% HAc3 water-acetic acid 85:15 6 4 Forestal acetic acid HCl-water 30:3:10 15 lBate- -§mith 1950, Gage, et 21 1951, Harborne 1958; 2Harborne 19553 Gage 6t 81 19513 —EBate-Smith 1954, Harborne 1958. *Top phase used— The BAW solution was the standard solvent used in chro- matographic separation of the flower pigments in the raw plant extracts. The three other solvents were used for iden- tification of the purified pigments. BAW was equilibrated three days in a separatory funnel before the lower phase was removed and the upper phase distributed in 30 cc aliquots into the test tubes. The BAW in the tubes was replaced with 15 fresfli solution at least every two weeks, depending on the frequency of use. In descending chromatography the filter paper strips, spotted with the extract, were equilibrated 2; hours with the lower phase of the bhfi solvent system before the top phase was added to the top tray (Bate-Smith 1950). The running time for the descending saw was only 9 hours. For purposes of identification, the solvent systems ' I used were those for which published hf values of flavonoid )- pigments were available. The hf values for anthocyanins F. were determined in saw and flAc-HCl and the Rf values for the anthoxanthins in BAT and 15% Hac. For the aglycones of the anthocyanins, (anthocyanidins), Rf values were determined in "Forestal" solvent and on 1; H81 acid washed paper in BAH. The acid washilg was needed to prevent fading out of the an- thocyanidins (Harborne, 1958). The Rf values for the antho- xanthin aglycones were determined in the "forestal" solvent only. The "Forestal" solvent was particularly effective for aglycone analysis because any glycosioes remaining as a result of incomplete hydrolysis have a very high Bf value in this solvent and hence are not confused with the lower Rf values of the aglycones (Bate-Smith 1950). An important aid in locating the flavonol pigments on the chromatograms was a blank tube containing 30 cc of am- monium chloride in the bottom. When the chromatograms were inserted into the ammonia vapors, the flavonol spots were intensified to a bright yellow. Also, in some cases, ultra 16 -violet light was used in combination with ammonia vapor to locate very dilute spots. Purification of Pigments.—For purification of individual pigments, a series of 30 chromatograms were run of the concen- trated extracts from plants containing the appropriate pig- ments. The pigment bands were then cut from the dried chro- matograms and placed in 15 ml vials. To these 30 strips, approximately 3 cc of .1N HCl was added as an eluting sol- vent. Elution with the more commonly used 1% H01 or 1% methanolic HCl (Harborne 1958) did not prove satisfactory as the anthoxanthin glycosides were readily hydrolyzed in this solution at room temperature. Pigment eluted in .lN HCl re- mained primarily as the glycoside even after one month stor- age in the refrigerator. Hydrolysis of glycosides.-In order to identify the agly- cones, acid hydrolysis of the purified glycoside was used. Three ml of .1N HCl containing the eluted pigment was made 2N by adding concentrated dCl. This solution was then heated in a boiling water bath, in absence of light, for 10 minutes (Nordstrom 1956). Hydrolysis at 100°C for 20 minutes in the light, as recommended by Bate-Smith (1954) was too severe for the dilute solution of eluted anthocyanins, and the pig- ments were lost. The aglycone was separated from the hydro- lyzate by partitioning into a few drops of isoamyl alcohol. The isoamyl fraction was then used in spotting the chromato- grams (Bate-Smith l95h). 1? END? the anthoxanthins, hydrolysis of the eluted pig- UKWNS in.lN HCl for 10 minutes gave nearly complete hydroly- SiS (Roberts 1956). A more severe treatment of 2N H01 for 15 minutes destroyed the eluted pigment although it was sa- tisfactory for hydrolysis of the raw extract. Here again the isoamyl alcohol was used to separate the aglycone from the hydrolyzate. Identification of pigments.-For identification, the Rf value for each of the purified glycosides was determined in both an organic and an aqueous solvent. In addition, the Rf value for the aglycone of each pigment was determined in two solvent systems for the anthocyanidins and one solvent system for the anthoxanthin aglycones (Table 1). These Rf values were compared with previous reported Rf values and where available, were cochromatogrammed with authentic compounds. The compounds obtained and their sources are listed in table 2. 18 ‘Yablfii 2. Authentic compounds and their source '— Pigment Source Delphinidin 3,5 diglucoside J. B. Harborne, John Innes Delphinidin 3 monoglucoside Horticultural Institute Hartford, England malvidin 3,5 diglucoside Mann Chemical Company Kaempferol 135 Liberty Street New York 6, N. Y. Quercetin General Biochemicals Rutin Laboratory Park Crystalline Carotene Chagrin Falls, Ohio Delphinidin Hydrolysis of their respective malvidin standard glycosides Petunidin Hydrolyzed extract of maroon Cyanidin petunia and of red rose, respectively Carotenoid Intensity.-It was noted that there were sev- eral levels of intensity for the ether soluble carotenoid pigments. These were rated visually, at the time of extract- ion as 0, if no visible yellow was obtained in the ether ex- tract, Tr, if a trace of yellow pigment, and'+,++, or-r++ with increasing intensity of yellow pigment. The residual plastid pigments, extracted with isopropyl alcohol were found to be directly proportional to the intensity of the ether ex- tract. Partition test for carotenoids.-A simple test for the separation of carotene and xanthophylls is to partition the ether soluble pigment between petroleum ether and 80% methan- ol. Carotene, being more soluble in petroleum ether, remains in that phase, whereas xanthophyll will be accumulated in the 7‘ ” L _ . J . - . . .. u , . A. s . O . - I l _ 2 t . . . l . . . u . a . . v . . . . . . . . _ v . . . . . . I . . t . D t l \ a. . . v . u . U \ )i ..l\.. I w l9 mettumnol phase. however, if xanthophylls are esterified with fatty acids, which often may be the case in raw plant extract, this test can be misleading for the xanthophyll esters are retained in the petroleum ether phase, appearing to be carotene (Strain 1945). To prevent such an error in identi- fication, the raw petroleum ether extract was saponified for 15 minutes in alcoholic KOH* to break possible ester link- ages. To remove the hydrolyzed xanthophyll from the alcoholic K05, the solution was diluted in half with water and then shaken with an equal volume of petroleum ether. The xantho- phyll was partitioned into the ether phase where it was washed twice by shaking with an equal volume of 50% ethyl alcohol and once with water, to remove any residual KOH. This saponified pigment was then partitioned between petro— leum ether and 80% methanol to determine if the pigment was primarily xanthophyll or carotene. Adsorption chrgmatography.-Adsorption columns were pre- pared for separation of component pigments in the petroleum ether extract. Glass tubing, 1 cm in diameter was cut in 30 cm lengths and heated in a blow flame 2 inches from one end to form a constriction for support of a cotton plug. *Dissolve 10 gms of KOH in 5 m1 of water, cool, and then add 95 ml of absolute ethyl alcohol. 20 In all columns, the wet packing system described by Williams (1948) was used. For the separation of xanthophyll or xanthophyll esters from carotene, a column was prepared with the bottom 15 cm packed with a 2:1 mixture of Celite and Mgo (activated)*, the next 4 cm consisting of Ca (OH)2 and finally a two cm plug of Naasou. A highly concentrated solution of petroleum ether extract from 0.5 gm of g. falcata flowers was used and the pigment adsorption bands compared with that of B carotene from standard crystalline carotene. Xanthophylls and xanthophyll esters are adsorbed in the Ca (Ofi)2 layer while the carotene pigments move through the Ca (OH)2 into the MgO layer (Strain 19h2). For the resolution of xanthophyll components, a column of 1:1 Celite to activated MgO was prepared by wet packing under vacumn in dichloroethane (ethylene dichloride) Strain 1945). The saponified petroleum ether extract from 0.5 gm of E. falcata flowers was concentrated by evaporating the ether solution to near dryness and then dissolving in 3 m1 of dichloroethane. This solution was then added to the top of the Celite-mgO column and developed in dichloroethane. Analysis of Data.-The observed genetic ratios for pig- ment inheritance were tested against expected values, using the X2 method for determining goodness of fit. *Formerly Micron brand adsorptive magnesia, now call Sea Sorb hB, Fisher Scientific Company, Chicago 51, Illinois EXPERIMENTAL RESULTS Three anthocyanin and nine flavonol pigments were sep- arated from the .lN HCl extracts of diploid alfalfa flowers and identified by use of paper Chromatography. Carotenoid pigments were extracted with petroleum ether and were separated and identified on adsorption columns and by partition tests ( Figures 2a and 2b). Identification of anthocyanins Anthogyanin glycosides.-Three distinct anthocyanin glycosides were separated out on the paper chromatograms (Figure 2a). The Rf values for these pigments were determined in BAW and HAc-Hfll, and where available, the unknown pigment was cochromatogrammed (run side by side on the same chromato- gram) with an authentic compound, Based on these Rf values and comparison with known standards or previously reported Rf values, these three pigments were identified as delphidin 3,5 diglucoside, petunidin 3,5 diglucoside, and malvidin 3,5 diglucoside (Table 3). 21 in: (a) Purple (b) Variegated (c) Yellow (d) White Figure 1. Flower color types in diploid alfalfa (a)Anthocyanins(b)Anthoxanthins(c)B Carotene(d)Xanthophy11 Esters Fi ure 2. Paper chromatograms in BAW of (a) anthocyanins and (b anthoxanthins and Ca(OH) -Mg0 adsorption columns of (c) B carotene and (d) xanthophy 1 esters. fl at Table 3. RI. values of anthocyanins in BAW and HAc-HCl _/ Em W315) Delphinidin 3,5 diglucoside 0.15 0.15 0.15 0.11 Petunidin 3,5 Diglucoside* 0.21 - 0.24 - Malvidin 3,5 diglucoside 0.25 0.25 0.31 0.22 HAc-HCl (82:15:3 H20) Delphinidin 3,5 diglucoside 0.36 0.36 0.32 - Petunidin 3,5 diglucoside 0.43 - 0.32 — Malvidin 3,5 diglucoside 0.53 0.53 0.h2 - *Tentative identification (a) Harborne 1958 (b) Bate-Smith 1950 The delphinidin and petunidin glycosides formed a pur- ple band on the BAW chromatogram whereas the malvidin glyco- side was mauve in color. These observations were in agree- ment with those reported by Harborne (1958) for these three pigments. Although an authentic source of petunidin 3,5 di- glucoside was not available, the intermediate position of this pigment, between the delphinidin and malvidin glycosides, its color, and its Rf value in comparison to a previously re- ported Rf value in BAW (Harborne 1958), strongly supported this identification. Anthocyanin aglycones.-After hydrolyzing the purified anthocyanin glycosides 10 minutes in 2N HCl, three different aglycones were obtained. These were identified as delphini- din, petunidin, and malvidin, by cochromatogramming with authentic compounds using Forestal solvent and BAW as the developing solvents (Table 4). Table 4. hf values of anthocyanin aglycones in Forestal solvent and BAW Alfalfa Authentic Previous Values Extract Compounds (a) (b) (c) Forestal (10820:30HA0:3HCI) Delphinidin 0.32 0.32 0.32 '0.30 0.30 Petunidin 0.46 0.46 0.46 0.48 0.45 Malvidin 0.68 0.68 0.60 0.63 0.60 Cyanidin - 0.53 0.49 0.50 0.50 BAW (4:1:5) Delphinidin 0.36 0.36 0.42 - - Petunidin 0.48 0.48 0.52 - - Malvidin 0.58 0.58 0.58 - a Cyanidin - 0.65 0.68 - - ¥Run on acid washed paper (a) Harborne 1958 (b) Pecket 1958 (c) Bate-Smith 1954 The observed R values closely agree with those pre- f viously reported. Since Davies (1958) had reported that one of the anthocyanins in alfalfa was a cyanidin glycoside, the Rf values for cyanidin were also determined in the two solvent systems (Table 4). None of the three anthocyanins isolated from the alfalfa flowers gave Rf values in these two solvents corresponding to those of cyanidin. Identification of anthoxanthins Anthoxanthin glycosides.-Nine anthoxanthin glycosides were separated from alfalfa flower extract by paper chroma- tography in BAN and 15% HAc solvents. All but two glycosides had Rf values less than 0.5 (Table 5). of these Table 5. Br values of anthoxanthins in BAW and HA0 BAW 15% HAc Ascending Descending Ascendigg Authentic Compound Rutin (previous values)* 0.5? 0.57 0.62 Rutin (observed) 0.46 0.56 0.63 Quercetin glycosides A 0.15 0.12 0.69 B' 0.18 0.15 0.67 B 0.27 0.26 0.54 C' 0.29 0.28 0.59 C 0.33 0.34 0.65 D 0.36 0.42 0.51 Kaempferol glycosides 0.42 0.46 0.73 0.61 0.68 - G 0.76 0.79 - *Gage e_t_ L1 1951 26 27 These values are considerably lower than any Rf values previously reported for anthoxanthin glycosides. Since previously reported Rf values were obtained by descending chromatography this method was used as a check against the ascending method used in this Study. The des- cending method failed to significantly alter the Rf values 5" from those obtained by the ascending method (Table 5). ' The pigments were tentatively identified as flavonol )— or flavone glycosides based on their forming a bright yellow 5‘ appearance when exposed to ammonia vapors in visible light and a brown or brownish yellow appearance under ultraviolet light (Geissman 1955). Since there were no authentic flavone or flavonol compounds with such low Rf values for comparison, exact identity of the unknown anthoxanthin glycosides was not possible. However, each of the glycosides were charac- terized by their Rf values in the BAT and 15% HAc solvent systems (Table 5) and the aglycone of each glycoside deter- mined (Table 6). Anthoxanthin aglycones.-Each of the nine anthoxanthins were hydrolized ten minutes in 1N H01 and the aglycone par- titioned from the hydrolyzate with isoamyl alcohol. The six glycosides with the lowest Rf values (A thru D) were found to be quercetin derivatives while the other three (E thru C) were derivatives of the aglycone kaempferol (Table 6). These aglycones were identified by cochromatogramming with the authentic samples of quercetin and kaempferol. 28 Table.6. R values of anthoxanthin aglycones in Forestal sglvent Anthoxanthin Aglycone Authentic Compound (a) Quercetin A 0.51 0.52 B' 0.57 0.57 B 0.58 0.57 c' 0.60 0.60 C . 0.59 0.58 D 0.56 0.56 (b) Kaempferol E 0.76 0.76 F 0.74 0.75 0 0.73 0.73 The quercetin glycosides were yellow to orange yellow and formed less diffuse spots than the pale yellow kaempferol glycosides (Figure 2b). Iggntification of carotenoid pigments Xanthophyll Ester.-When the petroleum ether extract was added to a Ca(0H)2-Mg0 adsorption column and developed with petroleum ether, nearly all the yellow pigment was adsorbed in the Ca(OH)2 layer in a bright yellow band (Figure 2d), vindicating that the carotenoid pigment was primarily 29 xanthophyll (Strain 1945). A trace of orange yellow pig- ment moved through the Ca(0H)2 layer into the MgO zone and was identified as B carotene by comparison with an authentic sample of crystalline carotene (85% B carotene). The xanthophyll was eluted from the Ca(0.~i)2 layer into petroleum ether and shaken with an equal volume of 80% meth- anol. The pigment remained in the ether phase, indicating that the xantnophylls were in an esterified form. By hydro- lysis in alcoholic ROE, the ester linkages were broken and when the hydrolyzed pigment was partitioned between ether and 80% methanol, the pigment moved into the methanol phase, verifying its identification as xanthophyll. Separation of the hydrolyzed pigment on a column of MgO, using dichloroethane as the developing solvent, resulted in the development of at least two and possibly three separ- ate bands of xanthOpnyll. however, since no authentic sam- ples of xanthophyll were available for comparison, positive identification of these pigments was not possible. Segregation patterns of individual pigments The general procedure followed in the study of segre- gation patterns was to formulate a genetic hypothesis for inheritance of a given pigment based on the large 51 progen- ies of two variegated plants, 5346 (56 plants) and 183 (43 plants). This hypothesis was then further tested with the additional smaller families. A ' i _ ‘ ‘ I \ V. , , n V . .1 A i . , > . x ‘1 c ‘ A. . .- . .. x. k i . ". ' \ I . - . .. t 1. .L\, '3 . I . as. l ‘ ‘ x, . .. . u - 3 N Ls ~ . - L.\t 3O Anthocyanins.-The three anthocyanin pigments were in- herited as a unit with either all or none of the three pig- ments present. There was some evidence of differences in relative amounts of the three anthocyanin pigments, but no definite segregation pattern or phenotypic effect was ob- served. The ratio of plants having anthocyanin (P) to those without (p) fit a 3:1 ratio in the selfed progeny of two var- iegated plants and either a 1:1 or 1:0 in crosses of blue or variegated plants to yellow or white flowered plants. No anthocyanin containing progeny were obtained from crosses between yellow x yellow, yellow x white, or white x white plants (Table 7). 31 Table 7. Observed segregation for anthocyanin production 2 Cross P p X Probability 8346 (X7 1L0 16 183 (X) _3__1 12 Observed total 71 ?8 Expected (3:1) 74.3 2A.? 0.587 O.SO>’P7'O.3O H 7h X 63 13 9 f 3' X 183 5 8 (- 3gs x 22506319 15 9 l s X 73 S 5 3 Hh7h x 59-101-(2) 6 g .g 38 X 103 3% ' Observed total EH Expected (1:1) 47.5 47.5 0.501 O.SO7P70.3O to x 163 22 0 4'10 X 1&8 15 O 40 X 3 s 11 O to x 59-101-2 10 0 30 X 73 11 0 C73 X 7s 17 0 :Nb GD 15 _2 Observed total 101 O 383 x 59-101-2 o lO 383 X 8-2 0 12 as X 35 O 10 35 X falcata-l O 19 35 __c_>_ 8 Observed total 0 3? Kaempferol glycoside F.-Of the nine anthoxanthin gly- cosides, only kaempferol glycosides F and G exhibited segre- gation in enough families to permit genetic anlaysis. The ratio of plants containing kaempferol glycoside F to those without, (f), showed a good fit to expected 3:1 and 1:1 ratios (Table 8). Other families were non-segregating either for presence or absence of this pigment. 32 183 OD 183 X 73 Table 8. Observed segregation of kaempferol glycoside F 2 Cross F f X Probability 3356 ® MI- H474 x 63 17 40 1 38s 9 383 X 383 9 388 X 103 10 Observed total 8? Expected (3:1) 84.7 .3 0.87 0.507 P7'O.30 40 x 163 10 38 x 183 6 30 x 73 6 C73 X 78 32 Observed total 1 Expected (1:1) 31.5 .5 0.016 O.90>P70.80 0 _g 0 Observed total 40 x 14s 40 x 59-101-2 38s x 59-101-2 H474 x 59-101-2 3h3 X 33506319 35 X falcata-l #3 X 35 35 36 Observed total 1% OIOOOOOOOOO 335 fiflmmus wWWN-F‘RJJ Kaempferol glycoside G.-Segregation for kaempferol glycoside G is similar to that observed for glycoside F. The ratio of present to absent fit very well with expected values of 3:1 and 1:1. Also there were non-segregating families in which all plants either contained the pigment or none contained the pigment. 33 Table 9. Observed segregation of kaempferol glycoside G Cross G g X2 Probability 5346 3) 1+2 14 40 X 385 9 2 C73 X 75 12 5 38 X lOs _% 2% Observed total 7 Expected (3:1) 72.? 24.3 0.16 0.707 P7 0.50 383 X 8-2 5 7 H%74 X be 11 10 3 X 183 9 38 X33 1 :4 3} serve tota Expected (1:1) 2 .5 28.5 1.42 0.30) P7 0.20 183 GD 0 43 348 x 22506319 _9 £4 Observed total 0 7 40 X lbs 22 0 40 X 148 15 O 40 X 59-101-2 10 0 383 X 59-101-2 lO 0 H474 x 59-101-2 13 o 35 X falcata-l 19 0 48 X 35 10 O 35 ® 8 0 36 m g _2 Observed total 1 O Xanthophyll.-The xanthophyll pigments were extracted with petroleum ether and given a quantitative rating of o, Tr,+,++ and+++ , depending on the intensity of yellow pig- ment in the extract. The frequency distribution within these classes approached a 1:4:6:4:1 (+++:++:+~: Tr: 0) for the #3 51 progeny of clone 183, and a 1:2:1 (+:Tr:0) for the 56 S1 progeny of clones g346 (Table 10). 34 Table 10. Observed segregation of xanthophyll Cross YYYY YYY YY Y y X2 Probability 183 ® 0 10 20 12 2 4 Expected 2.6 10.8 16.2 10.8 2.6 4.35 0.507P>O.30 (1:4:6:4:1) s346 ® 13 26 17 Expected 14 28 14 0.85 0.707 P 7 0.50 1:2:1 Most families were too small for determining a fit to a five class distribution, but when 0 and Tr plants were grouped into one class and +,++Iand+++'p1ants into a second. class, a good fit to expected ratios was obtained. The 18s progeny gave a very good fit to a 11:5 (2+:Tr) ratio and the 8346 progeny fit at 1:3 distribution. _Other ratios observed for smaller families were 381, 1:1, and non-segregation for a11-+ or greater or all Tr or 0 plants (Table 11). 35 Table 11. Observed segregation of xanthophyll into combined classes Cross ZYY f Y X2 Probability 183 GD 30 13 Expected (11:5) 29.6 13.4. 0.017 0.957 P > 0.90 346cm _ 13 43 ED X 383 3 9 30 X 75 _3 8 Observed total 19 6'6 Expected (1:3) 19.8 59.2 0.043 0.007 P70.80 38 X 183 8 5 38 X 103 _2 _4 Observed total 17 2 Expected (3:1) 19.5 .5 0.82* 0.50>P>0.30 383 X 8-2 6 6 H474 X 63 11 10 183 X 73 :5 Observed total 1 Expected (1:1) 21.5 21.5 0.024 0.90>'P rO.80 40 X 59-101-2 10 O 3484 X g9-101-2 10 O H 7 X 9-101-2 13 0 343 x 225061319 24 o 35 X falcata-l 1% _Q Observed total 7 0 40 X 163 0 22 40 x 143 0 15 C73 X 73 O 17 X 35 0 8 35 0 10 36 _9 %5 Observed total 0 7 *Yates correction for sample size less than 40 and with 1 degree of freedom. There was some evidence that the level of refinement of the extraction procedure permitted the occasional error of placing some Tr plants into the O classification resulting in an excess of 0 type plants. However, grouping the Tr I I. .l' I r a I ‘- a fi. .. .1 a . ..5 r3 36 and 0 plants into one class overcame this difficulty, re- sulting in a good fit to expected ratios. Joint segregation of pigments In order to detect possible linkage relationships in the inheritance of the pigments studied, their joint segre- gation patterns were analyzed. Because of the multiple classes involved, only the large 8 progenies of clone g346 l and 183 were used for these analyses. Anthocyanin and Kaempferolpglyposide E.-The joint segre- gation for the presence or absence of anthocyanin (P) and kaempferol glycoside F gave a very good fit to the 9:3:3:1 ratio expected from independent inheritance of these two pigments (Table 12). Table 12. Joint segregation of kaempferol glycoside F. and anthocyanin P. Cross F-P- F-pp ffP- ffpp Total e346 ® 31 12 10 3 , 56 Expected(9:3:3:1) 31.5 10.5 10.5 3.5 x2 - 0.195 D.F. = 3 0.987P70.95 Anthocyanin and kaempferol glycoside G.-The observed ratio fit very closly the 9:3:3:1 ratio predicted on the assumption of independent inheritance of anthocyanin produc- tion and kaempferol glycoside G (Table 13). <2. .A‘I 37 Table 13. Joint segregation of kaempferol glycoside G. and anthocyanin P. Cross G-P- G-pp ggP- ggpp Total 8346(3) 30 12 11 3 56 Expected 31.5 10.5 10.5 3.5 X2 = 0.380 D.F. = 3 0.95>P790 Kaempferol glycosides F and G.-A good fit to the ex- pected 9:3:3:l dihybrid ratio was obtained for the joint segregation of the two kaempferol glycosides F and G (Table 14). Table 14. Joint segregation of kaempferol glycosides F. and G. Cross F-F- F-gg ffG- ffgg Total g31i6® 32 10 12 2 56 40 2: 38s 6 2 2 1 11 Total observed 33 T2 TE. 3 '57 Expected (9:3:3:l)37.8 12.5 12.5 4.2 x2 = 0.551 D.F. = 3 0.95>P70.90 Anthocyanin, kaempferol glycosides F and G.-In the 56 81 progeny of clone gBAb, a very good fit was obtained to the predicted trihybrid ratio 27:9:9:9:3:3:3:l, based on in- dependence of inheritance of anthocyanin production (P), kaempferol glycoside F, and kaempferol glycoside 0 (Table 15). m‘ .m‘. .- —1 n 38 Table 15. Joint segregation of kaempferol glycosides F and G and anthocyanin P. Cross PFG PFg PfG pFG Pfg ng pr pfg g346® 22 9 8 9 2 3 0 Expected 23.6 7.9 7.9 7.9 2.6 2.6 2.6 .875 (27:9:9:9:3:3:3:l) x2 = 1.84 D.F. = 7 0.98>P>o.95 Xanthophyll and anthocyanin.-Progenies were too small to give a valid approximation to the ten possible (l:4:6:4:l K 3:1) or 6 possible (1:2:1 X 3:1) classes. Also, the occas- ional misclassification of plants containing trace amounts of xanthophyll (Tr) as (0) type plants caused marked devia- tions from predicted joint segregation. However, by group- ing the xanthophyll classes into those with at least two plus factors (ZYY) for xanthophyll production and those with one or none (SY), a good fit was obtained to the 6.6: 3:2.2:l ratio (ie 11:5 K 3:1) expected, assuming independent inheri- tance of xanthophyll and anthocyanin production (Table 16). Table 16. Joint segregation of xanthophyll and anthocyanin Cross ZYYP éYP ZYYp éYp X Probability 183 ® 21 1o 9 . 3 Expected 22.1 10 7.4 3.4 0.829 0.90) P'>O.80 (6.6:3:2.2:1) g3 6® 9 31 4 12 Exgected 10.5 31.5 3.4 10.5 0.50 0.957’P70.90 (3:9:l:3) 39 Xanthophyll and kaempferol glycoside F.-Using the com- bined classes for xanthophyll, the joint segregation of xanthophyll and kaempferol glycoside F approximated the 3:9: 1:3 ratio expected assuming independent inheritance of these two pigments (Table 17). Table 17. Joint segregation of xanthophyll and kaempferol glycoside F Cross ZYYF {YF ZlYf .éYf X2 Probability 23469 13 30 o , 13 Expected 10.5 31.5 3.5 10.5 4.75 0.20>P70.10 (339:1:3) _fi Xanthophyll and kaempferol glycoside G.-The predicted 3:9:l:3 ratio, assuming independent inheritance was closely approximated by the joint segregation of xanthophyll and kaempferol glycoside G (Table 18). Table 18. Joint segregation of xanthophyll and kaempferol glycoside G Cross zyye 5Y0 ZY‘ig 513 x2 Probability £3346 ® 12 3o 1 13 . Expected 10.5 31.5 3.5 10.5 2.66 0.407 P70.30 (3:9:lz3) 40 Additional evidence of segregation Anthocyanin intensity.-In certain crosses, there was marked variability in anthocyanin intensity. However, be- cause of a mixture with anthoxanthin pigments in the raw ex- tract and the separation of the anthocyanin pigment into three distinct bands on the paper chromatograms, it was difficult to give a reliable visual rating for anthocyanin intensity pg; g3. In some plants the relative amounts of the three pigments seemed to vary, but in general there ap- peared to be a rather constant equilibrium between the con- centration of the three pigments. There was not enough data to formulate a definite inheritance pattern, but the differ- ences in anthocyanin concentration could account in part for some of the phenotypic variation within the proposed geno— typic classes. Bud color.-Among the yellow and white flowered plants, there appeared to be segregation for a trace of anthocyanin in the buds (pink buds). In nearly all crosses having yellow or white progeny, both pink budded and yellow or white budded types appeared and in about equal numbers. There was some evidence of an environmental effect on penetrance, however. In field classification of pink budded versus non-pink budded plants, it was observed the same plants which had pink buds on the first reading, failed to show the pink buds on a later reading or vice versa. Also, one of the white flow- ered plants which had white buds all summer in the field, developed pink buds when brought into the greenhouse. Because of this possible penetrance effect and the fact that the maximum number of yellow flowered plants in any one fam- ily was only 16, no genetic hypothesis for the inheritance of this trait was proposed. Intensity of quercetin_glycosides.-In most families, all plants contained the quercetin glycoside A, but in a few fam- ilies segregation for presence or absence of this pigment oc- curred. In one cross, 40 X 168, neither of the parents nor their 21 prOgeny contained this pigment. In other crosses there was evidence of quantitative differences in the amount of glycoside A among the progeny but no segregation for pre- sence or absence of this pigment. Also, there seemed to be a correlation between the intensity of quercetin glycoside A and the amount or presence of quercetin glycosides B', B, and C'. In the progeny from crosses containing a E. falcata par- ent, the glycoside A was usually intense and the quercetin glycosides B', B and 0' were also present. In most other plants, quercetin glycoside A was present in the complete absence of or only trace amounts of B', B, and C'. In no plants, however, were the glycosides B', B and 0' found in the absence of A. The quercetin glycosides C and D, along with kaempferol glycoside E, were found in every plant examined, including 18 white flowered plants. however, when the intensity of querce- tin glycosides A, C, and D, in the white flowered Sl progeny of clone 35, were compared with the intensity of these pig- ments in yellow flowered progeny from a cross between clone 35 and fl. falcata, there was a marked higher intensity in the yellow flowered plants. Phenotypic effects of pigments Anthocyanins.-The anthocyanins impart the reddish-blue or purple color to alfalfa flowers. The apparent wide range in anthocyanin intensity results in a corresponding range in flower color. In flowers, hOWever, the phenotypic effect of anthocyanins are also modified by anthoxanthin copigments and the background effect of the yellow xanthophyll pigment. No definite correlation between a certain phenotype and cer- tain quantitative balance of the three anthocyanin pigments was observed. Flower color ranged from very light blue, with only a small amount of anthocyanin, to dark blue or purple in wnich the concentration of the anthocyanin was considerably higher. In some yellow and white flowered plants, a trace amount of anthocyanin was evident in the buds, giving a pink color. As the flowers opened the trace of anthocyanin disappeared. In plants with variegated flowers, the buds or freshly opened flowers contained considerably more anthocyanin than flowers several days old. The decreased anthocyanin inten- sity, resulting from aging, permitted the yellow xanthophyll pigment to show through. In this transition period from nearly purple to nearly yellow, the variegated flower types are observed, and at a certain ratio of blue to yellow, a green phenotype often occurs. This green gradually fades out to a smudgy yellow in many flowers. Quercetin glycosides.-There was some evidence that higher concentrations of quercetin glycoside gave a pheno- typic effect. In plants containing only a trace amount of xanthophyll pigment, there were two levels of yellow inten- sity. Quercetin glycoside A was absent in the lighter yellow flowers but present in the more intense yellow flowers. At higher concentrations of xanthophyll, no marked phenotypic effect of the quercetin glycosides was evident. Although the white flowered plants contained quercetin glycosides A, C, and D, there were no phenotypic effects. However, when these flowers were placed in ammonia vapors, they turned yellow indicating presence of flavonol pigments. No "true white" flowers were found in which the ammonia vapor test was negative. Yellow flowers and variegated flowers, how- ever, contained considerably nigher amounts of quercetin glycosides than did white flowers. fl. falcata clones or prOgeny of a cross with m. falcata parentage were notice- ably higher in the concentration of quercetin glycosides. Additional evidence of the phenotypic effect of querce- tin glycosides is the appearance of a trace of light yellow color in blue flowers which do not contain xanthophyll pig- ment. These flowers have a slight bluish green appearance, but do not fade out to the smudgy yellow of variegated flow- ers containing xanthophyll. gggmpferol glycosides F and G.-Although the pale yellow color of kaempferol glycosides did not appear to impart a Kaempferolgglycosides F and G.-Although the pale yellow color of kaempferol glycosides did not appear to impart a direct yellowing effect on the phenotype, there was some evidence for copigmentation with anthocyanin, particularly for glydoside G. The maroon or red wine phenotype nearly always was associated with the absence of glycoside G and the presence of at least one factor for xanthophyll. In the non-xanthophyll plants (yyyy), presence of both kaempferol glycosides F and G was associated with blue flowers whereas absence of F or G or both was associated with reddish blue types. These latter two classifications were not absolute since some reddish blue and blues occurred in the same geno- types. However, the relative frequency of the reddish blue types in comparison to blue was considerably higher in plants with f, g or fg genotypes. Xanthophyll.-The concentration of xanthophyll in the flower extract appeared to be directly correlated with the intensity of yellow color in the flowers. The five levels of xanthophyll pigment intensity,+++ ,f+-,+, Tr, and 0 had four corresponding phenotypic classes of orange yellow, bright yellow, yellow, light yellow, and white. However in plants containing a trace (Tr) level of xanthophyll, two different yellow intensities were observed, indicating the presence of other yellow pigments with phenotypic effect. In plants containing‘F or greater amounts of xanthophyll, only one phenotype per level of xanthophyll was observed. '45 In plants containing both anthocyanins and xanthophyll, the xanthophyll has an important background effect on the phenotype of the flowers. Plants high in both anthocyanin and xanthophyll pigments appeared very dark purple to almost black. In other types the yellow background effect produced a maroon or reddish wine color. As the anthocyanin fades on aging, the yellow xanthophyll becomes more evident and the phenotype change from purple to a variegated type. Each level of xanthophyll produces a somewhat different phenotypic effect in combination with anthocyanin, but because of the continuous variation in the degree that the anthocyanin had faded, specific genotypic-phenotypic relationships were difficult to establish. High intensity of xanthophyll pig- ment produces an orange-yellow background effect in varie- gated flowers, while at lower xanthophyll intensities, the background color is yellow to pale yellow. DISCUSSION Because of the difficulty in collecting flowers, each sample of flowers was used for the extraction of both caro- tenoids and flavonoid pigments. In most flowers, all the carotenoid pigment was removed by the petroleum ether extract, but in plants containing considerable amounts of carotenoids, a yellow residue was observed following the ether and acid extractions. This yellow pigment could readily be removed by isopropyl alcohol, leaving a white residue. The amount of residual pigment extracted by the isopropyl alcohol was directly proportional to the intensity of yellow pigment obtained in the ether extract. Since, in this study, the relative amounts of pigment from plant to plant were more important than the absolute amount it was not deemed nece— sary to combine the carotenoid extracts. Ascending chromatography was used in preference to des- cending chromatography for this study. In preliminary tests, ascending chromatography in the 15" test tube gave better separations of pigments and more consistent results than descending chromatography in a large glass chromatogra- phic jar. An additional advantage of using the test tubes was that the solvent could be used for several runs whereas with the descending method, it was necessary to remove the solvent after each run and clean the equipment. Also the 46 #7 small air volume of the test tubes permits rapid saturation of the atmosphere by the volatile solvent, eliminating an equilibrating period often needed when larger containers are used. The simplicity of both set-up and use, with good sep— arations and consistent results made this procedure part1- cularly well suited for the'chromatographic analysis of large numbers of plants from segregating populations. Because of the dilute concentration of pigments obtain- ed, it was necessary to concentrate the pigments by repeated spotting of the chromatogram. The light from the mimeoscope used was beneficial in locating spots for repeated applica- tion and speeding up evaporation of the solvent. The glass rod used in spotting the chromatograms gave fairly consis- tent size drops at each application and was satisfactory for the purpose of this study. However, for a strictly quanti- tative study micropipettes should be used for spotting chro- matograms. Identification Anthocyanins.-Three anthocyanin pigments were found in all purple flowered plants examined and were identified as the 3,5 diglucosides of delphinidin, petunidin and mal- vidin. These findings are in agreement with Lesins (1956) who found delphinidin, petunidin and malvidin glycosides in both tetraploid and diploid alfalfas. Davies (1958), using Forestal solvent, reported similar results with the except- ion of finding cyanidin instead of petunidin. Because of 48 this discrepancy in the literature, both petunidin and cyan- idin authentic samples were used in this study. Rf values of these two pigments were very close in Forestal solvent but their marked separation in BAW solvent positively iden- tified the anthocyanidin from alfalfa as petunidin. The small difference between the Rf value of cyanidin and of petunidin in Forestal solvent (the only solvent used by Davies) suggests the possibility of a missidentification. The occurrence of anthocyanin mixtures derived from delphinidin and its methylated derivatives, petunidin and malvidin is not uncommon and has been reported in several species (Lawrence 23 El 1939). Beale (1939) first reported this combination of pigments in Lathyrus odoratus (sweet pea) and more recently Pecket (1960) in a survey of Lathyrus species, found several species containing mixtures of delph- inidin, petunidin and malvidin glycosides. Lesins (1956) reported the separation of four anthocy- anins and a trace of the aglycone delphinidin from raw ex— tract of alfalfa flowers. The anthocyanins were identified as two delphinidin glycosides, a petunidin glycoside and a malvidin glycoside. These results differ from those reported here by the presence of the aglycone delphinidin and a second glycoside of delphinidin. This could be due to actual dif- ferences in plant material or differences in technique used. guercetin.-Six quercetin glycosides were identified based on their characteristic yellowing in ammonia vapor, A9 brownish yellow color under ultra violet light, and posi- tive identification of their aglycone as quercetin. These pigments ranged in Rf values from 0.12 for quercetin glyco- side A to 0.25 for quercetin glycoside D. All of theseRf values are considerably below any previous reported values for quercetin glycoside (Bate-Smith 1950, Gage and Wender 1951). A satisfactory explanation for these observations is difficult. when the chromatographic procedure used (BAH, ascending) was checked against and authentic sample of rutin (quercetin 3-rhamnodiglucoside), the observed Rf values were only slightly below previously reported Rf values. Also this same procedure was used for anthocyanin separations and gave good approximations to Rf values reported for these pig- ments. Altering the technique by using descending chromato- graphy in a large chromatographic jar and equilibrating the paper 24 hours with the lower phase of the BAW solvent (as used by Bate-Smith 1950) failed to significantly alter the Rf values. Extraction procedure, paper and solvent used were similar to those used by previous investigators. There- fore, the possibility that the low Rf values obtained were due to differences in technique does not seem likely. The basic 015 molecule, common to all flavonoid pig- ments, is elaborated into different classes by the processes of hydroxylation, methylation, glycosidation and acylation (Lawrence and Price 1940, Harborne 1956). In a study of .‘1 5 . RT"? 50 the relationship between the structure of flavonoid mole- cules and their Rf values in EAW, Bate-Smith and Westall (1950) observed that in general, as the number of hydroxyl groups and/or number of glucose units increase, Rf is decreased, whereas methylation and acylation increase the Rf values of the basic pigments. Since the hydroxylation and methylation patterns are determined at the aglycone level and acylation increases Rf values, the only structural modification that can explain the low Rf values obtained is different glycosidation pat- terns. Anthoxanthins in general have considerably more gly- cosidic variability than anthocyanins. At least five glyco- sidic derivatives of quercetin have been reported in the li- terature. In the BAW solvent, these pigments range in Rf values from 0.57 for rutin (quercetin-3-rhamnoglucoside) to 0.82 for quercetrin (quercetin-3-rhamnoside) (Gage and Wen- der 1951). The only yellow to orange flavonoid pigment re- ported with an Rf value within the 0.12 to 0.28 values ob- tained in this study was a glycoside of aueresidin, an orange aurone pigment found in Antirrhinum ggigg (Geissman 1954). Chromatogramming of an extract from yellow flowers of snap- dragon verified this report, but the bright orange color of this pigment in ammonia vapor readily distinguished it from the bright yellow color of the quercetin glycosides obtained from alfalfa. 51 An alternative explanation for such low Rf values is the possibility that the quercetin glycosides occur in a complex molecule in alfalfa. The analytical procedures needed to deter- mine such a complex, or the chemical structure of the glyco- sides beyond the identification of their aglycone, was beyond the scope of this study and must await further investigation. Kaempferol.-Three kaempferol glycosides were identified by their characteristic color reactions in ammonia vapor and under ultra-violet light, and by positive identification of their aglycone as kaempferol. The kaempferol glycosides formed pale yellow spots on the chromatograms, usually visible only when intensified in ammonia vapor. This was in contrast to the brighter yellow of the quercetin glycosides. The Rf values of the three kaempferol glycosides in BAW solvent were 0.35, 0.54 and 0.71. Two previously reported Rf values in BAW were 0.51 for robinin (kaempferol-B-dirhamnoga- lactose) and 0.75 for kaempferitrine (3—dirhamnoside). The appoximation of these two Rf values, by the kaempferol glycosides F and G respectively, suggests a tentative identification of these pigments. However, since authentic samples were not avail- able for comparison, these tentative identifications could not be verified. At least three other kaempferol glycosides have been reported in the literature, 3-glucoside (astragalin), 3-rhamnoglucoside, and 3-rhamodiglucoside, but their Rf values in BAW had not been determined (Roberts 1956). The possibility exists, however that one of these corresponds to the third un- identified kaempferol glycoside. 52 Of interest is that robinin (kaempferol-3dirhamnoga- lactose, sometimes called kaempferol-3-robinoside, was first identified from the flowers of fipbinia pseudoacacia (Honey locust, a legume tree) and that kaempferitrin (kaempferol-3- dirhamnoside) as well as other kaempferol rhamnosides were obtained from Acacia linifolia, and Acacia decurrens, two members of a large genus of flowering trees and shrubs belong- ing to the legume family (Sannie 1952). If the tentative iden- tification of these pigments is correct, this suggests a pos- sible taxonomic significance of these kaempferol glycosides. Carotenoids.-The carotenoid pigments in alfalfa flowers were identified as primarily xanthophyll esters with a trace of B carotene. This identification was based on partition tests between petroleum ether and 80% methanol, before and after saponification, and their differential adsorption in a Ca(0d)2-Mg0 column. Xanthophyll esters behave as carotene in the partition test, remaining in the ether phase, and could be mistaken for carotene if this were the only test used. This is a possible explanation of the previous report (Twamley 1955) that the carotenoid pigment isolated from alfalfa was primarily car- otene. The xanthophyll pigment was separated into two and pos- sibly three components on a mgO column, using dichloroethane as the developing solvent. Lack of authentic samples of xantho- phylls did not permit identification of these components. However, the traces of B carotene found suggest that these xanthophyll esters may he hydroxylated derivatives of B carotene. One of 53 the components, with the least adsorbancy on the MgO column was pale orange in color, approximating the color of zeaxanthin (3,3'-dihydroxy B carotene), in Strain's (1945) color plate. Zeaxanthin and two other B carotene derivatives, cryptoxanthin (3-hydroxy B carotene) and eschscholtzanthin (3,3'-dihydroxyde- hydro B carotene) have been isolated from flowers (Goodwin 1955). Numerous other carotenoid pigments have been isolated from flow- ers, however, so that the above possibilities are not conclusive. 5(4- Pigment inheritance Anthocyanins.-Production of anthocyanin was controlled by a single dominant gene. The absence of complementary gene action as reported by Twamley (1956) and others is probably due to homozygosity of one of the complementary factors for antho- cyanin production. The three anthocyanin pigments identified were inherited as a unit, all three being present in every plant contain- ing anthocyanin. These observations were similar to those re- ported by Lesins (1956) with the exception that in two tetra- ploid crosses, segregation for the malvidin glycoside occurred. However, it was noted that malvidin was only present in the plants which also contained delphinidin and petunidin deriva- tives. Since the only anthocyanidins obtained from alfalfa flowers were trihydroxylated in the 2 phenol ring of the benzopyrilium nucleus, alfalfa must be homOZygous for a gene or genes determin- ing the trihydroxylating pattern. In addition, since every plant with anthocyanin, contained delphinidin plus its two methylated derivatives, petunidin and malvidin, alfalfa must also be homozygous for a methylating gene or genes. The occur- rence of all three anthocyanidins in a mixture has been attri- buted to incomplete methylation by Lawrence 23 51 (1939). According to Lawrence, this methylating process may be either the straight forward methylation of delphinidin, or of a 55 precursor which in the absence of complete methylation gives rise to delphinidin. The incomplete methylation theory seems applicable to the situation in diploid alfalfa and in general to tetraploid alfalfa, with the possible exception of a gene segregating for malvidin production in tetraploid alfalfa (Lesins 1956). Genes determining the glycosidation pattern for antho- cyanins appeared to be in a homozygous condition, and there was no evidence of acylating genes. Since anthoxanthins were found in every plant examined, the gene affecting the production of anthocyanin must exert its influence in the biosynthetic pathway after formation of the C15 precursor, which is common to both anthocyanin and anthoxanthins. Whether this also applies to both complemen- tary factors reported in diploid alfalfa by Twamley (1955), and reported in tetraploid alfalfa by various investigators, depends on the finding of a "true white" plant, indicating a block preventing formation of 015 nucleus and hence of all flavonid pigments. "True white" flowers are defined as white flowers which fail to turn yellow in ammonia vapor (Bate-Smith 1955). If even a trace of anthoxanthin is present, flowers will turn a pale yellow. Some supposedly "true white" tetra- ploid alfalfa flowers, obtained from Clements (1961), turned pale yellow in the ammonia vapor test and on chromatogramming showed at least two of the quercetin glycosides (A and D). Based on these observations it appears doubtful, that "true whites" in 56 the sense used by Bate-Smith (1955) exist in alfalfa. guercetin.-Two of the six quercetin glycosides were pre- sent in every plant examined. The four other glycosides ex- hibited segregation in some crosses, but data was insufficient to formulate a definite inheritance pattern. It was observed that quercetin glycoside A (fig. 2b) was present in nearly all plants. However, a few plants were ob- tained in which this glycoside was absent. When such plants were crossed, none of their progeny contained this pigment. This suggested that the absence of quercetin glycoside A was controlled by a gene or genes in the homozygous recessive con- dition. Additional observations indicated that the intensity levels of glycoside A could be divided into at least four classes ( O, Tr, + ,-++-), and that as the intensity of quercetin glycoside A increased, there was also an increase in intensity of the other quercetin glycosides. Quercetin glycosides B', B, and C' were very dilute pigments and were found only in plants contain- ing fairly intense levels (2+) of quercetin. The simultaneous increases in intensity of all six quercetin glycosides suggests an intensifier gene or genes for the aglycone, quercetin. From these observations, two hypothesis for inheritance of quercetin glycosides have been proposed, (1) the inheritance of quercetin glycosides is controlled by quantitative factors at two loci, affecting the intensity of the aglycone, quercetin and segregating l:4:6:4:1. An alternative hypothesis (2) is that production of quercetin glycoside A is controlled by a 57 single dominant gene and that quercetin intensity is control- led by a single quantitative factor segregating 1:2:1. The presence of quercetin glycosides A, C, and D in white flowers, indicated that not all levels of intensity produce a phenotypic effect. To test the hypotheses for quercetin inheritance and to determine genotypic-phenotypic relation- ship, crosses should be made between non-xanthophyll (yyyy) yellow flowered plants and the distribution of yellow inten- sity types observed in the segregating generations. In this manner, the epistatic effect encountered in this study re- sulting from the presence of yellow xanthophyll pigments, can be eliminated. Kaempferol.-One of the kaempferol glycosides, E, was present in every plant examined while the other two, F and G exhibited independent segregation in a 3:1 ratio. Thus, it was concluded that production of each of the kaempferol glycosides F and G was controlled by a single dominant and independent gene. Since the three kaempferol glycosides differ only in their glycosidic pattern, it suggests that genes F and G must produce their effect by determining the kind and/or position of glycoside units on the kaempferol molecule. Glycosidation genes have been reported for anthocyanins (Law- rence 1940), but to the author's knowledge, no specific ex- amples of gene action controlling the glycosidic nature of 58 kaempferol have previously been reported. Determining the exact action of the genes involved here, depends on positive identification of the kaempferol glycosides, which was not possible in this study. Carotenoids.—The carotenoid pigments, identified pri- marily as esters of xanthophyll, segregated into a l:h:6:4:l ratio for intensity in the 43 51 progeny of a variegated flowered clone 183. Based on this evidence plus supplemen- tary evidence from twenty other crosses, it appeared that xanthophyll inheritance in flowers of alfalfa is controlled by quantitative factors at two loci, ( Yllegyg ). Of particular interest is the close agreement between the observed segregation for xanthophyll and quercetin pig- ments in diploid alfalfa, and the hypothesized segregation for these pigments proposed by Twamley (1955), who based his hypotheses on the segregation of yellow color intensity in 65 yellow F2 plants from a purple x yellow diploid cross. That a similar inheritance pattern occurs for yellow pigments in tetraploid alfalfa is indicated by the 13 inten- sity classes Twamley was able to discern in 257 F2 plants of a yellow x white tetraploid cross. Earlier reports of a single factor Y for yellow pigment production in tetraploid alfalfa could be an artifact from the bulking of all purple and variegated plants and of all gradations of yellow into single classes. The 38:lO:l (purple:yellow:white) ratio 59 obtained from the F2 of a purple times yellow tetraploid was explained as approximating a 12:3:1 ratio (Lepper and Odland 1939). However, the observed ratio of 48 non white: 1 white, more nearly approaches the 63:1 disomic ratio expected based on two quantitative genes segregating for yellow and a 3:1 segregation for anthocyanin production. Under this hypothe- sis, all of the blue plants with exception of 3/64 would con- tain at least one factor for xanthophyll production. Weihing's (1948) observations of F2 family segregation from a white times yellow tetraploid cross is also in accord with a two factor hypothesis for xanthophyll produCtion. Joint segregation.-The joint segregation data indicated that genes segregating for the production of anthocyanin, kaempferol glycosides F and G, and xanthophyll were indepen- dently inherited. Based on the segregation patterns of these genes in 21 families, genotypes were proposed for the 23 par- ent clones. These genotypes together with observed pheno- types are listed in table 19. In some crosses, there was some indication that intensity of quercetin glycosides was associated with intensity of xanthophyll, but this could be due to the homozygous condition in E. falcata of the plus factors for the production of each of these pigments. Other factors.—A1though a quantitative study of antho- cyanin intensity was not made, marked differences were appar— ent. It was noticed however, that some of these apparent dif- ferences were actually due to the presence of anthoxanthins 60 Table 19. Phenotypes and genotypes proposed for parental clones Parental Phenotype Genotype General' Hort. Color Chart* 15p* reddish blue Imperial purple 33 PPffFFylylygye 3p"" variegated Pod green 61 be‘_G_Ylle2y2 g3h6 variegated Pod green 61 PprGleylyZy2 | P 188 variegated Pod green 62/2 prfrngyl‘i'zY2 )fii H474 maroon Violet purple 733/1 Pprgngyly2y2 30 maroon Imperial purple 33/1 PPngngylyZy2 D n to light blue Mineral violet 635/2 PPFngylylyay2 lOs blue-yellow Orchid purple 31/3 PprGngle2y2 073 blue Spectrum violet 735 PPFngylylyEy2 36 blue green Lavender violet 637 PPFFGGylylyZy2 3&3 violet Spectrum violet 735 PpPFggylyly2y2 383 yellow Sulphur yellow 1/2 pprGngleZy2 38 yellow Sulphur yellow 1/1 pprGleYly2y2 68 yellow Sulphur yellow 1/2 pprGngleEy2 73 light yellow Dresden yellow 64/2 ppffdgilyly2y2 143 light yellow Dresden yellow 64/2 ppEEGeYlylyeyg l6s** pale yellow Primrose yellow 601/3 pprgngylyay2 o-2** pale yellow Primrose yellow 601/3 pprgngylygyE 59-101-2 orange yellow Lemon yellow 4 ppFFGGYlYlY2Y2 falcata-l orange yellow Lemon yellow h ppffggElYlY2Y2 22506B19 orange yellow Orange yellow pprgngYlYgYZ 35 white White PPFFGGY1Y1y2Y2 43 white White PPffGGY1y1Y2Y2 lghOriginal Parents: 15p= g. sativa; 3p= (g. gaetula X‘fl. falcata) ‘aabiffer from light yellow types by absence of quercetin glycoside A _”,n' 61 causing blending or copigmenting effects, or to the back- ground effects of the xanthophyll pigments. The inverse relationship of anthocyanin production ver- sus anthoxanthin production, as proposed by Robinson (1931) was not especially evident. In fact, very definite contra- diction of this hypothesis was obtained by finding many very ’ pale blue flowers with very small amounts of anthoxanthin. * Because of the blending and copigmentation effects of anthoxanthins, it would be desirable to separate the antho- F xanthins from the anthocyanins by exhaustive extraction of F the acid extract with ethyl acetate (Bate-Smith 1950), be- fore quantitative measuring of anthocyanin intensity. This was not done by Twamley (1955) and, thus, could account for some of the variability observed. All 14 families which contained yellow and white pro- geny segregated for a trace of anthocyanin in the buds and in general approximated a 1:1 ratio. One exception was a 3:1 segregation for presence of pink buds in the 16 non-purple progeny of g346 A. However, because of the small numbers of non-purple plants in each family, and possible penetrance effect, influenced by environment, a definite genetic hypo- thesis was not possible. Phenotypic correlation Anthocyanins.-The presence of the gene for anthocyanin production is expressed in the light blue to purple color of alfalfa flowers. Anthocyanin intensity ranged from very di- lute, almost white flowers to deep purple. It was evident, I. I.II| 62 however, that much of the phenotypic variability in purple color was due to the blending or copigment effects of antho- xanthins and background effects of xanthophyll. Pink buds on yellow or White flowers have been observed in many genera and usually are attributed to the accumula- tion of anthocyanin diluter genes (Paris 33 31 1960). Accord- ing to Lawrence (1940) such genes produce their effect by al- tering the rate of pigment synthesis. This serves as a possi- ble explanation of the penetrance effect of pink buds in alfal- fa and points out the need to adjust the environment to permit expression of these effects in an inheritance study. Kaempferol.-The dilute yellow color of kaempferol gly- cosides, and the presence of all three in white flowers in- dicate these pigments impart little or no yellow color to alfalfa flowers. There is evidence, however, that these pigments produce an important phenotypic effect by copigmen- tation with anthocyanins. Of 30 plants with a maroon phenotype, observed in the segregating progenies, none were found which contained both kaempferol glycosides F and G. Seven contained only glyco- side F, three only glycoside G and twenty contained neither glycoside F or G. In addition, all maroon phenotypes con- tained at least one plus factor for xanthophyll. In plants having no xanthophyll, there was a significantly greater number of reddish blue flowered plants in the Fg, f0, and fg 63 classes than was found in the F0 genotypes. There was some overlapping of phenotypes between these classes, however. Thus, based on these observations, it is proposed that glycosides F and 0 function in a copigment action, complex- ing with anthocyanins to give a bluing effect. This would explain the predominance of reddish blue types in the fg genotypes and the low frequency of these types in F0 geno- types. Likewise, the maroon phenotype is a result of the combination of reddening effect by absence of kaempferol glycoside and the background effect of xanthophyll. Quercetianhe exact phenotypic effects of the querce- tin glycosides are difficult to determine because of the epistatic effect of the yellow xanthophylls. however, the appearance of traces of yellow in some phenotypes absent for a plus factor for xanthophyll indicated a phenotypic effect. Additional evidence was observed in the segregation of yel- low color intensity within a single genotypic class for xan- thophyll. The presence of quercetin glycosides A, C and D in white flowered plants indicates that there must be quanti- tative differences in the amount of quercetin pigment and that only when the concentration is above a certain level, do quercetin glycosides exhibit a phenotypic effect. Xanthophyll.-In general, there was a very high corre- lation between the intensity of xanthophyll pigment and the yellow in the flowers. It appears that xanthophyll is the most important yellow pigment in determining the yellow 61+ in alfalfa flowers. no white plants were found which con- tained xanthophyll, and all intense yellow flowers, especi- ally fl. falcata flowers contained a high intensity of xantho- phyll pigment. At a low intensity of xanthophyll pigment there was some indication of a phenotypic effect from quercetin glycosides. At higher intensities however, xanthophyll may be epistatic to the effect of the quercetin pigment. It was also evident that most of the variegated types observed were a result of xanthophyll background effects and the fading of anthocyanins. No variegated types were observed in plants which were absent in xanthophyll pigment. In a cross of g. falcata x E. sativa, the progeny were very dark to al- most black in color, exhibiting the color subtraction pheno- menon mentioned by Lesins (1956). Based on the information obtained from this study, a proposed inheritance chart for flower color in diploid alfal- fa is proposed (Figure 3). Figure 3. Flower Color Inheritance in Diploid Alfalfa YYYY P F G dark variegated 1 F G orange yellow P I F G purple variegated 'YYy I F G bright yellow P P G purple variegated YYyy maroon variegated 8 G purple variegated f I maroon variegated 8 F G yellow P P F G purple variegated YYY I maroon 8 G purple variegated f l maroon r 8 F G pale yellow P P F G blue yyyy reddish blue g G blue & reddish-blue f I reddish-blue 8 F G white P Y gives xanthophyll In Addition: P gives anthocyanin l) diluter genes for anthocyanin F gives kaempferol glycoside 2) segregation for quercetin in- G gives kaempferol glycoside tensity All variegated types have a yellow background effect from the xanthophyll and change from purple, to green, to smudgy yellow with fading of the anthocyanin. CONCLUSIONS An attempt was made to determine the pattern of flower color inheritance in alfalfa by studying the inheritance of flower pigments at the diploid level. By chromatographic techniques, three anthocyanin pig- ments, nine flavonol pigments, and xanthophyll esters were identified in alfalfa flowers. The inheritance of the anthocyanin pigments, two kaemp- ferol glycosides, and of xanthophyll pigments was determined. Two hypotheses were proposed for the inheritance of quercetin glycosides. The segregation patterns of 5 pigment genes in 23 fami- lies, enabled the determination of genotypes for 22 parent clones. From the inheritance patterns for these pigments and their phenotypic correlation, an inheritance chart for flow- er color in diploid alfalfa is proposed. 66 Alston, LITERATURE CITED R. and C. Hagen.l958. Chemical aspects of the in- heritance of flower color in Impatiegg balsamina. Genetics 43: 35-47 Armstrong, J. m. and D. 3. Gibson. 1941. Inheritance of ‘Bates—Smith, E. 19h8. Paper chromatography of anthocyanins certain characters in the hybrid of medicaio media and g. glutinosa. Scientific AgricuIture 22: - O. and related substances in petal extracts. Nature, Lond. 161: 835—838. and R. hestall. 1950. Chromatographic behaviour and chemical structure I. Some naturally occurring phenolic substances. Biochimica et Biophysics, Acta 4:427-4h0- 1950. Anthocyanins, flavones and other ghenolic compounds. Biochemical Symposium No. 3 2-71. 1954. Leuco-anthocyanins I. Detection and identi- fication of anthocyanidins formed from leuco-antho- cyanidins in plant tissues. Biochemical Journal 58:122. ,,T. Swain, and C. Nordstrom 1955. Chemistry and in- heritance of flower color in Dahlia. Nature, Lond. 176:1016-1018. “"" 1956. The commoner phenolic constituents of plants and their systematic distribution. Royal Dublin Society Scientific Proceedings 27:375-387. Beale, G. H. 33 a1. 1939. Genetics and chemistry of flower Bonner, Britisn Brozek, color. journal of Genetics 37:375-387. W. D. 1952. Plant Biochemistry. Academic Press Inc. 1‘}. Yo, N. Y. Colour Council, 1941. Horticultural Colour Chart 2 Vol. Banbury, England. A. 1932. Mendelian analysis of the red-orange-yellow group of flower—colors in himulus cardinalis. Hort. Praha. 11:16-25. 67 68 Clements, W. 1961. By personal correspondence. Davies, W. L. 1958. Welsh Plant Breeding Station Report October 1956-September 1958. Dodds, K. and h. Long. 1955. The inheritance of colour in diploid potatoes. Journal of Genetics 53:136. Douwes, H. 1943. Ein genetisch-chemisch enderzeek van Eschsch- oltzia californica Cham. Genetica 23:353-464. (English Summary71 Endo, T._1959. Biochemical and genetical investigation of flower color in Swiss Giant Pansy, Viola x Willbrokiana Gams. The Botonical Magazine, Tokyo 42:10-19. Frimmel, F. 1932. Die genetischen Grundlagen die Farbenzuch- tung der Gartenprimel. Zeithschr. Pff.-Zuchtg. 17: 173-185 0 Gage, T. and S. Wander 1950. Quantitative determination of certagn flavonol-3glycosides. Analytical Chemistry 22:70 . C. D. Douglas and S. h. Sender. 1951 Identification of flavonoid compounds by filter paper chromatography. AnalytiCal Chemistry 23:1582-1584. Geissman, T. E. Jorgensen and b. Johnson. 1954. The chemis- try of flower pigmentation of Antirrhinum ma'us color genotypes. Arch. Biochem. biophys. 49:368-338. , and J. Harborne 1955. The chemistry of flower pig- mentationgin Antirrhinum maius IV. Arch. :iochim and Biophy. 55:447. 1955. Anthocyanins,chalcones, aurones, flavones and related water—soluble plant pigments. From Paech, K and m. Tracey. modern methods of Plant Analysis. Springer-Verlog Heidelberg, Germany. Vol. 3:450-198. Gortner and Gortner 1949. Outlines of Biochemistry. John Wiley and Sons Inc. N. Y. 3rd Ed. pp 868-881. Goodwin, T. 1955. tarotenoids. From Paech, K. and m. Tracey. Modern Methocs of Plant Analysis. Springer-Verlog. Heidelberg, Germany. Vol. 3:272—311. Hagen, C. 1959. Influence of genes controlling flowerecolor and relative quantities of anthocyanins and flavonols in petals of Impatiens balsamina. Genetics h4:787. 69 Haney, W. J. 1954. Color genotypes of greenhouse snapdragons. Journal of Heredity 45:146-148. harborne, J. and H. Sherratt. 1957. Variations in the gly907 sidic patterns of anthocyanins II. hxperientia 13:480- 457- and d. Sherratt. 1958. Flavonoios of the Primulaceal. Nature 161:25-27. 1958. The chromatographic identification of antho- cyanin pigments. Journal of Chromatography 1:473- 487. 1958. Late stages in the biosynthesis of flavonoids. Biocnemical Journal 68:12p. iayashi, K. 1954. A review of the research on anthocyanins in Japan with special consideration of the natural plant pigments. Pharmazie 9:584-588. Lawrence, H. J., and d. Scott-Moncrieff. 1935. The genetics and chemistry of flower color in Dahlia, a new theory of specific pigmentation. Journal of Gentics 30:155. , W. J. Price, G. Robinson and G. hobinson. 1939. The distribution of anthocyanins in flowers, fruits, and leaves. Royal Society of London, Philosophical Trans- actions Series B 230:1h9-178. and Price 1940. The genetics and chemistry of flower color variation.' Cambridge Philosophical Society Bio- logical deview 15:35-57. , Sturgis, V. 1957. Genetics and chemistry of flower colour in the garden forms, species, and hybrids of Streptocarpus. Herdity 11:303-336. Lepper, J. R. and T. E. Odland. 1939. Inheritance of flower color in alfalfa. Journal American Society of Agron- omy 31:209-216. LeHosen, A., F. Kent and L. Zeichmeister l9hl. Relation be- tween genes and carotenoid of the tomato. Proceed- ings National Academy of Science 27:236. Lesins, K. 1956. Somatic flower color mutations in alfalfa. Journal of Heredity 47:171-179. 7O Mangelsdorf, P. C. and U. Fraps. 1931. A direct quantitative relationship between vitamin A in corn and the number of genes for yellow pigmentation. Science 73:241. moffett, A. A. 1936. The genetics of Tropaeolium maju . Journal of Genetics 33:151-161. Nordstromé g.h§. and T. Swain 1953. The flavonoid glycosides o a ia variabilis Part 1. Journal Chemical Society 2764-277‘.' ‘ 1956. The flavonoid glycosides of Dahlia variabilis Part IV. Acta Chem. Scand. 10:1491. Paris, 0., a. J. Haney, and G. B. Wilson. 1960. A survey of the interactions of genes for flower color. Michigan State Univ. Tech. Bull. 281. Pecket, R. 1960. The nature of the variation in flower colour in the genus Lathyrus. The New Phytalogist 59:138-144. Robinson, G. and Robinson R. 1931. A survey of anthocyanins 1. Biochemical Journal 25:1687. ,Robinson R. 1933. A survey of anthocyanins III. Biochemical Journal 27:20 . and R Robinson. 1934. A survey of anthocyanin IV. Biochemical Journal 28:1712-1720. 1936. The formation of anthocyanins in plants. Nature, Lond. 137:172. Roberts, E., H. A. Cartwright and D. Wood. 1956. Les Couleurs Les Fleurs et Des Fruits anthocyanins et Flavones. Memoires Du museum National D' distaire Naturelle, Series B, Botanique Tome II. Fascicule Unique Syste- matique: 188-234. Scott-honerieff, A. 1932. A biochemical survey of some mende- 1ian factors for flower color. Journal of Genetics 32:117. 1936. A biochemical survey of some mendelian factors for flower colour. Journal of Genetics 32:117-170. Shrock, Otto. 1943. Genetische beobachtungen an luzerge (medicago media). Ztchr. fur Phlanzenzuchtung 5: Stanford, E. d. 1951. Tetrasomic inheritance in alfalfa. Agronomy Journal 43:222-225. .r., 71 Stephens, S. 1948. A biochemical basis for the pseudo-allelic anthocyanin series in Gossypium. Genetics 33, 191-214. Stewart, C. 194 . Lffects‘of inbreeding on variability in alfalfa. Journal Agricultural desearch 49:669-694. Sutton, Eileen. 939. The genetics of Trgpaeolium majus 11. Journal of Genetics 38:161-176. Storgaard, A. K. 1957. Genetic studies in alfalfa medicapo sativa. Dissertation Abstracts 324:203-204. Strain 1945. Chromatographic Adsorption Analysis. Inter- science Publishers, Inc. N. Y., N. Y. Twamley, B. E. 1955. Flower color inheritance in diploid and tetraploid alfalfa. Canadian Journal of Agricul— tural Science 35:461-476. Vickery, n. K. and R. L. Olson. 1956. Flower color inheritance in Mimulus cardinalis complex. Journal of heredity 47:194-199. Reining, A. m. 1946. Flower color inheritance in alfalfa. Journal American Society of Agronomy 40:746-750. 1948. Williams, A. J. and h. Kirby. 1948. Paper chromatography using capillary ascent. science 107:401-4d3. williams, T. I. 1948. An Introeuction to Chromatography Chemical Publishing Company, Brooklyn N. 1. fi._.___._—-4 if. j. x '. I ~_ _' ‘. \5 ¢ ' . J: .9 -. ‘ i .4” |‘ - n s. h 1': . ' l S} ‘ . 1’ ... '\ ’l . ., ,. .6 :7 -. ~.q_ .I g . . -‘r 'L" ' 0., . :- _. q '35:, £35 ‘ Room USE ear Giff}??? USE (3532.1