Or Et Al.: Genetics of Flowering Time in Chickpea

C timing of flowering is a major component of crop environmental adaptation, particularly when Water availability is a major yield-limiting factor in semi-arid regions. Hence, efficient utilization of soil water for grain production the growing season is restricted by climatic factors such depends on correct timing of flowering. Following winter (Decemberas drought and high temperature (Subbarao et al., 1995). January) sowing, modern Israeli chickpea (Cicer arietinum L.) cultiFlowering date is generally dependent on germination vars begin flowering during the last week of March and their reproducdate, the seasonal temperature profile, and the phototive period extends throughout April to June. In the Middle East, thermal response of the plant. In a given genotype, at April, May, and June are often dry and hot months. The objectives a given latitude and altitude, flowering date may be of this study were (i) to assess the potential range of chickpea germmodified by sowing date (Garner and Allard, 1920, plasm as a source for early flowering, and (ii) to study the inheritance 1930). However, a wider phenological variation may be of the time to flowering trait. Germplasm evaluation was carried out achieved by the introduction of allelic variation for day by measuring days from germination to flowering and calculating length and/or temperature response. phenotypic correlations between days to first flower, grain weight, In most Mediterranean, Asian, American, and Afriand pod number along main branches. A number of early-flowering genotypes were identified, and weak association between flowering can chickpea-growing areas, the duration of the reprotime genes and seed weight loci was observed. Crosses were made ductive phase of the crop is delimited by the initiation between types of contrasting photoperiod response. In F2 populations of flowering and the summer drought that terminates derived from crosses between an early-flowering breeding line (desi) seed set. Furthermore, because of the indeterminate with weak photoperiodic response and a late-flowering high-yielding (kabuli) cultivar with a strong photoperiod response, a 3:1 ratio of E. Or, Institute of Horticulture, Agric. Res. Organization, The Volcani late-flowering : early-flowering types was observed. This segregation Center, P.O. Box 6, Bet Dagan, Israel; R. Hovav and S. Abbo, Dep. is consistent with action of a major photoperiod response gene (Ppd ) of Field Crops, Vegetables and Genetics, Faculty of Agric., Food and affecting time to flowering. Considerable genotype 3 environment Environmental Quality Sci., The Hebrew Univ. of Jerusalem, Rehovot interaction was observed among F3 progeny of these crosses. The 76100, Israel. Received 18 Aug. 1997. *Corresponding author (abbo relatively simple inheritance of the photoperiodic response observed @agri.huji.ac.il). in this study suggests that the early-flowering trait may be easily introduced into popular late-flowering genetic backgrounds. Published in Crop Sci. 39:315–322 (1999). 316 CROP SCIENCE, VOL. 39, MARCH–APRIL 1999 growth habit of chickpea, the duration of its flowering combe, 1984). On the basis of these findings, it was postulated that early-flowering chickpea types might be period is a major yield determinant (Bonfil and Pinthus, 1995; Eshel, 1967). Early flowering combined with other found among Ethiopian varieties which are grown at latitude 108N under smaller seasonal day-length changes desirable plant traits might make it possible to prolong the reproductive phase in the various target envicompared with those of other growing areas. Another potential source for early flowering is the relatively dayronments. length insensitive type ICC5810, identified by Roberts In the closely related legumes, pea (Pisum sativum et al. (1985). L.) and lentil (Lens culinaris Medik.), early-flowering Earlyand super-early flowering genotypes have been habit is often associated with day-length insensitivity identified in the ICRISAT program (Pundir et al., 1988; (Arumingtyas and Murfet, 1994; Erskine et al., 1994). Kumar and Rao, 1996). The multi-location testing sysIn lentil, a close association between the geographic tem operated by ICRISAT is very useful for such purorigin of cultivars and their flowering time has been poses, since lines that attain early flowering at the ICRInoted. Cultivars originating from higher latitudes are SAT center in Hyderabad (178N) are most likely to mostly late flowering, whereas types originating from flower early in the Mediterranean spring. Despite its lower latitudes are early flowering (Erskine and Witeavailability to breeders, the authors are unaware of any utilization of such germplasm for improvement of kaTable 1. Days from emergence to 50% flowering of chickpea lines buli chickpea. (and F1 plants) in nethouse experiments. The objectives of this study were (i) to assess the Days to flower potential range of chickpea germplasm as a source for Line Origin 1994–1995 1995–1996 1996–1997 Seed type early flowering, and (ii) to study the inheritance of the ICC7526 Ethiopia 110 117 –† D‡ time-to-flowering trait in chickpea. ICC7528 Ethiopia 110 111 – D ICC7534 Ethiopia 108 94 – D MATERIALS AND METHODS ICC7537 Ethiopia 109 78 82 D ICC7545 Ethiopia 110 72 – D In the first season (1994-1995), all lines (Tables 1 and 2) ICC8448 Ethiopia 125 94 – D were grown in a bird-proof nethouse at the Weizmann InstiICC8454 Ethiopia 108 80 – D ICC8573 Ethiopia 104 85 – D tute of Science, Rehovot, Israel. During the 1995-1996 and ICC8575 Ethiopia 110 71 – D 1996-1997 seasons, plants were grown in an insect-proof netICC8578 Ethiopia 117 84 – D house at the experimental farm of the Faculty of Agriculture, ICC8584 Ethiopia 110 107 – D Rehovot. Cultivated forms were sown and irrigated to ensure ICC8605 Ethiopia 112 85 – D ICC8625 Ethiopia 107 85 – D uniform germination during the third week of November in all ICC8628 Ethiopia 117 83 – D growing seasons. In each experiment, all genotypes emerged ICC8631 Ethiopia 110 80 – D within a week of sowing. Seeds of the wild forms were scarified ICC8634 Ethiopia 128 76 64 D to overcome seed dormancy, germinated in Petri dishes, and ICC12578 Ethiopia 110 78 – D ICC12734 Ethiopia 104 99 – D planted after root emergence during the last week of NovemICC12769 Ethiopia 117 96 – D ber of the three growing seasons. Four plants of every line ICC12825 Ethiopia 104 71 – D were grown in the 1994-1995 season. Two replicates, each ICC12830 Ethiopia 117 87 84 D consisting of eight plants, were grown in the following season ICC13940 Ethiopia 127 84 – D ICC14088 Ethiopia 117 75 – D 1995-1996. Three replicates, of eight plants each, were grown ICC14091 Ethiopia 117 85 – D during the third season. In all nethouse experiments, plants ICC14093 Ethiopia 114 85 – D were grown in two row plots along trickle-drip irrigation lines, ICC14098 Ethiopia 119 69 – D (40 cm between the rows and 2 m between the irrigation ICC14102 Ethiopia 125 83 – D ICC14107 Ethiopia 110 83 85 D ICC14144 Ethiopia 117 91 – D Table 2. Days from emergence to flowering and number of pods M-2487 Israel – 105 – K along the main branches of modern Israeli chickpea cultivars M-2496 Israel – 94 – K and selected Ethiopian landraces, in the field with February M-2516 Israel – 125 – K 1997 sowing at Kama farm, Israel. M-2582 Israel – 97 – K M-2856 Israel – 100 – K Days to 50% No. of pods on Dwelley USA – – 125 K Line Origin flowering† main branches‡ Sanford USA – – 119 K CECH1 Mexico – – 84 K ICC8573 Ethiopia 54 6 0.29 9.5 6 0.76 ICC8575 Ethiopia 53 6 0.25 11.8 6 1.61 ICC5810 India 79 59 39 D Hadas Israel 140 121 115 K ICC8625 Ethiopia 53 6 1.11 11.9 6 0.83 ICC8631 Ethiopia 53 6 0.41 11.5 6 0.73 F1-Hadas 3 ICC5810 – 96 87 – ICC13940 Ethiopia 53 6 0.5 12.1 6 0.76 ICC14088 Ethiopia 53 6 0.41 11.5 6 0.22 F1-ICC5810 3 Hadas – 96 87 – ICC14098 Ethiopia 53 6 0.5 12.6 6 1.38 M2581 Israel 56 6 0.58 7.9 6 0.1 Ayelet Israel 141 – 125 K Bulgarit Bulgaria 107 117 K Ayala Israel 55 6 1.0 5.3 6 0.65 Bulgarit Bulgaria 64 6 0.25 9.3 6 1.0 Spharadit Mexico 122 83 84 K 205 Turkey 142 134 130 W Hadas Israel 61 6 0.67 7 6 0.51 F7-31 Israel 64 6 3.43 5.3 6 0.38 231 Turkey – 143 – W PI510655 Turkey – 140 – W 13816 Israel 69 6 0.25 8 6 0.51 WIR32 Russia 70 6 0.63 7.7 6 0.71 89-09-01 Turkey 142 145 – W † Means and standard error for number of days from emergence for 50% † Missing data indicates that the specific accession was not grown during the respective season. of the plants to have at least one flower in four replicates. ‡ Mean and standard error for pod number on four branches of eight ‡ D 5 desi seed type, K 5 kabuli seed type, W 5 wild C. reticulatum seed type. random individual plants in four replicates. OR ET AL.: GENETICS OF FLOWERING TIME IN CHICKPEA 317 lines), with spacing of 10 plants meter2 within the row, on RESULTS AND DISCUSSION Rhodoxeralf soil. Data were recorded for the date at which In all three growing seasons in Rehovot, a large numthe first fully opened flower appeared on 50% of the plants ber of the tested Ethiopian landraces flowered earlier for each genotype. In the second growing season, the number than the standard Israeli cultivars (Table 1). The three of pods along the main branches and mean grain weight were recorded in a sub-set of the germplasm collection. (Rehovot) growing seasons differed in temperature Crosses were made between the late-flowering cv. Hadas (Fig. 1) and precipitation profiles; hence, the betweenand the early-flowering line ICC5810 by Retig’s (1971) techyear variation that was observed for most lines (Table nique. The parental lines differ in growth habit and flower 1). The 1994-1995 winter was colder than the 1995-1996 color, hence the hybrid nature of the F1 plants grown during winter, while the 1996-1997 season was characterized the 1995-1996 and 1996-1997 seasons was easily verified. F2 by an atypically warm December causing earlier flowplants from the Hadas 3 ICC5810 cross (210 individuals and ering of most lines compared with the two preceding 179 of the reciprocal cross) were grown (same planting arwinters (Table 1). The temperature profile of the 1997rangement as above) with nine replicates of the parental lines 1998 season (Kedma farm) was more typical of the Is(minimum of eight plants/replicate) in the nethouse during the 1996-1997 season. All plants germinated during the last raeli winter and clearly different from the previous seaweek of November 1996. Data on flowering date of F2 plants son (Fig. 1). were collected on an individual plant basis. F2 plants were The screening results show the potential of the Ethioharvested individually and the number of pods along the most pian chickpea gene-pool as a source for early-flowering fertile branch was recorded. Upon threshing and removal of alleles. Having been grown for millennia around 108N, shriveled grains, mean grain weight was determined for every where seasonal day-length changes are minimal, these F2 plant. lines are presumed to be less sensitive to photoperiod F3 populations of the Hadas 3 ICC5810 and the reciprocal than Mediterranean types. Another genotype selected crosses were grown in a replicated field experiment in the on the basis of its reduced day-length response Kedma farm, about 35 km south of Rehovot on Vertic Haploxeralf soil. Five to 15 individual F3 progeny were grown from (ICC5810) was the earliest to flower during all 3 yr each F2 plant of the respective populations. The material was (Table 1). This line started flowering 79, 59, and 39 d hand-sown during the last week of November and the first after emergence during the 1994-1995, 1995-1996, and week of December 1997. Following 75 mm of rain, the material 1996-1997 seasons, respectively, compared with 140, germinated during the third week of December 1997. Flow121, and 115 d for the standard Israeli cultivar Hadas ering scores were given on an individual plant basis upon the during the respective seasons. These two lines were seappearance of first open flower. The plot was adjacent to a lected for further study. commercial chickpea field of cultivar Spharadit. In both fields, The correlations (based on accessions means) bepest and disease control as well as irrigation were all done according to local practice. tween days to flowering during one season and days to A set of early-flowering lines was sown in a four-replicated flowering during the other two seasons in the nethouse field test at the Kama farm, about 25 km east of Rehovot on experiments were as follows: r 5 0.879 (first season 3 13 Feb. 1997 on Chromoxerert soil. Each plot included two second season), r 5 0.907 (second season 3 third searows, 1.5 m long, 1 m apart (about 20 plants m2). The test son), and r 5 0.858 (first season 3 third season). All plot was surrounded by a commercial field of cv. Hadas, sown three r values differ from zero (P , 0.01). These r 10 d earlier. Standard base fertilization, irrigation, chemical values were obtained despite the fact that only a limited weed, and ascochyta control throughout the season were done number of plants was grown during the first season (four according to local practice. Flowering dates and pod numbers at maturity were recorded as described for all germplasm. individuals/accession) and during the second season Fig. 1. Mean monthly temperature profiles of three growing seasons (1994-1997) in Rehovot, and the fourth season (1997-1998) in Kedma, Israel. 318 CROP SCIENCE, VOL. 39, MARCH–APRIL 1999 Table 3. Parental and F1 mean values (6S.E.) of days from emerindividuals commenced flowering 87 d after emergence gence to 50% flowering in the Hadas 3 ICC5810 and ICC5810 (Table 3). The early parental line ICC5810 started flow3 Hadas crosses. ering after 39 d, while the late-flowering parent started Line Days to first flower 6 S.E. flowering 115 d after emergence. The early-flowering Hadas 115 6 0.53 parent flowered throughout the season, whereas the F1 ICC5810 39 6 0.66 plants stopped flowering after about 10 d and produced Hadas 3 ICC5810 87 6 0.00 undeveloped white buds, known as ‘‘pseudo-flowers’’ ICC5810 3 Hadas 87 6 0.00 (Summerfield and Roberts, 1985) during a 3-wk period of lower than average night temperatures (ranging be(two replicates of eight plants each). This may be taken tween 10–28C). As a result, the first developing pods as evidence for the relative accuracy of the evaluation on the F1 plants appeared only during the first week of procedure. April. Similar cessation of flowering followed by undeEight F1 individuals of the Hadas 3 ICC5810 combiveloped flower buds was observed, during the same nation and five F1 individuals of the reciprocal cross period, on many F2 individuals grown in the same were grown with the parental lines during the 1996nethouse. 1997 season. Chromosomal aberrations have been noted The distribution pattern of the days to first flower in between and within annual Cicer species (Ladizinsky the F2 populations (Fig. 2A, B) suggest an involvement and Adler, 1976). Pollen viability of all F1 plants, estiof a major gene affecting the flowering time trait. Classimated by staining with 2% aceto-carmine, was found fication of the days-to-flowering data into two classes— to range between 95 and 100%. Time to flowering of late (after 46 d) and early (up to 45 d)—resulted in a x2 value of 2.53 for the Hadas 3 ICC5810 and 0.58 for all F1 plants was monitored on a daily basis. All F1 Fig. 2. A. Frequency distribution of days to first flower in the Hadas 3 ICC5810 F2 population expressed as percentage of individuals flowering on given days after emergence. B. Frequency distribution of days to first flower in the ICC5810 3 Hadas F2 population expressed as percentage of individuals flowering on given days after emergence. OR ET AL.: GENETICS OF FLOWERING TIME IN CHICKPEA 319 the reciprocal cross, with probabilities of 0.11 and 0.45, slightly affected (115 d in 1996-1997 and 105 d in 19971998), ICC5810 was strongly affected (39 d and 84 d, respectively. The F1 days to flowering values (87 d) of the two reciprocal combinations coincide with the peaks respectively). This compaction of the phenotypic groups caused the frequency distribution to lose its bimodal of the late groups in the respective F2 populations (Fig. 2A, B). patterns in the F3 generation of the Hadas 3 ICC5810D cross, and nearly to lose its bimodal patterns in the Inspection of the distribution histograms of days to first flower (Fig. 2A, B) shows that nearly no F2 individreciprocal F3 population (Fig. 3). Despite the strong environmental effect on the days-to-flowering phenoual was as late as the late parent in both populations. The absence of such late segregants may be accounted type, the strong genotypic effect of the PPd gene is evident when testing the means of the F3 families acfor by the fact that F2 plants were tagged when the first perfect flower appeared until the third week of February cording to the grouping used to test the 3:1 segregation in the F2 generation. In the Hadas 3 ICC5810 cross, 1997, when it was no longer possible to detect the newly flowering individuals because of intermingling of the means of all F3 families progeny of F2 individuals with days-to-flowering values of up to 45 d and those branches and canopy development along the rows. As a result of a different seasonal temperature profile with days-to-flowering values 46 d and after were 88 6 0.3 and 90 6 0.2, (different from each other at P , (Fig. 1) and repeated predation by hares and larks that delayed the vegetative development of the plants, mean 0.0001). In the ICC5810 3 Hadas population, the corresponding values were 87 6 0.8 and 95 6 0.3, respectively days to first flower of the F3 families of the two reciprocal populations were delayed as expressed by higher values (different from each other at P , 0.0001). ICC5810 has been characterized previously by Robof the parental means and the smaller population ranges (Fig. 3A, B). This change of the environment affected erts et al. (1985) as ‘‘virtually day-neutral’’, and since Hadas is considered to be highly photoperiod sensitive the parental lines differently. While Hadas was only Fig. 3. A. Frequency distribution of days to first flower (F3 family means) in the Hadas 3 ICC5810 F3 generation expressed as percentage of families flowering on given days after emergence. B. Frequency distribution of days to first flower (F3 family means) in the ICC5810 3 Hadas F3 generation expressed as percentage of families flowering on given days after emergence. 320 CROP SCIENCE, VOL. 39, MARCH–APRIL 1999 (Retig, pers. comm.), the 3:1 Mendelian ratio of late:early F2 individuals is probably a result of the above photoperiod-response gene effect. A gene symbol of Ppd is suggested for this gene, with ppd for the recessive day-neutral allele. Roberts et al. (1985) provided evidence of genetic differences in flowering photoperiod response between chickpea cultivars, but the underlying genetic mechanism was not characterized. A similar genetic system of photoperiod response was identified in pea (Weller et al., 1997). In common bean (Phaseolus vulgaris L.), monogenic (Coyne 1970) as well as digenic control (Kornegay et al., 1993) of photoperiod response was reported. In pea, it was possible by means of defined genetic stocks and test-crosses to dissect the interaction between the photoperiod-response genes and other genes affecting time to flowering (Weller et al., 1997). At the moment, we are unable to relate the suggested Ppd effect with other loci responding to additional environmental factors (e.g., temperature). However, the differences between the days-to-flowering values in the different seasons (Table 1) suggest the involvement of additional loci (modifiers) in determining the flowering phenotype. Roberts et al. (1985) who analyzed the flowering response of nine chickpea cultivars (ICC5810 included) concluded that ‘‘for all genotypes, the rates of progress towards flowering . . . were linear functions of mean temperature’’. The differential photoperiod response recorded by Roberts et al. (1985) did not correspond with the temperature response, suggesting that the photoperiod loci are (at least partially) independent from the temperature-response loci. Such temperature-response loci may account for part of the seasonal variation of the tested parental lines and their progeny in Fig. 4. A. Frequency distribution of mean seed weight (g) in the the F2 and F3 generations. Hadas 3 ICC5810 F2 population expressed as percentage of individPhenotypic correlations between days to first flower, uals with respective mean seed weights. B. Frequency distribution pod number, and mean grain weight were calculated of mean seed weight (g) in the ICC5810 3 Hadas F2 population in the two F2 populations. In F2 generations with no expressed as percentage of individuals with respective mean seed weights. replicated progeny, it is impossible to subtract the environmental variance component. Still, the correlation values may indicate if a strong genetic association exists produced larger numbers of pods along their main between the tested traits. In the two F2 populations, no branches than standard large-seeded kabuli cultivars correlations were detected between pod-set and days (Table 4). However, the pod number of some desi lines to flower. A low, but significant correlation (r 5 0.289, (ICC7537, ICC14088) was quite similar to those of the P , 0.00001) between days to flowering and grain kabuli lines (Table 4). weight was observed in the Hadas 3 ICC5810 populaEarly lines sown in the field on 13 Feb. 1997 flowered tion. However, in the reciprocal F2 population, the re7 to 16 d earlier than the standard kabuli Israeli cultivars spective r value did not differ from zero. Interestingly, (Table 2), whereas the difference in flowering time bethe two populations exhibited different distribution pattween the two germplasm groups was 2 to 5 wk after terns of mean grain-weight values (Fig. 4A, B). In the November sowing (Table 1). The correlation between ICC5810 3 Hadas cross, a shift towards lower graindays to 50% flowering and pod number along the main weight values was evident with a mean of 0.244 g, while branches for the field-tested lines was negative (r 5 in the Hadas 3 ICC5810 cross no such shift was seen 20.48, P , 0.001). A different relationship between with an average grain weight of 0.272 g. A t test indicated days to flowering and pod number was observed within that the two means differed with P(t) , 0.0001. Since each of the two cultivar groups. The mean of days to the desi line ICC5810 is the parent with the smaller 50% flowering of the desi lines was 53.4 d, and the mean grains, this result suggests a cytoplasmic effect on grain value of pods along the main branch was 11.5 d. The weight. Further experiments are required to study the mean values for the kabuli cultivars were 62.3 d and 7.3 cause of the different associations between grain weight pods. A t test indicated that cultivar-group means for and time to flowering in the reciprocal F2 populations. both parameters were different at P(t) , 0.0001. However, no significant correlation was found between days In general, small-seeded early-flowering desi lines OR ET AL.: GENETICS OF FLOWERING TIME IN CHICKPEA 321 Table 4. The number of pods (mean and range on main branches) and the early-summer high air temperatures that termiof selected lines and F1 plants in the 1995–1996 nethouse exnate seed set. Therefore, it was hypothesized that a periment. longer flowering period, brought about by early-flowNo. of pods ering alleles, may enhance seed yield. In the present Line designation Range Mean 6 S.E. Seed type study, this simple assumption proved incorrect. Among the range of kabuli and desi cultivars, and in the tested Hadas 5–10 7 6 0.82 K† ICC5810 9–22 16.1 6 1.8 D F2 populations derived from early 3 late crosses, no F1-Hadas 3 ICC5810 9–14 12 6 0.86 – correlation was found between days to flowering and F1-ICC5810 3 Hadas 6–9 8 6 0.44 – number of fertile pods at maturity. Nevertheless, the Bulgarit 5–11 8.7 6 1.1 K M-2582 4–13 9.3 6 1.6 K data (Tables 2 and 4) suggest that early flowering may M-2856 9–15 12.2 6 1.15 K be associated, in certain genetic backgrounds, with high ICC7528 8–14 11 6 1.0 D pod set. Several causes may account for the lack of a ICC7537 6–11 9 6 0.89 D ICC8573 9–19 14.6 6 1.8 D simple relationship between flowering time and pod set ICC8575 9–15 12.2 6 1.16 D in this work. ICC8625 9–16 12.4 6 1.29 D ICC12825 11–19 15.2 6 1.6 D The realization of the full yield potential of early ICC12830 8–16 11.8 6 1.91 D flowering under Mediterranean growing conditions reICC13940 9–21 16.4 6 2.31 D quires a certain degree of chilling tolerance during the ICC14088 6–12 10.2 6 1.11 D fertilization process. Low rates of pollen germination † D 5 desi seed type, K 5 kabuli seed type. and slow pollen-tube elongation have been mentioned among the factors limiting pod set in chickpea below to 50% flowering and number of pods among the differ208C (Savithri et al., 1980). Indeed, under the Israeli ent lines within each of the desi or kabuli groups. This winter conditions of 1997, the early-flowering line indicates that earlyand late-flowering types with either ICC5810 flowered 60 d before the first developing pods low or high number of pods exist in both cultivar groups, were observed in mid-March. Certain temperature rewhich agrees with the nethouse data (Table 4). quirements for proper pollen germination and subseThe pod number data of the F1 plants of the Hadas 3 quent embryo development may have been among the ICC5810 (and reciprocal) from the 1995-1996 season reasons for the absence of correlation between days to (Table 4) may indicate a cytoplasmic effect on the reproflowering and pod set in the analyzed F2 populations. ductive potential. This result was not confirmed in the A small-scale survey of a few F2 individuals of in vivo following season when pod numbers along the main pollen-tube germination and elongation into the stylar branches of the two reciprocal F1 combinations did not columns supports this assumption. Moreover, as mendiffer. A mean (eight plants) of seven pods/branch for tioned earlier, the Hadas 3 ICC5810 (and reciprocal) the Hadas 3 ICC5810 F1 and a mean of 6.8 pods/branch F1 plants, as well as a considerable number of F2 plants, (five plants) for the reciprocal combination was obpractically stopped flowering during a 3-wk period of served. low temperatures in early spring 1997. Since the earlyThe simple genetic basis of the photoperiod response flowering parent (ICC5810) did not show this behavior, observed in this study suggests that it might be fairly while it was observed on the F1 plants, it is suspected easy to incorporate the early-flowering habit into modto have been transmitted from the late-flowering parent ern high-yielding cultivars, either by backcross breeding cv. Hadas. With commercial planting, this trait is hardly or by simple selection in F2 and subsequent generations. detected because plants commence flowering later in At the moment, there is no evidence that other genetic the season (late March to early April) when temperadifferences between the desi and the kabuli gene-pools tures are higher. However, in the 1996-1997 season, may prevent successful introgression of the early-flowduring a different field experiment, cv. Hadas and Ayala ering trait. (sown on 1 Dec. 1996) produced undeveloped buds In semi-arid habitats, the timing of flowering is of (‘‘pseudo flowers’’) for more than 30 successive days. great adaptive value for both wild and crop plants. In the Because planting in this experiment took place about wild, optimal flowering should ensure seed maturation 40 d prior to the recommended sowing date, the plants before onset of the summer drought. The growth cycle passed the juvenile stage earlier than usual, while avof wild legumes, like chickpea or lentil, often ends with erage temperatures were still lower than the required maturation of a small number of seeds (Ladizinsky, minimum for proper floral development (Abbo, unpublished data). Successful exploitation of the early-flow1989; 1993). This is a disadvantage for a farmer interested in maximizing the seed yield. Water availability is ering trait to extend the reproductive period will depend on combining the reduced temperature requirements a major yield-limiting factor in semi-arid regions; hence, efficient utilization of soil water for grain production for floral development, as seen in ICC5810, with chilling tolerance of the fertilization process. depends on correct timing of flowering. Comparison between spring wheat (Triticum aesitvum L.) and The F2 populations analyzed in the above experiments were irrigated until the latest plants to flower produced chickpea have led Bonfil and Pinthus (1995) to suggest that because of its indeterminate growth habit, the duramature pods. This was done to ensure a genetically balanced F3 generation comprising the widest range of tion of the chickpea flowering period is a major yield determinant. The duration of the reproductive period flowering habits possible. Had drought occurred earlier in the season, as often happens under dryland condiis restricted between the commencement of flowering 322 CROP SCIENCE, VOL. 39, MARCH–APRIL 1999 and a comparison of the factors affecting chickpea seed yield with tions, the latest plants to flower might not have produced those affecting wheat grain yield. Exp. Agric. 31:39–47. a single seed. Such a situation could have resulted in a Coyne, D.P. 1970. Genetic control of photoperiod-temperature renegative correlation between days to flowering and pod sponse for time of flowering in beans (Phaseolus vulgaris L.). Crop set. In fact, a simulation exercise, using flowering values Sci. 10:246–248. Erskine, W., A. Hussain, M. Tahir, A. Bakash, R.H. Ellis, R.J. Sumof March 1997, for the untagged F2 individuals resulted merfield, and E.H. Roberts. 1994. Field evaluation of a model of in a significant negative correlation between days to photothermal flowering response in a world lentil collection. Theor. first flower and pod number along the main branch. To Appl. Genet. 88:423–428. address this point, F4 families of the above populations Erskine, W., and J.R. Witecombe.1984. Lentil germplasm catalog. ICARDA, Aleppo, Syria. will be tested under dryland conditions in the 1998Eshel, Y. 1967. Effect of sowing date on growth and seed yield compo1999 season. nents of chickpea. Israel J. Agric. Res. 17:193–197. While it is clear that selection for early flowering Garner, W.W., and H.A. Allard. 1920. Effect of the relative effect of will not directly increase productivity, the possibility of day and night and other factors of the environment on growth and reproduction in plants. J. Agric. Res. (Washington, DC) 17: combining early flowering with yield-promoting alleles 553–606. was demonstrated in desi chickpea (Siddique and Khan, Garner, W.W., and H.A. Allard. 1930. Photoperiodic response of 1996). The feasibility of such an exercise in desi 3 kabuli soybeans in relation to temperature and other environmental facprogeny still remains to be explored. tors. J. Agric. Res. (Washington, DC) 41:719–735. Kornegay, J., J.W. White, J.R. Dominguez, G. Tejada, and C. Cajiao. 1993. Inheritance of photoperiod response in Andean and MesoamACKNOWLEDGMENTS erican common bean. Crop Sci. 33:997–984. Kumar, J., and B.V. Rao. 1996. Super-early chickpea developed at the We thank F.J. Muehlbauer (USDA-ARS, Washington State ICRISAT Asia Center. Int. Chickpea Pigeonpea Newsl. 3:17–18. University, Pullman WA) for providing a collection of AmeriLadizinsky, G. 1989. Lentil domestication facts and fiction. Econ. can and Ethiopian chickpea cultivars and wild Cicer reticulaBot. 43:131–132. Ladizinsky, G. 1993. Lentil domestication: On the quality of evidence tum Ladiz. accessions; B. Retig (Agriculture Research Organiand arguments. Econ. Bot. 47:60–64. zation, Volcani Center, Bet-Dagan, Israel) for donating seeds Ladizinsky, G., and A. Adler. 1976. The origin of chickpea Cicer of Israeli cultivars; G. Ladizinsky (The Hebrew University arietinum L. Euphytica 25:211–217. of Jerusalem, Rehovot, Israel) for supplying two additional Pundir, R.P.S., K.N. Reddy, and M.H. Mengesha. 1988. ICRISAT C. reticulatum accessions; the Israeli Gene-Bank (Volcani Chickpea Germplasm Catalog: Passport Information. ICRISAT, Center, Bet-Dagan) for seed of Israeli landraces; R.J. SumPatancheru, India. merfield (Agricultural Botany, Reading University, Reading, Retig, B. 1971. Hybridization methods in chickpea. II. crossing without emasculation. Israel J. Agric. Res. 21:113. UK) who provided seeds of cv. ICC5810. This research was Roberts, E.H., P. Hadley, and R.J. Summerfield. 1985. Effect of temsupported, in part, by grants from the Scheinbroon Foundaperature and photoperiod on flowering in chickpeas (Cicer arietition, the Sandeground Endowment for Biological Science, the num L.). Ann. Bot. 55:881– 892. Indian-Israeli Biotechnological Cooperation (grant no. 8969Savithri, K.S., P.S. Ganapathy, and S.K. Sinha. 1980. Sensitivity to 1-97), and the Chief Scientist’s Foundation of the Israeli Minislow temperature in pollen germination and fruit-set in Cicer arietitry of Agriculture, grant no. 811-0204-97. Thanks are due to num L.J. Exp. Bot. 31:475–481. Mr. Eli Lior and Mr. Yaa’kov Yinon for their help in the Siddique, K.H.M., and T.N. Khan. 1996. Early-flowering and highyielding chickpea lines from ICRISAT ready for release in Western field experiment. We extend our gratitude to I.C. Murfet, M. Australia. Int. Chickpea Pigeonpea Newsl. 3:22–24. Ambrose, L.A Morrison, and two anonymous reviewers for Subbarao, G.V., C. Johansen, A.E. Slinkard, R.C. Nageswara Rao, their useful comments on the manuscript. N.P. Saxena, and Y.S. Chauhan. 1995. Strategies for improving drought resistance in grain legumes. Critic. Rev. Plant Sci. 14: 469–523.REFERENCESSummerfield, R.J., and E.H. Roberts. 1985. Cicer arietinum. p. 92–99.Arumingtyas, E.L., and I.C. Murfet. 1994. Flowering in Pisum: AIn A.H. Halevy (ed.) CRC Handbook of Flowering. CRC Press,further gene controlling response to photoperiod. J. Hered.Boca Raton, FL.85:12–17.Weller, J.L., J.B. Reid, S.A. Taylor, and I.C. Murfet. 1997. The genetic control of flowering time in pea. Trends Plant Sci. 2:412–418.Bonfil, D.J., and M.J. Pinthus. 1995. Response of chickpea to nitrogen,