Escape and tolerance to high temperature at flowering in groundnut (Arachis hypogaea L.)

Transcription

1 Journal of Agricultural Science, Cambridge (2000), 135, Cambridge University Press Printed in the United Kingdom 371 Escape and tolerance to high temperature at flowering in groundnut (Arachis hypogaea L.) P. Q. CRAUFURD*, T. R. WHEELER, R. H. ELLIS, R. J. SUMMERFIELD AND P. V. VARA PRASAD Plant Environment Laboratory, Department of Agriculture, University of Reading, Cutbush Lane, Shinfield, Reading RG2 9AD, UK (Revised MS received 19 June 2000) SUMMARY Groundnut is an important crop of the semi-arid tropics where potential yields are frequently reduced by heat and water stress. Eight groundnut genotypes varying in heat tolerance were grown in controlled environments and exposed to either high (40 28 C) or near-optimum (30 24 C) temperature from 32 days after sowing (DAS) to maturity. There was significant variation among genotypes in mainstem leaf number and total flower number at and C and rates of appearance were faster at C than at C. Days from sowing to first flowering varied among genotypes from 28 to 41 days and therefore the time plants were exposed to high temperature relative to first flowering ranged from 4 to 9 days. Fruit number for seven out of eight genotypes at C was linearly and negatively related to the time of first flowering relative to the onset of high temperature (r 0 93; n 7;P 0 001), indicating that escape was an important component of heat tolerance in this experiment. Further analysis showed that fruit number in all genotypes at C was closely associated with the cumulative number of flowers that had opened between first flowering and 3 days after the onset of the high temperature regime (r 0 95; n 8; P 0 001). Variation in fruit number was therefore due both to the timing of flowering and the initial rate of flower production. These data also suggest that the most sensitive stage of development to high temperature in groundnut occurred around 3 days before flowers opened. Therefore, it was the timing of flowering, rather than heat tolerance or susceptibility, that was the dominant attribute determining fruit number. INTRODUCTION Groundnut (Arachis hypogaea L.) is an important oilseed and forage crop of the semi-arid tropics. Average yield (groundnut in shell) in the semi-arid tropics is less than 800 kg ha and drought is a major constraint to groundnut production (ICRISAT 1994). Drought is a complex combination of stresses involving both water-deficit and high temperature. Considerable research has been conducted on drought tolerance in groundnut (Wright & Nageswara Rao 1994; Williams & Boote 1995), but there has been little research on high temperature per se. Given likely global climate warming scenarios and recent research at the ICRISAT Sahelian Centre in Niger which has shown that temperature tolerance is the dominant attribute of the adaptation of groundnut to the drought-prone Sahelian regions of West Africa (Greenberg et al. * To whom all correspondence should be addressed. p.q.craufurd reading.ac.uk 1992), it is clearly important to investigate the effects of high temperature on groundnut growth and development. The optimum temperature (T o ) in groundnut is between 25 and 30 C for growth processes and 28 to 33 C for developmental progress to flowering and maturity (Ketring et al. 1982; Ong 1986; Williams & Boote 1995), though T o has only been precisely determined for a few aspects of growth and development. Daily and mean diurnal temperatures warmer than 30 C are frequently experienced in semi-arid tropical environments, particularly in the Sahelian zone of West Africa (Sivakumar et al. 1993), and may occasionally be experienced in the southern USA (Ketring 1984). Continuous exposure to supraoptimal temperature has been reported to reduce apparent photosynthesis (Ketring et al. 1982); vegetative growth and leaf area (Ketring 1984); partitioning of dry matter between roots and shoots (Wood 1968) and between vegetative and reproductive

2 372 P. Q. CRAUFURD ET AL. structures (Ong 1986; Nigam et al. 1994); and wateruse efficiency (Craufurd et al. 1999). Reproductive processes in groundnut are also sensitive to temperature; high temperature increases flower production (Fortanier 1957; Wood 1968; Vara Prasad et al. 1999a) but reduces pollen viability and number (de Beer 1963; Vara Prasad et al. 1999a), and fruit-set (Ketring 1984; Vara Prasad et al. 1999a, b, 2000). The critical day temperature for these processes is about 36 C (Vara Prasad et al. 1999a, 2000). In other legumes the timing of high temperature during reproductive development is also critical (Hall 1992); for example, in cowpea (Vigna unguiculata) the most sensitive stage of development to heat stress occurs 3 to 5 days before anthesis (Hall 1992). In cowpea, heat tolerant genotypes have been identified and the mechanism elucidated (Hall 1992). Heat tolerant genotypes of groundnut have been identified from field screening in West Africa (Greenberg et al. 1992; ICRISAT 1994; Wheeler et al. 1997). However, no studies on the effects of heat stress on the reproductive processes determining fruit number in tolerant and susceptible genotypes have been made. The objectives of the research reported here were, first, to examine the effect of high (40 28 C, mean 34 C) relative to near-optimum (30 24 C, mean 27 C) temperature imposed between flowering and reproductive maturity on leaf and flower appearance, and fruit number, in eight genotypes of groundnut varying in heat tolerance; and, second, to investigate the effect on fruit number of the time of exposure relative to first flowering of high temperature. MATERIALS AND METHODS The research was carried out in two controlledtemperature polyethylene-covered tunnels (polytunnels) at the Plant Environment Laboratory, University of Reading, Reading, UK (51 27 N lat. and W long.) between May and September Each polytunnel has a volume of 633 m and is aligned in an east west direction. They are naturally lit and the transmission measured at solar noon on a clear day was 77%. The average daily photosynthetic photon-flux density during the experiment, estimated from daily values of solar radiation measured outside the polytunnels and the percentage transmission, was about 500 µmol m per s. Day and night temperatures were controlled in the polytunnels by heating and venting. Each tunnel had an 88 kwh capacity heater which blows air down both sides of each tunnel through 0 3 m diameter plastic ducts with holes approximately every 0 2 m. Air was constantly recirculated within the polytunnel. In the roof of the tunnels were three polythene ducts connected to three fans for venting. The venting fans were switched on in two stages (to avoid dropping the temperature too rapidly) and created a partial vacuum in the polythene ducts, causing air to be sucked into the ducts at the opposite end of the tunnel. An aspirated and shielded thermocouple mounted at 1 5 m height in the centre of each tunnel and connected to a solid-state controller (Nobel Engineering, Bognor Regis) was used to control temperature by adding or losing (by venting) heat. Air temperatures were monitored continuously at canopy height in the centre and at both ends of the tunnels using aspirated and shielded thermocouples connected to a datalogger (Delta-T Devices, Cambridge, UK). The difference in mean temperature within the tunnel was 0 5 C. The target vapour presure deficit (VPD) during the day was 2 kpa in both tunnels. Relative humidity was monitored continuously by an RH sensor connected to the datalogger in the centre of each tunnel, and water added to the tunnel from a network of minisprays at ground-level when RH dropped below the target value in each tunnel. Genotypes Seeds of five Spanish (ssp. fastigiata) botanical type (796, , ICGV 86015, ICGV and ICGV 87003) and three Virginia (ssp. hypogaea) botanical type (47-16, and ICGV 87282) genotypes (ICRISAT 1994; Wheeler et al. 1997; Craufurd et al. 1999) of groundnut were provided by the ICRISAT Sahelian Centre, Niger. Four genotypes (796, , ICGV and 47-16) were classed as heat tolerant; the remainder were classed as heat susceptible (ICRISAT 1994; J. H. Williams, personal communication). Plant husbandry Seeds were inoculated with Rhizobium strain NC92 (Biocare Technology Pty, Somersby, NSW, Australia) before sowing into a general purpose compost in module trays (each cell mm). At 15 days after sowing (DAS) uniform seedlings were transplanted into 15 litre pots containing 14 kg of a mixture of dry sand, gravel, vermiculite and soilless compost mixed in proportions of 4:4:2:1 (v v). A commercial controlled-release fertilizer (0 15 kg kg N, 0 10 kg kg P, 0 12 kg kg K, 0 02 kg kg MgO plus trace elements: Osmocote, Scotts UK Ltd, UK) was incorporated into the mixture at a rate of 5 g l. Pots were weighed each day and hand watered to maintain them at field capacity (Craufurd et al. 1999). Sporadic infestation of red spider mite (Tetranychus urticae) was prevented using the predatory mite Phytoseulus persimilis and Cercospera leaf spot was controlled with chlorothalonil (2,4,5,6-tetrachloro-1,3-benzenedicarbonitrile).

3 Heat tolerance in groundnut 373 Temperature treatments All plants were grown at C from sowing to 32 DAS, when 50% of the genotypes had flowered. At 32 DAS, plants were then either kept at C (control) or transferred to an adjacent polytunnel maintained at C (high temperature) until harvest. Target daily temperatures were maintained throughout the experiment (S.D C) and mean temperatures in the C control and C high temperature regimes were 27 8 and 34 1 C, respectively. Plants were harvested at 98 and 119 DAS in the and C temperature regimes, respectively, when 2100 Cd (base temperature 10 C: Ong 1986) had been accumulated in each regime. The photoperiod and thermoperiod were both 12 h day. Relative humidity during the day was maintained between 50 to 70 and 70 to 90% in the and C treatments, respectively, in order to maintain vapour pressure deficit close to 2 kpa in both temperature regimes. Observations and data analysis Observations were made on all plants. The respective durations from sowing to first flower, peg, fruit (pod) and seed appearance were recorded. At flowering and at reproductive maturity, plants were separated into roots, stems (including petioles), leaves, pegs and fruits (seeds and fruit walls) and oven-dried at 80 C for at least 48 h to determine their dry matter. Total fruit dry matter values at harvest were adjusted by a factor of 1 76 to allow for the oil content of the seeds (Duncan et al. 1978). The numbers of pegs and fruits were also counted. The cumulative number of leaves on the mainstem was counted every 3 4 days from 25 to 80 DAS. A leaf was counted when the petiole was visible. The date the first flower opened was noted and thereafter the number of flowers opening each day was recorded until 80 DAS. The proportions (angular transformed) of flowers forming pegs (peg-set) and pegs forming fruits were derived from total cumulative flower number, and total peg and fruit number at maturity. The experiment comprised two unreplicated temperature treatments (tunnels), within each of which eight genotypes were arranged in five blocks. Pots were re-randomized every day following hand watering (Craufurd et al. 1999). There were no guard rows. Leaf, flower and fruit number, and parameters derived from them, were analysed by ANOVA (Genstat 5 Committee 1987) with temperature treatments as unreplicated mainplots and genotypes as subplots. The effects of temperature and genotype on the rates of leaf and flower appearance were analysed by comparison of regressions of cumulative leaf and flower number against days from sowing. RESULTS AND DISCUSSION Leaf appearance and number Leaves on the mainstem appeared more rapidly (P 0 001) between 25 and 80 DAS at C than at C and more leaves (P 0 001) were produced at C in all genotypes as a result (Fig. 1). There were also significant differences between genotypes at C in leaf appearance rate and mainstem leaf number at 80 DAS, which ranged from 0 19 to 0 31 leaves day and 15 6 to22 0, respectively. The common relationship between leaf appearance rate and leaf number at and C for all but three genotypes in Fig. 1 suggests that a temperature of C (mean 34 C) was not supraoptimal for leaf appearance in most genotypes, i.e. the thermal time per leaf was the same at and C. A comparison of leaf appearance rates at and C against thermal time or growing degree days (GDD, calculated using the mean daily temperature above a base temperature of 10 C; Ong 1986) confirmed that there was no significant effect (P 0 50) of temperature on thermal time for the appearance of successive leaves (θ) in four genotypes, values of θ ranging from 66 to 78 GDD leaf. However, in ICGV 86021, 796, and 47-16, θ was either increased or decreased by high temperature. In this experiment most of the groundnut genotypes were clearly able to develop vegetatively and grow as well at C asat30 24 C. Several other studies have shown similar results. For example, Talwar et al. (1999) reported significantly greater leaf number and vegetative biomass in three genotypes of groundnut grown at mean temperatures of 32 5 compared to 25 C, while Leong & Ong (1983) reported a constant value of 56 GDD leaf for cv. Robut 33-1 (now known as Kadiri-3) grown at mean temperatures from 19 to 31 C. Net photosynthesis rates have also Number of mainstem leaves Mainstem leaf appearance rate (leaves/day) Fig. 1. Relation between mainstem leaf number and leaf appearance rate per day in eight genotypes of groundnut grown at C ( ) and C ( ). Fitted line: y 5 1( 1 58) 54 0( 5 36)x, r 0 88, P Bar shows S.E.D.

4 374 P. Q. CRAUFURD ET AL. Table 1. Duration from sowing to first flowering at C and the effect of temperature on the total number flowers plant, the proportion of flowers forming pegs and the total number of fruits plant in eight genotypes of groundnut Proportion of flowers forming pegs Number of fruits plant Total number of flowers plant C C C C C C Duration (days) from sowing to first flowering Genotype ICGV-SM 87003* * * ICGV 86015* ICGV 86021* ICGV S.E.D. (7,64D.F.) * Spanish types. Virginia types. Cumulative flower number/plant Time (DAS) Fig. 2. Cumulative flower number plant against days from sowing (DAS) in the Spanish genotype ICGV-SM (, ) and the Virginia genotype (, ) grown at C (open symbols) and C (closed symbols). been reported to be 26% greater at 32 5 compared to 25 C (Talwar et al. 1999), and only reduced by 25% at 40 relative to 30 C (Ketring et al. 1982). However, Nigam et al. (1994) found plant growth rates in three Spanish genotypes to be lower at than at C. Nonetheless, these results support the view that groundnut is well adapted to warm climates and that the optimum temperature for vegetative development and growth is probably between 30 and 33 C, as Williams & Boote (1995) suggest. Flower number The total number of flowers plant produced between flowering and maturity at C ranged from 100 to 218 (Table 1). In general, Spanish genotypes produced fewer flowers than Virginia genotypes, and this was associated with fewer reproductive nodes and a shorter flower production period (Fig. 2) brought about by internal competition (Bunting & Elston 1980; Reddy 1988). High temperature significantly (P 0 001) increased the rate of flower production (e.g. Fig. 2) and the total number of flowers produced per plant (Table 1, Fig. 3) in several genotypes by more than 300%. However, thermal rates of flower production (flowers GDD) were significantly (P 0 05 to 0 001) less at than at C in all genotypes except ICGV and ICGV 86015, where rates were the same (P 0 10), and 47-16, where the rate was significantly (P 0 001) higher (not presented). Across both temperature regimes there was a good relation (r 0 77) between total flower number and mainstem leaf number (Fig. 3), and greater flower numbers at C can be partly explained by the increase in node (leaf) and axillary bud number per plant at this temperature (Bagnall & King 1991). Increased flower production at high temperature is

5 Heat tolerance in groundnut 375 Total flower number/plant Number of mainstem leaves Fig. 3. Relation between total flower number plant and mainstem leaf number in eight genotypes of groundnut grown at C ( ) and C ( ). Fitted line: y ( 141 4) 45 9 ( 6 73)x, r 0 77, P Bars show S.E.D. also associated with reproductive sterility; groundnut flowers are borne on an inflorescence, but usually the first fertilized ovary exerts dominance over later formed flowers on the inflorescence, preventing their development. However, high temperature prevents fruit set (see below) and so flowers continue to develop on each inflorescence resulting in an exponential increase in flower production, i.e. both rate and duration of flowering are increased by high temperature. Similar observations have been made by Bolhuis (1958) from studies in which flowers were removed every day. Duration to flowering, fruit set and fruit number The duration from sowing to first flowering at C varied from 28 to 31 days among Spanish genotypes and 31 to 41 days in Virginia genotypes (Table 1). The high temperature treatment was imposed at 32 DAS, and only in the latest flowering genotype, , was flowering delayed by high temperature, from 40 6 to45 2 DAS (P 0 001). There were significant differences between genotypes (P 0 001) in the proportion of flowers forming pegs (peg-set) at C, which varied from 0 46 to 0 84 (Table 1). High temperature significantly reduced peg-set in all genotypes and at C peg-set ranged from 0 40 to The Virginia genotype ICGV was particularly sensitive to high temperature. The proportion of pegs forming fruits also varied among genotypes (P 0 01), from 0 17 in to 0 61 in (not presented). High temperature, however, had no significant effect (P 0 20) on the proportion of pegs forming fruits among genotypes. There were significant differences between genotypes (P 0 001) at C in the number of fruits plant at harvest, which ranged from 25 to 52, Fruit number at 40/28 C (% 30/24 C) Time of first flowering relative to the start of the 40/28 C temperature regime (days) Fig. 4. Relationship between fruit number plant at C (as a proportion of fruit number at C, angular transformed) and the time the first flower opened relative to the start of the C regime in Spanish ( ) and Virginia (, ) genotypes. Fitted line, excluding ICGV ( ): y 38 9 ( 1 84) 3 6 ( 0 43)x; r 0 93, P and Virginia genotypes had more fruits than Spanish genotypes (Table 1). High temperature significantly reduced fruit number, particularly in Virginia genotypes where only 2 to 7 fruits plant were formed, compared with 13 to 26 in Spanish genotypes. There was no difference in fruit number between tolerant and susceptible Spanish or Virginia genotypes at either or C. Similar effects of high temperature on fruit-set and fruit number have been observed in a number of previous studies (e.g. Ketring 1984; Talwar et al. 1999; Vara Prasad et al. 1999a, b). Relationship between flower production and fruit yield It is clear that Virginia genotypes were more sensitive to high temperature than Spanish genotypes both in terms of absolute and relative effects on fruit number, i.e. irrespective of differences due to botanical type. However, Virginia genotypes were generally later flowering (Table 1) and initial rates of flower production are lower than in Spanish genotypes. Escape through earlier and more rapid development often contributes to genotypic differences in sensitivity to heat and drought stress (Harris et al. 1988), particularly if certain stages of reproductive development are very sensitive to stress. To examine the effect of escape, the proportion (%) of fruits (angular transformed) at relative to C was calculated within each genotype, to remove differences in yield potential between genotypes, and regressed against the duration after which individual genotypes first flowered relative to the start of the C temperature treatment (Fig. 4). There was a good relationship between relative fruit number and the timing of high temperature relative to

6 376 P. Q. CRAUFURD ET AL Vaiance (s 2 ) Number of fruits/plant Days from start of the 40/28 C temperature regime Fig. 5. Variance (s ) of regressions of fruit number at C against the cumulative number of flowers plant opening at C between first flowering and 0 to 9 days after the start of the C regime. flowering, with the later flowering genotypes having fewer fruits. Excluding ICGV 87282, this relationship was strongly linear (r 0 93; n 7; P 0 001) and relative fruit number decreased by 3 6% day as high temperature was imposed between 4 and 9 days after flowering. Quite clearly, therefore, greater fruit number in the Spanish genotypes was associated with earlier flowering, i.e. with escape from the deleterious effects of high temperature. Similarly, any putative differences in heat tolerance susceptibility were confounded with the overriding effect of the timing of flowering in relation to the onset of heat stress. Thus a susceptible genotype, ICGV-SM 87003, had the greatest fruit number because it was the earliest flowering genotype. Although early flowering and escape from stress were clearly important, the Virginia genotype ICGV flowered early (31 DAS), yet had a much smaller fruit number and yield than would be expected on the basis of time to flower. This may have been due to differences in the initial rate of flower production (Fig. 2). In order to verify that differences in the number of fruits produced at C were in fact due both to the timing of flowering and the number of flowers produced by the onset of the high temperature treatment, the number of fruits produced at C was regressed against the cumulative number of flowers produced in each genotype by the start of the C regime (day 0), and by each of the next 9 days, and the variances compared to determine the best relationship between fruit number and flower production (Fig. 5). The lowest variance was found when the number of fruits was regressed against the number of flowers produced between first flowering and 3 days after the start of the C treatment. There was a strong and significant relationship across all genotypes (r 0 95, n 8, P 0 001) between fruit number and the number of flowers that Cumulative number of flowers opening between first flowering and 3 days after the start of the 40/28 C temperature regime Fig. 6. Relationship between number of fruits plant at C and the cumulative number of flowers plant that opened between first flowering and the third day (35 DAS) of the C regime in Spanish ( ) and Virginia ( ) genotypes. Fitted line: y 4 0 ( 1 27) 0 64 ( 0 062)x; r 0 95, P opened between flowering and the start of the high temperature treatment (Fig. 6). The intercept of this relation was significantly (P 0 05) greater than zero, and in genotypes and 47-16, both of which flowered 3 days after the start of the heat stress treatment, four or more fruits were produced. The slope of this relationship gave a mean fruit-set (ratio of fruits to flowers) for all genotypes from these initial flowers of 64%. In contrast, more than 90% of later appearing flowers set no fruits. Clearly, fruit number and yield were closely related to the number of flowers produced within 3 days of the start of the high temperature treatment. This was due both to the timing of first flowering and the initial rate of flower production. The relationship shown in Fig. 6 can also be interpreted in terms of a critical or sensitive stage of development to high temperature. It has been shown in a number of crop species that it is not flowering per se that is the most sensitive stage of development to heat and water stress, but microsporogenesis (Hall 1992), which in groundnut occurs between 3 and 6 d before anthesis (Xi 1991). Recent research on responses to high temperature stress in groundnut (Talwar 1997; Vara Prasad et al. 1999b) has confirmed that the critical stage of development occurs from 3 to 5 days before anthesis. Thus, any flower buds at this critical stage (i.e. at anthesis 3 to 5 days) that experienced C would be unlikely to have set fruit. Conversely, any flower buds at the same stage of development that experienced C would probably have set fruit, even if they then subsequently experienced a warmer temperature during anthesis, i.e. all flowers which opened from buds where the critical stage occurred at the cooler temperature should have escaped the effect of high temperature.

7 Heat tolerance in groundnut 377 In conclusion, groundnut plants produced leaves faster at C than at C and were therefore well adapted in terms of vegetative growth and development to high temperature. Fruit number at high temperature was strongly related to the time of first flowering and to the number of flowers produced from flower buds that escaped high temperature during the critical period, around 3 days before flowers opened. Although genotypes varied in heat tolerance, they also varied in the time of first flowering, and this was the dominant attribute determining fruit number and yield in this temperature regime. This experiment therefore demonstrates again how important it is in abiotic stress studies to: (a) determine the critical stage of plant development to the stress; and (b) to understand and quantify the relationship between the timing of stress and the timing of the critical stage of development. We thank Messrs S. Caswell, S. Gill, K. Chivers and A. Pilgrim for technical support. This research was funded by the Plant Sciences Research Programme of the Department for International Development of the United Kingdom. However, the Department for International Development can accept no responsibility for any information provided or views expressed. BAGNALL, D. J.& KING, R. W. (1991). Response of peanut (Arachis hypogaea) to temperature, photoperiod and irradiance 2. Effect on peg and pod development. Field Crops Research 26, BOLHUIS, G. C. (1958). Observation of the flowering and fructification of the groundnut (Arachis hypogaea). Netherlands Journal of Agricultural Science 6, DE BEER, J. F. (1963). Influence of temperature on Arachis hypogaea L. with special reference to its pollen viability. Ph.D. Thesis, State University of Agriculture, Wageningen. BUNTING, A. H. & ELSTON, J. (1980). Ecophysiology of growth and adaptation in the groundnut: an essay on structure, partition and adaptation. In Advances in Legume Science (Eds R. J. Summerfield & A. H. Bunting), Vol. 1, pp London: HMSO. CRAUFURD, P. Q., WHEELER, T. R., ELLIS, R. H., SUMMERFIELD, R. J.& WILLIAMS, J. H. (1999). Effect of temperature and water deficit on water-use efficiency, carbon isotope discrimination and specific leaf area in groundnut. Crop Science 39, DUNCAN, W. G., MCCLOUD, D. E., MCGRAW, R. L. & BOOTE, K. J. (1978). Physiological aspects of peanut yield improvement. Crop Science 18, FORTANIER, E. J. (1957). Control of flowering in Arachis hypogaea L. Medelelingen van de Landbouwhogeschool Wageningen 57, GENSTAT 5COMMITTEE. (1987). Genstat 5 Reference Manual. Oxford: Clarendon Press. GREENBERG, D. C., WILLIAMS, J. H. & NDUNGURU, B. J. (1992). Differences in yield determining processes of groundnut (Arachis hypogaea L.) genotypes in varied drought environments. Annals of Applied Biology 120, HALL, A. E. (1992). Breeding for heat tolerance. Plant Breeding Reviews 10, HARRIS, D.H.,MATTHEWS, R.B.,NAGESWARA RAO, R.C. & WILLIAMS, J. H. (1988). The physiological basis for yield differences between four genotypes of groundnut (Arachis hypogaea) in response to drought. III. Developmental processes. Experimental Agriculture 24, ICRISAT (1994). ICRISAT West African Programs Annual Report Niger: ICRISAT. REFERENCES KETRING, D. L. (1984). Temperature effects on vegetative and reproductive development of peanut. Crop Science 24, KETRING, D. L., BROWN, R. H., SULLIVAN, G. H. & JOHNSON, B. B. (1982). Growth physiology. In Groundnut Science and Technology (Eds H. E. Pattee & C. T. Young), pp Yoakum, Texas: American Groundnut Research and Education Society. LEONG, S. K. & ONG, C. K. (1983). The influence of temperature and soil water deficit on the development and morphology of groundnut (Arachis hypogaea L.). Journal of Experimental Botany 34, NIGAM, S. N., NAGESWARA RAO, R. C., WYNNE, J. C., WILLIAMS, J. H., FITZNER, M. & NAGABHUSHANAM, G. V. S. (1994). Effect and interaction of temperature and photoperiod on growth and partitioning in three groundnut (Arachis hypogaea L.). genotypes. Annals of Applied Biology 125, ONG, C. K. (1986). Agroclimatological factors affecting phenology of groundnut. In Agrometeorology of Groundnut. Proceedings of an International Symposium, Aug 1985, Niamey, Niger (Eds M. V. K. Sivakumar, S. M. Virmani & S. R. Beckerman), pp Patancheru, India: ICRISAT. REDDY, P. R. (1988). Physiology. In Groundnut (Ed. P. S. Reddy), pp New Delhi: ICAR. SIVAKUMAR, M. V. K., MAIDOUKIA, A. & STERN, R. D. (1993). Agroclimatology of West Africa: Niger. Information Bulletin No. 5. Patancheru, India: ICRISAT. TALWAR, H. S. & YANAGIHARA, S. (1999). Physiological basis for heat tolerance during flowering and pod setting stages in groundnut (Arachis hypogaea L.). JIRCAS Working Report 14, TALWAR, H.S.,TAKEDA, H.,YASHIMA, S.&SENBOKU, T. (1999). Growth and photosynthetic responses of groundnut genotypes to high temperature. Crop Science 39, VARA PRASAD, P. V., CRAUFURD, P. Q. & SUMMERFIELD, R. J. (1999a). Fruit number in relation to pollen production and viability in groundnut exposed to short episodes of heat stress. Annals of Botany 84,

8 378 P. Q. CRAUFURD ET AL. VARA PRASAD, P. V., CRAUFURD, P. Q. & SUMMERFIELD, R. J. (1999b). Sensitivity of groundnut to timing of heat stress during reproductive development. Crop Science 39, VARA PRASAD, P.V.,CRAUFURD, P.Q.,SUMMERFIELD, R.J. & WHEELER, T. R. (2000). Effects of short episodes of heat stress on flower production and fruit-set of groundnut (Arachis hypogaea L.). Journal of Experimental Botany 51, WHEELER, T.R.,CHATZIALIOGLOU, A.,CRAUFURD, P.Q., ELLIS, R. H.& SUMMERFIELD, R. J. (1997). Dry matter partitioning in groundnut exposed to high temperature stress. Crop Science 37, WILLIAMS, J. H. & BOOTE, K. J. (1995). Physiology and modelling predicting the unpredictable legume. In Advances in Groundnut Science (Eds H. E. Pattee & H. T. Stalker), pp Stillwater, OK: APRES. WOOD, I. M. W. (1968). The effect of temperature at early flowering on the growth and development of groundnuts (Arachis hypogaea L.). Australian Journal of Agricultural Research 19, WRIGHT, G. C. & NAGESWARA RAO, R. C. (1994). Groundnut water relations. In The Groundnut Crop: A Scientific Basis for Improvement (Ed. J. Smartt), pp London: Chapman and Hall. XI, X. Y. (1991). Development and structure of pollen and embryo sac in peanuts (Arachis hypogaea L.). Botanical Gazette 152,