Crop Science Journal of Natural Resources and Life Sciences Education
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bissuel-Belaygue, C.
Right arrow Articles by Wery, J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Bissuel-Belaygue, C.
Right arrow Articles by Wery, J.
Agricola
Right arrow Articles by Bissuel-Belaygue, C.
Right arrow Articles by Wery, J.
Related Collections
Right arrow Water Stress
Right arrow Crop Physiology & Metabolism
Right arrow Seed Production
Crop Science 42:414-422 (2002)
© 2002 Crop Science Society of America

CROP PHYSIOLOGY & METABOLISM

Reproductive Development of White Clover (Trifolium repens L.) is Not Impaired by a Moderate Water Deficit That Reduces Vegetative Growth

II. Fertilization Efficiency and Seed Set

Christine Bissuel-Belaygue*,a, Alexander A. Cowanb, Athole H. Marshallb and Jacques Weryc

a ENSAR-INRA, UMR Sols-Agronomie-Spatialisation, 65 rue de Saint-Brieuc, 35042 Rennes, France
b Institute of Grassland and Environmental Research, Plas Gogerddan, Aberystwyth, Ceredigion, SY23 3EB, UK
c Agro.M-CIRAD-INRA, UMR SYSTEM, 2 place Viala, 34060 Montpellier, France

* Corresponding author (Christine.Bissuel{at}agrorennes.educagri.fr)


   ABSTRACT
TOP
 NOTES
 ABSTRACT
INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSION
 REFERENCES
 
The reproductive potential of white clover (Trifolium repens L.) plants is favored by moderate water deficits that reduce vegetative growth without impairing development of inflorescences, florets, and ovules. We examined the effects of this type of water deficit on reproductive characters involved in pollination, ovule fertilization, seed set, and seed filling. Two experiments were conducted under controlled conditions with four genotypes either in soil columns or vermiculite, and with various levels of water deficit that were maintained for 68 d. Treatments were classified as moderate (M) and severe (S) water deficit on the basis of the reduction of leaf relative water content (RWC) compared with the well-watered plants (C) and of previously established relationships between RWC, soil water potential and vegetative growth. Florets were hand pollinated and seed set was determined in all possible cross-pollinations between pollen and florets from plants subjected to C, M, or S treatments. Pollen viability declined as water deficit increased. Control and M plants had similar high levels of nectar production. In S plants, the proportion of florets containing nectar was reduced by 60 to 70%. Pollen from S plants induced a lower fertilization efficiency than pollen from C or M plants. Most of the reduction of seed set by severe water deficit, however, was linked to the decrease in female plant water status resulting in abortion of embryos. These results show that pollination, fertilization, seed setting, and seed filling can be limited by a severe water deficit applied during reproductive development. A moderate water deficit characterized by an avoidance of leaf dehydration and a reduction of vegetative growth did not impair fertilization efficiency and resulted in maximum seed set and thousand seed weight, compared to C and S plants.

Abbreviations: NFv, number of viable florets per inflorescence • NI, inflorescence number per reproductive stolon • NO, ovule number per floret • NSr, reproductive stolon number per plant • NSd, number of seeds per pod • RWC, relative water content • Y, seed number per plant


   INTRODUCTION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSION
 REFERENCES
 
IN INDETERMINATE LEGUMES, maximal seed yield is generally obtained by plants subjected to a moderate water deficit, in comparison with well-watered plants of white clover (Trifolium repens L.; Danyach-Deschamps and Wery, 1988; Oliva et al., 1994), alfalfa (Medicago sativa L.; Steiner et al., 1992), pea (Pisum sativum L.; Turc et al., 1990), or chickpea (Cicer arietinum L.; Wakrim-Mezrioui and Wery, 1995). This yield improvement can be explained by a greater number of inflorescences for white clover (Danyach-Deschamps and Wery, 1988; Oliva et al., 1994; Bissuel-Belaygue et al., 2002), or a lower level of seed abortion for pea (Turc et al., 1992). A moderate water deficit in white clover plants was characterized by a reduction of soil water potential (to -80 x 10-3 MPa) and avoidance of leaf wilting, but with significant reductions in stolon branching, leaf expansion and leaf production (Belaygue et al., 1996). On the other hand, it increases the proportion of reproductive to vegetative organs and the number of inflorescences compared with well watered plants (Bissuel-Belaygue et al., 2002). The development of inflorescences, florets, and ovules were largely unaffected by moderate water deficit (Bissuel-Belaygue et al., 2002). These results show that a moderate soil dehydration produces a crop canopy with a high reproductive potential (high ratio of reproductive vs. vegetative organ numbers and thus a high number of ovules per unit area). Nevertheless, subsequent phases of reproductive development in white clover (i.e., flower pollination, ovule fertilization, seed setting, and seed filling) could be adversely affected by water deficit (Thomas, 1987).

White clover is a cross-pollinated perennial with a gametophytic self-incompatibility system. For successful pollination, fertilization and seed set, a foraging insect must transfer compatible pollen from one plant onto the stigma of another. First, water availability may modify nectar production or quality and both affect pollination success. Pedersen (1953) found positive correlations in alfalfa between nectar production and honey bee visitation, and between honey bee visitation and seed production. A relationship was shown between nectar volume and number of seeds per pod (Teuber et al., 1983). Intensity and composition of emitted volatiles may indicate flower readiness for visitation or availability of nectar to pollinators (Robacker et al., 1983, 1988). Heinrich (1979) and Jakobsen and Olsen (1994) reported some field observations in white clover that support this hypothesis as they noticed that a bee will hover for a few seconds in front of the floret before entering or rejecting it without landing. Second, water stress could reduce pollen efficiency as observed on rice (Oryza sativa L.), (Ekanayake et al., 1990) and maize (Zea mays L.) (Westgate and Boyer, 1986). At last, one or several processes may be affected at the stigma level including stigma receptivity, pollen germination, and pollen tube growth through the style and ultimately to the ovary, and so play a major role in determining the success of fertilization, as shown in maize (Bassetti and Westgate, 1993a, 1993b). The duration of stigma receptivity in maize was reduced by water stress (Bassetti and Westgate, 1993a). The loss of turgidity of the papillae decreases the stigma receptivity, which determines the rate of rehydration of pollen grains (Schoper et al., 1987). The reduction of the silk water potential could also reduce the ability of style cells to sustain pollen tube growth as shown in maize (Bassetti and Westgate, 1993b).

Water deficit could result in increased ovule sterility, which is proposed by Pasumarty et al. (1993) as the major cause of the failure of 50% of the ovules to produce a mature seed at harvest in white clover seed crops (Zaleski, 1961; Van Bogaert, 1977; Pasumarty et al., 1993). Water deficit induces the abortion of younger seeds and ultimately the reduction of seed weight in grain legumes. The percentages of pod and seed abortion can be increased by water deficit in soybean [Glycine max (L.) Merr.] (Westgate and Peterson, 1993), and pea (Pisum sativum L.) (Ney et al., 1994). Research conducted on these self-fertile grain legumes has shown that seed abortion can only occur from fertilization to the beginning of seed filling (Ney et al., 1993). When a seed has reached this stage, which corresponds to the end of the cell division phase (Ney et al., 1993), a water deficit cannot induce abortion and the seed becomes a dominant sink for the plant (Ney et al., 1994).

The objectives of this study were (i) to analyze the response of seed yield of white clover plants over a range of water deficit treatments and (ii) to evaluate how and when yield was limited by respective male and female factors. We evaluated the contribution of the four post-flowering phases: (i) flower pollination, (ii) ovule fertilization, (iii) seed setting, and (iv) seed filling, to the drought-induced reduction of seed yield per plant.


   MATERIALS AND METHODS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
RESULTS
 DISCUSSION
 CONCLUSION
 REFERENCES
 
Experimental Conditions and Water Deficit Treatments
The two experiments reported in this paper, A and B, correspond to Experiments III and IV described in the companion paper by Bissuel-Belaygue et al. (2002). The experiments were conducted in a controlled environment chamber under the following conditions: 16 h photoperiod at 20°C with photosynthetic photon flux of 400 mmol m-2 s-1, 8 h at 15°C in darkness, and an air vapor deficit of 0.6 kPa. Four genotypes (G2 and G3 for Exp. A, G4 and G5 for Exp. B) differing in number of ovules per ovary (four for G3 and G4 or six for G2 and G5) and leaf type (small for G2 and G4 or intermediate for G3 and G5) were selected from cultivar Olwen (see Table 1 in Bissuel-Belaygue et al., 2002). The cross-compatibility between genotypes used in each experiment was confirmed before the required number of plants were cloned from each genotype.


View this table:
[in this window]
[in a new window]
  		Table 1. Cross combinations performed between two genotypes subjected to various water treatments: well–watered control (C), moderate (M), or severe (S) water deficit. Each genotype was used both as female and as pollen donor. Cross combination termed S x C, indicates that inflorescences from S plants (female) were pollinated with pollen taken from inflorescences of C plants (pollen donor).

 
Experiment A: Small Pot with Vermiculite–Moderate and Severe Water Deficits
Twelve plants of genotype G2 (small leaf type, 6 ovules per ovary) and 12 plants of G3 (intermediate leaf type, 4 ovules per ovary) were grown in vermiculite in plastic pots (0.10 m diameter, 0.20 m height). When the plants were 30 d old, the pots were placed on top of columns of foam blocks following a system similar to that described by Snow and Tingey (1985) to impose moisture stress. Their base was sealed with a fine nylon gauze which prevented root growth but allowed free movement of water. Plants were supplied with Jensen's nutrient solution (Jensen, 1942) minus nitrogen through a float valve that controlled the level of nutrient solution in the column. Three moisture treatments (4 plants per genotype and per treatment) were maintained for 68 d by adjusting the level of nutrient solution below the nylon gauze at 0.05 m in control treatments (C), at 0.15 m and at 0.25 m respectively for the first and the second levels of water deficit.

Experiment B: Small Column with Compost–Moderate Water Deficits
Eighteen plants of genotype G4 (small leaf type, 4 ovules per ovary) and 18 plants of G5 (intermediate leaf type, 6 ovules per ovary) were grown in columns (0.15 m diameter, 0.45 m height) filled with John Innes No. 1 compost (IGER Aberystwyth, Wales, UK). Plants were irrigated daily to maintain soil moisture content at 100% (treatment C), 65% (first level of water deficit), and 44% (second level of water deficit) of the available soil water for 68 d. Soil moisture content was checked by weighing columns daily.

Plant Water Status
Leaf relative water content (RWC) of most-recently fully expanded leaves was measured once a week at midday on four or six leaves per treatment (one per plant) as described by Belaygue et al. (1996). Bissuel-Belaygue et al. (2002) established that a range of reasonably constant water deficits were imposed during 68 d for both experiments, and that each treatment may be characterized by the average RWC over the water deficit period. Each water deficit was classified a posteriori as moderate (M) or severe (S), on the basis of the reduction of RWC compared with the control (C). Moderate water deficits, characterized by a reduction of mean RWC of less than 12% compared with the control were obtained in both experiments, while severe water deficits, characterized by a higher reduction of RWC and symptoms of leaf wilting, were obtained only in experiment A (Bissuel-Belaygue et al., 2002).

Nectar Volume and Sucrose Concentration
Nectar volume and sucrose concentration were measured for 12 inflorescences per treatment (10 florets per inflorescence). Only the florets, opened for at least 24 h, were sampled to ensure an optimal nectar secretion (Vansell, 1951, cited by Thomas, 1987). Nectar was collected by inserting a capillary tube (0.5 mm diameter) in 10 successive florets and nectar volume was estimated from the height of the nectar column in the tube, as described by Norris (1984). Nectar sucrose concentration was determined by placing the extracted nectar on a hand held refractometer (40 to 85% sugar refractometer, Bellingham & Stanley, Ltd, Tunbridge Wells, UK).

Pollen Viability, Pollen Germination and Pollen Tube Growth
Pollen from recently opened florets was tested for viability by staining with fluorescein diacetate (Heslop-Harrison and Heslop-Harrison, 1970). This procedure tests the integrity of the plasmalemma of the vegetative cell and the presence of an active esterase. Pollen viability estimated by this method was correlated with the results of a pollen germination test (Heslop-Harrison et al., 1984). Pollen samples were collected from 10 to 15 florets (opened for 24 h to avoid taking immature or old pollen) and dispersed on a glass slide on a drop of fluorescing medium. Observation of 200 pollen grains from 10 different areas of the slide was carried out under UV light to calculate the percentage of viable grains per sample. Five observations were carried out per treatment.

A few pollinated florets were collected at 0, 2, 4, 6, and 8 h after pollination to assess pollen germination on the stigma (2 h after pollination) and pollen tube growth within the style (completed 4 h after pollination). The styles were excised at their base, gently squashed on a microscope slide in 1:1 glycerol: 0.1% decolorized aniline blue prepared in 0.1 N K3PO4 (Martin, 1959) and observed under UV light. Four observations were carried out for each cross-pollination between genotypes G2 and G3 in Exp. A.

Cross-Pollinations
Within each experiment, the two genotypes submitted to the various levels of water deficit were crossed. All possible cross combinations (9) were made between G2 and G3 in Exp. A, and only five selected cross combinations were made between G4 and G5 in Exp. B (Table 1). For each cross, two inflorescences (one per genotype) at a similar stage of development were selected and 20 open florets were marked with a felt-tipped pen. Pollen was collected from florets of the first inflorescence (chosen for example on a control plant, C) and placed on the stigmas of the 20 selected florets of the second inflorescence (chosen for example on a plant submitted to severe water deficit, S), which resulted in the cross S x C. At the same time, a reciprocal cross C x S was made by pollinating the first inflorescence (C) with pollen collected from the second one (S). Twelve inflorescences were pollinated for each cross-combination (e.g., S x C).

Seed Set
Seed development was examined 10 d after pollination in Exp. A in a subsample of 20 florets per cross-combination. In white clover, most of the seed abortion occurs during this period of development (Marshall and Hides, 1993) and the number of aborted seed recorded 10 d after pollination is similar to the number of aborted seeds at harvest. Ten days after pollination, two types of abortion could be distinguished. The first type, termed early abortion, included non-fertilized ovules and seeds that aborted in the first few days following fertilization. Unfertilized ovules and early aborted seeds were completely dry and shriveled and so could not be distinguished. The second type, termed late abortion, corresponded to the abortion of developing seeds. These were larger in size than unfertilized ovules (allowing distinction with early aborted seeds) and were partly shriveled (allowing distinction from non aborted seeds).

Ripe inflorescences were harvested one month after pollination. In Exp. A, for each cross-combination, all pollinated florets were pooled and counted together. Total seed number and total seed weight were determined so that the average number of seeds per pod (NSd) and the thousand seed weight could be calculated. The percentage of seed abortion was calculated from the ratio between NSd and the average number of ovules per pod (NO). In Exp. B, five of the 20 pollinated florets per inflorescence were dissected under a microscope. For each sample, the mean number of seeds per pod, the mean number of aborted seeds and the mean thousand seed weight were determined.

Statistical Analysis
Experiments involved completely randomized designs with four (Exp. A) or six replications (Exp. B). Analysis of variance was performed for each genotype within each experiment. Differences between treatment means were evaluated by genotype by Fisher's least significance difference at P <= 0.05 (LSD0.05). Statistical analysis was not possible for nectar volume and sucrose content in Exp. B because these variables were determined from a single nectar sample collected with the same capillary tube from successive 120 florets of different inflorescences. Similarly, NSd (and percent seed abortion) and 1000 seed-weight in Exp. A were determined from a single sample of pooled seeds collected from the 12 inflorescences studied per cross-combination. Then variability of 1000 seed-weight cannot be statistically tested. Differences between treatments for NSd and percent seed abortion at harvest time have been tested with the LSD0.05 calculated on these variables 10 d after pollination and after demonstration of significant correlation between measurements made at harvest and 10 d after pollination (r = 0.83, P < 0.001).


   RESULTS
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
DISCUSSION
 CONCLUSION
 REFERENCES
 
Nectar Volume, Sucrose Content and Pollen Viability
Considerable variability was found in the proportion of florets containing nectar in control plants (from 43% in G5 to 96% in G3) (Table 2). Moderate water deficit (M) increased this proportion, while severe water deficit (S) greatly reduced it. Moderate water deficit was the only treatment in which the mature florets produced similar amounts of nectar (not shown), and where the average volume of nectar per floret, calculated from the volume collected from a sample of ten florets and from the proportion of florets containing nectar, was representative of each floret production (Table 2). This nectar production per floret cannot be statistically analyzed but it was higher in M than in C plants in all cases where it was measured. Severe water deficit plants had some florets with high nectar production but a significant proportion (up to 71% in G2) of the florets within an inflorescence had no nectar. Nectar sucrose concentration was higher in M and S plants compared with C plants for G2 and G3, while moisture deficit tended to reduce sucrose content for genotype G5 (Table 2).


View this table:
[in this window]
[in a new window]
  		Table 2. Effect of water deficit on the percentage of florets containing nectar, on nectar volume per 10-florets sample, on nectar volume per floret, on nectar sucrose content and on pollen viability.

 
Within each genotype, the more severe the water deficit, the more pollen viability was reduced (Table 2). There were great differences, however, between genotypes and experiments in the proportion of viable pollen produced.

Seed Set
No significant differences in ovule number per floret (NO) were observed in Exp. B while moderate water deficits resulted in slightly less ovules per floret compared with corresponding control or severe water deficit for both genotypes in Exp. A (Table 3). For each cross-combination described in Table 1, we examined final seed number per pod (NSd) in comparison with ovule number per ovary at flowering time (NO). In all cross-combinations, the number of aborted seeds per pod (NO - NSd) represented more than 30% of NO (Table 4). This occurred despite hand pollination which ensured that all selected florets were pollinated. The analysis of the cross-combinations most likely to occur under field conditions (i.e., those with same treatment for female and male plants: C x C, M x M, and S x S), reveals that moderate water deficits (M x M) tended to decrease the percentage of seed abortion compared with the control (C x C) for all genotypes in both experiments (Table 4). Although in Exp. A the number of ovules per floret (NO) was slightly reduced by moderate water deficit, the seed number per pod (NSd) at harvest time was not impaired or even increased by M treatments compared with the control (Table 3). In contrast, severe water deficit (obtained only in Exp. A) increased the percentage of seed abortion per pod (Table 4) and, consequently, reduced the final seed number per pod (NSd) (Table 3) compared with C x C and M x M. The negligible percentage of late seed abortions on C x C and M x M cross-combinations (Table 4) indicate that seed abortion resulted mainly from ineffective fertilization or cessation of seed development at an early stage. For the S x S cross-combinations, there was a high proportion of late abortion of developing embryos, especially for genotype G3 (Table 4).


View this table:
[in this window]
[in a new window]
  		Table 3. Effect of water deficit on number of ovules per ovary (NO) at pollination, on number of seeds per pod (NSd), and on thousand seed weight at harvest time. Data are reported for cross-combinations, C x C, M x M, and S x S.

 

View this table:
[in this window]
[in a new window]
  		Table 4. Comparison of female and pollen donor treatment on percentage of seed abortions per pod at harvest time and on percentage of late seed abortions per ovary examined 10 d after pollination time. Within each experiment, each genotype subjected to various water treatments (C, M, S) was used both as female and as pollen donor.

 
The number of seeds per inflorescence was not impaired or slightly increased by a moderate water deficit compared with the control (Fig. 1) , because the number of ovules per inflorescence was only slightly reduced by M treatment (Fig. 1) and the proportion of ovules per pod forming seeds was increased in M x M compared with C x C (Table 4). Severe water deficits decreased the number of ovules per inflorescence (Bissuel-Belaygue et al., 2002) and increased the percentage of seed abortion per pod (Table 4), resulting in a marked decrease in the final number of seeds per inflorescence in S x S (Fig. 1).



View larger version (49K):
[in this window]
[in a new window]
  		Fig. 1. Number of ovules (total bar) and number of seeds (black part of each bar) per inflorescence for cross-combinations performed between control plants (C x C), moderate water deficit plants (M x M) and severe water deficit plants (S x S). Data were obtained from the product of mean values of NO or NSd (Table 3) by the number of viable florets per inflorescence (Bissuel-Belaygue et al., 2002) and therefore could not be statistically analyzed.

 
The number of seeds per plant (Y) was estimated by the product of the measured components: reproductive stolon number per plant (NSr), inflorescence number per reproductive stolon (NI), number of viable florets per inflorescence (NFv), and seed number per floret (NSd), taken from Table 2 and 3 in Bissuel-Belaygue et al. (2002) and from Table 3 in this paper. Yield was closely related to the number of inflorescences per plant (Fig. 2) with a slightly different relationship for the genotypes with four (G3 and G4) and six ovules per floret (G2 and G5). Severe water deficit decreased both the proportion of ovules forming seed and inflorescence number per plant, resulting in a marked decrease in Y (Fig. 2). Yield reductions in M plants (compared with C plants) were observed only in Exp. B, but they were of limited extent because the large reductions of NSr were balanced by an increase of the other yield components.



View larger version (23K):
[in this window]
[in a new window]
  		Fig. 2. Contribution of inflorescence number to seed yield (i.e. number of seeds per plant) under various water deficit treatments for each type of genotypes, having four (filled symbols, G3 and G4) or six (open symbols, G2 and G5) ovules per floret. Each point represents one treatment (well-watered, circle; moderate water deficit, triangle; severe water deficit, diamond) in one experiment.

 
Thousand seed weight differed more between experiments and genotypes than between water deficit treatments (Table 3). Nevertheless, the trends show that M treatments were consistently associated with higher thousand seed weights.

Effect of Water Deficit on Male and Female Components of Seed Set
Crosses between plants with contrasting moisture treatments (C x M, M x C, C x S, S x C, M x S, and S x M) were examined in order to distinguish the contribution of the male component (pollen fertility) and the female component (ovule fecundity) to the reduction of the percentage of seed set by water deficit.

Dissection of the florets 4 h after pollination revealed no difference either in pollen germination on the stigma or in pollen tube growth in the style for cross-combinations made between M and C plants (i.e., C x C, C x M, M x C, and M x M). For the C x S and S x C cross-combination, a large proportion of pollen grains failed to germinate on the stigma and pollen tubes had grown more slowly in the style compared with what was observed in cross-combinations between C or M plants (data not shown).

The percentage of seed abortion per pod was similar for cross-combinations C x M and M x C and tended to be slightly lower (not always significantly) than in C x C (Table 4). In general, seed abortion was increased when one parent plant was an S plant. Seed abortion was always increased when severe stress was applied on the female plant (S x C or S x M) but may also be increased when severe stress was only applied on the male plant (C x S or M x S in G2 x G3).

The percentage of late seed abortion was negligible for all crosses made between C and M plants (Table 4). Severe water deficit applied on the female plant (S x C, S x M, and S x S) resulted in a higher and significant percentage of late seed abortion (15% in G2 x G3 and 20 to 41% in G3 x G2) compared with all the other cross-combinations.


   DISCUSSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
CONCLUSION
 REFERENCES
 
Factors of Pollination Success
The success of pollination in white clover seed crops depends upon pollinator activity, which is governed by the necessity for the insects to find nectar and/or pollen. Our experiments show that water deficit influenced both quality and quantity of nectar, but the effect was dependent on water deficit intensity (Table 2). A moderate water deficit favored nectar production compared with well watered conditions, as it produced inflorescences with more florets containing nectar and more nectar per floret, and resulted in nectar with a higher sucrose content (except for genotype G5 in Exp. IV). Plants submitted to a severe water deficit had greater average nectar content per floret (calculated only on florets containing nectar) and similar sucrose content compared with moderate water stressed plants, but a large proportion of their florets contained no nectar. Pedersen (1953) postulated that increased nectar production would enhance effectiveness of pollination in alfalfa and Holtkamp et al. (1992) reported that seed set per pod was correlated with both nectar volume and sucrose concentration. Intensity and composition of emitted volatiles may indicate flower readiness for visitation or availability of nectar to pollinators (Robacker et al., 1983; 1988). Despite the variability between genotypes and between experiments, an observation also reported by Norris (1984), our results suggest that a moderate water deficit would not limit the success of insect visitation and, therefore, pollination in white clover, while a severe water deficit may affect the number of pollinated florets through a high proportion of florets without nectar.

Pollen/Stigma Interaction
Increasing water deficit reduced significantly pollen viability (Table 2). Similar results have been obtained in rice by Ekanayake et al. (1990) who reported that a reduction of plant water potential contributed to reduced pollen viability. Microscopic observations of the stigmas (not shown), following pollination of florets of control plants by pollen from severe water deficit plants showed that the germination of pollen grains was inhibited or delayed, compared with pollen from C or M plants. This indicates that severe moisture deficits, applied during the whole period of male gamete development, have a harmful effect on the subsequent capacity of pollen to germinate. Whether a reduction in pollen germination as observed in our study would be sufficient to reduce fertilization efficiency in a white clover seed crop remains uncertain as very few viable pollen grains are required to achieve fertilization.

Water deficit may affect stigma receptivity, pollen germination and pollen tube growth as shown on maize (Bassetti and Westgate, 1993a, 1993b). No literature is available on these effects in white clover, but our microscopic examinations (not shown) of the stigmas, following application of pollen from control plants, indicated that both pollen germination and pollen tube growth were inhibited on stigmas of severe water stressed plants, compared with well-watered or moderate water stressed plants.

Ovule Fertilization and Seed Set
Despite reduced pollen viability by a moderate water deficit (Table 2), seed abortion per pod was slightly reduced when M pollen was applied to control plants (C x M) or to moderate water deficit plants (M x M) compared with the C x C cross-combination (Table 4). This indicates that a moderate water deficit applied to a pollen donor plant has no harmful effect on fertilization efficiency and seed development. But pollen grains produced by severe water stressed plants are less efficient for germination, pollen tube growth and/or ovule fertilization. In our experiments, the low viability of pollen from S plants (Table 2) may have accounted for a significant proportion of the increase in seed abortion recorded in C x S and M x S cross-combinations compared with C x C, C x M, M x C, and M x M (Table 4). Most of the lack of seed set recorded in C x S and M x S crosses, occurred during the days immediately following pollination and concerned what we termed early abortion. This suggests that the physiological status of pollen grains was sufficiently damaged by a severe water deficit to decrease ovule fertilization and/or the production of viable embryos. This contradicts results obtained in maize by Schoper et al. (1986) where low water status of pollen donor plants did not affect seed set. Nevertheless, their water deficit treatment is difficult to compare with ours, as it was imposed for a shorter time, and only during the latter period of pollen maturation.

Both stigma receptivity and water status of female plants during seed development are important components of the marked decrease in the number of seeds per pod (NSd in Table 3) in severe water deficit conditions. A major part of this decrease can be explained by an increase in the percentage of abortion of developing seeds (late abortion, Table 4) rather than by a lack of fertilization or of development of fertilized embryos (early abortion). Whether this increase in seed abortion was due to a lack of assimilates during seed development as postulated by Khrbeet et al. (1993), or was due to an alteration of seed water status is not clear. Severe water deficits applied in our experiments considerably reduced plant leaf area (Belaygue et al., 1996), photosynthetic rate, and stolon dry matter accumulation (Bissuel-Belaygue et al., 2002), which could have led to a lack of assimilates for the reproductive sinks, as shown in soybean (Egli and Yu, 1991) and in maize (Schussler and Westgate 1991; 1994). Low water status of ovaries may also have reduced their ability to metabolize the assimilates (Zinselmeier et al., 1995), thereby decreasing their sink activity as shown on soybean (Westgate and Peterson, 1993) and maize (Schussler and Westgate, 1995). The low water status of the female plants would contribute directly (through ovary water status) or indirectly (through reduced carbon assimilation) to the decrease in seed number per pod by severe water deficit. In contrast, a moderate water deficit imposed on female plants increased the percentage of ovules forming seeds and the number of seeds per floret compared with the control treatment, indicating that neither fertilization efficiency nor seed development were impaired. This positive effect of a moderate water deficit on the percentage of ovules forming mature seeds may be related to resource allocation. Recent investigations on soybean showed a positive relationship between an increase in pod and seed number per node and increase in nodal carbon supply (Bruening and Egli, 1999). Since a moderate water deficit reduces vegetative growth (Belaygue et al., 1996) without impairing dry matter accumulation or reproductive development of growing stolons (Bissuel-Belaygue et al., 2002), it can be postulated that more resources are available to sustain the development of seeds. Field results support this hypothesis as minor reductions in levels of irrigation can increase white clover seed yield by increasing seeds per floret and inflorescence density (Clifford, 1986; Danyach-Deschamps and Wery, 1988; Oliva et al., 1994).

Seed Filling
No detrimental effect of water deficit was observed on 1000 seed-weight in our experiments (Table 3) or in field studies (Clifford, 1986; Danyach-Deschamps and Wery, 1988; Oliva et al., 1994). The trend for a greater 1000 seed-weight in moderate and severe water deficit plants compared with control ones might be attributed to a higher availability of assimilates for the filling seeds, resulting from the reduction of vegetative sinks (moderate water deficit) or from the reduction of the number of seeds (severe water deficit). These results confirm that in indeterminate legumes, seed weight is less affected by water deficit than seed number, as plants respond to water deficit by maintaining only the seeds that can be filled with the available assimilates, as shown in pea by Ney et al. (1994).

Implication for Seed Production
Our results show that the potential seed number per plant established at flowering (Bissuel-Belaygue et al., 2002) can be restricted by water deficit at any subsequent phase of reproductive development: pollination, fertilization, seed setting, and seed filling. The response of seed set and thousand seed weight depends on the intensity of the water deficit and on whether it is applied to the pollen donor plant and/or to the female plant. An unlimited water supply results in excessive vegetative growth which is detrimental to seed yield more in terms of inflorescence production (Bissuel-Belaygue et al., 2002) than in seed set per floret. A severe water deficit limits both vegetative growth (Belaygue et al., 1996) and reproductive potential (Bissuel-Belaygue et al., 2002), and impairs fertilization efficiency, seed set, and seed development by detrimental effects on pollen viability, stigma receptivity, and assimilate availability during seed development. A moderate water deficit results in maximal numbers of inflorescences per stolon (Bissuel-Belaygue et al., 2002) and in a higher percentage of ovules developing into mature seeds with slightly higher seed weight. This is consistent with results obtained in numerous field studies on white clover (Clifford, 1986; Danyach-Deschamps and Wery, 1988; Oliva et al., 1994) showing that an optimum level of soil dehydration is beneficial to seed yield. Our results have led us to define this level as a reduction of soil water potential in the rooting zone sufficient to inhibit stolon branching and induce a reduction of leaf growth without causing leaf wilting in the mature leaves (Belaygue et al., 1996) and without reducing net photosynthetic rates or dry matter accumulation in growing stolons (Bissuel-Belaygue et al., 2002).

Our studies were designed above all to examine the effect of soil water availability on the processes that may be important determinants of the seed yield of white clover. The results showed that the individual genotypes could be more or less drought sensitive for vegetative and reproductive components (Belaygue et al., 1996; Bissuel-Belaygue et al., 2002), and for the impact of male and female factors on success of fertilization and seed set (Table 2, 3, and 4). This suggests a white clover crop grown for seed production under a range of water limited conditions could be associated with some genetic shift compared with a crop grown under well-water conditions. Further studies are required to evaluate if the genetic composition of the cultivar could be modified by the conditions of water deficit experienced during the process of seed production.


   CONCLUSION
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSION
REFERENCES
 
The analysis of yield components of the cross-pollinations made between plants submitted to various levels of water deficit provides a useful tool to understand how yield was limited by respective male and female factors. Pollen was sufficiently damaged by a severe water deficit to decrease seed set by reducing the fertilization efficiency and/or the production of viable embryos. A severe water deficit applied to female plants increased abortion of developing seeds rather than reducing fertilization efficiency. Moderate water deficit did not impair fertilization efficiency and slightly increased seed set compared with the control. Although the use of cross-pollinations adds much to understanding reproductive ecology as related to water stress, further investigations are needed to establish the physiological basis of the differential effect of soil water deficit on growth of vegetative and reproductive organs and to understand how resource allocation between these organs is influenced by moisture deficit.


   ACKNOWLEDGMENTS
 
Helpful suggestions from the editor and the anonymous reviewers are gratefully acknowledged. The authors thanks Philippe Naudin for technical assistance in preparation and maintenance of greenhouse experiments.


   NOTES
TOP
 NOTES
ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSION
 REFERENCES
 
This work was supported by INRA, IGER via Biotechnology and Biological Science Research Council (BBSRC), RAGT company (Rodez, France), and European Union (EC Air 3 CT92-0279).

Received for publication February 18, 2001.


   REFERENCES
TOP
 NOTES
 ABSTRACT
 INTRODUCTION
 MATERIALS AND METHODS
 RESULTS
 DISCUSSION
 CONCLUSION
 REFERENCES
 




This article has been cited by other articles:


Home page
 Crop Sci.Home page
C. Bissuel-Belaygue, A. A. Cowan, A. H. Marshall, and J. Wery
Reproductive Development of White Clover (Trifolium repens L.) is Not Impaired by a Moderate Water Deficit That Reduces Vegetative Growth: I. Inflorescence, Floret, and Ovule Production
Crop Sci., March 1, 2002; 42(2): 406 - 414.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available
Right arrow Figures Only
Right arrow Full Text (PDF) Free
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Services
Right arrow Similar articles in this journal
Right arrow Similar articles in ISI Web of Science
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via ISI Web of Science (4)
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Bissuel-Belaygue, C.
Right arrow Articles by Wery, J.
Right arrow Search for Related Content
PubMed
Right arrow Articles by Bissuel-Belaygue, C.
Right arrow Articles by Wery, J.
Agricola
Right arrow Articles by Bissuel-Belaygue, C.
Right arrow Articles by Wery, J.
Related Collections
Right arrow Water Stress
Right arrow Crop Physiology & Metabolism
Right arrow Seed Production

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
The SCI Journals Agronomy Journal Vadose Zone Journal
Journal of Plant Registrations Soil Science Society of America Journal
Journal of Natural Resources
and Life Sciences Education		 Journal of
Environmental Quality