Reproduction and offspring competitiveness of Senna obtusifolia are influenced by nutrient availability

Authors

  • Kimberly D. Tungate,

    1. Department of Crop Science, North Carolina State University, Raleigh, NC, 27695–7620, USA;
    Search for more papers by this author
  • David J. Susko,

    1. Department of Plant Sciences, University of Western Ontario, London, ON, N6A 5B7, Canada
    Search for more papers by this author
  • Thomas W. Rufty

    Corresponding author
    1. Department of Crop Science, North Carolina State University, Raleigh, NC, 27695–7620, USA;
      Author for correspondence:Thomas W. Rufty Jr Tel: +919 515 3660 Fax: +919 515 5315Email: tom_rufty@ncsu.edu
    Search for more papers by this author

Author for correspondence:Thomas W. Rufty Jr Tel: +919 515 3660 Fax: +919 515 5315Email: tom_rufty@ncsu.edu

Summary

• Senna obtusifolia (sicklepod) is a troublesome weed in many agronomic cropping systems in the southeastern USA with varying fertility regimes. This series of experiments was initiated to determine the impact of nutrient availability on reproductive output and maternal effects.

• Experiments were conducted with plants growing in soil in growth chambers for 120 d. Offspring growth was evaluated in hydroponics.

• Plants grown under higher nutrient conditions had greater reproductive biomass, number of seeds and total seed biomass. Individual seed mass distributions were slightly affected, with negative skewness decreased at higher nutrition. Seed germination rates were influenced primarily by seed size, as small seeds germinated faster than large seeds. The nitrogen content of seeds increased with increasing seed size and was higher in the high nutrition treatment. Seedlings arising from maternal plants grown under high nutrition grew more rapidly than did seedlings arising from maternal plants grown in a low nutrient regime.

• The results suggest that S. obtusifolia reproduction and offspring competitiveness can be strongly influenced by the fertilization regimes used in different agronomic crops and rotations.

Introduction

The reproductive success of an individual plant is determined not only by prevailing environmental conditions during its life cycle, but also by the environmental conditions experienced during growth and maturation of its maternal parent. During parental development, environmental factors such as temperature, photoperiod, light quality, water availability and nutrient levels can modify the size, elemental composition, percentage and rate of germination, and longevity of seeds, as well as the growth and survivorship of offspring in the seedling and/or adult stage of development (Roach & Wulff, 1987; Fenner, 1992; Gutterman, 1992; Donohoe & Schmitt, 1998).

With the extensive use of fertilizers in agroecosystems, the maternal effects of nutrition on seed quantity and quality have been studied intensively in agronomic and vegetable crops. In numerous studies with cultivated species, increased mineral nutrition has been shown to increase both seed mass and the concentrations of mineral nutrients and amino acids in seeds (Gray & Thomas, 1982; Altman et al., 1983; Fenner, 1992; Bruckner et al. 1998; Oikeh et al., 1998).

In contrast, with weed species the maternal effects associated with increasing parental nutrition have been inconsistent. Seed size has been observed to increase (Parrish & Bazzaz, 1985; Stratton, 1989; Sills & Nienhuis, 1995) or remain constant (Fenner, 1986). The nutrient content of seeds has increased markedly with increased maternal nutrition in some cases (Fawcett & Slife, 1978; Saini et al., 1985), but not in others (Fawcett & Slife, 1978; Fenner, 1986;Van Andel & Vera, 1977). Variable effects have also been observed on seed germination, with germination being unaffected (Fawcett & Slife, 1978; Parrish & Bazzaz, 1985; Sills & Nienhuis, 1995) and decreased (Galloway, 2001) in different experiments.

Senna obtusifolia is a nonnodulating, annual monocarpic weed in the family Fabaceae. Numerous flowers are produced on indeterminate inflorescences in late summer and early fall. Multi-seeded fruits are long (6–15 cm), dehiscent, sickle-shaped pods. Senna obtusifolia is widespread in the southeastern United States and ranked as one of the most troublesome weeds in major crops such as maize, peanut and soybean (Dowler, 1995). Populations of S. obtusifolia can experience a wide range of fertilization regimes in varying crop rotations. With maize, for example, nitrogen fertilizer is supplied in the 150–200 kg ha−1 range, while it is common for nitrogen fertilizer to be omitted entirely with N2-fixing peanut and soybean. In this study, we examine the influence of nutrition on reproduction and maternal effects of S. obtusifolia. The intent of the research was to begin assessing whether fertilization patterns associated with specific crop rotations can lead to differential reproductive performance and competitiveness of this agronomically important weed species.

Materials and Methods

Plant culture

Senna obtusifolia seeds were obtained in spring 1998 from Azlin Seed Service (Leland, Mississippi). The seeds were sown in 0.6-l (11 × 10 cm) plastic pots containing a standard greenhouse soil mix and placed in walk-in growth chambers in the Southeastern Plant Environment Laboratory at North Carolina State University (Downs & Thomas, 1991). The chambers were programmed for 12 h photoperiods and day/night temperatures were maintained at 32/27°C. A combination of metal halide, high-pressure sodium, and incandescent lamps supplied a PPFD (400–700 nm) of c. 1100 µmol m−2 s−1.

Emerging seedlings were allowed to establish for 1 wk before plants were thinned to a single seedling per pot. At the time of thinning, 150 plants were assigned at random to one of three nutrient addition treatments: low, 250 ml solution added once every 7 d; medium, 250 ml solution added once every 2 d; or high, 250 ml solution added once every day. The treatment solution contained 7.6 mM N, 0.3 mM P, 2.8 mM K, 1.4 mM Ca, 0.5 mM Mg and micronutrients at 1/4 strength Hoagland’s solution. In addition, each pot was flushed with deionized water prior to nutrient additions to minimise residual nutrient accumulation. Plants were rearranged on benches every week to minimise within chamber location effects. Seven weeks after the start of the experiment, plants were transplanted into larger 9.0-l pots (25 × 19 cm) to lessen pot binding of root systems.

Measures of growth and reproduction

Beginning 2 wk after the start of the experiment, four plants per treatment were harvested at 3–4 d intervals for the following 3 wk. Thereafter, harvests were made every 5–10 d until wk 9, and a set of 10 plants (per treatment) was carried until maturity 120 d after the start of the experiment. At harvest, leaf area was measured using a Li-Cor Model LI-3100 Area Meter, and shoot tissues were placed in a drying oven at 60°C for 48 h and then weighed. After the onset of reproduction, above-ground biomass was separated into reproductive (pedicels, flowers, fruits, seeds) and vegetative tissues (stem and leaves), which were dried separately. Fruits were collected as they ripened (but prior to dehiscence) during the reproductive period. At the time of collection, fruit position on an individual plant was recorded. Also, seed number per fruit, and total seed mass per fruit were determined. Subsequently, cumulative fruit and seed production and seed mass were calculated on a per plant basis. One-way ANOVA (SYSTAT, 1997) was performed on untransformed data to test for the effect of fertilization level on various measures of growth and fecundity. Post-hoc Tukey tests were used to contrast treatments when significant effects were encountered.

Bulk collections of seeds were obtained by pooling seeds from the 10 plants grown to maturity in each nutrient treatment condition. In order to determine the distribution of seed mass in each treatment, 300 seeds were subsampled randomly and weighed to the nearest 0.01 mg using a microbalance (Mettler AT20, Mettler Instrument Corp., Hightstown, NJ, USA). The distributions of seed masses were compared with normal frequency distributions using t-tests (Zar, 1996).

Germination studies

A mesh sieve was used to separate groups of seeds from the bulk collections into small (seed < 2.4 mm in width) and large (seed > 2.4 mm in width) size classes, with 300 seeds from each seed-size class within each nutrient treatment randomly selected and weighed individually. Across nutrient treatments, mean seed mass of small and large seeds was 16.17 ± 0.44 mg and 26.42 ± 0.51 mg, respectively. Seeds were stored at room temperature (22–25°C) prior to use in germination experiments and analyses of chemical composition. Seeds exhibiting visual signs of damage by pathogens were not used. A two-way factorial experiment was initiated to determine the effects of seed size and maternal nutrient regime on seed germination. The germination experiment consisted of three replications of 50 seeds from each combination of seed-size class (small, large) and nutrient treatment (low, medium, high). Seeds were placed on two sheets of blotter paper in 9 cm plastic Petri dishes. The blotter paper was moistened with 5 ml of distilled water. Additional distilled water was added as needed to ensure adequate moisture over the duration of the experiments. The Petri dishes were placed in polyurethane bags to slow desiccation. Germination studies were conducted in temperature- and light-controlled environmental growth chambers. Day/night temperatures of 30/20°C were maintained for 12 h each. A 14-h daily photoperiod was maintained in each chamber, with the light period extending from 1 h before to 1 h after exposure to the daily high temperature. Fluorescent lamps produced a PPFD of 150 µmol m−2 s−1. Seed germination was monitored daily for the first 10 wk, and once per week thereafter. The experiment was terminated after 140 d. The criterion for seed germination was radicle emergence. Prior to statistical analysis, final cumulative percentage germination values were arcsin-square-root transformed. The coefficient of velocity was calculated as an estimate of the rate of germination as per the formula of Horak & Wax (1991). Separate, two-way ANOVAs (SYSTAT, 1997) were used to assess the effects of maternal nutrient level and seed-size class on germination rate and final percentage germination.

Nitrogen composition of seeds

Eighty seeds from the low and high maternal nutrient treatments were randomly selected from a range of seed mass classes, established from the mass distribution curves, and then weighed individually. The seeds were analyzed for total N with an automated CHN analyzer (Perkin Elmer 2400). Individual seed mass was regressed against seed N content (%N × mass) using linear regression analysis (SYSTAT, 1997). Comparisons of slope and elevation of the two linear regression equations were performed as per the recommendations of Zar (1996). Nonlinear regressions were also performed, but they explained no more variance than linear models, so only linear regression results are reported.

Seedling growth experiments

The effects of maternal nutrition and concurrent nutrient regime on offspring growth were evaluated in a 21-day experiment using a two-way factorial design. One hundred seeds from the bulk seed collection for each maternal nutrient treatment were separated into large and small size classes as above and germinated in paper rolls in a germination chamber for 72 h in the dark. The paper rolls were placed vertically into 4-l beakers and kept moist by capillary action from 200 ml of 0.1 mM CaSO4 solution. Upon radicle emergence, 40 seedlings of uniform size were transferred into 50-l continuous flow, solution culture systems. Twenty seedlings originating from each seed-size class (small vs large) were grown in a continuous flow hydroponics system containing a modified Hoagland’s nutrient solution with or without 0.6 mM KNO3. Solution pH was automatically monitored and maintained at 5.8 ± 0.2, and temperature was maintained at 30 ± 0.5°C. Seedlings were harvested 21 d after placement into the solution culture systems. Leaf area was measured, and plant tissues were oven dried and weighed. Separate three-way ANOVAs (SYSTAT, 1997) were used to determine the effects of seed-size class, maternal nutrient regime, offspring nutrient regime, and their interaction on total leaf area, root mass, and shoot mass of seedlings.

Results

Growth and reproduction

The nutrient treatments of the parental plants were successful in establishing different growth rates and plant sizes. Increased application of nutrients resulted in significant increases in total leaf area (Fig. 1; F2,9 = 1228.4, P < 0.001 at 66 d) and total plant biomass (Fig. 2; F2,9 = 63.6, P < 0.001 at 66 d). The onset of reproduction began about 35 d after the start of the experiment. The percentage of biomass allocated to reproductive structures did not differ significantly among nutrient treatments, as they accounted for 26.0 ± 4.3%, 22.5 ± 1.5%, and 23.6 ± 7.7% of the whole plant biomass in the low, medium, and high nutrient treatments, respectively, at the 66 d harvest (F2,9 = 0.13, P= 0.883).

Figure 1.

Effect of parental nutrient regime on leaf area of individuals of Senna obtusifolia over a 67-d growth period. Triangles, low; diamonds, medium; circles, high nutrient addition treatment.

Figure 2.

Effect of parental nutrient regime on reproductive and vegetative biomass of individuals of Senna obtusifolia over a 67-d growth period. (a) low, (b) medium and (c) high nutrient addition treatments. Filled area, reproductive tissues; hatched, vegetative tissues.

Total fruit and seed production, and the total biomass of seeds per plant differed significantly among nutrient treatments (Table 1). Exposure to the lowest level of nutrients led to plants with fewer fruits and seeds, and lower total seed biomass. In contrast, the number of seeds per fruit decreased as nutrient levels increased. Mean individual seed mass was similar among nutrient treatments.

Table 1.  Means (± SE) of reproductive components for plants of Senna obtusifolia grown under three soil nutrient conditions. Also shown are results of one-way ANOVA testing for effects of nutrient treatment
ParameterMaternal nutrient conditionsF
LowMediumHigh
  1. Values with different superscripts (within a row) differ significantly using Tukey tests at P < 0.05. **P < 0.01; *P < 0.05; NSP > 0.05.

N1077 
No. of fruits per plant 6.2b± 1.4 14.0ab± 2.9 20.3a± 3.48.8**
No. of seeds per plant 90.5b± 16.4 163.1ab± 31.7 190a± 33.64.2*
Total seed mass per plant (g) 2.70b± 0.50 4.17ab± 0.82 5.51a± 0.904.0*
No. of seeds per fruit 14.5a± 0.7 11.7b± 0.6 10.5b± 0.412.1***
Mean seed mass (mg)29.30 ± 1.0625.80 ± 1.1330.39 ± 2.022.6NS

Seed mass, germination, and composition

Even though individual seed masses were similar among treatments, random sampling of the seed populations indicated that differences in seed mass distribution existed (Fig. 3). Mass distributions were significantly negatively skewed in each nutrient treatment (low g1 = −1.008, t = −7.148, P < 0.001; medium, g1 = −0.878, t=−6.227, P < 0.001; high g1 = −0.520, t= −3.688, P < 0.001), but the degree of negative skewness decreased as nutrient level increased. The mass distribution of seeds from the low nutrient treatment (g2 = 0.718, t = 2.555, P < 0.01) was leptokurtic, while the degree of kurtosis of the seed mass distributions for medium (g2 = 0.504, t = 1.794, P > 0.05) and high nutrient treatments (g2 = −0.520, t=−1.850, P > 0.05) did not differ significantly from normality. Seed mass ranged 2.6, 3.1 and 2.2-fold within low, medium, and high nutrient treatments, respectively.

Figure 3.

Effect of parental nutrient regime on distribution of individual seed mass of Senna obtusifolia. (a) low, (b) medium and (c) high nutrient addition treatments.

The nutrient treatments had no statistically significant effect on germination. Instead, the dominant influence on seed germination was seed size. Germination began earlier and was more rapid in the small seed-size class than the large seed-size class across all nutrient conditions (Table 2; Fig. 4). Over time, the percentage germination of large seeds tended to exceed that of small seeds, although final percentages differed significantly only in the high nutrient treatment (as demonstrated by a seed-size class × nutrient treatment interaction; Table 2).

Table 2. F-values of two-way analyses of variance of the effects of maternal nutrient levels (N) and seed-size class (S) on germination characteristics of Senna obtusifolia
 NSN × S
  • ***

    P < 0.001;

  • *

    P < 0.05;

  • NS

    P > 0.05.

df2, 1239.61***1.32NS
Coefficient of velocity1.36NS23.95***4.30*
Final percentage germination2.17NS23.95***4.30*
Figure 4.

Effect of parental nutrient regime and seed-size class (triangles, large seeds; circles, small seeds) on germination of seeds of Senna obtusifolia. (a) Low, (b) medium and (c) high nutrient addition treatments.

Total N content in seed increased as seed mass increased in low (y = 0.038x − 0.147; r2 = 0.938) and high nutrient treatments (y = 0.039x – 0.099; r2 = 0.930) (low F1,61 = 940.5, P < 0.001; high F1,51 = 691.0, P < 0.001 (Fig. 5)). The regression coefficients (slopes) of the linear regression equations for N content vs seed mass did not differ significantly (t112 = 1.504, P > 0.05), so a similar increase occurred. However, the elevations (y-intercepts) of the linear regressions were different (t113 = 3.731, P < 0.001), indicating that seeds of the same mass had higher N content when produced by maternal plants grown in a high nutrient treatment condition.

Figure 5.

Scatterplot showing the relationship between nitrogen content and individual seed mass for seeds produced by parental plants grown under low (circles) and high (triangles) nutrient regimes. Linear regressions of N content vs seed mass for seeds from plants grown in low (solid line: y= 0.038x − 0.147; r2 = 0.938; F1,61 = 940.5) and high nutrient regimes (dashed line: y= 0.039x – 0.099; r2 = 0.930; F1,51 = 691.0) were significant at P < 0.001.

Seedling growth experiments

A series of hydroponics experiments was conducted to determine whether the maternal nutrient regime affected vigour of offspring. After germination, seedlings were grown in the presence and absence of a nitrogen source. After 21 d of growth, significant maternal nutrient effects could be detected (Fig. 6, Table 3). Offspring of plants from the high maternal nutrient regime had greater biomass and total leaf area than did those from plants grown in the low maternal nutrient regime. The maternal effects were apparent within both seed-size classes. The maternal effects were amplified in offspring grown in the presence of nutrients, where seedling growth was considerably greater than that in the absence of nutrients (note different scales on y-axis). The maternal nutrient regime × seed-size class, seed-size class × offspring nutrient regime, and maternal nutrient regime × offspring nutrient regime interaction terms were significant for all growth parameters (Table 3). Hence, the expression of maternal nutrient effects depended on both seed size and the nutrient environment of offspring. The three-way interaction between maternal nutrient regime × seed-size class × offspring nutrient regime was also significant for all growth parameters (Table 3). Thus, while larger seeds produced larger plants than did small seeds, the effect was greatest for offspring of high nutrient parents grown in high nutrient conditions.

Figure 6.

Effect of seed-size class (open bars, small seeds; closed bars, large seeds), parental nutrient regime and offspring nutrient regime on shoot and root biomass of Senna obtusifolia.

Table 3. F-values of three-way analyses of variance of the effects of seed-size class, maternal nutrient regime and offspring nutrient regime on leaf area, shoot mass, and root mass of 3-wk-old-seedlings of Senna obtusifolia
 dfLeaf areaShoot massRoot mass
  • ***

    P < 0.001;

  • **

    P < 0.01;

  • *

    P < 0.05.

Seed-size class1, 73 62.2***103.0***46.8***
Maternal nutrient regime1, 73 49.7*** 78.1***47.9***
Offspring nutrient regime1, 73409.8***325.0***32.2***
Seed-size class × Maternal nutrient regime1, 73  7.4** 16.4***15.8***
Seed-size class × Offspring nutrient regime1, 73 54.2*** 60.5***20.0***
Maternal nutrient regime × Offspring nutrient regime1, 73 43.5*** 44.8***10.4**
Seed-size class × Maternal nutrient regime × 1, 73  8.5** 13.0** 6.6*
Offspring nutrient regime

Discussion

The results indicate that nutrient availability strongly influences reproductive performance of S. obtusifolia. Increases in leaf area and plant biomass with increased nutrient supply were associated with increases in fruit production and total seed mass and number. The growth and fecundity responses to increased resource conditions are typical for many species (Winn & Werner, 1987; Schwaegerle & Levin, 1990; Wulff et al., 1994; Weiner et al., 1997). The proportion of biomass allocated to reproduction did not differ significantly among nutrient treatments in S. obtusifolia (Fig. 2). Similarly, the proportion of biomass produced as seed did not increase with increased soil nutrient concentrations for several early successional species in an old-field community in the midwestern United States (Parrish & Bazzaz, 1982). Furthermore, empirical support was found for a model that predicted that species of open, disturbed environments, where most weedy species occur and similar to the typical habitats of S. obtusifolia, allocated a relatively constant proportion of biomass to reproduction among individuals (Hara et al., 1988). Since variation in reproductive allocation is less than that of plant biomass, reproductive output of S. obtusifolia, like that of many species, is largely a function of plant size (Samson & Werk, 1986; Aarssen & Taylor, 1992; Susko & Lovett-Doust, 2000a).

Effect of maternal nutrient regime on seed mass

Unlike other reproductive components, mass of individual S. obtusifolia seeds did not differ significantly with increased mineral nutrition. Studies with Senecio vulgaris (Fenner, 1986), Hydrophyllum appendiculatum (Wolfe, 1995), and Centaurea maculosa (Weiner et al., 1997) have also shown stability in seed mass in nutritional experiments, which is consistent with seed mass being the least plastic component of reproductive yield (Harper et al., 1970). Some studies have shown noticeable increases in seed mass with increased external nutrient supply, although in those cases the magnitude of increases in mass were small compared with changes in plant size, fruit production and seed production (Parrish & Bazzaz, 1985; Roach & Wulff, 1987; Aarssen & Burton, 1990; Sills & Nienhuis, 1995).

In S. obtusifolia, seed mass variation was greater than twofold among plants in each nutrient treatment (Fig. 3), a level of intraspecific variation commonly seen among and within plants (Mehlman, 1993; Obeso, 1993; Méndez, 1997; Susko & Lovett-Doust, 2000b). Weiner et al. (1997) stated that plants typically respond to changing environmental conditions by producing the same distribution of seed masses while varying the number of seeds. They hypothesised that plants vary seed mass distribution only when seed production is limited by developmental constraints (e.g. inherent architectural variation in the structure and function of flowers within inflorescences). In S. obtusifolia, we found that seed mass distributions were similar in shape but did change slightly with increased nutrient conditions. The distribution of seed masses became less negatively skewed and less leptokurtic. Since S. obtusifolia produces numerous indeterminate inflorescences until its resources are exhausted (thereby allowing for considerable plasticity in the number of seeds made) it would seem that developmental constraints would have little impact on seed mass distributions. It is more likely that mass distributions would be influenced by other maternal factors, such as the timing and location of flower and seed production. The present study did not control for genotypic differences among maternal plants. Hence, small differences in seed mass distribution among nutrient treatments also could be due to genetic variation.

Effects of maternal nutrient regime on germination

The results clearly showed that germination rate and final percentage germination of seeds of S. obtusifolia were largely unaffected by the maternal nutrient treatments (Fig. 2). Similarly, no significant differences in germination rates or percentages were found among seeds from maternal plants of Arabidopsis thaliana grown at different soil nutrient levels (Sills & Nienhuis, 1995). In that study, as in ours, mean seed mass did not differ significantly with respect to nutritional treatments. However, plants of Senecio vulgaris grown in high soil nutrients produced seeds that had greater individual mass and germinated earlier (Aarssen & Burton, 1990). In most species, germination is influenced by differences in seed mass (Dawson & Ehleringer, 1991; Baskin & Baskin, 1998). Likewise, once we separated seeds of S. obtusifolia into small and large size classes, we found significant effects. Small seed populations germinated earlier, more rapidly and to an equivalent or slightly lower final percentage than large seeds. Earlier germination of small seeds has also been found in Cakile edentula (Zhang, 1993), Erodium brachycarpum (Stamp, 1990), and Alliaria petiolata (Susko & Lovett-Doust, 2000b). Earlier germination of small seeds of E. brachycarpum was attributed to more rapid uptake of water due to their higher surface area to volume ratios (Stamp, 1990). Seeds of S. obtusifolia are known to possess physical dormancy (Baskin et al. 1998), and therefore require that seed coats be made permeable to water prior to germination. In addition to a higher surface area to volume ratio, small seeds may possess thinner seed coats than large seeds, thereby allowing quicker imbibition. Because of the more rapid germination of smaller seeds, the slight shift in mass distribution favouring small seeds in low nutrient conditions (Fig. 3) could have agroecological implications.

Seed composition

In general, the N concentration and total N content in seeds from maternal plants grown in high nutrients was somewhat greater than that of seeds from maternal plants grown in low nutrients (Fig. 5). Other studies have also shown that nutrient additions can elevate seed nutrient concentrations (Austin, 1966a; Austin, 1966b), and increased additions of nitrogenous fertilizers in environmentally controlled growth rooms (Saini et al., 1985) and in the field (Fawcett & Slife, 1978) increased the seed N content of Chenopodium album. In some cases increased nutrient content of seeds is positively correlated with percentage germination (Saini et al., 1985; Dawson & Ehleringer, 1991), while in other cases, as was found in S. obtusifolia, percentage germination is little affected (Vaughton & Ramsey, 1998).

The relationship between seed mass and seed nutrient content is less ambiguous. As seed mass increased in S. obtusifolia, N content in seeds increased significantly (Fig. 5, positive slopes). A similar relationship has been observed in other studies (Dawson & Ehleringer, 1991; Vaughton & Ramsey, 1998). Typically, seed mass and nutrient content are positively correlated with seedling size and survivorship. Hence both of the seed characteristics have the potential to influence seedling establishment, particularly when plants are grown in resource-poor environments or under competitive conditions (Parrish & Bazzaz, 1985; Stratton, 1989; Miao & Bazzaz, 1990; Wulff & Bazzaz, 1992). The seeds of S. obtusifolia in our experiments averaged between 3 and 4% N across nutrient treatments (data not shown), which is within the ranges of percentage N in seeds of species of Fabaceae (3.39–6.38%; Tanji & Elgharous, 1998) and other plant families reported elsewhere (Mattson, 1980; Henderson, 1990).

Maternal effects

One of the most interesting aspects of our study was how both offspring nutrient environment and seed size influenced the expression of maternal nutrient effects (Fig. 6). A number of studies have found a positive correlation between seed size and seedling growth and survivorship (Weis, 1982; Aarssen & Burton, 1990; Wulff & Bazzaz, 1992; Schmid & Dolt, 1994) and, indeed, we found that larger seeds yielded larger seedlings than did smaller seeds. By controlling for differences in seed size, significant effects of maternal nutrition on seedling growth became apparent. Seedlings from the high maternal nutrient regime grew larger and had greater leaf area than did seedlings from the low maternal nutrient regime. These results agree with the general notion that progeny growth, at least early in development, is correlated with maternal nutrient status (Harper et al., 1970; Stanton, 1984; Parrish & Bazzaz, 1985; Schmid & Dolt, 1994). The logical physiological linkage between maternal nutrition and the growth of offspring was the increase in seed N content, which may reflect a general increase in nutrient reserves. Seed nutrient availability, as well as associated factors (e.g. hormonal levels), presumably led to enhanced seedling growth over the 21-d experiment.

In comparison with the many studies of maternal environmental effects (Roach & Wulff, 1987; Wulff, 1995), relatively few studies have shown that the expression of maternal effects may vary according to the nutrient supply encountered by the progeny (Stratton, 1989; Wulff et al., 1994). Significant maternal nutrition effects were found with Pisum sativum, but only when seedlings were grown at a low nutrient supply (Austin, 1966b). In velvetleaf, A. theophrasti, the response to maternal nutrition did not differ with progeny nutrient status at early stages of growth, but did differ later in development (Wulff & Bazzaz, 1992). For S. obtusifolia, although the maternal effect was always present, the expression was magnified by larger seed size and increased seedling nutrient levels, both of which enhanced seedling growth.

Collectively, the results of this study indicate that seeds and seedlings of S. obtusifolia can have variable responses to heterogeneous edaphic nutrient availability in both parental and offspring generations, which suggests considerable phenotypic plasticity. Such plastic responses potentially have ecological significance, and could explain why this species is so prolific and widespread in the diverse array of agroecosystems with different nutrient inputs in the southeastern USA. We note, however, that although differences in early seedling development may be ephemeral when plant densities are low, the development of a hierarchy of plant sizes during the establishment phase is likely to influence survivorship and intra- and inter-specific interactions throughout the growing season when plant densities are high and competition is size-asymmetric. Hence, early differences in seedling size could persist, influencing adult fitness and persistence.

Acknowledgements

This work was supported by USDA-NRI grant #99-35315-7711. The authors thank Dr Jan Spears for the use of germination chambers.

Ancillary