Heat shock, mass-dependent germination, and seed yield as related components of fitness in Cistus ladanifer

Environmental and Experimental Botany 46 (2001) 11 – 20 www.elsevier.com/locate/envexpbot

Heat shock, mass-dependent germination, and seed yield as related components of fitness in Cistus ladanifer Juan A. Delgado *, Jose´ M. Serrano, Francisco Lo´pez, Francisco J. Acosta Departamento de Ecologı´a, Facultad de Biologı´a, Uni6ersidad Complutense de Madrid, 28040 Madrid, Spain Received 1 September 2000; received in revised form 2 January 2001; accepted 3 January 2001

Abstract The different weight-number strategies of seed production displayed by individuals of a Mediterranean fire-prone plant species (Cistus ladanifer) were investigated in relation to seed germination responses to pre-germination heating. A control (no heating), a high temperature during a short exposure time (100°C during 5 min) and a high temperature during a long exposure time (100°C during 15 min) were applied to seeds from different individual plants with different mean seed weight. These pre-germination treatments resemble natural germination scenarios for the studied species, absence of fire, typical Mediterranean shrub fire, and severe fire with high fuel load. Seed germination was related to heat treatments and seed mass. Seed heating increased the proportion of seeds germinating compared with the control treatment. Mean seed weight was positively correlated to the proportion of germinated seeds but only within heat treatments. These results suggest that in periods without fire, the relative contributions to the population dynamics are equal for all seeds, regardless of their mass, whereas heavier seeds would be the main contribution after wildfire events. Since lighter seeds can be produced in higher quantities than heavier ones within a given fruit, the number of seedlings produced per fruit depended strongly on the germination conditions. In the absence of wildfire, fruits producing lighter seeds gave rise to more seedlings; nevertheless, they were numerically exceeded by those producing heavy seeds after a wildfire. The implications of these results are discussed in relation to their consequences on the population dynamics of this species, considering also additional information on stand flammability and changes in seed mass with plant age. © 2001 Elsevier Science B.V. All rights reserved. Keywords: Evergreen Mediterranean shrub; Fire; Pre-germination heating; Reproductive allocation; Seed mass

1. Introduction Seed mass within a plant species (if measured as mean seed weight of a seed pool per individual * Corresponding author. Tel.: + 34-1-3945085; fax: + 34-13945081. E-mail address: [email protected] (J.A. Delgado).

plant) has been reported as a relatively constant feature, even when individuals differed largely in the amount of resources, as it could be inferred from their huge differences in total weight of seeds produced (Harper et al., 1970; Fenner, 1985). This result has been generally used to support the idea that seed mass has likely evolved in response to environmental pressures (Moore,

S0098-8472/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved. PII: S 0 0 9 8 - 8 4 7 2 ( 0 1 ) 0 0 0 7 6 - 4

12

J.A. Delgado et al. / En6ironmental and Experimental Botany 46 (2001) 11–20

1993). Differences in seed mass between species within a community or within a taxonomic group have been found to be related to seed predation (Janzen, 1969; Westoby et al., 1992), seed dispersal (Troumbis and Trabaud, 1986) and seedling establishment and survival (Armstrong and Westoby, 1993). This trade is, therefore, likely to have important ecological consequences, and several correlations have been found between mean seed mass and habitat. For instance, heavy seeds have been reported to be associated to rich (Maran˜ o´ n and Grubb, 1993) or shaded environments (Westoby et al., 1992; Leishman and Westoby, 1994) and light seeds to soils prone to waterlogging (Westoby et al., 1992). Nevertheless, these patterns cannot be easily interpreted as adaptations to different life histories because seed mass is correlated with other plant features (Primack, 1987; Lord et al., 1995; Van Hinsberg, 1988). Furthermore, seed size within a species (if measured as mean seed mass of a seed pool per individual plant) is a relatively constant feature, even when individuals differ largely in the number of seed produced. Despite the relative constancy of mean seed weight, variation in seed mass within species, populations, and even within individual plants is common (Michaels et al., 1988). Seed size differences within a given species are reflected in seed fitness through correlations between seed mass and predation (heavier seeds are more likely to be predated, Bradford and Smith, 1977), dispersion (heavier seeds disperse poorly, Howe and Westley, 1986; Troumbis and Trabaud, 1986; Willson, 1992), seed germination and seedling survival (heavier seeds produce larger, more competitive seedlings, Gross and Kromer, 1986; Andersson, 1990; Tripathi and Khan, 1990; Shipley and Parent, 1991; Leishman and Westoby, 1994). These results suggest that, at least within a species, seed mass is strongly related in a positive way to fitness and is likely a compromise between predispersal seed predation and dispersion on one hand and seedling establishment on the other. Parental plants must also face a trade-off between seed mass and seed number because although heavier seeds could produce fitter seedlings, few heavy seeds could be produced if

resources are limited. From a parent perspective, therefore, it is not the fitness of seeds, but rather the product of the number of seeds and their survival to independent seedlings, which must be maximised (Smith and Fretwell, 1974; Lloyd, 1987). This trade-off is obvious when there is a fixed amount of resources to partition in few heavy seeds or in many light seeds. This relation, however, can hardly be recorded in the real world, where different parental plants differ largely in the amount of resources available to them (Michaels et al., 1988; Silvertown, 1989; Mojonnier, 1998). Nevertheless, this trade-off is usually detected within a species when differences in resources are controlled in some way, e.g. by using fruits as the units of comparisons, since they are more constant in size than individual plants (Hainsworth et al., 1984; Mazer et al., 1986; A, gren, 1989; Mazer, 1989), or by using seed number per g of vegetative mass of a plant (Klinkhamer and Jong, 1992). Assuming this basic scenario wherein seed fitness increases with seed mass but reduces the number of seeds that can be produced, Smith and Fretwell (1974) proposed an optimisation model of allocation of resources where parents maximise their fitness by producing seeds of a homogeneous ‘optimal’ mass. This optimum is dependent on the shape of the curve that relates seed mass with seed fitness. Departures from this optimum within an individual or a population have been attributed to physiological, developmental or morphological constraints (e.g. McGinley et al., 1987), parentoffspring conflict and sibling rivalry (see e.g. Ganeshaiah and Uma Shaanker, 1988; Uma Shaanker et al., 1988), variation in mass or quality of parental resources (e.g. McGinley, 1989) or adaptation to variable environments (see e.g. Capinera, 1979; Silvertown, 1989). This latter adaptative explanation of seed mass variation has been mainly based upon the existence of environmental heterogeneity, each mass being fitted in their particular appropriate microhabitat. This is a plausible explanation for species with seed dimorphism, in which seeds with different mass could be selectively dispersed to their favourable environment, but not for species with continuous seed mass variation (McGinley et al., 1987; Silvertown, 1989). Nevertheless, some continuous variation

J.A. Delgado et al. / En6ironmental and Experimental Botany 46 (2001) 11–20

could be also adaptative if safe sites for seed germination or seedling establishment are limited and heavier seeds win the competition but have lower probabilities of reaching a safe site because of their small number (Geritz, 1995). The present work deals with the analysis of different size-number strategies exerted by individuals of a woody perennial plant species (Cistus ladanifer). Species of Cistus show a low germination rate if not exposed to fire (Roy and Sonie´ , 1992; Trabaud and Renard, 1999), but heating of seeds could result in reduced growth of seedlings and even in seed mortality (Valbuena et al., 1992a,b; Hanley and Fenner, 1998). Since a larger mass could be a relatively good protection from heating (lower surface/mass ratio, thicker coat), we expect that, after heating exposure, light seeds should have a low germination rate when compared with heavier ones. Although light seeds can likely be produced in large numbers, fruits producing few heavy seeds could finally produce more seedlings if a severe fire event takes place, because it could result in the death of most light seeds. The aim of the study is to assess whether there is a reproductive strategy exerted in fruits that are advantageous in all circumstances or, alternatively, different strategies perform distinctly in different environments. We have specifically assessed the germination of seeds that deferred in mass under different experimental conditions (with and without heating) that resemble natural, but contrasting germination scenarios for the species under study. In order to assess the strategy that could produce the largest number of seedlings, we have taken into account not only germination ability of seeds that differed in mass but also the potential number of seeds of each specific mean seed mass that can be produced within fruits.

13

(40°35%N; 3°43%W). All fieldwork was carried out in a natural patch of C. ladanifer L. containing some dispersed trees of Mediterranean holm oak (Quercus rotundifolia Lam.). C. ladanifer fruits are globular lignified capsules with 7– 10 locules. Each fruit produces many seeds (500–1000), which remain inside during several weeks, before the seeds are released. All species of the genus Cistus are fire-sensitive, having small hard seeds (these features can be extended to almost all the members of the Cistaceae; Thanos et al., 1992). Hardseededness is a specific type of seed dormancy characterised by a seed coat impermeable to water (Murdoch and Ellis, 1992). C. ladanifer patches are typically almost monospecific and re-establish themselves after fire events through seedling recruitment (Martı´n and Guinea, 1949; Valbuena et al., 1992a). Plants of C. ladanifer are easily flammable, due to the terpen-derived resin that impregnates leaves and branches. In a typical fire, most individuals are destroyed (Martı´n and Guinea, 1949). C. ladanifer follow an obligatory seeding strategy, therefore, depending on soil-stored seeds for post-fire seedling establishment. Soil seed banks of this kind of obligate seeders and nonsprouters, must be large enough to guarantee the regeneration of the population after fire. In order not to deplete the soil seed bank, most germination must be delayed until a fire occurs. Dormancy breaking promoted by fire in hard seeds has been suggested to take place through the rupture of the seed coat by means of cell expansion and contraction (Murdoch and Ellis, 1992). The survival rate of these seeds in the soil is low, compelling the plants to have a high production rate in order to compensate for losses in the soil seed bank (Troumbis and Trabaud, 1986, 1987).

2.2. Sampling, experiments and analysis 2. Materials and methods

2.1. Study area and species The study area was located near the village of Tres Cantos, 20 km North of Madrid, Spain

The sampling work was carried out in September 1997, around the time of the opening of fruits. There is no special dispersal trait and when the fruit is open, seeds just fall to the ground. Within the study area, 50 individual plants of C. ladanifer were randomly selected and removed.

14

J.A. Delgado et al. / En6ironmental and Experimental Botany 46 (2001) 11–20

Only a single non-predated fruit per individual plant was randomly selected and all the seeds were gathered in order to obtain a seed pool for each individual plant. Given the modular structure of plants, this type of sampling avoids pseudoreplication and ensures independence of observations in statistical comparisons (Hurlbert, 1984). In addition, this procedure reduces differences in germination ability within a group of seeds due to genetic variability, or maternal (nongenetic) environmental effects unrelated to seed mass (see e.g. Roach and Wulff, 1987; Van Hinsberg, 1988; Mazer, 1989). The total number of seeds and total seed weight of each fruit was recorded. Mean seed weight of each fruit was estimated as total seed weight of fruit divided by total seed number. From each of these seed pools, three subsamples of 25 seeds were randomly selected, weighed and later stored in paper bags in a dry place. Germination experiments were performed from August to September 1997. From earlier studies, we had a prior knowledge of the responses of C. ladanifer seeds to different heat treatments (Valbuena et al., 1992a,b). According to this, we selected the following pre-germination treatments, (1) control (no heating); (2) 100°C during 5 min; and (3) 100°C during 15 min. The first treatment (control) provides the germination rate of seeds in the absence of fire, which is important in order to establish the contribution of different seed mass to the dynamic of C. ladanifer populations in periods between fires. The second treatment, high temperature (100°C), short duration (5 min) is effective in stimulating a quick germination in response by the seeds. It was considered to be an optimum pre-germination treatment and was used to assess the contribution of different seed mass to the population dynamics of this species in a favourable fire event. The third treatment, high temperature (100°C), long duration (15 min) is slightly longer than the heat levels reached in a typical Mediterranean shrub fire (Trabaud, 1979); and had been reported to be less effective inducing germination of C. ladanifer seeds, probably because of embryo mortality by overheating. This was considered as a

severe heat treatment and it was thus used to assess heat resistance of different seed mass and their consequent contribution to population dynamics after an unfavourable fire event. Germination experiments were initiated the same day in which treatments were applied (1st August). Every treatment was randomly assigned to each subsample of 25 seeds in such a way that all treatments were applied to each individual plant (50 plants, three treatments per plant, and 25 seeds per treatment, a total of 3750 seeds). This experimental design follows a complete randomised block experiment, where individual plants act as blocks. In this type of analysis, the basic assumption is that there is no interaction between the effects of blocks and the effects of the treatment, i.e. besides random variation, differences between heat treatments are the same in all individual plants (Manly, 1992). Since our focus on this investigation is not on the effect of heat treatment per se but rather its interaction with seed mass, data on germination were compared with a two-factor Analysis of Covariance (ANCOVA) (Zar, 1984; Wilkinson, 1987), using individuals and heat treatment as factors, the proportion of germinated seeds as dependent variable and the mean seed weight of the group of seeds as covariate. Mean seed weight and proportion of germinated seeds were log- and arcsinetransformed, respectively, to achieve normality. Each treated subsample was set to germinate in plastic dishes on cotton disks saturated with distilled water during the whole experiment. The arrangement of dishes was determined randomly in order to avoid spurious results caused by position effects (Hurlbert, 1984). The germination experiments were conducted in a germination chamber at constant temperature, 20°C, with a 10-h day:14-h night period. White light was provided by four white fluorescent tubes (Sylvania, 36W, Germany) and far-red light by a redcoloured incandescent lamp (40W, Claude) with a sheet of blue Plexiglas filter (3 mm thick, Ro¨ hm GmbH, Darmstadt, Germany). Newly germinated seeds were counted daily. Germination was considered to occur when the radicle emerged from the testa. Seedlings were removed from the plastic dishes in order to avoid

J.A. Delgado et al. / En6ironmental and Experimental Botany 46 (2001) 11–20

interferences with future germination. The experiment was finished after 45 days, when no further germination was observed over a period of 2 consecutive days. In order to consider seed mass but along with seed number, we also assessed differences in the expected number of seedlings between fruits producing different seed mass. Expected number of seedlings within each treatment was estimated using the expected proportion of seeds that germinated, multiplied by the expected number of seeds produced. Least square regression analyses were performed using mean seed weight of each subsample as independent variable and the proportion of germinated seeds as dependent variable. Expected proportions of germination were estimated as the predicted values for the mean seed weight of each fruit, according to the significant linear equations obtained. In the same way, expected number of seeds produced per fruit was estimated as the fitted values for the mean seed weight of each fruit, according to the significant linear equation obtained from least square regression analysis of mean seed weight of each fruit (independent variable) and total number of seeds produced (dependent variable).

3. Results Pre-germination heat treatments (optimal and severe heating) increased the proportion of seeds germinating compared with the control treatment (78916%, 679 12% and 199 17%, respectively). Although differences are clear (P B 0.001), their significance cannot be ascertained directly from the results on Table 1 because of the significant Table 1 Results of two-way ANCOVA showing the effects of seed mass and heat treatment on germination of Cistus ladanifer seeds Variables

d.f.

F-Value

P-Value

Individuals Treatments Seed weight Treatment×seed weight

49 2 1 2

0.574 9.732 8.422 3.746

0.9832 0.0001 0.0046 0.0272

15

interaction between the effects of seed mass and the effects of treatments on seed germination (Zar, 1984; Wilkinson, 1987). Mean seed weight is positively correlated to the proportion of germinated seeds (Table 1), suggesting that, in broad terms, heavier seeds of C. ladanifer have higher germination probabilities than lighter ones. Nevertheless, the existence of a significant interaction between seed mass and treatments indicates that differences in germination found in seeds that differed in mass varies with treatment. The separated analyses of the effects of mean seed weight on germination within each treatment showed that seed mass was only positively correlated with probability of germination when seeds were heated (Fig. 1). This indicates that heavier seeds germinated better than lighter ones only under heat conditions. Comparisons of seed germination data between both heat treatments revealed no interaction between treatments and seed weight (Table 2). This result indicates that (i) differences between treatments were independent of seed mass, and (ii) the relative advantage of heavier seeds remained constant between treatments. Despite this finding, our results also showed that lighter seeds could be produced in higher quantities than heavier ones within a given fruit (Fig. 2). Contributions of different seed mass to population dynamics are therefore, likely to change considerably within each treatment, if the influence of seed mass on the number of seeds produced is taken into consideration. Since there was no significant relationship between germination and seed mass in the control treatment (Fig. 1), the proportion of mean germination was used as an estimate of expected germination probabilities. The results showed that expected contributions of different seed mass varied with treatment (Fig. 3). In the absence of pre-germination heating, lighter seeds account for the largest number of seedlings. When subjected to heat treatments, heavier seeds are clearly expected to produce more seedlings, although, there is some tendency for stabilisation of the relationship for heaviest seeds in the optimal heat treatment.

16

J.A. Delgado et al. / En6ironmental and Experimental Botany 46 (2001) 11–20 Table 2 Results of two-way ANCOVA showing the effects of seed mass and heat treatment on germination of Cistus ladanifer seedsa Variables

d.f.

F-Value

P-Value

Treatments×seed weight Individuals Treatments Seed weight

1 49 1 1

0.182 1.323 9.443 17.310

0.6719 0.1671 0.0035 0.0001

a

In this analysis, data on unheated controls were removed.

Hardseededness is a relatively common trait in Mediterranean-type vegetation. The presence of species with hard seeds in a community may result in little regeneration in unburned vegetation, following the decline in plant populations caused by the natural ageing and death of individual plants (Martı´n and Guinea, 1949; Wells, 1962; Probert, 1992). Nevertheless, a few studies have dealt directly with the regeneration of this type of communities in the absence of fire, suggesting different mechanisms for community regeneration (see e.g. Lloret and Zedler, 1991; Keeley, 1992; Midgley, 2000). Despite the ability of C. ladanifer seeds to germinate without fire, competition with adult plants makes seedling recruitment possible only with the existence of gaps, when senescent individuals die or after clearing or ploughing (Valbuena et al., 1992a). Since only a low percentage

Fig. 1. Relationship between mean seed weight (mg) and proportion of C. ladanifer seeds germinating after applying three different pre-germination treatments (a) control treatment (not significant relationship), (b) optimal heating (r 2 = 0.10; F = 5.33; P= 0.02) and (c) severe heating (r 2 =0.15; F = 8.72; P = 0.005). Variables were log or arcsine-transformed in order to normalise data.

4. Discussion Our results showed that C. ladanifer seeds could germinate in the absence of fire. In the field, this may occur through the influence of external agents responsible for breaking seed dormancy through alteration of the impermeable seed coat (Thanos et al., 1992; Valbuena et al., 1992a,b).

Fig. 2. Relationship between mean seed weight (mg) and number of seeds produced per fruit in the studied C. ladanifer population (r 2 =0.16; F=18.89; P B0.0001). Variables were log-transformed for normality.

J.A. Delgado et al. / En6ironmental and Experimental Botany 46 (2001) 11–20

Fig. 3. Relationship between mean seed weight (mg) and the number of expected germinating seeds in a fruit. Models combine both germination capability of seeds related to their weight and also the relationship between mean seed weight and number of seeds produced per fruit in the studied C. ladanifer population. Variables were log-transformed for normality.

of seeds could germinate, the amount of seeds produced in a population could be large enough to fill every gap produced by the death of individual plants. Furthermore, the demography of natural patches of Cistus species show a relatively wide range of ages with seed production only 2 or 3

17

years after fire (Roy and Sonie´ , 1992), suggesting that Cistus species have at least a small regeneration ability in the absence of fire. Although some seeds of C. ladanifer germinate without experiencing heat shock, most remain dormant and they are likely to remain viable for several years until wildfire takes place (Martı´n and Guinea, 1949; Troumbis and Trabaud, 1986; Thanos et al., 1992). Our study also shows that dormancy is broken by heat shock supporting evidence from studies on the population dynamics of several Cistus species where quick regeneration after fire has been reported (Arianoutsou and Margaris, 1981; Valbuena et al., 1992a,b). In the C. ladanifer population we studied, seeds showed the same probability of germination in periods without fire regardless of mass. Assuming there were no seed mass-related differences in seedling survival (which have not been assessed in this investigation), different seed mass showed the same relative contribution to total recruitment. This lack of correlation between seed mass and seed fitness (measured as probability of germination) could be a consequence of seed mass not being related to embryo vitality or to the amount of reserves. It is also possible that this feature is not important in such an early ontogenetic stage. As a result of this, and assuming no mass-dependent post-dispersal mortality of seeds, the distribution of seed mass in the soil seed bank would be similar to that produced in the population. Our experiment also revealed that prolonged heating, and also higher temperatures (Valbuena et al., 1992b) may reduce the total amount of germinating seeds. Since C. ladanifer plants are fire-sensitive, population regeneration after fire depends only on seeds in the soil (Martı´n and Guinea, 1949; Valbuena et al., 1992a,b). The rupture of dormancy reveals the advantage of heavier seeds over lighter ones. This advantage is likely to be enhanced by the severity of heating, which could result in the preferential survival of heavier seeds. Within the C. ladanifer population we studied, there was a trade-off between seed mass and seed number i.e. within a given fruit, mean seed mass can only be increased at the expense of the number of seeds. Since germination rate is not related

18

J.A. Delgado et al. / En6ironmental and Experimental Botany 46 (2001) 11–20

to seed mass, it could be suggested that most inter-fire seedling recruitment come from fruits with light seeds. Survival of these seedlings in a well-stabilised stand seems to be dependent on competition-free space, which could be made available through plant senescence and also by occasional disturbances other than fire. Such disturbance may promote the disappearance of mature individuals, without causing massive dormancy breakage amongst the soil-stored seed bank (e.g. cutting, ploughing or clearing activities, Valbuena et al., 1992a). C. ladanifer plants reach reproductive maturity relatively early and a fraction of the population can produce seeds just 2 years after germination (Montgomery and Strid, 1976, as cited in Roy and Sonie´ , 1992, and personal observations). Early maturation could be related to the extreme dependence of this species upon the soil seed bank, but also to early opportunities for recruitment. Recently-developed stands of C. ladanifer are relatively open, since massive post-fire seedling recruitment is followed by heavy seedling mortality, caused mainly by drought or freezing (Martı´n and Guinea, 1949; Arianoutsou and Margaris, 1981; Gallego Barrera et al., 1985a,b). Young individuals probably benefit from the production of fruits with many light seeds, because there is enough free space for seedling during the first years after fire. During periods between fires, light seeds are dispersed further away from the parent plant than heavy ones (see e.g. Troumbis and Trabaud, 1986), thus decreasing the probability of parent-offspring competition and increasing that of seeds to reach better conditions (i.e. competition free areas) for seedling establishment. This fruiting strategy also promotes the quick production of a bank of seeds in the soil. Flammability of recent C. ladanifer stands is high, due to the presence of terpenes in live plants (Martı´n and Guinea, 1949; Roy and Sonie´ , 1992). Although seeds are light, and therefore, vulnerable to severe heating, an unfavourable (i.e. long duration) wildfire event is unlikely, because of the small amount of burnable material accumulated on the soil (Davis and Burrows, 1989; Riggan et al., 1989). As the population ages, there is an increase in the accumulation of burnable material

in the soil (mainly dead branches) which is typical of Cistus species (see e.g. Martı´n and Guinea, 1949; Roy and Sonie´ , 1992). Therefore, wildfire in old stands burn slowly, increasing the exposure of seeds to heat. Heavy seeds are being dispersed closely to the parent plants, which have accumulated dead branches beneath them, the production of heavy seeds that resist prolonged heating, could be relevant for the dynamic of these stands. Our investigation has shown that although heavy seeds have usually a higher fitness with regard to heating processes than light ones, their contribution to the parental fitness must be filtered to their cost in terms of the reduced number of seeds that could be produced, suggesting a seed number compensation effect. The significance of such a seed number compensation could vary under different environmental conditions. The approach that we have followed in this study assumes that the results of these size-versusnumber strategies are the result of allocation and partitioning strategies exerted in fruits (reproductive modules). These results could be easily, further integrated through the hierarchical structure of plants and even extended to the population level.

Acknowledgements Thanks are debt to Angel Herna´ ndez and Alicia Delgado for their help during the sampling works. This study was sponsored by the Spanish Inter Ministerial Commission of Science and Technology (Research Project AMB-0777-C02-02).

References A, gren, J., 1989. Seed size and number in Rubus chamaemorus between-habitat variation, and effects of defoliation and supplemental pollination. J. Ecol. 77, 1080 – 1092. Andersson, S., 1990. Paternal effects on seed size in a population of Crepis tectorum (Asteraceae). Oikos 59, 3 – 8. Arianoutsou, M., Margaris, N.S., 1981. Early stages of regeneration after fire in a phryganic ecosystem (east Mediterranean). I. Regeneration by seed germination. Biol. Ecol. Me´ diterr. 8, 119 – 128.

J.A. Delgado et al. / En6ironmental and Experimental Botany 46 (2001) 11–20 Armstrong, D.P., Westoby, M., 1993. Seedlings from large seeds tolerate defoliation better: a test using phyllogenetically independent contrasts. Ecology 74, 1092 –1100. Bradford, D.F., Smith, C.C., 1977. Seed predation and seed number in Scheelea palm fruits. Ecology 58, 667 – 673. Capinera, J.L., 1979. Qualitative variation in plants and insects: effect of propagule size on ecological plasticity. Am. Nat. 114, 350 –361. Davis, F.W., Burrows, D.A., 1989. Spatial simulation of fire regime in Mediterranean-climate landscapes. In: Moreno, J.M., Oechel, W.C. (Eds.), The Role of Fire in Mediterranean-Type Ecosystems. Springer Verlag, New York, pp. 117 – 139. Fenner, M. (Ed.), 1985. Seed Ecology. Chapman & Hall, London. Gallego Barrera, J., Mata Moreno, C., Sa´ nchez Rodrı´guez, M., Zamora Lozano, M., Peinado Lucena, E., 1985a. Desarrollo de Cistus ladanifer tras su corte meca´ nico. Arch. Zootec. 34, 151 – 158. Gallego Barrera, J., Peinado Lucena, E., Zamora Lozano, M., 1985b. Productividad de la UTH en el desbroce de la jara negra (Cistus ladanifer). Arch. Zootec. 34, 27 –34. Ganeshaiah, K.N., Uma Shaanker, R., 1988. Seed abortion in wind-dispersed pods of Dalbergia sissoo: maternal regulation or sibling rivalry? Oecologia 77, 135 –139. Geritz, S.A.H., 1995. Evolutionary stable seed polymorphism and small-scale spatial variation in seedling density. Am. Nat. 146, 685 –707. Gross, K.L., Kromer, M.L., 1986. Seed weight effects on growth and reproduction in Oenothera biennis L. Bull. Torrey Bot. Club 113, 252 – 258. Hainsworth, F.R., Wolf, L.L., Mercier, T., 1984. Pollination and predispersal seed predation: net effects on reproduction and inflorescence characteristics in Ipomopsis aggregata. Oecologia 63, 405 – 409. Hanley, M.E., Fenner, M., 1998. Pre-germination temperature and the survivorship and onward growth of Mediterranean fire-following plant species. Acta Oecol. 19, 181 – 187. Harper, J.L., Lowell, P.H., Moore, K.G., 1970. The shapes and sizes of seeds. Annu. Rev. Ecol. Syst. 1, 327 – 356. Howe, H.F., Westley, L.C., 1986. Ecology of pollination and seed dispersal. In: Crawley, M.J. (Ed.), Plant Ecology. Blackwell Scientific Publications, Oxford, pp. 185 – 215. Hurlbert, S.H., 1984. Pseudoreplication and the design of ecological field experiments. Ecol. Monogr. 54, 187 – 211. Janzen, D.H., 1969. Seed eaters versus seed size, number, toxicity and dispersal. Evolution 23, 1 –27. Keeley, J.E., 1992. Demographic structure of California chaparral in the long-term absence of fire. J. Veg. Sci. 3, 79 – 90. Klinkhamer, P.G.L., Jong, T.J., 1992. Phenotypic gender in plants: effects of plant size and environment on allocation to seeds and flowers in Cynoglossum officinalis. Oikos 67, 81– 86. Leishman, M.R., Westoby, M., 1994. The role of seed size in seedling establishment in dry soil conditions-experimental evidence from semi-arid species. J. Ecol. 82, 249 – 258.

19

Lloret, F., Zedler, P.H., 1991. Recruitment pattern of Rhus integrifolia populations in periods between fire in chaparral. J. Veg. Sci. 2, 217 – 230. Lloyd, D.G., 1987. Selection of offspring at independence and other size-versus-number strategies. Am. Nat. 129, 800 – 817. Lord, J., Westoby, M., Leishman, M., 1995. Seed size and phylogeny in six temperate floras: constraints, niche conservatism, and adaptation. Am. Nat. 146, 349 – 364. Manly, B.F.J. (Ed.), 1992. The Design and Analysis of Research Studies. Cambridge University Press, Cambridge. Maran˜ o´ n, T., Grubb, P.J., 1993. Physiological basis and ecological significance of the seed size and relative growth rate relationship in Mediterranean annuals. Funct. Ecol. 7, 591 – 599. Martı´n, M., Guinea, E., 1949. Jarales y jaras (Cistografı´a hispa´ nica). Instituto Forestal de Investigaciones Experimentales, Madrid. Mazer, S.J., 1989. Family mean correlations among fitness components in wild radish: controlling for maternal effects on seed weight. Can. J. Bot. 67, 1890 – 1897. Mazer, S.J., Snow, A.A., Stanton, M.L., 1986. Fertilization dynamics and parental effects upon fruit development in Raphanus raphanistrum: consequences for seed size variation. Am. J. Bot. 73, 500 –511. McGinley, M.A., 1989. Within and among plant variation in seed mass and papus size in Tragopogon dubious. Can. J. Bot. 67, 1298 – 1304. McGinley, M.A., Temme, D.H., Geber, M.A., 1987. Parental investment in variable environments: theoretical and ampirical considerations. Am. Nat. 130, 370 – 398. Michaels, H.J., Benner, B., Hartgerink, A.P., Lee, T.D., Rice, S., Willson, M.F., Bertin, R.I., 1988. Seed size variation: magnitude, distribution, and ecological correlates. Evol. Ecol. 2, 157 – 166. Midgley, J., 2000. What are the relative costs, limits and correlates of increased degree of serotiny? Aust. Ecol. 25, 65 – 68. Mojonnier, L., 1998. Natural selection on two seed-size traits in the common morning glory Ipomoea purpurea (Convolvulaceae): patterns and evolutionary consequences. Am. Nat. 152, 188 – 203. Montgomery, K.R., Strid, T.W., 1976. Regeneration of introduced species of Cistus (Cistaceae) after fire in Southern California. Madron˜ o 23, 417 – 427. Moore, P.D., 1993. For richer, for poorer. Nature 366, 515. Murdoch, A.J., Ellis, R.H., 1992. Longevity, viability and dormancy. In: Fenner, M. (Ed.), Seeds: The Ecology of Regeneration in Plant Communities. CAB International, Wallingford, pp. 193 – 229. Primack, R.B., 1987. Relationships among flowers, fruits and seeds. Annu. Rev. Ecol. Syst. 18, 409 – 430. Probert, R.J., 1992. The role of temperature in germination ecophysiology. In: Fenner, M. (Ed.), Seeds: The Ecology of Regeneration in Plant Communities. CAB International, Wallingford, pp. 157 – 191.

20

J.A. Delgado et al. / En6ironmental and Experimental Botany 46 (2001) 11–20

Riggan, P.J., Franklin, S.E., Brass, J.A., Brooks, F.E., 1989. Perspectives on fire management in Mediterranean ecosystems of Southern California. In: Moreno, J.M., Oechel, W.C. (Eds.), The Role of Fire in Mediterranean-Type Ecosystems. Springer Verlag, New York, pp. 141 – 162. Roach, D.A., Wulff, R.D., 1987. Maternal effects in plants. Annu. Rev. Ecol. Syst. 18, 209 –235. Roy, J., Sonie´ , L., 1992. Germination and population dynamics of Cistus species in relation to fire. J. Appl. Ecol. 29, 647 – 655. Shipley, B., Parent, M., 1991. Germination responses of 64 wetland species in relation to seed size, minimum time to reproduction and seedling relative growth rate. Funct. Ecol. 5, 111 –118. Silvertown, J., 1989. The paradox of seed size and adaptation. Trends Ecol. Evol. 4, 24 –26. Smith, C.C., Fretwell, S.D., 1974. The optimal balance between size and number of offspring. Am. Nat. 108, 499 – 506. Thanos, C.A., Georghiou, K., Pantazi, C., 1992. Cistaceae: a plant family with hard seeds. Isr. J. Bot. 41, 251 – 263. Trabaud, L., 1979. Etude du comportement du feu dans la garrigue de Cheˆ ne kermes a` partir des tempe´ ratures et des vitesses de propagation. Ann. Sci. For. 36, 13 – 38. Trabaud, L., Renard, P., 1999. Do light and litter influence the recruitment of Cistus spp. stands? Isr. J. Ecol. 47, 1 – 9. Tripathi, R.S., Khan, M.L., 1990. Effects of seed weight and microsite characteristics on germination and seedling fitness in two species of Quercus in a subtropical wet hill forest. Oikos 57, 289 –296.

.

Troumbis, A., Trabaud, L., 1986. Comparison of reproductive biological attributes of two Cistus species. Acta Oecol. Oecol. Plant. 7, 235 – 250. Troumbis, A., Trabaud, L., 1987. Dynamique de la banque de graines de deux espe`ces de Cistes dans les maquis grecs. Acta Oecol. Oecol. Plant. 8, 167 – 179. Uma Shaanker, R., Ganeshaiah, K.N., Bawa, K.S., 1988. Parent-offspring conflict, sibling rivalry, and brood size patterns in plants. Annu. Rev. Ecol. Syst. 19, 177 – 205. Valbuena, L., Alonso, I., Ta´ rrega, R., Luis, E., 1992a. Influencia Del calor y del aclarado sobre la germinacio´ n de Cistus laurifolius y Cistus ladanifer. Pirineos 140, 109 – 118. Valbuena, L., Ta´ rrega, R., Luis, E., 1992b. Influence of heat on seed germination of Cistus laurifolius and Cistus ladanifer. Int. J. Wildland Fire 2, 15 – 20. Van Hinsberg, A., 1988. Maternal and ambient environmental effects of light on germination in Plantago lanceolata: correlated responses to selection on leaf length. Funct. Ecol. 12, 825 – 833. Wells, P.V., 1962. Vegetation in relation to geological substratum and fire in the San Luis Obispo quadrangle, California. Ecol. Monogr. 32, 79 – 103. Westoby, M., Jurado, E., Leishman, M., 1992. Comparative ecology of seed size. Trends Ecol. Evol. 11, 368 – 372. Wilkinson, L. (Ed.), 1987. SYSTAT: The System for Statistics. Evanston, Illinois. Willson, M.F., 1992. The ecology of seed dispersal. In: Fenner, M. (Ed.), Seeds: The Ecology of Regeneration in Plant Communities. CAB International, Wallingford, pp. 61 – 85. Zar, J.H. (Ed.), 1984. Biostatistical Analysis. Prentice-Hall, New Jersey.