*Present address: Prof. Dr Diethart Matthies, Department of Biology, Plant Ecology, University of Marburg, Karl-von-Frisch Str., D-35032 Marburg, Germany.
Reduced fecundity and offspring performance in small populations of the declining grassland plants Primula veris and Gentiana lutea
Article first published online: 24 DEC 2001
Journal of Ecology
Volume 88, Issue 1, pages 17–30, February 2000
How to Cite
Kéry, M., Matthies, D. and Spillmann, H.-H. (2000), Reduced fecundity and offspring performance in small populations of the declining grassland plants Primula veris and Gentiana lutea. Journal of Ecology, 88: 17–30. doi: 10.1046/j.1365-2745.2000.00422.x
- Issue published online: 24 DEC 2001
- Article first published online: 24 DEC 2001
- genetic drift;
- inbreeding depression;
- pollinator limitation
- Top of page
- Study species and area
1 We studied reproduction and offspring performance in relation to population size in the declining self-incompatible perennials Primula veris and Gentiana lutea. In both species, reproduction was strongly reduced in small populations, where plants produced fewer seeds per fruit and per plant. Total seed mass per plant was higher in large populations, but individual seeds were smaller, indicating a trade-off between seed number and size. Reproduction was depressed most strongly in populations consisting of less than c. 200 (P. veris) and c. 500 plants (G. lutea), respectively.
2 The inclusion of plant size (an integrated measure of habitat quality) in the statistical models did not change the relationships between fecundity and population size. Pollen limitation or inbreeding depression in small populations are therefore more likely explanations for these patterns than is habitat quality.
3 Germination rate and survival of seedlings in a common environment was not related to population size in either species, although P. veris developed into larger rosettes when seeds were derived from large populations. This suggests that inbreeding depression occurs in small populations of P. veris.
4 In a factorial fertilizer-by-competition experiment with P. veris, offspring from larger populations grew significantly larger and responded more strongly to fertilizer. For this declining species genetic deterioration as a result of habitat fragmentation may therefore aggravate the effects of environmental changes such as habitat eutrophication.
5 Our results suggest that small populations may face an increased short-term risk of extinction because of reduced reproduction, and an increased long-term risk because they are less able to respond to environmental changes.
- Top of page
- Study species and area
Land-use changes have led to the widespread destruction of the habitats of many plants. The remaining patches are often fragments in a matrix of unsuitable environments (Jennersten et al. 1992; Vitousek 1994), and many formerly common species have been reduced to small and isolated populations (Landolt 1991; Korneck et al. 1996).
Small size and isolation may have a number of negative effects on populations (Saunders et al. 1991). For example, small habitat patches may be particularly susceptible to deteriorating environmental conditions, and reduced seed set in small populations of both Senecio integrifolius (Widén 1993) and Gentiana pneumonanthe (Oostermeijer et al. 1994) has been attributed to such lower environmental quality. Small populations of plants are also more vulnerable to the effects of demographic and environmental stochasticity (Menges 1991a, 1992; Menges & Dolan 1998).
Interactions with mutualists, for example seed dispersers and pollen vectors, may be interrupted (Olesen & Jain 1994; Matthies et al. 1995) and, in particular, small patches of flowering plants may be less attractive to pollinators and receive less pollen (Sih & Baltus 1987). Insufficient pollen quantity or quality has been suggested as the cause of reduced fecundity in small populations of several species (Menges 1991b; Petanidou et al. 1991; Lamont et al. 1993; Aizen & Feinsinger 1994; Fischer & Matthies 1997) and has been demonstrated experimentally for others (Byers 1995; Ågren 1996; Groom 1998). Self-incompatible plants are more likely to be affected by pollen limitation in small populations than those that are self-compatible (Byers 1995).
Sometimes local population density rather than overall population size is important for pollinator behaviour and successful reproduction (van Treuren et al. 1994; Roll et al. 1997). Thus, in experimental stands of Brassica kaber, density but not the size of populations had a strong effect on both the visitation rate of pollinating insects and on seed set (Kunin 1997). However, there are few studies of the effect of density on plant fecundity in the field.
Small populations may be subject to increased inbreeding and to the loss of genetic variation due to genetic drift (Lacy 1987; Barrett & Kohn 1991; Ellstrand & Elam 1993), both of which are expected to lead to a reduction in fitness. Such genetic deterioration has been found in several species (Ouborg & van Treuren 1994; Raijmann et al. 1994; Fischer & Matthies 1998a). Self-incompatible plants may be particularly vulnerable to genetic deterioration, and the negative effects of the loss of mating-type alleles (a particular form of genetic drift) have been demonstrated in small populations of the self-incompatible Hymenoxys acaulis (DeMauro 1993) and Eupatorium resinosum (Byers 1995). Furthermore, out-crossing species may be more vulnerable to increased inbreeding (Ellstrand & Elam 1993). Although in the long term loss of genetic variability may decrease the potential of a population to persist in the face of environmental change (Huenneke 1991), few studies have shown either that it is triggered by small population size or that it leads to reduced plant fitness (Oostermeijer et al. 1994; Fischer & Matthies 1998a,b).
Fecundity in small populations of short-lived species may be reduced by pollen limitation or genetic deterioration, and this may cause a rapid further decline in population size and an increased risk of extinction (Fischer & Matthies 1998b). In long-lived species, however, the negative consequences of small population size and isolation may not become obvious for a long time, because the mortality of established perennial plants is often very low (Summerfield 1972; Tamm 1972; Harper 1977) even though reproduction may be affected much sooner than survival (Oostermeijer et al. 1992; Lamont et al. 1993).
We studied reproduction in relation to population size and population density in two declining perennials, Primula veris L. and Gentiana lutea L., in fragments of nutrient-poor grassland, a type of habitat that has declined strongly in many parts of Europe (Zoller & Wagner 1986; Keymer & Leach 1987). Both species are self-incompatible and P. veris is distylous, making them particularly sensitive to pollen limitation and genetic deterioration in small populations (Husband & Schemske 1996). To analyse possible genetic effects of small population size on offspring fitness, we studied the performance of offspring over 2 years in a common garden.
In addition, we carried out a factorial fertilizer-by-competition experiment with P. veris to investigate the potential of populations of different size for coping with changes in environmental conditions.
Study species and area
- Top of page
- Study species and area
Primula veris is a small perennial rosette plant typical of nutrient-poor grasslands and the edges of woodlands, from Spain to eastern Asia (Hegi 1927). In spring it produces umbels with up to 20 yellow flowers. Like other species of Primula (Clapham et al. 1987), P. veris is distylous and strictly self-incompatible: only cross-pollination between the two genetically determined pin and thrum flower morphs results in seed set (Grime et al. 1988). Gentiana lutea is a large perennial that grows in nutrient-poor grasslands from the Pyrenees to Asia minor (Hegi 1927). Its typical habitat in central Europe is pasture grazed by cattle. In summer the plant produces one tall inflorescence (up to 200 cm) that carries four to 10 pairs of pseudo-umbels, each consisting of c. 20 yellow flowers, spaced at 5–10 cm intervals. Gentiana lutea is self-incompatible and thus depends on pollination by insects to produce any seeds (Hegi 1927; Bucher 1987). Both G. lutea and P. veris are pollinated mainly by Hymenoptera and Diptera (Hegi 1927; M. Kéry, personal observation). Neither species has special features to facilitate dispersal.
We studied P. veris in the French Sundgau foothills of the Jura mountains west of the city of Basel, Switzerland (47°30′ N, 7°35′ E), and G. lutea in grasslands in the Swiss Jura mountains south of Basel. Most of the studied populations of P. veris were situated on south-facing slopes between 340 and 440 m a.s.l., whereas most of the populations of G. lutea were situated between 700 and 900 m a.s.l. on north-facing slopes. The study populations were selected randomly and were therefore representative of the range of conditions and population sizes found in the study area. The habitats of both species have declined strongly and become fragmented since the 1950s (Zoller & Wagner 1986), and it is therefore likely that the extant populations are remnants of formerly much larger and less isolated populations. Dispersal in both species is very restricted, and colonization of new habitats a rare event. Both species are long-lived (Hegi 1927; Tamm 1972) and the number of generations since fragmentation must therefore have been small.
- Top of page
- Study species and area
Relationship between population size and fecundity in natural populations
In both P. veris and G. lutea, groups of plants separated by 50 m or more from the next conspecific were taken as a population. Most populations, however, were separated by at least several hundred metres.
In April 1994, as part of a study on heterostyly in populations of P. veris (Kéry 1995), we counted the number of flowering rosettes in 76 populations. An initial subsample of 27 populations (nine to 13 060 flowering rosettes) was chosen to investigate the relationship between fecundity and population size. In each population, 20 flowering rosettes were randomly selected and individually marked unless population size was less than 20, in which case all plants were marked (425 plants in total). For each plant the number of flowers was recorded, and the length of the longest leaf measured to the nearest millimetre.
To characterize the environmental conditions at the P. veris sites, vegetation relevees were made in May at all sites except one that had been subjected to grazing by cattle. Five 2 × 2 m plots were established at random, and identity and cover of all plant species was determined. These data were used to establish the main environmental gradients, using correspondence analysis (version 4.0 of program canoco; ter Braak & Smilauer 1998).
When the P. veris populations were visited again at fruiting time (June), only 80 of the 425 marked plants in 12 of the initial 27 populations could be located. The rest had been lost to grazing cattle. We therefore selected an average of 14 fruiting plants from each of an additional eight populations of known size (114 additional plants). The total size of the sample at the time of fruiting was thus 194 plants in 20 populations.
For all plants we counted the number of flowers and fruits and measured the length of the longest flower stalk, and collected, counted and weighed all their seeds. The number of seeds per fruit was calculated as the total number of seeds per plant divided by the number of fruits per plant. Mean individual seed mass (‘seed mass’) was calculated as total seed mass per plant divided by the total number of seeds. In addition, all above-ground plant parts were harvested in 16 of the 20 study populations. For conservation reasons no biomass samples were taken in the four smallest populations. The material was dried for 2 days at 70 °C, divided into reproductive and vegetative structures, and weighed.
In June 1994 we counted the number of flowering plants in 27 populations of G. lutea (one to 3420 flowering individuals). At the time of fruit ripening in August, we randomly selected an average of 12 plants per population (335 plants in total). We measured the height of each plant, counted the number of fruits and collected three fruits per plant at random. Developed and aborted seeds were distinguished in the laboratory by their size, shape and colour. ‘Seeds per fruit’ refers to the number of developed seeds per fruit, and ‘seeds initiated per fruit’ to the total number of developed and aborted seeds per fruit.
All developed seeds from each of the three sampled fruits per plant were weighed and used with the total number of fruits to calculate the total seed mass per plant.
Mean individual seed mass was calculated in the same way as in P. veris. Because of heavy grazing, no vegetation data could be recorded at the G. lutea sites.
For both species, we marked the location of the marginal plants in each population on a map. The area that each population occupied was then determined by cutting out the minimum-convex polygon area and measuring it with a photoplanimeter. To calculate population density, we divided population size by area.
Effects of population size on germination of seeds, and on size, survival and growth of seedlings
The performance of offspring of P. veris and G. lutea was studied in 1995 using the seeds collected in the field in 1994. We assessed germination rates in seed samples taken in May from up to 10 randomly selected mother plants in each of the 17 P. veris populations that had produced any seeds (117 mothers) and in June from each of the 27 populations of G. lutea studied (211 mothers).
The seed samples were weighed and placed on filter paper in Petri dishes, to give 343 and 416 dishes, for P. veris and G. lutea, respectively. To break dormancy, 5 ml of a solution of gibberellic acid (1 mg GA3 ml−1 of water) was added. The seeds were kept in the dark at 17–20 °C. Germination was assessed after 90 (P. veris) and 137 days (G. lutea). Germination rates were affected by fungal infection. We therefore recorded presence or absence of fungi for each dish.
In July 1995 we planted, on average, 25 seedlings per seed family of P. veris into pots (10 cm diameter) containing commercial seedling compost, at a maximum density of five plants per pot. We estimated the size of each seedling (using four size classes) and placed the pots into flowering beds in the experimental garden. A total of 2800 P. veris seedlings originating from 113 seed families from 17 populations was planted into 845 pots. The plants were watered if necessary. In March 1996 we recorded the number and size (rosette diameter and width of the largest leaf) of the surviving plants.
In November 1995 we measured (using four size classes), on average, 18 seedlings per seed family of G. lutea and planted them into pots at a maximum density of four plants per pot. The pots were kept well-watered in an unheated but frost-free glasshouse (minimum temperature 7 °C). In total, 2949 G. lutea seedlings from 166 seed families from 27 populations were planted into 1184 pots. In June 1996 we recorded the number and size (length of the longest leaf to the nearest millimetre) of the surviving plants. In both species, size data for the individual seedlings within a pot were averaged for the analyses.
Effects of population size on plasticity of offspring
To study the effects of population size on the subsequent performance of the offspring, we performed a factorial 2 × 2 fertilizer-by-competition experiment using all the P. veris plants that survived to the end of the seedling growth experiment. This experiment could not be conducted with G. lutea because too few plants remained.
In March 1996, we planted 340 1-year-old P. veris plants (descended from 79 mothers from 15 populations) singly into larger pots and measured the width of the largest leaf and the rosette diameter of all plants. The available plants from each population were assigned randomly to the four treatment combinations in a balanced way. About 100 seeds of the grass Lolium perenne L. were sown into each pot of the competition treatment. Plants in the fertilizer treatment received 1 g of slow release fertilizer (Osmocote) at the start of the experiment and 50 ml of a 0.12% solution of a commercial fertilizer (HF2 Typ K, Hauert Wädenswil, Switzerland) in April, May and June 1996. (Osmocote and HF2 Typ K both distributed by HBG Düngervertrieb-AG, Grossaffoltern, Switzerland.) Plants in the control group received equivalent amounts of water. In July 1996 we recorded whether the plants flowered. The above-ground parts of the plants were harvested, dried at 70 °C for a week, and weighed.
Integrated fitness functions
Plant fitness depends both on the number of seeds produced and on offspring performance. To analyse the effects of population size on plant fitness we calculated three multiplicative fitness functions. The mean number of seeds germinated per mother plant was calculated as the mean number of seeds produced per plant in a population × the mean germination rate of the seeds originating from that population. The mean number of surviving offspring per mother plant was calculated as the mean number of seeds germinated × mean survival until September 1995 (P. veris) or February 1996 (G. lutea). Cumulative offspring size was calculated as the mean number of surviving offspring per mother plant × mean offspring size (leaf length in G. lutea, rosette diameter in P. veris).
The relationship between reproduction in natural populations and population size was analysed by nested anova models (Steel & Torrie 1980). The effects of population density and size were tested against the variation among populations, and the effect of plant size was tested against the residual variation among plants within populations.
To achieve homoscedasticity and normally distributed residuals in P. veris, population size and number of flowers were log-transformed, and fruits per plant, seeds per fruit, seeds per plant, seed mass and total seed mass per plant were square-root transformed prior to analysis. In G. lutea, population size, fruits per plant and seeds initiated per fruit were log-transformed, and seeds per plant and total seed mass per plant square-root transformed prior to analysis.
Effects on germination rate were analysed by nested anova. Because setting up the germination experiment in P. veris and G. lutea took 2 and 7 days, respectively, the starting day was included as a blocking factor in the analysis of the germination experiment. Infection with fungi was included as a covariate. Seed mass was much smaller in large populations (see the first part of the Results section). Therefore, we also fitted seed mass as a covariate to separate the effects of population size from maternal effects through seed mass. The effect of population size was tested against variation among populations and the variation among populations was tested against variation among seed families within populations. All other effects were tested against the residual.
Effects on growth and survival of seedlings were analysed by nested anova. Initial size and planting density of seedlings in the pots were used as covariates to eliminate any size and crowding effects on survival and further growth. In addition, time of planting was used as a blocking variable, because planting was spaced across several days. The effect of population size was tested against variation among populations and the variation among populations was tested against variation among seed families within populations. The variation among seed families was tested against variation among Petri dishes within seed families. All other effects were tested against the residual. Data on seedling survival were angular-transformed prior to analysis.
Above-ground biomass in the plasticity experiment was analysed with a mixed-model anova with population and seed family within population as random effects, and fertilizer and competition as fixed effects. Effects were tested according to the standard rules for the analysis of mixed models (Steel & Torrie 1980; see skeleton analysis in Table 3).
|Source||(a) Above-groun namest="col3" nameend="col5">d biomass MS||(b) Flowering probability Variance ratio||d.f.||MS||F||MD||Approx. F|
|Seed family within population||MSSF||MSSF/MSR||64||1.16||3.26***||1.54||1.45*|
|Fertilizer × competition||MSFC||MSFC/MSPFC||1||13.33||50.28***||1.24||0.93|
|Population size × fertilizer||MSPSF||MSPSF/MSPF||1||2.30||6.07*||2.72||2.61|
|Population size × competition||MSPSC||MSPSC/MSPC||1||0.03||0.05||0.14||0.10|
|Population size × fertilizer × competition||MSPSFC||MSPSFC/MSPFC||1||0.32||1.19||5.70||4.29(*)|
|Population × fertilizer||MSPF||MSPF/MSSFF||10||0.38||0.85||1.04||1.45|
|Population × competition||MSPC||MSPC/MSSFC||11||0.51||1.00||1.35||0.71|
|Population × fertilizer × competition||MSPFC||MSPFC/MSSFFC||10||0.27||0.43||1.33||12.17|
|Seed family within population × fertilizer||MSSFF||MSSFF/MSR||36||0.45||1.25||0.72||0.68|
|Seed family within population × competition||MSSFC||MSSFC/MSR||29||0.51||1.43(*)||1.91||1.80*|
|Seed family within pop. × fertilizer × competition||MSSFFC||MSSFFC/MSR||9||0.62||1.72(*)||0.11||0.10|
Effects on first-year flowering probability were studied by analysis of deviance using a logistic regression model (McCullagh & Nelder 1989). Mean deviances due to a factor were divided by the appropriate error mean deviances, analogous to the calculation of F-values in the analysis of biomass. These quasi-F-values follow an approximate F-distribution (Francis et al. 1993). All models were implemented in Genstat 5.3. (Payne 1993) using the procedures for general and generalized linear models.
- Top of page
- Study species and area
Relationship between population size and fecundity in natural populations
In P. veris (Fig. 1 and Table 1a) the number of flowers per plant was not influenced by population size. In contrast, a much higher proportion of flowers developed into fruits in large than in small populations, and consequently the number of fruits per plant was higher. Moreover, the number of seeds per fruit was also higher in large populations. As a result, the number of seeds per plant strongly increased with population size. While plants in large populations produced several hundred seeds each, plants in some of the very small populations produced no seeds at all. On average, a 10-fold increase in population size was associated with an additional 80 seeds per plant. In contrast, mean individual seed mass decreased with increasing population size. However, the reduced size of the seeds was more than compensated for by their higher number. Total seed mass per plant therefore increased strongly with population size.
|Flowers per plant||Fruits per flower||Fruits per plant||Seeds per fruit||Seeds per plant||Mean seed mass||Total seed mass per plant|
|(a) Primula veris|
|Fruits per plant||Seeds initiated per fruit||Proportion of seeds developed||Seeds per fruit||Seeds per plant||Mean seed mass||Total seed mass per plant|
|(b) Gentiana lutea|
In G. lutea (Fig. 2 and Table 1b) the number of fruits per plant and the number of seeds initiated per fruit did not vary systematically with population size. However, in small populations more seeds were aborted and thus the proportion of seeds developed per fruit was lower. As a consequence the number of seeds per fruit and per plant was much higher in large than in small populations. On average, plants in the largest populations produced c. 8000 seeds, and those in the smallest only c. 4000. A 10-fold increase in population size was, on average, associated with an additional 1400 seeds per plant. Seed size decreased with increasing population size. Because the higher number of seeds per plant in large populations more than compensated for the reduction in seed size, total seed mass in G. lutea increased with population size.
We thus found very similar patterns in both species. Plants in small populations produced fewer seeds per fruit and per plant, but individual seeds were larger. Plots of the number of seeds per fruit against population size on a linear scale (Fig. 3) indicated an increase up to a population size of c. 200 plants in P. veris and c. 500 flowering plants in G. lutea. Further increases in population size had little effect on the number of seeds per fruit.
The increase of reproduction with population size in both species was not the result of higher habitat quality at sites with large populations. Plant size, a measure of local habitat quality, did not increase with population size in either species. In P. veris, plant size, as estimated by leaf length, was actually smaller in large than in small populations (P < 0.05). However, in P. veris, population size was significantly correlated with the first axis of a correspondence analysis of the vegetation data (rs = −0.48, d.f. = 24, P = 0.01, Spearman rank correlation).
Plant size strongly affected most components of reproduction in both species (Table 1), but the relationships between population size and reproduction were significant independent of the effects of plant size.
Another factor that may influence fecundity in natural populations is population density, which in P. veris was negatively correlated with population size (P < 0.001) but was not correlated with it in G. lutea. Population density in P. veris had a positive effect on the number of fruits per plant and a negative effect on the number of seeds per fruit and per plant (Table 1a). In contrast, no effects of population density were found in G. lutea (Table 1b). The inclusion of population density as a covariate in the models did not change the relationships between population size and reproduction (Table 1). The effect of population size on reproduction was therefore also independent of the effect of population density.
Effects of population size on germination rate, and on size, survival and growth of seedlings
In P. veris 93.5%, and in G. lutea 57.5%, of all seeds germinated. In both species there was significant variation in germination both among populations and among seed families within populations. However, there was no effect of population size on germination rate in either species (Table 2), either with or without adjustment of the germination rate for initial differences in seed mass. Seed mass had a significant effect on germination in both species, but in opposite directions, with more large seeds germinating in P. veris and more small seeds in G. lutea.
|Primula veris||Gentiana lutea|
|Population size||1||0.001||< 0.01||1||0.136||0.49|
|Seed family within population||99||0.077||4.74***||183||0.091||6.24***|
Seedlings from large populations were smaller in P. veris (F1,15 = 3.98, P = 0.06). However, contrary to expectation, this effect was not mediated by seed size, which was smaller in large populations. Maternal population size had an effect on seedling size in P. veris, even if seedling size was adjusted for seed mass in the model (F1,15 = 3.36, P = 0.09). In contrast, seedling size in G. lutea was unrelated to maternal population size (F1,25 = 0.38, P > 0.54).
In P. veris, offspring survival until March 1996 was not affected by maternal population size (F1,15 = 0.31, P = 0.59) when adjusted for variation in initial seedling size. However, in March 1996 rosettes of offspring from large populations were larger than those from small populations (F1,13 = 7.52, P = 0.02). This effect was even more pronounced if initial seedling size was not adjusted for (F1,15 = 9.44, P < 0.01). In contrast, in G. lutea maternal population size had no effect on offspring survival until June 1996 (F1,25 = 0.31, P > 0.58), or on offspring size (F1,22 = 0.68, P > 0.42).
Integrated fitness functions
In both P. veris and G. lutea, plants from large populations had a higher fitness than plants from small populations, as estimated by three different multiplicative fitness functions. In P. veris, there were significant positive correlations between population size and the number of seeds germinated (r = 0.79, d.f. = 18, P < 0.001), the number of seedlings surviving until September 1995 (r = 0.57, P < 0.01), and cumulative offspring size per parent plant (Fig. 4; r = 0.57, P < 0.01).
Similar relationships were found in G. lutea, although some of them were only marginally significant. The number of seeds germinated (r = 0.38, d.f. = 25, P < 0.05), the number of seedlings surviving until February 1996 (r = 0.32, P < 0.11), and cumulative offspring size (Fig. 4; r = 0.35, P < 0.08) per parent plant increased with population size.
Effects of population size on phenotypic plasticity of offspring
As expected, the addition of fertilizer increased and competition with the grass L. perenne reduced the growth of P. veris (Table 3a). There was also a significant fertilizer-by-competition interaction. The growth of P. veris plants was reduced in particular when they were both fertilized, and raised when in competition with the grass. The size of the maternal population had a positive overall influence on plant size. In addition, plants from large populations responded more strongly to the fertilizer treatment than plants from small populations (Fig. 5 and Table 3a; significant population size-by-fertilizer interaction). In contrast, there were no differences in the reaction of plants from large and small populations to competition.
First-year flowering probability was higher for fertilized plants and lower for plants growing in competition with L. perenne (Table 3b). The size of the maternal population had no overall effect on the flowering probability of the offspring. However, there was a marginally significant three-way interaction between population size, fertilizer and competition (Table 3b). The response of offspring from large maternal populations to high resource availability (fertilizer addition, no competition) tended to be stronger.
We also found evidence for genetic variation among seed families in both studied traits. Seed families differed in above-ground biomass and in their probability of flowering.
Moreover, among seed families, there was marginally significant variation in the growth response to fertilizer and competition, and significant variation in the response to competition of the probability of flowering in the first year.
- Top of page
- Study species and area
Reproduction and offspring performance in relation to population size
We found a strong reduction in the number of seeds produced in small populations of the declining grassland plants P. veris and G. lutea, and in P. veris this was accompanied by reduced offspring performance. These patterns could be due to lower habitat quality, to pollen limitation, or to increased inbreeding and loss of genetic variation in small populations (Barrett & Kohn 1991; Widén 1993; Ågren 1996; Fischer & Matthies 1998a,b).
Several lines of evidence suggest that the lower fitness of plants in small populations of P. veris and G. lutea was not a consequence of lower habitat quality in these populations. In both species, differences among populations in plant size, an integrated measure of habitat quality, did not account for the relationships between fecundity and population size. This indicates that the effects of population size and of habitat quality on reproduction were independent. Moreover, population density did not increase with population size in either species, and the inclusion of density in the statistical models did not affect the relationships between population size and reproduction.
Obligate out-crossing species like P. veris and G. lutea (Bucher 1987; Grime et al. 1988) may be particularly sensitive to the disruption of pollinator–plant mutualisms in small populations. Bumble-bees, which are important pollinators for both species, have been shown to pollinate a higher proportion of flowers in large than in small populations (Sih & Baltus 1987). Insufficient quantity and quality of pollen is therefore a likely explanation for the reduced fecundity in small populations of both study species (Jennersten & Nilsson 1993; Byers 1995; Ågren 1996). Severe pollinator limitation is known from other species of Primula (Washitani et al. 1994), and in P. veris, which is distylous and intramorph incompatible (Clapham et al. 1987), the probability of plants receiving suitable pollen is highest if the two different morphs are present in equal numbers. Deviations of the morph ratio from 1 : 1 have been found to be much larger in small than in large populations (Kéry 1995), and reproduction in populations with large deviations was indeed significantly reduced, suggesting that the imbalance may have contributed to a lack of compatible pollen and thus reduced reproduction in small populations of P. veris.
Genetic factors may also have contributed to reduced reproduction in small populations of P. veris and G. lutea. Although reduced genetic variation due to genetic drift and inbreeding depression has been found to reduce fecundity in small populations of other species (Oostermeijer et al. 1994; Fischer & Matthies 1998a,b), both study species are long-lived perennials, and as the studied populations have probably been isolated for only a few generations (Spillmann 1998) it is unlikely that there has been strong genetic drift. However, the higher seed abortion (G. lutea) and lower number of seeds per fruit (P. veris) in small populations could be the result of increased mortality of developing seeds due to inbreeding, whose deleterious effects are often expressed during seed development (Charlesworth & Charlesworth 1987; Barrett & Kohn 1991; Husband & Schemske 1996). Although the observed larger size of seeds in small populations of both species might be considered to be inconsistent with the hypothesis of inbreeding depression, there is often a trade-off between number and size of offspring. Seeds compete for maternal resources, and a reduction in seed number may thus result in greater resource availability per seed and larger seed size (Lee 1988; Matthies 1990; Venable 1992; Oostermeijer et al. 1995). In conclusion, either or both reduced pollination and genetic deterioration could explain the lower reproduction in small populations of the two study species.
The common garden experiment provides additional evidence for genetic effects of small population size in P. veris. The growth of offspring from small populations was lower than that of plants from large populations. Similarly, negative effects of small population size on offspring performance have been found in a few other studies (Menges 1991b; Oostermeijer et al. 1994), although these effects are not necessarily genetic. For instance, in the rare perennial Gentiana pneumonanthe, several of the reductions in measures of plant performance in plants from small populations grown in a common garden were due to maternal carry-over effects (Oostermeijer et al. 1994). However, because of the longer duration of our study and the fact that seeds from large populations were actually smaller, environmental maternal effects are unlikely to be responsible for the observed differences in offspring performance in P. veris. In the rare short-lived Gentianella germanica, offspring performance after 2 years in a common garden was similarly positively correlated with population size (Fischer & Matthies 1998b).
A particularly interesting result of this study is the lower plasticity shown by small populations of P. veris. Plants from small populations were less able to increase growth and flower early in response to increased nutrient availability. It has been suggested that small population size may reduce the ability of a population to adapt to changes in environmental conditions because its genetic variation is lower (Huenneke 1991). Our results suggest that fragmentation may also reduce the ability of a population to react plastically to environmental changes. Genetic variability and phenotypic plasticity are different means of coping with environmental change and heterogeneity, but as phenotypic plasticity itself has a genetic basis and may vary among individuals and populations, there is not necessarily a trade-off between plasticity and genetic variation (Schlichting 1986; Schmid 1992). The reduced plasticity of offspring in small populations of P. veris, like their reduced performance, is therefore probably due to increased inbreeding, although effects of drift cannot be excluded. A reduced ability to tolerate changes in environmental conditions and simulated herbivory in plants from small populations has been claimed for Ipomopsis aggregata (Heschel & Paige 1995), although the validity of these results is questionable for statistical reasons (Ouborg & van Groenendael 1996).
Consequences of reduced reproduction and offspring performance in small populations
The reductions in both reproduction and performance of offspring from plants in small populations are likely to have both short-term and longer term negative consequences. Reduced sexual reproduction may have negative consequences even for the short-term population dynamics of long-lived species. In a stable Swedish population of P. veris with an annual population growth rate estimated at 1.03, 53% of all newly recruited plants recorded during 25 years of study were derived from seedlings (data from Tamm 1972). In an artificially founded expanding population close to our study area, recruitment by seeds was even more important. Here the annual population growth rate during the first 5 years was estimated at 1.60, with 61% of recruits being due to sexual reproduction rather than clonal propagation (M. Schaub, personal communication). In the medium term, reduced reproduction may reduce the chance to colonize unoccupied habitat patches. This is especially serious in fragmented landscapes. A reduction in the number of seeds produced may therefore increase the extinction risk of metapopulations (Hanski & Gilpin 1991).
A further consequence of reduced sexual reproduction is that fewer genetic variants are available in each generation for natural selection to act upon. This could compromise the ability of a small population to respond to changed environmental conditions in the long term and therefore increase its extinction risk. Moreover, the reduced plasticity that we observed in P. veris may strongly affect fitness under changed environmental conditions. Survival and reproduction of plants is usually positively correlated with size (Solbrig 1981; Antonovics & Primack 1982; see also the effect of plant size in Table 1a), and age at first reproduction is usually one of the traits in the life cycle of a species to which the population growth rate is most sensitive (Stearns 1992). Eutrophication, as a result of fertilizer run-off or nitrogen emissions, is a major threat to P. veris, a species of nutrient-poor grasslands, and our results suggest that fragmentation may aggravate the negative consequences of other anthropogenic changes.
Monitoring population size may not be sufficient to assess the risk of extinction of rare plants, in particular of perennial species with low adult mortality. Oostermeijer et al. (1994) have shown that in the rare plant Gentiana pneumonanthe habitat alterations may cause a lack of recruitment and thus threaten the long-term survival of populations, although there are no short-term effects on the number of flowering plants. They therefore suggested that the stage structure of populations should be monitored to assess their vitality. We show that reproductive success can be reduced strongly in small populations of perennial plants, probably long before a further decline in the size of these populations can be measured. Monitoring of reproduction in populations of rare plants may therefore provide important indicators of future population trends.
- Top of page
- Study species and area
M. Fischer, M. Dudash, M. Paschke, H. Prentice, M. Schaub, B. Schmid and an anonymous referee made valuable comments on the manuscript. U. Kienzle, T. Brodtbeck and M. Zemp helped with the selection of field sites. The late Karl Altenbach and Stephan Peter saved two P. veris populations from being mown. Miguel Schaub made available census results from one population of P. veris. Thanks to them all. This study was supported by the Swiss National Science Foundation, Priority Programme Environment (grant no. 5001-35231 to D. Matthies and B. Schmid, and no. 5001-44626 to D. Matthies, B. Schmid and P.J. Edwards).
- Top of page
- Study species and area
- 1996) Population size, pollinator limitation and seed set in the self-compatible herb Lythrum salicaria. Ecology, 77, 1779–1790. (
- 1994) Forest fragmentation, pollination and plant reproduction in a Chaco dry forest, Argentina. Ecology, 75, 330–351. & (
- 1982) Experimental ecology and genetics in Plantago. VI. The demography of seedling transplants of P. lanceolata. Journal of Ecology, 70, 55–75. & (
- 1991) Genetic and evolutionary consequences of small population size in plants: implications for conservation. Genetics and Conservation of Rare Plants (eds D.A.Falk & K.H.Holsinger), pp. 3–30. Oxford University Press, Oxford, UK. & (
- 1998) canoco Reference Manual and User's Guide to canoco for Windows: Software for Canonical Community Ordination (Version 4). Microcomputer Power, Ithaca, NY. & (
- 1987) Biosystematische Untersuchungen an Gentiana lutea L., Gentiana purpurea L. und deren Hybriden[Biosystematic investigations on Gentiana lutea L., Gentiana purpurea L. and their hybrids]. Diploma Thesis. University of Zürich, Zürich, Switzerland. (
- 1995) Pollen quantity and quality as explanations for low seed set in small populations exemplified by Eupatorium (Asteraceae). American Journal of Botany, 82, 1000–1006. (
- 1987) Inbreeding depression and its evolutionary consequences. Annual Review of Ecology and Systematics, 18, 237–268. & (
- 1987) Flora of the British Isles. Cambridge University Press, Cambridge, UK. , & (
- 1993) Relationship of breeding system to rarity in the Lakeside Daisy (Hymenoxys acaulis var. glabra). Conservation Biology, 7, 542–550. (
- 1993) Population genetic consequences of small population size: implications for plant conservation. Annual Review of Ecology and Systematics, 24, 217–242. & (
- 1997) Mating structure and inbreeding and outbreeding depression in the rare plant Gentianella germanica (Gentianaceae). American Journal of Botany, 84, 1685–1692. & (
- 1998a) RAPD variation in relation to population size and plant performance in the rare Gentianella germanica. American Journal of Botany, 85, 811–819. & (
- 1998b) The effect of population size on performance in the rare plant Gentianella germanica. Journal of Ecology, 86, 195–204. & (
- 1993) Glim4. The Statistical System for Generalized Linear Interactive Modelling. Clarendon Press, Oxford, UK. , & (
- 1988) Comparative Plant Ecology – A Functional Approach to Common British Species. Unwin, London, UK. , & (
- 1998) Allee effects limit population viability of an annual plant. American Naturalist, 151, 487–496. (
- 1991) Metapopulation dynamics: brief history and conceptual domain. Biological Journal of the Linnean Society, 42, 3–16. & (
- 1977) Population Biology of Plants. Academic Press, London, UK. (
- 1927) Illustrierte Flora Von Mitteleuropa. J.F. Lehmanns-Verlag, München, Germany. (
- 1995) Inbreeding depression, environmental stress, and population size variation in scarlet gilia (Ipomopsis aggregata). Conservation Biology, 9, 126–133. & (
- 1991) Ecological implications of genetic variation in plant populations. Genetics and Conservation of Rare Plants (eds D.A.Falk & K.E.Holsinger), pp. 31–44. Oxford University Press, New York, NY. (
- 1996) Evolution of the magnitude and timing of inbreeding depression in plants. Evolution, 50, 54–70. & (
- 1993) Insect flower visitation frequency and seed production in relation to patch size in Viscaria vulgaris (Caryophyllaceae). Oikos, 68, 283–292. & (
- 1992) Conservation biology in agricultural habitat islands. Conservation Biology by Ecological Principles (ed. L.Hansson), pp. 396–425. Elsevier, London, UK. , , , & (
- 1995) The effect of population size on reproduction of the grassland species Gentiana lutea L. and Primula veris L. Diploma Thesis. University of Basel, Basel, Switzerland. (
- 1987) Calcareous grassland – a limited resource in Britain. Calcareous Grasslands – Ecology and Management. Proceedings of a BES/NCC Symposium at the University of Sheffield, 1987 (eds S.H.Hillier, D.W.H.Walton & D.A.Wells), pp. 11–17. Bluntisham Books, Bluntisham, UK. & (
- 1996) Rote Liste der Farn- und Blütenpflanzen (Pteridophyta et Spermatophyta) Deutschlands. Schriftenreihe für Vegetationskunde, 28, 21–187. , & (
- 1997) Population size and density effects in pollination: pollinator foraging and plant reproductive success in experimental arrays of Brassica kaber. Journal of Ecology, 85, 225–234. (
- 1987) Loss of genetic diversity from managed populations: interacting effects of drift, mutation, immigration, selection, and population subdivision. Conservation Biology, 1, 143–158. (
- 1993) Population fragmentation may reduce fertility to zero in Banksia goodii– a demonstration of the Allee effect. Oecologia, 94, 446–450. , & (
- 1991) Gefährdung der Farn- und Blütenpflanzen in der Schweiz. Bundesamt für Umwelt, Wald und Landschaft, Bern, Switzerland. (
- 1988) Patterns of fruit and seed production. Plant Reproductive Ecology: Patterns and Strategies (eds J.Lovett Doust & L.Lovett Doust), pp. 179–201. Oxford University Press, New York, NY. (
- 1989) Generalized Linear Models. Chapman & Hall, London, NY. & (
- 1990) Plasticity of reproductive components at different stages of development in the annual plant Thlaspi arvense. Oecologia, 83, 105–116. (
- 1995) The importance of population processes for the maintenance of biological diversity. Gaia, 4, 199–209. , & (
- 1991a) The application of minimum viable population theory to plants. Genetics and Conservation of Rare Plants (eds D.A.Falk & K.E.Holsinger), pp. 47–61. Oxford University Press, New York, NY. (
- 1991b) Seed germination percentage increases with population size in a fragmented prairie species. Conservation Biology, 5, 158–164. (
- 1992) Stochastic modeling of extinction in plant populations. Conservation Biology: The Theory and Practice of Nature Conservation, Preservation and Management (eds P.L.Fiedler & S.K.Jain), pp. 253–276. Chapman & Hall, New York, NY. (
- 1998) Demographic viability of populations of Silene regia in midwestern prairies: relationships with fire management, genetic variation, geographic location, population size and isolation. Journal of Ecology, 86, 63–78. & (
- 1994) Fragmented plant populations and their lost interactions. Conservation Genetics (eds V.Loeschcke, J.Tomiuk & S.K.Jain), pp. 417–426. Birkhäuser, Basel, Switzerland. & (
- 1995) Effects of outcrossing distance and selfing on fitness components in the rare Gentiana pneumonanthe (Gentianaceae). Acta Botanica Neerlandica, 44, 257–268. , & (
- 1994) Offspring fitness in relation to population size and genetic variation in the rare perennial plant species Gentiana pneumonanthe (Gentianaceae). Oecologia, 97, 289–296. , & (
- 1992) Population biology and management of the marsh gentian (Gentiana pneumonanthe L.), a rare species in the Netherlands. Botanical Journal of the Linnean Society, 108, 117–130. , , & (
- 1996) Demography, genetics, or statistics: comments on a paper by Heschel & Paige. Conservation Biology, 10, 1290–1291. & (
- 1994) The significance of genetic erosion in the process of extinction. IV. Inbreeding load and heterosis in relation to population size in the mint Salvia pratensis. Evolution, 48, 996–1008. & (
- 1993) genstat 5, Version 3. Reference Manual. Clarendon Press, Oxford, UK. (
- 1991) Comparative pollination ecology of two rare Dutch Gentiana species, in relation to population size. Acta Horticulturae, 288, 308–312. , & (
- 1994) Genetic variation and outcrossing rate in relation to population size in Gentiana pneumonanthe L. Conservation Biology, 8, 1014–1026. , , , , & (
- 1997) Reproductive success increases with local density of conspecifics in a desert mustard (Lesquerella fendleri). Conservation Biology, 11, 738–746. , , & (
- 1991) Biological consequences of ecosystem fragmentation: a review. Conservation Biology, 5, 18–32. , & (
- 1986) The evolution of phenotypic plasticity in plants. Annual Review in Ecology and Systematics, 17, 667–693. (
- 1992) Phenotypic variation in plants. Evolutionary Trends in Plants, 6, 45–60. (
- 1987) Patch size, pollination behavior, and pollinator limitation in Catnip. Ecology, 68, 1679–1690. & (
- 1981) Studies on the population biology of the genus Viola. II. The effect of plant size on fitness in Viola sororia. Evolution, 35, 1080–1093. (
- 1998) Studies on the population biology of Primula veris L. s.str. and Gentiana lutea L. Diploma Thesis. University of Zürich, Zürich, Switzerland. (
- 1972) Biological inertia – an example. Journal of Ecology, 60, 793–798. (
- 1992) The Evolution of Life-Histories. Oxford University Press, Oxford, UK. (
- 1980) Principles and Procedures of Statistics, 2nd edn. McGraw-Hill, Auckland, New Zealand. & (
- 1972) Survival and flowering of perennial herbs. III. The behaviour of Primula veris on permanent plots. Oikos, 23, 159–166. (
- 1994) Relationships between plant density, outcrossing rates and seed set in natural and experimental populations of Scabiosa columbaria. Journal of Evolutionary Biology, 7, 287–302. , , & (
- 1992) Size-number trade-offs and the variation of seed size with plant resource status. American Naturalist, 140, 287–304. (
- 1994) Beyond global warming: ecology and global change. Ecology, 75, 1861–1902. (
- 1994) Patterns of female fertility in heterostylous Primula sieboldii under severe pollinator limitation. Journal of Ecology, 82, 571–579. , , & (
- 1993) Demographic and genetic effects on reproduction as related to population size in a rare, perennial herb, Senecio integrifolius (Asteraceae). Biological Journal of the Linnean Society, 50, 179–195.DOI: 10.1006/bijl.1993.1053 (
- 1986) Rückgang und Gefährdung von Mesobromion-Arten im Schweizer Jura. Veröffentlichungen Des Geobotanischen Institutes der ETH, Stiftung Rübel, 87, 239–259. & (
Received 25 March 1999revision accepted 13 July 1999