| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
2Centre d'Ecologie Fonctionnelle et Evolutive, CNRS, UPR 9056, 1919 route de Mende, 34293 Montpellier Cedex 5, France; and 4Department of Ecology and Evolutionary Biology, University of Arizona, Tucson, Arizona 85721 USA
Received for publication February 1, 2000. Accepted for publication June 22, 2000.
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
Key Words: Ficus • Florida • inbreeding depression • mutualism • phenology • pollination • seasonal environment • self-compatibility • selfing
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
; Bawa and Beach, 1981
; Lloyd and Webb, 1986
; Charlesworth and Charlesworth, 1987
; Bertin, 1993
). Temporal separation of female and male reproductive phases is frequent in tree species with bisexual individuals in neotropical forests (Bawa, Perry, and Beach, 1985
). In many cases, however, separation between phases is very short (3–5 d), and its effectiveness as a barrier to self-fertilization may be limited (Bertin, 1993
; Navarro, 1997
), particularly when floral phenology allows pollen transfer from one flower to another of the same individual (geitonogamy). In such cases, other barriers to self-fertilization may exist, and advantages of dichogamy related to sexual selection, such as reduction of pollen wastage and lowered incidence of stigma clogging by incompatible self pollen (Palmer, Travis, and Antonovics, 1989
), may be more important. In monoecious species of Ficus, however, separations between male and female phases, both at the level of inflorescences and of the individual tree, are much longer than in many other dichogamous plants. Here, the potential for geitonogamy is usually very limited and outcrossing strongly predominates (Janzen, 1979
; Wiebes, 1979
; Nason, Herre, and Hamrick, 1996
).
Figs are also well known for their obligate mutualism with pollinating fig wasps. Agaonid wasps are the only pollinators of fig inflorescences, and the wasps are only able to reproduce within the fig they pollinate. In mutualisms between monoecious figs and their pollinators, the separation of sexual phases is due to different mechanisms at the inflorescence and individual level. The extreme dichogamy within each inflorescence (protogyny, with female and male phases separated by a period of 4–12 wk, depending on the species [Hill, 1967
; Janzen, 1979
; Bronstein, 1989
; Bronstein et al., 1990
; Anstett, Hossaert-McKey and Kjellberg, 1997
]) is due to the fact that female pollinating wasps are attracted to inflorescences (syconia, or figs) in receptive (female) phase, while pollen is removed within the inflorescences by the pollinator offspring that emerge 1–3 mo later. At the level of the individual tree, the separation of sexual phases is due to intracrown synchrony in fig development, in the majority of fig species. In some monoecious fig species, intracrown asynchrony does occur (Baijnath and Ramcharun, 1983, 1988
; Bronstein, 1989
; Cook and Power, 1996
), but appears rarely to be sufficient to permit intratree overlap in sexual phases (Bronstein and Patel, 1992a
; Smith and Bronstein, 1996
). Thus when the new generation of pollinating wasps, loaded with pollen, are ready to exit male-phase figs, they must leave their natal fig tree in search of another tree with receptive, female-phase figs. Asynchrony among conspecific individuals, with some trees in flower at any time of the year, increases the probability of such encounters and thus of outcrossing. Due to these features of floral phenology at the inflorescence, individual, and population levels, outcrossing is expected to predominate in fig populations.
In fact, population genetic data for several species of monoecious figs show them to be indeed highly outcrossed, and breeding population size of monoecious fig species has been reported to be surprisingly high (Nason and Hamrick, 1997
). It is thus a reasonable hypothesis that, as in other highly outcrossed plant species (Husband and Schemske, 1996
), fig populations may possess deleterious recessive alleles that lead to inbreeding depression when selfing does occur. Avoidance of inbreeding depression may thus be an important selective factor maintaining the separation of sexual phases at the individual level.
By virtue of their specialized pollination biology, figs might also be subject to a novel type of inbreeding depression, the depression of the selfed tree's male reproductive fitness. Pollen produced by a fig is transported by the pollinators that develop in the fig. According to the leading hypothesis that attempts to explain the occurrence of active pollination in many pollinating fig wasps (Galil and Eisikowitch, 1971
; Pellmyr, 1997
), wasp larvae eat seeds. Wasp development thus depends on seed development. If selfing negatively affects seed development, it might thereby affect development of pollinator offspring and thus the plant's male reproductive fitness. This factor might reinforce selection maintaining the separation of sexual phases at the individual level.
Some aspects of reproductive phenology of monoecious figs show phenotypic plasticity, being directly affected by environmental variation (e.g., Damstra, Richardson, and Reeler, 1996
). However, such variation has never been shown to be sufficient to lead to intracrown overlap in sexual phases.
While it is clear that phenological patterns typical of monoecious figs should result in primarily outcrossed pollination, it is not clear that avoidance of selfing is the evolved function of those patterns. Strong protogyny, mass flowering, and intratree synchrony play several functionally important roles in fig pollination mutualisms (Janzen, 1979
; Wiebes, 1979
; Bronstein, 1992
). For example, synchronous production of receptive female-phase figs may be important in that it results in higher concentrations of the volatile compounds necessary for long-distance attraction of pollinators to the tree (Hossaert-McKey, Gibernau, and Frey, 1994
; Gibernau et al., 1998
). Synchronous production of male-phase figs may be important in satiating predators of emerging female pollinators (Bronstein, 1987
) or in mass attraction of seed dispersers when figs mature shortly after male phase (Kalko, Herre, and Handley, 1996
). Whether fig phenology has been driven by selection to avoid inbreeding or whether inbreeding avoidance is a secondary effect of an adaptation produced by other selective forces is open to question.
In either case, the hypothesis that these phenological patterns can reduce inbreeding depends on the assumption that monoecious figs are self-compatible. Also, the hypothesis that avoidance of inbreeding increases fitness assumes that inbreeding depression occurs in monoecious figs. To our knowledge, neither of these fundamental assumptions has ever been tested.
The species studied here, F. aurea, offers the possibility of examining both assumptions. As in a few other species of monoecious figs (Hill, 1967
; Bronstein, 1989
; Berg, 1990
; Smith and Bronstein, 1996
), overlapping of sexual phases at the level of the individual trees, although rare, does occur. Bronstein and Patel (1992a)
censused 17 trees of F. aurea at 2–4 d intervals over a period of 2 yr. In only 
5% of all census dates did one or more of the trees show intratree overlap in sexual phase. The rarity of intracrown sexual-phase overlap has been confirmed by our observations in following years (Bronstein and Hossaert-McKey, 1996
). During these periods of overlapping sexual phases, controlled experiments comparing self- and cross-pollination can be performed. In this paper we report the results of such experiments, designed to address the following questions: (1) Is F. aurea self-compatible? (2) If so, how does inbreeding affect syconium abortion, seed and pollinator production, and seed germination? (3) Does selfing affect the contribution of a syconium to a tree's male fitness, which depends on the number of female pollinators it produces? We then discuss whether evidence exists for inbreeding depression in monoecious fig species.
|
METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). In Florida, it grows abundantly in patches of woodland (locally known as hammocks) on the mainland and in the Florida Keys, as well as along canals and other watercourses. Trees are large, reaching a height of 15 m and a canopy spread of 25 m or more, and they often show an hemiepiphytic, strangling habit (Bessey, 1908
; Condit, 1969
). Syconia (or figs) are axillary and paired, and nearly sessile. They are subglobose and somewhat puberulent, with large and conspicuous bracts and a small ostiole (the bract-covered pore through which pollinators enter the fig). Syconium size is variable (6–11 mm at maturity), as is color at maturity (orange to purple). When some syconia of one tree are in female phase, the pollinating wasps, oriented to volatile chemicals released by receptive syconia (Ware et al., 1993
; Hossaert-McKey, Gibernau, and Frey, 1994
), are attracted and enter the syconia through the bract-lined ostiole. Once inside the syconia, wasps pollinate female florets and oviposit into some of the ovaries. After the interfloral phase (from 3 wk to 2 mo), during which seeds and wasp larvae complete their development, all the syconia enter male phase. At this phase pollen matures and becomes available, and a cohort of newly matured wasps emerges from their galls. The newly inseminated female wasps then actively collect pollen in special structures on their mesothoracic sternites called "pollen pockets" and leave the syconia. Due to the synchrony of each sexual phase within a crop, the short-lived wasps must usually leave their natal fig tree in search of other receptive syconia in which to oviposit and pollinate.
Ficus aurea is pollinated exclusively by Pegoscapus mexicanus (Grandi) Wiebes (Hymenoptera: Agaonidae) (Wiebes, 1995
), identified as P. jimenezi (Grandi) in earlier literature. The biology of this wasp and its interactions with F. aurea in south Florida have been detailed elsewhere (e.g., Frank, 1984
; Bronstein and Patel, 1992a, b
; Bronstein and Hossaert-McKey, 1995, 1996
). At least eight other highly specialized wasps—gallers, parasitoids, and inquilines—also live and feed within F. aurea syconia in this region (Nadel, Frank, and Knight, 1992
; Bronstein and Hossaert-McKey, 1996
; Bronstein, 1999
).
The study was conducted at two sites in Coral Gables, Dade County, Florida, USA (25°43' N, 80°16' W). All but one of the experiments were conducted on the University of Miami campus, using trees growing in hammock remnants along a canal. This site is described in more detail elsewhere (Bronstein and Patel, 1992a
). An additional experiment was conducted in a hammock remnant on the grounds of the Montgomery Foundation, located 8 km from the campus site. Phenological patterns within the campus population were studied intensively between 1988 and 1996 (e.g., Bronstein and Patel, 1992a
; Bronstein and Hossaert-McKey, 1995, 1996
); phenologies of the Montgomery Foundation trees are apparently quite similar to those of trees of the campus population (M. Hossaert-McKey, personal observation).
Pollination experiments
Breeding experiments in fig trees were performed using wasps as "pollination tools." In actively pollinated species of figs, wasps emerging from male-phase figs carry pollen in their "pollen pockets." If they are manually deposited on the surface of figs in female phase, they will enter the figs and pollinate the flowers within the figs (Khadari et al., 1995
). In other monoecious fig-pollination systems, each wasp may carry between 75 and 250 pollen grains in its pockets (Kathuria Gupta, 1999
).
A total of eight wasp introduction experiments were performed between 1991 and 1995 on seven trees. Trees were used if they showed the potential to self-pollinate in the immediate future, i.e., if they bore both (a) large numbers of syconia approaching the female or receptive phase, and (b) some syconia approaching male phase (containing nearly mature wasps ready to exit with pollen). This level of within-tree flowering asynchrony is found in F. aurea only rarely and sporadically during the year (Bronstein and Patel, 1992a
).
Once a suitable tree was located, 3–8 branches bearing prefemale-phase syconia were haphazardly selected for the experiment. After removing syconia in other phenological stages, a map was drawn of each branch, with each syconium and leaf assigned a unique number. The branch was then encased in a fine-mesh nylon bag to prevent entry by pollinators that might arrive naturally at the tree. Each bag was sealed tightly around the branch with string tied around a band of thick cotton and Tanglefoot® (Tanglefoot Company, Grand Rapids, Michigan, USA) insect glue.
Bags were removed daily to assess whether the protected syconia were large enough (4–6 mm) to suggest that they might be receptive to pollinators. Once they had attained this size, receptivity was assessed using a bioassay method. Pollinators were collected from male-phase figs and placed individually on a syconium (methods are described in more detail below). If the wasps exhibited behaviors indicating their imminent entry (see below), they were removed and it was concluded that the syconium had reached receptivity. In this case, the introduction experiment was begun on the same or subsequent day. If they did not, the bag was resealed and the bioassay was repeated the next day.
Once syconia had reached receptivity, we collected pollen-carrying wasps from two trees: the experimental tree itself (for the selfing treatment, in this case geitonogamy), and another tree in the population (for the outcrossing treatment). Wasps for the selfing treatment were collected from branches on the same part of the trees. We examined trees carefully to avoid trees or branches that might actually be the product of anastomosis of two different individuals (in fact, this phenomenon seems to be more rare than previously stated; see Thomson et al. [1997]
for discussion). No tree was used as a pollen/wasp source for more than one outcrossing experiment. In each case, we removed 50–100 early male-phase syconia (i.e., those containing mature female wasps ready to collect pollen and to leave in search of another tree), recognizable by a characteristic softening and enlargement of the syconium to 8–11 mm in diameter. Syconia were placed intact into a fine-mesh nylon bag (one bag for each source tree), which was sealed and left overnight. During this time, newly mated female pollinators within the figs collect pollen from the dehiscing anthers while male pollinators dig an exit tunnel through the syconium wall, allowing the females to escape. Thus, the following morning, pollen-carrying females were available in the bag for use in the experiments.
All introductions were carried out the morning after wasp collection; the wasps live only 24 h and in Florida become torpid and die in the afternoon heat (M. Hossaert-McKey, personal observation). At the experimental tree, one protective bag was first removed. At this stage, and throughout the subsequent introduction period, the experimental branch was continually scanned to make sure no naturally arriving wasps were on the leaves or syconia. One wasp from among those collected from one of the two source trees was then picked up with a fine camel-hair brush and deposited on a syconium. If that syconium was receptive, she would circle it, orient her head towards the ostiole, push up the bracts of the ostiole and enter it, usually within 5 min. If the wasp left the syconium or had not entered after 15 min, she was removed and a different wasp was used. If three consecutive wasps failed to enter, we concluded that the syconium was not yet receptive to pollination, and it was not used for the experiment that day. When an entry was successful, we recorded the syconium number, its diameter (measured with calipers), and the pollen source used. Only one wasp was introduced per syconium. This procedure reflects the usual number of foundresses observed within figs of this species. In open-pollinated figs of F. aurea, the number of foundresses per fig ranged from 0 to 24, and the modal value (43% of all cases) was one foundress (Bronstein and Hossaert-McKey, 1996
). We attempted as much as possible to assign equal numbers of syconia on each branch to each treatment (selfed vs. outcrossed pollen) and to have adjacent syconia assigned to contrasting treatments (to control for environmental effects such as light and nutrient availability).
When as many introductions as possible had been carried out, the protective bag was replaced. Introductions were then conducted on the next bagged branch. Introductions were concluded for the day either when wasps had been introduced to all bagged syconia, when no more wasps were available for introductions, or when the weather became too hot, windy, or wet for the wasps to behave normally. If the experiment had to be stopped before completion, it was resumed on the subsequent day or whenever the remaining syconia reached receptivity (usually within 3–5 d). In total, 10–43 introductions were carried out per experiment; sample sizes were constrained in some cases by limited availability of either wasps or receptive syconia.
Protective bags were left on the experimental branches until the syconia were mature (1–2 mo). Aborted syconia were removed periodically from the bags and their identities noted, based on the branch maps. In the same bags, figs to which no pollinators were introduced were used as controls to confirm that our bags effectively excluded pollinators. These figs invariably were aborted, showing that presence of wasps is required for fig development, and that bagging was effective in excluding pollinators. Bags were checked daily once pollen-carrying wasps were observed to be emerging from unbagged syconia. We assessed whether each experimentally pollinated syconium had reached early male phase. If it had, its diameter was measured and recorded and it was removed from the tree.
Within an hour of collection, syconia were split open and placed individually into labeled, mesh-covered vials. After 24 h, all wasps were removed from the vial and counted. The syconium remains were then dissected, and the numbers of male flowers, seeds, vacant female flowers, and female flowers exited by wasps were recorded. In some cases, extracted seeds were air-dried and retained for germination tests.
Germination experiments
Germination was tested for seeds from three trees within one month of their collection. All seeds from all syconia containing 
10 seeds were used. Seeds were first washed in a 10% chlorine solution to prevent fungal contamination, a treatment shown previously to have no effect on germinability (M. Hossaert-McKey, unpublished data). They were then rinsed for 1 h in distilled water. Up to 50 seeds from a given syconium were then counted out and placed on filter paper in a glass-bead-filled petri dish maintained at a constant level with distilled water. Dishes were haphazardly arranged in a growth chamber set for a 14-h light/30°C and 10-h dark/20°C cycle. Dish positions in the growth chamber were switched every 2–3 d.
The number of newly germinated seeds was recorded for each dish every 2–3 d. Germination was scored if the tip of the radicle protruded from the seed coat. The experiment ended when no germination occurred for a week. Total length of the germination period ranged from 8 to 55 d.
Estimate of inbreeding depression
Finally, for each individual tree we estimated the inbreeding depression for different stages of the reproductive cycle: seed production, wasp production, and seed germination. Inbreeding depression for each stage was estimated as 
= 1 - ws/wo with ws and wo being the mean relative fitness (estimated by percentage of seeds or wasps produced on the total number of flowers, and number of seeds germinated) for selfed or outcrossed progeny (Holsinger, 1988
). It was not possible to estimate inbreeding depression for one tree (ID) where the total number of flowers in each fig was not recorded.
Data analysis
Syconium abortion rates were compared between selfed and outcrossed treatments using nonparametric Wilcoxon signed-rank tests. Development time was calculated as the average number of days between pollination and wasp departure (i.e., between when we introduced a pollinator and when we removed the mature syconium from the tree) for all the figs within a tree. Difference in this development rate due to treatment was tested by a nonparametric Mann-Whitney test.
A global analysis was performed using a fixed-model ANOVA (Type III SS; GLM on SAS, version 7.0) examining the effect on seed production, wasp production, and number of vacant ovaries, of the following factors: tree (or date), branch, total number of flowers within each fig, and treatment (selfing vs. outcross). As the tree (or date) effect was always highly significant, statistical comparisons between selfed and outcrossed syconia were completed by a separate analysis on a tree-by-tree basis (nonparametric Mann-Whitney tests). The two factors "tree" and "date" are confounded, because we were only able to perform the experiment on one date for each tree. Previous studies have shown that the individual tree effect is one of the main factors determining patterns of fig and wasp production in F. aurea (Bronstein and Hossaert-McKey, 1996
). In fact, the only tree we studied for two different years (tree 25) showed very similar production of seeds and wasps for both years studied. Also, syconium developmental rate is a function of ambient temperature in F. aurea (Bronstein and Patel, 1992b
), and date (experiments were conducted over different seasons and years) may thus also contribute. Comparisons between syconia entered by wasps carrying selfed vs. outcrossed pollen were made on six variables: abortion rate during development; development time; numbers of seeds; pollinating wasps; male flowers produced per matured syconium; and proportion of germinated seeds.
In these analyses, we used number of female flowers exited by wasps, rather than agaonid wasp numbers per se, as a measure of pollinator wasp production. Over 99% of all wasps developing in these bagged figs of F. aurea were the pollinator species; hence, the number of exited flowers should have almost exactly equaled the number of pollinator wasps. In practice, several sources of error, e.g., collection of the syconium after some or all female wasps had escaped, tended to lower the number of wasps retrieved, making counts of exited flowers a more reliable estimate of pollinator production. Because we used this method, however, it was not possible to separate male and female agaonids in our wasp counts and subsequent analyses. Such separation would have been useful because only the females are the tree's pollen vectors. However, the sex ratio in single-foundress figs is 11 females to 1 male (0.093 ± 0.006) and was not found to differ significantly between selfed and outcrossed syconia in the cases where this comparison was possible (Mann-Whitney, U = 325.5, P = 0.826). Our total counts should thus be very strongly correlated with number of female agaonids produced.
At maturity, about one quarter of the experimental syconia unexpectedly contained Anidarnes bicolor (Ashmead) Bou
ek (3.3 ± 2.9 individuals, range 1–12), a non-pollinating fig wasp that develops within sterile galls formed on the inner wall of the syconium (Bronstein, 1999
). Females of this species oviposit into very small, prereceptive syconia (Bronstein, 1999
); and it is highly probable that they had attacked the experimental syconia before they were encased in protective bags. We did not exclude Anidarnes-infested syconia from statistical analyses, since preliminary analyses showed that (a) selfed and outcrossed syconia on a given tree were infested equally (Wilcoxon signed-rank test, Z = -0.809, P = 0.418, df = 6) and (b) on one tree with an infestation rate equal to the population mean, neither seed nor pollinator production was affected by presence of Anidarnes (Mann-Whitney tests, all P > 0.70, df = 27).
Germination rate was calculated on a per-syconium basis, then averaged within treatment for each tree. Nonparametric comparisons (Mann-Whitney) between treatments were made tree by tree.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
|
0.15), except in a single case where selfed syconia took significantly longer to develop (tree 9, P < 0.003). As expected, there were large differences in development time across trees, reflecting the season in which the experiments were conducted. The overall difference including all the experiments was tested using the combined probabilities (Fisher combined probabilities) obtained for each tree (Sokal and Rohlf, 1969
, p. 780). This test showed that proportion of syconia aborted was not significantly different (P > 0.05) between selfed and outcrossed treatments.
Contents of syconia
The general ANOVA performed on all trees combined showed that effects of individual tree and of number of female flowers per fig were significant factors in determining production of seeds, pollinators, and vacant ovaries (Table 2). Individual branch significantly affected production of pollinators and of vacant ovaries and tended (P < 0.07) to affect production of seeds. In contrast, type of cross, self or outcross, only had a marginally significant effect on number of pollinators produced and no effect on production of seeds and of vacant ovaries.
|
0.14) between the selfed and outcrossed treatment in any of the eight experiments. There were more wasps per selfed syconium in six cases and more per outcrossed syconium in the other two. In only the one previously mentioned case did number of female flowers differ significantly between syconia assigned to different treatments (Table 3). Similarly, number of male flowers differed significantly (P < 0.05) between treatments in only one experiment (Table 3).
|
0.40). In each experiment there was a weak trend towards higher germination among outcrossed seeds. However, the combined probabilities (Fisher test) showed that the overall difference in germination rate was not significant (P > 0.05) between selfed and outcrossed treatments.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
). The ancestors of figs may thus have already possessed phenologies that led to only weak selection, if at all, for incompatibility mechanisms. Studies in this area, as well as phylogenetic studies to identify ancestral states of phenological and breeding-system characters in the lineage that led to figs, could provide important insights for interpreting breeding systems of figs.
Effects of selfing on female reproductive success
Postzygotic mechanisms to avoid selfing could not be demonstrated on either seed production or seed germination. Our results (cumulative inbreeding depression near to zero) are similar to values reported by Husband and Schemske (1996)
as typical of self-fertilizing species. They contrast strongly with those obtained for Yucca, another plant genus with a highly specific active pollinator, by Richter and Weis (1998)
, who report a cumulative inbreeding depression of 0.68, even if deposition of self-pollen is quite frequent in this obligate pollination system (Marr et al., 2000)
. As in most other studies of long-lived woody angiosperms (Husband and Schemske, 1996
), we lack data on effects of inbreeding on later life history stages. However, in most highly outcrossed species, inbreeding depression affects both early (seed production) and late (growth/reproduction) life history stages (Husband and Schemske, 1996
). Early-acting inbreeding depression is usually due to lethal recessive alleles, which are expected to be more frequent in highly outcrossed populations. The absence of demonstrable inbreeding depression at the early life history stages we studied is thus surprising.
Why do individuals of this species, whose pollination biology suggests that it should be strongly outcrossed (like all other figs that have been studied), fail to exhibit inbreeding depression when experimentally selfed? One possibility is that inbreeding depression occurs, but is expressed only at later life history stages that we did not study. As noted above, we cannot exclude this hypothesis, but the absence of inbreeding depression at the seed formation stage is in any case surprising for a characteristically outcrossed species (Husband and Schemske, 1996
).
Another possibility is that individuals used in the experiments were all extremely closely related, so that the experiments provided little discrimination between selfing and outcrossing. Genetic variation in this population might be limited, for two reasons. First, ornamental trees in an anthropogenically altered habitat may have been planted from limited genetic stock, possibly even clonally propagated. However, the individuals we used were large, old trees, mostly along watercourses and are remnant elements from native vegetation that began to be altered in the 1920s, not planted trees. Secondly, bottlenecks during colonization of Florida may have reduced genetic diversity. Ficus aurea is a Caribbean species, and southern Florida is the northern edge of its range. Founder effects in the colonization of Florida, or severe weather events that reduced populations of wasps and/or trees, may have reduced within-population genetic diversity to the point where selfing and outcrossing have similar effects. The history of these populations is unknown, and no data exist on genetic variation in this species, either in Florida or elsewhere, that would shed light on mating systems. However, we consider it unlikely that purging or reduction of genetic variation by previous bottlenecks can explain the absence of demonstrable inbreeding depression at early stages of development. The requirement, for establishment and maintenance of fig pollinators, of a critical minimum size of the tree population (Janzen, 1979
; Bronstein et al., 1990
) would be expected to have limited any genetic drift during colonization. In the population we studied, the density of trees is very high, and natural recruits are abundant. Furthermore, F. aurea does not possess the phenological traits that would allow year-round cycling of pollinator populations in one or a very small number of trees (Janzen, 1979
). The intracrown asynchrony that permits occasional selfing in this species is restricted to the summer rainy season, when pollinator population densities are highest (Bronstein and Patel, 1992a
; Bronstein and Hossaert-McKey, 1996
), and thus cannot act to ensure selfing when low pollinator densities make outcrossing most difficult.
Effects of selfing on male reproductive success
Selfing also appeared not to depress male reproductive fitness. As already mentioned, selfing did not affect the probability of syconium abortion, and in syconia borne to maturity neither production of male flowers nor of pollinating wasps, which together determine the contribution of a syconium to male fitness, was affected by selfing.
A negative effect of selfing on wasp development could have been expected if selfing affected seed set and if, as according to the commonly accepted explanation for active pollination (Galil and Eisikowitch, 1971
), development of wasps depends on development of seeds. However, our results show that selfing appears not to affect seed set or seed development. Furthermore, recent findings show that development of the pollinator of F. aurea does not depend on pollination of the ovule in which the larva grows, and retention of a syconium does not depend on pollination of some of its female flowers (Jousselin, Hossaert-McKey, and Kjellberg, unpublished data). Because in these experiments, contrary to expectations and results from other species (Anstett, Hossaert-McKey, and Kjellberg, 1997
; Herre and West, 1997
), pollination has limited effect on pollinator development, it is not surprising that source of pollen, self or outcross, also had no effect.
This finding implies that wasps entering female-phase syconia of their natal tree would not suffer reduced fitness as a result of self-pollination. This further implies that selection acting on wasp behavior would not oppose the spread of plant genotypes that allow selfing.
Self-compatibility, adaptation of figs to seasonal environments, and colonizing ability
Janzen (1979)
proposed that in highly seasonal environments, crop failure due to periodic reductions of pollinator densities might favor individual fig trees with relaxed intracrown synchrony in sexual phases, allowing pollinators to cycle within the crown of a single tree. This hypothesis implicitly assumed that pollinators can develop in self-pollinated figs. Our data are the first test of this assumption. Janzen (1979)
also noted that by reducing the number of trees necessary to maintain the pollinator, such a phenology could facilitate establishment of small populations in newly colonized sites. In species with the appropriate phenology, self-compatibility would contribute to the colonizing ability of figs. Again, our results supply the first hard data on self-compatibility in figs. As noted above, the mechanisms postulated by Janzen (1979)
appear unlikely to apply to F. aurea, because selfing is impossible at the time of year when it would be most crucial and rare at other times. However, his hypotheses may apply to other species (Smith and Bronstein, 1996
), and our study suggests that the key assumptions behind them are plausible.
Why are phenologies that permit selfing not more frequent in monoecious figs?
The fig–fig wasp mutualism is horizontally transmitted and the encounter of both partners must be renewed at each generation. This phase of encounter of both partners is a crucial constraint in the maintenance of the mutualism. In fact, from empirical studies, Herre (1989)
and Bronstein (1992)
pointed out the huge mortality of the wasps in flight between trees. They estimated a wasp mortality of >99% during dispersal. Wasps that could stay within the tree would be almost assured of having at least some reproductive success (Cook and Power, 1996
), but the vast majority of wasps that leave the tree do not. In light of the transmission advantage always associated with selfing, and the absence of demonstrable inbreeding depression, why are these populations of F. aurea not invaded by trees with phenologies that permit frequent selfing? One likely answer is that selection pressures other than the avoidance of inbreeding enforce intratree synchrony. As pointed out in the introduction, intracrown synchrony may be important in several phases of fig development, enhancing pollinator attraction, minimizing predation of emerging pollinators, and attracting seed dispersers to mature figs. Also, reproductive success of both tree and wasp may be higher if wasps disperse. For the tree, male fitness is higher if its pollen is dispersed to sire other crops, rather than competing locally for ovules of the same individual. Furthermore, passage of newly emerged females to receptive figs on their natal trees would likely increase the number of foundresses per fig. The proportion of female wasp progeny, which alone contribute to the tree's male fitness, is highest when foundresses enter figs singly (Herre, 1989
). Trees also leave the most seeds, and wasps leave the most offspring, when foundresses enter figs singly (Anstett, Bronstein, and Hossaert-McKey, 1996
; Nefdt and Compton, 1996
). It may thus be in the interests of both fig and wasp if foundresses do not compete within figs (Bronstein, Vernet, and Hossaert-McKey, 1998
), and this is more likely to be achieved if wasps depart their natal trees.
Thus, in figs, as in other plants, while dichogamy and other phenological traits may reduce the likelihood of selfing, the selection pressures responsible for the origin and maintenance of these traits might well lie elsewhere (Palmer, Travis, and Antonovics, 1989
; Bertin, 1993
).
Our results by no means demonstrate that outbreeding is an unimportant consequence of fig phenology. As already noted, inbreeding depression may be expressed only in later developmental stages. Furthermore, the consequences of inbreeding might not be seen in a benign greenhouse environment, perhaps becoming evident only when pathogens or herbivores attack a highly homozygous plant or spread in a local population stripped of genetic diversity.
In conclusion, these surprising results need to be completed by a study on the genetic structure of this population. Does genetic relatedness between trees used as pollen donors and maternal parents partly explain our results? Such studies could also profitably focus on other fig species that are known to have enough individual-level asynchrony to permit selfing, such as F. microcarpa. Such studies must combine investigations of phenology, pollinator population dynamics, breeding system, and population genetic structure. Too often, data are available only on one or the other aspect, rarely on all aspects for the same population. Only a combined approach will enable us to understand the system.
|
|
FOOTNOTES |
|---|
|
LITERATURE CITED |
|---|
|
TOPABSTRACTINTRODUCTIONMETHODSRESULTSDISCUSSIONLITERATURE CITED |
|---|
———, M. Hossaert-McKey, and F. Kjellberg. 1997 Figs and fig pollinators: evolutionary conflicts in a coevolved mutualism. Trends in Ecology and Evolution 12: 94–99[CrossRef]
Baijnath, H., and S. Ramcharum. 1983 Aspects of pollination and floral development in Ficus capensis Thunb. (Moraceae). Bothalia 14: 883–888
———, and ———. 1988 Reproductive biology and chalcid symbiosis in Ficus burtt-davyi (Moraceae). Monographs in Systematic Botany of the Missouri Botanical Garden 25: 227–235
Bawa, K. S., and J. H. Beach. 1981 Evolution of sexual systems in flowering plants. Annals of the Missouri Botanical Garden 62: 254–274
———, D. R. Perry, and J. H. Beach. 1985 Reproductive biology of tropical lowland rain forest trees. I. Sexual systems and incompatibility mechanisms. American Journal of Botany 72: 331–345[CrossRef][ISI]
Berg, C. C. 1990 Reproduction and evolution in Ficus (Moraceae): traits connected with the adequate rearing of pollinators. Memoirs of the New York Botanical Garden 55: 169–185
Bertin, R. I. 1993 Incidence of monoecy and dichogamy in relation to self-fertilization in angiosperms. American Journal of Botany 80: 557–560[CrossRef][ISI]
Bessey, E. A. 1908 The Florida strangling figs. Missouri Botanical Garden Annual Report 19: 25–33
Bronstein, J. L. 1987 Predators of fig wasps. Biotropica 20: 215–219
———. 1989 A mutualism at the edge of its range. Experientia 45: 622–636[CrossRef][ISI]
———. 1992 Seed predators as mutualists: ecology and evolution of the fig/pollinator interaction. In E. A. Bernays [ed.], Insect–plant interactions, vol. IV, 1–44. CRC Press, Boca Raton, Florida, USA
———. 1999 The biology of Anidarnes bicolor (Hymenoptera, Agaonidae, Sycophaginae), a galler of Ficus aurea. Florida Entomologist 82: 454–461[CrossRef][ISI]
———, P. H. Gouyon, C. Gliddon, F. Kjellberg, and G. Michaloud. 1990 The ecological consequences of flowering asynchrony in monoecious figs: a simulation study. Ecology 71: 2145–2156[CrossRef][ISI]
———, and M. Hossaert-McKey. 1995 Hurricane Andrew and a Florida fig pollination mutualism: resilience of an obligate interaction. Biotropica 27: 373–381[CrossRef][ISI]
———, and ———. 1996 Variation in reproductive success within a subtropical fig/pollinator mutualism. Journal of Biogeography 23: 433–446[CrossRef][ISI]
———, and A. Patel. 1992a Causes and consequences of within-tree phenological patterns in the Florida strangling fig, Ficus aurea (Moraceae). American Journal of Botany 79: 41–48[CrossRef][ISI]
———, and ———. 1992b Temperature sensitive development: consequences for local persistence of two subtropical wasps species. American Midland Naturalist 128: 397–403[CrossRef][ISI]
———, D. Vernet, and M. Hossaert-McKey. 1998 Do fig wasps interfere with each other during oviposition? Entomologia Experimentalis et Applicata 87: 321–324[CrossRef][ISI]
Charlesworth, D., and B. Charlesworth. 1987 Inbreeding depression and its evolutionary consequences. Annual Review of Ecology and Systematics 18: 237–268[CrossRef][ISI]
Condit, I. J. 1969 Ficus, the exotic species. Division of Agricultural Sciences, Berkeley, California, USA
Cook, J. M., and S. A. Power. 1996 Effects of within-tree flowering asynchrony on the dynamics of seed and wasp production in an Australian fig species. Journal of Biogeography 23: 487–494
Damstra, K. St. J., S. Richardson, and B. Reeler. 1996 Synchronized fruiting between trees of Ficus thonningii in seasonally dry habitats. Journal of Biogeography 23: 495–500[CrossRef][ISI]
Darwin, C. R. 1876 The effects of cross and self fertilization in the vegetable kingdom. John Murray, London, UK
Frank, S. A. 1984 The behavior and morphology of the fig wasps Pegoscapus assuetus and P. jimenezi: descriptions and suggested behavioral characters for phylogenetic studies. Psyche 91: 289–308
Galil, J., and D. Eisikowitch. 1971 Studies on mutualistic symbiosis between syconia and sycophilous wasps in monoecious figs. New Phytologist 70: 773–787[CrossRef][ISI]
Gibernau, M., M. Hossaert-McKey, J. E. Frey, and F. Kjellberg. 1998 The evolution of entomophilous pollination: insights from the study of the attraction of the mutualistic fig pollinators by fig trees. Ecoscience 5: 306–311[ISI]
Herre, E. A. 1989 Coevolution of reproductive characteristics in twelve species of new world figs and their pollinator wasps. Experientia 45: 637–647[CrossRef][ISI]
———, and S. A. West. 1997 Conflicts of interest in a mutualism: documenting the elusive fig wasp–seed trade-off. Proceedings of the Royal Society of London Series B 264: 1501–1507
Hill, D. S. 1967 Figs of Hong Kong. Hong Kong University Press, Hong Kong
Holsinger, K. E. 1988 Inbreeding depression doesn't matter: the genetic basis of mating system evolution. Evolution 42: 1235–1244[CrossRef][ISI]
Hossaert-McKey, M., M. Gibernau, and J. Frey. 1994 Chemosensory attraction of fig wasps to substances produced by receptive figs. Entomologia Experimentalis et Applicata 70: 185–191[CrossRef][ISI]
Husband, B. C., and D. W. Schemske. 1996 Evolution of the magnitude and timing of inbreeding depression in plants. Evolution 50: 54–70
Janzen, D. H. 1979 How to be a fig. Annual Review of Ecology and Systematics 10: 13–51
Kalko, E. K., E. A. Herre, and C. D. Handley. 1996 Fruit characteristics and their relation to bats. Journal of Biogeography 23: 565–576[CrossRef][ISI]
Khadari, B., M. Gibernau, M. C. Anstett, F. Kjellberg, and M. Hossaert-McKey. 1995 When figs wait for pollinators: the length of fig receptivity. American Journal of Botany 82: 992–999[CrossRef][ISI]
Kathuria Gupta, P. 1999 The pollination biology of monoecious fig trees in India. Ph.D. dissertation, University of Leeds, School of Biology, Leeds, UK
Lloyd, D. G., and C. J. Webb. 1986 The avoidance of interference between the presentation of pollen grains and stigmas in angiosperms. I. Dichogamy. New Zealand Journal of Botany 24: 163–178[ISI]
Marr, D. L., J. Leebens-Mack, L. Elms, and O. Pellmyr. 2000 Pollen dispersal in Yucca filamentosa (Agavaceae): the paradox of self-pollination behavior by Tegeticula yuccasella (Prodoxidae). American Journal of Botany 87: 670–677
Milton, K. 1991 Leaf changes and fruit production in six neotropical Moraceae species. Journal of Ecology 79: 1–26[CrossRef][ISI]
Nadel, H., J. H. Frank, and R. J. Knight. 1992 Escapees and accomplices: the naturalization of exotic Ficus and their associated faunas in Florida. Florida Entomologist 75: 29–38[CrossRef][ISI]
Nason, J. D., and J. L. Hamrick. 1997 Reproductive and genetic consequences of forest fragmentation: two case studies of neotropical canopy trees. The Journal of Heredity 88: 264–276
———, E. A. Herre, and J. L. Hamrick. 1996 Paternity analysis of the breeding structure of strangler fig populations: evidence for substancial long-distance wasp dispersal. Journal of Biogeography 23: 501–512[CrossRef][ISI]
Navarro, L. 1997 Is the dichogamy of Salvia verbenacea (Lamiaceae) an effective barrier to self-fertilization? Plant Systematics and Evolution 207: 111–117[CrossRef][ISI]
Nefdt, R. J. C., and S. Compton. 1996 Regulation of seed and pollinator production in the fig–fig wasp mutualism. Journal of Animal Ecology 65: 170–182[CrossRef][ISI]
Palmer, M., J. Travis, and J. Antonovics. 1989 Temporal mechanisms influencing gender expression and pollen flow within a self-incompatible perennial, Amianthium muscaetoxicum (Liliaceae). Oecologia 78: 231–236[CrossRef][ISI]
Pellmyr, O. 1997 Pollinating seed eaters: why is active pollination so rare? Ecology 78: 1655–1660[CrossRef][ISI]
Richter, K. S., and A. E. Weis. 1998 Inbreeding and outcrossing in Yucca whipplei: consequences for the reproductive success of plant and pollinator. Ecology Letters 1: 21–24
Smith, C., and J. L. Bronstein. 1996 Site variation in reproductive synchrony in three neotropical figs. Journal of Biogeography 23: 477–486
Sokal, R. R., and F. J. Rohlf. 1969 Biometry. Freeman, New York, New York, USA
Thomson, J. D., S. Dent-Acosta, P. Escobar-Páramo, and J. D. Nason. 1997 Within-crown flowering synchrony in strangler figs, and its relationship to allofusion. Biotropica 29: 291–297[CrossRef][ISI]
Ware, A. B., S. G. Compton, P. T. Kaye, and S. van Noort. 1993 Fig volatiles: their role in attracting pollinators and maintaining pollinator specificity. Plant Systematics and Evolution 186: 147–156[CrossRef][ISI]
Wiebes, J. T. 1979 Co-evolution of figs and their insect pollinators. Annual Review of Ecology and Systematics 10: 1–12
———. 1995 Agaonidae (Hymenoptera, Chalcidoidea) and Ficus (Moraceae): fig wasps and their figs, XV (Meso-American Pegoscapus). Proceedings of the Koninklijke Nederlandse Akademie van Wetenschappen 98: 167–183
This article has been cited by other articles:
|
|
 |
|
  J. L. Bronstein The Costs of Mutualism Integr. Comp. Biol., August 1, 2001; 41(4): 825 - 839. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
0.01, Mann-Whitney test). Means ± 1 SD are listed