Acta Oecologica xxx (2013) 1e6 Contents lists available at SciVerse ScienceDirect Acta Oecologica journal homepage: www.elsevier.com/locate/actoec Original article The effect of fig wall thickness in Ficus erecta var. beecheyana on parasitism Hsy-Yu Tzeng a, Chern-Hsiung Ou a, Fu-Yuan Lu b, Anthony Bain c, d, Lien-Siang Chou d, *, Finn Kjellberg c a Department of Forestry, National Chung-Hsing University, 250 Kuokwang Road, Taichung 40227, Taiwan Department of Forestry and Natural Resources, National Chiayi University, 300 University Rd., Chiayi 60004, Taiwan Centre d’Ecologie Fonctionnelle et Evolutive, CEFE-CNRS, UMR 5175, 1919 route de Mende, F-34293 Montpellier Cedex 5, France d Institute of Ecology and Evolutionary Biology, College of Life Sciences, National Taiwan University, 1, Sec. 4, Roosevelt Road, Taipei 10617, Taiwan b c a r t i c l e i n f o a b s t r a c t Article history: Received 29 February 2012 Accepted 26 June 2013 Available online xxx Fig wasp communities constitute a model system to analyse determinants of community complexity and to investigate how biological interaction networks are maintained. It has been suggested for monoecious figs, that fig pollinating wasps avoid ovipositing in flowers located close to the fig wall because of strong parasitic pressure by wasps ovipositing through the fig wall. This behaviour could help explain why mainly seeds are produced in flowers located close to the fig wall, thus stabilizing the fig-pollinating wasp mutualism. In this contribution we explore, for dioecious figs, whether ovipositor length of parasitic species may really be limiting. In dioecious figs, functionally male figs produce pollinating wasps and pollen while female figs produce only seeds, facilitating selection of traits favouring pollinator reproduction in male figs. We show in Ficus erecta that fig walls are thicker in male figs than in female figs. Male figs presenting thick walls, thicker than the length of the parasites’ ovipositors, went unparasitized while male figs presenting thinner walls were systematically parasitized. Hence, in F. erecta, ovipositor length of the parasites is limiting access to some figs. However, we also show that in another dioecious species, Ficus formosana, presenting thin walled male figs, no fig is protected against oviposition by its two parasites. Hence in dioecious as well as in monoecious figs, in some Ficus species, ovipositors of the parasites are limiting access to ovules, while in other Ficus species all ovules are exposed to parasitism. Ó 2013 Elsevier Masson SAS. All rights reserved. Keywords: Parasitism Ficus erecta Ficus formosana Sycoscapter Co-evolution 1. Introduction Explaining the maintenance of complex interaction networks has been a major scientific challenge. The theoretical basis of the maintenance of interaction networks including mutualism and parasitism is still beyond our reach and only simplistic models of mutualism-parasitism are available (Wilson et al., 2003). The communities of wasps associated with figs may serve as a good empirical model for disentangling species interactions and exploring possible mechanisms for stabilizing community structures. Indeed, each of the 735 species of fig trees (Ficus, Moraceae)(Berg and Corner, 2005) is pollinated by one or a few species of mutualistic host-specific agaonid wasps (Hymenoptera: Agaonidae sensu Cruaud et al., 2010; for exceptions to species specificity * Corresponding author. E-mail address: chouls@ntu.edu.tw (L.-S. Chou). see e.g. Cornille et al., 2012) and may be associated with 0e30 species of, generally specific, non-pollinating fig wasps parasitic on the system (Cook and Rasplus, 2003; for exceptions to specificity see Marussich and Machado, 2007). Further the phylogenies of both figs and fig pollinating wasps are relatively well known (Cruaud et al., 2012) and the worldwide history of non-pollinating fig wasps is progressively becoming available (Cruaud et al., 2011). Finally all these wasps emerge at the same time from figs allowing relatively easy collection of complete communities. As such the figefig wasp interaction constitutes a unique model for the comparative investigation of interactions within more or less complex communities, and for investigating the historical and ecological determinants of community structure and its long term stability. For instance, preliminary results have shown 1) frequent instances of co-speciation within the communities (Jousselin et al., 2008) despite some very clear cases of lack of specificity and recent host shifts (Marussich and Machado, 2007; Jousselin et al., 2008) and world-wide radiations that followed major host shifts (Cruaud 1146-609X/$ e see front matter Ó 2013 Elsevier Masson SAS. All rights reserved. http://dx.doi.org/10.1016/j.actao.2013.06.007 Please cite this article in press as: Tzeng, H.-Y., et al., The effect of fig wall thickness in Ficus erecta var. beecheyana on parasitism, Acta Oecologica (2013), http://dx.doi.org/10.1016/j.actao.2013.06.007 2 H.-Y. Tzeng et al. / Acta Oecologica xxx (2013) 1e6 et al., 2011, 2012), and 2) lack of community saturation (Hawkins and Compton, 1992; Compton et al., 1994). Ficus species produce closed inflorescences called figs (or syconia), the inside of which is lined by female flowers and male flowers. When the female flowers become receptive, the figs produce volatile compounds which attract the female pollinating wasps (Proffit et al., 2008). After entering through an ostiole closed by bracts, the wasps oviposit into some flowers between the inner integument and the nucellus by introducing their ovipositor through the style (Jansen-Gonzalez et al., 2012). The wasps pollinate while ovipositing. The wasp larvae develop within the uniovulate galled flowers and some weeks later adult offspring wasps emerge into the fig cavity, become loaded with pollen before leaving their natal fig in search of a new receptive one. In monoecious fig trees (approximately half of the species), each female flower may produce either a wasp or a seed. Flowers are stacked in several layers at fig receptivity and wasps preferentially oviposit into flowers located close to the fig lumen, furthest away from the outside. Flowers located closest to the fig wall are more likely to produce seeds (for a full description of the interaction and its variation, see Kjellberg et al., 2005). The stability of the mutualism in monoecious Ficus species depends on this behaviour of the wasps avoiding oviposition in a subset of flowers. Four non-exclusive hypotheses have been proposed to explain this wasp behaviour. Firstly, it was proposed that many flowers within figs had styles longer than pollinator ovipositors and thus protected them against oviposition. It was subsequently shown that in many Ficus species, pollinating wasps had ovipositors long enough to access most ovaries within a fig (Nefdt and Compton, 1996). Second, it was proposed that some flowers simply did not allow wasp production (West and Herre, 1994). Later work has shown that the assumptions on feeding regime and oviposition behaviour of the non-pollinating fig wasps on which the idea was based were wrong (Elias et al., 2012). Thirdly, when the insects become adult, young female wasps that have developed in galls located close to the cavity will be more easily mated (they are mated while still within their natal galls) and they will more easily emerge into the fig cavity before leaving the fig through a hole cut by males, resulting in higher fitness (Anstett, 2001). Finally, it has been proposed that non pollinating fig wasps parasitic on the pollinators could play an important role (Dunn et al., 2008). These wasps oviposit from outside the fig. They insert their ovipositor through the fig wall and deposit an egg in an ovule within which a wasp larva is already developing. The parasite larva develops either as a cleptoparasite consuming the gall initiated by a pollinator (or another wasp) and killing the wasp larva (Joseph, 1958), or it may develop as a true parasitoid (Tzeng et al., 2008). Data on Ficus rubiginosa show that wasps developing in flowers located close to the fig wall were mainly non pollinating fig wasps while wasps developing in flowers located close to the fig lumen were mainly pollinators. The authors assumed that the non pollinating figs wasps were parasitic on the pollinator larvae and had killed their host. Therefore they concluded that pollinator larvae located closer to the fig wall were more exposed to parasitism than larvae located closer to the fig lumen (Dunn et al., 2008). Further, measurements of length of ovipositors of cleptoparasites and/or parasitoids demonstrated that pollinator larvae located close to the lumen were out of reach from parasites (Al-Beidh et al., 2012). They suggested that this may provide a selection mechanism for pollinating wasps avoiding ovaries located close to the fig wall and as such would stabilize the mutualism between Ficus and their pollinating agaonid wasps. However, earlier data demonstrated that in another fig species, with small figs, there was no enemy-free space and levels of parasitism by cleptoparasites and parasitoids on galling wasps (pollinators and non pollinating fig wasps) where independent of flower position (Compton et al., 1994). Further, in two Ficus species with larger figs, data showed that parasite (galler, cleptoparasite and parasitoid) ovipositor lengths matched the thickness of the fig wall plus distance to the ovules at the time of oviposition. These studies suggested that there was no enemy-free space (Compton et al., 1994; Kerdelhué et al., 2000). Therefore we may question the generality of the presence of an enemy free space within figs. Is parasite ovipositor length sometimes or often limiting, resulting from a trade-off between cost of longer ovipositors and proportion of potential hosts that are accessible for a given ovipositor length? Potential costs of longer ovipositors include a penetration problem (longer ovipositors are more difficult to insert), a strain on the egg delivery system through the ovipositor, wind resistance in flight, and vulnerability to predators (Sivinski and Aljua, 2001) and indeed there are many observations of ants predating non pollinating fig wasps that could not flee because their ovipositor was inserted into the fig (Kerdelhué and Rasplus, 1996; Schatz et al., 2008; Compton et al., 2009). The functionally dioecious Ficus species may provide a simpler system to investigate whether some pollinator larvae may be protected against parasitism because the ovipositors of cleptoparasites or parasitoids are too short to reach them from outside the fig. In these species, some trees produce wasps and no seeds in their female flowers and serve as a pollen donor: they are functionally male. In functionally male trees all female flowers have very short styles through which the pollinators can easily oviposit. Other trees do not produce any wasps within their female flowers and produce seeds only: they are female. In female trees all female flowers have very long styles forbidding oviposition (for a full description see Kjellberg et al., 2005). In receptive male dioecious figs, female flowers form a single layer which can serve as a control group in the variation of ovary thickness, because the distance from the fig surface to the ovules is rather homogeneous within individual figs. This distance, however, may vary within population among figs depending on fig wall thickness. Further, the presence of distinct sexes may allow selection towards thicker, spongy, fig walls in male figs resulting in some protection against parasitism, while other traits involved inbreeding or seed dispersal may be selected in female figs (Harrison and Yamamura, 2003). Hence we may expect to observe exacerbated traits in dioecious figs that can enlighten our understanding of monoecious fig evolutionary biology. In order to explore the potential effect of fig wall thickness on parasitism prevention, we investigated differences between male and female figs in a species, Ficus erecta var. beecheyana, in which fig wall thickness differed between sexes, and we tested whether pollinating wasps in male figs presenting thicker fig walls were less parasitized. Then, to evidence that different dioecious Ficus species develop different strategies we compared this data with data on fig wall thickness and parasite ovipositor length for another dioecious Ficus species, Ficus formosana, presenting thin fig-walls. 2. Materials and methods 2.1. Study species F. erecta var. beecheyana is a dioecious shrub or treelet belonging to section Ficus subsection Frutescentiae (Berg and Corner, 2005). Figs in this species are located at the axils of the leaves or in the scars of leaves from the previous growth season. Male figs are generally borne below the leaves, at the axils of the previous year’s leaves, while female figs mostly develop at the axils of the leaves on the current year’s shoot (Tzeng et al., 2001). Such a phenology is typical of dioecious Ficus species adapted to highly seasonal climates (Kjellberg et al., 1987), such as the local climate at the Guandaushi Forest Station (Tzeng et al., 2003, 2004). Sycoscapter Please cite this article in press as: Tzeng, H.-Y., et al., The effect of fig wall thickness in Ficus erecta var. beecheyana on parasitism, Acta Oecologica (2013), http://dx.doi.org/10.1016/j.actao.2013.06.007 3 H.-Y. Tzeng et al. / Acta Oecologica xxx (2013) 1e6 inubiae is the sole non-pollinating fig wasp species associated with F. erecta and its pollinator Blastophaga nipponica in Taiwan (Wu, 1996; Tzeng et al., 2006). We have never observed this Sycoscapter species in figs that did not contain pollinators, and it oviposited in figs after pollination: it is hence a cleptoparasite or a parasitoid of the pollinator. F. formosana, a small shrub, is a close relative of F. erecta and also belongs to subsection Frutescentiae. In places where F. formosana co-occurs with F. erecta, it is parasitized by two species of Sycoscapter. One is most probably Sycoscapter inubiae, the same wasp observed on F. erecta, and the other is a Sycocapter species only recorded from F. formosana (Tzeng et al., 2008). The Sycoscapter species developing in F. formosana figs are the sole non pollinating fig wasps for which direct evidence is available showing them to be true parasitoids of the pollinator (Tzeng et al., 2008), thus confirming the parasitoid status of Sycoscapter inubiae. Because male figs do not produce seeds, these wasps can not be facultative seed eaters. 2.2. Study site The study was conducted at the Guandaushi Forest Station of the Hue-Sun Experimental Forest Station (24 40 N, 12 800 E), one of the Long-Term Ecological Research sites at altitudes of 500e800 m in central Taiwan. Mean annual rainfall, relative humidity, and yearly temperature during the study period were 2597 mm, 79%, and 21  C, respectively (Tzeng et al., 2003). Most of the rainfall (96%) occurred during the rainy and typhoon seasons (mainly May to September). 2.3. Methods A regular survey of the phenology of F. erecta var. beecheyana was carried out from October 1995 to February 1997. Two-three branches on 17 mature male trees and 13 mature female trees were followed at 5e9 day intervals (Tzeng et al., 2003, 2006). Figs were collected in 1996 during the main fruiting periods of both male (February to June) and female (April to July) trees. Ripe female figs (27) and male figs (28) from which wasps were about to emerge were collected and preserved in formalin-acetic acid fixative solution (FAA fixative solution). In addition, 62 post receptive male figs, at the stage of when parasitoids oviposit, were collected to establish whether fig wall thickness varied between the time at which parasitoid wasps oviposited and the time of offspring wasp emergence from figs. Figs were collected from the 17 male trees and 13 female trees. Fig diameters and wall-thicknesses were measured to the nearest millimetre. Each fig was then cut into four equal parts. Of these, one was randomly picked, and the number of florets and galls/seeds in it were counted. The numbers were then multiplied by four to obtain an estimate of the total numbers of florets and galls/seeds of the given fig. All fig wasps were extracted from the figs, identified, and counted. Total ovipositor length was measured for a subset of wasps of Sycoscapter inubiae emerging from F. erecta figs by extracting the ovipositor from the body, in addition to measuring ovipositor sheath length (n ¼ 37, from 3 figs). Data was analysed using SPSS Statistics 11.0. Pearson correlation analyses were used to assess the relationships between diameter, wall-thickness, and the number of florets. Because male figs of F. formosana presented obviously thinner fig walls than male figs of F. erecta var. beecheyana, we also measured fig wall thickness on that species, and compared it with ovipositor lengths of its two parasitic Sycoscapter species. Thirty four figs were collected from 1996 to 1999 at the same study site from a group of 5 male shrubs interspersed with the F. erecta var. beecheyana treelets (Tzeng et al., 2008). Ovipositor sheath length was measured for 250 Sycoscapter inubiae and Sycoscapter sp. emerging from 10 F. formosana figs. The aim was to show that fig wall thickness can hardly be seen as an obstacle to parasite oviposition in that species. 3. Results 3.1. Ficus erecta Twenty-eight male figs at wasp emergence and 27 ripe female figs were collected. The morphological characteristics of these male and female figs are listed in Table 1. Male figs were 1.7 times larger than female figs (t ¼ 15.147, p < 0.001). Male figs also contained three times more flowers (Table 1). Male flowers represented 31.4  9.1% of all flowers in male figs, a figure very similar to the 32.5  5.9% neuter flowers (sterile male flowers) observed in female figs (Table 1). In male figs, 36% of female flowers produced insects against 69% producing seeds in female figs (t ¼ 5.992, p < 0.001). Beyond the sexual differences in fig size and flower composition, we also noted a difference in traits correlated with fig size. The size of female figs correlated strongly and significantly with the total number of flowers, and most strongly with the number of female flowers (r2 ¼ 0.612, p < 0.001, r2 ¼ 0.679, p < 0.001, n ¼ 27), but not with the number of seeds (r2 ¼ 0.372, p ¼ 0.061, n ¼ 27). In contrast, the size of male figs was not significantly correlated with the total number of flowers, female flowers, nor with the number of galls (r2 ¼ 0.266, p ¼ 0.171, r2 ¼ 0.194, p ¼ 0.324, and r2 ¼ 0. 074, p ¼ 0.707, respectively, n ¼ 28). Fig wall-thickness differed between sexes. At the time of wasp emergence, the fig-walls of male figs were 1.5 time thicker than the walls of female figs (t ¼ 5.497, p < 0.001). Furthermore, fig size was best explained by fig wall-thickness in male figs (r2 ¼ 0.701, p < 0.001) and by the number of female flowers in female figs (r2 ¼ 0.679, p < 0.001). Although we were interested in the correlation between parasitism by Sycoscapter inubiae and fig wall-thickness of figs at the time of parasitoid oviposition, the wall thickness of figs for which we quantified parasitism could only be measured much later, at the time of wasp emergence. However, wall thickness did not vary between post receptive figs and figs at the time of offspring wasp emergence (mean 4.31  0.72 mm and range 2.6e6.1 mm for post receptive figs, n ¼ 62; mean 4.19  1.27 mm and range 2.5e7.0 mm for figs at the time of wasp emergence, n ¼ 28, t ¼ 0.535, p ¼ 0.594). Inside the figs of F. erecta var. beecheyana, only one species of non pollinating fig wasp, the parasitoid Sycoscapter inubiae, was found besides the pollinator, Blastophaga nipponica. Nine figs (32%) hosted only pollinators and no parasitoid while 19 figs (68%) hosted 1to 385 parasitoids, which represented17.6  15.2% (n ¼ 19) of the total number of wasps found on average in a parasitized fig. The number of parasitoids increased significantly with total wasp Table 1 Fig characteristics for F. erecta var. beecheyana at Guandaushi Forest Station, Taiwan. Male figs contain female flowers in which wasps develop and no seeds are produced. Female figs contain sterile male flowers. Male figs are larger, contain more flowers and have thicker walls. Size (diameter  height) Wall-thickness Female flowers Undeveloped female flowers Bladders (empty galls) Healthy galls/seeds Male flowers Total flowers Male figs (n ¼ 28) Female figs (n ¼ 27) 24.9  3.0  26.2  3.0 4.2  1.2 934  303 524  231 63  61 347  215 428  137 1362  362 14.9  1.8  15.4  1.4 2.7  0.6 288  68 88  69 203  90 135  20 423  74 Please cite this article in press as: Tzeng, H.-Y., et al., The effect of fig wall thickness in Ficus erecta var. beecheyana on parasitism, Acta Oecologica (2013), http://dx.doi.org/10.1016/j.actao.2013.06.007 4 H.-Y. Tzeng et al. / Acta Oecologica xxx (2013) 1e6 500 60 50 400 Proportion parasites (%) Sycoscapter number a 300 200 100 40 30 20 10 0 0 200 400 600 800 1000 0 2 3 60 5 6 7 8 Fig wall thickness (mm) b Proportion parasites (%) 4 Fig. 2. Relationship between fig wall-thickness and the proportion of parasites among offspring wasps. A threshold effect is apparent. 50 40 between the two groups of wall-thickness, globally significantly longer than the wall-thickness of all parasitized figs, but shorter than those of almost all unparasitized figs. As predicted from the correlation between fig-size and fig-wall thickness, unparasitized male figs (27.56  1.97 mm) were larger than parasitized ones (23.56  2.54 mm; t ¼ 5.942, p < 0.001). 30 20 10 3.2. Ficus formosana 0 0 200 400 600 800 1000 Total number of fig wasps Fig. 1. Relationship between the total number of fig wasps produced by a fig (including pollinators and non-pollinator wasps) and (a) the number of Sycoscapter inubiae, (b) the proportion of S. inubiae. The proportion of parasites is independent of total number of wasps. production (r2 ¼ 0.664, p < 0.001, n ¼ 19, Fig. 1a), while the proportion of parasites produced did not vary with total wasp production (Fig. 1b). The average ovipositor length of parasitoids was 5.24  0.49 (n ¼ 37). It was significantly longer than average male fig wall thickness (t ¼ 6.159, p < 0.001 for post receptive figs; t ¼ 3.873, p < 0.001, for figs at wasp emergence stage). However, the variance of parasitoid ovipositor length was significantly smaller than that of male fig wall thicknesses (F1,98 ¼ 9.257, p ¼ 0.003; F1,64 ¼ 20.564, p < 0.001; for post receptive and wasp emergence stage figs respectively). In addition, the proportion of parasitoids produced was negatively correlated with fig size and fig wall-thickness, and the correlation with fig wall thickness was strongest (correlation analysis, r2 ¼ 0.335, df ¼ 26, p < 0.001; r2 ¼ 0.416, df ¼ 26, p < 0.001, respectively). To further assess the effect of fig-wall thickness, we separated the male figs into two groups based on whether they were parasitized. Unparasitized figs presented thicker walls than parasitized ones (5.68  0.83 mm versus3.48  0.70 mm, t ¼ 7.356, p < 0.001). There was a marked threshold effect (Fig. 2). Indeed out of 19 parasitized figs only one had a wall thicker than 4.5 mm while out of 9 unparasitized figs, none had a wall less than 4.5 mm thick. The pattern observed with fig size instead of fig wall thickness did not present such a marked threshold. Average ovipositor length of the parasitoids was 5.24 mm and presented an intermediate distribution The fig walls of F. formosana male figs are much thinner (0.75  0.11 mm, range 0.5e0.93 mm, 36 figs from 5 trees)than those of F. erecta var. beechayana. Two parasitoids of F. formosana were observed abundantly on the study site, Sycoscapter sp. and Sycoscapter inubiae. The ovipositor sheath of Sycoscapter inubiae in F. erecta (length 4.56  0.61 mm, range 3.0e5.5, n ¼ 37) was longer than that of Sycoscapter sp. (1.82  0.19 mm, range 1.0e2.5, n ¼ 87) but was similar to that of Sycoscapter inubiae in F. formosana (4.60  0.61 mm, range 1.0e2.5, n ¼ 163). Hence the size of Sycoscapter inubiae seems to have been almost independent of host plant despite large differences in fig shape and size. In addition, the ovipositor sheath lengths of both Sycoscapter parasitic on F. formosana were significantly longer than fig-wall thickness (t ¼ 38.761 for Sycoscapter sp. and t ¼ 74.071 Sycoscapter inubiae, p < 0.001). The variances in the ovipositor length of both Sycoscapter species were larger than that of the fig wall thickness of F. formosana (Fig. 3). 4. Discussion In this contribution we addressed a simple question: can there be an enemy free space for developing pollinator larvae within figs, i.e. can ovipositor length be too short for parasitoid or cleptoparasitic non pollinating fig wasps to access all ovules occupied by pollinator larvae? The results presented by Dunn et al. (2008) suggested that this may sometimes be the case in monoecious figs. We demonstrate here formally that ovipositor length is limiting in Sycoscapter inubiae, a parasitoid of the pollinator of the dioecious F. erecta var. beecheyana. Indeed, in figs presenting thick fig walls, thicker than the length of the wasps’ ovipositors, there was no parasitism with a sharp transition in wall-thickness between parasitized and non-parasitized figs. While average ovipositor length was similar to average wall thickness, higher variance in Please cite this article in press as: Tzeng, H.-Y., et al., The effect of fig wall thickness in Ficus erecta var. beecheyana on parasitism, Acta Oecologica (2013), http://dx.doi.org/10.1016/j.actao.2013.06.007 H.-Y. Tzeng et al. / Acta Oecologica xxx (2013) 1e6 20 Wall thickness at parasitoid oviposition time Ovipositor sheath of Sycoscapter inubiae a Number 15 10 5 0 0 1 2 3 4 5 6 7 8 100 Wall thickness at parasitoid oviposition time Ovipositor sheath of Sycoscapter sp. Ovipositor sheath of Sycoscapter inubiae b Nnumber 80 60 40 20 0 0 1 2 3 4 5 6 7 8 Length/fig wall thickness (mm) Fig. 3. Relationship between the ovipositor sheath lengths of parasitoid fig wasps and fig wall-thickness at the stage of parasite ovipositon for (a) F. erecta var. beecheyana and (b) F. formosana. In F. formosana, no fig is protected against parasitoid oviposition by fig wall thickness. In F. erecta some figs are protected against oviposition by parasitoids by fig wall thickness. fig-wall thickness resulted in some figs presenting walls thicker than ovipositor length. To our knowledge this is the second direct demonstration that ovipositor length limits access to ovaries for a non pollinating fig wasp and it is the first for a dioecious fig (AlBeidh et al., 2012). We have qualitatively observed the same phenomenon in the spring crop of male Ficus carica. Fig trees with thicker fig walls presented no or almost no parasitism by Philotrypesis caricae (F/K, unpublished personal observations). The difference in fig wall thickness between male and female figs and the observation that thick walls limit parasitism may suggest that fig wall thickness of male figs in F. erecta and F. carica is an evolutionary response to parasitism. Indeed no alternative explanation has been suggested for male figs presenting larger fig walls than female figs (Harrison and Yamamura, 2003). Spongy fig walls reminiscent of those observed here in male figs have been reported in the monoecious Ficus cyathistipula, and it has been suggested to be a trait related to fig dispersal by water, hence a trait related to seed dispersal (Berg and Wiebes, 1992). Analysing genetic variation within population for fig wall thickness in male figs is non-trivial. Indeed we have observed variation of fig size and shape over the years within individual trees, depending on the development of a main trunk and the presence of more or less numerous young vigorous shoots stemming from the base of the tree. 5 The hypothesis of an enemy-free internal space within the fig in ovules located far from the fig wall was at the center of the proposition by Dunn et al. (2008) on how parasites could stabilize the mutualism between figs and pollinating fig-wasps. Our results confirm that ovipositor length may be limiting for parasitoids of pollinating wasps in some dioecious figs, in that case due to fig wall thickness. Nevertheless this can not be generalized to all dioecious figs. Indeed, the data on F. formosana reveals that the fig wall in that species is thin, and the ovipositors of both parasitoids of its pollinator are always longer than the thickness of the fig wall. This is similar to the situation documented for monoecious figs as ovipositor length of non pollinating wasps parasitic on the pollinators seemed to be leaving an enemy free space in F. rubiginosa (Dunn et al., 2008) but did not in Ficus burtt-davyi (Compton et al., 1994). One can only speculate on the relative frequency of limiting ovipositor lengths in both monoecious and dioecious figs. Harrison and Yamamura (2003) provided some data on the relative sizes of male and female figs in terms of diameter at maturity but also in terms of numbers of flowers. No clear trend was visible in terms of which sex had largest figs and which sex had more female flowers per fig, suggesting a diversity of evolutionary answers to the different selective pressures exerted on male and female figs. Constraints or lack of constraints on the capacity of fig wasps to evolve a response to fig wall thickness and their consequences on the mutualism between figs and their pollinators have often been suggested or discussed (e.g. Bronstein, 1991; Zhen et al., 2005). Our results suggest that a long ovipositor is costly for Sycoscapter inubiae. Indeed, in our sample, one third of the figs were protected against parasitism because the ovipositors of S. inubiae were not long enough. Wasps with longer ovipositors would have had access to many more figs to parasitize and hence, there was probably strong selection for longer ovipositors. The lack of evolutionary response implies another selective force is counter-selecting long ovipositors. Indeed it has been proposed that long ovipositors may affect flight prowess, increase handling time during insertion, and may increase vulnerability to predators (Sivinski and Aluja, 2001). Data on F. formosana further supports the idea that ovipositor size is tuned to fig size. In F. formosana the fig wall is thin. The specialist parasitoid Sycoscapter sp. has a short ovipositor. On the other hand, the non-specialist, Sycoscapter inubiae, has a much longer ovipositor. Nevertheless, both species have been observed ovipositing at the same time (Tzeng et al., 2008). 5. Conclusion The results obtained here suggest that in some species of Ficus, both monoecious and dioecious, ovipositor length of nonpollinating fig was parasitic on pollinators seems to be the product of an evolutionary compromise between selection for longer ovipositor to increase the number of accessible pollinator larvae and the cost of long ovipositors. Such trade-offs could stabilize some figefig wasp systems. However in other Ficus species, ovipositor lengths of non-pollinating fig wasps parasitic on pollinators are not limiting and others processes have to be invoked to explain the stability of the system. Acknowledgement We would like to thank the National Science Council of the Republic of China, Taiwan for financially supporting this study under Contract No. NSC 87-2613-B-005-085 A07 and No. NSC 992923-B-002-001-MY3, and grant ANR 09 BLAN-03392-CFD7. In addition, we are grateful to the officials at Hue-Sun Forest Station for their logistics support during field work and providing research Please cite this article in press as: Tzeng, H.-Y., et al., The effect of fig wall thickness in Ficus erecta var. beecheyana on parasitism, Acta Oecologica (2013), http://dx.doi.org/10.1016/j.actao.2013.06.007 6 H.-Y. Tzeng et al. / Acta Oecologica xxx (2013) 1e6 permits. Finally, we extend our thanks to Dr. Liang-Yi Chou for identifying the fig wasp species. References Al-Beidh, S., Dunn, D.W., Sally, A., Power, S.A., Cook, J.M., 2012. Parasites and mutualism function: measuring enemy-free space in a fig-pollinator symbiosis. Oikos 121, 1833e1839. Anstett, M.C., 2001. Unbeatable strategy, constraint and coevolution, or how to resolve evolutionary conflicts: the case of the fig/wasp mutualism. Oikos 95, 476e484. Berg, C.C., Corner, E.J.H., 2005. Moraceae (Ficus). In: Nooteboom, H.P. (Ed.), Flora Malesiana. National Herbarium Nederland, Leiden. Berg, C.C., Wiebes, J.T., 1992. African Fig Trees and Fig Wasps. Koninklijke Nederlandse Akademie van Wetenschappen, Amsterdam, pp. 1e298. Bronstein, J.L., 1991. The nonpollinating wasp fauna of Ficus pertusa: exploitation of a mutualism? Oikos 61, 175e186. Compton, S.G., van Noort, S., McLeish, M., Deeble, M., Stone, V., 2009. Sneaky African fig wasps that oviposit through holes drilled by other species. African Natural History 5, 9e15. Compton, S.G., Rasplus, J.Y., Ware, A.B., 1994. African fig wasp parasitoid communities. In: Hawkins, B., Sheehan, W. (Eds.), Parasitoid Community Ecology. Oxford University Press, pp. 343e368. Cook, J.M., Rasplus, J.Y., 2003. Mutualists with an attitude: coevolving fig wasps and figs. Tree 18, 241e248. Cornille, A., Underhill, J.G., Cruaud, A., Hossaert-McKey, M., Johnson, S.D., Tolley, K.A., Kjellberg, F., van Noort, S., Proffit, M., 2012. Floral volatiles, pollinator sharing and diversification in the fig-wasp mutualism: insights from Ficus natalensis, and its two wasp pollinators (South Africa). Proceedings of the Royal Society B 279, 1731e1739. Cruaud, A., Jabbour-Zahab, R., Genson, G., Cruaud, C., Couloux, A., Kjellberg, F., van Noort, S., Rasplus, J.Y., 2010. Laying the foundations for a new classification of Agaonidae (Hymenoptera: Chalcidoidea), a multilocus phylogenetic approach. Cladistics 26, 359e387. Cruaud, A., Jabbour-Zahab, R., Genson, G., Kjellberg, F., Kobmoo, N., van Noort, S., Da-Rong, Y., Peng, Y.Q., Ubaidillah, R., Hanson, P.E., Santos-Mattos, O., Farache, F.H.A., Pereira, R.A.S., Kerdelhué, C., Rasplus, J.Y., 2011. Phylogeny and evolution of life-history strategies in the Sycophaginae nonpollinating fig wasps (Hymenoptera, Chalcidoidea). BMC Evolutionary Biology 11, 178. Cruaud, A., Rønsted, N., Chantarasuwan, B., Chou, L.S., Clement, W.L., Couloux, A., Cousins, B., Genson, G., Harrison, R.D., Hanson, P.E., Hossaert-McKey, M., Jabbourzahab, R., Jousselin, E., Kerdelhué, C., Kjellberg, F., Lopez-Vaamonde, C., Peebles, J., Peng, Y.Q., Pereira, R.A.S., Schramm, T., Ubaidillah, R., Van Noort, S., Weiblen, W.D., Yang, D.R., Yodpinyanee, A., Libeskind-Hadas, R., Cook, J.M., Rasplus, J.Y., Savolainen, V., 2012. An extreme case of plant-insect co-diversification: figs and fig-pollinating wasps. Systematic Biology 61, 1029e1047. Dunn, D.W., Segar, S.T., Ridley, J., Chan, R., Crozier, R.H., Yu, D.W., Cook, J.M., 2008. A role for parasites in stabilising the fig-pollinator mutualism. PLoS Biology 6, 1e7. Elias, L.G., Teixeira, S.P., Kjellberg, F., Pereira, R.A.S., 2012. Diversification in the use of resources by Idarnes species: bypassing functional constraints in the fig-fig wasp interaction. Biological Journal of the Linnean Society 106, 114e122. Harrison, R.D., Yamamura, N., 2003. A few more hypotheses for the evolution of dioecy in figs (Ficus, Moraceae). Oikos 100, 628e635. Hawkins, B.A., Compton, S.G., 1992. African fig wasp communities-undersaturation and latitudinal gradients in species richness. Journal of Animal Ecology 61, 361e372. Jansen-Gonzalez, S., Teixeira, S.P., Pereira, R.A.S., 2012. Mutualism from the inside: coordinated development of plant and insect in an active pollinating fig wasp. Arthropod-plant Interactions 6, 601e609. Joseph, J.K., 1958. Recherches sur les chalcidiens, Blastophaga psenes (L.) et Philotrypesis caricae (L.) du figuier (Ficus carica L. Annales Des Sciences Naturelles Zoologie et Biologie Animale 10, 197e260. Jousselin, E., van Noort, S., Rasplus, J.Y., Rønsted, N., Erasmus, C., Greeff, J., 2008. One tree to bind them all: host conservatism in a fig wasp community unraveled by cospeciation analyses among pollinating and nonpollinating fig wasps. Evolution 62, 1777e1797. Kerdelhué, C., Rasplus, J.Y., 1996. The evolution of dioecy among Ficus (Moraceae): an alternative hypothesis involving non-pollinating fig wasp pressure on the fig-pollinator mutualism. Oikos 77, 163e166. Kerdelhué, C., Rossi, J.-P., Rasplus, J.Y., 2000. Comparative community ecology studies on old world figs and fig wasps. Ecology 81, 2832e2849. Kjellberg, F., Jousselin, E., Hossaert-McKey, M., Rasplus, J.Y., 2005. Biology, ecology, and evolution of fig-pollinating wasps (Chalcidoidea, Agaonidae). In: Raman, A., Schaefer, W., Withers, T.M. (Eds.), Biology, Ecology and Evolution of Gallinducing Arthropods. Science Publishers, Inc., Enfield (NH) USA Plymouth (UK), pp. 539e572. Kjellberg, F., Gouyon, P.H., Ibrahim, M., Raymond, M., Valdeyron, G., 1987. The stability of the symbiosis between dioecious figs and their pollinators: a study of Ficus carica L. Evolution 41, 693e704. Marussich, W.A., Machado, C.A., 2007. Host-specificity and coevolution among pollinating and nonpollinating New World fig wasps. Molecular Ecology 16, 1925e1946. Nefdt, R.J.C., Compton, S.G., 1996. Regulation of seed and pollinator production in the figefig wasp mutualism. The Journal of Animal Ecology 65, 170e182. Proffit, M., Schatz, B., Bessière, J.-M., Chen, C., Soler, C., Hossaert-McKey, M., 2008. Signalling receptivity: comparison of the emission of volatile compounds by figs of Ficus hispida before, during and after the phase of receptivity to pollinators. Symbiosis 45, 15e24. Schatz, B., Kjellberg, F., Nyawa, S., Hossaert-McKey, M., 2008. Fig wasps: a staple food for ants on Ficus. Biotropica 40, 190e195. Sivinski, J., Aluja, M., 2001. The evolution of ovipositor length in the parasitic hymenoptera and the search for predictability in biological control. Florida Entomologist 86, 143e150. Tzeng, H.Y., Ou, C.H., Lu, F.Y., 2001. Morphological study on the syconia of Ficus erecta var. beecheyana. Taiwan Journal of Forest Science 16, 295e306 (in Chinese with English summary). Tzeng, H.Y., Ou, C.H., Lu, F.Y., 2003. Syconium phenology of Ficus erecta var. beecheyana at the Hue-Sun Forest Station. Taiwan Journal of Forest Science 18, 273e282 (in Chinese with English summary). Tzeng, H.Y., Ou, C.H., Lu, F.Y., 2004. Phenology and sexual differentiation of Ficus erecta var. beecheyana at Guandaushi Forest Station, Taiwan. Quarterly Journal of Forest Research 26, 61e78 (in Chinese with English summary). Tzeng, H.Y., Lu, F.Y., Ou, C.H., Lu, K.C., Tseng, L.J., 2006. Pollinational-mutualism strategy of Ficus erecta var. beecheyana and Blastophaga nipponica in seasonal Guandaushi Forest Ecosystem, Taiwan. Botanical Studies 47, 307e318. Tzeng, H.Y., Tseng, L.J., Ou, C.H., Lu, K.C., Lu, F.Y., Chou, L.S., 2008. The confirmation of the parasitoid feeding habit in Sycoscapter, and their impact on pollinator abundance in Ficus formosana. Symbiosis 45, 129e134. West, S.A., Herre, E.A., 1994. The ecology of the New World fig-parasitizing wasps Idarnes and implications for the evolution of the fig-pollinator mutualism. Proceedings of the Royal Society B 258, 67e72. Wilson, W.G., Morris, W.F., Bronstein, J.L., 2003. Coexistence of mutualists and exploiters on spatial landscapes. Ecological Monographs 73, 397e413. Wu, H.F., 1996. The Symbiosis between Ficus erecta Thunb. var. beecheyana and Blastophaga nipponica at Yang-Ming Shan. MS Thesis. Graduate Institute of Zoology, National Taiwan University, Taipei, Taiwan, p. 54 (in Chinese with English summary). Zhen, W.Q., Huang, D.W., Xiao, J.H., Yang, D.R., Zhu, C.D., Xiao, H., 2005. Ovipositor length of three Apocrypta species: effect on oviposition behavior and correlation with syconial thickness. Phytoparasitica 33, 113e120. Please cite this article in press as: Tzeng, H.-Y., et al., The effect of fig wall thickness in Ficus erecta var. beecheyana on parasitism, Acta Oecologica (2013), http://dx.doi.org/10.1016/j.actao.2013.06.007