INTRODUCTION
Seeds are the principal means of reproduction and dispersal for most vascular plants (Vander Wall et al. Reference VANDER WALL, FORGET, LAMBERT, HULME, Forget, Lambert, Hulme and Vander Wall2005), but seeds are also the stage when the plants are susceptible to heavy predation by animals (Janzen Reference JANZEN1971). Therefore, seed predators can affect the population dynamics of individual plant species, and ultimately, diversity and species composition of plant communities (Lewis & Gripenberg Reference LEWIS and GRIPENBERG2008).
Many plant species are known to synchronize their flowering or fruiting phenology and vary the amount of crop year to year (van Schaik et al. Reference VAN SCHAIK, TERBORGH and WRIGHT1993). The intermittent production of large seed crops by a population of plants is called mast seeding or mast fruiting (Kelly Reference KELLY1994). Because of the variations in availability, seeds might be unreliable resources to the predators.
The Dipterocarpaceae, the dominant family in lowland rain forests of South-East Asia, present one of the extreme examples of mast fruiting; they flower synchronously among many species at irregular intervals from 2 to 10 y while few trees flower in years between the events (Appanah Reference APPANAH1985, Ashton et al. Reference ASHTON, GIVNISH and APPANAH1988, Numata et al. Reference NUMATA, YASUDA, OKUDA, KACHI and NOOR2003, Sakai Reference SAKAI2002). Several hypotheses both in terms of the proximate and the ultimate cause have been proposed to explain the mechanisms of general flowering. The proximate cause is the environmental cues that trigger flowering. An increase in sunshine hours (Ng Reference NG1977), a drop in night-time temperature (Ashton et al. Reference ASHTON, GIVNISH and APPANAH1988, Yasuda et al. Reference YASUDA, MATSUMOTO, OSADA, ICHIKAWA, KACHI, TANI, OKUDA, FURUKAWA, RAHIM NIK and MANOKARAN1999), or prolonged droughts (Brearley et al. Reference BREARLEY, PROCTOR, SURIANTANA, DALRYMPLE and VOYSEY2007, Medway Reference MEDWAY1972, Sakai et al. Reference SAKAI, HARRISON, MOMOSE, KURAJI, NAGAMASU, YASUNARI, CHONG and NAKASHIZUKA2006) have been proposed. The ultimate cause is the adaptive significance for the evolution of synchronous flowering and/or fruiting, e.g. seed predator satiation (Janzen Reference JANZEN1974), pollination enhancement (Sakai et al. Reference SAKAI, MOMOSE, YUMOTO, NAGAMITSU, NAGAMASU, HAMID and NAKASHIZUKA1999), resource matching (Kelly Reference KELLY1994), and environmental prediction (Kelly Reference KELLY1994, Williamson & Ickes Reference WILLIAMSON and ICKES2002). Among them, the most plausible and well-known explanation is seed predator satiation (Curran et al. Reference CURRAN, CANIAGO, PAOLI, ASTIANTI, KUNSNETI, LEIGHTON, NIRARITA and HAERUMAN1999). This hypothesis suggests that the mast fruiting of the Dipterocarpaceae is an adaptive strategy for starving their seed predators during non-fruiting periods, while satiating them during synchronous mast-fruiting events.
During the fruiting event, huge amounts of seed are initiated, however, most seeds die prior to full maturation or germination. Aside from the heavy abortion soon after flowering, the main mortality factor is predation by insects or vertebrates and infestation by fungi (Curran & Leighton Reference CURRAN and LEIGHTON2000, Curran & Webb Reference CURRAN and WEBB2000, Maycock et al. Reference MAYCOCK, THEWLIS, GHAZOUL, NILUS and BURSLEM2005, Sun et al. Reference SUN, CHEN, HUBBELL, WRIGHT and NOOR2007). Among them, pre-dispersal seed predation by insects is common (Maycock et al. Reference MAYCOCK, THEWLIS, GHAZOUL, NILUS and BURSLEM2005, Naito Reference NAITO2008, Sun et al. Reference SUN, CHEN, HUBBELL, WRIGHT and NOOR2007) and has a negative effect on viable seed production (Nakagawa et al. Reference NAKAGAWA, TAKEUCHI, TANAKA and NAKASHIZUKA2005).
Most insect seed predators of dipterocarps are weevils (e.g. Curculionidae or Nanophyidae: Coleoptera), moths (e.g. Tortricidae or Pyralidae: Lepidoptera), and barkbeetles (Scolytidae: Coleoptera) (Chey Reference CHEY2002, Daljeet-Singh Reference DALJEET-SINGH1974, Nakagawa et al. Reference NAKAGAWA, ITIOKA, MOMOSE, YUMOTO, KOMAI, MORIMOTO, JORDAL, KATO, KALIANG, HAMID, INOUE and NAKASHIZUKA2003, Toy Reference TOY1988). Weevils and moths oviposit in the seeds and on fruits, respectively, when fruits are still on the tree. Normally a single larva grows up in the seed and feeds on its cotyledons, thereby killing the seed (Daljeet-Singh Reference DALJEET-SINGH1974, Toy Reference TOY1988). Bark beetles attack fruit on the ground, though some possibly attack fruit on the tree, and lay a number of eggs in a seed. Hatched larvae, as well as adult beetles, tunnel into the seeds and feed voraciously (Chan Reference CHAN1977, Daljeet-Singh Reference DALJEET-SINGH1974, Momose et al. Reference MOMOSE, NAGAMITSU and INOUE1996).
Although insect seed predators are often considered to be highly host-specific, especially in the tropics (Janzen Reference JANZEN1980), dipterocarp predators are known to feed on multiple hosts belonging to different genera or even other families (Lyal & Curran Reference LYAL and CURRAN2000, Reference LYAL and CURRAN2003; Nakagawa et al. Reference NAKAGAWA, ITIOKA, MOMOSE, YUMOTO, KOMAI, MORIMOTO, JORDAL, KATO, KALIANG, HAMID, INOUE and NAKASHIZUKA2003, Robinson et al. Reference ROBINSON, ACKERY, KITCHING, BECALLONI and HERNÁNDEZ2001). However, most previous studies were limited to nearly or already mature fruits though some studies implied the possibility of seed predator succession during seed development of dipterocarps (Kokubo Reference KOKUBO1987, Nakagawa et al. Reference NAKAGAWA, TAKEUCHI, TANAKA and NAKASHIZUKA2005, Toy Reference TOY1988). The abortion of damaged fruits occurs before maturation in many plant species (Stephenson Reference STEPHENSON1981), thus sequential sampling throughout seed development is definitely needed. The host range of seed predators is a key to understanding their likely impact on tropical plant diversity (Lewis & Gripenberg Reference LEWIS and GRIPENBERG2008). Particularly, in the case of dipterocarp predators, predator satiation would work in the synchronized fruiting by congeners if they share important predators (Silvertown Reference SILVERTOWN1980).
In the present study, we examine the composition of pre-dispersal seed-predator insects on 11 Shorea species (Dipterocarpaceae) and some dipterocarp species in other genera in Peninsular Malaysia by sequential seed sampling to test the following hypotheses: (1) a succession of predator species occurs during seed development; (2) dipterocarp species share the same seed predators.
MATERIALS AND METHODS
Study site and tree species examined
The field work was conducted at the permanent 50-ha plot in Pasoh Forest Reserve, Negeri Sembilan, Peninsular Malaysia (2°59′N, 102°18′E and 75–150 m asl). The reserve covers 2450 ha of lowland dipterocarp forest, which is dominated by species of Shorea and Dipterocarpus. Of the 13 Shorea spp. known in the plot, 11 species belonging to six sections were examined (Table 1). Although two relatively rare species – S. ochrophloia Strugnell ex Sym. and S. guiso (Blanco) Bl. – were not included, examined species represented the Shorea community in Pasoh well in terms of abundance, comprising 97.5% of total stems of Shorea spp. that exceeded 10 cm in dbh in the plot (Kochummen Reference KOCHUMMEN1997). Three common Dipterocarpus spp. and Neobalanocarpus heimii (King) Ashton (Dipterocarpaceae) were also examined to compare the species composition of seed predators with those of Shorea.
Table 1. Tree species and the number of trees and seeds examined, and the maximum weekly mean weight of fresh fruits (without wings). No weight is given for the species that did not bear mature fruit.
Two consecutive mass-flowering events were used for sampling of seed predators. One occurred from late August 2001 to February 2002 (F-01), and the other occurred from April 2002 to September 2002 (F-02). Since each event had a clear flowering peak and few dipterocarps flowered for several weeks between two events, we treat these events separately in the present study. Detailed information on these flowering events is available in Numata et al. (Reference NUMATA, YASUDA, OKUDA, KACHI and NOOR2003) and Sun et al. (Reference SUN, CHEN, HUBBELL, WRIGHT and NOOR2007).
Sequential fruit sampling and insect rearing
The fruits used for rearing insects were collected from the ground by hand. Newly fallen fruits were randomly sampled at up to 50 fruits from under each target tree every week during the fruiting events: for 16 wk from November 2001 to February 2002 in F-01 and for 21 wk from April 2002 to September 2002 in F-02. Soon after collection, 50 fruits per tree had their wings cut off, their fresh weight measured and were then stored in 10 plastic boxes (≤5 fruits per box). The insect rearing was conducted for at least 4 mo with appropriate moisture at the laboratory near Pasoh. Emerging adult insects were collected and identified. The host species and the date of seed sampling and insect collecting were recorded together. Our methodology inevitably missed the predators leaving fruits before the fruits dropped. However, the fruits which had an exit hole when sampled were few, suggesting that the majority of insect predators emerged after dispersal.
Comparison in predator frequency between immature and mature seeds
The seed-fall period was divided into three phases based on the relative weight of dropped fruits. The first phase is defined as the period when the mean weight of sampled fruits in a week was 0–15% of its maximum and named ‘aborted phase’. Similarly, the second phase and the third phase is the period when the relative fruit weight was 15–60% and 60–100%, respectively. Since fruits were still immature with bright red or light green wings at the second phase while nearly mature to mature with ochre or brown wings at the third phase, we named the second phase ‘immature phase’ and the third phase ‘mature phase’.
The emergence frequency of insects (the number of individuals divided by the number of seeds sampled in each phase) was compared between immature and mature phase by Fisher's exact test for four and five host species with sufficient number of sampled seeds in F-01 and in F-02, respectively. No comparison was made with the aborted phase, since few adult insects emerged at that phase (Figure 2).
Measurement of host specificity level for predators
The level of host specificity of predator species with more than one individual obtained was measured by the Kullback–Leibler distance (or Kullback–Leibler divergence: di) derived from Shannon entropy (Blüthgen et al. Reference BLÜTHGEN, MENZEL and BLÜTHGEN2006), which is denoted as
where, aij is the number of insect i from host j. The index considers not only the diversity of hosts but also respective host availability (e.g. the number of sample seeds) (Blüthgen et al. Reference BLÜTHGEN, MENZEL and BLÜTHGEN2006), and is standardized as
where, d min and d max is the theoretical minimum and maximum of di. This standardized Kullback–Leibler distance (d′i) ranges from 0 for the most generalized to 1.0 for the most specialized case. Calculation of the index was performed by online Monte Carlo statistics on R × C matrices (http://itb.biologie.hu-berlin.de/~nils/stat/).
Similarities of seed predator composition among Shorea trees
In order to examine the similarity/dissimilarity of predator composition among hosts, an ordination, Detrended Correspondence Analysis (DCA) was carried based on the frequency of predators. The frequency was expressed as the number of emerging individuals from 100 seeds in immature phase and mature phase (excluding abortion phase) for each predator. The predators other than the five abundant species were combined as ‘others’. Thus, there were six variables on insect frequency for each host species. The data were square-root transformed prior to the analysis.
The host species that fruited in both events were analysed independently. Three Dipterocarpus species were excluded from the analysis since no predators overlapped with those of Shorea. Shorea bracteolata in F-02 was also excluded since no mature fruits were sampled in the event. A program package PC-ORD ver.4.0 (MJM Software, Gleneden, USA) was used for the DCA.
RESULTS
Species composition of pre-dispersal seed predators
In total, 5211 insects were obtained from 14 162 fruits of the Shorea species in F-01 (Appendix 1), and 3166 insects were obtained from 12 748 fruits in F-02 (Appendix 2). Among these, 2144 individuals in F-01 and 1655 individuals in F-02 were weevils or moths, both of which were considered pre-dispersal seed predators. The other insects were hymenopteran parasitoids (Braconidae and Ichneumonidae), orthopteran nymphs, flies (Diptera) and bark beetles (Scolytidae). The flies and bark beetles were not regarded as pre-dispersal seed predators in this study because they were seldom found in fresh fruits. They were probably post-dispersal seed predators or scavengers (bark beetles even sometimes fed on wetted paper used to maintain humidity in the rearing boxes), although the numbers of individuals were sometimes high even in a single fruit. In the present paper, we refer to ‘seed predator’ or ‘predator’ only when referring to the pre-dispersal seed predator, i.e. weevils and moths.
Among the predators, two nanophyid weevils: Nanophyes shoreae Marshall (Coleoptera: Nanophyidae) (1075 and 1278 individuals in F-01 and F-02, respectively) and Nanophyidae gen. & sp. indet. 1 (nanophyid sp. 1) (209 and 138) and two Alcidodes weevils: A. dipterocarpi Marshall (Coleoptera: Curculionidae) (245 and 53) and A. humeralis Heller (367 and 36), and one moth Andrioplecta shoreae Komai (Lepidoptera: Tortricidae) (134 and 93) were dominant in terms of the number of individuals; they comprised more than 80% of all predators in each host except S. bracteolata in F-01 and S. hopeifolia in F-02 (Figure 1a), and accounted for 95.5% and 97.0% of all predators.
Figure 1. The composition of individual numbers of seed predatory insects emerged from seeds of Shorea species (a) and from seeds of other genera (b). The total number of seed predators is shown in the parentheses following the code name of hosts. No data for Dipterocarpus costulatus is presented here since only one individual weevil was obtained from the host (Appendix 2). * P < 0.05, *** P < 0.001 (Fisher's exact test). Ns: Nanophyes shoreae, Nu: nanophyid sp. 1, Ad: Alcidodes dipterocarpi, Ah: A. humeralis, Ma: Andrioplecta shoreae: OW: other weevils, OM: other moths. See Table 1 for the code names of the hosts.
Neobalanocarpus heimii also hosted nanophyid sp. 1 and Andrioplecta shoreae, as well as unique clearwing moth Synanthedon nautica Meyrick (Figure 1b). None of the seed predators of Shorea spp. emerged from the three Dipterocarpus spp., but other weevils of the genus Damnux (Nanophyidae) and two pyraustine moth species (Pyralidae) were obtained.
The predator species composition differed significantly between two events in all the five hosts that fruited in both events (Fisher's exact test, P < 0.001 for S. acuminata, S. leprosula and S. macroptera, P = 0.038 for S. pauciflora) except S. bracteolata (P = 0.467) (Figure 1a). The relative abundance of N. shoreae to the total predators was significantly larger in F-02 than F-01 in S. leprosula and S. macroptera (Fisher's exact test, P < 0.001), and this was true for that of nanophyid sp. 1 in S. acuminata (P < 0.001). By contrast, that of Alcidodes dipterocarpi and A. humeralis was significantly smaller in F-02 than F-01 (P < 0.001) in S. acuminata, S. leprosula and S. macroptera. Andrioplecta shoreae was also significantly more abundant in F-01 than F-02 in S. acuminata (P < 0.001) and S. leprosula (P < 0.05).
Temporal pattern of predator species composition during seed development
The predator species composition changed over the seed developmental period. For example, Shorea acuminata in F-01 (Figure 2), the nanophyid spp. began to emerge from fruits dropped 7 wk after flowering and continued to emerge for another 7 wk. On the other hand, Alcidodes spp. started to emerge from fruits dropped 10 wk after flowering and continued until just before the end of seed fall. The frequency of nanophyid spp. was significantly higher during the immature phase than the mature phase (Fisher's exact test, P < 0.001) while that of Alcidodes spp. was significantly higher in the mature phase than the immature phase (P < 0.001). The predator species succession during seed development occurred in the same way in other hosts (Table 2). Andrioplecta shoreae was also found significantly more frequently in the mature phase than the immature phase for some hosts (P < 0.05) (Table 2). However, this species seemed to feed on seeds (or fruits) of various stages and its larvae were often observed even earlier than those of nanophyid spp.
Figure 2. Temporal patterns of insect predator frequency in dropped fruits (bar) and relative fruit weight (circle) during seed developmental period of Shorea acuminata in F-01. Fruiting period was divided into three phases based on the relative fruit weight. Ns: Nanophyes shoreae, Nu: nanophyid sp. 1, Ad: Alcidodes dipterocarpi, Ah: A. humeralis, Ma: Andrioplecta shoreae: OW: other weevils, OM: other moths.
Table 2. The result of comparison in predator frequency between immature and mature phase for six host species in two fruiting events. The code name for insect species is shown in the column in which phase it emerged significantly more frequently. ns = no significant difference between phases. The comparison in frequency was made by Fisher's exact test (*P < 0.05, **P < 0.01, ***P < 0.001). Ns: Nanophyes shoreae, Nu: nanophyid sp. 1, Ad: Alcidodes dipterocarpi, Ah: A. humeralis, Ma: Andrioplecta shoreae.
The level of host specificity of predators
The degree of host specificity (d′) ranged from 0.03 to 0.94 among predator species, with the weighted average of 0.16 and 0.20 in F-01 and F-02, respectively. Most predator species were below the average in host specificity in both events (Figure 3). As for the five abundant predators, Andrioplecta shoreae showed the lowest d′ of the five. The host specificity of N. shoreae, Alcidodes dipterocarpi and A. humeralis were around the average, and nanophyid sp. 1 was higher than the average in both events. No correlation was found between d′ and relative abundance (Spearman's rank correlation, n = 14, r = −0.07, P = 0.982 in F-01; n = 11, r = 0.141, P = 0.679 in F-02). The average of d′ was higher in F-02 than F-01 and it was true for all the five abundant predators except N. shoreae.
Figure 3. Relative abundance (all Shorea combined) and the degree of host specificity measured by standardized Kullback–Leibler distance (d′) of insect predators (N ≥ 2) in F-01 (a) and F-02 (b). The five abundant predators are indicated as follows, Ns: Nanophyes shoreae, Nu: nanophyid sp. 1, Ad: Alcidodes dipterocarpi, Ah: A. humeralis, Ma: Andrioplecta shoreae.
Similarity of seed predator composition among Shorea species
The Axis 1 and Axis 2 extracted by DCA based on the abundance of predators explained 56.2% and 10.3% of the total variance, respectively (Figure 4). Axis 1 was correlated with the frequency of N. shoreae negatively (r=−0.911) and with ‘others’ positively (r=0.855). Axis 2 was correlated with the frequency of Andrioplecta shoreae negatively (r=−0.242) and with nanophyid sp. 1 (r=0.892) and Alcidodes dipterocarpi (r=0.835) positively.
Figure 4. Ordination of host species by DCA based on the abundance of each predator species (number of predators per seed). See Table 1 for the code names of the hosts.
Many of the Shorea species were clumped in the lower scores of Axis 1 due to the dominance of N. shoreae (Figure 4, Group 1). However, S. acuminata and S. parvifolia were outstanding in higher scores of Axis 2 since they had nanophyid sp. 1 and Alcidodes dipterocarpi abundantly as well as Nanophyes shoreae (Group 2). Shorea hopeifolia, S. bracteolata and Neobalanocarpus heimii had higher scores in Axis 1 since they had little or no N. shoreae but more ‘others’ (Group 3). The same tree species of different events were in the same group, suggesting that abundant predator species did not differ largely within the same host between events.
DISCUSSION
Pre-dispersal seed predator composition
In the present study, we found four weevil species: Nanophyes shoreae, nanophyid sp. 1, Alcidodes dipterocarpi and A. humeralis, and one moth species Andrioplecta shoreae were abundant on Shorea seeds. Principally, an individual larva of the predators feeds and kills one individual seed, thus dominance of these predators indicates their importance in terms of the number of seeds destroyed. Shorea species examined in the present study are typical and widely distributed throughout lowland rain forests of Peninsular Malaysia. Thus, the dominance of these predators on Shorea seeds is likely to be common in that region. On the other hand, Nakagawa et al. (Reference NAKAGAWA, ITIOKA, MOMOSE, YUMOTO, KOMAI, MORIMOTO, JORDAL, KATO, KALIANG, HAMID, INOUE and NAKASHIZUKA2003) reported from Borneo that weevils in the genera Orchestes (Curculionidae) or Aracerus (Anthribidae) were also abundant as well as Nanophyes and Alcidodes (Sternuchopsis in their paper) in Shorea seeds. Reflecting the higher number of Shorea species in Borneo: 138 Shorea in Borneo (Ashton Reference ASHTON, Soepadmo, Saw and Chung2004) and 58 in Peninsular Malaysia (Symington Reference SYMINGTON2004), seed predator weevils might be more diverse in Borneo than Peninsular Malaysia.
The year-to-year variation in relative abundance of seed predators indicated that N. shoreae and nanophyid sp. 1 were proportionately more abundant in F-02 than F-01 (Figure 1). Their biology, e.g. reproductive cycle, might be more suitable for consecutive fruiting events like F-01 and F-02 than other predator species.
Species succession of seed predators during seed development
The sequential sampling over seed development revealed many seeds were attacked and aborted before seed maturation. Predator species changed considerably during the seed developmental period; the nanophyid weevils were found in immature aborted seeds while Alcidodes weevils were found in nearly mature to mature dropped seeds (Figure 2). Species succession of insect predators during seed development has been reported in other plants (Fukumoto & Kajimura Reference FUKUMOTO and KAJIMURA2001, Igarashi & Kamata Reference IGARASHI and KAMATA1997). Fukumoto & Kajimura (Reference FUKUMOTO and KAJIMURA2001) successfully categorized insect seed predators of Japanese oaks into three guilds: pistillate-flower-feeding guild, immature-acorn-feeding guild, and mature-acorn-feeding guild. They bagged canopy twigs with fruits sequentially during seed development in order to determine oviposition period of insects. The nanophyid weevils, Alcidodes weevils and Andrioplecta shoreae in the present study probably correspond to immature seed-feeding, mature (or nearly mature) seed-feeding, and all seed-feeding predators, respectively, based on the information when they aborted. The oviposition timing of dipterocarp predators has not been revealed except that of N. shoreae, which oviposit within 15–48 d after peak anthesis (Toy Reference TOY1991). It is essential to determine the oviposition period for other predators to fully understand the timing of predation.
In dipterocarps, most initiated fruits, whether damaged or undamaged, are known to be aborted before maturation (Momose et al. Reference MOMOSE, NAGAMITSU and INOUE1996, Naito Reference NAITO2008, Nakagawa et al. Reference NAKAGAWA, TAKEUCHI, TANAKA and NAKASHIZUKA2005, Sakai et al. Reference SAKAI, MOMOSE, YUMOTO, NAGAMITSU, NAGAMASU, HAMID and NAKASHIZUKA1999). Only a small proportion of initiated fruit can grow with sufficient resources provided by the mother tree. Momose et al. (Reference MOMOSE, NAGAMITSU and INOUE1996) observed that only 3.6% of the total number of flowers received 47% of the total reproductive investment in dry weight and matured to germination. In this context, loss of resource and reduction of sound seed production due to predation, would differ considerably with the timing of predation, i.e. the greater impact on sound seed production occurs during late predation. Immature seed predation by nanophyid spp. was observed more frequently than that by other predators, but the fruit loss due to this predation might be compensated for by undamaged fruits that would otherwise be aborted later. Insect exclusion did not affect the seed survivorship at the earlier stages of seed development of Dryobalanops aromatica while it did affect negatively at the later stage (Nakagawa et al. Reference NAKAGAWA, TAKEUCHI, TANAKA and NAKASHIZUKA2005). Therefore, predation by Alcidodes spp., rather than by nanophyid spp., might cause greater resource loss and reduce the total number of sound seeds produced.
Host specificity of predators
The degree of host specificity (d′) of most predator species was below the low average (0.16 in F-01 and 0.20 in F-02) (Figure 3), suggesting that insect predators have low host specificity to Shorea seeds. The exceptions were Alcidodes sp. 1 (d′i = 0.86) in F-01 and A. shoreaphilus (0.94) in F-02, which emerged only from S. bracteolata and S. hopeifolia, respectively. Among the five abundant predators, nanophyid sp. 1 showed higher host specificity than others. The weevil mostly emerged from S. acuminata or S. parvifolia, even though it had at least six Shorea hosts (Appendices 1 and 2), suggesting its strong preference to these hosts. When S. parvifolia did not fruit well in F-02, the weevils aggregated to S. acuminata (d′ = 0.62).
Since the fruiting frequency is different even among closely related Shorea species (Numata et al. 2001), the species composition of fruiting trees is often different between fruiting events (Brearley et al. Reference BREARLEY, PROCTOR, SURIANTANA, DALRYMPLE and VOYSEY2007). The low host specificity would be advantageous to the predators to maximize the opportunity for utilizing fruiting events.
Similarity of seed predator composition among Shorea species
With the dominance of N. shoreae, the scores in DCA Axis 1 were low for many Shorea species in Pasoh (Figure 4, Groups 1 and 2). Host species in group 1 shared three of the five abundant predators: N. shoreae, Alcidodes humeralis and Andrioplecta shorae while those in group 2, S. acuminata and S. parvifolia, had all of the five abundant predators (Figure 2). The hosts in group 3 had only one or two of the five abundant predators but often had unique ones, e.g. Alcidodes shoreaphilus Lyal from S. hopeifolia or A. sp. 1 from S. bracteolata (Appendices 1 and 2).
All trees in group 1 and 2 belong to a timber group Red Meranti which consists of sections Mutica, Brachypterae and Ovalis, except S. maxwelliana (section Shorea) belonging to a timber group Balau. On the other hand, S. hopeifolia (section Richetioides) and S. bracteolata (section Anthoshorea) in group 3 belong to a timber group Yellow Meranti and White Meranti, respectively. Kamiya et al. (Reference KAMIYA, HARADA, TACHIDA and ASHTON2005) reported nuclear gene PgiC of Red Meranti was closely related to Balau, and they were closer to Yellow Meranti than White Meranti. Furthermore, White Meranti is very close to Neobalanocarpus. This is roughly consistent with the scores of DCA Axis 1 (Figure 4), thus differences in compositions of insect seed predators may well reflect the phylogenetic distance between sections or timber groups of the hosts.
Implications for predator satiation of insect predators
Silvertown (Reference SILVERTOWN1980) pointed out three interdependent elements for predator satiation to work: (1) production of enough seeds to satiate predators, (2) sufficient interval between fruiting events to reduce the population of predators, (3) sharing the same seed predators among sympatric species fruiting synchronously. Wide overlap of seed predators among many Shorea spp. in Pasoh ensures the third element. Synchronized mast seeding by congeneric trees sharing the same seed predators has been reported in Chionochloa spp. (Poaceae) from New Zealand (Kelly et al. Reference KELLY, HARRISON, LEE, PAYTON, WILSON and SCHAUBER2000, McKone et al. Reference MCKONE, KELLY, HARRISON, SULLIVAN and CONE2001) and the cycad Encephalartos spp. from South Africa (Donaldson Reference DONALDSON1993).
On the other hand, there were two findings that might contradict the second element. First, predators probably have alternative hosts that are available between mast fruiting years. For example, nanophyid sp. 1 and Andrioplecta shoreae were also obtained from Neobalanocarpus heimii in the present study (Figure 1b). Unlike other dipterocarps, N. heimii is known to fruit almost every year at the population level (Burgess Reference BURGESS1972, Marzalina et al. Reference MARZALINA, JAYANTHI, ANG, Okuda, Manokaran, Matsumoto, Niiyama, Thomas and Ashton2003). Furthermore it takes 6 mo to grow seeds and another 1 y to complete seed dispersal (M. Yasuda, pers. comm.). One can thus find its fruits in various developmental stages at almost any time of the year (Hosaka et al. Reference HOSAKA, ARITA and KIRTON2007). Therefore, the two insect species certainly rely on N. heimii during the non-fruiting period of Shorea trees, the second most abundant dipterocarp species in our plot (Kochummen Reference KOCHUMMEN1997). Nakagawa et al. (Reference NAKAGAWA, ITIOKA, MOMOSE, YUMOTO, KOMAI, MORIMOTO, JORDAL, KATO, KALIANG, HAMID, INOUE and NAKASHIZUKA2003) found some of the dipterocarp seed predators including Nanophyes shoreae, Alcidodes dipterocarpi and Andrioplecta shoreae, from trees of other families such as Moraceae, Myrtaceae, Celastraceae and Sapotaceae. The rather wide host range of these insects might enable them to reproduce and keep their populations between mast-fruiting events.
Second, weevil predators might have an extended diapause. We found many weevil larvae still alive and probably in diapause after 4-mo rearing. Some of the larvae lived more than 1 y in the rearing boxes. A similar observation was reported by Nakagawa et al. (Reference NAKAGAWA, ITIOKA, MOMOSE, YUMOTO, KOMAI, MORIMOTO, JORDAL, KATO, KALIANG, HAMID, INOUE and NAKASHIZUKA2003). The extended diapause is known in many seed predators both in larval and adult stages (Donaldson Reference DONALDSON1993, Hanski Reference HANSKI1988, Janzen Reference JANZEN1971, Kelly et al. Reference KELLY, HARRISON, LEE, PAYTON, WILSON and SCHAUBER2000, Maeto & Ozaki Reference MAETO and OZAKI2003), and might be effective to track years of masting by responding to the same weather cue as the plants (Kelly et al. Reference KELLY, HARRISON, LEE, PAYTON, WILSON and SCHAUBER2000).
However, it should be noted that, although seeds of N. heimii are available throughout the year, its density was not comparable to that of Shorea spp. in mast years. In addition, since carnivorous animals and disease micro-organisms are active throughout the year in the aseasonal tropics, mortality is likely to be high during years of diapause. Thus, population size of predators would become smaller after a mast fruiting event even they have alternative hosts and/or diapause. Moreover, Kelly et al. (Reference KELLY, HARRISON, LEE, PAYTON, WILSON and SCHAUBER2000) argued that such counter-adaptations of predators might even contribute to maintaining or strengthening the selection for extreme masting, since predators would become harder to satiate by normal levels of masting.
It is difficult to explain fruiting synchrony at family level by satiation of pre-dispersal insect seed predators alone since no predator overlap was found between Shorea and Dipterocarpus in the present study, in contrast to the results from Borneo (Nakagawa et al. Reference NAKAGAWA, ITIOKA, MOMOSE, YUMOTO, KOMAI, MORIMOTO, JORDAL, KATO, KALIANG, HAMID, INOUE and NAKASHIZUKA2003). The original idea of predator satiation in the mast fruiting of dipterocarps assumed polyphagous post-dispersal vertebrate predators (Janzen Reference JANZEN1974), and some evidence to support the idea has been presented (Curran & Leighton Reference CURRAN and LEIGHTON2000, Curran & Webb Reference CURRAN and WEBB2000). However, considering the wide overlap of important insect predators among Shorea, the synchronized mast fruiting by congenerics is possibly also effective for avoiding pre-dispersal seed predators. Thus it would be important in order to produce and disperse a large number of viable seeds sufficient for swamping post-dispersal predators.
ACKNOWLEDGEMENTS
We are grateful to Dr Laurence G. Kirton of the Forest Research Institute of Malaysia (FRIM) for permission to conduct the research in Pasoh. We thank Dr Chris Lyal (Natural History Museum London) for helping weevil identification and providing information on them. Special thanks are due to Dr Naoya Osawa (Kyoto University), and Dr Toshinori Okuda (Hiroshima University) for their kind support. Dr Shinya Numata (Center for Research and Development Strategy), Dr Sen Nishimura (Kyoto University), Dr Akihiro Konuma (National Institute for Agro-Environmental Sciences), Dr Masatoshi Yasuda (Forestry and Forest Products Research Institute), Dr Yoko Naito (Kyoto University), Dr Yu-Yun Chen (Georgia University), Dr I-Fang Sun (Tunghai University), and Malaysian assistants kindly provided us with some insect samples and much information on Shorea and other dipterocarp flowering in Pasoh. The field work and insect rearing could not have been completed without kind help of Mr Mohd Widradul and FRIM staff in Pasoh. Parasitoid wasps were identified by Dr van Achterberg (Nationaal Natuurhistorisch Museum) and Dr Y. F. Ng (University Kebangasaan Malaysia), and nanophyid weevils were by Mr Junnosuke Kanto (Tokyo University of Agriculture). The comments from two anonymous reviewers greatly improved the manuscript. This study was partly supported by a joint research project of FRIM and the National Institute for Environmental Studies, Japan: Grant No. E-4, Global Environmental Research Program, Ministry of Education, Culture, Sports, Science and Technology of Japan.
Appendix 1. A list of insects that emerged from dipterocarp fruits in 2001 flowering. The codes for plant species are as follows. AC: Shorea acuminata, BR: S. bracteolata, DS: S. dasyphylla, LR: S. leprosula, MC: S. macroptera, PR: S. parvifolia, PU: S. pauciflora.
Appendix 2. A list of insects that emerged from dipterocarp fruits in the 2002 flowering. HP: S. hopeifolia, LI: S. lepidota, MX: S. maxwelliana, OV: S. ovalis, DI: Dipterocarpus crinitus, DR: D. cornutus, DS: D. costulatus, NH: Neobalanocarpus heimii.
1Including fruits originally from the 2001 flowering.
