Chaos of Wolbachia Sequences Inside the Compact Fig Syconia of Ficus benjamina (Ficus: Moraceae)

Figs and fig wasps form a peculiar closed community in which the Ficus tree provides a compact syconium (inflorescence) habitat for the lives of a complex assemblage of Chalcidoid insects. These diverse fig wasp species have intimate ecological relationships within the closed world of the fig syconia. Previous surveys of Wolbachia, maternally inherited endosymbiotic bacteria that infect vast numbers of arthropod hosts, showed that fig wasps have some of the highest known incidences of Wolbachia amongst all insects. We ask whether the evolutionary patterns of Wolbachia sequences in this closed syconium community are different from those in the outside world. In the present study, we sampled all 17 fig wasp species living on Ficus benjamina, covering 4 families, 6 subfamilies, and 8 genera of wasps. We made a thorough survey of Wolbachia infection patterns and studied evolutionary patterns in wsp (Wolbachia Surface Protein) sequences. We find evidence for high infection incidences, frequent recombination between Wolbachia strains, and considerable horizontal transfer, suggesting rapid evolution of Wolbachia sequences within the syconium community. Though the fig wasps have relatively limited contact with outside world, Wolbachia may be introduced to the syconium community via horizontal transmission by fig wasps species that have winged males and visit the syconia earlier.


Introduction
Wolbachia (Alphaproteobacteria) are maternally inherited endosymbiotic bacteria that are found in arthropods and filarial nematodes [1,2]. A meta-analysis estimates that the proportion of Wolbachia-infected arthropod species is 66% and that there are some 1 million infected species in insects alone [3]. Recently, with a more appropriate data set from both a broad range of species and a sufficient number of individuals per species, this number was reduces to 40%, however confirming that Wolbachia are the most abundant endosymbionts among arthropod species [4]. Besides the pandemic distribution, host reproductive manipulations are also important features of Wolbachia. The bacteria spread and persist in some host species by manipulating host reproduction via methods including feminization of genetic males [5], parthenogenesis induction [6], killing of male progeny from infected females [7] and cytoplasmic incompatibility [8]. These all provide relative reproductive advantages to infected females, thereby favoring the persistence and spread of the Wolbachia strain in host populations [9].
Vertical transmission from mother to offspring via the egg cytoplasm is thought to be the predominant mode of Wolbachia transmission within a host species [10]. However, horizontal transfer of Wolbachia has also been detected both within and among different host species in many cases. Various findings suggest the existence of horizontal transfer. For example, closely related bacterial strains can be found in taxonomically unrelated hosts [11,12,13]; the phylogeny of Wolbachia is often strongly incongruent with that of their hosts [12,13,14,15,16,17]. Further, it is possible to transfect Wolbachia from native hosts to other species, although a newly acquired strain often cannot maintain high infection rates over many generations [18,19,20]. A case study in the wasp genus Nasonia indicated that the four closely related species were infected with eleven different Wolbachia strains, and that horizontal transmission, hybrid introgression and codivergence during host speciation all contributed to the current infection patterns [21]. Though the exact transfer mechanisms are not well understood, many factors may play a part in the process, including close phylogenetic relationship, physical intimate contact or ecological association [17,19,22,23]. Recent work has also shown that host plants may be conduits of bacterial symbionts between insect herbivores that feed upon them [22,24].
In order to investigate the transmission of Wolbachia or its effects on hosts, many studies have compared samples from different localities, in some cases even among different continents [2,3,25,26]. Other studies have focused on ecological communities or at least interacting host species, such as social hosts [27], insects with host-parasite relationships [19], or otherwise linked through feeding on the same food [17,22]. These species are phylogenetically or ecologically closely related, and even form communities in some cases. However the environment they occupy is open, so they have many opportunities to interact with a wide range of species that is hard to define. This can make it difficult to define the boundaries of a relevant panel of potential hosts from which they may acquire Wolbachia infections. However, the diverse but  [28,29,30,31]. We suggest that they may also provide a good model to study infection patterns and transfer routes of Wolbachia.  [33,34]. Most of the wasp species are chalcidoids from the families Agaonidae, Pteromalidae, Eurytomidae, Torymidae and Orymidae. Different wasp species can survive within the same syconium by employing different living strategies, such as using different flower layers [35,36], laying eggs at different phases of fig maturity [30,37] and having diverse larvae diets (e.g. gallers and parasitoids) [30,38,39,40,41,42,43,44]. Upon maturation, the wingless males of most species mate with females inside the syconia, which they do not leave. Meanwhile, other non-pollinator species have fully winged males, which emerge from the syconia and search for females after dispersal.
Previous surveys of fig wasps have proved that the incidence of Wolbachia infection in fig-associated wasps is significantly higher than for a broad collection of insects [11,25,26]. The compact and closed world of the syconium is one in which intimate physical contact and close ecological association between species are inevitable. Together with the relatively close phylogenetic relationships of these fig wasps species, this suggests that horizontal transmission of Wolbachia is more likely to occur in this species community than other less closed systems. The frequent horizontal transmission of Wolbachia may also lead to the high evolutionary rates of Wolbachia genes, for example the recombination events of outer membrane protein genes, due to the immune responses of the hosts. In the present studies, we are very interested in the following questions: In the present study, we focused on the 17 species of chalcidoid wasps living on Ficus benjamina, covering 4 families, 6 subfamilies, and 8 genera. We made a thorough survey of Wolbachia infection patterns by using the Wolbachia Surface Protein gene (wsp) as a marker. We also observed and investigated the biological characteristics on these fig wasps, based on which we tried to reconstruct horizontal transfer routes of wsp. determined by morphological diagnostics by using Nikon SMZ80 microscopes, with most of the species described in a previous study in our lab which have corresponding data for both morphological and molecular characters [45]. We tried to maximize the independence of samples by choosing samples from different batches of syconia so as to avoid sampling with an unusual Wolbachia distribution. The fig wasp species used are listed in table S1.

DNA Extraction
Total genomic DNA was isolated from the whole body for each individual sample by using Easypure genomic DNA Extraction kits (TransGen, Beijing, China). The quality of the templates were confirmed by amplification of a partial fragment of COI gene with around 700 bp length (Primers: LCO1490: 59-GGTCAACAAAT-CATAAAGATATTGG-39 and HCO2198: 59-TAAACTT-CAGGGTGACCAAAAAATCA-39), and the DNA templates with poor quality were discarded. All of the specimen and DNA vouchers were deposited in College of Plant Protection, Shandong Agriculture University in China.

PCR Amplification and Sequencing
The samples were first screened for Wolbachia infection by PCR amplification with the primers wsp81F (59-TGGTCCAATAAGT-GATGAAGAAAC-39) and wsp691R (59-AAAAATTAAACGC-TACTCCA-39) that amplified part of the Wolbachia Surface Protein gene (wsp) [46]. Genomic DNA from Ceratosolen solmsi (Hymenoptera: Chalcidoidea), a pollinator of Ficus hispida, which had high infection titre of Wolbachia was used as positive control, while ddH 2 O was used as template for a blank control. Because in some cases, the amplification of wsp81F and wsp691R could not get a single band, another pair of primers were used, wsp1f (59-AATAAGTGATGAAGAAACTAG-39) and wsp691R. The locality of wsp1f is on the right of wsp81F by 3 bp, with 18 nucleotides exactly the same as wsp81F. Both of them do not have degenerate sites. A BLAST search on wsp81F of NCBI GenBank, we found subjects with identical wsp81F sequences also have the same inner 3 bps as wsp1f. So we consider them to have almost the same sensitivity. Touchdown PCR programs were used: 5 min at 94uC, followed by 20 cycles of 30 s at 94uC, 45 s at 56uC (with 0.5uC reduction every cycle), 1 min at 72uC, and then 8 cycles of 30 s at 94uC, 45 s at 46uC, and 1 min at 72uC; and a final elongation step of 10 min at 72uC. PCR products were purified by EasyPure PCR purification Kit (TransGen, Beijing, China) and directly sequenced by ABI3730 capillary autosequencer (Biosune, Shanghai, China). If the sequencing results showed multiple peaks, the PCR products were cloned with Peasy-T1 vector (TransGen, Beijing, China), and 3 positive clones were sequenced. In case the clones showed high sequence diversity, more clones were sequenced. On average, 5-6 clones were sequenced in each multiple infected individual. The length of wsp sequences is from 552 to 573 bp with primers of wsp1f and wsp691R and from 555 to 576 bp with primers of wsp81F and wsp691R. In order to acquire accurate infection frequency in each species, negative samples of wsp were double-checked by amplification of another two genes, ftsZ and 16S rDNA.
For multilocus sequence typing (MLST), we only selected the samples whose wsp sequence was consistent with a single Wolbachia infection, and a single clear single gel band was obtained. We used a system of MLST primer F3/R3 nested with MLST standard primers, with the primer sequence and PCR conditions as described on the Wolbachia MLST website (http://pubmlst.org/ wolbachia/info/protocols.shtml). We also tried to acquire MLST sequences from low Wolbachia titre samples, but it was difficult to successfully amplify all the five gene fragments.

Sequence Analysis
Raw sequences treatments, wsp typing and ascertainment of recombinants. If multiple identical sequences were obtained for each species, we chose only one to represent them. The sequences were then submitted for wsp typing (http://pubmlst.org/wolbachia/wsp/). According to the typing results, the sequences were clustered into different groups of nonrecombinant or recombinant wsp types. The suggested recombinant evidence of the recombinant wsp sequences were further ascertained by using the programs of RDP, GENECONV, MaxChi, BootScan, and SiScan in RDP3 software [47]. Phylogenetic analyses. The non-recombinant wsp sequences were aligned to relevant sequences previously published on NCBI with Clustal W in Bioedit [48]. Neighbor-Joining trees were constructed by using p-distance model in Mega 5 [49]. Gaps and missing data were pairwise deleted. Bootstrap values were calculated from 1000 replications.
Statistic analyses. All of the statistic comparison analyses were based on Chi square test with P values larger than 0.05 indicating insignificant divergences. We sampled a total of 17 fig wasp species, which belong to 4 families and 8 genera. Eupristina koningsbergeri Grandi is a pollinator, while the others are non-pollinators (table S1) [45]. Wasp number and species composition are very similar among the trees observed. In principle, more wasps are produced in raining season (May to October) than in dry season (November to April). However, the species composition is relatively stable all the year round. Pollinators occupy half of the community population, while the rest are non-pollinators. Many species are easily found whereas some species are rare species, such as Sycoscapter sp. 2, Ormyrus sp. 1, and Sycophila sp. 3. Up to 12 species co-occurred in individual syconia, while up to 14 species were collected on the same batch of a fig tree.

Ecological Observations on the Fig Wasps Associated with Ficus benjamina
The species assemblage has diverse oviposition behaviors with different species arriving at the figs at different stages of development and using different oviposition sites ( Figure 1 and Table S1). Seven species from four genera visit at pre-female stage; 2 species come at female stage, including pollinator Eupristina koningsbergeri; 7 species visit at interfloral stage. The detailed order has been described in Figure 1. Except Eupristina koningsbergeri lay eggs from inside the syconium, all the others oviposite from outside the syconium. Most of the earlier visitors are prone to be gall makers, while the later are mostly inquilines or parasitoids (Table  S1). The species assemblage has diverse larval diets of gall-makers, parasitoids or inquilines, which can indicate the parasitic relationships of some of these species within the same syconium (Table S1).

High Wolbachia Infection Incidence in Ficus benjamina Wasps
We screened 870 wasps of the 17 species associated with F. benjamina for Wolbachia infection. The detailed results are listed in table 1. Fourteen out of 17 species were found to be infected. This infection incidence (84%) is higher than in wider screenings (from many fig species) of fig wasps in Australia (67%) and Panama (59%), but the differences are not significant (X 2 = 0.031, P.0.05).

Wsp genotypes of Wolbachia in this Fig Wasp Community
After removing repeated identical sequences within each species, we obtained 103 effective wsp sequences from the infected species. These were aligned based on the amino acid motifs of the four hypervariable regions (HVRs) and arranged into four groups: group 1, 45 non-recombinant sequences, 7 genotypes; group 2 & group 3: 57 recombinants, 28 genotypes (the detailed comparison of these two groups were described in the following sections); group 4: one unique sequence (data not shown), 1 genotypes. Evidence of recombination for the 57 recombinant sequences was further confirmed by RDP software (data not shown). All of the sequences were submitted to Genbank with accession numbers JQ342186-JQ342231, JQ361490-JQ361513 and JX192925-JX192928. Within each non-recombinant type, the percentage divergence is from 0 to 0.004, however, between each type, the net percentage divergence is from 0.086 to 0.480. Six belong to Wolbachia supergroup A, while wben-2 belongs to supergroup B. Previous studies on Wolbachia infection of fig wasps found that Wolbachia strain are region-specific. Strains in Asia and Australia are more similar, while they have more divergence from those in the new world, for example, Panama [25,50]. However in our research, wben-1 is very similar to the wsp type infected several fig wasp species in Panama, in contrast to previous results that Wolbachia strains infecting fig wasps in Panama and Australia were different and region-specific [25].

Detection of Seven non-recombinant Wolbachia wsp Genotypes
Four of the wsp types were common ones with wide distributions in other insect hosts besides the fig wasps in Ficus benjamina. These were wben-1, wMors, wMel and wHaw, which were found respectively in previous studies [23,51,52,53,54,55,56,57]. Wben-2, wben-3, and wben-4 were unique wsp types first detected here in our F. benjamina community. Wben-2 was widely distributed in this community and often co-infects with other types, while wben-3 existed only in some samples from Hainan. Wben-4 was only detected in Ormyrus and this host species had no other wsp types (Table 1). We attempted multilocus sequence typing (MLST) of all nonrecombinant wsp strains that occurred as single infections. This involved use of the five housekeeping genes (gatB, coxA, hcpA, ftsZ, and fbpA). However, due to the low titre of Wolbachia, we only obtained full MLST data for wHaw and wben-2, which can now be identified as ST-19 and ST-274 respectively (table S2).

Abundant Recombinant Sequences Exist in Ficus benjamina
We detected a large number of recombinant wsp types in this study. Of the 57 recombinant sequences, 38 can be assigned to non-recombinant ''parents'' detected in the F. benjamina community (group 2). The origins of the other 19 (group 3) are less clear and involve parent sequences not detected in the fig wasp community. Considering the HVRs composition and origin of all recombinants, we classified them into 28 types, named rec-1,rec-28 (data not shown). By MLST methods, we only obtained full strain information for rec-28 (Table S2). The distribution of recombinant wsp types in the different host species are listed in table 1. All may be functional since they can be translated without frame-shifts or early stop codons.
The recombination patterns of the 28 recombinant types according to the origin of the four hypervariable regions are described in Figure 3. Most recombinants (18 of 28) are composed of sequence blocks from two non-recombinant types; 9 recombinants stem from three different non-recombinant genotypes; and rec-24 is composed of sequences from four non-recombinant types. The ubiquitous and Ficus benjamina-specific non-recombinant wsp types differ in their involvement in recombination. The ubiquitous types with wide range of hosts (wben-1, wMores, wMel and wHaw) are also frequently involved in recombination events. In particular, recombinants involving wHaw sequences were found in all species harboring recombinant sequences. However, the three F. benjamina specific non-recombinant wsp types are either not involved in any recombination events (wben-4) or involved in few cases (wben-3 and wben-2) (Figure 3).

Infection Patterns of Wolbachia in Ficus benjamina Community
Based on a survey of 143 infected wasps, we detected that 63 insects have multiple infections, while 80 were single infected. Superinfected samples were mainly from species Philotrypesis sp. 1 The species that harbor more recombinant sequences also have more non-recombinants; for example, in Philotrypesis sp. 1, P. sp. 4, and Sycophila sp. 2, implying a positive correlation of recombinants and non-recombinants. Species with multiple wsp genotypes mostly harbor both recombinant and non-recombinant types. There is only one exception in Philotrypesis sp. 5, which harbors only non-recombinants (wben-3, wben-1). More wsp sequences from Philotrypesis sp. 5 are required to ascertain this infection appearance. However, there is no indication that the recombinants come from the combination of the non-recombinant sequences hosted in the same individual or species. For example, Eupristina koningsbergeri in Hainan hosts non-recombinant wben-3, while it has recombinant rec-3 which derives from wHaw and wMel, not from wben-3 ( Figure 3 and table 1).
The infection patterns are independent of sampling localities (X 2 = 0. 19

Discussion
Our study made a thorough survey on the Wolbachia infection patterns of all the fig wasp species living inside the closed syconium community of Ficus benjamina. We detected high infection incidence in this compact community, as expected. In addition, complicated infection patterns were revealed, including frequent recombination and horizontal transfer, which are signature features of fast evolution of the wsp gene [58].
We sequenced 3-6 clones of wsp of each multiple infect individual. This method would underestimate the number of multiple infect hosts and the number of strains that co-infect in the individual. Fortunately, large sample numbers will compensate this drawback to some degree. Despite the underestimate, common existing multiple infections in individuals and species is a peculiar character in this system.

Frequent Recombination Events within the Syconia
WSP in Wolbachia is confirmed to be an outer membrane protein [59]. WSP may play a key role in an arms race between arthropod hosts and Wolbachia, by inducing host immune responses [60] in which the WSP protein and its surface domains may have a considerable influence. In the dynamic host-bacterium immune interactions, the bacteria can take advantage of the rapidly evolved outer membrane proteins, e.g. WSP, to produce new phenotypes to avoid the host immunity. Recombination may provide an effective means to create new variants and promote the evolution of novel phenotypes, and so far wsp sequences have shown the most remarkable pattern of recombination among the Wolbachia proteins studied so far [58].
In Additionally, most of the recombinants were descended from other sequences within the same community, which indicated that most of these recombinants were produced in the syconium community, and further suggested that intimate insect species interactions within the Ficus syconia provided a platform for horizontal transfer and recombination of Wolbachia.
Several lines of evidence help us to confirm that the recombinants we detected are real recombinants in the cells and not PCR recombinants. First, we have acquired the same recombinants from different host individuals and even from different host species for many wsp genotypes. Second, we can confirm recombinants with repeated PCR reactions on the same template. Third, All the recombinants sequences can be translated correctly; if these sequences were PCR errors it is unlikely that all of them have right open reading frame without interruption.
In previous studies, Werren and Bartos reported the first example of recombination and they detected one type of recombinants. The parasitoid wasp and the flies it parasitizes may provide ecological context for this recombination [61]. In Reuter and Keller's studies, all ants co-infected with four or five distinct Wolbachia strains, three of which were arisen by Figure 2. NJ tree constructed with the obtained non-recombinant wsp sequences and some related sequences downloaded from GenBank (The accession IDs of the downloaded sequences are annotated on the tree). Two sequences from supergroup C and D are used as outgroups. All of the seven non-recombinant wsp types are indicated. The strains to which some wsp types belonged with previously reported names are also indicated. The three wsp types unique to Ficus benjamina are pointed out with m. Wben-2 belongs to supergroup B while the other wsp types belong to supergroup A. doi:10.1371/journal.pone.0048882.g002 homologous recombination [62]. In Spalangia spp. (Hymenoptera: Pteromalidae), 24 out of 48 strains were recombinants [53]. Some statistic analyses predicted that the rate of recombination might be approximately equal to the rate of horizontal transmission [63]. In our studies, 28 types of recombinants occupied 80% (28/28+7) of all genotypes detected within the Ficus benjamina community. These comparison results indicated that the recombination events of wsp happened frequently in the system of Ficus benjamina.
We do not exactly know how recombinants are produced; however, the contact of different wsp DNA sequences is a prerequisite. Studies indicate that wsp occurs as a single copy in the Wolbachia genome, so the co-existence of different wsp sequences in a host is evidence of co-infections of divergent Wolbachia strains in a host, and the widespread occurrence of co-infections provides a suitable arena for recombination [60].
While the ubiquitous non-recombinant strains were extensively involved in recombination, some non-recombinant strains were involved in no or few recombination events. Why do these strains have different appearances? We suggest that the ubiquitous strains have broad hosts both within the syconium and the outside world, which means that they may be more likely to transfer, so coinfection events are easily formed. However for the unique strains only found within the syconium, they were often present in a host individual with high concentration alone, which may be a function of the host biology so that they do not have chance to recombine.

Prevalent Horizontal Transfer of wsp in Ficus benjamina Community
Due to the fast evolution, especially the frequent recombination events, of wsp, it is misleading to classify Wolbachia strains based only on wsp sequences [58,64]. As wsp can be horizontally transferred as a single gene, horizontal transfer of Wolbachia strains can not be inferred solely on the basis of wsp sequences [58,64] Horizontal transfer also appears prevalent in the syconia of Ficus benjamina. As shown in Figure 4 Though the exact horizontal transfer routes can only be speculated upon, there is plentiful evidence that horizontal transfer is common. First, identical wsp sequences exist in species with long distance in phylogenetic relationship. Considering the fast mutation rate of wsp sequences [60], it is unlikely that vertically transmitted Wolbachia accumulate no mutation in wsp sequences among species that diverged a long time ago. Second, there is no clue that the recombinants come from the combination of the nonrecombinant sequences hosted in the same individual or species, so horizontal transfer of recombinants among different species seems more likely. It is common that some recombinant sequences exist in different species in the community, but, it is highly unlikely that identical recombinant sequences form independently and repeatedly in different species. Finally, intimate ecological relationships occur among the different fig wasp species such as host-parasitoid, phytophagous insects-food, co-habitant, or close phylogeny relationship and provide good chances for the wsp sequences to horizontally transfer within the closed system. Previous studies on dipteran insects associated with mushrooms indicate that the mushroom habitat may provide a somewhat comparable (if more open) ecological arena for the exchange of Wolbachia [17]. We summarized previous studies on the Wolbachia infection patterns. These concerned different species, included parasitic insects, social insects, insects within the same community and other phylogenetically closely related insects. The rate of horizontal transfer events of our community was 37.1% (13/37), lower than that in pumpkin community 57% (4/7) and in Nasonia 45.4%. If we only consider the non-recombinants in our system, in which 5 out of 7 were involved in horizontal transfer events, the rate was the highest (71.4%).
We also compared the average strain numbers per species harbored in the surveyed insects: the highest one is Nasonia (11/ 4 = 2.75), second parasitoid wasp (48/21 = 2.29), and the third one is our Ficus benjamina community (35/27 = 2.06). The above data indicated that parasitic insects were prone to harbor more strains.
Fig wasps in Ficus benjamina had 9 species with multiple infections, the rate of which was only lower than the four species of Nasonia.
In summary, our species within the syconia of Ficus benjamina had abundant events of horizontal transmission, a high number of strains and multi-infections. Compared to those species in the outside world, the close relationship of fig wasps within this closed community may accelerate horizontal transfer events (Detailed iformation was listed in supplementary Table S4).

Origin of the Wolbachia in Ficus Benjamina
Wolbachia as an endosymbiont presents with a pandemic distribution in arthropods. In our study, we detected that they infected many species of fig wasps. Irrespective of the transmission modes are vertical or horizontal, the origin of the Wolbachia in Ficus benjamina, especially the non-recombinant strains should be outside of the syconia and they have been introduced inside via certain agents at some time.
Most of the recombinant wsps are descended from the nonrecombinants existing in Ficus benjamina, so we speculate that these recombinants are original within the syconia, considering the high frequency of recombination events occurring between wsp sequences.
How are the non-recombinant Wolbachia strains introduced into the fig/fig wasp system Why are species with winged males more likely the agents helping the transmission of Wolbachia from outside into the syconia? In the closed system of fig wasps, most of the males are wingless and they never leave the syconia where they were born. Only very few species have winged males which will emerge from the syconia and mate on the outside of syconium. The mating process is as following: firstly, winged males come out earlier than females. They linger around on the outer surface of the syconia and wait for females. This process can last for about ten minutes (field survey results of our lab). Many other insects such as ants and flies will attack these males and this intimate contact may make them infected by Wolbachia from other insects. Then during the mating process, Wolbachia can transfer to the reproductive system of the females, which will finally introduce Wolbachia into the syconia through their eggs. Though female may spend more time away from syconium, where they will finish the process of mating, fig searching and oviposition, the introduction route of Wolbachia is more likely to be: other insects R winged males R females R syconia, rather than: other insects R females R syconia. Because the key factor of successful transmission to the offspring is through the reproductive system of females. The mating behavior outside of the syconia can help the introduction of Wolbachia into the syconia through the eggs. Many insects attract females when they are searching for figs or laying eggs, but in these conditions, it may be difficult for Wolbachia to infect the eggs in the ovaries and transmit to the next generation.
Current theory favors the idea that females are important factors influencing Wolbachia transmission, while males are dead ends of Wolbachia vertical transmission. In our studies, only females were included in the PCR screening. However, by combining infection patterns and system ecological data, we suggested that winged males is a proper route to transfer outside strains to the inside world of syconia. Males may also participate in regulating the Wolbachia horizontal transmission within the syconia, by the behavior of fighting within the syconium. We suggest that the closed system of