Natural History

One of the goals of the group who studied spiders in the project "Mechanisms of maintenance of high diversity in the Tropics" (see previous chapter of book) was to recognize biotic and abiotic variables acting on spider diversity. One of the lines within this objective was the evaluation of the coexistence of a restricted group of species of the family Ctenidae that occur in “Reserva Ducke”.

The study of coexistence is complex. On the one hand, it is not sufficient to find a difference between two species (e.g., in the diet or habitat use) to guarantee that this difference is a major contributing factor for coexistence. Even more difficult would be to sustain that the difference has evolved as a result of the coexistence based solely on descriptive studies. Moreover, the experimental verification of the importance of even a single factor can be very complicated in practice, and there are many factors to consider. This does not mean that a study is impossible, but that it depends on complementary approaches. We chose to compare various aspects of biology, evaluating how the differences between species and the characteristics of the environment could contribute to coexistence.

We believe that some of the interactions between the selected species are important, more than with most other species that live with them, and that it is appropriate to define this group as a guild. Competition within guilds can be high, however, in the guilds of wandering spiders the most important interaction would generally be "intraguild predation" (Polis, Myers & Holt 1989), more than the exploitation competition (Schaefer 1972). Intraguild predation means that the larger wandering spider (species) tend to capture the smaller ones, but also that the young of the larger species are prey for the adults of smaller species. It is unlikely that the interactions of these species are symmetrical, that is, a species having the same effect on the other as the other on the former. Therefore, understanding the coexistence in this system involves raising factors that can prevent competitive exclusion or predation exclusion.

In this part of the project, we studied only species of the genera Ancylometes, Ctenus, Phoneutria, Centroctenus, Cupiennius (Ctenidae) and Aglaoctenus (Lycosidae). A greater number of genera would make the analysis of coexistence too complex, and these genera seemed sufficient to address the diversity of interactions under the aspect of coexistence. The choice of genera was based on phylogenetic proximity, on a size spectrum of spiders from medium to large and on conspicuousness and abundance of the species. Another reason was that most of them do not build any kind of web to capture prey, and there are few studies on the coexistence of wandering spiders (including ambushers). Furthermore, with the exception of Cupiennius (see citations in Schuster et al. 1994), there was little knowledge on the natural history of these genera until recently (Schiapelli & Gerschman 1970; Höfer et al. 1994, Gasnier & Höfer 2001; Salvestrini & Gasnier 2001; Gasnier et al. 2002; Rego et al. 2007; Mestre & Gasnier 2008).

Höfer et al. (1994) recorded seven species of the genus Ctenus in Reserva Ducke, however only four of them are abundant. One of the species described by the authors as a new species (C. tapereba) was subsequently synonymized with Centroctenusauberti (Caporiacco) by Brescovit (1996). Ctenus minor was never found in the study area; individuals of this species recorded from Reserva Ducke were probably sampled outside of our study site, in an area of more open vegetation of the “campina” type. The species Ctenus inaja was described by Höfer et al. (1994) in Ctenus, but its position within this genus should be revised. This species is morphologically similar but differ in the color pattern from the other species of Ctenus at Reserva Ducke and shows a very different habitat use. It is usually found on fallen logs or living trees, and apparently the females tend to be close to refuges. Perhaps by living in trees, or because the species really is less abundant, it is much less commonly found than the other species that live on the ground. Three individuals of this species were found on the river island "Ilha do Baixio " (Iranduba), where the ground is completely flooded during part of the year by the waters of the Solimões River (Gasnier, personal observation); in the same collection no other species of Ctenus were found. An individual of this species was found in the canopy of a tree in an urban garden in Manaus (Höfer, personal observation). Considering these observations, and also that we have rarely found small spiderlings on the forest floor, we suppose that this species lives mainly on trees and only occasionally (or in some part of the reproductive cycle) individuals go down to lower strata of vegetation.

The four most abundant species of Ctenus in Reserva Ducke occur throughout all the study area, however the relative abundances vary among habitats (Gasnier & Höfer 2001). Ctenus amphora is the most abundant in heath forests (locally called campinarana) on sandy soils, but are also relatively common in dense forests on yellow latosol and occur in swamp forests (locally called baixio), a habitat with hydromorphic soil near streams. Ctenuscrulsi and C. manauara are the predominant species in dense forest. They are relatively common in swamp forest areas, since these are close to dense forest areas, but are infrequent in heath forests. Ctenusvillasboasi shows a more homogeneous distribution, but its frequency seems to be higher in swamp forests. A more detailed evaluation of abundance patterns reinforces the relationship between abundance and soil type. We found a transition area between the yellow clayey latosol and sandy soils which coincides with a gradual transition from predominance of C. crulsi and C. manauara to a predominance of C. amphora. We also found a small area in heath forest where soils were not sandy as usual, but slightly more clayey, and here we found the highest numbers of C. crulsi and C. manauara for this vegetation type.

The abundance of Ctenus varies greatly in time, apparently seasonally, and in a similar way in the four abundant species (Gasnier & Höfer 2001). Although the temporal variation in abundance has been studied intensively for only one year, we have observed in the following years that the patterns tended to repeat with a strong reduction in abundance at the end of one rainy season and increasing again at the beginning of the next rainy season. The cause of seasonal variation in abundance is not clear and may be a direct effect of the lack or excess of rains, or an indirect effect of the rains, perhaps related to seasonality in the amount of litter. During the manipulation of spiders in predation experiments we noticed that they are sensitive to desiccation, but even during the dry season there seems to be enough moisture for the spiders in the forest. On the other hand, very small juveniles seem to be very fragile and there may be a significant mortality due to waterlogging during heavy rainfall. This would explain why the abundance of these spiders tends to decrease during the second half of the rainy season. However, the amount of litter is also reduced in this period, and this could have an effect on the abundance of spiders (see previous chapter).

Although there is seasonality in abundance, we found no evidence of reproductive seasonality in C. amphora, C. crulsi or C. manauara. We believe that the lifetime of these species is shorter than one year, preventing the possibility of a reproductive season. Only in C. villasboasi, the largest of the four species, we find a structure of individual sizes in the population indicating seasonal reproduction.

The diet of the four species consists mainly of arthropods and other invertebrates such as crickets, termites, cockroaches, beetles, spiders and worms, and occasionally small lizards. However, they differ in the frequency of consumption of the termite workers of the genus Syntermes. These termites live in underground nests and reach the surface to consume leaves of the litter. Syntermes forage in large groups, which make them an attractive prey for spiders, but the workers are defended by a cast of aggressive soldiers. These termites constitute about 50 % of the diet C. crulsi and C. manauara, but less than 10 % of the diet C. amphora and C. villasboasi. This difference in consumption of termites holds when we restrict the analysis to only spiders that preyed on dense forest on latosol, where the four species have access to large numbers of these termites, indicating that there is a difference in the ability to capture termites. In an experiment with spiders in captivity, we have seen that C. crulsi really captured these termites with greater efficiency than C. amphora, although both have similar sizes.

Army ants are probably the main predators of Ctenus. Vieira & Höfer (1994) showed that the spiders of this genus are the main prey of these ants in reserves of the “Biological Dynamics of Forest Fragments” project, covered mostly by dense forests on latosol, about 100 km North of Reserva Ducke. These ants hunt in groups of thousands of individuals and can substantially reduce the number of spiders in the areas through which they pass. Gasnier & Höfer (2001) showed that the probability of a spider to be attacked is 92 % during 3 months in dense forest areas and 21% in heath forests. Therefore, its effect on the populations of Ctenus should be fairly large, at least in the first habitat.

Based on these data, we suggest the following integrated model for coexistence within the genus Ctenus: C.manauara and C. crulsi probably have both a greater ability to capture Syntermes termites. This advantage would be the cause of their greater abundance in dense forests, where these termites are more abundant. However, predation and/or some factor directly or indirectly related with the seasonal rainfall could be keeping the abundances of these two species below a level that would allow one of these species exclude the other two. This factor could be, for example, the excess of accumulated rainfall or variation in the thickness of litter caused by climatic seasonality, which causes a seasonal variation in the amount of prey and refuges. Therefore, the existence of C. amphora and C. villasboasi in dense forests depends on a factor that limits the other two species. We believe that the low frequency of C. crulsi and C. manauara in heath forests is also related to its ability of these species to capture Syntermes. Since Syntermes is much less abundant in heath forest, individuals of these species could actively avoid these locations. In the near absence of C. crulsi and C. manauara, the other two species, C. amphora and C. villasboasi could reproduce more, due to the lower pressure of competition and intraguild predation. Therefore, unlike the previous situation, we propose that the abundance of the dominant species, C. amphora, would not have been the key factor that determined the low density of the other species.

The most important factors for the coexistence of C. crulsi and C. manauara are less clear. The two differ greatly in adult sizes. Being bigger, C. crulsi tend to perform a higher intraguild predation and exclude C. manauara. Moreover, there are two species of Syntermes more abundant in the Ducke Reserve, S. molestus and S. spinosus, the latter being larger, and it might be more difficult to be captured by a smaller spider. Possibly, the smaller species C. manauara has a shorter life cycle and a fast reproduction, which could enable them to rapidly recolonize after a population reduction (affecting both species) by climatic factors or by army ants.

Ctenusvillasboasi is the largest species of the genus in Reserva Ducke, and also has a lower density of adults. Although their adults are too large to be preyed upon by other species of Ctenus, the life cycle should be longer and the risk to die before reaching adulthood higher, i.e. by predation of the young by other Ctenus. It is unclear whether C. villasboasi has some strategies to reduce this predation pressure. There is evidence that this species is more abundant in the swamp forest, where the abundance of the other Ctenus is lower, however, while predation by Ctenus may be lower there, predation by Ancylometesrufus may increase, as we will discuss bellow.

The two species of Phoneutria, P. fera and P. reidyi are among the largest spiders in Reserva Ducke. Both are found almost exclusively in the vegetation when they are small juveniles. Only after they grow relatively large they can be found foraging both on the ground and on the vegetation (Sanchez-Torres 2000). We suggest that this behavioral change during development may have originated as a way for juveniles to avoid a higher risk of predation in the litter, and one of the main predators to be avoided would be the spiders of the genus Ctenus, due to their abundance. Evidence for this comes from the size of the juveniles which begin to descend to the ground, which is about the size of the largest Ctenus, and from this size they cease to be potential predators and become potential prey of Phoneutria. Other evidence comes from the observation that juveniles of Phoneutriaboliviensis were common on the ground in wetlands of the Solimões River, where Ctenus were in short supply. Sanchez Torres (2000) suggests that P. reidyi, which is a little smaller, needs palms as safer place of encounter between the sexes to avoid predation by P.fera, and this would be the cause of the first species to be rare in places with few palms. Furthermore, there is evidence that P. reidyi has seasonal reproduction, which can also be a mean of reducing predation by first, limiting the risk to a single season. Therefore, we advocate the hypothesis that the interaction between species of the family was a key factor in the evolution of this neotropical genus. We speculate that due to its size, only vertebrates would remain as important predators of adult spiders of this genus, and this would be the cause of their aggression and the reason for their venom being so potent to mammals.

The three species of Centroctenus are generally found close to burrows or natural shelters, with the exception of adult males which are wandering. Centroctenus auberti, the species with larger individuals, can be found in burrows in logs or in soil, possibly dug by themselves. Centroctenus ocelliventer has an intermediate size and often uses natural refuges among the leaves accumulated at the bases of palms with subterranean stems or clumps of branches of fallen trees or logs, and generally near the ground. Centroctenus acara is found in locations similar to those used by C. ocelliventer, but are more often in logs at a higher position, using even crevices in the trunks as refuges. The use of refuges can be efficient to avoid many predators, but, depending on the nature of the refuge, can be inefficient against army ants, so that it may be important to have a way to escape by climbing to the vegetation. Possibly there is a dispute between them by the use of the best refuges; however, the density of these spiders is apparently not high enough to consider this interaction as a key factor in the biology. We believe that, because of the use of burrows, the interactions with other species of the family may be relatively small.

Cupiennius celerrimus [synonym of C. bimaculatus] is the only species of the genus in South America. It is a sedentary species found on low vegetation, usually on small plants with large leaves. We did not do regular counts of this species, but we noted that it has a seasonal reproductive cycle, as there were moments of greatest abundance of young and others with larger abundance of adults. There should be a low risk of predation by other ctenid spiders due to this sedentary behavior. However, males of this genus often need to move between plants in search of females (Schmitt et al. 1990), which probably represents a great risk of predation by Ctenus. We believe that this seasonality may be a way to optimize reproductive meetings, reducing the risk of predation by restricting the reproductive season. Apparently the species from Central America show no seasonal reproduction, which could be related to a lower abundance of Ctenus, but this remains to be investigated.

Aglaoctenus castaneus (Lycosidae) build sheet webs in the vegetation; when young and small the spiders build a small web without refuge close to the ground (30 cm); as they grow webs gradually increase in size and position (usually between 0.5 and 1.5 meters above ground), and then include a refuge. Spiders often use palm trees with spines rich in structures. They are relatively abundant and apparently more common on sandy than on clayey soil. The species show seasonal reproduction, many small spiderlings found in webs in July 1995, reaching maturity in October and November. As for Cupiennius, the risk of predation on juveniles and females by ctenid spiders should be almost zero, but males need to move to find the females, and the risk of this displacement could also be the cause of the well marked reproductive seasonality. It is the only species of the spider family Lycosidae found in this forest, but it is also interesting to note that there are species of this family living in open areas under anthropogenic influence, like the administration buildings and at the margins of the forest reserve. While Ctenus rarely move more than a few meters from the forest edge, lycosids never penetrate the undisturbed forest.

There are two species of Ancylometes (Ctenidae) in Reserva Ducke, A. rufus and A. terrenus. These are large spiders hunting on the ground. The first, as most species of the genus (Höfer & Brescovit 2000), is almost always close to water bodies such as ponds and streams. It is the predominant species in the swamp forests, but can also be found far from streams. Although they often use burrows as refuge, both males and females occasionally move hundreds of meters in a few days, (Azevedo 2000), so it is unlikely that they use the same burrow for a long time. They can dive into the water to escape predators and to capture prey. Their diet is more diverse than the diet of other wandering spiders, because in addition to arthropods and earthworms, they are known to consume small fishes and amphibians. Apparently, they migrate in the rainy season to small ponds in the forest, and return to the streams when these dry up, probably because it is safer. There must be a strong interaction between A. rufus and Ctenus as encounters of these species on the ground must be frequent; in fact, we recorded several predatory incidents. The size and the use of water bodies as areas protecting the spiders against predators, i.e. army ants and as a hunting ground make this interaction asymmetrical in favor of A. rufus, and possibly the relatively low frequency of Ctenus spp. in some swamp forest areas, especially near streams, is the result of high abundance of Ancylometes. Among Ctenus species, only C. villasboasi seems to be relatively more abundant in the swamp forest, but it remains unclear what are the factors that contribute to the coexistence of this species and Ancylometes. Ancylometesterrenus is relatively rare, lives in burrows dug into the ground, most likely by them. The holes should be effective refuges against Ctenus spp. Males are too large to be consumed by Ctenus spp. and probably Phoneutria spp. have a density too low to have a great influence on the population of this species.

The coexistence of genera seems to be mainly related to differences in habitat use that would decrease the chance of intraguild predation. The construction of webs by Aglaoctenus above the soil level, the use of residential burrows by Centroctenus spp. and Ancylometes terrenus, the use of vegetation by Cupiennius,Ctenusinaja and young Phoneutria and the preference for water bodies by Ancylometesrufus appear to be important strategies to prevent predation in general and intraguild predation in particular. Possibly, the main predators avoided by these habitat use strategies are the army ants and the abundant Ctenus spiders foraging on the ground.

The model of coexistence of these species is still under construction. However, it is a model more complex than we thought at the beginning, because it includes differences in habitat use, intraguild predation, adult sizes and life cycles, seasonal migrations between habitats, changes in microhabitat preferences during development, differences in the ability to capture certain prey, prey distribution in relation to soil types, seasonal effects and climate effects on key predators. Any model is a simplification of reality, but we are still restricted to the primary aspects of the biology of these species. Authors arguing that the experimental approach is the safest way to assess the coexistence of species recognize the importance of basic knowledge to plan and interpret experiments (e.g. Hairston, 1989 and Wise, 1993). However, it is possible that the time necessary to collect basic data before the experimental approach in the tropics may need longer than is usually admitted.

Further studies could help in the expansion and testing of this model of coexistence, for example: (a) Is there is an increase in the populations of C. amphora in an area in which C. crulsi and C. manauara are systematically removed? (b) Is there a spatial pattern of abundance in C. villasboasi and A. rufus indicating a division in habitat use in swamp forests? (c) C. manauara has a greater capacity for recolonization after the passage of army ants that C. crulsi, and after some time the arrival of the second species influences the abundance of first? (d) The size of adult C. amphora in heat forest is higher than in dense forests on latossol, and the opposite occurs with C. crulsi, indicating that each species has a better development in the habitat in which they are more abundant? (e) What is the factor that determines the seasonal variation in the abundance of Ctenus spp.? ; (f) Is there a difference in the spatial patterns of abundance of Centroctenus? (g) What are the indications that seasonal reproduction of P. reidyi, Cupiennius celerrimus, Ctenus villasboasi and Aglaoctenus castaneus could be an adaptation to reduce predation risk? And after they answered, others will surely arise. The importance of models like this is to raise the possibility of interactions between different factors on coexistence, to give a direction for testing, and to reveal factors not usually considered important for coexistence. Therefore, they are important tools for understanding factors that contribute to the high diversity in the tropics.

The development of integrated models of coexistence of arthropods, important for understanding the issue of biodiversity is only possible in places where the fauna is relatively well known and where data can be collected natural history and tests carried out intensively. There are few places like the Reserva Ducke, which provide these conditions in tropical areas. This is, therefore, one more item to the list of the importance of this reserve to study the Amazon.

see Reference page of Ctenidae