Neotropical tree species and their faunas of Xylophagous longicorns (Coleoptera: Cerambycidae) in French Guiana.
Tropical rain forests are the repository for much of the Earth's biological diversity. Although a mere 7% of the land surface supports tropical rain forests (Wilson, 1988), they are home to a disproportionate number of the world's plant and animal species. Projections have been made that, given current rates of tropical deforestation, most of these forests will have been either cleared or significantly degraded by 2135 A.D. It is thus presumed that we are in the midst of a mass extinction event unparalleled within the last 65 million years (Wilson, 1988). The magnitude of this event cannot be described without a clear understanding of the rate of habitat loss, which has been the subject of varying estimates (Lugo, 1988), and the relationship between geographic area and the number of species that can be supported (Wilson, 1988). Other poorly understood variables include the relative species richness of different types of tropical forest (Lugo, 1988) and the distribution ranges of tropical forest species (Wilson, 1988), the majority of which are insects.
Even should data permit the calculation of credible extinction rates (see Lugo, 1988, for a summary), it would be difficult to generate reliable figures predicting how many species might be lost, because we lack baseline documentation of many species that currently exist. Current high-end estimates of the potential number of extant species (30-50 million) were formulated on the basis of the results of experiments in which insecticides were used to knock insects down from the canopies of tropical trees (Erwin, 1982, 1988). These estimates have engendered great controversy, in part due to the assumption that many tropical arthropods are dependent on a single host plant for survival (Erwin, 1982). Subsequent studies have compared the insect faunas of conspecific trees with those of unrelated tree species and have suggested that Erwin may have overestimated the host fidelity of tropical insects (Kitching et al., 1997; Mawdsley & Stork, 1997). These studies, also based on canopy-fogging experiments, do not adequately sample concealed feeders such as leaf-miners and bark- or wood-borers. In addition, due to the chaotic profusion of arthropod species and individuals harvested after the release of insecticides, very few trees can actually be sampled. The state of knowledge regarding the host specificity of tropical insects is still rudimentary at best.
This paper presents the results of an extended study of cerambycid beetles and their associated host plants in French Guiana, undertaken by the first author. French Guiana is part of the Guayana floristic province in northeastern South America (Mori, 1991), a species-rich region with between 7000 and 10,000 angiosperms (Lindeman & Mori, 1989). Although numerous forest types are represented, including mangrove, marsh, swamp, and montane forest, much of French Guiana is covered with intact, and relatively well-investigated, seasonal evergreen forest (Granville, 1986).
Cerambycids (longicorns, long-horned beetles, timber beetles) play an important role in the recycling of nutrients in the forest. The females of most species oviposit their eggs into freshly killed or damaged wood with persistent bark that protects the immature stages of the beetles (Linsley, 1959, 1961). After about a week, the eggs hatch into larvae which, feeding on the wood, gain weight and pass through five or six larval instars (Hequet & Tavakilian, 1996). As they feed, the larvae create systems of tunnels and galleries throughout the wood, which may then be colonized by other animals or microorganisms. The larvae eventually excavate pupal chambers underneath the bark. After metamorphosis, the adult beetles chew exit holes through the bark of the host and commence the search for mates and appropriate host plants. The feeding habits of the rather short-lived adult longicorns can be quite variable. They may feed on numerous plant parts (Linsley, 1959, 1961) not necessarily belonging to the host plant in which the larvae developed. Therefore, when we discuss the host specificity of longicorns, we refer to reproductive-host specificity.
Cerambycidae is one of the largest insect families and is particularly diverse in the tropics. As recently as 1983, only 330 cerambycid species had been recorded in French Guiana, but well over 1400 species have now been documented, including 600 that are currently being described (Hequet & Tavakilian, 1996; Tavakilian, in prep.). Although metropolitan France occupies an area more than five times as large as French Guiana, French Guiana supports 10 times as many tree species. France has a predictably diminished cerambycid fauna of only 235 species (Tavakilian, 1993), and fewer than 1000 species are found in all of the United States and Canada (Arnett, 1988).
The only way to provide incontrovertible documentation of host-plant association is to actually rear an insect from an accurately identified host. More than 1000 host-plant records have been established for the cerambycids of French Guiana via the arduous process of rearing adult beetles from freshly fallen wood. Host-plant associations for 348 cerambycid species reared from more than 200 tree species are presented in Table I. While the host-plant associations for the 235 cerambycid species found in France were revealed over a period of 200 years (Tavakilian, 1993), most of the data presented here were assembled between 1991 and 1993. During this time, a massive study was conducted to document the biodiversity of an area of the Sinnamary River Basin subsequently inundated by a reservoir formed behind the Petit Saut Dam. Table I also incorporates the results from additional rearing experiments in French Guiana, as well as selected literature references. We propose a quantitative system to designate the reproductive-host specificity of cerambycid species reared, and discuss patterns of host utilization by beetle guilds associated with well-represented plant taxa.
III. Materials and Methods
A. SINNAMARY RIVER BASIN STUDY SITE
Most of the beetle-plant interactions reported on here were obtained from trees felled on the Sinnamary River in an area now inundated by the reservoir formed after the closure of the Petit Saut Dam.
The Sinnamary River Basin, covering 6000 [km.sup.2], is located in north central French Guiana. Because the Sinnamary Basin encompasses little relief and does not include savannas, the species richness of the basin is not as great as it is in areas of comparable size in other parts of French Guiana (Hoff, 1994). Variation in forest vegetation in the Sinnamary Basin is dictated by soil moisture, ranging from swamp forests dominated by Euterpe oleracea Martius at some places along rivers and streams to mixed forest on well-drained soil. The mixed forest does not differ fundamentally in structure and composition from this type of forest in other parts of French Guiana (Hoff, 1994, 1996).
Annual rainfall is usually less than 3000 mm. However, some areas may receive as little as 2000 mm and others as much as 3600 min. There are two dry seasons each year, a long one from July to November and a short one in February and March. Winds are light throughout the year (Hoff, 1994).
Trees were felled at two localities in close proximity to one another. One study was located at the mouth of Crique Plomb and the other was situated between Crique Plomb and Crique Tigre at about 500 meters above Saut Tigre. The former site was subject to periodic inundation, whereas the latter site was situated at roughly 80 meters above the river and was not subject to inundation before the closure of the dam.
B. REARING EXPERIMENTS
Between 1991 and 1993, 690 trees and lianas were felled (a complete list of these can be made available upon request). The cuts were made during different parts of the dry season (July-November), when the adults of most cerambycid species are active. The plants belonged to 38 plant families. Abundant plant taxa included Moraceae, Malvales, Lecythidaceae, Sapotaceae, Fabales, and Apocynaceae. Although these taxa were represented by numerous samples, the selection of individuals was not random. Rare tree species and lianas were favored because it was anticipated that they might give rise to poorly known cerambycid species.
[TABULAR DATA FOR TABLE I OMITTED]
Voucher specimens were collected from the plants, identified by specialists, and deposited in major herbaria (NY, P, CAY). Collections in the 1991 and 1992 field seasons were made by Denis Loubry (collection numbers preceded by "L" in Table I), and collections in 1993 were made by Scott Mori (collection numbers preceded by "M" in Table I). Wood specimens were also collected and deposited at CTFT (the French research organization responsible for the study of wood) and the United States Forest Products Laboratory wood collection at Madison, Wisconsin. During the 1992 and 1993 field seasons, additional small wood samples were collected, preserved in methanol, and delivered to the laboratory of Barbara Meurer-Grimes (Lehman College, CUNY) for chemical analysis.
The felled trees were left on the forest floor for a period of about four months, and during this interval cerambycids had the opportunity to lay their eggs. Several site visits were made after each cut, to observe and collect cerambycid visitors. After four months the dead trees were surveyed for signs of insect attack, such as oviposition scars or piles of ejected frass. Sections of trunk or thick branch (80 cm long) were removed, along with some thin twigs, and placed into cages. The cages were monitored twice a week, newly emerged adult beetles were identified, and representatives were drymounted (see Hequet & Tavakilian, 1996, for the preparation protocol). The cerambycid specimens are deposited at ORSTOM, Cayenne.
C. ADDITIONAL DATA
Data from earlier experiments at Sinnamary and additional experiments in other parts of French Guiana are also included in Table I. These data bring the total number of plant families investigated to 48. Branches were collected by Christian Feuillet, Daniel Sabatier, and MarieFrancoise Prevost (abbreviated as "F," "S," and "P," respectively, in Table I), and rearing experiments were conducted as described above. Certain cerambycid host-plant associations described in Duffy, 1960, confirm or supplement results from the rearing experiments, and are therefore included in Table I.
IV. Results and Discussion
Table I provides a list of plant specimens yielding beetles, collection voucher numbers of the plants, associated species of cerambycids, and the tribes to which the beetles belong. Plant families are listed following the nomenclature and systematic concepts of Cronquist (1981). For each plant species, the associated cerambycid species are listed following Monne and Giesbert (1994), with tribes arranged to reflect presumed phylogeny, but with genera and species arranged alphabetically. Longicorn species that have not yet been described (n = 90, [greater than]25% of the species reared) are listed under their working numbers to facilitate identification for the analysis of host specificity. These species will be described in a forthcoming monograph (Tavakilian, in prep.). Collections that did not originate in the Sinnamary River Basin are marked with an asterisk, and entries derived from the literature are thus noted.
A. CLASSIFICATION OF HOST SPECIFICITY
An additional column in Table I proposes a hypothetical host utilization strategy for each cerambycid species. Herbivores may be broadly classified as generalists (those which attack numerous unrelated plant species) or specialists (those which attack a limited set of plants, usually taxonomically related, but also sometimes unrelated plants that produce similar chemicals). The terms "oligophagous" and "monophagous" have been used to describe specialists showing progressively greater fidelity in their choice of host plant(s), but unfortunately these terms have been used with little consistency (see summary in May & Ahmad, 1983).
To facilitate the interpretation of the data in Table I, we have used an alternate list of emergences (Tavakilian, unpubl.data), which records all documented host plants of each cerambycid species, to classify the host fidelity of each longicorn species reared. Generalists are denoted "G," and specialists are categorized according to the taxonomic level of the host plant(s) utilized. "S/ORD" indicates that a beetle species is associated with plants belonging to a particular plant order, and "S/FAM," "S/GEN," and "S/SP" refer to beetle species reproductively restricted to a particular plant family, genus, or species, respectively. There is, in addition, a small group of longicorns which reproduce in trees that are not taxonomically related but that belong to families characterized by the production of milky latex. These beetles are considered chemical specialists, and classified "S/LAT."
A longicorn has been tentatively assigned to a particular category if there are at least two host-plant records directly derived from rearing experiments in French Guiana. When there are numerous host plant records, 90% of them must be in accord for a beetle to be retained in a category. For instance, ira cerambycid species has been reared from nine plant specimens belonging to the same genus and only once from an alternate genus, that species is considered "S/GEN," or a specialist associated with the first genus. However, if the cerambycid species has been reared from eight plant specimens belonging to a single genus and twice from an alternate genus in the same plant family, that beetle is considered "S/FAM." Should the cerambycid species be reared eight times from plant specimens belonging to a single genus and twice from specimens belonging to a different plant family, that beetle would be downgraded to "G." These classifications are considered hypothetical, because many of them are supported by very little data, but they nevertheless enable us to perceive meaningful trends in host-plant fidelity.
Many cerambycid species have not been classified as either generalists or specialists, because even after the intensive sampling in the Sinnamary River Basin, they are represented by a single host-plant record. In Table I these beetles are classified "ISD," for insufficient data. This is actually the largest class [ILLUSTRATION FOR FIGURE 1 OMITTED], because almost 47% (n = 163) of the cerambycid species reared are represented by a single host-plant record! These beetles might conceivably be rare specialists that are principally associated with plant taxa that were not sampled, or they might optimally reproduce in living wood, wood at a different stage of decompositon, or at a different time of the year.
Fewer than 13% (n = 44) of the cerambycid species were classified as generalists [ILLUSTRATION FOR FIGURE 1 OMITTED]. Many of the beetles in this category have numerous host-plant records confirming successful reproduction in trees belonging to many different unrelated plant families. Some examples are Chlorida festiva [ILLUSTRATION FOR FIGURE 2c OMITTED], Macropophora trochlearis, Mecometopus triangularis, Nyssodrysternum signiferum, Oreodera bituberculata, and Toronaeus virens. Some cerambycids classified as generalists are disproportionately represented in association with a particular plant taxon, but fewer than 90% of the records are from that taxon. Species such as Macronemus antennator and Oedopeza ocellator are associated primarily with legumes, which have been extensively sampled, but also reproduce in some unrelated trees. Other longicorns, such as Chrysoprasis chlorogaster, are represented by only two host records from unrelated plants. Pending additional data, they are also provisionally classified as generalists.
The remaining cerambycid species reared ([greater than]40%; n = 14]) display a higher degree of host specificity and usually reproduce in a group of taxonomically related plants. When all plant taxa are included, specialists outnumber generalists [greater than]3:1, but different plant taxa give rise to cerambycid faunas with quite different ratios of specialist to generalist [ILLUSTRATION FOR FIGURE 1 OMITTED]. In addition, the highest taxonomic rank of the selected host plant(s) differs from one plant taxon to the next. When all plant taxa are included, more than one-half of the specialists are associated with a particular plant family, and fewer than one-quarter are associated with a particular plant genus. Specialist cerambycids regularly recognize only two groups of plants at the ordinal level: Fabales and Malvales. This might be an indication either of little chemical divergence among the families composing the order (not likely to be the case for the legumes but possible for Malvales) or of confusion in the current taxonomic concept. Other closely related pairs of plant families, such as Moraceae/Cecropiaceae and Burseraceae/Anacardiaceae, support longicorn guilds that are largely independent. Of all 348 cerambycid species reared, only 13 species ([less than]4%) appear to be reproductively dependent on a single plant species. Eight of these associations are supported by only two host records, and might not be confirmed if sampling was continued!
B. BEETLE GUILDS OF ABUNDANTLY REPRESENTED PLANT TAXA
The plant taxa treated in Fig. 1 (Moraceae, Malvales, Lecythidaceae, Sapotaceae, Fabales, and Apocynaceae) were among those especially well represented in the Sinnamary River Basin, in terms of both numbers of species and individuals present. These and other abundant plant taxa were visited by large and often diverse beetle guilds (see [ILLUSTRATION FOR FIGURE 2 OMITTED] for representative cerambycids). These guilds are usually composed of numerous taxonomically unrelated cerambycids, although most guilds do include a few genera represented by several species. Cerambycid guilds associated with some of these plant taxa are discussed below.
The 36 plant specimens (approximately 16 spp., 8 genera) investigated gave rise to 32 cerambycid species. The host specificity of almost one-third of these species (31%) cannot be determined, due to insufficient data, and the remaining species are almost evenly divided between specialists (38%) and generalists (31%). Half of the specialists were associated with the genus Brosimum, represented by 21 plant specimens. These beetles were Alphus malleri, Alphus sp. 1366, Nyssodrysternum caudatum, N. fiavolineatum, N. simulatum, and Toronaeus magnificus, all of which belong to the related tribes Acanthoderini and Acanthocinini in the subfamily Lamiinae. These large tribes are composed of small to medium-sized, cryptic, nocturnal beetles. Two beetle species were specialists at the family level, three species were associated with several plant families characterized by milky latex, and Alphus aurivilli was the sole species reared from both Moraceae and the closely related family Cecropiaceae. In the past, some authors have treated the Cecropiaceae as part of the Moraceae (e.g. Takhtajan, 1969), but in our study Cecropiaceae gave rise to an independent longicorn guild including two species of Drycothaea, Anisopodus phalangodes, and Nyssodrysina binoculata.
The Malvalean families Eleaocarpaceae, Tiliaceae, Sterculiaceae, Bombacaceae, and Malvaceae were represented by 29 plant specimens (at least 12 spp., 8 genera) which gave rise to 33 cerambycid species. Of the longicoms associated with this plant group, almost 24% of the species had only a single host record, and 53% were generalists. There were eight specialists, but half of these were associated with more than a single family in the order. Three specialists emerged solely from Catostemma fragrans, belonging to the Bombacaceae. This tree species, represented by six specimens, not only is relatively common but, if the name is a clue, also may be particularly easy for the beetles to locate. Once again, the majority of the specialists belong to the tribes Acanthoderini or Acanthocinini.
We noted that Steirastoma breve, S. melanogenys, and Lepturges sp. 710 all emerged from trees belonging to both the Bombacaccae and the Sterculiaceae. There are, in addition, several literature references (Duffy, 1960) reporting an association between Steirastoma breve and the Malvaceae. A fourth cerambycid species, Oreodera undulata, emerged from trees belonging to the Sterculiaceae and the Tiliaceae. Only Eleaocarpaceae, represented here by numerous undetermined trees belonging to the genus Sloanea, seemed to lack convincing evidence for a close relationship with the other families constituting the Malvales.
To our surprise, we recently learned that a two-step parsimony analysis on 125 rbcL sequences resulted in trees that strongly support the monophyly of the core malvalian families (Bombacaceae, Malvaceae, Sterculiaceae, and Tiliaceae), but that "Eleaocarpaceae, hypothesized by most authors as a sister group to the four core malvalean families, is shown to not fall close to these taxa" (Alverson et al., in press). It was gratifying to find the beetles and the molecules in accord!
The 41 plant specimens investigated (19 spp., 4 genera) gave rise to 17 cerambycid species. Lecythidaceae was one of two plant families that not only gave rise to an unusually well-defined guild of specialists but also was conspicuously avoided by generalists. Five of the cerambycid species were represented by a single host record, although some of these longicorns are closely related to beetles that are clearly Lecythidaceae specialists. Of the remaining cerambycid species, only one was a generalist. Nine of the 11 specialist species emerged from two or more tree genera, including Neoeutrypanus incertus, Oedopeza leucostigma, Palame spp., Periboeum pubescens, Xylergates elaineae, and Xylergatina pulchra. All of these belong to the tribe Acanthocinini, with the exception of P. pubescens, which belongs to the tribe Elaphidionini in the subfamily Cerambycinae. The Lecythidaceae guild is currently the subject of a more detailed study (Berkov & Tavakilian, in prep.; Berkov et al., in prep.).
The 73 plant specimens investigated (at least 27 spp., 6 genera) gave rise to 36 cerambycid species. The beetle guild associated with Sapotaceae showed certain patterns reminiscent of the Lecythidaceae guild, although the species composition was totally different. The majority of cerambycids that emerged from Sapotaceae were also family-level specialists, and once again generalists were rare. Although several family-level specialists belong to the subfamily Lamiinae, the majority belong to the subfamily Cerambycinae. These include four species of Tomopterus [ILLUSTRATION FOR FIGURE 2B OMITTED], Acorethra zischkai, Ischasia sp. 926, and Ommata sp. 311 (all belonging to the tribe Rhinotragini), as well as Callichroma auricomum [ILLUSTRATION FOR FIGURE 2i OMITTED] and three species of Mionochroma (belonging to the tribe Callichromatini). In addition, Callichroma velutinum was reared from Manilkara, and three species of Tomopterus were associated with particular plant species. Two additional cerambycid species were associated with plant families characterized by the production of milky latex.
The tribes Rhinotragini and Callichromatini both comprise conspicuous diurnal species with flower-visiting adults. Many species belonging to Rhinotragini are brightly colored, but with reduced elytra, and mimic wasps or predatory flies. Callichromatini is composed of beautiful iridescent beetles, all of which emit a distinctive musky aroma. The Sapotaceae guild is composed largely of aposematic beetles, and the association between these longicorns and their chemically distinctive host plants certainly merits further attention.
The legumes - including Caesalpiniaceae, Fabaceae, Mimosaceae, and the tribe Swartzieae (commonly treated as a member of either Caesalpiniaceae or Fabaceae, and treated as Fabaceae for the classification of host specificity in Table I) - are without doubt the most abundantly sampled plants. The 141 plant specimens (at least 73 species, 38 genera) gave rise to an astounding 123 cerambycid species. No host-utilization strategy can be proposed for the 43% of these species which were represented by a single host record. Almost 19% of the cerambycids reared were generalists. The remaining 38% of the cerambycid species were specialists.
The longicorns associated with the legumes have been the subject of additional study (Meurer-Grimes & Tavakilian, this issue). The legumes were the second plant group with a substantial number of specialists (14 out of 47) that reproduced in the trees of more than one family. In addition to these less demanding species, it has been determined that there are actually a number of different longicorn guilds, each associated with only a few plant genera.
The tribe Swartzieae has its own distinct beetle guild, including Agaone notabilis, Cycnidolon approximatum, and Odontocera molorchoides. Considering the disputed taxonomic status of the tribe Swartzieae, it is interesting to note that its longicorn guild shows some resemblance to guilds associated with both the Caesalpiniaceae and the Fabaceae. Odontocera sp. 1018 was regularly associated with trees belonging to the Swartzieae but also emerged from several trees belonging to the Caesalpiniaceae. Additional beetles belonging to the genus Odontocera emerged in abundance from trees belonging to the tribe Dipteryxeae (Fabaceae). The generalist cerambycid Colobothea hirtipes has a broad host range, but within the Fabales it emerged from trees belonging to the Swartzieae and the tribe Dalbergieae (Fabaceae).
Several distinct guilds can be recognized within Caesalpiniaceae and Fabaceae. Only the Mimosaceae appears to serve a more or less consistent guild of longicorns, including Chrysoprasis moerens, Hemilissa catapotia, Thoracibidion ruficaudatum, and T. striatocolle.
Legume specialists belong to a great diversity of cerambycid tribes. The subfamily Cerambycinae is represented by tribes including Eburiini, Elaphidionini, Piezocerini, Ibidionini, Rhinotragini, and Heteropsini. The subfamily Lamiinae is represented by tribes including Onciderini, Desmiphorini, Polyraphidini, Acanthoderini, Acanthocinini, and Colobotheini. There is a predictable mix of diurnal and nocturnal, cryptic and aposematic species.
The 26 plant specimens investigated (15 spp., 9 genera) gave rise to 17 cerambycid species. The guild was almost evenly divided between generalists and specialists, but in this case four of the specialists were cerambycids associated with various families producing milky latex, and only four longicorn species were restricted to Apocynaceae. The specialist which generated the most host records was Hemilissa cornuta [ILLUSTRATION FOR FIGURE 2H OMITTED], belonging to the tribe Piezocerini and associated with the plant genus Aspidosperma. Although the congener H. opaca was classified as a generalist, it was also most frequently associated with Aspidosperma.
7. Latex-Producing Plants
Three of the plant families discussed above are characterized by the production of milky latex: Moraceae, Sapotaceae, and Apocynaceae. Euphorbiaceae, another latex-producing family, was also investigated for its cerambycid fauna. An ecological study of the forest of La Fumee Mountain in central French Guiana revealed that 42.7% of the trees had a distinctive latex, resin, or sap (Mori & Boom, 1987). Given the number of specimens collected from latex-producing taxa at Sinnamary, they must also have been quite abundant at the project site.
There was a small guild of longicorns which utilized hosts from two or more of the above families. Acrocinus longimanus was associated most frequently with various Moraceae but was also reared from Apocynaceae. Hylettus coenobita was associated primarily with Moraceae but was also reared from Apocynaceae and Euphorbiaceae. Nealcidion triangulare had one host record from Euphorbiaceae and a second from Apocynaceae. Nyssodrysina pulchella emerged from Moraceae, Euphorbiaceae, and Apocynaceae. Oreodera albata and O. basiradiata emerged from both Sapotaceae and Euphorbiateae. These beetle species were associated with no other plant families. Although the chemical constituents of the latices produced by the plants in the above families may be variable (Cronquist, 1981), they apparently provide the requisite cues to longicorns seeking oviposition sites.
Specialization on latex-producing plants does not appear to be restricted to cerambycids. Caterpillars belonging to 17 species were documented feeding on the leaves of Manilkara chicle, a tree belonging to the Sapotaceae which is commonly found in the Santa Rosa National Park in Costa Rica. Although this appeared to be the sole host plant of most of the caterpillars, the sphingid Erinnyis ello was reported to feed upon three or more latex-rich tree species (Janzen, 1988).
It has been suggested that cerambycids, as wood-boring beetles, would not be especially prone to develop host-specific relationships with plants (Bassett, 1992; Bassett et al., 1996). The longicorns investigated in this study lay their eggs in freshly fallen wood, and the longest and most vulnerable part of the life cycle is spent, as larvae, in intimate contact with host tissues. When a host-utilization strategy can be determined, specialists outnumber generalists by more than 3:1, although this ratio is very different from one plant taxon to the next. Specialists are, however, seldom associated with a single plant species; rather, they are most frequently able to locate and successfully reproduce in the wood of related species. Extreme specialism might be a more reliable strategy in temperate ecosystems, typically dominated by fewer species represented by more individuals (Bassett, 1992, and references therein; Beaver, 1979; Hammond, 1992).
The fact that some longicorn species are associated with unrelated latex-producing plants suggests that plant chemistry plays an important role in determining the range of suitable host plants for cerambycids. Intrafamilial variability in plant chemistry also appears to have a profound influence on beetle attack. For instance, certain members of the Lecythidaceae produce putrid-smelling compounds which appear to act as oviposition deterrents to Lecythidaceae specialists. The possible influence of chemical characteristics on host specificity in longicorns is the subject of two additional studies (Berkov et al., in prep.; Meurer-Grimes & Tavakilian, this issue). Abundantly represented plant taxa are associated with cerambycid guilds composed of both related and unrelated cerambycid species, indicating that, in most cases, specialization on a particular plant group has arisen repeatedly. These guilds are sometimes extremely well defined, and on at least one occasion the beetles reared from a log provided the clues required to identify a sterile herbarium voucher! Careful analysis of the longicorn guilds could also provide useful data for taxonomic and phylogenetic studies of tropical plants at both the intra- and interfamilial levels.
This study supported the observation that when insects are sampled, relatively few insect species are abundantly represented and most insect species are quite sparsely represented (Elton, 1975). The implication is that there are many insect species that we simply haven't learned how to locate, or that many species are able to persist at what seem to be impossibly low population densities.
Forest fragments left in the wake of tropical deforestation are unlikely to retain a full complement of the species present within intact forests. Relict populations within isolated tracts of forest would face increased competition for dwindling resources. These populations, which could not be replenished via recolonization from adjacent forest, would find it difficult to rebound from periods of adverse conditions. Tropical insects with narrow host ranges would be at a particular disadvantage. Those dependent on a single host plant would face local extinction if their host was eliminated or its abundance too drastically reduced. Most tropical cerambycids do not appear to have the excessively narrow reproductive/larval feeding niches that might make them particularly vulnerable to disruption and habitat loss.
We thank Electricite de France for funding the Sinnamary River project. We are also grateful to Christian Feuillet, Daniel Sabatier, and Marie-Francoise Prevost for additional wood collections used in rearing experiments. We appreciate William Alverson's contribution regarding the phylogeny of the plant families constituting the Malvales, and Lee Herman's comments on the manuscript.
VII. Literature Cited
Alverson, W. S., K. G. Carol, D. A. Baum, M. W. Chase, R. McCourt & K. J. Sytsma. In press. Circumscription of the Malvales and relationships to other Rosidae. Amer. J. Bot.
Arnett, Ross H. 1988. Present and future of systematics of the Coleoptera in North America. Pp. 165-173 in M. Kosztarab & C. W. Schaefer (eds.), Systematics of North American insects and arachnids: status and needs. Virginia Polytechnic Institute and State University, Blacksburg, Virginia.
Basset, Y. 1992. Host specificity of arboreal and free-living insect herbivores in rainforests. Biol. J. Linn. Soc. 47: 115-133.
-----, G. A. Samuelson, A. Allison & S. E. Miller. 1996. How many species of host-specific insects feed on a species of tropical tree? Biol. J. Linn. Soc. 59: 201-216.
Beaver, R. A. 1979. Host specificity of temperate and tropical animals. Nature 281: 139-141.
Cronquist, A. 1981. An integrated system of classification of flowering plants. Columbia University Press, New York.
Duffy, E. A. J. 1960. A monograph of the immature stages of neotropical timber beetles (Cerambycidae). British Museum (Natural History), London.
Elton, C. L. 1975. Conservation and the low population density of invertebrates inside neotropical rain forest. Biol. Conservation 7: 3-15.
Erwin, T. R. 1982. Tropical forests: their richness in Coleoptera and other arthropod species. Coleopterists's Bull. 36: 74-75.
-----. 1988. The tropical forest canopy: the heart of biotic diversity. Pp. 123-129 in E. O. Wilson (ed.), Biodiversity. National Academy Press, Washington, DC.
Granville, J.-J. de. 1986. Flore et vegetation. Saga, Cayenne, French Guiana.
Hammond, P. M. 1992. Species inventory. Pp. 17-39 in B. Groombridge (ed.), Global biodiversity, status of the Earth's living resources: a report compiled by the World Conservation Monitoring Centre. Chapman & Hall, London.
Hequet, V. & G. Tavakilian. 1996. Longicornes de Guyane. Silvolab, ORSTOM, Cayenne, French Guiana.
Hoff, M. 1994. Biodiversite floristique d'un bassin fluvial tropical: le Sinnamary (Guyane Francaise). Ecologie 25:189-200.
-----. 1996. Les berges du fleuve Sinnamary (Guyane Francaise): geomorphologie et groupements vegetaux. Hydroecol. Appl. Tome 7 Vol. 1-2: 151-183.
Janzen, D. H. 1988. Ecological characterization of a Costa Rican dry forest caterpillar fauna. Biotropica 20: 120-135.
Kitchling, R. L., H. Mitchell, G. Morse & C. Thebaud. 1997. Determinants of species richness in assemblages of canopy arthropods in rainforests. Pp. 131-150 in N. E. Stork, J. Adis & R. K. Didham (eds.), Canopy arthropods. Chapman & Hall, London.
Lindeman, J. C. & S. A. Mori. 1989. The Guianas. Pp. 376-390 in D. G. Campbell & H. D. Hammond (eds.), Floristic inventory of tropical countries. The New York Botanical Garden, Bronx.
Linsley, E. G. 1959. The ecology of the Cerambycidae. Annual Rev. Entomol. 4: 99-138.
-----. 1961. The Cerambycidae of North America: Part I. Introduction. University of California Press, Berkley.
Lugo, A. E. 1988. Estimating reductions in the diversity of tropical forest species. Pp. 58-76 in E. O.
Wilson (ed.), Biodiversity. National Academy Press, Washington, DC.
Mawdsley, N. A. & N. E. Stork. 1997. Host-specificity and the effective specialization of tropical canopy beetles. Pp. 104-130 in N. E. Stork, J. Adis & R. K. Didham (eds.), Canopy arthropods. Chapman & Hall, London.
May, M. & S. Abroad. 1983. Host location in the Colorado potato beetle: searching mechanisms in relation to oligophagy. Pp. 173-179 in S. Ahmad (ed.), Herbivorous insects: host-seeking behavior and mechanisms. Academic Press, New York.
Meurer-Grimes, B. & G. Tavakilian. 1997. Chemistry of cerambycid host plants, part I: survey of Leguminosae - a study in adaptive radiation. Bot. Rev. (Lancaster) 63: 356-394.
Monne, M. A. & E. F. Giesbert. 1994. Checklist of the Cerambycidae and Disteniidae (Coleoptera) of the Western Hemisphere. Wolfsgarden Books, Burbank, California.
Mori, S. A. 1991. The Guayana lowland floristic province. C. R. Soc. Biogeogr. 67: 67-75.
----- & B. M. Boom. 1987. The forest. In S. A. Mori & collaborators, The Lecythidaceae of a lowland neotropical forest: La Future Mountain, French Guiana. Mere. New York Bot. Gard. 44: 9-29.
Takhtajan, A. 1969. Flowering plants: origin and dispersal. Oliver and Boyd, Edinburgh.
Tavakilian, G. 1993. Etude exhaustive des insectes xylophages de la famille des Cerambycides (Coleoptera) sur un site amene a disparaitre lots de la mise en eau du barrage de Petite-Saut. Progress Report for Electricite de France.
Wilson, E. O. 1988. The current state of biological diversity. Pp. 3-18 in E. O. Wilson (ed.), Biodiversity. National Academy Press, Washington, DC.
|Printer friendly Cite/link Email Feedback|
|Author:||Tavakilian, Gerard; Berkov, Amy; Meurer-Grimes, Barbara; Mori, Scott|
|Publication:||The Botanical Review|
|Date:||Oct 1, 1997|
|Previous Article:||Mineral nutrition of carnivorous plants: a review.|
|Next Article:||Survey of Leguminosae - a study in adaptive radiation.|