Skip to main content
NCBI home page
Proceedings of the Royal Society B: Biological Sciences logoLink to Proceedings of the Royal Society B: Biological Sciences
. 2006 Apr 19;273(1596):1969–1975. doi: 10.1098/rspb.2006.3521

Beetle assemblages from an Australian tropical rainforest show that the canopy and the ground strata contribute equally to biodiversity

Nigel E Stork 1,*, Peter S Grimbacher 1
PMCID: PMC1634767  PMID: 16822759

Abstract

There remains great uncertainty about how much tropical forest canopies contribute to global species richness estimates and the relative specialization of insect species to vertical zones. To investigate these issues, we conducted a four-year sampling program in lowland tropical rainforest in North Queensland, Australia. Beetles were sampled using a trap that combines Malaise and flight interception trap (FIT) functions. Pairs of this trap, one on the ground and a second suspended 15–20 m above in the canopy were located at five sites, spaced 50 m or more apart. These traps produced 29 986 beetles of 1473 species and 77 families. There were similar numbers of individuals (canopy 14 473; ground 15 513) and species (canopy 1158; ground 895) in each stratum, but significantly more rare species in the canopy (canopy 509; ground 283). Seventy two percent of the species (excluding rare species) were found in both strata. Using IndVal, we found 24 and 27% of the abundant species (n≥20 individuals) to be specialized to the canopy and the ground strata, respectively, and equivalent analyses at the family level showed figures of 30 and 22%, respectively. These results show that the canopy and the ground strata both provide important contributions to rainforest biodiversity.

Keywords: biodiversity, canopy crane, Coleoptera, species richness, specialization, tropical forest insects

1. Introduction

Tropical forest insects dominate the number of described and estimated species on Earth. Understanding the degree of insect specialization to vertical zones, host plants and other resources within tropical forests is of central importance to global species richness estimates (Erwin 1982; Stork 1988, 1993; Hammond 1995; Ødegaard 2000; Novotný et al. 2002). The first use of knockdown insecticides (canopy fogging) to collect samples of tropical forest canopy insects (Erwin & Scott 1980; Adis et al. 1984) revealed a previously unparalleled level of species richness of Coleoptera and subsequently led to Erwin's supposition that the canopy is twice as rich in species as the ground (Erwin 1982). However, to date there have been few studies that have accurately quantified the vertical stratification of insects within tropical forests.

One such study of 4025 beetle species collected over 12 months from an Indonesian tropical forest (Hammond 1990; Stork & Brendell 1990) suggested that only 8–13% of these beetles are canopy specialists (Hammond et al. 1997). Other canopy–ground comparisons within tropical rainforest have suggested that most insect activity is located in the top few metres of the canopy (Sutton et al. 1983), and that insects are more abundant and speciose in the canopy than the understory (Basset et al. 2001; Charles & Basset 2005). However, this contrasts with other studies on Lepidoptera and Formicidae which suggest that the ground layer is as speciose as the canopy (DeVries et al. 1997; Brühl et al. 1998; Schultze et al. 2001).

Despite poor consensus as to whether the canopy or the ground contributes more to global biodiversity, numerous studies on arthropod groups including Collembola (Rogers & Kitching 1998), Lepidoptera (DeVries et al. 1997; Devries & Walla 2001; Schultze et al. 2001), Formicidae (Longino & Nadkarni 1990; Brühl et al. 1998), Chrysomelidae (Charles & Basset 2005), Drosophila (Diptera; Tanabe 2002) and a range of insect orders (Basset et al. 2001), have unequivocally shown that arthropod assemblages in the canopy are very distinct from those that inhabit the ground zone. These differences are likely to be determined by a range of factors including resource availability, microclimate preferences and predator avoidance (Haddow et al. 1961; Kato et al. 1995; Brühl et al. 1998; Rogers & Kitching 1998; Schultze et al. 2001; Tanabe 2002; Basset et al. 2003). The degree of specialization to resources and physiological tolerances to microclimate are, therefore, likely to be particularly important in structuring arthropod assemblage differences between the canopy and the ground.

Our inability to determine how different vertical strata contribute to global biodiversity is the result of three major methodological problems with previous studies. First, few studies have incorporated methodologies that allow equal sampling effort in the canopy and near the ground. For example, although canopy fogging can sample vast numbers of arthropods in the canopy, such a technique cannot be used to sample the insect fauna of the ground layer. Second, few studies have used ecologically diverse or species-rich taxa to assess arthropod assemblages and it is difficult to extrapolate these results with confidence for other insect taxa. Third, few have tested vertical stratification data using robust and appropriate statistics that incorporate abundance. As far as we are aware, no study has adequately addressed all three areas of concern.

This study focuses on the distribution of insects between the canopy and the ground of an Australian tropical lowland rainforest. Here we have been careful to use a methodology that allows equal sampling of the canopy and the ground. We have used Coleoptera as our target taxon as this is by far the most species-rich order of insects (Hammond 1992; Nielson & Mound 1999) and, since this is arguably the most ecologically diverse order of insects (Lawrence et al. 2000), extrapolations for other insect taxa may be possible. We have also used appropriate statistics that incorporate abundance to test the hypotheses that (i) the canopy is richer in species than the ground and (ii) that a higher proportion of these species are strata specific.

2. Material and methods

This study was conducted in lowland tropical rainforest (complex mesophyll vine forest type 1a (Tracey 1982)), under or close to the Australian canopy crane (40 m asl, 16°17′ S, 145°29′ E), approximately 140 km North of Cairns in North Australia. The canopy is topographically quite varied ranging from 10 to 35 m in height and at least 90 species of tree of 10 cm dbh of 35 families have been identified in the hectare of forest directly below the canopy crane. Insects were sampled using a 1700 mm long, 1700 mm high Malaise trap modified to also act as a flight interception trap (FIT). Collecting pans containing a 40% solution of propylene glycol were placed either side of the base of the central net to provide the FIT samples. In this way insects flying into the central net would either fly upwards and be caught in the Malaise bottle or fly/drop downwards and collect in the FIT. Pairs of traps, one suspended 15–20 m in the canopy and one directly underneath on the ground, were located at five sites, each 40–60 m from each other. Canopy traps were suspended from different species of trees. Aluminium frames with plastic covers were built for the traps to ensure that they maintained their shape and sampling efficiency. Malaise traps were washed once a year and replaced if damaged. Traps were run for two weeks a month from March 2000 to February 2004 (except July and October 2001, March 2002 and May 2003, while two samples were obtained in August 2000). This strategy generated five replicate canopy FIT, canopy Malaise, ground FIT and ground Malaise samples, during each of the 45 sampling periods. Beetles were removed from the samples, point mounted, labelled and sorted to families and morphospecies by both authors, using reference collections and expert advice where necessary (see acknowledgments).

For most analyses temporal data were pooled and differences among strata (and sometimes trapping methodology) were compared. Differences in the number of individuals, total species richness, rare species (singletons, and doubletons) and species (excluding rare species), were compared between strata using traps as replicates (but pooling trapping methodology), ANOVA statistics, and log (x+1) transformations of the data. Multi-dimensional scaling ordination (MDS), and cluster analysis were used to represent the patterns of similarity in beetle species composition among traps from the ground and the canopy (based on log (x+1) transformations of the entire species abundance data and a Bray–Curtis distance measure). The statistical significance of beetle assemblage differences between canopy and ground samples was tested with analysis of similarity using the data described above. The identification of species specialist to the canopy or ground strata was made using the indicator value, IndVal procedure (Dufrêne & Legendre 1997; McGeoch et al. 2002). This method combines species abundances with their frequency of occurrence at various groups of sites and in this study is based on their relative (square root-transformed) abundances and fidelity to the canopy or ground. IndVal is a symmetric indicator and is maximum when all individuals of a species are found at a particular type of site. The statistical significance of the IndVal statistics was assessed using a randomization procedure (5000 randomizations). IndVal statistics were separately calculated for each trapping method, trapping methods combined, each of the four years of sampling (year=March−February), cumulatively across years and for beetle families. Species accumulation curves and species richness estimators (Chao 1, ACE, ICE), were performed using EstimateS (v. 7.5, Colwell 2005).

3. Results

Over four years, the canopy–ground samples from the Australian tropical lowland rainforest site produced 29 986 beetles sorted to 1473 species within 77 families (table 1). There were surprisingly similar numbers of individuals in each stratum (canopy 14 473; ground 15 513), and these differences were not statistically significant (ANOVA, d.f.1,4, p=0.783). Of all the species caught, 78.6% (1158) were in canopy traps and 60.8% (895) were in ground traps, however, these differences were not statistically significant (ANOVA, d.f.1,4, p=0.131). A high proportion of species (39.4%) were shared by both strata, and when singletons and doubletons were excluded, this figure was proportionally even higher (72.0%). There were significantly more rare species in the canopy than near the ground (ANOVA, d.f.1,4, singletons p=0.048, doubletons p=0.011) and in particular more in canopy FITs than ground FITs (ANOVA, d.f.1,4, singletons p=0.027, doubletons p=0.003). Rare species largely contributed to the differences in species totals among strata, as when they were removed, the number of species in ground and canopy samples were very similar (table 1), (ANOVA, d.f.1,4, p=0.322). The FIT component of the traps collected a higher proportion of the individuals and species than the Malaise trap component (figure 1). Non-parametric species richness estimators Chao 1, ACE and ICE (Colwell 2005) predict total species richness figures of 2118–2200 species suggesting that 67–70% of the beetle fauna has so far been sampled. However, total species richness estimators continued to rise with year by year increasing sample size (data not presented here) demonstrating the unstable nature of all of the species richness estimators used in EstimateS (v. 7.5, Colwell 2005).

Table 1.

Total number of individuals and species caught in five combined Malaise/FIT traps over four years from the ground and the canopy.

canopy ground
individuals 14 473 15 513
species 1185 895
singletons 340 186
doubleton species 169 97
species excluding singletons and doubletons 649 612
Open in a new tab

Figure 1.

Figure 1

Open in a new tab

Species accumulation curves (dark lines) and 95% confidence intervals (light lines) for each trapping method, strata and combining both factors. Samples represent two week trapping intervals, conducted almost every month, over four years. Curves are the product of 50 randomizations (Colwell 2005).

Ordination of this dataset by stratum and sampling method shows the striking differences between canopy and ground beetle assemblages, irrespective of differences between trapping method (figure 2). The differences between beetle assemblages from the canopy and the ground were highly significant (ANOSIM Global R=0.424, p<0.001). A cluster analysis based on degree of similarity (Bray–Curtis) among the beetle assemblages collected in both the Malaise traps and FITs, showed clear separation of sites into canopy and ground groups while there was no evidence of spatial autocorrelation among sampling sites (figure 3). Despite the clear partitioning of assemblages by vertical strata, canopy and ground assemblages showed compositional similarity in the 30–40% range. Up to 238 species (84.3% of all individuals) surpassed the abundance threshold (n≥20 individuals) to be included in IndVal analyses. We found 24% (range 24–26) of the canopy species and 27% (range 16–29) of the ground species were identified as specialists of those strata, depending on trapping method used (figure 4a). These proportions remained fairly constant over the four years sampling was conducted (figure 4b). Over time, additional species reaching the analysis threshold were identified as stratum specialists at a similar proportional rate (figure 4c) to overall species stratum-specialist allocations (figure 4a,b).

Figure 2.

Figure 2

Open in a new tab

Multi-dimensional scaling ordination of the Australian rainforest beetle assemblage according to trap position and type (temporal data pooled). Each point represents the species composition of a trap. Points spaced closer together are more similar in species composition.

Figure 3.

Figure 3

Open in a new tab

Dendrogram of a cluster analysis showing the degree of similarity in the beetle assemblage sampled over 4 years, among the five sampling sites and vertical strata for (a) FITs, and (b) Malaise traps.

Figure 4.

Figure 4

Open in a new tab

Proportions of species that are statistically significant indicators of the canopy, ground or not statistically significant (i.e. are neither canopy or ground stratum specialists) using IndVal, (a) for all species where number of individuals n≥20 (total number of species is over each column), (b) as (a) but separately for each year (year 1 n=58, year 2 n=58, year 3 n=96, year 4 n=61), and (c) as (a) but cumulatively across years for ‘new’ species only (species not previously classified from previous year(s), year 1 n=58, year 1, 2 n=47, year 1, 2, 3 n=84, year 1, 2, 3, 4 n=49).

A similar IndVal analysis at the family level (and at subfamily level for Scarabaeidae, Staphylinidae and some Curculionidae; table 2) showed that of 64 Families (including 19 subfamilies) with 20 or greater individuals, 19 (30%) were canopy specialists and 14 (22%) were ground specialists. These specialist families usually had species that were also specialists of the particular stratum, but collectively were from diverse feeding guilds (allocation of families to feeding guild according to Hammond (1990) and Lawrence et al. (2000)) including herbivores, predators, fungivores and saprophages and mixtures of these guilds. There were few stratum specialist herbivore families and surprisingly more fungivorous (or largely fungivorous families).

Table 2.

Numbers of individual beetles, the ratio of those collected in canopy to ground traps, and significance levels (those in bold p<0.05) of IndVal analyses at the family level (and subfamily for Scarabaeidae, Staphylinidae and some Curculionidae). Number of species, number of canopy specialists and number of ground specialists record the total number of individual species and those species recognized as indicators of the canopy or the ground using IndVal. Trophic guild provides an assessment (or possible trophic guilds) for each family and is based on Hammond (1990) and Lawrence et al. (2000); H=herbivore; P=predator; F=fungivore; X=xylophage; S=saprophage; ?P=possibly predatory; F, (P) most of the family are fungivores but a few species are predators; ? unknown feeding guild.

family number of individuals canopy : ground ratio p-value (canopy) p-value (ground) number of species number of canopy specialists number of ground specialists trophic guild
Rhysodidae 6 83 : 17 0.1102 0.9778 2 F
Carabidae 113 41 : 59 0.7730 0.2422 26 P
Hydrophilidae 25 28 : 72 0.7378 0.2730 7 S, (P)
Histeridae 510 54 : 46 0.3102 0.6996 12 2 P
Hydraenidae 3 0 : 100 1.0000 0.0844 1 H
Ptiliidae 171 20 : 80 0.9702 0.0338 8 1 F
Leiodidae 987 19 : 81 1.0000 0.0040 15 5 F
Scydmaenidae 521 27 : 73 1.0000 0.0052 31 3 P
Staphylinidae
 Aleocharinae 674 39 : 61 0.8352 0.1724 109 1 P, F,
 Euaesthetinae 6 33 : 67 0.7310 0.4918 1 P
 Omaliinae 2 0 : 100 1.0000 0.2236 2 P, S
 Osoriinae 275 53 : 47 0.2912 0.6994 17 1 1 S
 Oxytelinae 20 30 : 70 0.7732 0.2976 3 S
 Paederinae 211 38 : 62 0.8874 0.1148 16 2 P
 Pselaphinae 521 27 : 73 1.0000 0.0044 86 3 P
 Scaphidiinae 578 17 : 83 1.0000 0.0038 15 6 F
 Staphylininae 551 25 : 75 0.9540 0.0592 34 1 P
 Tachyporinae 194 37 : 63 0.8672 0.1466 20 1 P, S, F
Lucanidae 3 100 : 0 0.2154 1.0000 3 X
Passalidae 1 0 : 100 1.0000 0.5064 1 X
Bolboceratidae 26 12 : 88 0.9970 0.0306 3 S
Ceratocanthidae 2 50 : 50 0.7708 0.7904 1 S?, F?
Scarabaeidae
 Aphodiinae 5 60 : 40 0.5000 0.7820 2 S
 Cetoniinae 120 96 : 4 0.0030 1.0000 6 2 H, (S, X)
 Dynastinae 2 100 : 0 0.2228 1.0000 1 H, S, X
 Melolonthinae 343 56 : 44 0.2146 0.7972 14 2 2 H
 Rutelinae 35 91 : 9 0.0134 0.9910 4 H (S, X)
 Scarabaeinae 672 18 : 82 1.0000 0.0074 9 3 S
Eucinetidae 3 100 : 0 0.0850 1.0000 1 F
Clambidae 1 0 : 100 1.0000 0.4988 1 F
Scirtidae 35 80 : 20 0.0106 0.9940 8 S
Buprestidae 43 63 : 37 0.4060 0.6524 9 H, (X)
Byrrhidae 1 0 : 100 1.0000 0.4976 1 H
Limnichidae 5 20 : 80 0.9972 0.0978 2 ?S/H
Psephenidae 6 100 : 0 0.0822 1.0000 2 H
Ptillodactylidae 38 24 : 76 0.8104 0.1996 3 F, (S)
Chelonariidae 2 100 : 0 0.2206 1.0000 1 ?S
Callirhipidae 2 100 : 0 0.2350 1.0000 1 X
Eucnemidae 748 43 : 57 0.6682 0.3246 34 2 1 F/X
Throscidae 759 80 : 20 0.0038 1.0000 3 2 ?F, ?X
Elateridae 722 32 : 68 0.9710 0.0286 38 1 1 H, X, P, S
Lycidae 30 37 : 63 0.8806 0.1460 5 ?P, ?F/X
Lampyridae 1 100 : 0 0.5022 1.0000 1 P
Cantharidae 43 42 : 58 0.7888 0.2658 6 P, (H)
Nosodendridae 106 20 : 80 0.9972 0.0110 1 1 S/F
Dermestidae 325 94 : 6 0.0042 1.0000 8 2 1 S
Anobiidae 944 88 : 12 0.0202 0.9824 23 4 X, (F)
Jacobsoniidae 10 70 : 30 0.1794 0.9182 1 S?, F?
Trogossitidae 5 40 : 60 0.7968 0.2626 2 ?
Cleridae 219 78 : 22 0.0138 0.9890 14 2 P, (H)
Melyridae 44 59 : 41 0.1096 0.8988 7 P
Sphindidae 390 16 : 84 0.9972 0.0110 2 1 F
Nitidulidae 1905 25 : 75 1.0000 0.0030 46 4 4 F, S, H, (P)
Monotomidae 147 63 : 37 0.0890 0.9064 3 S, F
Silvanidae 32 41 : 59 0.8630 0.1636 5 F, S
Passandridae 4 100 : 0 0.0816 1.0000 2 P
Cucujidae 1 100 : 0 0.4938 1.0000 1 F, (P)
Laemophloeidae 60 95 : 5 0.0044 1.0000 18 F, (S)
Propalacticidae 26 85 : 15 0.0270 0.9958 6 F
Phalacridae 200 45 : 56 0.7754 0.2226 17 1 2 H, F
Cryptophagidae 56 54 : 46 0.5216 0.4994 5 1 F
Languriidae 101 93 : 7 0.0032 1.0000 7 1 F, H
Erotylidae 5 80 : 20 0.3542 0.9166 3 F
Biphyllidae 127 33 : 67 0.9862 0.0254 4 1 F
Bothrideridae 2 100 : 0 0.5044 1.0000 2 P, (F)
Cerylonidae 244 45 : 55 0.8490 0.1552 7 1 F
Discolomatidae 5 20 : 80 0.9770 0.1940 2 F
Endomychidae 104 19 : 81 0.9918 0.0116 10 1 F
Coccinellidae 294 71 : 29 0.0040 1.0000 35 2 P, (H)
Corylophidae 1884 45 : 55 0.8026 0.2022 23 2 F
Latridiidae 346 40 : 60 0.5238 0.4906 6 2 F, (H)
Mycetophagidae 5 100 : 0 0.0198 1.0000 1 F
Ciidae 67 31 : 69 0.9258 0.0794 8 F
Melandryidae 29 90 : 10 0.0126 1.0000 8 F, (X)
Mordellidae 1368 75 : 25 0.0080 0.9962 32 1 1 H, (F, X)
Rhipiphoridae 5 100 : 0 0.0202 1.0000 2 P
Zopheridae 187 58 : 42 0.1898 0.8270 21 1 1 F, (P)
Tenebrionidae 103 78 : 22 0.0232 0.9816 39 S, F
Oedemeridae 23 35 : 65 0.8454 0.2200 5 H, (X)
Pyrochroidae 18 17 : 83 0.9562 0.0758 1 S?, F?
Salpingidae 20 95 : 5 0.0134 1.0000 8 ?F
Anthicidae 110 7 : 93 1.0000 0.0048 4 1 S, (?P)
Aderidae 272 56 : 44 0.3000 0.7072 16 1 ?S, ?H
Scraptiidae 193 80 : 20 0.0048 0.9972 6 2 ?X, ?F
Cerambycidae 219 60 : 40 0.1140 0.8940 52 1 1 X, (H)
Chrysomelidae 2547 68 : 32 0.0886 0.9192 70 5 1 H
Anthribidae 404 66 : 34 0.0314 0.9686 57 1 1 F, (H)
Attelabidae 4 75 : 25 0.3730 0.9084 2 H
Brentidae 65 65 : 35 0.0716 0.9434 12 X
Apionidae 2 100 : 0 0.2110 1.0000 2 H
Curculionidae
 other Curculionidae 2918 57 : 43 0.1832 0.8066 241 3 13 H, (X)
 Platypodinae 413 56 : 44 0.3370 0.6758 8 3 X, X/F
 Scolytinae 4412 39 : 61 0.8480 0.1682 53 7 3 X, X/F
Open in a new tab

4. Discussion

Our results unequivocally show that (i) both the canopy and the ground are equally species-rich, and (ii) they share many species. The evidence presented here therefore does not support the notion that most species are exclusively found in the canopy (Erwin 1982). Insect associations with the canopy or ground are likely to vary along a sliding scale, with some entirely associated with a particular stratum and others equally split between strata. These patterns are likely to reflect the specific biology of the species concerned and may even change with season (Devries & Walla 2001), although we have not addressed this possibility here. Our finding that there was vertical stratification in the beetle assemblage concurs with those for other tropical forest insect taxa (Longino & Nadkarni 1990; DeVries et al. 1997; Brühl et al. 1998; Rogers & Kitching 1998; Basset et al. 2001; Schultze et al. 2001; Tanabe 2002; Charles & Basset 2005).

Our results show that similar proportions (24–27%), of those species common enough to assess, were sufficiently strongly associated with the canopy or ground layer to be considered to be specialists of these strata. However, even among the 115 ‘non-specialist’ species, 42 species were three times more abundant in one stratum than the other. The absence of any spatial autocorrelation among assemblages caught from canopy–ground pairs of sites suggests that the vertical stratification patterns among beetle assemblages are robust. The statistical method we have used to determine stratum specialization (IndVal) provides an unbiased assessment of species' fidelity to vertical strata but was only used for species with 20 or more individuals. Even though this accounted for 84.3% of the individuals, it included only 16.2% of all the species. However, there is some evidence that the patterns observed for the abundant species may also hold true for rarer species as figure 4c shows that the proportions of species classified as canopy or ground specialists did not change when additional species reached the analysis threshold of 20 or more individuals each year.

What are the causes of vertical stratification? First, there are differences between the canopy and the ground in the kinds of resources and their quality and quantity. For example, the canopy is where most leaves, flowers and fruits appear, while the ground is where these resources and dead wood accumulate and decompose. As a result, canopy insects should be more likely to be associated with key canopy processes such as herbivory, and pollination, whereas ground-based insects should be more likely to be associated with decomposition. Family and feeding ecology data from this study, however, do not support this theory. Second, the microclimate of the canopy layer differs strongly from the ground. The upper canopy is exposed to more solar radiation, experiences much stronger wind velocities and temperature extremes, and is less humid than the ground (Pinker 1980; Walsh 1996; Szarzynski & Anhuf 2001; Turton & Siegenthaler 2004). Third, behavioural traits (e.g. differences in predator avoidance; Schultze et al. 2001) may be important.

Our finding that 72% of the more common species are found in both strata shows that the beetle faunas are in many ways much more similar than might have been expected. We have two possible explanations for this. First, although there are some differences in the quality and quantity of resources between the canopy and the ground, there are also some similarities. We found support for this theory as among the families identified as canopy- or ground-specialists (52% of families with n≥20 individuals), there was no clear discrimination between vertical strata based on feeding ecology. At the resource level living wood is found in both strata and the ground zone contains other plant resources such as leaves, although in lower volumes than the canopy. Fungi are also present in both canopy and ground strata, although sometimes in different forms. Second, life-history constraints may force many canopy-dwelling species to spend time near the ground. The forest canopy has a much greater level of variation in humidity and temperature than the ground, and is therefore a challenging environment for beetle larvae, which are usually less protected from desiccation than adult beetles. This is reflected in the observation that some of the adult canopy-dwelling beetles have ground-based beetle larvae (Lawrence et al. 2000).

We were surprised to find that although more species were caught in the canopy than the ground, this difference was the consequence of there being many more rare species in the canopy. To our knowledge, this is the first time that this has been demonstrated although rare species are an important component of all tropical insect assemblages (Novotný & Basset 2000). We explore three possible explanatory hypotheses. First, it is possible that greater wind velocities present in the canopy (Walsh 1996; Szarzynski & Anhuf 2001) may have resulted in canopy traps capturing more ‘tourists’ (sensu Moran & Southwood 1982) just blowing by. We are unable to assess the likelihood of this hypothesis due to the absence of species' distributional data. The second possibility, which is related to the first, is that the difference in the abundance of rare species in the canopy and ground could be due to possible differences in the flight strategies of beetles in these strata. The presence of more rare species in the canopy could be the consequence of long-distance migratory flight and therefore, these species represent a much larger, geographically speaking, species pool, whereas the more abundant species represent local dispersal for new food resources. We have not tested this hypothesis further here. The third hypothesis is that the canopy is more heterogeneous in structure than the ground, and thus there are more rare species. However, the species composition data (Bray–Curtis measure of similarity) do not support this interpretation as the degree of variation within canopy and ground groups was very similar.

The tropical forest canopy is poorly known and yet is of significance to so many fields of environmental concern (Ozanne et al. 2003). Our results show that the canopy does have a large number of species, although probably not exclusively two-thirds of all insect species as had been suggested (Erwin 1982), and is therefore home to an important component of forest biodiversity. These results also confirm that the ground stratum provides an equally important component of biodiversity that is often overlooked (Hammond 1990). Insect assemblages from the canopy and ground strata of tropical rainforest share many species, and certainly cannot be considered independent from each other.

Acknowledgments

We thank V. Novotný, R. L. Kitching, R. Colwell, S. Pimm, L. Rapley and two anomous reviewers for comments on an earlier version of this manuscript. We thank the Australian Canopy Crane for use of the crane to access the canopy and M. Cermak, R. Rader, J. Hill, K Goodall and Green Corps volunteers for help in making the insect collections and sorting and mounting of the beetles.

References

  1. Adis J, Lubin Y.D, Montgomery G.G. Arthropods from the canopy of inundated and Terra firme forests near Manaus, Brazil, with critical considerations on the Pyrethrum-fogging technique. Stud. Neotrop. Fauna Environ. 1984;19:223–236. [Google Scholar]
  2. Basset Y, et al. Stratification and diel activity of arthropods in a lowland rainforest in Gabon. Biol. J. Linn. Soc. 2001;72:585–607. 10.1006/bijl.2001.0518 [Google Scholar]
  3. Basset Y, Hammond P.M, Barrios H, Holloway J.D, Miller S. Vertical stratification of arthropod assemblages. In: Basset Y, Novotny V, Miller S, Kitching R.L, editors. Arthropods of tropical forests: spatio-temporal dynamics and resource use in the canopy. Cambridge University Press; Cambridge, UK: 2003. pp. 17–27. [Google Scholar]
  4. Brühl C.A, Gunsalam G, Linsenmair K.E. Stratification of ants (Hymenoptera, Formicidae) in a primary rain forest in Sabah, Borneo. J. Trop. Ecol. 1998;14:285–297. 10.1017/S0266467498000224 [Google Scholar]
  5. Charles E, Basset Y. Vertical stratification of leaf-beetle assemblages (Coleoptera: Chrysomelidae) in two forest types in Panama. J. Trop. Ecol. 2005;21:329–336. 10.1017/S0266467405002300 [Google Scholar]
  6. Colwell, R. K. 2005 EstimateS: statistical estimation of species richness and shared species from samples. URL http://viceroy.eeb.uconn.edu/EstimateS
  7. Devries P.J, Walla T.R. Species diversity and community structure in neotropical fruit-feeding butterflies. Biol. J. Linn. Soc. 2001;74:1–15. 10.1006/bijl.2001.0571 [Google Scholar]
  8. DeVries P.J, Murray D, Lande R. Species diversity in vertical, horizontal, and temporal dimensions of a fruit-feeding butterfly community in an Ecuadorian rainforest. Biol. J. Linn. Soc. 1997;62:343–364. 10.1006/bijl.1997.0155 [Google Scholar]
  9. Dufrêne M, Legendre P. Species assemblages and indicator species: the need for a flexible asymmetrical approach. Ecol. Monogr. 1997;67:345–366. [Google Scholar]
  10. Erwin T.L. Tropical forests: their richness in Coleoptera and other arthropod species. Coleopt. Bull. 1982;36:74–82. [Google Scholar]
  11. Erwin T.L, Scott J.C. Seasonal and size patterns, trophic structure, and richness of Coleoptera in the tropical arboreal ecosystem: the fauna of the tree Luehea seemannii in the Canal Zone of Panama. Coleopt. Bull. 1980;34:305–355. [Google Scholar]
  12. Haddow A.J, Corbet P.S, Gillett J.D. Entomological studies from a high tower in Mpanga forest, Uganda. Part I. Introduction. Trans. R. Entomol. Soc. 1961;113:249–256. [Google Scholar]
  13. Hammond P.M. Insect abundance and diversity in the Dumoga-Bone National Park, N. Sulawesi, with special reference to the beetle fauna of lowland rain forest in the Toraut region. In: Knight W.J, Holloway J.D, editors. Insects and the rain forests of South East Asia (Wallacea) Royal Entomological Society; London, UK: 1990. pp. 197–254. [Google Scholar]
  14. Hammond P.M. Species inventory. In: Groombridge B, editor. Global biodiversity: status of the Earth's living resources. Chapman & Hall; London, UK: 1992. pp. 17–39. [Google Scholar]
  15. Hammond P.M. Described and estimated species numbers: an objective assessment of current knowledge. In: Allsopp D, Colwell R.R, Hawksworth D.L, editors. Microbial diversity and ecosystem function. CAB International; Wallingford, UK: 1995. pp. 29–71. [Google Scholar]
  16. Hammond P.M, Stork N.E, Brendell M.J.D. Tree-crown beetles in context: a comparison of canopy and other ecotone assemblages in a lowland tropical forest in Sulawesi. In: Stork N.E, Adis J, Didham R.K, editors. Canopy arthropods. Chapman & Hall; London, UK: 1997. pp. 184–223. [Google Scholar]
  17. Kato M, Inoue T, Hamid A.A, Nagamitsu H, Merdek M.B, Nona A.R, Itino Y, Yamane S, Yumoto T. Seasonality and vertical structure of light attracted insect communities in a dipterocarp forest in Sarawak. Res. Popul. Ecol. 1995;37:59–79. [Google Scholar]
  18. Lawrence J.F, Hastings A.M, Dallwitz M.J, Paine T.A, Zurcher E.J. CSIRO; Collingwood, Victoria, Australia: 2000. Beetles of the World. [Google Scholar]
  19. Longino J.T, Nadkarni N.M. A comparison of ground and canopy leaf litter ants (Hymenoptera: Formicidae) in a neotropical montane forest. Psyche. 1990;97:81–93. [Google Scholar]
  20. McGeoch M.A, Van Rensburg B.J, Botes A. The verification and application of bioindicators: a case study of dung beetles in a savanna ecosystem. J. Appl. Ecol. 2002;39:661–672. 10.1046/j.1365-2664.2002.00743.x [Google Scholar]
  21. Moran V.C, Southwood T.R.E. The guild composition of arthropod communities in trees. J. Anim. Ecol. 1982;51:289–306. [Google Scholar]
  22. Nielsen E.S, Mound L.A. Global diversity of insects: the problems of estimating numbers. In: Raven P.H, Williams T, editors. Nature and human society: the quest for a sustainable world. National Academy Press; Washington, DC: 1999. pp. 213–222. [Google Scholar]
  23. Novotný V, Basset Y. Rare species in communities of tropical insect herbivores: pondering the mystery of singletons. Oikos. 2000;89:564–572. 10.1034/j.1600-0706.2000.890316.x [Google Scholar]
  24. Novotný V, Basset Y, Miller S.E, Weiblen G.D, Bremer B, Cizek L, Drozd P. Low host specificity of herbivorous insects in a tropical forest. Nature. 2002;416:841–844. doi: 10.1038/416841a. 10.1038/416841a [DOI] [PubMed] [Google Scholar]
  25. Ødegaard F. How many species of arthropods? Erwin's estimate revisited. Biol. J. Linn. Soc. 2000;71:583–597. 10.1006/bijl.2000.0468 [Google Scholar]
  26. Ozanne C.M.P, et al. Biodiversity meets the atmosphere: a global view of forest canopies. Science. 2003;301:183–186. doi: 10.1126/science.1084507. 10.1126/science.1084507 [DOI] [PubMed] [Google Scholar]
  27. Pinker R. The microclimate of a dry tropical forest. Agric. Meteorol. 1980;22:249–265. 10.1016/0002-1571(80)90005-9 [Google Scholar]
  28. Rogers D.J, Kitching R.L. Vertical stratification of rainforest collembolan (Collembola: Insecta) assemblages: description of ecological patterns and hypotheses concerning their generation. Ecography. 1998;21:392–400. [Google Scholar]
  29. Schultze C.H, Linsenmaier E, Fiedler K. Understorey versus canopy: patterns of vertical stratification and diversity among Lepidoptera in a Bornean rain forest. Plant Ecol. 2001;153:133–152. 10.1023/A:1017589711553 [Google Scholar]
  30. Stork N.E. Insect diversity: facts, fiction and speculation. Biol. J. Linn. Soc. 1988;35:321–337. [Google Scholar]
  31. Stork N. How many species are there? Biodivers. Conserv. 1993;2:215–232. 10.1007/BF00056669 [Google Scholar]
  32. Stork N.E, Brendell M.J.D. Variation in the insect fauna of Sulawesi trees with season, altitude and forest type. In: Knight W.J, Holloway J.D, editors. Insects and the rain forests of south east Asia (Wallacea) Royal Entomological Society of London; London, UK: 1990. pp. 173–190. [Google Scholar]
  33. Sutton S.L, Ash C.P.J, Grundy A. The vertical stratification of flying insects in lowland rain forests of Panama, Papua New-Guinea and Brunei. Zool. J. Linn. Soc. 1983;78:287–297. [Google Scholar]
  34. Szarzynski J, Anhuf D. Micrometerological conditions and canopy energy exchanges of a neotropical rain forest (Surumoni–Crane Project, Venezuela) Plant Ecol. 2001;153:231–239. 10.1023/A:1017536512887 [Google Scholar]
  35. Tanabe S.-I. Between-forest variation in vertical stratification of drosophilid populations. Ecol. Entomol. 2002;27:720–731. 10.1046/j.1365-2311.2002.00469.x [Google Scholar]
  36. Tracey J. CSIRO; Melbourne, Australia: 1982. Vegetation of the humid tropical region of north Queensland. [Google Scholar]
  37. Turton S.M, Siegenthaler D.T. Immediate impacts of a severe tropical cyclone on the microclimate of a rain-forest canopy in north-east Australia. J. Trop. Ecol. 2004;20:583–586. 10.1017/S0266467404001622 [Google Scholar]
  38. Walsh R.P.D. Microclimate and hydrology. In: Richards P.W, editor. The tropical rain forest an ecological study. Cambridge University Press; Cambridge, UK: 1996. pp. 206–236. [Google Scholar]

Articles from Proceedings of the Royal Society B: Biological Sciences are provided here courtesy of The Royal Society

RESOURCES