CONSERVATION BIOLOGY AND BIODIVERSITY Diversity of Tropical Montane Forests: Diversity and Spatial Distribution of (Coleoptera: ) Inhabiting Leaf Litter in Southern Mexico

ROBERT W. JONES,1 CHARLES W. O’BRIEN,2 LORENA RUIZ-MONTOYA,3 4 AND BENIGNO GO´ MEZ-GO´ MEZ Downloaded from https://academic.oup.com/aesa/article/101/1/128/8463 by guest on 29 September 2021

Ann. Entomol. Soc. Am. 101(1): 128Ð139 (2008) ABSTRACT The abundance, diversity, and spatial distribution of the fauna (Coleoptera: Curculionidae) of leaf litter were compared between primary and successional cloud forests at the Biological Reserve Cerro Huitepec in the highlands of Chiapas, Mexico. In total, 2,102 individuals and 57 species of Curculionidae were collected and identiÞed, of which Ͼ90% represented undescribed species. Total weevil densities averaged 8.51 Ϯ 9.04 individuals per 0.5 m2 of leaf litter. Estimates of the total number of leaf litter weevils in the reserve were Ͼ20 million. Data suggest that even small fragments of cloud forest (Ͻ10 ha) may maintain viable populations of the more common species of leaf litter weevils for extended periods. However, because of their low powers of dispersion and their high levels of endemism, leaf litter weevils, and presumably other leaf litter , are particularly vulnerable to extinction when whole forest fragments are lost or when recent climate changes push cloud forests to higher elevations.

KEY WORDS tropical montane forests, cloud forests, biodiversity, Curculionidae, conservation

Tropical montane cloud forests (TMCFs) are consid- ranges that produces climatic archipelagos conducive ered one of the most threatened and least understood to allopatric speciation (Va´zquez-Garcõ´a 1993, Chal- terrestrial habitats in the world (Hamilton et al. 1993, lenger 1998, Luna Vega et al. 1999). Bruijnzeel and Veneklaas 1998). The total area cov- In Mexico, cloud forests or the broader category ered by TMCFs has always been small compared with “bosque meso´Þlo de montan˜ a” (sensu Rzedowski other habitat types, and it is rapidly being reduced by 1978) potentially occupy Ϸ0.86% of the total area of deforestation and the drying effects from deforested the country, corresponding to 17,400 km2 (Rzedowski upwind lowlands (Lawton et al. 2001). Despite their 1993, Challenger 1998). The total area is scattered imperiled state, these forests are vital in the capture among at least 100 isolated forests associated princi- and storage of water and often form the headwaters of pally with the eastern and western Sierra Madre, the the most important watersheds of countries within highlands of northern and southern Oaxaca, and the tropical latitudes. Additionally, these forests are im- various mountain systems of Chiapas (Bubb 1991, portant areas of species endemism of many plant and Challenger 1998). These ecosystems are apparently of groups (Peterson et al. 1993, Rzedowski 1993). considerable antiquity as evidenced by the presence Although the much-publicized lowland tropical rain of fossils of many of the characteristic ßoral elements forests generally are considered to contain greater of these forests in Miocene and Oligocene deposits numbers of species within a given area (alpha diver- (Challenger 1998). In addition, these forests may have sity) than TMCFs, these latter forests usually have extended over greater areas and at lower elevations relatively high species diversity within speciÞc groups during cooler periods of the geological history of the and greater differences in species between sites (beta region (Graham 1993, Martinez-Torres et al. 1988). diversity) (Peterson et al. 1993). This trend is gener- Presently, estimates of the remaining area of cloud ally considered the result of the natural fragmentation forests in Mexico are calculated to be less than half of of cloud forest habitats along and between mountain the potential area, or roughly 8,000 km2 (Challenger 1998), and these forests are considered to be highly threatened (Bubb 1991, Challenger 1998). 1 Corresponding author: Facultad de Ciencias Naturales-Biologõ´a, Universidad Auto´noma de Quere´taro, Centro Universitario, Quere´- Many taxonomic groups have high diversity and taro, Qro, 76010, Mexico (e-mail: [email protected]). endemism in cloud forests. These include epiphytes, 2 2313 W. Calle Balaustre, Green Valley, AZ 85614-8047. bryophytes, ferns, amphibians, and leaf litter inverte- 3 El Colegio de la Frontera Sur, Carretera Panamericana y brates (Richards 1952, Wake 1970, Frith and Frith Perife´rico Sur s/n, San Cristo´bal de las Casas, Chiapas, Mexico. 4 El Colegio de la Frontera Sur, Tapachula, Carr. Antiguo Aeropu- 1990, Nadkarini and Longino 1990, Hamilton et al. erto, km. 2.5, Tapachula, Chiapas, Mexico. 1993, Wolf 1993, Anderson and Ashe 2000). Inverte-

0013-8746/08/0128Ð0139$04.00/0 ᭧ 2008 Entomological Society of America January 2008 JONES ET AL.: LEAF LITTER WEEVILS IN TROPICAL MONTANE FORESTS 129 brates are very diverse in the ever-moist accumulated quantify the richness, diversity, distribution, and tem- leaf and ground litter of TMCFs and most species are poral variation of leaf litter weevils species (Co- as yet unnamed. Nadkarini and Longino (1990) report leoptera: Curculionidae) in TMCFs that vary between that invertebrates were 2.6 times more abundant in successional and primary forest and to generate data, ground litter than in the canopy in Monteverde Cloud which will provide insight into the conservation biol- Forest of Costa Rica, a pattern that contrasts with ogy of leaf litter fauna of TMCFs. lowland tropical forests, where density and diversity of insects in the canopy are estimated to be higher Methods and Materials (Erwin 1988, 1997). Because of the high diversity of leaf litter inverte- Study Site. The Biological Reserve “Cerro Huite- brates in tropical forests, they have been used to in- pec” (BRCH) is a 136-ha reserve that was declared a dicate conservation priorities and assess restoration in private reserve in 1987 and managed by the nonproÞt

tropical forests (Anderson and Ashe 2000, Nakamura conservation organization PRONATURA A. C. It is Downloaded from https://academic.oup.com/aesa/article/101/1/128/8463 by guest on 29 September 2021 et al. 2003). Anderson and Ashe (2000) analyzed the located 4.5 km west of San Cristo´bal de las Casas, in diversity and distribution of two families, Cur- the central part of the Highlands of Chiapas, Me´xico. culionidae (weevils) and Staphylinidae (rove bee- The coordinates of the reserve at the northern corner tles), from leaf litter of cloud forests and used results are 16Њ 44Ј 38Љ N and 92Њ 40Ј 15Љ W. The reserve is made to determine conservation priorities for these habitats up of a series of steep slopes of between 40 and 60% in Honduras. Cloud forest sites within Honduras were with an altitude range of 2,230Ð2,710 m. The mean ranked based on richness and diversity measures. annual temperature is between 14 and 15ЊC, and the They found high total species diversity (Ͼ500 species) mean annual precipitation is 1,300 mm with heavy and that the species of both these beetle families were cloud cover common above 2,550 m from May through highly endemic, with Ͼ50% of species restricted to a November (Ramõ´rez-Marcial et al. 1998). The western single site. The combined measures produced a broad and eastern boundaries of the reserve are sharply range of values and proved sensitive to differences in delineated between forested vegetation of the reserve overall diversity and levels of local endemism. Other and cleared agricultural areas outside of the reserve. advantages of using Curculionidae and Staphylinidae Ramõ´rez-Marcial et al. (1998) analyzed the ßora and are the ease of sampling, combined with a large and categorized the vegetation of the reserve. They re- robust data set that can be generated. corded 315 plant species, representing 32% of the Of these two coleopteran families, weevils inhabit- ßoristic elements estimated for the Central Highlands ing leaf litter of cloud forests are of special interest for of Chiapas above 2,000 m. They also recognized six several reasons. First, the leaf litter habitat is generally distinct communities within the reserve: grassland, not thought of as the principal habitat of this taxon. shrubland, early successional oak forest, mid-succes- Whereas most weevils are strictly phytophagous and sional oak forest, oak forest, and evergreen cloud for- usually found in association with their host plants est. Of the vegetation types sampled in the current (Anderson 2004), many weevils inhabiting leaf litter study, (early successional oak forest, mid-successional apparently feed in dead plant material. Second, high oak forest, oak forest, and evergreen cloud forest), the diversity of weevils in leaf litter is primarily a phe- shared ßoral elements were Ͻ58% between all com- nomenon of the tropics, and especially in tropical munities. Although the oak forests had greater overall montane forests. Although some representatives of richness of plant species, the cloud forest has a greater the genera common in leaf litter from cloud forests are number of species (20%) that are restricted to this found at higher latitudes in temperate habitats, the forest type (Ramõ´rez-Marcial et al. 1998). number of species is comparatively low (OÕBrien and Sampling Methodology. In total, 248 leaf litter sam- Wibmer 1982). And Þnally, weevils in leaf litter are ples was taken within the study site on the northern among the least studied of this enormous taxon and and eastern slopes of Cerro Huitepec. These samples most species are undescribed (Anderson and Ashe were divided among four altitudinal transects with 2000). four study sites each (Fig. 1). At each study site, eight Given their high endemism and diversity in cloud samples of 0.5-m2 leaf litter were taken once during forests and their ease of sampling, leaf litter weevils late June to early August (wet season) and from late have potential for serving as important indicator taxa November to early February (dry season). Samples in the conservation of tropical montane forests and were taken among four transects ϫ 4 study sites ϫ 8 their phylogenetic relationships between species may samples ϫ 2 seasons ϭ 256 (minus nine samples from provide detailed, small scale, biogeographic patterns. one site that were accidentally lost to Þre) ϭ 247. However, much needs to be learned about these or- Within each sample transect, two sample sites each ganisms, including their biology, the patterns of in-site were in primary and successional forests, respectively. diversity, and how past disturbances of TMCFs may At each sample site, eight individual circular 0.5-m2 affect these patterns. To establish the groundwork for samples of leaf litter were taken at 2-m intervals per- such studies and understand differences among sites, pendicular to the slope of the terrain. At each sample it is important to understand the variation in densities location, the slope was measured with a simple cli- and distribution of species within a single site and what nometer and the humidity of the leaf litter catego- factors may be responsible for explaining these ob- rized. Humidity categories were 1) dry, without any served patterns. The purpose of the current study is to evidence of humidity; 2) barely moist, with moisture 130 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 101, no. 1 Downloaded from https://academic.oup.com/aesa/article/101/1/128/8463 by guest on 29 September 2021

Fig. 1. Location of Huitepec reserve in the Central Highlands of Chiapas, Mexico and sampling scheme for leaf litter samples. Numbers refer to transects along an altitudinal gradient and habitat types coded as follows: S, successional oak forest; P, primary oak and cloud forest. January 2008 JONES ET AL.: LEAF LITTER WEEVILS IN TROPICAL MONTANE FORESTS 131 evidenced by touch and slightly lowered temperature; Quere´taro, Quere´taro, Me´xico (UAQE), Colegio de la 3) moist, with some clumping and softening of leaf Frontera Sur, San Cristo´bal de las Casas, Chiapas, fragments due to obvious moisture; 4) very moist, with Mexico and in CWOB, Collection of C. W. OÕBrien, obvious clumping and softening of leaf fragments, wa- 2313 W Calle Balaustre, Green Valley, AZ, USA. ter appearing when squeezed; and 5) wet, leaf litter Diversity Measures and Statistical Analysis. The shiny wet with moisture and water readily appearing numbers of individuals for each species were recorded when squeezed. The depth of the leaf litter was esti- along with corresponding ecological data for each mated by clearing a small hole and measuring from the sample and sample site. These data were used to es- soil to the top layer of debris with a plastic ruler. All timate total weevil diversity and diversity in both pri- organic debris (e.g., leaves, twigs, fungi, wood debris) mary and successional habitat categories of the re- down to, but not including the soil, within the 0.5-m2 serve. The categories were based on the classiÞcations circle was collected and placed in a litter reducer of Ramõ´rez-Marcial et al. (1998) and included the

(Martin 1977). The reduced debris was transferred to following: 1) primary forest consisting of cloud forest Downloaded from https://academic.oup.com/aesa/article/101/1/128/8463 by guest on 29 September 2021 labeled cloth bags and transported to the laboratory. and primary oak forest and 2) successional oak forest In addition, at each sample site the characteristics of consisting of intermediate successional oak forest in the vegetation in the vicinity of the site were recorded the park interior and early to intermediate succes- once within plots of 300 m2. These plots were 10 by sional oak forest at the reserve edge. Species accumu- 30 m and parallel to the line of eight leaf litter sample lation curves and estimators of diversity were calcu- locations. Data recorded in these plots included the lated using EstimateS 6.0 computer program (Colwell following: 1) number of ferns Ͻ50 cm in height, 2) 1997). Estimators included Chao 2, and second order number of ferns Ͼ50 cm in height, 3) number of herbs jackknife (Jack 2) (Colwell and Coddington 1994). Ͻ50 cm in height, 4) number of herbs Ͼ50 cm in These estimators were calculated for the total sampled height, 5) number of bushes Ͻ50 cm in height, 6) diversity and for each habitat type and for primary and number of bushes from 1 to 1.5 m in height, 7) number successional habitats. GreensÕs Index of aggregation of trees of 1Ð5 cm in diameter at breast height (dbh), (Green 1966) was used to compare dispersion of in- 8) trees 5.1Ð10 cm dbh, 9) trees 10.1Ð20 cm dbh, 10) dividual species among samples (N ϭ 248) and sam- trees 20.1Ð40 cm dbh, 11) trees 40.1Ð80 cm dbh, 12) ples sites (N ϭ 31). Diversity indices (Shannon and trees Ͼ80 cm dbh, and 13) number of fallen logs. SimpsonÕs) also were calculated using EstimateS 6.0 In the laboratory, leaf litter samples in cloth bags for all samples and for each of the habitat categories were weighed and placed on wire mesh over cheese- and tested for signiÞcance difference (t-test) between cloth in metal Berlese funnels (one sample per funnel) habitats using methods given in Magurran (1988) and under 60-W light bulbs for 3 d. Escaping weevils were Ludwig and Reynolds (1988). The pattern of total collected in 80% ethyl alcohol. To evaluate the efÞ- species abundance was tested for the lognormal and ciency of Berlese funnels in extracting weevils, visual log series model of species abundance by using tests revision of 16 samples processed with funnels vali- for signiÞcance given by Magurran (1988). dated that all weevils were thus collected. Samples Analysis of the patterns of species densities with were observed under a stereo microscope, and weevils that of physical and biological variables was initially were separated and then mounted on points and la- explored using the principal component analysis beled as to sample number, site and date of collection. (PCA) program of Multi-Variate Statistical Package, All weevils found in samples were considered as part version 3.01 (Kovach Computing Services 1998). Spe- of the leaf litter fauna and incorporated in the analysis, cies with variable loading Ͼ0.10 for any of the Þrst six with some notable exceptions. These exceptions were axes by using standardized data were then analyzed species of weevils that general sweep net samples using a canonical correlation analysis to determine the indicated occurred only on herbaceous vegetation and abiotic and biotic variables of most importance in the were sometimes inadvertently dislodged from plants abundance of these weevil species by using Statistical while sampling and incorporated in leaf litter samples. Analysis Software (SAS Institute 1985). After testing The most common examples of these individuals were for normal distributions, differences in densities of from the family Apionidae and of the genera Phyllotrox individuals (all combined species of weevils and for and Smicronyx (Curculionidae). Once pinned and la- selected species) were tested using KruskalÐWallis beled, all weevil samples were sorted to morpho-spe- nonparametric procedures with the statistical soft- cies. First, samples were sorted to genera and then split ware MINITAB, release 14 (Minitab, Inc., State into distinct morpho-species. Species were identiÞed College, PA). when possible based on comparison with identiÞed material in the collection of C.W.O.B. All morpho- Results species were examined for the presence of wings using one or a combination of the following methods: 1) In total, 2102 weevils was recovered from 248 sam- direct removal of elytra (used for at least one species ples of 0.5 m2 each of leaf litter from the Cerro Huite- of each genus), 2) presence (winged) or absence pec Reserve in the state of Chiapas, Mexico (Appendix (apterous) of elytral humeri, and 3) examination for 1; Table 1). Total weevils densities averaged 8.51 Ϯ evidence of fusion between elytra without dissection. 9.04 weevil individuals per 0.5 m2 of leaf litter. Weevil Vouchers specimens were deposited in the ento- individuals were categorized into 57 species, repre- mological collections of the Universidad Auto´noma de senting 18 genera. Of these, only three could be iden- 132 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 101, no. 1

Table 1. Species list, aggregation indices, and comparison of densities of the more common (>30 individuals captured) species of weevils (Curculionidae) in leaf litter in Huitepec Reserve, Chiapas, Mexico

GreenÕs index of dispersiona Comparison of densities No. Species list Point samples Sample sites individuals Habitatb Seasonc (N ϭ 247) (N ϭ 31) COSSONINAE Rhyncholini Heptarthrum sp. 1 69 0.0648 0.2063 *Primary forest *Dry season Cryptorhynchini Eurhoptus sp. 2 372 0.0087 0.0339 *Primary forest NS Eurhoptus sp. 3 121 0.0087 0.0339 NS NS Eurhoptus sp. 4 109 0.0107 0.0304 *Primary forest NS Eurhoptus sp. 5 117 0.0152 0.0234 *Successional forest NS Cryptorhynchinae new genus 1 sp. nov. 2 185 0.0102 0.0274 NS NS

Cryptorhynchinae new genus 1 sp. nov. 3 51 0.0156 0.0477 *Primary forest S NS Downloaded from https://academic.oup.com/aesa/article/101/1/128/8463 by guest on 29 September 2021 Cryptorhynchinae new genus 1 sp. nov. 4 89 0.0054 0.0291 NS NS Tylodinus sp. 6 50 0.0090 0.0350 *Primary forest NS Sciaphilini Sciomias sp. 1 380 0.0228 0.1139 *Successional forest *Dry season Anchonini sp. 1 57 0.0267 0.0431 *Primary forest *Wet season Anchonus sp. 2 33 0.0187 0.0391 *Primary forest *Wet season Levites sp. 1 94 0.0129 0.0479 *Primary forest *Wet season sp. 2 32 0.0080 0.0438 *Successional forest NS Conotrachelus sp. 3 30 0.0259 0.0451 *Successional forest NS Conotrachelus sp. 6 32 0.0567 0.0979 *Successional forest NS Lepilius sp. 1 47 0.0219 0.0957 *Primary forest NS Lymantini Theognete sp. 1 46 0.0376 0.1174 NS NS Totals (including all species) 2,102 0.0041 0.0143 NS NS

Systematic arrangement according to Alonso-Zarazaga and Lyal (1999). a GreenÕs dispersion index ϭ (s2/x) Ϫ 1/(n Ϫ 1) (Green 1966). b Asterisk indicating signiÞcant difference (P Ͻ 0.05) between habitats with higher densities in either primary forest or in successional forests by using KruskalÐWallis nonparametric test of means (adjusted for ties); NS indicates nonsigniÞcant H values (P Ͼ 0.05). Species with signiÞcant values between habitats (df ϭ 1, P Ͻ 0.05) were as follows: Heptarthrum sp. 1, H ϭ 18.03; Eurhoptus sp. 2, H ϭ 4.86; Eurhoptus sp. 4, H ϭ 10.95; Eurhoptus sp. 5, H ϭ 8.96; Cryptorhynchinae new genus 1 sp. nov. 3, H ϭ 5.31; Tylodinus sp. 6, H ϭ 6.75; Sciomias sp., H ϭ 19.00; Anchonus sp. 1, H ϭ 6.31; Anchonus sp. 2, H ϭ 5.56; Levites sp. 1, H ϭ 5.85; Conotrachelus sp. 2, H ϭ 13.02; Conotrachelus sp. 3, H ϭ 5.74; Conotrachelus sp. 6, H ϭ 4.08; and Lepilius sp. 1, H ϭ 8.09. c Asterisk indicating signiÞcant difference (P Ͻ 0.05) between wet and dry season with higher densities using KruskalÐWallis nonparametric test of means (adjusted for ties); NS indicates nonsigniÞcant H values (P Ͻ 0.05). Species with signiÞcant values between seasons (df ϭ 1, P Ͻ 0.05) were as follows: Heptarthrum sp. 1, H ϭ 8.66; Sciomias sp., H ϭ 10.89; Anchonus sp. 1, H ϭ 8.30; Anchonus sp. 2, H ϭ 5.62; and Levites sp. 1, H ϭ 4.23. tiÞed to species, and although several of these mor- the subfamily Molytinae of the tribe Conotrachelini pho-species may be later identiÞed with further study, (Alonso-Zarazaga and Lyal 1999). These winged in- most (Ͼ90%) of the individuals collected in the cur- dividuals accounted for only 5.4% of the total sampled rent study represent undescribed taxa. A species of the individuals. genus Sciomias (sp. 1) was the most frequently col- The total number of species collected (57) was rela- lected species, with 380 individuals, and a species tively close to that predicted by the Þnal richness esti- of Eurhoptus (sp. 2) had a slightly lower total number mators (Fig. 2). The estimators Chao 2 richness esti- of individuals, 372. There was a fairly large number of mator (Chao2), the abundance based coverage species that had moderate total abundances, with Ͼ30 estimator of species richness (ACE) and the second individuals (31.5% of all species). There were 12 sin- order Jackknife richness estimator (Jack2) calculated gletons (21%) and 10 doubletons (17.5%) (Appendix Þnal total species for the reserve to be 63.05, 67.56, and 1; Table 1). Of the genera, Eurhoptus had the most 71.96 total species, respectively. However, all estima- individuals (737; 35.1% of total) with the genus Scio- tors were not very successful in predicting the Þnal mias, second in total numbers (382; 18.2% total) and total estimate from few samples, and the number of species of the undescribed Cryptorhynchinae genus 1 predicted species increased at a rate similar to the ranked third (329; 15.7% of total). The genera Tylodi- actual randomized species accumulation curve (Fig. nus and Anchonus both had nine species, whereas 2). Chao 2 was the Þrst estimator to plateau at its Þnal Eurhoptus and Conotrachelus both had six species. estimated richness value, but required Ϸ100 samples Greater than 85% of the genera (16; 88.8%) and to reach this Þnal Þgure (Fig. 2). The relatively large species (49; 86.0%) were apterous. All apterous spe- number of samples needed for these estimators to cies had no remnants of the hind wings, and elytra reach an asymptote is assumed to be the result of the were fused along the midline, sometimes heavily so. high numbers of singletons and doubletons in the The only winged species belonged in two genera, samples, a characteristic of many studies of insects in Conotrachelus and , and they are both placed the tropics (Novotny´ and Basset 2000). January 2008 JONES ET AL.: LEAF LITTER WEEVILS IN TROPICAL MONTANE FORESTS 133

70 densities in the successional habitats. The remainder of the common weevil species also had similar or 60 higher densities in primary than in successional hab- 50 itats. Further indication of the difference in the overall patterns of these groupings of weevils is supported by 40 comparison of total weevil densities between succes- sional and primary habitats with and without the in- 30 Mean Accumulation clusion of Sciomius and Contrachelus in the analysis. Chao 2 No signiÞcant differences were found in total weevil 20 Jack 1 ACE densities among habitats (H ϭ 3.58, df ϭ 1, P ϭ 0.58). Species Accumulation 10 However, excluding Sciomius and Contrachelus from the comparison resulted in signiÞcantly higher den- 0 sities of combined leaf litter weevils in primary forest Downloaded from https://academic.oup.com/aesa/article/101/1/128/8463 by guest on 29 September 2021 0 50 100 150 200 250 than in successional forests (H ϭ 12.91, df ϭ 1, P Ͻ Sample Number 0.001). The majority (76.5%) of the common weevil species Fig. 2. Species accumulation curve and richness esti- (Ͼ30 collected individuals) showed little variation in mates of leaf litter weevils in the Huitepec Reserve, Chiapas, Mexico (N ϭ 244 samples). Mean accumulation is the mean densities between the wet season (JuneÐAugust) and cumulative number of species (random 50 pooled samples). the dry season (NovemberÐFebruary) (Table 1), de- Richness estimates: 1) Chao 2, Chao 2 richness estimator spite signiÞcant differences in the humidity of leaf (Chao 1984); 2) ICE, incidence-based coverage estimator of litter (mean humidity rating 4.06 versus 2.28 for wet species richness (Chazdon et al. 1998, Chao et al. 2000) and and dry season, respectively; ␹2 ϭ 100.62, df ϭ 7, P Ͻ 3) Jack2, second-order jackknife richness estimator (Palmer 0.01). Differences between seasons of the combined 1991). total densities of weevils were insigniÞcant (KruskalÐ Wallis, H Ͻ 1.24, df ϭ 1, P Ͼ 0.05), and only Þve of the All weevil species with Ͼ30 sampled individuals had 18 common species showed signiÞcant differences be- 2 aggregated distributions within 0.5-m samples and at tween seasons. Of these, three species, all of the tribe 2 sample sites (4 m among eight samples) (Table 1). Anchonini, had higher densities during the wet season GreenÕs index indicated that aggregations were (Anchonus sp. 1, Anchonus sp. 2, and Levites sp. 1: H ϭ greater at the sample site level for all common species, 8.30, df ϭ 1, P Ͻ 0.01, H ϭ 5.62, df ϭ 1, P Ͻ 0.05, and suggesting that the size of aggregations were greater ϭ ϭ Ͻ 2 H 5.62, df 1, P 0.05, respectively). The most than the 0.5-m sample size. Although the sampled common species, Sciomius sp. 1, was more common species were spatially clumped, all of these more com- during the dry season (H ϭ 10.89, df ϭ 1, P Ͻ 0.01), mon species were present in all of the forest habitats and the strict cloud forest inhabitant, Heptarthrum sp. samples and GreenÕs index of aggregation was rela- 1(H ϭ 8.66, df ϭ 1, P Ͻ 0.01). tively low for most species, suggesting populations were only loosely aggregated (Table 1). A notable Of the habitats, the successional forests had lower exception to this pattern was Sciomius sp. 1 and three species richness and greater relative dominance of one species of the genus Conotrachelus (species 2, 3, and 6) species, Sciomias sp. 1 (Table 2). This was reßected in that had the highest aggregation indices of all species. the lower richness values and signiÞcantly lower di- The majority of weevil species in leaf litter (69%) versity indices of both ShannonÕs and SimpsonÕs indi- were more common in primary habitats with three ces for these habitats (Table 2). The relatively large important exceptions (Table 1). The most common differences in the SimpsonÕs diversity index between and most aggregated species, Sciomius sp. 1 had sig- primary and successional forests, which is sensitive to niÞcantly higher densities in the successional than dominance relations, further suggested that succes- primary habitats (H ϭ 9.92, df ϭ 1, P Ͻ 0.002). Like- sional forests were more dominated by a few species wise, species of Conotrachelus (one of only two genera than primary forests. Besides greater diversity, pri- with winged species) also had signiÞcantly higher mary forests also had greater numbers of unique,

Table 2. Comparisons of richness and diversity measures between habitats of weevils in leaf litter in montane forests of central Chiapas, Mexico

Diversity indices No. species Estimated richness Singletons Habitat N SimpsonÕsa (richness) (Chao2) ShannonÕsa (of total) (1/D) Primary oak and cloud forest 127 50 70.22 Ϯ 17.9 2.88a 11.12a 10 Successional oak forest 120 39 40.7 Ϯ 2.5 2.67b 7.41b 2 Totals 247 57 63.66 2.88 10.92 12

a Different letters indicating signiÞcant difference in the diversity indices of ShannonÕs and SimpsonÕ SD between primary and successional habitat types determined with t-test by using methods given by Magurran (1988): t ϭ 3.82, df ϭ 1913, P Ͻ 0.001 for Shannon; and t ϭ 5.21, df ϭ 2102, P Ͻ 0.001 for SimpsonÕ SD. 134 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 101, no. 1

1000 cated that the most common species Sciomius sp. 1, Total was positively correlated with the presence of herba- Primary Forest ceous vegetation, bushes, and medium trees, whereas Succesional Forest the correlations of the second most common species 100 (Eurhoptus sp. 2 and Levites sp. 1) with these variables, were strongly negative. The third coefÞcient rein- forced the correlation of Sciomius sp. 1 and Eurhoptus sp. 5 with the presence of bushes and large trees 10 suggested by correlations with previously mentioned coefÞcients. Also, the third coefÞcient supported a positive trend between Lepilius sp. 1 with the pres- Number of Individuals Number ence of small trees, but in this case by indicating a 1

corresponding strong negative correlation with large Downloaded from https://academic.oup.com/aesa/article/101/1/128/8463 by guest on 29 September 2021 0 10 20 30 40 50 60 trees. The fourth coefÞcient was similar to the Þrst in Most Common Least Common Sequence of Species indicating strongest correlations with the weevil spe- cies, Lepilius sp. 1, and Cryptorhynchinae new genus Fig. 3. Rank abundance plots of weevil species in the 1 sp. nov. 2 and herbaceous vegetation, but in contrast Huitepec Reserve, Chiapas, Mexico. All curves Þt the log- to the Þrst coefÞcient, were negatively correlated with normal distribution: total weevil species: ␹2 ϭ 5.14, df ϭ 8, P Ͼ the presence of trees of all sizes. Also in this last 2 0.05; primary forest: ␹ ϭ 9.41, df ϭ 8, P Ͼ 0.05; and succes- coefÞcient, Levites sp. 1 was positively correlated with ␹2 ϭ ϭ Ͼ sional forest:, 8.10, df 8, P 0.05. the presence of trees of all sizes, but negatively so with herbaceous vegetation. species represented by one individual (singletons) (Table 2). Discussion The pattern of total species abundance Þt the log- normal model of species abundance (␹2 ϭ 8.36, df ϭ The richness and densities of weevils encountered 6, P Ͼ 0.05). Species abundance for both primary and in leaf litter during the current study was notable for successional habitats also Þt closely to the lognormal a relatively small area (Ͻ130 ha) and for the speciÞc model (Fig. 3), suggesting that although the richness habitat sampled. Using a sample size of 0.5 m2, the of species between habitats was different, the patterns accumulation of species was relatively slow and re- of abundance of species within the communities quired at least 25 samples to record half of the species within each habitat were similar. actually collected. A sample size of 100 was needed Canonical correlation analysis indicated several before richness estimators reached an asymptote patterns in the spatial distribution of weevils in leaf (Chao 2) and estimators indicated that from 6 to 13 litter (Table 3). CoefÞcient 1 indicated that three species were yet to be collected. Of the collected species (Lepilius sp. 1, Levites sp. 1, and Crypto- species, Ͼ90% were undescribed, and this emphasizes rhynchinae new genus 1 sp. nov. 2) had strong cor- the unknown nature of this fauna. At the genus level, relations with the presence of small trees and herba- however, all genera found in the current study (with ceous vegetation, and negatively correlated with the the exception of two singletons: Oxypteropsis and Oo- presence of large trees. The second coefÞcient indi- pterinus) were reported previously as leaf litter in-

Table 3. Canonical correlation analysis between densities of most important species of weevilsa (Curculionidae) in leaf litter and variables in vegetation in the Hutepec Reserve, Chiapas, Mexico

Standardized canonical coefÞcientsb Variable 1234 Eurhoptus sp. 2 Ϫ0.457 Ϫ0.622 0.227 Ϫ0.326 Eurhoptus sp. 5 Ϫ0.133 0.327 0.360 0.247 Cryptorhynchinae genus nov. 1 sp. nov. 2 0.522 0.227 Ϫ0.209 Ϫ0.478 Cryptorhynchinae genus nov. 1 sp. nov. 4 0.162 Ϫ0.077 0.216 0.312 Sciomius sp. 1 Ϫ0.413 0.402 0.485 0.164 Anchonus sp. 1 0.157 Ϫ0.036 Ϫ0.156 Ϫ0.156 Levites sp. 1 0.449 Ϫ0.367 0.082 0.580 Lepilius sp. 1 0.628 0.071 Ϫ0.518 Ϫ0.518 with ferns Ϫ0.301 0.125 Ϫ1.127 0.134 with herbaceous vegetation 0.595 0.545 0.142 Ϫ0.343 with bushes 0.095 0.565 0.956 0.126 with small trees 1.018 Ϫ0.351 0.178 0.483 with medium trees Ϫ0.034 0.249 Ϫ0.390 0.392 with large trees Ϫ0.368 0.027 0.906 0.365

a Species used in analysis had variable loading Ͼ0.10 for any of the Þrst six axes of PCA with standardized data. Likewise, these species accounted for 63% of the total number of individuals. b The statistics WilkÕs Lambda, PillaiÕs Trace, HotellingÐLawley Trace, and RoyÕs Greatest Rott all signiÞcant, P Ͻ 0.0001. January 2008 JONES ET AL.: LEAF LITTER WEEVILS IN TROPICAL MONTANE FORESTS 135 habitants from Honduras by Anderson and Ashe other species of Eurhoptus, as well as Tylodinus, An- (2000). However, no species are shared with those chonus, and Theognete, and Cryptorhynchinae new found in Honduras by Anderson and Ashe (2000), and genus 1 Þt this pattern with low dispersion indices, and most are in fact unique to the Cerro Huitepec area with greater densities in primary than in successional (R. S. Anderson, personal communication). These re- forests. The second pattern was that of Sciomius sp. 1 sults support results from Anderson and Ashe (2000) and species of the one of the two winged genera, that although the genera of weevils in leaf litter of Conotrachelus (sp. 2, sp. 3, and sp. 6), which are ap- TMCF of Mesoamerica are widespread, species dis- parently favored by disturbance, because they were tributions are highly endemic. Accumulation curves the only species that were more common in succes- and estimation of total diversity of leaf litter inhabiting sional than in primary forests (Table 1). In addition, weevils also were comparable with results reported by Sciomius sp. 1 was clearly an outlier in other respects, Anderson and Ashe (2000) for cloud forest sites in because it had the highest densities and the highest

Honduras. Many of these sites in Honduras were con- aggregation indices at all sample size levels and more Downloaded from https://academic.oup.com/aesa/article/101/1/128/8463 by guest on 29 September 2021 siderably larger than the Huitepec Reserve reported common during the dry season. This species also was here, suggesting that despite past disturbance and strongly correlated with the presence of herbaceous some reduction of the original extension of cloud vegetation and bushes suggesting that, like many forest, the Huitepec Reserve is still sufÞciently large to members of the Entiminae (Anderson and Howden maintain much of the original diversity of the weevil 2004), Sciomius sp. 1 is a generalist, nocturnal, fauna in leaf litter. aboveground plant feeder in the adult stage, with The slow accumulation of species richness and the larvae feeding below ground on roots. Although apter- relatively large number of samples needed before ous, these results suggest that Sciomius sp. 1 may not richness estimators reached an asymptote was appar- be a true, leaf litter obligate such as species of Eu- ently the result of a large number of singletons and rhoptus, Tylodinus, Anchonus, and Theognete. How- doubletons. It is also noteworthy that although the ever, more study is needed of the habitat preferences abundance data Þt the lognormal, the number of spe- and biology of species of Sciomius to validate this cies represented by only one individual (singletons) conclusion. The other species with similar spatial and was the value furthest from the predicted value. This temporal patterns belonged to the genus Conotrach- pattern in greater than expected numbers of single- elus, that are different from the other weevils found in tons for tropical has been reported previously leaf litter of the study because of their winged con- by several studies (Novotny´ and Basset 2000, Lucky et dition. The capacity to ßy would suggest that the al. 2002, Carlton et al. 2004). These latter studies clas- factors governing the larger scale geographical distri- siÞed the possible types of rare species in insect di- butions and levels of endemism of these species of versity studies, and considered “tourists” (species not Conotrachelus are distinct from the rest of the apterous normally associated with the habitat being sampled) weevil species in leaf litter. Also, information from to be an important component and problematical in other species of the genus indicate that larvae of Cono- determining valid diversity measures. However, the trachelus are principally fruit feeders or feed on dead proportion of species that are tourists in the current or dying wood (Kissinger 1964), suggesting that a study is probably low, because the sampling method- signiÞcant portion of their populations may be outside ology was relatively speciÞc and well deÞned (samples the leaf litter habitat. These combined characteristics of 0.5 m2 of forest litter) and the generally apterous suggest that, as concluded for Sciomius, species of condition of leaf litter weevils suggests that individuals Conotrachelus should not be classiÞed as true leaf litter rarely stray far from the leaf litter habitats. This con- obligates. The third spatial pattern was displayed by clusion is also supported by the high percentage of Lepilius sp. 1 and species of Anchonini (Anchonus and species (83%) that are within genera that are known Levites) that were both more common in primary leaf litter inhabitants. In addition, other conspeciÞcs forest but had unique and signiÞcant patterns in their of the same genera of the singletons and doubletons association with components of the vegetation, sug- were found in greater numbers in the samples (Ap- gesting that their biology was distinctly different from pendix 1; Table 1). Results from the current study that of the majority of the other leaf litter weevils. In support Lucky et al. (2002) in that sampling effort may addition, species of Anchonini were the only species be underestimated when studying richness and diver- that were more common during the wet season, sug- sity of tropical beetles. gesting distinct adaptations in life history traits to the The analysis of densities of weevils between habi- seasonal patterns in the humidity of cloud forests. tats, seasons, dispersion indices and the canonical cor- Considering the small size of these organisms and relation analysis of species and vegetation suggested the relatively homogenous and ubiquitous nature of three general patterns in the spatial distribution of the leaf litter they inhabit, a small reserve probably weevils. The most common spatial pattern is exem- contains enough resources and area to maintain viable pliÞed by Eurhoptus sp. 2, the second most common populations of the majority of weevil species for many species that are wingless, with adults restricted to the years. The mean density of all weevils recorded in this leaf litter habitat and can be considered to be a true, study was 17.02 weevils per m2 that converts to obligate leaf litter inhabitants of moist primary forests. Ϸ170,000 weevils per ha or roughly 22 million total This species had low dispersion values and was neg- weevils in the forested area of the reserve. Even ex- atively associated with herbaceous vegetation. All cluding weevils that results suggested are not true leaf 136 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 101, no. 1 litter obligates (species of Sciomius and Conotrach- presumably other apterous leaf litter invertebrates elus), the total is still Ͼ16 million leaf litter weevils in restricted to cloud forests will be stranded on moun- the reserve. Validation for this simple extrapolation is tain peaks in ever smaller populations where extinc- supported by the observation that all forested areas of tion is increasingly more likely. It is unknown how the reserve had considerable leaf litter accumulations, many species of leaf litter invertebrates have already and presumably capable of maintaining at least some been lost because we know very little of what is there of the leaf litter species. Thus, population estimates for and only beginning to catalog their diversity. If larger all but the rarest weevil species in the reserve are vertebrates are indeed indicator species, then it seems considerably greater than estimates of minimum via- reasonable to assume that the numbers of extinctions ble population sizes calculated for other insect species of insects, arachnids and other macroinvertebrates of (Biedermann 2000, Reed and Bryant 2000) and greater what Dunn (2005) calls the “neglected majority” is, or than that of 5,500 individuals calculated as a composite shortly will be, of several orders of magnitude greater

estimate based on various populations of small verte- than the numbers of extinctions of the ßagship ver- Downloaded from https://academic.oup.com/aesa/article/101/1/128/8463 by guest on 29 September 2021 brates and invertebrates (Thomas 1990). Two addi- tebrate species. tional factors also support the conclusion that small areas can support viable populations of weevils. First, aggregation indices suggested that the more common Acknowledgments weevils (Ͼ30 individuals; 30% species total), although loosely aggregated, were widely spread throughout We thank Olga Go´mez Nacumendi and Manuel Giro´n who the reserve and relatively common in both primary helped in the collection of samples, separation of material, and the mounting of specimens. Jesu´ s Luna Cozar curated and successional habitats. Second, differences in the material and provided comments. Jan Wolf of the Universiteit densities between the wet and dry season of the ma- van Amsterdam, Amsterdam, and Robert Anderson of the jority of the common weevil species (76.5%) and total Museum of Nature, Ottawa, provided comments on the Þnal weevil densities were nonsigniÞcant. This suggests manuscript. Thanks go to Biologa Rosa Marõ´a Vidal and Bi- that population sizes of weevil species in leaf litter are ologo Romeo Domõ´ngues Barrasas of PRONATURA, Chia- relatively stable, without marked seasonal ßuctuations pas, for support of this research project. El Colegio de la which are conducive to local extinction. Although Frontera Sur, San Cristo´bal and the Universidad Auto´noma these results indicate that even moderately disturbed de Quere´taro provided funds and material for the Þeld and montane forest sites can maintain viable and relatively the curatorial work of this project, and IDEA WILD provided collecting material. diverse populations of the more common leaf litter weevils, these conclusions cannot be extended to the rarer species, which were generally restricted to the primary cloud forest areas, and probably have more References Cited limited environmental tolerances and greater re- Alonso-Zarazaga, M. A., and C.H.C. Lyal. 1999. A world cat- source specialization than the more common species. alogue of families and genera of Curculionoidea (Insecta: Although evidence suggests that small areas could Coleoptera) (excepting Scolytidae and Platypodidiae). retain common species of leaf litter weevils for sus- The Natural History Museum, London y el Museo Na- tained periods, this conclusion is based on the assump- cional de Ciencias Naturales (CSIC), Entomopraxis, Bar- tion that environmental conditions remain relatively celona, Espan˜ a. Anderson, R. S. 2004. Family 131. Curculionidae, pp. 722Ð unchanged. However, signiÞcant changes in cloud for- 815. In R. H. Arnett, Jr., M. C. Thomas, P. E. Skelley, and est weather patterns and cloud height have already J. H. Frank [eds.], American beetles. CRC, New York. been documented (Lawton et al. 2001, Lawton et al. Anderson, R. S., and A. T. Howden. 2004. Family 131. Cur- 2001, Pounds 2001, Pounds and Puschendorf 2004), culionidae. XII. Entiminae Schoernherr. 766Ð767. In R. H. and numerous vertebrate species, such as the golden Arnett, Jr., M. C. Thomas, P. E. Skelley, and J. H. Frank toad of Monteverde Reserve of Costa Rica, have al- [eds.], American beetles. CRC, New York. ready become extinct (Pounds et al. 1999). Weevils in Anderson, R. S., and J. S. Ashe. 2000. Leaf litter inhabiting leaf litter of the montane cloud forests of Mexico and beetles as surrogates for establishing priorities for con- Central America are also clearly at risk from climate servation of selected tropical montane cloud forests in Honduras, Central America (Coleoptera: Staphylinidae, change because of several factors. First, they have low Curculionidae). Biodivers. Conserv. 9: 617Ð653. powers of dispersion (86% were ßightless in the cur- Biedermann, R. 2000. Metapopulation dynamics of the rent study). Thus, even if nearby cloud forests exist as froghopper Neophilaenus albipennis (F., 1798) (Ho- sites for recolonization, it is doubtful that leaf litter moptera, Cercopidae)Ðwhat is the minimum viable met- weevils are capable of, or disposed to, crossing areas apopulation size? J. Insect Conserv. 4: 99Ð107. without accumulations of humid litter, and even less Bruijnzeel, L. A., and E. J. Veneklaas. 1998. Climatic con- probable that they would cross the pastures or culti- ditions and tropical montane forest productivity: the fog vated plots that presently surround most remaining has not lifted. Ecology 79: 3Ð9. cloud forests. Second, leaf litter weevils of cloud for- Bubb, P. 1991. The current situation of the cloud forest in northern Chiapas, Mexico. Final report, Hyden Heath, ests are at risk because they are highly endemic Amersham, Bucks, United Kingdom. (Anderson and Ashe 2000) and restricted to cloud Carlton, C. E., M. Dean, and A. K. Tishechkin. 2004. Diver- forest habitats. As cloud forests are pushed to higher sity of two beetle taxa at a western Amazonian locality elevations due to drying of moisture bearing trade (Coloeptera: Staphylinidae, Pselaphinae). Coleopt. Bull. winds and higher cloud cover, leaf litter weevils and 58: 163Ð170. January 2008 JONES ET AL.: LEAF LITTER WEEVILS IN TROPICAL MONTANE FORESTS 137

Challenger, A. 1998. Utilizacio´n y conservacio´n de los eco- ican cloud forests: a preliminary vicariance model apply- sistemas terrestres de Me´xico. Pasado, presente y futuro. ing parsimony analysis of endemicity to vascular plant Comisio´n Nacional para el Conocimiento y Uso de la taxa. J. Biogeogr. 26: 1299Ð1305. Biodiversidad (CONABIO), Me´xico, D.F. Magurran, A. E. 1988. Ecological diversity and its measure- Chao, A. 1984. Non-parametric estimation of the number of ment. Princeton University Press, Princeton, NJ. classes in a population. Scand. J. Stat. 11: 265Ð270. Martin, J. H. 1977. Collecting, preparing and preserving in- Chao, A., W.-H. Hwang, Y.-C. Chen, and C.-Y. Kuo. 2000. sects, mites, and spiders. The insects and arachnids of Estimating the number of shared species in two commu- Canada. Part 1. Agriculture Canada, Ottawa, ON, Canada. nities. Stat. Sin. 10: 227Ð246. Martı´nez-Torres, E., S. Lozano-Ga´rcia, C. Tomasini, and H. Chazdon, R. L., R. K. Colwell, J. S. Denslow, and M. R. Herna´ndez. 1988. Polen fo´sil en Me´xico. pp. 9Ð12. Sim- Guariguata. 1998. Statistical methods for estimating spe- posio sobre Diversidad Biolo´gico de Me´xico (Oaxtepec, cies richness of woody regeneration in primary and sec- Morelos del3a7deoctubre de 1988) Resu´ menes de las ondary rain forests of NE Costa Rica, pp. 285Ð309. In F. ponencias. Instituto de Biologõ´a, UNAM, Me´xico.

Dallmeier and J. A. Comiskey [eds.], Forest biodiversity Nadkarini, N. M., and J. T. Longino. 1990. Invertebrates in Downloaded from https://academic.oup.com/aesa/article/101/1/128/8463 by guest on 29 September 2021 research, monitoring and modeling: conceptual back- canopy and ground organic matter in a Neotropical mon- ground and Old World case studies. Parthenon Publish- tane forest, Costa Rica. Biotropica 22: 286Ð289. ers, Paris, France. Nakamura, A., H. Proctor, and C. P. Cattrall. 2003. Using soil Colwell, R. K. 1997. EstimateS: statistical estimation of spe- and litter to assess the state of rainforest res- cies richness and shared species from samples. Version 5. toration. Ecol. Manag. Restor. 4 (Suppl): S20ÐS28. UserÕs guide and application. (http://viceroy.eeb. Novotny´, V., and Y. Basset. 2000. Rare species in communi- uconn.edu/estimates). ties of tropical insect herbivores: pondering the mystery Colwell, R. K., and J. A. Coddington. 1994. Estimating ter- of singletons. Oikos 89: 564Ð572. restrial biodiversity through extrapolation. Philos. Trans. O’Brien, C. W., and G. J. Wibmer. 1982. Annotated check- R. Soc. Lond. B 345: 101Ð118. list of the weevils (Curculionidae sensu lato) of North Dunn, R. R. 2005. Modern insect extinctions, the neglected America, Central America and the West Indies (Co- majority. Conserv. Biol. 19: 1030Ð1036. leoptera: Curculionidae). Mem. Am. Entomol. Inst. 34: Erwin, T. L. 1988. The tropical forest canopy: the heart of iÐix, 1Ð382. biotic diversity. pp. 123Ð129. In E. O. Wilson [ed.], Biodi- Palmer, M. W. 1991. Estimating species richness: the sec- versity. The National Academies Press, Washington, DC. ond-order jackknife reconsidered. Ecology 72: 1512Ð Erwin, T. L. 1997. Biodiversity at its utmost: tropical forest 1513. beetles, pp. 27Ð40. In M. L. Reaka-Kudla, D. E. Wilson, Peterson, A. T., O. A. Flores-Villela, L. S. Leo´n-Paniagua, J. E. and E. O. Wilson [eds.], Biodiversity II. Joseph Henry Llorente-Bousquets, M. A. Luis-Martinez, A. G. Navarro- Press, Washington, DC. Siguenz, M. A. Torres-Chavez, and I. Vargas-Ferna´ndez. Frith, D., and C. Frith. 1990. Seasonality of litter inverte- 1993. Conservation priorities in Mexico: moving up in the brates in an Australian upland rain forest. Biotropica 22: World. Biodivers. Lett. 1: 33Ð38. 181Ð190. Pounds, J. A. 2001. Climate and amphibian declines. Nature Graham, A. 1993. Historical factors and biological diversity (Lond.) 410: 639Ð640. in Mexico, pp. 109Ð128. In T. P. Ramamoorthy, R. Bye, A. Pounds, J. A., and R. Puschendorf. 2004. Clouded futures. Lot, and J. Fa [eds.], Biological diversity of Mexico: or- Nature (Lond.) 427: 107Ð109. igins and distribution. Oxford University Press, New York. Pounds, J. A., M.P.L. Fogden and J. H Campbell. 1999. Bi- Green, R. H. 1966. Measurement of non-randomness in spa- ological response to climate change on a tropical moun- tial distributions. Res. Popul. Ecol. 8: 1Ð7. tain. Nature (Lond.) 398: 611Ð615. Hamilton, L. S., J. O. Juvik, and F. N. Scatena. 1993. The Ramı´rez-Marcial, N. S. Ochoa-Gaona, M. Gonza´lez-Espinosa. Puerto Rico Tropical Cloud Forest Symposium. Intro- 1998. Ana´lisis ßoristico y sucesional en la Estacio´n Bio- duction and workshop synthesis, pp. 1Ð16. In L. S. Ham- lo´gica, Cerro Huitepec, Chiapas, Me´xico. Acta Bot. Mex. ilton, J. O. Juvik, and F. N. Scatena [eds.], Tropical mon- 44: 59Ð85. tane cloud forests. Proceedings of an international Reed, D. H., and E. H. Bryant. 2000. Experimental tests of symposium. San Juan de Puerto Rico. East-West Center minimum viable population size. Anim. Conserv. 3: 7Ð14. Program on Environment, UNESCO International Hy- Richards, P. W. 1952. The tropical rain forest. Cambridge drological Programme, and U.S. International Institute of University Press, Cambridge, United Kingdom. Tropical Forestry. Rzedowski, J. 1978. La vegetacio´ndeMe´xico, Prensa Kissinger, D. G. 1964. Curculionidae of America north of Limusa, Me´xico, D.F. Mexico. A key to genera. Taxonomic Publications, South Rzedowski, J. 1993. Diversity and origins of the phanero- Lancaster, MA. gamic ßora of Me´xico, pp. 129Ð144. In T. P. Ramamoorthy, Kovach Computing Services. 1998. Multi-variate statistical R. Bye, A. Lot, and J. Fa [eds.], Biological diversity of package, version 3.01 Anglesey, Wales, United Kingdom. Mexico: origins and distribution. Oxford University Press, Lawton, R. O., U. S. Nair, R. A. Pielke, Sr., and R. M. Welch. New York. 2001. Climatic impact of tropical lowland deforestation SAS Institute. 1985. SAS/STAT guide for personal comput- on nearby montane cloud forests. Science (Wash., D.C.) ers, version 6. SAS Institute, Cary, NC. 294: 584Ð587. Thomas, C. D. 1990. What do real population dynamics tell Lucky, A., T. L. Erwin, and J. D. Wittman. 2002. Temporal us about minimum viable population sizes? Conserv. Biol. and spatial diversity and distribution of arboreal Carabi- 4: 324Ð327. dae (Coleoptera) in a western Amazonian rainforest. Va´quez-Garcı´a, J. A. 1993. Cloud forest archipleagos: Biotropica 34: 376Ð386. preservation of fragmented montane ecosystems in Ludwig, J. A., and J. F. Reynolds. 1988. Statistical ecology: a tropical America, pp. 203Ð216. In L. S. Hamilton, J. O. primer on methods and computing. Wiley, New York. Juvik, and F. N. Scatena [eds.], Tropical montane cloud Luna Vega, I., O. Alca´ntara Ayala, D. Espinosa Organista, and forests. Proceedings of an international symposium. J. J. Morrone. 1999. Historical relationships of the Mex- San Juan de Puerto Rico. East-West Center Program on 138 ANNALS OF THE ENTOMOLOGICAL SOCIETY OF AMERICA Vol. 101, no. 1

Environment, UNESCO International Hydrological Wolf, J.H.D. 1993. Epiphyte communities of tropical mon- Programme, and U.S. International Institute of Tropical tane rain forest in the northern Andes. I. Lower montane Forestry. communities. Phytocoenologia 22: 1Ð52. Wake, D. B. 1970. The abundance and diversity of tropical salamanders. Am. Nat. 104: 211Ð213. Received 27 October 2006; accepted 9 September 2007. Downloaded from https://academic.oup.com/aesa/article/101/1/128/8463 by guest on 29 September 2021

Appendix 1. Species list and numbers and densities of weevils (Curculionidae) in leaf litter in the Huitepec Reserve, Chiapas, Mexico (systematic arrangement according to Alonso-Zarazaga and Lyal 1999)

No. individuals Total (N ϭ 247 samples Mature oak Cloud forest Successional Successional Species list 2 of 0.5 m ) forest total total forest edge forest total Total Mean STD (N ϭ 64) (N ϭ 80) total (N ϭ 64) (N ϭ 39) Otidocephalini Oopterinus laevigatus Champion 1 0.004 0.064 0 1 0 0 COSSONINAE Rhyncholini Heptarthrum sp. 1 69 0.279 1.229 20 48 1 0 CRYPTORHYNCHINAE Cryptorhynchini sp. 1 1 0.004 0.064 1 0 0 0 Eurhoptus sp. 1 15 0.061 0.300 10 1 1 3 Eurhoptus sp. 2 372 1.506 2.524 174 94 44 60 Eurhoptus sp. 3 121 0.490 1.059 31 41 36 13 Eurhoptus sp. 4 109 0.441 1.080 38 50 11 10 Eurhoptus sp. 5 117 0.474 1.019 30 17 51 19 Eurhoptus sp. 6 2 0.008 0.090 0 0 2 0 Cryptorhynchinae new genus 1 sp. nov. 1 4 0.016 0.155 0 4 0 0 Cryptorhynchinae new genus 1 sp. nov. 2 185 0.749 1.704 42 57 21 65 Cryptorhynchinae new genus 1 sp. nov. 3 51 0.206 0.512 14 24 6 7 Cryptorhynchinae new genus 1 sp. nov. 4 89 0.360 0.930 32 22 24 11 Cryptorhynchinae new genus 1 sp. nov. 5 29 0.117 0.421 0 15 8 6 Cryptorhynchinae new genus 2 sp. nov. 1 3 0.012 0.110 2 0 1 0 Cryptorhynchinae new genus 2 sp. nov. 2 1 0.004 0.064 1 0 0 0 Cryptorhynchinae new genus 3 sp. nov. 1 1 0.004 0.064 1 0 0 0 Oxypteropsis sp. 1 1 0.004 0.064 0 1 0 0 Tylodinus sp. 1 7 0.028 0.166 2 4 0 1 Tylodinus sp. 2 13 0.053 0.224 1 1 6 5 Tylodinus sp. 3 1 0.004 0.064 0 0 1 0 Tylodinus sp. 4 2 0.008 0.090 1 1 0 0 Tylodinus sp. 5 10 0.040 0.235 5 2 1 2 Tylodinus sp. 6 50 0.202 0.541 24 16 5 5 Tylodinus sp. 7 15 0.061 0.313 5 1 8 1 Tylodinus sp. 8 1 0.004 0.064 0 0 1 0 Tylodinus sp. 9 1 0.004 0.064 0 1 0 0 ENTIMINAE Sciaphilini Sciomias sp. 1 380 1.538 3.853 58 26 159 137 Sciomias sp. 2 2 0.008 0.090 0 0 2 0 Sciomias sp. 3 1 0.004 0.064 1 0 0 0 Trachyphloeini Trachyphloeomimus sp. 1 7 0.028 0.245 0 7 0 0 Trachyphloeomimus sp. 2 3 0.012 0.110 1 0 2 0 MOLYTINAE Anchonini Anchonus sp. 1 57 0.231 0.759 32 16 4 5 Anchonus sp. 2 33 0.134 0.462 13 15 4 1 Anchonus sp. 3 4 0.016 0.126 0 4 0 0 Anchonus sp. 4 11 0.045 0.226 8 3 0 0 Anchonus sp. 5 1 0.004 0.064 1 0 0 0 Anchonus sp. 6 4 0.016 0.180 2 0 2 0 Anchonus sp. 7 5 0.020 0.190 2 0 3 0 Anchonus sp. 8 2 0.008 0.090 2 0 0 0 Anchonus sp. 9 3 0.012 0.142 3 0 0 0 Levites sp. 1 94 0.381 0.916 38 32 12 12

Continued on following page January 2008 JONES ET AL.: LEAF LITTER WEEVILS IN TROPICAL MONTANE FORESTS 139

Appendix 1. Continued

No. individuals Total (N ϭ 247 samples Mature oak Cloud forest Successional Successional Species list 2 of 0.5 m ) forest total total forest edge forest total Total Mean STD (N ϭ 64) (N ϭ 80) total (N ϭ 64) (N ϭ 39) Conotrachelini Conotrachelus sp. 1 4 0.016 0.126 1 1 0 2 Conotrachelus sp. 2 32 0.130 0.403 2 6 20 4 Conotrachelus sp. 3 30 0.121 0.461 2 9 14 5 Conotrachelus sp. 4 5 0.020 0.141 2 1 2 0 Conotrachelus sp. 5 2 0.008 0.090 1 1 0 0 Conotrachelus sp. 6 32 0.130 0.598 3 5 21 3 Lepilius sp. 1 47 0.190 0.618 7 29 3 8 Lepilius sp. 2 1 0.004 0.064 1 0 0 0 Downloaded from https://academic.oup.com/aesa/article/101/1/128/8463 by guest on 29 September 2021 Microhyus sp. 1 2 0.008 0.090 0 0 1 1 Microhyus sp. 2 6 0.024 0.154 2 1 2 1 Lymantini Dioptrophorus prolexus Champion 3 0.012 0.142 0 0 3 0 Dioptrophorus verruciger Champion 11 0.045 0.243 2 2 5 2 Theognete sp. 1 46 0.186 0.708 0 20 19 7 Theognete sp. 2 2 0.008 0.090 2 0 0 0 Theognete sp. 3 1 0.004 0.064 0 1 0 0 Totals 2,102 8.510 9.038 620 580 506 396