Thormann, Birthe: Biodiversity of leaf beetles (Coleoptera: Chrysomelidae) in a tropical montane rainforest ecosystem assessed with DNA barcoding. - Bonn, 2016. - Dissertation, Rheinische Friedrich-Wilhelms-Universität Bonn.
Online-Ausgabe in bonndoc:
author = {{Birthe Thormann}},
title = {Biodiversity of leaf beetles (Coleoptera: Chrysomelidae) in a tropical montane rainforest ecosystem assessed with DNA barcoding},
school = {Rheinische Friedrich-Wilhelms-Universität Bonn},
year = 2016,
month = aug,

note = {The aim of the present study was the assessment of an unknown tropical insect fauna without traditional taxonomy. For this purpose, the diversity of leaf beetles (Coleoptera: Chrysomelidae) in the montane rainforest of the Reserva Biológica San Francisco (RBSF) and parts of the Podocarpus National Park in southern Ecuador was investigated. Beetles were sampled at three different elevations, ’1000 m’ (Bombuscaro; 1020–1075 m a.s.l.), ’2000 m’ (Estación Científica San Francisco – ECSF; 1913–2089 m a.s.l.), and ’3000 m’ (Cajanuma; 2805–2891 m a.s.l.) with a set of different collection methods. Beetles were mainly sampled with sweep netting, beating, and hand-collection from the lower understorey vegetation of 36 sampling plots (12 per elevation, six of them in the valleys, six on the ridges) following a standardized sampling protocol. A total of 4286 leaf beetles have been collected, 1775 of these (usually one of each morphospecies per sample) were sorted into 515 different morphospecies, DNA barcoded, and assigned to molecular operational taxonomic units (MOTUs).
The study covers aspects of community structure and its changes with increasing elevation. Methodological aspects of rapid biodiversity assessment are evaluated: Different collection methods and morphological and sequence-based methods for species delimitation are compared.
General leaf beetle diversity patterns in an Andean mountain forest
Leaf beetle assemblages showed patterns typical for tropical arthropods: They were species-rich, with few common species but a high percentage of rare species. 1583 specimens were sorted into 473 morphospecies, and for 1334 of them a DNA barcode could be obtained. They belong to 416 morphospecies and were grouped into 459 MOTUs. Species accumulation curves showed no saturation indicating a further increase in species numbers with additional sampling. Species number estimates ranged up to 916 morphospecies (chao2) for the 1583 analysed individuals, and 705 morphospecies, respectively 805 MOTUs for the 1334 barcoded individuals. The higher MOTU number compared to morphospecies number suggests a high level of potential cryptic diversity that was not recognized by the morphospecies approach alone. The leaf beetle community showed an uneven distribution of incidence and abundance with very few common morphospecies (5% found in more than ten samples, 10% represented by more than ten individuals) and a high percentage of uniques (morphospecies found in one single sample; 50% of all morphospecies), respectively singletons (one single individual found; 45% of all morphospecies). The singleton curve did not reach saturation. Most morphospecies were restricted to one single elevational level (91%), indicating a high turnover of communities with elevation. This pattern was even more apparent for MOTUs (96%) and haplotypes (99%). More than half of the morphospecies belonged to Alticinae (53%), 21% were Galerucinae, 14% Eumolpinae, 5% Hispinae, and 4% Cassidinae. Criocerinae, Chrysomelinae, Lamprosomatinae, and Cryptocephalinae together accounted for 3% of all morpho- species. Rank order remained the same when number of individuals was considered. Composition of the subgroups changed slightly with elevation.
Diversity patterns along an elevational gradient inferred with DNA barcode data
Leaf beetle assemblages from the 36 study plots were sampled and differences between the three elevations and the two microhabitats (forest on ridges and in valleys) were analysed based on DNA barcode data. The importance of small-scale topography for elevational diversity patterns was evaluated: It was tested whether elevational diversity differs between ridge and valley forests and if the species turnover between and within habitats varies with elevation and changes patterns of elevational diversity when scaling up from the local (sampling plot) to the regional (elevational belt) level. MOTUs were determined using PTP modelling and data was analysed using permutational MANOVA analysis and ordinary linear models. When study sites of both habitats were pooled, local leaf beetle diversity showed a clear mid- elevational peak pattern. However, only leaf beetle diversity in ridge forests peaked at mid- elevations, while the diversity in valley forests was similarly high at 1000 and 2000 m a.s.l. and declined at highest elevations. When scaling up to the regional scale, levels of diversity were generally similar at the two lower elevations and declined at 3000 m a.s.l. The scale-dependent shift in diversity patterns was caused by a higher turnover of species communities between and within habitats at lower than at mid-elevations, suggesting more specialized herbivore communities in the more productive lower elevations. The study underscores the importance of topography and spatial scale for the inference of diversity patterns. Changes in ecosystem productivity but also area and temperature with elevation might also influence the genetic diversity within species, however, levels of genetic diversity (haplotype diversity per MOTU) did not differ among elevational levels. Biodiversity patterns along the elevational gradient were revealed by MOTUs and morphospecies in the same way.
Comparison of morphospecies sorting and DNA barcoding
1475 barcoded individuals were assigned to MOTUs and the results were compared with the morphospecies sorting. The barcode approach estimated 10% higher species numbers (448 morphospecies, 493 MOTUs). This was caused by a higher number of splittings than lumpings of morphospecies. The similar numbers of morphospecies and MOTUs arose partly due to the fact that splittings and lumpings compensated one another. However, the number of perfect matches was comparatively low: 63% of all morphospecies corresponded exactly with one MOTU. Most lumpings united individuals of two morphospecies in one MOTU (76%), in some cases, individuals of up to five morphospecies (4%) were lumped. Similarly, most splittings divided a morphospecies in two networks (69%), only once a morphospecies was split into six MOTUs (1%). The subgroups most challenging for morphospecies sorting were Galerucinae and especially Alticinae. Difficulties most probably arose due to the large number of specimens and species. DNA barcoding showed to be a valuable tool in cases were morphospecies sorting is exacerbated by pronounced intraspecific variation in colour, shape, or size, and may reveal cryptic diversity. Especially in species that are small and/or lack conspicuous external characters barcoding is a useful tool to complement morphospecies sorting. Particularly in large, specimen- and species-rich data sets DNA barcoding can facilitate morphospecies sorting and can result into a more accurate species delimitation.
Influence of different species delimitation methods on species richness estimates
For a subset of 674 barcoded specimens, a set of four different DNA-based species delimitation methods and their influence on species richness estimates were compared. Distance-based clustering, statistical parsimony analysis, GMYC-, and PTP modelling led to highly similar results. The reason probably lies within the structure of the underlying data set: It is geographically restricted and undersampled with a high proportion of singletons what turns it insensitive against differences in species delimitation methods. Several cases of splittings and lumpings led to discrepancies between morphospecies and MOTU assignment and generally MOTU numbers were ~8% higher than morphospecies numbers. Morphospecies sorting and DNA barcoding allow similar conclusions on leaf beetle diversity: The leaf beetle fauna is species-rich with a strong turnover among elevations. Most morphospecies where found only at a single elevational level, also when singletons and doubletons have been excluded. This pattern was even more visible for MOTUs and haplotypes. The high turnover between leaf beetle communities at the different elevations is also visible in the species accumulation curves: If to the specimens of one elevation the specimens of a second elevation where added, the curves showed once more a further increase.
Comparison of sampling methods
Within the present study a total of 1174 samples were taken. They varied considerably in size and effort as different sampling methods were used. The focus was on standardized sampling with sweep netting, beating, and hand-collection on the sampling plots. Malaise trapping, light trapping, and additional hand-collection completed the sampling. In sweep netting-, beating-, hand-collection-, and light trap samples on average only few individuals and morphospecies were caught per single sample (less than five). In contrast, the Malaise traps were highly efficient on a per sample basis: They yielded a mean of 31 individuals and 15 morphospecies per sample. Collection efficiency for certain subgroups slightly differed between the different methods. Even after 298.5 sampling hours the species accumulation curve of the standardized plot samples showed no saturation indicating that a further increase of morphospecies number is expected with further sampling.},

url = {}

Die folgenden Nutzungsbestimmungen sind mit dieser Ressource verbunden: