We sampled 14,603 geometrid moths along a forested elevational gradient from 1020–3021 m in the southern Ecuadorian Andes, and then employed DNA barcoding to refine decisions on species boundaries initially made by morphology. We compared the results with those from an earlier study on the same but slightly shorter gradient that relied solely on morphological criteria to discriminate species. The present analysis revealed 1857 putative species, an 80% increase in species richness from the earlier study that detected only 1010 species. Measures of species richness and diversity that are less dependent on sample size were more than twice as high as in the earlier study, even when analysis was restricted to an identical elevational range. The estimated total number of geometrid species (new dataset) in the sampled area is 2350. Species richness at single sites was 32–43% higher, and the beta diversity component rose by 43–51%. These impacts of DNA barcoding on measures of richness reflect its capacity to reveal cryptic species that were overlooked in the first study. The overall results confirmed unique diversity patterns reported in the first investigation. Species diversity was uniformly high along the gradient, declining only slightly above 2800 m. Species turnover also showed little variation along the gradient, reinforcing the lack of evidence for discrete faunal zones. By confirming these major biodiversity patterns, the present study establishes that incomplete species delineation does not necessarily conceal trends of biodiversity along ecological gradients, but it impedes determination of the true magnitude of diversity and species turnover.
Citation: Brehm G, Hebert PDN, Colwell RK, Adams M-O, Bodner F, Friedemann K, et al. (2016) Turning Up the Heat on a Hotspot: DNA Barcodes Reveal 80% More Species of Geometrid Moths along an Andean Elevational Gradient. PLoS ONE 11(3): e0150327. doi:10.1371/journal.pone.0150327
Editor: Jesus E. Maldonado, Smithsonian Conservation Biology Institute, UNITED STATES
Received: November 6, 2015; Accepted: February 11, 2016; Published: March 9, 2016
Copyright: © 2016 Brehm et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: Details on analysed specimens, including voucher images, specimen data with GPS coordinates and collection details are publicly available in BOLD and can be accessed under (dx.doi.org/10.5883/DS-ECUA). All further relevant data are within the paper and its Supporting Information file.
Funding: This study was supported by Basler Stiftung für biologische Forschung: (http://www.moneyhouse.ch/u/basler_stiftung_fur_biologische_forschung_CH-270.7.000.069-8.htm); Genome Canada through the Ontario Genomics Institute: (http://www.ontariogenomics.ca/about-us/partners/genome-canada); Deutsche Forschungsgemeinschaft, FOR 816, FOR 402, Fi 547/10-1 and 10-2: (http://www.dfg.de/en); Synthesys, GB TAF1048: (http://www.synthesys.info); CAPES Ciência sem Fronteiras: (http://www.cienciasemfronteiras.gov.br/web/csf). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Changing patterns in plant and animal communities along elevational gradients have interested biogeographers and macroecologists since Humboldt’s pioneering studies in the Andes 200 years ago. It was long assumed that species richness would decline monotonically with elevation, reflecting decreases in temperature and primary productivity  However, species richness has actually been shown to peak at intermediate elevations in 70% of past investigations [1, 2]. Only 20% of prior studies have shown the predicted monotonic richness declines with increasing elevation or otherwise divergent patterns. The dominance of hump-shaped richness patterns was long overlooked due to confounding area effects  and because richness often peaks at relatively low elevations but not at the lowest sites as, for example, in dung beetles  and ants .
 found a different pattern in the Andes of southeastern Ecuador: Richness and diversity of geometrid moths showed no change along an elevational gradient from 1020m to 2677m. Although this study did not examine diversity in the lowlands or at the highest elevations, the lack of an association with elevation was novel; it has not been observed in moths at other sites, over similar partial gradients, or in other groups of organisms. For example, geometrid species richness and diversity showed a pronounced hump-shape along a gradient from 40m to 2730m in Costa Rica . It has also become evident that the study area in Ecuador represents, to date, the most species-rich region for geometrid moths worldwide, with the latest regional count at 1445 species .
The analysis of very species-rich assemblages of tropical arthropods confronts several challenges. Among other problems, for example, the identification of species is often impeded by the lack of experienced taxonomists [9, 10, 11] and by cryptic species that cannot be discriminated morphologically [12, 13], particularly if specimens are damaged or worn. New methodological approaches can help to overcome these problems. In particular, the analysis of sequence diversity in the barcode region of the mitochondrial cytochrome c oxidase I gene (COI) has proven to be a powerful tool for clarifying species boundaries when combined with traditional morphospecies sorting (e.g. [14, 15, 16]), including for studies of Geometridae [17, 18].
The present study relies upon Barcode Index Numbers (BINs), a persistent species-level taxonomic registry based on the analysis of patterns of sequence variation in the barcode region, to delineate species . The study does not focus on the comparison of the results from BIN analysis with those obtained through other delineation methods, such as studies employing different sequence thresholds or the Yule-coalescent (GMYC), because results from the application of these methods are usually very similar [20, 21].
The use of DNA barcoding usually increases species richness and diversity values beyond those recognized by morphological analysis, but the extent of this increase varies considerably. Studies on groups with well-developed taxonomy have typically revealed less than a 10% increase in species richness [17, 22, 23], but two-fold or greater increases have been reported in other groups, especially those with cryptic differences in morphology [24, 25]. Large changes in species counts might shift ecological or biogeographical patterns from those based on morphology alone, especially if the incidence of cryptic species varies among taxa or along environmental gradients.
Our study evaluates the impact of using differing species delineation methods to examine patterns of species richness and turnover in geometrid moths along an elevational gradient in Ecuador. We achieve this goal by comparing the results from two sampling programs that examined the diversity of these moth communities along the same gradient with identical sampling methods. In particular, this study sought to determine the extent of the increase in local richness and diversity resulting when morphological analysis was augmented with DNA barcoding. Furthermore, we examine how figures describing beta diversity shift with the two methods. Finally, the study considers how DNA barcode results may affect perceived patterns of species turnover along this elevational gradient, testing if diversity is stable with increasing elevation, as reported in the first study.
Material and Methods
Moths were sampled by light-trapping along a continuously forested elevational gradient in southeastern Ecuador using methods previously described e.g. by . All sampling sites were in the northwestern section of Podocarpus National Park or in the adjacent Reserva Biológica San Francisco (RBSF) in the provinces of Zamora-Chinchipe and Loja (Fig 1). Collection of specimens occurred under research permits No. 023-2011-IC-FLO-DPL-MA issued by the Ministerio del Ambiente of Ecuador. None of the species is protected by international regulations such as CITES.  and  provide an introduction to the ecosystems and vegetation types in the study region. The maximum distance between the sampling sites was about 35 km, and most sites were clustered in the RBSF area in the north (Fig 1). The first study examined 22 sites [6, 29, 30] while the second sampled 28 sites, reflecting new high elevation locations that extended the elevational gradient by about 20%.
See S1 Table for detailed information on all sites. “+” indicates that samples from these sites were pooled for analysis.
The first study, which analysed specimens collected on three field trips in 1999 and 2000, used only morphology to discriminate species. The second study examined specimens collected on three field trips in 2008, 2011 and 2013 and employed both morphology and DNA barcoding to discriminate species. Specimens from both collections were used to assess gamma diversity, and other results are analyzed both with and without the new high elevation sites.
All specimens were assigned to a subfamily, most to a genus, and some to a species. However, most species identifications were provisional because of the large numbers of undescribed species , and the presence of many species with close morphological similarity (see below).
Ten sites were used for quantitative analyses of local diversity, as well as eight pairs of adjacent sites from which samples were combined because local catches were small or a site was sampled just once. These merged sites were always from the same habitat type and differed less than 100 m in elevation. S1 Table details all sampling sites. All sites were undersampled, as many species expected to be present at each site were not collected, but this situation is almost inescapable in extremely species-rich assemblages . Therefore, measures of richness and diversity need to be employed that reduce undersampling bias. However, we did not expect any serious qualitative effects of undersampling on the results presented in this study since all sites are similarly affected (sample coverage sensu , mean ± 1 SD: 0.78 ± 0.10).
Specimens in the second study were first sorted morphologically, and several representatives (when available) from each morphological group were selected for barcode analysis. A total of 3871 specimens were analysed at the Canadian Centre for DNA Barcoding, where a 658bp barcode region of the COI gene was amplified and sequenced using standard protocols  from DNA extracts prepared from single legs (small species) or tarsi (large species) . Because of the extremely high number of species and limited resources, we used a cost-effective, stepwise approach designed to gather as many different operational taxonomic units (OTUs) as possible with a minimum number of specimens analysed. Further details, including voucher images, specimen data with GPS coordinates and collection details are publicly available in BOLD and can be accessed under dx.doi.org/10.5883/DS-ECUA. Most specimens (98%) generated a sequence >600bp and 90% delivered a 658bp read. Barcode Index Numbers (BINs) were generated from 91% of the morphospecies collected in the second study. In the remaining 9%, most species were sorted in correspondence with morphospecies recognized in the first study or from reared material. Very few species were only sorted by morphology, mostly uncommon taxa where efforts to recover a barcode from a single specimen failed. Many specimens assigned to different BINs possessed conspicuous phenotypic differences that were recognized in the initial morphological studies, but others possessed subtle differences that became apparent only after barcode analysis.
Once the barcode results were available, all specimens from the second study that were not barcoded (approximately 74% of the total catch) were sorted according to the revised species delineations. In the few cryptic species complexes where no morphological differences could be detected (< 5% of the species), specimens were assigned to a particular species based on the elevational distribution of barcoded specimens from that species complex. This approach undoubtedly led to some incorrect assignments, but such cryptic species complexes were so infrequent that they could not affect the overall results in a substantial way.
We consider three sets of records: all data from the first study, all data from the second study, and all data from the second study excluding the new high elevation sites (the third approach was used only for regional comparisons). Diversity was examined at two spatial scales: individual sites (alpha), and all sites (gamma). Exponential Shannon diversity (classic formula) and Fisher's alpha were calculated for each site, and aggregated over all sites in a data set, using the software SPADE ). Both these parameters tend to be relatively robust even with undersampled, rich communities . We additionally compared species richness across all sites with two rarefaction approaches. First, we rarefied species richness to the smallest of the three aggregated data sets (i.e., to 10,306 individuals). Second, we rarefied richness to a fixed sample coverage of 0.96. Coverage-based rarefaction makes use of information more effectively than individual-based approaches if samples are drawn from communities of very different richness . At the level of local sites, species richness was estimated at a fixed coverage of 0.80. This level was selected since it meant that a similar number of sites needed to be extrapolated or rarefied. Calculations were done using iNEXT .
Gamma diversity was partitioned  into its alpha and (multiplicative) beta components (sensu ) based on observed species richness (diversity of the order 0) and exponential Shannon diversity (diversity of the order 1, following . The relationships between local richness or diversity and elevation were examined using standard linear regression. Finally, we assessed species turnover using an unconstrained ordination method . We expressed pairwise faunal similarities between sites as Chao-Soerensen index values to account for the large numbers of unseen species expected at all sites  (calculations performed using EstimateS 9.1.0: ). The resultant similarity matrix was then ordinated by means of non-metric multidimensional scaling (NMDS) using the software PRIMER . Two-dimensional representations were employed due to their low stress values (< 0.09). The scores of sites along the first ordination axis were then extracted and tested for their correlations with elevation.
Regional richness and diversity
The number of specimens collected in the second study was just 5% higher than in the first study (Table 1). Despite this small difference, 80% more species were recognized in the second study (1857 versus 1010). The 21% increase in length of the gradient in the second study accounted for little of this difference. When the high elevation sites were excluded, and sample sizes were rarefied to 10,306 specimens for both studies, the increase in species richness was still 75% (Table 1, Fig 2). The number of singletons in the second study was very high, with 545 among 1857 species (29%). Despite the extremely high species numbers, the high proportion of singletons suggests a far higher true richness than observed values. As a consequence of the improved species resolution enabled by DNA barcoding, the single most common species (Eupithecia duena (Dognin)) included only 179 individuals, just 1.2% of all samples combined. By comparison, the most common species in the first study was represented by 379 individuals, but the barcode analysis revealed that the species involved, Eois azafranata (Dognin), is a complex. Diversity measures that are less prone to sample-size effects showed an even greater increase in the second study. Fisher’s alpha, exponential Shannon diversity as well as species richness rarefied by sample coverage were all more than twice as high in the second study (Table 1).
Symbols in each curve refer to observed total sample size. The grey vertical line indicates a rarefaction level of 10,306 specimens.
Partitioning this exceptionally high gamma diversity into its alpha and beta components revealed two patterns (Fig 3). First, when considering species richness (diversity of the order 0), the local (alpha) fraction was always considerably less than the corresponding Shannon measures (diversity of the order 1). This pattern largely reflects the impact of the many singletons, which necessarily occur at only one site. Although they strongly affect richness, they have little impact on Shannon diversity values. Second, with both measures the local (alpha) fraction substantially declined in the second study. This decline indicates that the elevational ranges of individual species were smaller in the second study, increasing the beta component of diversity relative to alpha. With species richness, the beta component increased by 43% in the new data, whereas for Shannon diversities the increase was 51%.
For both data sets, the local alpha component is, on average, smaller for species richness than for Shannon diversity. In the second data set, with far more observed species and far more species recorded as singletons, the relative contribution of the beta component is substantially larger than in the first data set, which was based on morphospecies delimitations only.
Local richness and diversity
When considered at the level of local sampling sites, all metrics (richness, Shannon diversity, Fisher’s alpha) were substantially greater in the second study than in the first (Table 2). When plotted against elevation, none of the three richness or diversity measures was significantly associated with elevation (Fig 4), in the second study (max(|r|) = 0.29; p > 0.24) nor in the first (max(|r|) = 0.36; p > 0.10). There was also no evidence for either a hump-shaped diversity pattern or a monotonic decline in richness or diversity with elevation. On the contrary, for exponential Shannon diversity, there was a slight (although insignificant) increase at higher elevations. Regression lines fitted to both datasets were parallel for all three measures.
The slopes of all regression lines (fitted by ordinary least squares regression) are not significantly different from zero, indicating that geometrid moth diversity is not related to elevation. Sites that were sampled in both studies are outlined in black. a) expected species richness (at coverage 0.8), b) exponential Shannon diversity, exp (H’), c) Fisher’s alpha.
Given are means and standard deviations.
Species turnover patterns
Unconstrained ordinations of both data sets revealed a clear elevational segregation of samples along dimension 1, which was nearly identical in strength irrespective of the approach used for species recognition (second study: r = 0.99; first study: r = 0.98, p < 0.001). A horseshoe pattern, a typical artifact of ordinations dominated by one single environmental factor (e.g., [42, 46]) was visible in both ordinations (Fig 5). Neither data set indicates the existence of discrete faunal zones, i.e. a narrow clustering of sites in reduced space.
A hotspot even hotter
This study reinforces earlier evidence for the unmatched species richness of geometrid moths in southern Ecuador.  recognized 1266 species in southern Ecuador based on morphological studies, while  raised the total to 1445 putative species. The present study raises the count again to 1857 putative species, and the high fraction of singletons (almost 1/3 of the species) indicates that many more species await detection. In fact, the Chao1 estimator suggests the occurrence of 2350 ± 57 (SE) species in the elevational gradient from 1000 to 3000 m. More species are likely to occur at lower elevations and above 3000 m, and additional diurnal species from genera such as Erateina, Heterusia or Smicropus occur in the in the area . Hence, in this small area of the Ecuadorian Andes, richness of the regional species pool of Geometridae is more than twice as high than in all of Europe (<1000 species: ) or the 1098 species known to occur in Borneo . The application of DNA barcoding has again highlighted the extraordinary biological diversity of the tropical Andes. The new results reinforce earlier conjectures that this region is the global diversity hotspot for geometrid moths . Unfortunately, it is highly threatened, by the ongoing, rapid destruction of forest habitats in the region .
A taxonomic revision based on past collections is in progress (GB, unpublished results) and will eventually clarify the full observed richness of geometrid moths in the study area, based on the combined results of DNA barcoding and morphology. Whether other elevational gradients in the wet tropical Andean forests of Colombia, Ecuador, Peru, and Bolivia possess comparable richness remains to be explored. Despite the increase in species richness and diversity, the observed pattern of near-constancy in species richness between 1000 m and 3000 m remains unaltered. Conclusions on the role of environmental parameters , on species turnover , and on decreasing phylogenetic diversity along the elevational gradient  thus remain valid. This is remarkable, since an almost doubled recognized species pool as well as a far larger fraction of species represented only as singletons would have had the potential to substantially alter emerging patterns. We conclude that cryptic diversity occurred evenly across the elevational gradient and we are not aware of comparable studies that have shown the same. From the viewpoint of ecologists and biogeographers, our observation is reassuring since the validity of earlier results is supported, even though the magnitude of richness and diversity was severely underestimated.
Reasons for increased richness and diversity
The sharp increase in richness and diversity values in the current study is largely due to the adoption of DNA barcoding. This increase was observed both at the regional and, to a lesser extent, at the local scale. The smaller increase in local, than in regional diversity is linked to the increased and very high contribution of the beta component to regional species diversity, i.e. an even higher species turnover between sites than previously documented (Fig 3). Prior studies have established that cryptic species generate most beta diversity in European butterfly assemblages . It must be emphasized, however, that undersampling (as indicated by the large proportion of singletons) tends to bias species turnover upwards, overestimating faunal differences between sites .
DNA barcoding has revealed many new operational taxonomic units (OTUs) in previously unrecognised species complexes of cryptic or morphologically very similar species [52, 53]. In many cases, this reflects the split of what was previously thought to be a more broadly distributed species into two or more OTUs, often with differing elevational distributions, but frequently with some overlap. Often, only one species in these species complexes has been formally described, with no morphological differential diagnosis to separate it from the rest of the OTUs in the species complex. In such cases, the assignment of this name to a particular OTU is a challenge, but advances in sequencing old type specimens can hopefully help resolve this uncertainty [24, 54].
Complexes of morphologically similar species were often lumped in the first study, because the number of available specimens for comparison was low and there was less experience in species delineation at the beginning of the project. Moreover, larger series of specimens can better be judged than single specimens. There also has been a tendency to treat species with limited morphological differences as synonyms in major reference collections such as the Natural History Museum (London). Many complexes of lumped geometrid species can nowadays be segregated in the light of combined molecular and morphological results (e.g. ). Often, the members of such species complexes also differ in some other dimensions of their ecological niches such as larval hostplants (e.g. ).
Sampling design may have contributed slightly to the observed increase in species richness. In the second study, sites were more equally spaced, and the elevational range between 1020 and 1800 m was better covered than in the earlier study. Moreover, a greater variety of habitats was sampled than before, but only in a narrow band between 1800 m and 2000 m. Preliminary observations suggest some species turnover to exist between, for example, valley and ridge sites (for tropical butterflies see e.g. ). The role of such small scale gradients for moths still awaits quantification. However, the extension of the upper limit of the elevational gradient by 20% played only a minor role, as demonstrated by similar values of gamma diversity with and without these sites.
DNA barcoding as a powerful tool in biodiversity research
DNA barcoding has been validated as a powerful tool for species delineation in numerous studies, but it is particularly useful in species-rich tropical arthropod communities with their high numbers of undescribed and cryptic species. Decisions on species boundaries based on BINs or fixed sequence thresholds do not perfectly align with species boundaries when species hybridize  or when they are young . However, such cases are clearly uncommon, as evidenced by the results of work on taxonomically well-known faunas [17, 60, 61, 62].
To our knowledge, this dataset represents the largest number of (putative) species from a single animal order to be analysed from such a small geographic region. For example,  studied 6500 specimens of Trigonopterus weevils across New Guinea and found 270 morphospecies and 324 genetic clusters an increase of 20% species with barcoding.  barcoded 2597 parasitoid microgastrine wasps from northwestern Costa Rica, resulting in 171 species sorted by morphological analysis and a final delineation of 312 species (an increase of 83% species with barcoding).
DNA barcoding of poorly known tropical arthropod faunas undoubtedly increases the quality and reliability of biodiversity assessments. Additional costs are involved in barcoding, but they are justified by the greater accuracy in estimation of richness and diversity values compared with studies based solely on morphology. As DNA barcode workflows usually require the deposition of voucher specimens in natural history collections, together with metadata and photographs, this wealth of information will greatly facilitate subsequent detailed taxonomic study, allowing comparisons among studies and helping to resolve the fundamental question of how many species there are on Earth.
S1 Table. Supporting information on sites, species, and diversity measures.
Sheet 1: All sites (first and second studies) with coordinates and elevation, sheet 2: first study: species-site matrix, sheet 3: second study: species-site matrix, sheet 4: diversity measures.
We thank Scott Miller (Washington, D.C.) and an anonymous reviewer for helpful comments on our manuscript. Sequence analysis was enabled by funding from Genome Canada through the Ontario Genomics Institute in support of the International Barcode of Life Project and by the Basler Stiftung für biologische Forschung, Basel, Switzerland. Additional sequence data for the genus Eois were generated by P. Strutzenberger (Dresden) and B. Gottsberger (Vienna). Data analysis was aided by the Barcode of Life Data System, which is supported by the Ontario Ministry of Research and Innovation. We thank the Ministerio del Medio Ambiente of Ecuador for issuing research permits, the Universidad Técnica Particular de Loja for logistic support, and the foundation Nature and Culture International (Loja, Del Mar) for providing access to the Reserva Biológica San Francisco. Lisamarie Lehner (Vienna) and Yoko Matsumura (Kiel) kindly assisted with field work and with moth sampling and processing. Roman Link (Göttingen) and Shaun Levick (Jena) kindly helped produce the map of the study area. Support from DFG grants (FOR 816, FOR 402, Fi 547/10-1 and 10–2), through funds from the Faculty of Life Sciences (Univ. of Vienna), and the SYNTHESYS programme (GB-TAF1048) is acknowledged. RKC acknowledges support from CAPES Ciência sem Fronteiras (Brazil).
Conceived and designed the experiments: GB. Performed the experiments: GB FB K. Friedemann LM MA. Analyzed the data: GB RKC PDNH K. Fiedler. Contributed reagents/materials/analysis tools: GB K. Fiedler PDNH. Wrote the paper: GB PDNH RKC FB MA K. Fiedler LM K. Friedemann.
- 1. Rahbek C. The elevational gradient of species richness: a uniform pattern? Ecography 1995;19:200–205. doi: 10.1111/j.1600-0587.1995.tb00341.x
- 2. Guo Q, Kelt DA, Sun Z, Liu H, Hu L, Ren H, et al. Global variation in elevational diversity patterns. Sci Rep. 2013;3: 3007. doi: 10.1038/srep03007. pmid:24157658
- 3. Rahbek C. The relationship between area, elevation and regional species richness in Neotropical birds. Am Nat. 1997;149: 875–902. pmid:18811253 doi: 10.1086/286028
- 4. Herzog SK, Hamel-Leigue AC, Larsen TH, Mann DJ, Soria-Auza RW, Gill BD, et al. Elevational distribution and conservation biogeography of phanaeine dung beetles (Coleoptera: Scarabaeinae) in Bolivia. PLoS ONE 2013;8: e64963. doi: 10.1371/journal.pone.0064963. pmid:23717678
- 5. Longino JT, Branstetter MG, Colwell RK. How ants drop out: ant abundance on tropical mountains. PLoS ONE 2014;9: e104030. doi: 10.1371/journal.pone.0104030. pmid:25098722
- 6. Brehm G, Süssenbach D, Fiedler K. Unique elevational diversity patterns of geometrid moths in an Andean montane rainforest. Ecography 2003;26: 456–466. doi: 10.1034/j.1600-0587.2003.03498.x
- 7. Brehm G, Colwell RK, Kluge J. The role of environment and mid-domain effect on moth species richness along a tropical elevational gradient. Glob Ecol Biogeogr. 2007;16: 205–219. doi: 10.1111/j.1466-8238.2006.00281.x
- 8. Brehm G, Strutzenberger P, Fiedler K. Phylogenetic diversity of geometrid moths decreases with elevation in the tropical Andes. Ecography. 2013;36: 1247–1253. doi: 10.1111/j.1600-0587.2013.00030.x
- 9. Hebert PDN, Cywinska A, Ball SL, deWaard JR. Biological identifications through DNA barcodes. Proc R Soc B. 2003;270: 313–321. pmid:12614582 doi: 10.1098/rspb.2002.2218
- 10. Bortolus A. Error cascades in the biological sciences: the unwanted consequences of using bad taxonomy in ecology. AMBIO 2008;37: 114–118. pmid:18488554 doi: 10.1579/0044-7447(2008)37[114:ecitbs]2.0.co;2
- 11. Hortal J, de Bello F, Diniz-Filho JAF, Lewinsohn TF, Lobo JM, Ladle RJ. Seven shortfalls that beset large-scale knowledge on biodiversity. Ann Rev Ecol Evol Syst. 2015; 46: doi: 10.1146/annurev-ecolsys-112414-054400
- 12. Bickford D, Lohmann DJ, Sodhi NS, Ng PK, Meier R, Winker K et al. Cryptic species as a window on diversity and conservation. Trends Ecol Evol. 2007;22(3): 148–155. pmid:17129636 doi: 10.1016/j.tree.2006.11.004
- 13. Dincă V, Montagud S, Talavera G Hernández-Roldán J, Munguira ML, García-Barro E et al. DNA barcode reference library for Iberian butterflies enables a continental-scale preview of potential cryptic diversity. Sci Rep. 2015;5: 12395. doi: 10.1038/srep12395. pmid:26205828
- 14. Kekkonen M, Mutanen M, Kaila L, Nieminen M, Hebert PDN. Delineating species with DNA barcodes: a case of taxon dependent method performance in moths. PLoS ONE 2015;10: e0122481. doi: 10.1371/journal.pone.0122481. pmid:25849083
- 15. Hernandez-Triana LM, Chaverri LG, Rodriguez-Perez MA, Prosser SWJ, Hebert PDN, Gregory TR, et al. DNA barcoding of Neotropical black flies (Diptera: Simuliidae): Species identification and discovery of cryptic diversity in Mesoamerica. Zootaxa. 2015;3936: 93–114. doi: 10.11646/zootaxa.3936.1.5. pmid:25947423
- 16. Thormann B, Ahrens D, Armijos DM, Peters MK, Wagner T, Wägele JW. Exploring the leaf beetle fauna (Coleoptera: Chrysomelidae) of an Ecuadorian mountain forest with DNA barcoding PloS ONE 2016;11(2): e0148268. doi: 10.1371/journal.pone.0148268. pmid:26849826
- 17. Hausmann A, Haszprunar G, Hebert PDN. DNA barcoding the geometrid fauna of Bavaria (Lepidoptera): Successes, surprises, and questions. PLoS ONE 2011;6(2): e17134. doi: 10.1371/journal.pone.0017134. pmid:21423340
- 18. deWaard JR, Hebert PDN, Humble LM. A comprehensive DNA barcode library for the looper moths (Lepidoptera: Geometridae) of British Columbia, Canada. PLoS ONE 2011;6(3): e18290. doi: 10.1371/journal.pone.0018290. pmid:21464900
- 19. Ratnasingham S, Hebert PDN. A DNA-based registry for all animal species: the Barcode Index Number (BIN) system. PLoS ONE 2013;8: e66213. doi: 10.1371/journal.pone.0066213. pmid:23861743
- 20. Fernández-Flores S, Fernández-Triana JL, Martinez JJ, Zaldívar-Riveron . A DNA barcoding species inventory of Microgastrinae wasps (Hymenoptera, Braconidae) from a Mexican tropical dry forest. Mol Ecol Resour. 2014;13: 1146–1150. doi: 10.1111/1755-0998.12102
- 21. Tänzler R, Sagata K, Surbakti S, Balke M, Riedel A. DNA barcoding for community ecology—How to tackle a hyperdiverse, mostly undescribed Melanesian fauna. PLoS ONE 2012;7: e28832. doi: 10.1371/journal.pone.0028832. pmid:22253699
- 22. Dincă V, Zakharov EV, Hebert PDN, Vila R. Complete DNA barcode reference library for a country’s butterfly fauna reveals high performance for temperate Europe. Proc R Soc B. 2010;278: 347–355. doi: 10.1098/rspb.2010.1089. pmid:20702462
- 23. Janzen DH, Hallwachs W, Harvey DJ, Darrow K, Rougerie R, Hajibabaei M et al. What happens to the traditional taxonomy when a well-known tropical saturniid moth fauna is DNA barcoded? Invertebr Syst. 2012;26: 478–505. doi: 10.1071/is12038
- 24. Strutzenberger P, Brehm G, Fiedler K. DNA barcode sequencing from old type specimens as a tool in taxonomy: a case study in the diverse genus Eois (Lepidoptera: Geometridae). PLoS ONE 2012;7: e49710. doi: 10.1371/journal.pone.0049710. pmid:23185414
- 25. Young MR, Behan-Pelletier V, Hebert PDN. Revealing the hyperdiverse mite fauna of subarctic Canada through DNA barcoding. PLoS ONE 2012;7: e48755. doi: 10.1371/journal.pone.0048755. pmid:23133656
- 26. Brehm G, Axmacher JC. A comparison of manual and automatic moth sampling methods (Lepidoptera: Arctiidae, Geometridae) in a rain forest in Costa Rica. Environ Entomol. 2006;35: 754–764. doi: 10.1603/0046-225x-35.3.757
- 27. Beck E, Makeschin F, Haubrich M, Richter M, Bendix J, Valerezo C. The ecosystem (Reserva Biológica San Francisco). In: Beck E, Bendix J, Kottke I, Makeschin F and Mosandl R, editors. Gradients in a tropical mountain ecosystem of Ecuador. Ecological Studies. 2008;198: pp. 1–13.
- 28. Homeier J, Breckle SW, Günter S, Rollenbeck RT, Leuschner C. Tree diversity, forest structure and productivity along altitudinal and topographical gradients in a species-rich Ecuadorian montane rain forest. Biotropica 2010;42: 140–148. doi: 10.1111/j.1744-7429.2009.00547.x
- 29. Brehm G. Diversity of geometrid moths in montane rainforests in Ecuador. PhD Thesis, University of Bayreuth. 2002. Available at http://opus.ub.uni-bayreuth.de/volltexte/2003/20/pdf/Dissertation_Gunnar_Brehm.pdf
- 30. Brehm G, Homeier J, Fiedler K. Beta diversity of geometrid moths (Lepidoptera: Geometridae) in an Andean montane rainforest. Div Distrib. 2003a;9: 351–366. doi: 10.1046/j.1472-4642.2003.00023.x
- 31. Brehm G, Bodner F, Strutzenberger P, Hünefeld F, Fiedler K. Neotropical Eois (Lepidoptera: Geometridae): checklist, biogeography, diversity, and description patterns. Ann Entomol Soc Am. 2011;104: 1091–1107. doi: 10.1603/an10050
- 32. Coddington JA, Agnarsson I, Miller JA, Kuntner M, Hormiga G. Undersampling bias: the null hypothesis for singleton species in tropical arthropod surveys. J Anim Ecol. 2009;78: 573–584. doi: 10.1111/j.1365-2656.2009.01525.x. pmid:19245379
- 33. Chao A, Jost L. Coverage-based rarefaction and extrapolation: standardizing samples by completeness rather than size. Ecology 2012;93: 2533–2547. pmid:23431585 doi: 10.1890/11-1952.1
- 34. deWaard JR, Ivanova NV, Hajibabaei M, Hebert PDN. Assembling DNA barcodes: analytical protocols. In: Martin C, editor. Methods in Molecular Biology: Environmental Genetics. Totowa: Humana Press. 2007. pp: 275–293.
- 35. Strutzenberger P, Brehm G, Fiedler K. DNA barcoding based species delimitation increases species richness count of (Geometridae) moths in a well-studied tropical mountain forest by up to 50%. Insect Sci. 2011;18: 349–362. doi: 10.1111/j.1744-7917.2010.01366.x
- 36. Chao A, Shen TJ. Program SPADE (species prediction and diversity estimation). Program and user’s guide. 2010. Available at http://chao.stat.nthu.edu.tw.
- 37. Beck J, Schwanghart W. Comparing measures of species diversity from incomplete inventories: an update. Methods Ecol Evol. 2010;1: 38–44. doi: 10.1111/j.2041-210x.2009.00003.x
- 38. Hsieh TC, Ma KH, Chao A. iNEXT online: interpolation and extrapolation (Version 1.0). 2013. Available at: https://chao.shinyapps.io/iNEXT/
- 39. Jost L. Partitioning diversity into independent alpha and beta components. Ecology. 2007;88: 2427–2439. pmid:18027744 doi: 10.1890/06-1736.1
- 40. Tuomisto H. A diversity of beta diversities: straightening up a concept gone awry. Part 1. Defining beta diversity as a function of alpha and gamma diversity. Ecography. 2010;33: 2–22. doi: 10.1111/j.1600-0587.2009.05880.x
- 41. Hill MO. Diversity and evenness: a unifying notation and its consequences. Ecology. 1973;54: 427–432. doi: 10.2307/1934352
- 42. Brehm G, Fiedler K. Ordinating tropical moth ensembles from an elevational gradient: a comparison of common methods. J Trop Ecol. 2004;20: 165–172. doi: 10.1017/s0266467403001184
- 43. Chao A, Chazdon RL, Colwell RK, Shen T-J. A new statistical approach for assessing compositional similarity based on incidence and abundance data. Ecol Lett. 2005;8: 148–159. doi: 10.1111/j.1461-0248.2004.00707.x
- 44. Colwell RK EstimateS 9.1. Statistical estimation of species richness and shared species from samples. 2013. Available at: http://purl.oclc.org/estimates.
- 45. Clarke KR, Gorley RN. PRIMER version 7: user manual/tutorial. 2015. PRIMER-E, Plymouth, UK.
- 46. Legendre P, Legendre LF Numerical ecology. 3rd ed. Elsevier. 2012.
- 47. Brehm G, Pitkin LM, Hilt N, Fiedler K. Montane Andean rain forests are a global diversity hotspot of geometrid moths. J Biogeogr. 2005;32: 1621–1627. doi: 10.1111/j.1365-2699.2005.01304.x
- 48. Hausmann A. 2001. The Geometrid moths of Europe. Volume 1: Introduction, Archiearinae, Orthostixinae, Desmobathrinae, Alsophilinae, Geometrinae. Apollo Books, Stenstrup, DK.
- 49. Holloway J.D. The moths of Borneo: families Phaudidae, Himantopteridae, and Zygaenidae; revised and annotated checklist. Malayan Nature J. 2011;63: 1–548.
- 50. Tapia-Armijos MF, Homeier J, Espinosa CI, Leuschner C, de la Cruz M. Deforestation and forest fragmentation in South Ecuador since the 1970s –losing a hotspot of biodiversity. PLoS ONE 2015;10; e0133701. doi: 10.1371/journal.pone.0133701. pmid:26332681
- 51. Vodă R, Dapporto L, Dincă V, Vila R. Cryptic matters: overlooked species generate most butterfly beta-diversity. Ecography. 2014;38: 405–409. doi: 10.1111/ecog.00762
- 52. Rougerie R, Kitching IJ, Miller SE, Hausmann A, Hebert PDN. Australian Sphingidae–DNA barcodes challenge current species boundaries and distributions. PLoS ONE 2014;9: e101108. doi: 10.1371/journal.pone.0101108. pmid:24987846
- 53. Wilson JJ, Landry JF, Janzen DH, Hallwachs W, Nazari V, Hajibabaei M et al. Identity of the ailanthus webworm moth (Lepidoptera: Yponomeutidae), a complex of two species: evidence from DNA barcoding, morphology and ecology. ZooKeys 2010;46: 41–60. doi: 10.3897/zookeys.46.406
- 54. Speidel W, Hausmann A, Muller GC, Kravchenko V, Mooser J, Witt TJ et al. Taxonomy 2.0: Sequencing of old type specimens supports the description of two new species of the Lasiocampa decolorata group from Morocco (Lepidoptera: Lasiocampidae) Zootaxa 2015;3999: 401–412. doi: 10.11646/zootaxa.3999.3.5. pmid:26623584
- 55. Brehm G. Three new species of Hagnagora Druce, 1885 (Lepidoptera, Geometridae, Larentiinae) from Ecuador and Costa Rica and a concise revision of the genus. ZooKeys 2015;537: 131–156. doi: 10.3897/zookeys.537.6090. pmid:26798242
- 56. Seifert C, Bodner F, Brehm G, Fiedler K. Host plant associations and parasitism of South Ecuadorian Eois species (Lepidoptera: Geometridae) feeding on Peperomia (Piperaceae). J Ins Sci. 2015;15: 119
- 57. Binz H, Schulze CH, Linsenmair KE. Effects of topography on forest butterfly assemblages in the Pacific lowlands of Costa Rica. Ecotropica 2015;20:1–14.
- 58. Schmidt BC, Sperling FAH. Widespread decoupling of mtDNA variation and species integrity in Grammia tiger moths (Lepidoptera: Noctuidae). Syst Entomol. 2008;33: 613–634. doi: 10.1111/j.1365-3113.2008.00433.x
- 59. Wiemers M, Fiedler K. Does the DNA barcoding gap exist?–a case study in blue butterflies. Frontiers Zool. 2007;4: 8. doi: 10.1186/1742-9994-4-8
- 60. Hebert PDN, deWaard JR, Landry JF. DNA barcodes for 1/1000 of the animal kingdom. Biol Lett. 2010;6: 359–362. doi: 10.1098/rsbl.2009.0848. pmid:20015856
- 61. Huemer P, Mutanen M, Sefc KM, Hebert PDN. Testing DNA barcode performance in 1000 species of European Lepidoptera: Large geographic distances have small genetic impacts. PLoS ONE 2014;9: e115774. doi: 10.1371/journal.pone.0115774. pmid:25541991
- 62. Pentinsaari M, Hebert PDN, Mutanen M. Barcoding beetles: A regional survey of 1872 species reveals high identification success and unusually deep interspecific divergences. PLoS ONE 2014;9: e108651. doi: 10.1371/journal.pone.0108651. pmid:25255319
- 63. Smith MA, Rodriguez JJ, Whitfield JB, Deans AR, Janzen DH, Hallwachs W et al. Extreme diversity of tropical parasitoid wasps exposed by iterative integration of natural history, DNA barcoding, morphology, and collections. Proc Nat Acad Sci USA. 2008;105: 12359–12364. doi: 10.1073/pnas.0805319105. pmid:18716001