Molecular taxonomy and species delimitation in Andean Schistocerca (Orthoptera: Acrididae) Amir Yassin a,1 , Christiane Amédégnato a , Corinne Cruaud b , Michel Veuille a,c, * a UMR CNRS 7205 Origine Structure and Evolution of Biodiversity, Muséum National d’Histoire Naturelle, 16 rue Buffon, 75005 Paris, France b Centre National de Séquençage Genoscope, 2 rue Gaston Crémieux, 91057 Evry Cedex, France c Laboratoire de Biologie intégrative des populations, Ecole Pratique des Hautes Etudes, France article info Article history: Received 28 October 2008 Revised 18 June 2009 Accepted 19 June 2009 Available online 24 June 2009 Keywords: DNA barcoding Locust Species concept Gene flow Population structure abstract The acridian genus Schistocerca comprises about 50 species which are endemic to the New World, except the Old World locust S. gregaria. Their morphological identification is rendered difficult by phase polyphenism, geographical overlap due to migrations or swarming, the difficulty to easily differentiate genitalia and the occurrence of interspecific hybrids. The three species reported from Peru include the swarming species S. interrita, a pest that can be recognized only by taxonomists. We show that it can be unambiguously identified using a mitochondrial DNA fragment known to have barcoding properties in this genus. We used several methods to delimitate Peruvian species. While S. interrita and S. pallens were well characterized, S. piceifrons peruviana was split into several taxa by a phylogeny-based method, whereas a combination of population genetics methods led one to identify only the three nominal spe- cies. A tentative reconstruction of the species history shows that several populations of S. piceifrons peru- viana have recently increased in number, while exchanging some migrants, whereas an isolated population at the northern margin of the species range is substantially differentiated while exchanging no migrants with the others. This complex history has resulted in an atypical lineage pattern that appears to have confounded the standard assumptions underlying available species delimitation methods. Because of its behavioral property which tends to keep it panmictic, the identification of the swarming S. interrita remained unaffected. Ó 2009 Elsevier Inc. All rights reserved. 1. Introduction The introduction of DNA sequence analysis as a reference meth- od in taxonomy (Hebert et al., 2003, 2004) was innovative in two ways. Firstly, DNA barcoding may lead to an easier identification of already described species. Second, DNA taxonomy may acceler- ate the discovery of new species that would otherwise remain cryptic (Markmann and Tautz, 2005; Vogler and Monaghan, 2007). Part of the taxonomic work (the identification of individuals and the discovery of new taxa) can be carried out by non-taxono- mists, even though the eventual decision to designate a new spe- cies and to describe it remains a key step in the hands of taxonomic experts. While DNA barcoding relies on the definition of species from the classical morphological and geographical crite- ria (DeSalle et al., 2005), DNA taxonomy goes one step further and introduces molecular criteria. These new criteria must account for the fact that, before becoming a specific trait shared by all chromo- somes in a taxon and being diagnostic, a new mutation must first survive, increase in frequency during a long polymorphism phase and eventually be fixed. Thus, population genetic approaches enter the field of taxonomy. This addresses the classical question of the consistency of species concepts between the different fields of evo- lutionary biology (Agapow et al., 2004). A theoretical examination of this issue by Hudson and Coyne (2002) shows that while mtDNA is an efficient marker to identify already characterized taxa, pro- vided there was no horizontal transfer, it may be a poor predictor of new taxa compared to more efficient, yet more costly methods, such as the sequencing of a number of independent nuclear loci. New data analysis methods have been developed to estimate migration rates between populations (Hudson, 2000; Nielsen and Wakeley, 2001; Hey and Nielsen, 2004), and have been applied to DNA barcoding (Matz and Nielsen, 2005; Knowles and Carstens, 2007). In their current state, these methods are very time-consum- ing and can be applied only under standard assumptions. They can provide substantial information on a limited number of taxa. Simplified procedures have been developed to make taxonomic decisions in large-scale empirical studies. Hebert et al. (2003) 1055-7903/$ - see front matter Ó 2009 Elsevier Inc. All rights reserved. doi:10.1016/j.ympev.2009.06.012 * Corresponding author. Address: Muséum National d’Histoire Naturelle, UMR 7205, Bâtiment 39, Département Systématique et Evolution, 16 rue Buffon, 75005 Paris, France. Fax: +33 1 40 79 33 42. E-mail address: veuille@mnhn.fr (M. Veuille). 1 Present address: American Museum of Natural History (AMNH), Sackler Institute for Comparative Genomics, Division of Invertebrate Zoology; 79th St. Central Park West, New York, NY 10024-5192, USA. Molecular Phylogenetics and Evolution 53 (2009) 404–411 Contents lists available at ScienceDirect Molecular Phylogenetics and Evolution journal homepage: www.elsevier.com/locate/ympev