Corresponding author: Jörg Spelda (
Academic editor: R. Mesibov
We give a first account of our ongoing barcoding activities on Bavarian myriapods in the framework of the Barcoding Fauna Bavarica project and IBOL, the International Barcode of Life. Having analyzed 126 taxa (including 122 species) belonging to all major German chilopod and diplopod lineages, often using four or more specimens each, at the moment our species stock includes 82% of the diplopods and 65% of the chilopods found in Bavaria, southern Germany. The partial COI sequences allow correct identification of more than 95% of the current set of Bavarian species. Moreover, most of the myriapod orders and families appear as distinct clades in neighbour-joining trees, although the phylogenetic relationships between them are not always depicted correctly. We give examples of (1) high interspecific sequence variability among closely related species; (2) low interspecific variability in some chordeumatidan genera, indicating that recent speciations cannot be resolved with certainty using COI DNA barcodes; (3) high intraspecific variation in some genera, suggesting the existence of cryptic lineages; and (4) the possible polyphyly of some taxa, i.e. the chordeumatidan genus
Molecular species identification based on sequence diversities in the Folmer segment of mitochondrial COI DNA has been under intense study for some years (
Among the Chilopoda and Diplopoda, the 146 species known from Bavaria cover 73% of the fauna of Germany. Hence, the first aim of our study is to establish a barcode reference library for Bavarian Myriapoda that will be expanded step by step (the dataset treated in this paper can be accessed in Barcode of Life Data Systems (BOLD;
In the present paper we give an overview of our ongoing barcoding activities, which so far cover 73% of all Bavarian Chilopoda and Diplopoda. In addition to conventional analysis of the actual dataset based on our BOLD data, we give examples of how our barcodes will contribute to taxonomic revisions and to analyses of past and ongoing speciation processes.
To cover the variability within species, numerous samples from locations inside and outside Bavaria have been included. Besides the centipedes and millipedes known to occur in Bavaria, species that might be found there in the future have also been included, as well as close relatives of the known Bavarian species (
Map of sampled areas (dots). For checks of intraspecific variability of COI sequences, localities in Bavaria, but also elsewhere within the species’ areas of distribution, have been sampled and analyzed (sampling data from November 2008 to November 2010; a few specimens from northern Spain omitted).
Sequencing was carried out at the CCDB, using the standard protocols of IBOL (
Sequencing failed for about 30% of the species. Sometimes whole genera (
Resulting data for the myriapods treated here are taken from the respective tools included in BOLD, and calculated using the Kimura 2 Parameter (K2P) model. Sequences were imported into PAUP* (
At the moment our myriapod barcode library includes 320 specimens, 122 species, 56 genera and 24 families of Myriapoda (
Some Bavarian myriapods for which barcodes are now available.
The mean sequence compositions in our sequences are G = 16.32%, C = 21.75%, A = 30.04% and T = 31.87% in Chilopoda, and G = 17.64%, C = 17.67%, A = 26.21% and T = 38.29% in Diplopoda. This shows a pronounced bias towards A and T, which is characteristic of arthropods.
In Chilopoda (
Interspecific COI variability (K2P): distance to nearest neighbour; Chilopoda
Interspecific COI variability (K2P): distance to nearest neighbour; Diplopoda
Intraspecific distances in Chilopoda (K2P maximum pairwise distance) ranged between 0%, for five species, and 21.55% for
Intraspecific COI variability (K2P): maximum pairwise distances; Chilopoda
Intraspecific COI variability (K2P): maximum pairwise distances; Diplopoda
Analysis of our data resulted in the Neighbour-joining (NJ) trees shown in
Complete neighbour-joining tree of COI sequence divergences (K2P model) of studied myriapod orders; barcoded terminal taxa and clades above their basal nodes omitted. This tree serves for orientation in the detailed trees given in
Neighbour-joining tree of COI sequence divergences (K2P model) of studied Polyxenida, Polydesmida and Glomerida. Solid circles: examples of excellent resolution of very close species of the genus
Though the mitochondrial COI gene is generally not seen as adequate for resolving relationships at taxonomic levels higher than species or genus, all barcoded myriapod orders (Polyxenida, Polydesmida, Glomerida, Chordeumatida, Polyzoniida, Scolopendrida, Lithobiida and Geophilida) form single COI clades, except for the Julida, which form two clades (
Neighbour-joining tree of COI sequence divergences (K2P model) of studied Julida. Note well-supported COI groups for each species allowing for sequence-based species identification. Numbers above and below branches show bootstrap values of NJ analysis, branch length indicates sequence divergence in %.
Neighbour-joining tree of COI sequence divergences (K2P model) of studied Chordeumatida. Asterisk: deep barcoding divergence in
Most of the studied species appear as distinct COI clades. Barcoded species can overlap for two reasons. First, speciation may have been very recent, e.g. during Pleistocene glacial episodes, as is the case for the diplopod genera
At the species and genus levels, we found examples of both well and weakly supported species. For example,
Conversely, very low interspecific variation is found, e.g., in the chordeumatidan genera
Moreover, examples of high intraspecific variation can be found in several
Neighbour-joining tree of COI sequence divergences (K2P model) of studied Chilopoda. Asterisks: Deep divergences within
Despite the success of COI barcoding in so many species of centipedes and millipedes it has to be admitted that there are still technical problems with this method that make the success of the barcoding process for any single sample unpredictable. For reasons of cost efficiency the CCDB presently uses only one set of standard primers that are probably not optimal for all groups of centipedes and millipedes. For example, we have failed so far to get any sequences in the genera
Although COI barcoding has provided an excellent tool for the identification of all life stages in several species, there are some problems with this gene locus as it is of mitochondrial origin. This means that it only shows maternal inheritance; therefore different maternal lines might mock cryptic species. This mainly affects the Chilopoda, which show a much higher genetic variability than the Diplopoda. While the histogram of intraspecific distances of the Diplopoda (
Recent speciations of glacial or postglacial origin with ongoing hybridization and introgression are impossible to resolve using barcodes, as apparently shown by the genera
Our results show that DNA barcoding can be a highly effective tool for the identification of Chilopoda and Diplopoda, provided that the right primers are designed and the right protocol is used. Before it can be better used, a reference barcode library is needed, the genetic variation must be known, and a close partnership between researchers with taxonomic expertise and those with a background in molecular analysis should be established for the interpretation of the results.
The Barcoding Fauna Bavarica project is financially supported by the Bavarian Ministry of Science, Research and Art (Bayerisches Staatsministerium für Wissenschaft, Forschung und Kunst), Munich, Germany. But our work would still have been difficult without the generous consent to use their facilities from the Canadian Centre for Barcoding at the Department of Zoology, University of Guelph, Ontario, Canada, financed by Genome Canada through the Ontario Genomics Institute. We thus have to thank first and foremost the project leaders, Gerhard Haszprunar (Munich) and Paul D. N. Hebert (Guelph), for their support.
For lots of discussion, processing and coordination, our thanks go to Axel Hausmann, Lars Hendrich, Michael Balke, Stephan Schmidt, Michael Miller (all Munich), Greg Singer (Guelph), and David Porco (Rouen). Julian Augusteyns, Nathalie Bäumler, Stefan Friedrich (all Munich), Jürgen Gruber (Vienna), Karin Voigtländer (Görlitz), and Thomas Wesener (Bonn) are thanked for their help with providing specimens for our study. Special thanks go to Martin Spies (Munich) and Robert Mesibov (Launceston) for polishing the English.