Specimens were determined and collected by the authors of the study by hand, and either directly or after a few days transferred to vials containing 95% undenatured ethanol. The vials contain an individual GBOL number with which the specimens can be connected to the accompanying data. After conservation the specimens were either sent to the GBOL facility at the Museum Koenig, Bonn, Germany (ZFMK) or to the corresponding laboratory at the Bavarian State collection of Zoology, Munich, Germany (ZSM). Upon arrival, all specimens were photographed (images will be uploaded to BOLD, http://www.boldsystems.org/), and a tissue sample was removed for DNA extraction. All specimens will later be stored as vouchers in 95% undenatured ethanol, either at the ZFMK, the SMNG (Senckenberg Museum für Naturkunde, Görlitz) or the ZSM (see Table 1). For this specific GBOL subproject, DNA extraction was attempted for more than 35 specimens of Geophilus and 24 Stenotaenia, all specimens from Germany.

At the ZFMK, DNA was extracted from the tissue samples using the BioSprint96 magnetic bead extractor by Qiagen (Germany). After the extraction, samples were outsourced for PCR and sequencing (BGI China). For PCR and sequencing, HCO/LCO primer pairs (Folmer et al. 1994) were utilized. Because of a low PCR and sequencing success (<50%) for the Myriapoda, the degenerated primer pair HCOJJ/LCOJJ (Astrin and Stüben 2008) was used for further sequencing attempts, resulting in a much higher success rate (>75%). At the ZSM, a tissue sample was removed from each specimen and transferred into 96 well plates for subsequent DNA extraction at the Canadian Center for DNA Barcoding (CCDB) where they were processed using standard barcoding protocols. All protocols for DNA extraction, PCR amplifications and Sanger Sequencing procedures are available online under: http://www.dnabarcoding.ca/pa/ge/research/protocols.for DNA. DNA was extracted from the whole voucher at the CCDB. All samples were PCR amplified with modified Folmer primers CLepFolF, and the same primers were employed for subsequent Sanger sequencing. All voucher information and the DNA barcode sequences, primer pairs and trace files were uploaded to BOLD (http://www.boldsystems.org).

However, for more than five S. linearis and more than 10 Geophilus specimens no sequences could be obtained. Sequences were obtained for 19 Stenotaenia and 25 Geophilus specimens. Sequence identities were confirmed with BLAST searches (Altschul et al. 1997). All 44 new sequences were deposited in GenBank (see Table 1 for accession numbers). The only available COI sequence of Stenotaenia (KF569300.1), labelled as S. sorrentina, was added to the dataset. In order to rule-out the accidental amplification of nuclear copies of the mitochondrial COI gene, the whole dataset was translated into amino acids following the ‘invertebrate’ code in MEGA6 (Tamura et al. 2013); internal stop codons were absent in our dataset. There were a total of 658 positions in the final dataset, gaps were absent.

Sequences were aligned by hand in Bioedit (Hall 1999). The final dataset included 45 nucleotide sequences with 658 positions (44 newly sequenced and the one of S. ‘sorrentina’ from GenBank). Phylogenetic analyses were conducted in MEGA6 (Tamura et al. 2013). A Modeltest, as implemented in MEGA6 (Tamura et al. 2013), was performed to find the best fitting maximum likelihood substitution model. Models with the lowest BIC scores (Bayesian Information Criterion) are considered to describe the best substitution pattern. Codon positions included were 1st+2nd+3rd+Noncoding. Modeltest selected the Tamura-Nei model (Tamura and Nei 1993) with gamma distribution and invariant sites as best fitting model (lnL -4245.19958, Invariant 0.55674, Gamma 1.176355, R 3.46, Freq A: 0.288843, T: 0.282885, C: 0.262778, G: 0.16546).

The evolutionary history was inferred by using the maximum likelihood method based on the selected Tamura-Nei model (Tamura and Nei 1993). The tree with the highest log likelihood (-4247.0145) is shown (Nei and Kumar 2000). The percentage of trees in which the associated taxa clustered together is shown next to the branches. Initial tree(s) for the heuristic search were obtained automatically by applying Neighbor-Join and BioNJ algorithms to a matrix of pairwise distances estimated using the Maximum Composite Likelihood (MCL) approach, and then selecting the topology with superior log likelihood value. A discrete Gamma distribution was used to model evolutionary rate differences among sites (5 categories (+G, parameter = 1.1347)). The rate variation model allowed for some sites to be evolutionarily invariable ([+I], 55.5093% sites). The tree is drawn to scale, with branch lengths measured in the number of substitutions per site.

The number of pairwise base differences per site were calculated in MEGA6 (Tamura et al. 2013). Codon positions included were 1st+2nd+3rd+Noncoding. In the distance analysis, all positions containing ‘N’s were removed for each sequenced pair. There were a total of 658 positions in the final dataset. To further evaluate the divergence within the genera Geophilus and Stenotaenia, the frequency distribution of the pairwise intra- and inter-specific distances were analysed.

Geophilus is not clearly separated from Stenotaenia in our analysis (Fig. 1). The basal-most node of the tree supports three monophyletic groups: G. flavus (de Geer, 1778), a species formerly separated in a different genus, Necrophloeophagus Newport, 1842, all other Geophilus, and Stenotaenia. However, the other Geophilus receive little statistical support (34%). The monophyly of the individual Geophilus species, as well as the Stenotaenia lineages L1–L3, all receive 100% bootstrap support (Fig. 1).

Figure 1. 

Maximum likelihood tree, 1000 bootstrap replicates. L1–L3 = S. linearis lineages 1–3; NRW = North Rhine-Westphalia; Baden-W = Baden-Württemberg. S. ‘sorrentina’ comes from GenBank and might refer to S. forficularis. For exact locality data, see Table 1.

All 13 specimens of G. flavus show little genetic distance (0–2.4%) to one another. Within the group containing the remaining Geophilus species, G. ribauti Brölemann, 1908, a species formerly treated as a member of the genus Brachygeophilus Brölemann, 1908, is in a basal position to a weakly supported clade (64% statistical support) including G. electricus (Linné, 1758), G. carpophagus Leach, 1814, and G. alpinus Meinert, 1870. In this clade, G. electricus (100% statistical support) is opposed to the sister-taxa G. carpophagus and G. alpinus (83% statistical support). Inside G. electricus, the one specimen from western Germany is opposed to the three from Saxony-Anhalt (Table 1 and Fig. 1). G. alpinus is the only analyzed Geophilus species with widely separated intraspecific groups (Fig. 1). A basal trichotomy (Fig. 1) divides the five analyzed specimens into three groups that can not be separated geographically.

Within Stenotaenia, a basal trichotomy separates the specimens into (1) S. linearis L1, (2) S. ‘sorrentina’, and (3) the weakly supported (56% bootstrap support) S. linearis L2 (including the topotypes) together with S. linearis L3 (Fig. 1). S. linearis L1 includes three specimens, one from Bonn, another from Euskirchen, both in western Germany and one from Leipzig in eastern Germany. S. linearis L2 contains a single specimen from Dachau, one close to Ulm, as well as two topotypes from Regensburg, all in southern Germany, while the majority (12) of analyzed German S. linearis specimens are recovered in S. linearis L3 (Fig. 1). The L3 group is divided into two clusters (L3a and b), one including seven specimens representing a single haplotype from seven different localities in western and eastern Germany, and the other one including five specimens also representing a single haplotype from four different localities (Esslingen, Hegnach, Badenweiler, and Stuttgart) in south-western Germany.

The distance analysis shows a first cluster of intraspecific distances ranging from 0–2.8%, with a G. electricus outlier at 4.9% (Fig. 2), a second cluster at 9.4–10.2%, and a third cluster, which overlaps with the interspecific distances, at 13.7–16.7%. Interspecific distances inside German Geophilus and Stenotaenia are high, varying from 16.3–22.0%. The highest observed genetic distance is between Stenotaenia and Geophilus species (16.6–22.7%), while the Geophilus species differ from one another by 17.2–21.7%.

Figure 2. 

Frequency distribution of pairwise intraspecific (blue) and interspecific (red) distances. Blue circle = intraspecific distances of G. alpinus and among S. linearis L3; Red circle = interspecific distances and distances between S. linearis lineages. Basic table see Suppl. material 1.

Clear intraspecific distances in German Geophilomorpha range from 0–5% (Fig. 2). A potential barcoding gap, however, is filled by the relatively high intraspecific distances (Fig. 2) of G. alpinus and Stenotaenia L3 (9.4–10.2% range). The genetic distances (13.7–16.7%) between the different Stenotaenia lineages (L1, L2 & L3) fall partly in the interspecific range of variation of the German Geophilomorpha (Fig. 2). The large interspecific distances (16.6–22.7%) observed among German Geophilomorpha are an indication that all species can be easily separated using the COI barcode marker. The distance analysis is partly biased towards interspecific distances because only a few specimens per species were analyzed. To explain the high nucleotide variability, excluding cryptic species, the presence of the maternally inherited endosymbionts (Hurst et al. 2005), as well as the origin of the lineages from different glacial refugia (Babik et al. 2005) followed by a subsequent fusion to a single species, need to be checked.

The three German Stenotaenia lineages are only weakly geographically separated (Fig. 3). Stenotaenia L1 is represented in our dataset with one specimen from Bonn, one from Euskirchen, and a third specimen from Leipzig (Fig. 1), the first two localities are separated from the third by more than 400 km apart (Fig. 3). This clade can be described as the Stenotaenia specimen from central Germany. All three specimens show the same haplotype.

Figure 3. 

Map of S. linearis samples studied during GBOL (large dots), as well as other S. linearis records from Edaphobase, the ZSM and ZFMK collection (small dots, status 10.2014). Yellow = S. linearis L1; Blue = S. linearis L2; Green = S. linearis L3. (A) S. linearis in the field, photo: J. Spelda, specimen from Stuttgart-Hofen, Zuckerberg.

Stenotaenia L2 represents topotypic material from Regensburg, a specimen from the Kiesental near Ulm, as well as a single specimen from Dachau in southern Germany. All three localities are more than 100 km apart but only the specimen from Dachau differs by 1.4%. S. L2 differs significantly (13.7–16.7%) from other German Stenotaenia. This clade might be characterised as of southeastern German origin along the Danube river system.

Both clades of lineage 3, one from western and eastern Germany (L3a), the other from SW Germany (L3b) show identical haplotypes, but differ from one another by 9.4% (Fig. 3). The intraspecific difference is similar to the differences observed in some Geophilus species (9.4–10.2% in G. alpinus), but significantly larger than the differences observed in the widespread G. flavus (0.2–2.4%), which often come from the exact same localities as the Stenotaenia specimens (Table 1).

Whether or not the apparent sympatric distribution of the three different lineages of Stenotaenia in Germany (Fig. 3) might have been influenced by human-induced introduction or dispersal is not known. Virtually all collection localities are close to human habitats, but differ strongly in their current direct exposure to human activities.

Such a large project faces a set of predictable technical problems, which can potentially cause wrong results.

Specimen collections: According to the main aim of the project (get approx. 10 specimens from at least five localities for each species to capture the estimated German-wide COI variation), the different collectors preferred localities where they could find many myriapod specimens easily – a potential collection bias. The amount of successfully sequenced S. linearis specimens as well as G. alpinus specimens and their different positions and deep splits within the maximum likelihood tree (Fig. 1) tell us that a larger amount of specimens from many more regions in Germany (Fig. 3) would be a desirable object for future taxonomic and/or biogeographical studies on these species. Bergsten et al. (2012) showed that up to 70 individuals are required to sample 95% of the intraspecific variation.

Specimen determination: As done by Bonato et al. (2014) for all Geophilomorpha species, a data matrix of additional morphological characters, presumably morphometric characters, should be created for the detection of usable characters for determining the possible cryptic S. linearis taxa. However, it is not feasible to have such morphological studies as part of a large barcoding project like GBOL.

Our analysis shows the importance of COI barcode data in the detection of taxonomic problems inside the centipede order Geophilomorpha. However, it also illustrates that barcode data alone does not clarify taxonomic problems. Only a thorough morphological study of the Stenotaenia species, including the types, plus the addition of nuclear markers, may be able to solve the complex picture of this genus.

As a result of the voucher-based barcoding effort, all analyzed specimens, and even their DNA extracts, are available for loan and should be incorporated into any future study of Stenotaenia.