Results

Morphospecies and sequencing success

Fifty-six morphospecies could be determined from the Phu Phan mountain area. Twenty-seven of the morphospecies (48%) were supposedly undescribed species or could not yet be assigned to an existing species (the fauna of mainland Asia has been fully revised in terms of type specimen revision (Ahrens, unpublished data), however, several species from Indonesia are known to widely occur in the Oriental region. Some species might thus still be assigned to already described species, when taxonomic revisions are finished for all parts of Asia).
186 specimens were sequenced successfully. The length of the alignedcox1 sequences was 658 base pairs (bp). For Maladera sp 16 and Neoserica phuphanensis more than five individuals per morphospecies have been sequenced, since a few specimens were initially mistakenly assigned to other morphospecies. Of the 56 morphospecies 14 were singletons, i.e., only represented by one specimen per species.
Due to shared haplotypes in different morphospecies, lowest inter- and infrasepcific distances were both zero (Table 2), while maximum infraspecific distances were around 7%. The infraspecific mean distance was 1.5%, and the median even lower (0.8%). Nine morphospecies (i.e., 16% of the taxa) had infraspecific distances larger than 3%.

Species delimitation

All morphospecies included in the analysis resulted monophyletic with three exceptions (Fig. 2): 1) Microserica sp 11 andMicroserica sp 13 are nested within the clade ofMicroserica varians ; 2) one of the five specimens ofNeoserica sp 29 was within the clade of Neoserica martinui ; and 3) Maladera sp 27 is placed within the clade of the morphologically very similar Maladera sp 9.
Three morphospecies shared identical haplotypes (Maladera sp 3a, sp 3b; sp 4; Figs 2, 3). Branch support values (ultrafast bootstrapping) of morphospecies clades are high with values of 0.8 to 1. DNA-based species delimitation applying PTP, TCS, and ABGD resulted in different clusters. Thirty-one morphospecies showed congruent results for all DNA-based delimitation (Fig. 2). For 46 morphospecies the results of at least one method matched with the morphospecies assignment. All methods showed splitting and also lumping of morphospecies.
bPTP and mPTP subdivided the specimens into 70 and 65 MOTUs (Table 1), with 37 (bPTP) and 38 (mPTP) matches between the morphospecies and MOTUs. Deviations are caused by erroneously inferred splitting events (i.e., individuals of one morphospecies were separated into two or more different MOTUs). Match ratios of both PTP variants were relatively low: 0.59 and 0.63, for bPTP and mPTP, respectively. TCS resulted in 69 MOTUs and had same number of matches (37) as bPTP. The match ratio (0.60) was higher than bPTP, but lower than mPTP.
ABGD yielded 51 MOTUs and showed the highest match ratio of all delimitation methods (0.77). It was the species delimitation method that showed most lumping of different morphospecies (Fig. 2). Examples for lumping are: one MOTU for Maladera exima plus M.parexima ; Maladera sp 9 plus Ma. sp 27;Neoserica sp 37, N. martinui plus N. sp 29; as well as Microserica varians , Mi. sp 11 plus Mi. sp 13.
Distance based clustering at the 3% level yielded similar results to the previous methods. It found 62 MOTUs and matched with 40 morphospecies; the match ratio (0.68) was the second highest, after ABGD. Barcode Index Number (BIN) assignments revealed 65 MOTUs and matched as well with 40 morphospecies, however, its match ratio was lower (0.66) than that of 3% clustering.
In 21% of the morphospecies (n= 12) we found relatively deep coalescence (i.e., distinct infraspecific phylogenetic structure) (e.g.,Ma. sp 4, sp 6, sp 16, Ma. fuscipes , Ma. futschauana , Ma. obscurata , Ma. peregoi , N. sp 26,Mi. panzona , Mi. varians , G. marginalis , G. carolusi ). In all others, infraspecific branches were rather shallow. Taxa sampled with more than three specimens and that were represented by a single haplotype did not occur. For all those cases with deep coalescence, at least one of the DNA-based species delimitations split the morphospecies, which in turn decreased the match ratio.