Lasius precursor, Seifert, 2020

Seifert, Bernhard, 2020, A taxonomic revision of the Palaearctic members of the subgenus Lasius s. str. (Hymenoptera, Formicidae), Soil Organisms 92 (1), pp. 15-86 : 51-52

4.4.8 Lasius precursor sp. nov.

Etymology. From an evolutionary perspective, colony demography and behavior this species can be considered as a precursor of the situation in L. neglectus (Seifert 2010) . L. precursor sp. nov. probably represents a rather recent split-off from L. turcicus and may probably serve as a genetic model for transition from monogyny in L. turcicus to true supercoloniality.

Type material. Holotype worker plus 4 paratype workers and 3 alate gynes labelled ”TUR: 39.795° N, 26.681°E, Üzümlü- 1.1 km N, 167 m, rural grassland, leg. Cremer et al. 2004.06.08 -712“; 5 paratype workers plus 2 males labelled ”TUR: 39.795° N, 26.682°E, Üzümlü- 1.1 km N, 160 m, rural grassland, leg. Cremer et al. 2004.06.08 -710“; depository SMN Görlitz.

All material examined. A total of 51nest samples with 180 workers from Greece (1 sample) and 14 localities in Turkey (50 samples) were subject to NUMOBAT investigation. For details see supplementary information SI1.

Geographic range. So far only known from Anatolia between 26°E and 31°E and the Island of Kos (36.85°N, 27.08°E) close to coast of West Anatolia. The altitudinal distribution of 51 samples ranges from 5 to 1116 m with the median at 237 m and 86% of all findings below 400 m. This differs significantly from the situation in 77 sympatric L. turcicus samples with the median at 1007 m, 74% of all samples above 400 m and a range from 1 to 1170 m (one-tailed Mann-Whitney U-test p<0.0005). The situation in East Anatolia is not studied.

Diagnosis ( Tab. 2 , Figs. 15 –16; key):

Body size small (CS 769 µm). Number of mandibular dents low (MaDe 900 7.6). Clypeal pubescence dilute (sqPDCL 900 5.13). Pronotal setae rather short (PnHL/ CS 900 0.127), very little longer than gular setae (GuHL/ CS 900 0.125). Petiole scale in profile view rather thin with an acute dorsal tip. Pubescence hairs on frons rather long (PLF 900 34.5 µm). Dorsum of scape and hind tibiae without or few, occasional setae. It differs from L. neglectus by shorter scape (SL/CS 900 0.946), larger postocular distance (PoOc/CL 900 0.238), smaller torulo-clypeal distance (dClAn/CS 900 3.61%) and fewer setae on posterior margin of head (nOcc 900 5.8). The most significant differences to L. turcicus are smaller absolute size, shorter scape, larger postocular distance and smaller torulo-clypeal distance. Coloration: Head and gaster dark brown, mesosoma often suggested lighter; mandibles, antennae, lateral part of clypeus, tibiae and tarsae light yellowish-brown.

Biology. Cremer et al. (2008) found the following differences in the biologies of L. precursor sp. nov. and Lasius turcicus : Gynes of L. precursor sp. nov. show a trend to mate within colonies and reduced dispersal rates compared to L. turcicus which is in line with the smaller absolute size and shorter wings of the former. Gynes of L. turcicus are larger and longer-winged, never showed intranidal mating in 106 nest box trials, and are better dispersers. Combining L. turcicus gynes with males of L. precursor sp. nov. resulted in 38% matings in 29 tests. In contrast, L. turcicus males never mated L. precursor sp. nov. gynes in 51 tests, suggesting that the males of L. turcicus need to fly before being able to copulate. This suggests that part of the L. precursor sp. nov. populations are propagated by intra-nest mating and colony budding, but without being invasive. Intraspecific aggression among workers of different nests was high in L. turcicus , lower but still significant in L. precursor sp. nov. and absent in L. neglectus . Relatedness was moderately high (0.392 ± 0.070) in nests of L. precursor sp. nov., whereas it reached in L. turcicus levels consistent with territorial colonies headed mostly by a single queen (0.561 ± 0.034). L. precursor sp. nov. shows an intermediate social structure between the supercolonial invasive L. neglectus and the highly structured L. turcicus populations: in L. precursor , several ‘small-scale supercolonies’ coexist in most populations, but without coming close to the sizes of the extensive L. neglectus supercolonies. Long-chain hydrocarbons were relatively frequent in L. precursor sp. nov. and rare in L. turcicus . The less volatile long-chain hydrocarbons have been hypothesized to be less informative as recognition cues which is consistent with the lower aggression levels in L. precursor sp. nov. compared to L. turcicus .

Comments. Based on investigation of cuticular hydrocarbon patterns, nuDNA (microsatellite data), Cremer et al. (2008) unambiguously supported the separate species identity of L. neglectus from L. turcicus . This finding is in line with the results of NC-clustering reported above. Yet, Cremer et al. also reported that Lasius turcicus in the conception of Seifert (2000) can be divided into two entities which they called the highland form (= L. turcicus Santschi ) and the lowland form (= L. precursor sp. nov. introduced here). They found clear differences between the two entities in cuticular hydrocarbons and microsatellite data. The material used here in NC-clustering, altogether 127 samples, is identical for 122 samples with the material evaluated by Cremer et al. Five samples – the type series of L. turcicus plus 4 samples from the islands of Kos and Rhodos – were added here. Using the 10 standard characters CS, CL/CW 900, SL/ CS 900, nGen 900, nGu 900, sqPDCL 900, dClAn/CS 900, PoOc/ CL 900, EYE/CS 900) and MP6/CS 900, 51 nest samples of L. precursor are separated from of 77 nest samples of L. turcicus with the following error rates: 4.7% in NC-Ward, 4.7% in part.hclust (plus 4.7% outliers) 0.8% in NC-part. kmeans, and 0.8% in NC-NMDS-kmeans – in the mean of all four analyses 2.8%.This is below the 4%error threshold and justifies describing L. precursor sp. nov. as a separate species. Figure 111 shows the data of NC-Ward NC-part. hclust and NC-part.kmeans. If run as wild-cards in a linear discriminant analysis, the type sample of L. turcicus and the holotype sample of L. precursor sp. nov. are allocated to corresponding clusters with p=0.999 and 1.000. The classification error by the LDA was 7.6% in 396 worker individuals which is a normal value for cryptic species.

Figures

Figs. 15 and 16: Lasius precursor sp. nov.; holotype

Figs. 111: NC-clustering of 51 nest samples of Lasius precursor sp. nov.(black bars) and 77nest samples of L.turcicus (grey bars) from the East Aegean and Asia Minor. The hierarchical algorithms, NC-Ward (tree shown) and NC-part.hclust, showed a classification error of 4.7% whereas the non-hierarchical algorythms, NC-part.kmeans and NC- NMDS-k.means (not shown), misclassified only 0.8% of the samples giving an overall mean error of 2.8%. Outliers in NC-part.hclust are indicated by white bars (gaps).