Crematogaster difformis-subgroup, F. Smith, 1857

Hosoishi, Shingo, Maruyama, Munetoshi, Yamane, Seiki, Jaitrong, Weeyawat, Hashim, Rosli, Syaukani, Syaukani, Sokh, Heng, Itioka, Takao, Meleng, Paulus & Pham, Thai Hong, 2023, Multilocus phylogeny and historical biogeography of the Crematogaster inflata-group (Hymenoptera: Formicidae) in South-East Asia, Zoological Journal of the Linnean Society 198 (3), pp. 901-922 : 917

The Crematogaster difformis-subgroup

The relationships of the sister-species in the C. difformis -subgroup were variable with respect to their geographical distribution patterns. Crematogaster ampullaris and C. seaeardi display an allopatric distribution, whereas C. difformis and C. tanakai display a sympatric distribution. The closely related C. ampullaris and C. seaeardi illustrate an interesting distribution pattern with C. ampullaris in Sulawesi and C. seaeardi widespread elsewhere in South-East Asia. The estimated dating suggests that C. ampullaris and C. seaeardi diverged approximately 4 Mya ( Fig. 4 ; Table 2 ). The divergence implies a recent split of these populations and landmass distributions in that geological time ( Hall, 2002) suggesting that the ancestor of both species dispersed from the Sundaic region to Sulawesi west-to-east across Wallace’s Line. Sulawesi is considered to have taken its present form due to the collision of different islands c. 15–5 Mya ( Lohman et al., 2011). As the last possible connections between Borneo and Sulawesi were during the Pliocene or the Miocene ( Morales & Melnick, 1998; Mercer & Roth, 2003), the colonization of Sulawesi by the common ancestor may have occurred much later after the island collision event. The ancestor of C. seaeardi and C. ampullaris may have expanded its range to western Sulawesi (eastern Sunda Shelf) during low sea-level periods and evolved into C. ampullaris . As the Makassar Strait is a deep trench, Borneo and Sulawesi were not connected when sea levels dropped. Their diversification resulted from ‘an inter-island dispersal combined with climatic vicariance event’ ( Condamine et al., 2013).

Crematogaster difformis and C. tanakai are sympatric in Borneo, and their habitats are spatially close to each other. Crematogaster difformis has an obligate mutualistic relationship with Lecanopteris and Platycerium ferns in canopy trees of lowland rainforests, and C. tanakai inhabits the same nests with C. difformis ( Tanaka et al., 2009, 2012; Tanaka & Itioka, 2011). This restricted distribution demonstrates distinct endemism and is, therefore, of conservation significance. Lecanopteris fern species are found in the canopies of lowland tropical rainforests. The preservation of such primary forests will lead to the conservation of the endemic species. Molecular dating indicates that C. difformis and C. tanakai diverged approximately 1 Mya ( Fig. 4 ; Table 2 ). The low divergence implies a recent speciation.

As in C. aurita , the dispersal event and competition can explain similarly the position of C. mucronata , which is endemic to Sumatra ( Fig. 4 ). Klaus et al. (2013) suggested early divergence of freshwater crabs Parathelphusa and subsequent appearance on Sumatra, since much of Sumatra was still submerged during the Late Miocene. In contrast, the ancestor of C. mucronata already appeared on fragmented Sumatra during the Late Miocene. Freshwater crabs require more land area for their habitats, but ants presumably do not require large areas as Yamane (2013) listed 126 ant species having colonized Krakatau Island after eruption. Faunal and genetic similarities between the Malay Peninsula and Sumatran populations have been documented for several taxa ( Gorog et al., 2004; Lohman et al., 2011; Leonard et al., 2015), as the two areas were connected at sea level 20–30 m below the present level ( Voris, 2000).

Figures

Figure 4. A, historical biogeography of the Crematogaster inflata-group based on BEAST2 and BioGeoBEARS analyses using the dispersal multiplier [Scheme (ii)] and two time slices. Blue horizontal bars depict the 95% highest posterior probability (HPD). Node labels N1–N11 correspond with denotations in Table 2. Biogeographical analysis employed a DEC model, with eight regions. The left-bottom map represents the Indo-Australian Archipelago delimited into eight areas. Colours of squares correspond to the coloured area on the map. Coloured squares indicate the most likely ancestral area recovered at each node. The present distribution of each species is given by coloured squares. B–D, palaeogeographic maps were redrawn and modified from Hall (2013): B, 15 Mya. C, 10 Mya. D, 5 Mya.