Climate and soil type together explain the distribution of microendemic species in a biodiversity hotspot

Abstract

The grasshopper genus Caledonula, endemic to New Caledonia, was studied to understand the evolution of species distributions in relation to climate and soil types. Based on a comprehensive sampling of 80 locations throughout the island, the genus was represented by five species, four of which are new to science, of which three are described here. All the species have limited distributions in New Caledonia. Bioclimatic niche modelling shows that all the species were found in association with a wet climate and reduced seasonality, explaining their restriction to the southern half of the island. The results suggest that the genus was ancestrally constrained by seasonality. A molecular phylogeny was reconstructed using two mitochondrial and two nuclear markers. The partially resolved tree showed monophyly of the species found on metalliferous soils, and molecular dating indicated a rather recent origin for the genus. Adaptation to metalliferous soils is suggested by both morphological changes and radiation on these soils. The genus Caledonula is therefore a good model to understand the origin of microendemism in the context of recent and mixed influences of climate and soil type.

Figure 1. Caledonula species and habitat.

Caledonula fuscovittata (A-D): male (A), female (B); couple mating in Monts Koghis during day (C), view of habitat in Col d’Amieu (D); Caledonula amedegnatae (E-F): male (E), view of habitat in Forêt Nord (F) (photos by T. Robillard).

Figure 2. 50% majority-rule consensus tree for Caledonula obtained from Bayesian analysis of the combined data set (Cytb, CO2, EF1a and H3).

The Bayesian posterior probabilities are indicated below branches. The geographical distribution is given at the right of the taxon names. On the topology, pale blue lines represent the dating results obtained with the divergence rate of 4.22%/Ma, and violet lines represent the dating results obtained with the divergence rate of 1.1%/Ma. On the maps, the distribution of ultramafic rocks and corresponding metalliferous soils is indicated in grey, and the species distributions are indicated with the same colour as in the phylogeny.

Figure 3. Predicted distribution of the genus Caledonula (A) and of the species Caledonula fuscovittata (B) constructed from presence data using MAXENT.

Results are presented for logistic probabilities of occurrences ranging continuously from low to high. Warmer colours show areas with predicted better conditions. Dots show species’ sampling locations; triangles indicate all other sampled sites in New Caledonia, to indicate those where Caledonula was not found. For each map, a) shows the Regularized Training Gain under 100 Jackknife turns (green: without variable; blue: with only one variable; and red: with all variables); b) shows the response curves of BIO 15, the variable that contributes most to these models.

Figure 4. Caledonula fuscovittata, male (A) and female (B) habitus.

Scale bar: 5 mm. Drawings by Gilbert Hodebert (MNHN).

Figure 5. Caledonula species in dorsal and lateral view.

C. fuscovittata (A-D), C. amedegnatae n. sp. (E-H), C. humboldti n. sp. (I-L) and C. grandgousieri n. sp. (M-N). Left: ♂, right: ♀. Scale bar: 5 mm.

Figure 6. Caledonula species: face (upper part) and epiproct in dorsal view (lower part).

C. fuscovittata: (♂: A, C; ♀: B, D), C. amedegnatae n. sp. (♂: E, G; ♀: F, H), C. humboldti n. sp. (♂: I, K; ♀: J, L) and C. grandgousieri n. sp. (♂: M, N). Scale bars: 2 mm.

Figure 7. Caledonula species, male genitalia in dorsal view.

(an: ancorae; br: bridge; lo: lophi). Scale bars: 1 mm.