Table 1.

(a) Best (within a ΔAIC of 4) candidate models exploring language richness as a function of mean elevation, distance to the coast, latitude and mammal species richness at the 50 km spatial resolution, using a general linear model approach combined with a negative binomial distribution. (b) Best (within a ΔAIC of 4) candidate models exploring threatened language richness as a function of mean elevation, distance to the coast, latitude and threatened mammal species richness at the 50 km spatial resolution, using a general linear model approach combined with a negative binomial distribution.

model	AIC
(a)
language richness ∼ mammal richness × distance + elevation + latitude	1846.98
language richness ∼ mammal richness × distance + latitude	1848.05
(b)
threatened language richness ∼ distance × latitude + elevation	1055.84
threatened language richness ∼ elevation × latitude + distance	1056.98
threatened language richness ∼ elevation + distance + latitude	1057.46
threatened language richness ∼ distance × latitude + threatened mammal richness + elevation	1057.61
threatened language richness ∼ elevation × distance + latitude	1058.76
threatened language richness ∼ elevation × latitude + threatened mammal richness + distance	1058.95
threatened language richness ∼ threatened mammal richness + elevation + distance + latitude	1059.28
threatened language richness ∼ threatened mammal richness × latitude + elevation + distance	1059.62