Figure 4 . a) Percentage of polyploidy versus absolute latitude; b) Pruned regression tree results based on percentage of polyploidy, the total deviance explained is variance remaining in the sum of the leaves as a proportion of the variance in the entire data set. Cut off values for the most important environmental variable are given at each split: maximum temperature of the warmest month (TMX), total species richness (Stot) and annual precipitation (PCP). The numbers at each leaf are the average percentage of polyploidy of the grid square in that group, c) The partial dependency plots showing the shape of the relationship between percentage of polyploidy and its best predictor variables: Maximum temperature of the warmest month (TMX) and potential evapotranspiration (ET0).