No abstract

Aim We investigated whether cross‐cultural food plant selection in southern Africa is best explained by language ancestry, floristic environment or subsistence strategy. Location The flora of southern Africa region. Taxa All 1,740 edible plant taxa of southern Africa, representing 711 genera in 156 families. Methods Distribution data of plants were overlapped in ArcMap with 19 language maps, eight biomes and all taxa with nutritional data. Six correlations were estimated between five pairwise distance matrices (language ancestry, geographical proximity, floristic and edible environments and utilized species) with Mantel tests using the ‘vegan’ package in R. Regression analyses were used to identify floristic and cultural preferences in food plant selection. Results Spatial autocorrelation did not influence the selection of edible plants by the 19 language groups of southern Africa (r = −.078). The floristic and edible environments had a strong correlation (r = .9743) while the distance matrices of the edible and actually utilized plants had a low correlation for 13 of the language groups (r = .2174). Regression analyses between the floristic and edible environments for the FSA region and three languages, representing hunter‐gatherers (Ju│’hoan), pastoralists (Khoekhoe) and agrarians (Venda) were all significant (p < .001) with high R2 values (respectively .6181, .7702, .6654 and .7900), as were the relationship (p < .001) between what is edible and what was actually utilized. Surprisingly, the Apocynaceae had a much higher residual value than globally important food plant families. Vitamin C of fruits seems to have higher levels along the coastal regions, and carbohydrates in underground storage organs have higher levels in the summer‐arid western region. Main conclusions There is an apparent preference for certain food plant families in southern Africa. This selection appears to be driven by subsistence strategy, based on the categories of plants preferred by the three representative language groups.

No abstract

No abstract