Table 1.
Distribution of the five haplotypes (chloroplast) in the sampled populations.
Fig 1.
Point of known presence of O. pes-caprae in the Euro-Mediterranean area until 1800.
Only cultivated plants are known.
Fig 2.
Point of known presence of O. pes-caprae in the Euro-Mediterranean area until 1840.
Indications of the plant in the wild are known in south mediterranean.
Fig 3.
O. pes-caprae in the Euro-Mediterranean area until 1880.
More stations in the wild are known, but always limitedly to south mediterranean.
Fig 4.
O. pes-caprae in the Euro-Mediterranean area until 1950.
Some stations are indicated in a more northern position. Black stars: records of cultivated specimens. Blue circles: records of wild plants.
Fig 5.
Arrival and spread of Oxalis pes-caprae L.
General diachronic map of the presence of O. pes-caprae in the Euro-Mediterranean area (black stars: records of cultivated plants). The legends indicate the correspondence of the different colors of the circles on the map (the observation) with the first year of the record for a given population.
Fig 6.
Kmeans and geographic profiling analysis results on the data of 1840.
Two clusters were the most probable solution after the Silhouette criterion. The arrows show the most probable spread origin for the first cluster (left arrow) and for the second one (central arrows), this last compatible with the hypothesized origin from Malta. In red the areas of highest probability of the spread origin.
Fig 7.
Kmeans and geographic profiling analysis results on the data of 1880.
Three clusters were the most probable solution after the Silhouette criterion. The arrows show the most probable spread origin for the first cluster (left arrow), for the second cluster (central arrows), and for the third cluster (western coast of Turkey). In red the areas of highest probability of the spread origin. The numbers represent the observation number in the CSV file containing the geographical data. The Voronoi tessellation on the basis of the three clusters is leared with different colors.
Fig 8.
Kmeans and geographic profiling analysis results on the data until 2010.
Three clusters were the most probable solution after the Silhouette criterion. The arrows show the most probable spread origin for the clusters, confirming the hypothesis obtained with data until 1880. In red the areas of highest probability of the spread origin. Legend as in Fig 5.