The total number of taxa, mostly identified to species, was 554 (Table 1, Appendix Tables 1-3), 96 of which were fungi. Taxonomically, the difference between two food plant groups - garden versus wild - was strongly pronounced even at family level (Fig. 2). Only one plant species was bought in markets; some plants were found both in wild and garden (189 mentions, less than 1%). The great majority of mentions (>99%) were either from families found either in gardens (62%) or wild (37%). Over 41% of all mentions refereed to the use of fruits, 21% to leaves, about 7% to seeds, and 5% to fruiting bodies, leaves/stems and stems. Whole plants were only used very infrequently. Of all families Rosaceae, Apiaceae, Lamiaceae, Amaryllidaceae and Solanaceae showed highest importance. At a generic level, Allium, Pyrus, Malus and Brassica received the highest number of use report. Only 30 species (6% of the total) represented 46% of all use mentions, but only Malus orientalis (3.5%), Pyrus communis (3.2%), and Vitis vinifera (2.7%) had over 2% of mentions, and Chenopodium album and Urtica dioica were the only not cultivated plants reaching over 1% of mentions. In most regions at all altitudinal ranges the aboveground parts were mist frequently used (Fig. 3),
Most plants (65%) were simply eaten, many either raw (55%), or fried (e.g. 8% that were fungi). A full 5% of all mentioned plant-uses were for pickles (often stems), and a full 18% of all use reports were for Phkhali (boiled herb pie, especially in spring), and 4% were used as spices, and around 2% for the distillation of alcohol. All other use categories (35) had much fewer mentions.
The richness of plant families was 66 versus 97 families in garden and wild plants, respectively, and the difference was highly significant (Fig. 4). Other diversity indices also unequivocally pointed to a considerably more diverse family composition of wild versus garden plants as the differences between all the tested diversity indices appeared to be highly significant (Table 2).
The regions of Georgia can be divided into three groups by the similarity of garden food plants as can be seen on the nMDS ordination graph (Fig. 5). This ordination seems to be influenced on the presence of large markets: Adjara, Samegrelo, Guria, and Kakheti which are lowland regions with large cities are joined by minimum distance versus Tori, Zemo Svaneti, Khevsureti, Tusheti and Javakheti, which are the most remote places. Kvemo Svaneti, Lechkhumi, Meskheti, Kvemo Kartli, Zemo Imereti, Zemo and Kvemo Racha, Mtianeti are moderately remote from large markets. The grouping of the regions closer to large markets might however have another distinct reason: Adjara, Samegrelo, Guria, and Kakheti are also the climatically warmest regions in Georgia, with the longest growing seasons. This allows the production of food plants almost all year round, and greatly reduces the dependency on foraging wild species.
For comparison we assessedthe usage of plants between regions based on their family, genus, specific parts used (root, shoot, or both) used, reproductive stages used (vegetative, reproductive, or both) and their specific parts used (bark, branches, buds, bulb, cones, flowers, fruit, latex, leaves, resin, roots, seeds, shoots, silk, stem, timber, tuber, whole plant), but at regional level and within different altitudinal ranges through Non-Multidimensional Scaling (NMDS) followed by Permutational Multivariate Analysis of Variance (PERMANOVA) with 999 permutations and Euclidian distance. The detailed results are depicted in Table 3 and Appendix Tables 4-9.
At the same time, the regions vary strongly in their species richness (Fig. 6). These differences also might reflect the remoteness from large markets and severity of local climate. At this stage we however still have limited data to exactly explain the variation in garden species across the regions, this question requires further study.
Relationships among the regions in the case of wild food plants show a different and clearer pattern (Fig. 7). In particular, adjacent regions cluster together (Kvemo Zemo Racha, and Zemo Imereti; Samegrelo, Guria, Adjara, Lechkhumi and Kvemo and Zemo Svaneti; Meskheti, Javakheti, Kvemo Kartli; Mtianeti, Kakheti, Khevsureti, Tusheti). Like in the case of the garden food plants, species diversity of the wild food plants mentioned varied strongly (Fig. 8). Climate severity and traditions of the use of wild food plants might play role in this variation.
Pkhali and Pickles - emblematic foods of the Caucasus
Of all food preparations the use of plants as ingredient of boiled herb preparations (mostly as გაზაფხულის ფხალი - gazapkhuli pkhali = Spring Pkhkali, as the first vitamin source after winter), and as lacto-fermented or vinegar-based pickles are probably the most emblematic ones in the Caucasus, given that almost 50% of all food mentions were for phkhali, and almost 12% for pickled plants, and 8% for teas.
While the overall distribution of families, genera and their uses were similar between regions, although overall most species were used in Guria. However, the knowledge distribution was most uneven for these food categories (Fig. 9). The altitudinal range between 1001-1500m, followed by 1501-2000m were clearly predominant when it came to diversity of plants used as well as uses (Fig. 10). This very unequal distribution of the most important families/genera, as well as their respective uses is reflected in Fig. 11.
Only 60% of participants reported making pickles. Of these, over 16% each came from Zemo Imereti and Khevsureti, and 12% each from Zvemo Svaneti, the Javakheti-Plateau, and Guria. The first regions represent all high altitude - short growing season areas, where the population does need to preserve food for winter. Guria is relatively warm - but very wet and snow-rich, which also might explain the prevalence of pickles. No participants whatsoever from Adjara, Samegrelo (the most subtropical regions) and Mtianeti (close to the capital Tbilisi) reported making pickles. Unsurprising, Kakhetians were also not enthusiastic about this form of preparation, because Kahketi is also a region famous for its large agricultural production. In contrast, in Tori and Tusheti there are simply less products that actually can be pickled. Preferred species (of a total of 79) for pickles were mostly Amaranthaceae (Amaranthus, Chenopdium), Apiaceae (especially the stems of Anthriscus, Chaerophyllum and Heracleum were pickled, but also, astonishingly, stems of Conium maculatum), Amaryllidaceae (all Allium species), And Polygonaceae (Polygonum and Rumex). In addition, Aruncus vulgaris (Rosceae), Stapyllea colchica (Staphyleaceae). All of these were more important as pickles than "traditional European style species (Cucumis sativus, Capsicum etc.). The fermentation of the ferns Mattheucia struthiopteris (Onocleaceae) and Dryopteris filix-mas (Dryopteridaceae) was similar to what we observed e.g. in the Himalayas.
In case of Pkhali, over 93% of all participants - from all regions - reported to use such boiled herbs, normally in Spring. This was surprising, as we had expected much more limited use in the climatically favorable regions. Nevertheless, Zemo Imereti (19% of all Phkhali preparations), Tori and Kvemo Racha (16% each), Tusheti (15%) and Khevsureti (14%) - all mountain regions with long winters, stood out as the real "herb eater" areas. In contrast to the pickled species, essentially only young leaves were used for pkhali, with great emphasis on the same families indicated in pickles. (all pickled plant species were also used for phkhali). The overall number of species fused or pkhali was however much higher (197). The elaboration of phkhali often involves many steps in order to reduce the toxicity of species used, and in most cases a wide variety of herbs are included in each preparation. Interesting examples for the use of toxic species included the leaves of Solanum tuberosum, Veratrum lobelianum and Viola sp.