The database prototype has been implemented as a local relational database. Different implementation solutions were tested, according to the requirements and use cases. Present and ancient ethnobotanical data were organized in the same table structure, allowing some costs of redundancy in favor of the homogenization of data and queries. Because flexibility was a major issue for our archive, object oriented concepts were implemented, such as unlimited hierarchical structures (self-joined tables), particularly for categorized look-up tables and species taxonomy, and inheritance concepts, allowing e.g. information sources to be a reference, a person, an herbal or an archeological remain. Those techniques allowed us to develop a minimal data structure fitting different kind of information, and to perform queries at different levels of detail.

The structure of the developed repository is summarized in Figure 2 and in Table 1.

The archive currently allows the storage and management of following information:

• ethnobotanical information from interviews (informants), historical herbarium specimens and documents (recent, historic and ancient sources), according to a hierarchical catalogue of plant uses and preparations (medicinal, alimentary, ceremonial, veterinary, domestic). It is also possible to include ethnobotanical information from archeological remains, although the relationships to the relative databases have not yet been implemented.

• plant names information: plant popular names (local or ancient), and their relationship to species taxonomy 11. Information on reliability of species identification (also multiple options are possible, according to hypothesis of different authors);

• information on specimens in historical herbals.

Ethnobotanical data were organized in a relational structure: a relevée can group information on multiple plants, which can be used for different scopes, each being reached with one or a sequence of preparation steps. Traditional knowledge cannot be easily categorized and therapeutic uses are often intermingled with ceremonial and symbolic ones. Since attempts to categorize these kinds of uses are not always meaningful, the storage of a description that reports in detail a specific application of a plant is recommended. However, to allow a more effective search within the database, beliefs and practices dealing with folk uses of a species were grouped in use categories, ranging over medicinal (including affection and mode of administration of the plant), ceremonial (symbolic, ritual, religious including any belief concerning the plant), alimentary, domestic (from the control of undesired animal in the house, to house construction, or wooden utensils), veterinary uses. These categories have been organized in a hierarchical structure and may be assigned according to the detail level of the information gathered. A reliability degree of those use assignments is given, particularly for data from the ancient treatises and from archeological remains. The principle of storing information with different levels of accuracy has been implemented also for the preparation steps (mechanical breaking, cooking, drying, ...) and for the plant parts involved (above ground part, bark, leaf, young leaf, below ground part, root, bulb, tuber, ...).

This level of semantic detail allows fine tuned query possibilities and is needed if a future improvement to a multilingual application is planned.

The database is intended to store and manage data from pre-Linnean times to present. At the moment the oldest entries stored in the database, dealing with plant uses in Mediterranean region during Antiquity, come from studies carried out in the last decade on Corpus Hippocraticum 12 treatises, Dioscorides "De Materia Medica" 131415, Pliny's Naturalis Historia 16, and Pseudo-Apuleius Herbarium 17. Since ancient texts provides no iconography and limited description of plants; the identity of these species is always a matter of controversial 18. For example, according to Buenz 19, the identification of the plants quoted in the Corpus Hippocraticum is a largely overcome challenge, whereas Raven 20 is clearly skeptical about the possibility of resolving the identity of any Greek phytonym down to the species level, in accordance with Linnaean nomenclature. In order to explicitly implement in our database this uncertainty, a join table with the putative identifications of ancient phytonyms with one or more Mediterranean species has been implemented with the proposed identifications according to Aliotta et al. 12 and Stirling 21. Thus this allows a critical evaluation of ancient traditional knowledge about the uses of a plant. We used this specific structure to enable us also to manage synonym relationships of scientific names or identical phythonyms referring to different species, according to regions or authors. In this respect, our phythonym table contains popular, ancient, Italian and scientific names, allowing a maximum flexibility when entering new ethnobotanical records.

Since a consistent number of our entries come from plant labels in historical herbals, also detailed information on specimen and herbal collections was stored in the database.

Information sources can be of multiple types, namely bibliographic references, collectors, informants, historical herbals or archeological remains. All or subsets of these types can be referenced, according to the tables where a source information is needed.

As already mentioned, reliability of the information stored in the database is handled at multiple levels, namely: phytonym identification (Plant table), which in the case of informants or collectors is useful, tentative identifications (Taxon assignment table), used mainly for data from the ancient treatises, and use assignments (Plant use table), for data from ancient sources or archeological remains.

Figures 3 and 4 show two screenshots of the main data management mask, respectively with recent and ancient ethnobotanical data. At this development stage, neither administration of user privileges on data subsets nor multilingual support have been implemented.

Different query features allow the user to get a temporal drawing of the importance of a species within the life of a community. A query for a species or a genus returns all the available information, either from published works, unpublished interviews, labels of historical herbals or archeological remains. The hierarchical features implemented for uses, plant parts, preparations and sites (from national to municipality level) allows both generic or fine-grained queries for those parameters, according to the requirements of the user. Particular query settings can be specified, in order to extend the queries to rougher levels of detail, e.g. querying for a genus instead of a species, thus including all the species of that genus, or querying for neurology instead of headache, thus including all neurological affections. Such latter use comparisons can only be performed in the frame of the specified hierarchical structure, which implicitly defines the degree of similarity of the uses. In the future, different hierarchical classifications of the stored uses could be defined and added to the use table, thus enhancing comparative capabilities of the archive.

The performed queries return the appropriate reliability information. This is particularly important for the case of multiple putative identifications of ancient phytonyms, where a query for a given use produces multiple report lines for the same ethnobotanical information, according to the taxon assignments. The tuning of reliability settings allows restricting the query from all the record combinations (for a case by case analysis) to only records with a certain reliability degree or to only record combinations with the highest reliability.

Also historical summary information can be retrieved (analyzing the highest reliabilities), matching a plant binomial with a specific use, and showing for example how many times a given medicinal use of a species is reported in the database for a defined period.

The repository has been implemented with PostgreSQL8.3 22, while entry masks have been developed with Microsoft^® Access 2000 23.

At present time only data on flowering plants have been stored in the database, but it is possible to extend the archive also to non-vascular plants, algae and fungi. The interval of time was chosen because plant nomenclature from XVIII century onward is based on Linnean classification, which allows a more reliable approach to species identification. At the moment, all the data from modern (1950 – present days) ethnobotanical studies carried out in Campania have been stored in the database, together with the available information from different historical sources dating back from the first half of 1900's to XIX century. The majority of the records were extracted from hundreds of published works, both popular and scientific, which belong to three main types of sources:

I) floristic description of a territory, which often contained ethnobotanical information (i.e. Gussone 24; Pasquale 25; Pasquale and Pasquale 26).

II) the labels of the historical herbaria samples of Herbarium Porticense (PORUN), that in same case reported some indications about the use of a plant in the site where it was collected (collection Vincenzo and Francesco Briganti of Herbarium Porticense – PORUN)

III) medical treatises of physicians of past centuries (i.e. Petagna 27; Limoncelli 28; Pasquale and Avellino 29), which contains also the description of folk remedies, thus representing an important source of ethnopharmacological information.

The database currently includes 474 different species, which have been used in Campania between the end of XVIII century and the 2007.

1672 different uses have been recorded for the species listed in the database, ranging from medicinal, to alimentary, ceremonial, domestic and veterinary. Table 2 represents the distribution of the number of data (including number of sources of information and number of species) between XIX and XX Century. It is not surprising that the majority of data are related to recent studies, since previously folk knowledge was often largely disregarded. Caution should be adopted to the approach of such heterogeneous material. The studies dealing with the traditional plant knowledge in Campania which have been conducted approximately from 1950s onward, have been carried out according to the procedures adopted worldwide for ethnobotanical research. On the other hand, historical data stored in the archive allow to elicit information from sources which hardly represent a random sample of a statistic population. They can be derived from a single observation or from memories of very few individuals, and are only rarely a detailed recollection of uses coming from a population. The structure of information on plant uses from past centuries precludes any quantitative approach in comparison between past and present traditional knowledge of populations living in Campania. Notwithstanding this limitation, the historical ethnobotanical archive represents a useful tool to understand cultural transformations occurring on plant uses over the generations, particularly shedding light on continuity and changes in folk phytotherapy.

From the analysis of the information stored in the database, it is possible to make some generalizations on the evolution of ethnobotanical knowledge in Campania. Several trends remain unchanged over the last two centuries, such as the prevalence of data about medicinal plants, included those used in the so-called food medicine, which was often mentioned in XIX century sources. The occurrence of a relatively high number of medicinal citations is also due to the interests of the authors of that time, generally physicians or chemists.

The data were analyzed for both content and geographical distribution. In Figure 5a is shown the distribution of ethnobotanical information in the XIX century in Campania, and in Figure 5b that related to the XX century. Figures 5a and 5b reveal that some Campanian areas, which are ecologically homogeneous and share a common historic context, have been studied in some detail, whereas for other territories information is very scanty. In the last two centuries, a remarkable convergence of ethnobotanical studies has been carried out on Sannio, a territory which in part correspond to Samnium of Pre-Roman and Roman Age 10. Moreover, from the comparison between figure 5a and 5b a difference clearly emerges: some studies made in the first mid of the XX century report on ethnobotanical data attributed to large macro-areas of Campania, without any further detail. In figures 5b the areas interested by this kind of studies are marked with a pink colour. Although this reports can reveal meaningful information on the value of natural resources to local people, it is not possible to relate this information to a specific community. In our point of view these data should be usefully employed only as starting points for more detailed studies.

Table 3 lists the plants already employed in Campania at the beginning of XIX Century, which are still in use. Some species have been used for multiple purposes, including those related to food, medicine, ceremonies, and magic rituals. According to Turner 30, "the more widely or intensively a plant is used, the greater its cultural significance". If we take this assumption as a general rule, it is possible to derive from Table 3 a rank of significance for each plant included, although in a preliminary and qualitative way. In this respect, it is not surprising that the cultural importance of some species as Olea europea L. and Vitis vinifera L., followed by Brassica oleracea L. ranked highest in usage, and consequently in cultural significance, in the Campanian territory. The domestication and cultivation of these species was of pivotal importance in the Mediterranean civilization since Neolithic times, and represent still now an important economic resource for the Campania and the whole Mediterranean Region. The treatment of a range of internal affections with the so called food medicine often relies upon cultivated plants; in general, the collection of wild greens, although still practiced, has been considerably reduced in present times. This is probably due to the continuous decrease of people able to recognize plants in the wild, which, in turn, has been determined by the decline of agriculture in the region. However, cultivated plants correspond to about 30% of the species recorded in Table 3, the remaining being wild species, and it is possible to infer some degree of significance also for these species. For a detailed comparison, we selected three wild gathered species, Adiantum capillus-veneris L., Parietaria judaica L., and Artemisia absinthium L. frequently used in Campania (Table 4).

The Campanian historical database encompasses also information on plant uses in ancient Mediterranen cultures, particularly of Greek and Roman phytotherapy. This inclusion needs some methodological considerations: in a straightforward assay, the anthropologist Di Nola 42 categorized different kinds of ethnomedicine. The first one fits the Hoppal 43 definition: "ethnomedicine is the term for the practice of folk healing in the recent ethnographic literature". This definition embraces not only the study of cultures lacking written traditions, but also the body of traditional knowledge of European rural communities. According to Di Nola 42 also the corpus of knowledge belonging to the Ancient Classic medicine can be considered as a sort of ethnomedicine. It consists of treatises where popular phyto-remedies are described, although often tightly intermingled with a scholar approach to the therapy of some diseases. It cannot be denied that a thread linking Ancient and contemporary ethnobotany in Mediterranean Countries can be ideally traced: Stannard 44 stressed that not few curative practices elaborated in early times of Western civilization had remained alive. On the other hand, the caution evoked for the evaluation of ethnobotanical data coming from the last two centuries must be multiplied when one aims to analyze different uses of plants over a so large temporal period.

Paul 45 suggested that popular therapeutics is prevalently made by old medical prescriptions, which were subsequently abandoned by official Medicine. To shed light on this point, the medicinal uses of A. capillus-veneris, A. absinthium and P. judaica have been reported in table 5, according to some of the most important medical treatises of Antiquity and to a medical text of XVIII Century written by V. Petagna 27. In this way, it is possible to trace the possible influence on present popular medicine of historical recipes from different periods. As can be seen, during Antiquity A. capillus-veneris was administred mainly against pulmonar and kidney affections, and to treat gynecological diseases, A. absinthium was highly reputed as a tonic and digestive, and as a powerful antiworming agent and P. judaica had application as a diuretic, in the therapy of skin affections and to help the resolution of traumatic accidents. Similar therapeutic uses are also descripted by Petagna 27. A comparison with the present ethnopharmacological knowledge in Campania indicate that only some of the oldest applications are still in use, and that, as expected, there are more similarities between the indications reported by Petagna 27 and the present uses of the selected plants. However, no general rule can be traced; the story of uses of each plant follows hardly identifiable trajectories, which need a detailed study of all available sources. A point break is the transition from Antiquity to the Christian Era, when new symbolic meanings were introduced for many plants, renforcing old therapeutical applications or also relating them to unexplored medical applications, as in the case of Rosmarinus officinalis in Latin America 6.

In this respect, the comparison of the TK of a restricted territory as that of Campania with the corpus of Ancient phytotherapy can not be exhaustive and needs a more detailed scrutiny carried out on a larger set of historical data from other territories of the Mediterranean region.