Herbaria (and other collections) are vast store houses of information on the world’s biological diversity. At present several local initiatives have started to transfer the information of the collections into databases. This is a most important task as for larges-scale analyses the data is otherwise unavailable.
A major challenge, however, is the proper use of the specimen data for basic and applied science. As botanical specimens are generally collected for the sole purpose of taxonomic and floristic but not biodiversity studies, they present some serious drawbacks when they are used for that purpose. The major drawback is the non-random nature of the data, due to collectors’ bias for, for example, specific areas and specific taxa. This has led to the support of evolutionary processes such as speciation through refuges in the Amazon that were shown to be patterns merely based on artefacts of sampling intensity. However, many researchers have presented analysis based on collections that did not correct well for this bias. Also, collected specimens do not reflect the community structure well, as species are not collected in relation to their abundance – viz. common and rare species are collected generally in equal numbers. Therefore, diversity expressed as Fisher’s alpha or the Shannon index based on botanical collections tends to be an overestimation. Yet, herbarium collections are such vast stores of information, representing work of over three centuries in often extremely remote areas (where, not rarely, it is the only data available), that it is very attractive to find ways to extract useful information for the study of biodiversity patterns from them.