How many type specimens can be stored in old lesser-known herbaria with turbulent histories? – A Juncus case study reveals their importance in taxonomy and biodiversity research

Abstract Many herbarium sets in Europe are still being catalogued and it is likely that many old-type collections are yet to be discovered. This research has the potential to facilitate the study of the biodiversity of many regions, especially regions for which collections are extremely scarce. This has been confirmed by a case study using Juncus (Juncaceae) examining the turbulent history of botanical collections at the WRSL herbarium and the evaluation of its importance to the study of taxonomy and biodiversity since 1821. The analysis revealed that the WRSL collection is rich in types (ca. 3.6%) and we identified 76 (of 78) new, historically and nomenclaturally important specimens (types, original material and so-called “topotypes”). Some of these type specimens represent duplicates of these that were stored in Berlin and destroyed during World War II. Many of the type specimens are from the United States of America, South Africa, India, and Canada. The largest number of Juncus type specimens stored at WRSL originate from South Africa (42.3% of all type specimens), even though Juncus is rare in Africa. Our study highlights that uncatalogued old collections that are under-explored and under-exploited have the potential to facilitate the discovery of specimens important for the study of biodiversity, conservation, taxonomy and nomenclature.


Introduction
The Natural History Museum of Wrocław University (Muzeum Przyrodnicze Uniwersytetu Wrocławskiego) is the oldest natural history museum in Poland and its history dates back to 1814, when it was founded by Prof. Johann Ludwig Christian Gravenhorst as the Zoological Museum. Currently, it houses both the botanical and zoological collections. The beginning of the herbarium in its present form was the Herbarium Horti Botanici Universitatis Wratislaviensis, which was established by Prof. Ludolph Christian Treviranus in 1821 (Wiktor 2002;Wanat 2013). The Herbarium Silesiacum was independently founded by the Silesian Association of Native Culture (Schlesische Gesellschaft für Vaterländische Cultur) and, until 1945, it was housed on Tamka Island, Wrocław. It developed independently from the other botanical collections, but following the Second World War, it was merged with the main part of the herbarium.
Professor H.R. Goeppert expanded the botany collections and established the Botanical Museum (Botanisches Museum) in 1853 (Wanat 2013). The first known catalogue of the Museum (Goeppert 1884) included 26 different collections, including the Herbarium of the World, the Herbarium Silesiacum, the Herbarium Mycologicum, a wood collection and several fruit and seed sets. Goeppert also opened another museum in 1878 -The Museum of the Botanical Garden (Mularczyk 1998). In 1888, all these several botanical collections belonging to the University were moved to a building that is today located at 6/8 Kanonia Street. However, they still consisted of two separate collections (the Botanical Museum in charge of Prof. Cohn and collections of the Herbarium and the Museum of the Botanical Garden in charge of Prof. Engler). Due to the efforts of Prof. Engler, a private Silesian plant collection assembled by Rudolf von Uechtritz was purchased at this time and M. Winkler donated his herbarium to the Museum, which he had compiled for 30 years (Wiktor 2002).
At the end of nineteenth century, Ferdinand Pax (the elder) merged all the University botanical collections under the name of the Botanical Museum. His own collections were also included in the Museum at this time. Before merging, von Uechtritz's herbarium of Silesian plants was handed over to the Herbarium Silesiacum (then still independent) on his initiative. In exchange for Uechtritz's herbarium, the Botanical Museum later received the Herbarium Henschelianum (part of the Herbarium Silesiacum) with ca. 100,000 sheets.
As a result of these mergers and gifts, the Wrocław herbarium had acquired an extensive and significant collection of specimens from Europe (especially the Mediter-ranean) and the rest of the world. These were collected by botanists such as Hubert historically important specimens "hiding" in such under-appreciated collections are improtant for taxonomy, nomenclature and biodiversity studies. Using the WRSL herbarium, we address the importance of collections like WRSL as reservoirs of valuable data that are relevant to experts who are involved in taxonomic revision.

Assessing the significance of the WRSL collection
The WRSL herbarium is currently divided into three parts: the Herbarium Generale, the Herbarium Lauterbachi and the Herbarium Silesiacum. The Herbarium Generale (about 375,000 specimens including about 75,000 spore-bearing organisms) holds the plant and fungal material from around the world, excluding Lower Silesia, Poland, the Herbarium Lauterbachi (about 50,000 sheets) contains plants from New Guinea and Melanesia and the Herbarium Silesiacum (about 90,000 specimens) (K. Świerkosz, pers. comm., 2019) houses plants from Lower Silesia, Poland.
Generally, the importance of particular natural collections depends not only on their size, but also can be measured on the percentage or the absolute share of type specimen types (Sutory 1997). In 2017, digitalisation of the WRSL collection was initiated and was subsequently able to be accessed via GBIF.org (Świerkosz 2017); this work is on-going but only 25,000 specimens (4.9%) are currently listed in a database (K. Świerkosz, pers. comm., 2019). Therefore, we decided to assess the importance of using specimens of the genus Juncus (Juncaceae) stored in the Herbarium Generale (to date, no Juncus specimens from WRSL are included in GBIF.org database to facilitate this task). The reasons for this choice were: 1) type specimens of Juncus have never previously been assessed in the WRSL Herbarium; 2) the genus Juncus is rich in species from regions where the herbarium has geographical strengths, 311 are listed by Kirschner et al. (2002a, b) and 3) the first author of this paper is a specialist in Juncus taxonomy, which considerably aided the analysis of specimen status.
We evaluated the following factors (Sutory 1997): 1) the originality of the collection, including the number of types and other historically-important specimens; 2) the size of the collection, i.e. the total number of specimens; 3) the geographical scope of the collection; 4) the length of the period represented by the collection; 5) the number of duplicates and 6) the physical condition of the collection (well-prepared, well-preserved and undamaged and well-stored material with appropriate labels). Herbarium sheets with plants representing a single taxon that were gathered in the same locality and on the same date by the same collector, were regarded as duplicates. Additionally, we analysed the specimens with respect to: 1) the person who collected the material in the field; 2) the collection from which they came (i.e. to whom they belonged before accession in WRSL) and 3) the floras/exsiccatae from which they came.
We catalogued all Juncus specimens ourselves, paying particular attention to all types and other historical material, which we identified, based on the latest mono-graph (Kirschner et al. 2002a, b), from which we took the current nomenclature of the genus. The localities and dates of sets for historical collections, especially those of C.F. Ecklon & C.L.P. Zeyher and J.F. Drège, were deciphered from literature (Meyer 1832;Drège 1847Drège , 1848Buchenau 1875Buchenau , 1890Buchenau , 1906, which enabled us to recognise many Juncus types. The Juncus sets are stored in seven herbarium boxes indexed as separate fascicles, numbered 151-157 and an extra 43 herbarium sheets were kept in a separate folder. We analysed 2,192 herbarium sheets in total. We treated a separate collection with its own label as a separate herbarium sheet, as specimens from three different localities could have been mounted on one herbarium sheet (we treated these as three separate herbarium sheets). We identified 2,222 taxonomic records, since part of the material represents mixed sets. We conducted our research from scratch, since only two Juncus types identified in the Herbarium Generale had been previously labelled using a red label. Thus, no other Juncus types stood out from other herbarium sheets. Our results were also compared with those within the Global Plants Database (plants.jstor.org, accessed on 16 Apr 2020) and additional herbaria, not mentioned by Kirschner et al. (2002aKirschner et al. ( , 2002b) that store other type specimens/duplicates of names we assessed, are added to the last column of Table 1 and marked with an asterisk (*). Duplicates of selected type specimens stored at WRSL were also compared with those stored in other herbaria (present in the Global Plants Database). When comparisons were made, we considered the physical condition of specimens, quantity of materials, different annotations, kinds of labels and plant parts.

Type and other historically-important material
We found 78 specimens that are historically or nomenclaturally important (Table 1): two holotypes, 20 isolectotypes, 14 isotypes, 29 syntypes (including one probable syntype of Juncus exsertus Buchenau (1875: 435)), three paratypes, one isoneotype, five sheets of historically-relevant material (for names not validly published) or additional material from type localities collected by the author of the name (so-called "topotypes") and four sheets of probable original material to be analysed in the future (Fig. 1). Holotypes, isotypes and isolectotypes constitute 46.2% of all types (and other historically-and nomenclaturally-important specimens) of Juncus specimens recognised at the WRSL. The most significant discovery was the identification of the three following Juncus types in the WRSL Herbarium (see also remarks for them in Table 1 (Buchenau, 1875: 484). Thus, the lectotype was designated (in B, destroyed). Isolectotype (the only duplicate known) was rediscovered at WRSL (the specimen includes collection date (i.e. 5 Mar 1816), which corresponds to the date included in the protologue). The syntype (Bergius specimen at W) does not include a collection date.
The origin of Juncus type specimens at WRSL according to country is presented in Fig. 2.

Exsiccata series
In the WRSL Juncus set, the following exsiccatae are particularly well-represented (the number of herbarium sheets is shown in parentheses): Rchb.  (11), Flora des NW. Himalaya (10).

Discussion
The Herbarium Generale of the mid-sized WRSL herbarium is rich in specimens relevant to the nomenclature of Juncus and contains 78 specimens (3.6% of all Juncus specimens examined, see Table 1), with an average of 11 sheets per fascicle (i.e. herbarium box). Seventy-six (of 78) of these historically-important specimens (types, original material and specimens collected at the type locality by the author of the name) were not identified as such before our study. This significant number of types highlights the significance of the analysed set and of WRSL more broadly, for the study of taxonomy and nomenclature (Sutory 1997). To put this into context, the approximate percentages of types stored in other historically-important herbaria are as follows: K (5%), W (3.6%), BM (2.6%) [cited from herbaria websites, which include the total number specimens stored]. Our study revealed that the Juncus set at WRSL is a valuable collection globally with respect to the number of historicallyand nomenclaturally-relevant specimens. Often, specimens included are associated with research conducted involving a given group of plants in the academic centre housing collections. The majority of Juncus specimens (ca. 89.7%) date from before the Second World War and specialists studying the genus Juncus did not work at WRSL during that time. This suggests that the rest of the WRSL collection might also contain similarly high percentages of such historically-and nomenclaturallyrelevant specimens.
As the genus Juncus is rich in species (311 species, Kirschner et al. (2002a, b)), we consider that extrapolation of our results to other genera is appropriate. We assume that descriptions of taxa new to science before 1946 were equally common within most taxonomic groups and specimens belonging to different plant genera/families were sent to the WRSL herbarium equally often.
Only a small fraction of global herbarium specimens had been computerised by the end of last decade (Lughadha and Miller 2009). Despite the recent acceleration of the digitisation of herbarium collections (as of early 2015, the number of scanned specimens within the world's largest virtual herbaria was 18.4 million), we are far from fully digitising all collections (Seregin 2016). Even a small percentage (1-2%) of computerised specimens can drastically reduce research costs and help scientists focus on collections that are likely to contain the most information-rich specimens (O'Connell

Conclusions
The history of German-Polish herbaria, including WRSL, is very turbulent. A detailed examination of Juncus, as a case study, confirms the value of the WRSL collection in historical terms. That a significant number of historically-and nomenclaturallyimportant specimens at WRSL was acquired passively (Juncus was of no special interest to German or Polish scientists at the time) suggests that more such specimens may be found within the collection for other genera. Digitisation and taxonomic revision of material will facilitate the confirmation of the richness of the collection.
Other large type collections contain well-preserved specimens, well-prepared catalogues (often available on-line) and are well-known to scientists. However, the WRSL collection is not only unique, as confirmed here, but not well-known to date.
Some Juncus type specimens, listed here, can be found easily in a large number of other collections. However, some are preserved only at WRSL because many types, previously stored in Berlin, were destroyed during the Second World War. Although we researched only a few parts of the WRSL collection, we are convinced that duplicates of many type specimens destroyed in Berlin can be found in Wrocław. Uncatalogued herbaria like WRSL with turbulent histories can be a source of collections important for the study of biodiversity.
We selected Juncus as a case study since the collection at WRSL covers the entire distribution range of the genus. Therefore, it likely reflects the general situation in other groups of plants in the herbarium.
Currently, many herbarium sets in Europe are still being catalogued (and many remain undigitised). However, many old collections are indeed valuable and their type and other historical collections have the potential to facilitate taxonomy and nomenclature and, in addition, enhance our knowledge of biodiversity through application of correct names.