As to collate a uniform and validated index to the world's known species the Integrated Taxonomic Information System (ITIS), a partnership of federal agencies and other organizations from the United States, Canada and Mexico, with data stewards and experts from around the world and Species 2000, an autonomous federation of taxonomic database custodians, involving taxonomists throughout the world, joined forces to setup The Catalogue of Life (COL) in 2001 further developing it [DP 2009] up till today. As of October 2014 it contained 1.5 million accepted and provisionally accepted taxonomic names (341 thousand of Plantae). With a similar agenda but focused on plants the Royal Botanic Gardens Kew and the Missouri Botanical Garden enabled the creation of The Plant List (TPL) combining multiple checklist data sets held by these institutions and other collaborators in 2010. As of September 2013 and the release of version 1.1 TPL contained 1.3 million scientific plant names of which 351 thousand are accepted species names.

With sufficient information on the establishment of the accepted scientific name the story of a species unfolds and literature can be screened for relevant information, including names that are no longer in use. Beyond literature there are freely available scientific data repositories of different kinds containing additional information. Members of the International Nucleotide Sequence Database Collaboration (INSDC) have been collecting and providing sequence information for 30 years accumulating about 178 million sequences of 340 thousand species and infraspecific epithets. The Barcoding of Life Datasystems (BOLD) supporting the generation and application of DNA barcode data, as of May 2015 offering over 4 million DNA barcode sequences supporting specimen identification. The Global Biodiversity Information Facility (GBIF) provides a single point of access to more than 500 million records, shared freely by hundreds of institutions worldwide, making it the biggest biodiversity database on the Internet. Names are also critical when building priority lists, e.g. the Convention on International Trade in Endangered Species of Wild Fauna and Flora (CITES), the International Union for Conservation of Nature (IUCN) or invasive species lists (e.g. Global Invasive Species Information Network, GISIN or Delivering Alien Invasive Species Inventories for Europe, DAISIE).

The vision of building a species database combining names with all kinds of useful data [Wilson 2003] was realized after E. O. Wilson’s 2007 TED Prize speech by the release of the Encyclopedia of Life (EOL) in February 2008. Connected to scientific names derived from different sources are common names, diagnostic descriptions, media (images, sounds, videos and maps), ecological, geographical, molecular and usage data.

All mentioned data providers offer a website where the user can search for respective information. Some of them also offer the possibility to retrieve information through an application programming interface (API). By using dynamic high-level general-purpose languages (e.g. Perl, Phyton, PHP and others) stakeholders can include respective data in their own (web) application. Additionally, scientists are able to retrieve and analyze data from many different taxa in little time by using either the API or third party software that facilitates the API (e.g. Banbury and O'Meara 2014).

The primary source for all the described information are, nevertheless, physical beings - individual members of respective species that are found in their natural habitat, cultivated on experimental fields or in botanic gardens.

Using science as fundamental criteria for the definition of a botanic garden, the first European botanic gardens were established in the mid-16^th century [Palmer 1985] and focused solely on the study of medicinal plants. The model of such a "physic garden" became popular and resulted in the recreation throughout Europe in the following decades. In 1621 the University of Oxford botanic garden (Oxford BG) was the first to be established in the United Kingdom (www.bgci.org/resources/history/).

Today the role of botanic gardens is much more diverse. Support of scientific research and economic endeavors (e.g. Centre of Economic Botany at Royal Botanic Gardens Kew, founded in 1847 [Desmond 1995]), involvement in education, public relations, improvement of human well-being [Waylen 2006] and plant conservation [Ashton 1988, Hurka 1994, Maunder et al. 2001, Donaldson 2009] are some examples.

Botanic Gardens Conservation International (BGCI) is the world authority on botanic gardens and plant conservation. It represents about 700 members, mostly botanic gardens, from 118 countries. A traditional practice among botanic gardens is the exchange of plant (genetic) resources by annually offering seed catalogues (indices seminum) from which other gardens can order to develop their own collection. This practice is believed to have started in the late 16 nth century at the Oxford BG [Aplin et al. 2007] and has recently been critically discussed [Aplin and Heywood 2008, Heywood 1964, Howard et al. 1964]. A more recent development came with the dawn of the information age. Some gardens now provide access to their botanic collections through a web interface and interested parties can request material for their collection or research online. As a result of the Convention on Biological Diversity (CBD) the International Plant Exchange Network (IPEN) was endorsed by the EU-Consortium and approved by EU-gardens as model to meet the requirements of the CBD on access to genetic resources and benefit-sharing [Robbrecht and Bogaerts 2004].

The possibility to develop a collection or scientifically utilize respective species without the need for expensive expeditions and simultaneously complying to the CBD is appealing. However, the exchange and cultivation of plant species also has less favorable consequences. The introduction of exotic species by botanic gardens has frequently been associated with the potential for escape and evolution of invasiveness [Hulme 2011, Maunder et al. 2001]. The tremendous ecological and economic effects invasive species have [Pimentel et al. 2000CBD 2001] stimulated governments around the world to establish preventive measures (e.g. European Union Regulation No. 1143/2014, United States Federal Law: National Invasive Species Act). Other consequences of cultivation that are of great importance for species conservation with the ultimate goal of reintroduction are a range of threats to the genetic structure of respective species [Maunder et al. 2004,Husband and Campbell 2004].

The aim of this study was to gain an overview on the currently available genetic resources offered by botanic gardens using taxonomic names of seed catalogues, to gather relevant information (i.e. taxonomic status, IUCN Red List status, geographical distribution and invasiveness reports) and to evaluate the information and its sources in the context of improving facilities dealing with plant genetic resource.

PDF documents were converted into XML, parsed to extract taxa names which then were saved into a local database (IS Taxa, Fig. 1). The parser was restricted to names of species, subspecies, varieties, forms, cultivars and hybrids of the form "Genus x species". In a preliminary study of names contained in seed catalogues data from TPL was stored in a local database. This database was complemented by importing all names available in csv format from the TPL website to be used in this study.

To verify the existence and to retrieve the taxonomic status of each name I used data from TPL (The Plant List Version 1.1, available from http://www.theplantlist.org/), EOL (Encyclopedia of Life. Available from http://www.eol.org. Accessed March 2015) and COL (Catalogue of Life. Available from http://www.catalogueoflife.org/, Roskov et al. 2014). EOL and COL provide application programming interfaces (API) that I used to access the data. TPL on the other hand does not offer a true API thus was accessed through a locally installed database.

Name status information was retrieved by respective API functions summarized in Table 1. Since the status terminology of data providers was different I mapped the returned values to the status scheme used by TPL (Table 2).

In detail, the retrieval of status information was done in two steps. First, only exact matches were used to retrieve the taxon status. In case of EOL the parameter exact was set to true. For CoL and TPL only exact matches are returned by default. In case more than one exact match was returned the respective status was set to “ambiguous”. Secondly, all names that did not return any result (NA) were re-evaluated. In case of EOL, firstly, the parameter exact was set to false (default) and, if still no result was returned, the name of the next higher hierarchical level (i.e. species in case of infraspecific epithets or genus in case of species names) was searched for. If found, the respective taxon page was used to retrieve all available names of children (i.e. infraspecifics or species) stored in hierarchies available through that page. If any of the returned names was an exact match or very similar (i.e. levenshtein distance < 3) it was retained as alternative. In case more than one possible alternative was found the name status was set to “ambiguous” unless the closest match was a perfect match (see discussion for details). If a single perfect match or alternative match was found the corresponding status was used as result. For CoL and TPL the same approach was realized using the API or mySQL queries to retrieve and test respective data.

For subsequent analyses the results of the taxon name status evaluation were merged and a list of unique names was compiled (Unique Name List, Fig. 1). In addition to taxonomic status the source's unique identifier for the taxon was stored for later use (Unique Identifier List, Fig. 1).

To evaluate status discrepancies between the sources, the results from all three data providers were combined. The status of each unique name was compared between pairs of data sources (i.e. TPL vs. COL, TPL vs. EOL, COL vs. EOL). The status was either identical or not. In the latter case the comparison was saved and used to evaluate discrepancies in more detail.

The list of unique verified taxon names (UNL, Fig. 1) was used to determine the IUCN Red List status for each taxon. First, plant data was downloaded from the IUCN website and installed in a local database. Second, the database was queried and respective status information was saved.

To compile an overview of the geographic distribution of available genetic resources, I used two approaches. Firstly, I used the list of verified taxon names with unique identifiers from TPL, EOL and COL (UIL, Fig. 1) to retrieve distribution data using the respective data source. Secondly I used the list of unique verified taxon names (UNL, Fig. 1) to query a locally compiled flora database. In both instances the aim was to assess the distribution of the respective taxon according to the continental scheme of the International Working Group on Taxonomic Databases For Plant Sciences (TDWG) [Brummitt 2001] while tropical and temperate Asia were combined to Asia. For the first approach this meant to parse respective text content delivered by data providers for the occurrence of geographical terms listed in the World Geographical Scheme for Recording Plant Distributions by the TDWG (see Table 1 - 5, Brummitt 2001).

The Plant List Distribution Data

In case of TPL only some of the data sources offer information about distribution that I could use for this analysis: If the TPL data source was either the World Checklist of Selected Plant Families (WCSP), the International Organization for Plant Information (IOPI) or iPlants, the stored identifier (UIL) was used to access taxon information at WCSP which is based on TDWG scheme. In case the TPL data source was the International Legume Database & Information Service (ILDIS), I retrieved the taxon report from their website by using the stored identifier and parsed for geographical records with respective region names. If Tropicos was the TPL data source I used their web service to retrieve distribution data of the stored identifier. For TPL data provided by the The International Compositae Alliance (TICA), a locally installed database containing occurrence information of respective taxa was used (data was kindly provided by Kevin Richards).

The Catalouge of Life and EOL Distribution Data

The web services of COL and EOL could be accessed directly using stored identifiers (UIL).

Flora DB Approach

As a second approach and to supplement the geographical information retrieved from the previously mentioned sources, I compiled a flora database using information of different available floras and regional name checklists (Table 3). Processing the names stored in the list of unique taxa (UNL), each occurrence in a certain flora or regional checklist was used as indicator for the distribution of the respective taxon.

To determine the number and names of potentially invasive taxa cultivated by botanic gardens, I used the Global Invasive Species Information Network (GISIN) web service (http://www.gisin.org/GISIN/GISINWebService) to query each taxon name (UNL, Fig. 1) and retrieve occurrence datasets. If the dataset contained an entry that reported the taxon as exotic and harmful it was considered to be potentially invasive. Additionally, I compiled a database containing all Magnoliophyta, Gymnospemae, Pteridophyta and Bryophyta taxon names of DAISIE. Again, all unique verified names (UNL) were checked for occurrence in the local DAISIE database.

In the exact match pass, 13'764, 12'939 and 14'338 names could be verified using COL, EOL and TPL respectively, constituting a success rate of 79.8 to 88.4 % (Fig. 3a). Searching for similar names using a levenshtein distance of < 3 and using non-exact search resulted in additional 566, 1'894 and 686 verified names raising the success rate to 88.3 to 92.6 %. 1'894, 1'391 or 1'200 names (7.4 to 11.7 %) could not be verified using COL, EOL and TPL respectively (Fig. 3b). All verified unique and unambiguous names with their identifiers (UIL, Fig. 1) for each of the sources (COL: 13'454, TPL: 14'469, EOL: 13'960) were used to retrieve information on the geographical distribution of respective taxa. Combining verified names of all three online sources and removing redundant names resulted in 15'973 unique entries (98.5 % of the original 16'224) which were used to retrieve information on conservation status, geographical distribution and invasiveness (UNL, Fig. 1).

a

b

Figure 3.

Results of the taxonomic name verification and status check (Suppl. materials 2, 3) using COL, EOL and TPL.Suppl. materials 2, 3

a: Overview on the taxonomic name status (accepted, synonym, unresolved, ambiguous and NA = not available) of 16'224 unique taxon names retrieved from 135 indices seminum determined by inquiry using three online sources (TPL, EOL and COL). 
b: Overview on the recovery of taxonomic names and their status by using levenshtein distance < 3 and data from three online sources (TPL, EOL and COL). 

Comparing the status information retrieved during name verification from each of the sources 3'552 and 3'997 discrepancies were found between TPL and COL, and TPL and EOL respectively, while only 1'816 discrepancies were detected between COL and EOL (Fig. 4a). Among the 3'552 and 3'997 discrepancies between TPL and both, COL and EOL, the majority consisted of names accepted by COL and EOL but being considered ambiguous (2'048, 1'834), synonyms (612, 547) or unresolved (424, 566) by TPL (Fig. 4b). On the other hand names considered to be synonyms by COL and EOL were found to be accepted (532, 408), unresolved (222, 149) or ambiguous (167, 177) by TPL (Fig. 4c). Lastly a minority of ambiguous COL and EOL names were found to be accepted (95, 184), synonyms (45, 117) or unresolved (9, 15) by TPL (Fig. 4d).

a

b

c

d

Figure 4.

Taxon name status discrepancies between TPL, EOL and COL (Suppl. material 4).

a: Comparison of status discrepancy: Comparison of TPL with COL and EOL (green), COL with EOL and TPL (blue) and EOL with COL and TPL (red) are represented by a triangle (center = 0 point, two sides with a length determined by the number of discrepancies with the source denoted in the corner and one side connecting these two). 
b: Discrepancy details. Showing the TPL name status for names considered to be accepted by EOL (left bar) and COL (right bar) 
c: Discrepancy details. Showing the TPL name status for names considered to be synonyms by EOL (left bar) and COL (right bar) 
d: Discrepancy details. Showing the TPL name status for names considered to be ambiguous by EOL (left bar) and COL (right bar) 

Out of the 15'973 unique taxon names (UNL), 1'066 (7 %) were found to be assessed by the IUCN Red List. One of the taxon names returned the status extinct (EX). Euphrasia minima Jacq. ex DC., according to IUCN a synonym for Euphrasia mendoncae Samp., can be found in the seed list of BGU Lautaret. The same name with different authority (Euphrasia minima Schleich.) can be found in the seed list of CJB Geneva. Three of the taxon names lead to the status extinct in the wild (EW): Bromus bromoideus, Lysimachia minoricensis and Mangifera casturi. 267 (1.7 %) names fell into one of the IUCN Red List threatened categories (vulnerable, endangered and critically endangered) and the remaining 795 included 84 near threatened (27 in the lower risk category), 6 lower risk conservation dependent, 620 least concern (39 on the lower risk category) and 85 data deficient taxon names (Fig. 5). All taxon names and their IUCN Red List status are compiled in Suppl. material 5.

Figure 5.

IUCN Red List Status of 1'076 taxon names found in seed catalouges of 135 botanic gardens (Suppl. material 5). Status categories include Extinct: Extinct (EX) and Extinct in the wild (EW); Threatened: Vulnerable (VU), Endangered (EN) and critically Endangered (CR); Others and Lower Risk (LR): Near threatened (NT,nt), Least Concern (LC,lc), Conservation dependent (cd); and Data deficient (DD)

To retrieve information about the geographic distribution using a locally installed flora database, a list of unique names (UNL, Fig. 1) containing 15'973 entries was used. For the retrieval of respective information using COL, TPL and EOL, unique identifier lists (UIL, Fig. 1), containing 13'454 (COL), 14'469 (TPL) and 13'960 (EOL) entries, were used. About 62 % of the unique name list entries were detected in one or more of the floras (Table 3) and their geographic distribution could be approximated. Among the three external sources COL returned data on the distribution of 87 % of the names. Considerably less success was achieved using EOL (46 %) and TPL (36 %). Merging all information into one dataset slighty increased the overall success to 88 % (Fig. 6).

Figure 6.

Geographic distribution of studied taxa (Suppl. materials 6, 7). Number and ratio of taxon names connected to distribution data accessible via EOL, TPL and COL, occurring in a locally compiled flora database (FloraDB) and found by the one or the other approaches (Merged).

Most of the evaluated taxa occur in Eurasia (Fig. 7b, c) followed by Africa and the Americas (Figs 7d, 8a, bFigs 7c, d, 8a) and Australasia and the Pacific (Fig. 8c, d). The distribution after merging all information is displayed in Fig. 7a. Taxa exclusively distributed in one of the regions are also mostly from Eurasia and only little from the Americas and remaining regions.

a

b

c

d

Figure 7.

Geographic (Europe, Asia, Africa, North America, South America, Australasia and Pacific) distribution of plant genetic resources according to Flora, EOL, TPL and COL information (Suppl. materials 6, 7). Part I.

a: Taxa inclusively (dark) and exclusively (light) distributed in major regions according to combined flora, COL, EOL and TPL data.  
b: Taxa inclusively (dark) and exclusively (light) distributed in Europe by flora, COL, TPL and EOL data. 
c: Taxa inclusively (dark) and exclusively (light) distributed in Asia by flora, COL, TPL and EOL data. 
d: Taxa inclusively (dark) and exclusively (light) distributed in Africa by flora, COL, TPL and EOL data. 

a

b

c

d

Figure 8.

Geographic (Europe, Asia, Africa, North America, South America, Australasia and Pacific) distribution of plant genetic resources according to Flora, EOL, TPL and COL information (Suppl. materials 6, 7). Part II.Suppl. materials 6, 7

a: Taxa inclusively (dark) and exclusively (light) distributed in North America by flora, COL, TPL and EOL data. 
b: Taxa inclusively (dark) and exclusively (light) distributed in South America by flora, COL, TPL and EOL data. 
c: Taxa inclusively (dark) and exclusively (light) distributed in Australasia by flora, COL, TPL and EOL data. 
d: Taxa inclusively (dark) and exclusively (light) distributed in the Pacific by flora, COL, TPL and EOL data. 

COL appears as major contributor of distributional data in case of most regions while EOL is a major contributor primarily of Asian and American occurrences. The flora database makes similar contributions as COL (European, African and Australasian occurrences), EOL (Asian occurrences) and TPL (American occurrences).

Of the 15'973 unique taxon names (UNL) 3.9 % (616) have records in the GISIN database. 577 of these taxa have been reported as exotic, 183 also as harmful (invasive) and 39 neither as exotic nor as harmful. All exotic and ivasive taxa are contained in Suppl. material 8. DAISIE offers information on 6'091 European alien plant taxa. Evaluating the UNL 19.6 % of the names (> 50 % of the DAISIE taxa) appear to be European alien taxa. DAISIE also offers a list of the 100 worst invasive aliens in Europe. Among these there are 18 terrestial plant taxa and 12 of these are found in the UNL. All detected DAISIE taxon names are contained in Suppl. material 9.

Of 15'973 unique taxon names 51 % (8'489) can be found in the BOLD taxonomy database. 7'691 of these taxon names also have one or more public barcode record registered with BOLD.

Botanic gardens cultivate and store many different plant species for scientific research and education. They should maintain a high degree of consistency with current systematic research and should have adequately documented specimens of correct species identity. Factors that most likely compromise these points are insufficient financial commitment and the decline of expertise in classical botany due to reduction or omission of respective topics and training in the biological curriculum.

An approach aiming to resolve this issue, promising specimen identification on the species level using a small standardized region of the (plant) genome [DNA barcoding sensu Hebert et al. 2003], so far, has been shown to be very limited in the plant kingdom [Taberlet et al. 2007, Seberg and Petersen 2009, Roy et al. 2010]. With their particular life style and high degree of flexibility plants have been regularly shown not to apply to the conventional biological species concept (sensu Mayr) causing considerable difficulties for any authentication method. The facts that for only 48 % of the analysed taxa at least one DNA barcode sequence exists and the offical plant barcoding markers [CBOL 2009] with considerable limitations in species level identification [Federici et al. 2013] still represent a significant part of those, are pressing reasons to either complement better suited sequence markers or to start looking for a different practical solution. The development of sequence markers based on amplified DNA fingerprinting fragments is still too laborious to be practical. Using highly species specific functional DNA fingerprinting patterns [Bardini et al. 2004, Breviario et al. 2007, Poczai et al. 2013] obtained by fast and reliable methods [Gavazzi et al. 2012], however, might be a feasible alternative.

Botanic gardens are in their nature places where many different species are brought together that under normal circumstances would never meet (artificial sympatry). Additionally their cultivation inevitably means that they are put into a novel environment. Both factors entail complex consequences that are relevant for our understanding of plant evolution and for conservation biological projects [Ashton 1988, Husband and Campbell 2004, Maunder et al. 2004, Schaal and Leverich 2004].

How can botanic gardens maintain a high degree of consistency with current systematic research and provide authentic genetic resources ? With internet access a collection can easily be connected to all available data and thus provide local staff and scientists alike with information of their particular interest (e.g. information on natural habitat, taxonomy, invasiveness, conservation status, all known common and scientific names, traditional usage, medicinal potential, etc). As demonstrated in this study in most cases the different sources are in agreement on taxonomic names but there is also a significant number of conflicting cases. Creating brought awareness of these cases increases the chance of research that, ultimately, will resolve conflicts. In this regard it is favorable to integrate taxonomic name status information that clearly indicates if a name should be used with caution (e.g. the status "Unresolved" of TPL) that at the same time indicates the need for further study.

Judging from the analysed seed lists there are still many botanic collections with specimens of mostly unknown origin. This fact alone reduces scientific value considerably [Böcher and Hjerting 1964]. Effects of cultivation in an artificial environment (selection, hybridization) using a small sample of the genetic structure (genetic bottleneck) of a species [Barrett et al. 1991, Maunder et al. 2004] cannot be traced and certain key aspects would have to be studied to ascertain, that the individual is still representative for the respective species. For that a new adequate [Brown and Briggs 1991] sample would be necessary that at the same time would render the former sample obsolete. To increase the scientific value of botanic collections accessions with unknown origin should if possible be replaced with adequately documented ones. Having worked with several plant genera obtained from different botanic gardens misidentifications or mix-ups appear to be as common as proposed in other publications [Hurka 1994] and pose another problem for education and research.

Results of taxon name verification and status check are based on data provided by third parties to one or more of the data portals used in this study. In case of EOL the name status is determined by EOL curators. A name as part of one or more hierarchies is either the preferred name or it is not. Only COL and TPL provide a name with a status that is based on the respective data source. According to TPL 98 % of all status values were directly derived from the data source that supplied the name record while only 2 % are a result of automated conflict resolution processes. Information provided by COL is based on global species databases that "have been validated for inclusion by independent peer reviews". How many peers are necessary to validate databases that contain names of thousands of plant species from hundreds of families? Is it the few or the many peers that bring about an objective assessment of taxonomy ?

The coverage of taxon names is highest (92 %) using the plant specific portal (TPL) closely followed by that of EOL (91 %) and COL (88 %). Using the levenshtein distance as recovery approach is adequate for simple typos. In case of more complex spelling variants of a name the fuzzy algorithm [Rees 2014] would be more appropriate. Most of the 11.6 % of names verified by EOL in the second round of the name analysis are not spelling variants but names that were not found due to the used exact search option. This can be explained by an inconsistent name field format. It does not only contain the taxonomic name, like in the case of COL and TPL, but in the mentioned instances also includes the authority of the name. The similarity of results between COL and EOL, particularly the lower level of status discrepancies, is based on the fact that EOL is using COL as its taxonomic backbone since its launch in 2008.

Considering name statuses TPL apparently offers a more heterogeneous view. The status "Unresolved" which in most cases indicates that the respective name status has not yet been assessed by the data provider is unique and makes TPL a distinguished source for plant names. In addition, a much higher degree of name ambiguity can be detected using TPL. About 4 - 6 times more names are considered ambiguous because of the existence of homonyms, many of which are not found using COL and EOL. The primary concern of this study was to find names and associated information. Having an exhaustive name space increases the number of hits when mining for data. This, however, could also result in a more contaminated dataset. Having complete information on a name, never the less, means being able to take all information into consideration when evaluating associated data.

It is curious to find a presumably extinct species (Euphrasia mendoncae Samp.) in a seed list of a botanic garden. The species was described as Portuguese endemic in 1936 and was never found again. IUCN therefore classed the species as extinct. Along with this information doubt on the identity of the portuguese taxon is expressed in the IUCN dataset and Euphrasia minima Jacq. ex Dc. is introduced as synonym of E. mendoncae. The presumed synonym E. minima Jacq. ex DC. appears in the IS of BGU Lautaret 2014, reporting the collection site at 2500 m height in Col du Galibier, France. While COL lists E. mendoncae Samp. as accepted name, TPL lists it as Unresolved, supporting the notion of insufficient data on the identity of the species described in 1936. While E. minima Jacq. ex DC. is not listet in TPL the name appeared as a synonym of E. officinalis subsp. officinalis L. in COL (2014) and in the meantime (2015) is classed as doubtful synonym of E. officinalis subsp. officinalis Jacq. ex DC.. This example demonstrates that names, as important as they are, sometimes are too ambiguous to be meaningful.

1.7 % of the taxon names surveyed fell into a IUCN threatened category. This proportion appears to be very low and might suggest failure of conservation initiatives that aimed to draw botanic gardens closer into the conservation network (e.g. IUCN / WWF Plants Conservation Program, 1984). However, conservation efforts are not expressed by the quantity of threatened taxa maintained by a botanic garden alone (For performance indicators for ex situ conservation facilities see Havens et al. 2006). Conservation efforts start with promoting awareness [Williams et al. 2015]. Awareness that human civilization does have a major impact on ecosystems which sustain, amongst ourselves, many different interconnected species. The destruction of these ecosystems not only could mean the loss of those species and eventually endanger ourselves, but could result in a loss of diversity that one day may have lead to the discovery of a cure for a disease, a future crop that has a natural tolerance or an exceptional nutritional value. Public opinion, education and awareness are key aspects for the establishment of an evolutionary ethic as part of our societies which would lead to a broad public recognition that the existence of other species is an integral part of our own [Hurka 1994, Frankel 1974].

Reasons for a low proportion of threatened taxa distributed by botanic gardens might be that the distribution itself is considered to be counter productive by leading to extreme genetic depletion (i.e. gardens prefer to order seeds instead of collecting them from natural habitats). Another point may be the necessary expertise that only through considerable commitment of time and resources can be established. Additionally due to a lack of sufficient documentation and prior knowledge of conservational aspects [Hurka 1994], present collections of botanic gardens are for the most part still meaningless for conservation purpose. Finally the fact that about 4600 (42 %) of European threatened taxa have been detected in European botanic garden collections [Sharrock and Jones 2011] but not in my study may indicate that most of these taxa are not distributed as seeds or most of the gardens holding the taxa are not included in my sample.

Depending on the source the distribution of an taxon can be determined on the continental, country or even regional level. Unfortunately the form of geographic information is not all consistent. Besides the TDGW scheme [Brummitt 2001] there is also simple text that if not combined with additional markers results in ambiguity (e.g. Georgia is either a state of the USA or a country located between the Black and Caspian sea). Another issue is the undetermined completeness of the information. Does a source contain complete information on the geographical distribution of a taxa or is it limited? There are no indicators to answer that question. Much detail of distribution is still hidden. Datasets gathered on the regional level (geobotanic mappings) or civil science reports on occurrences could be integrated into respective repositories (e.g. GBIF). Studies on distribution of certain taxa including community patterns will be possible as soon as coverage is sufficiently high and data is freely accessible through one of the sources mentioned earlier. Similarly, EOL's TraitBank [Parr et al. 2014] - a new text mining approach inferring traits (including distribution, habitat, elevation, interactions, etc.) from the occurrence of specific patterns (Ontology of Biological Attributes, Dietze et al. 2014) in content of the respective species enriches the web of species knowledge. While in the past one had to search and filter information in the form of books manually, consuming a lot of time and patience, today, thanks to the efforts of EOL and others that prepare and provide data on a freely accessible platform, we can answer questions more quickly and, more importantly, approach questions differently [Hart 2014, Wright and Seltmann 2014, Barve 2014].

On the 1st of January 2015 an EU regulation on the prevention and management of the introduction and spread of invasive alien species (No. 1143/2014) came into force. It aims to address the adverse impact alien invasive species have on biodiversity, ecosystem services, human health and the economy in the EU Member States. Botanic gardens, without doubt, create artificial situations for species. Hybridization, as one possible consequence, has been shown to be an important factor in evolution [Whitham et al. 1994, Martinsen and Whitham 1994] and as promotor to invasiveness [Ellstrand and Schierenbeck 2000]. For the European Botanic Garden Consortium (EBGC) "it is vital that Botanic Gardens take steps to prevent future problem taxa from establishing through their collections" and direct stakeholders to "Initiatives such as Delivering Alien Invasive Species Inventories for Europe and North European and Baltic Network on Invasive Species as well as National Initiatives, such as Harmonia - Invasive species in Belgium" which provide detailed databases on respective species. One major problem mentioned is, that "it can be difficult for garden managers and curators to obtain summary lists that provide at a glance indications of problem taxa". This difficulty subsequently leads to seed lists containing potentially invasive species (e.g. Acacia dealbata Link , Ailanthus altissima (Mill.) Swingle, Echinocystis lobata (Michx.) Torr. & Gray., Rhododendron ponticum L.) without any note or indication. The identification of high-risk taxa is, however, the responsibility of those who introduce exotic species - may that be botanic gardens or the horticultural industry [Wittenberg and Cock 2004, Baskin 2002]. I demonstrated the integration of data from publicly available sources (GISIN and DAISIE) to provide at a glance indications of potentially invasive taxa. It would be possible to establish a publicly available web-tool that takes a simple list of taxa and in return produces a list of the supplied taxa and their invasive potential as demonstrated in this study (Suppl. material 8 and Suppl. material 9). Even more convenient would be if collection databases contained dynamic information provided by respective sources. Garden managers and scientists would be able to act with little latency evaluating and eliminating potential threats [Hulme 2014, Hulme 2011, Sharrock 2011Hulme 2011, Hulme 2014]. For this to become possible respective sources need to establish an API and collection managers or stakeholders should discuss the creation of an adminsitrative solution suitable for botanic gardens. Since many data providers already offer an API the most likely obstacle in implementing such a tool would be the transition of collection data from the "old" to the "new" system. In any case the "light" version offering a list matching service similar to that offered by COL for name statuses (http://www.catalogueoflife.org/content/list-matching-service) would still be an option and represent progress.

Botanic gardens represent one major sources of plant genetic resources. As to be useful for education or any other scientific research the material needs to be authentic and properly documented. Particularly in a highly specialized scientific world, where experience in the determination of species has become rare other methods need to be considered. The true (i.e. accepted) name of a taxon is essential in many ways. To verify the identity of a specimen the true name will lead to a diagnostic description. With a verified specimen studies can commence and DNA based authentication assays could be established. True names lead to alternative names that lead to additional information. To assess the conservational, horticultural or medicinal value of a collection true names lead to respective data that can be evaluated. To measure progress towards a certain goal only true names are relevant. In this study I evaluated names of taxa offered in seed catalogues demonstrating that there are still significant numbers of taxonomic issues which need to be resolved before true names will be available. Furthermore I demonstrated that The Plant List Version 1.1 is currently the most complete and informative checklist for plant species names. Providers of critical data (e.g. CITES, IUCN Red List, invasive or toxic species catalogues) should make sure to include a complete list of known names for a particular taxon to maximize data exposure.

With increasing level of publicly available data through portals like EOL and publishers supporting open data sharing [e.g. Smith et al. 2013, Kenall et al. 2014], a dynamic integration of this data will ultimately revolutionize the way botanic gardens promote their agendas and ecological and evolutionary science is approached.