The compilation, editing and review of WCVP spanned the digital revolution. Therefore, the format in which the data were stored and distributed, the format in which data were obtained and accessed changed radically over time. However, the key elements and core workflows stayed largely the same. Here we present an overview of these workflows and then provide more detail on each workflow in turn, before describing the approaches to standardization, taxon acceptance, alternative taxonomies and international collaboration adopted during the preparation of what became the WCVP dataset.
Overview of workflows
Four main workflows operated in parallel:
- (i)
The A-Z workflow in which each name was mapped to a taxon concept, if possible, and the correct name for each accepted taxon concept identified, the others being recorded as synonyms of an accepted name or unplaced (when not mapped).
- (ii)
The family review workflow whereby, once a family checklist was complete in draft, the checklist or portions thereof were sent for expert review by taxonomists with relevant expertise, whether at Kew or around the world. Once feedback from expert review had been considered, and incorporated where appropriate, family treatments were published on the WCSP website.
- (iii)
The geographic workflow focuses primarily on recording the global distribution of each accepted taxon in terms of its presence in the botanical countries of the world3.
- (iv)
The update workflow is a continuous process of updating the dataset and incorporating new information gleaned from new publications, directly or via IPNI, as well as from user feedback and expert review focused on particular subsets of the data (e.g. genera).
The parallel operation of these four workflows over decades resulted in data being checked and rechecked multiple times. For example, the widespread grass Poa annua has 264 country codes added and 67 references listed, indicating that the record was checked at least 67 times. All workflows use as a starting point standardised nomenclatural data from IPNI or by screening the literature during the workflows and adding standardised names missing from IPNI as they are encountered. This process is described under the A-Z workflow and in the Standards Used section. All workflows involve taxonomic decision-making processes described in the Taxon Acceptance section.
The A-Z workflow in detail
The A-Z workflow started in 1988 and was completed on 4 December 2019. Name data from Index Kewensis (IK), which in 2000 was incorporated into IPNI, was initially retyped into a Firefox database and digitally copied from 1995. These raw data contained different formats reflecting non-standard formatting throughout IK’s history and lacked many dates of publication. The data were therefore first standardised using the standards described below before they were imported. In the early years, the coverage of the name data was still incomplete as names were added from IK in five batches between 1995 and 2008, each batch being standardised before being added to WCVP. Compilation began with the genus Aa Rchb.f. and continued alphabetically through all the genera. The relevant literature on the genus was then consulted at Botanic Garden Meise and Kew to ascertain the taxonomic status of each name (see below) and to add any distribution data encountered, as well as some 190,000 names missing from IK/IPNI. The latter step was particularly important for infraspecific names, as these were not systematically recorded in IK before 1971. During the compilation process, names missing from WCVP are added when encountered and therefore the infraspecific names should be largely complete for those in current use. In parallel, infraspecific names from other databases have been imported and some historic literature important to particular families has been screened for all names. During this process duplicates were removed and names were also checked to make sure they complied with the ICN5. Despite the above, many validly published infraspecific names are still missing from WCVP, especially historic names.
Each name was assigned one of three basic taxonomic statuses: Accepted, Synonym or Unplaced.
If a name was accepted in a publication as a distinct species with a published species concept, then the name was given the status ‘Accepted’ and geographic distribution data were added from that source. The database differentiates two different kinds of accepted name, the most frequently assigned accepted name status is given to native plants that occur in the wild while the “Artificial Hybrid” status is assigned to names that are correct and can be used for cultivated or naturalised taxa that are either man-made and do not occur in the wild (not wild plants) or those that may have a combination of natural and human-influenced components such spontaneous hybrids occurring in gardens or between native and introduced taxa.
If a name was listed as a synonym in a publication or in the original volume of IK, the status given would be “Synonym” and the name would be linked to the published accepted name. Several different types of synonyms are recorded, depending on their nomenclatural status as defined by the ICN: legitimate synonyms, illegitimate synonyms, not validly published synonyms, orthographic variants and misapplied.
If a name was not encountered in any of the literature consulted it was assigned “Unplaced” status. This status is also used for names that would be accepted but for the fact that they are illegitimate or not validly published under the ICN and therefore cannot be used for taxa that should be accepted but do not have a correct name in an accepted genus. The most common occurrence of this last case are names published in genera that are not accepted in WCVP, but for which a validly published combination in an accepted genus does not exist. Distribution is also added for unplaced names as they may relate to distinct species concepts and may become accepted under a legitimate, validly published name in future or can be used as an aid to resolve them at regional level.
The Family Review workflow in detail
The Family review workflow started in 1994 when RG was first employed by RBG, Kew. The idea is simple, a basic checklist is completed for a particular family. Relevant parts are then sent for review by taxonomic experts based in many different institutes worldwide. Recommended changes are then incorporated, and the checklist is published as a book and/or online on WCSP.
The families selected as World Checklist foci in the first instance were chosen because Kew had a particular research interest in that family, and expertise acquired over decades of research could be captured before key senior scientists retired (e.g. World Checklist of Euphorbiaceae13). Publication of a global treatment of a family at genus level also prompted and facilitated some family checklists. For example, the availability of a genus level classification of palms14 facilitated compilation of the palm checklist originally published as part of WCSP and as a book15, which in turn formed the basis for the online resource, Palmweb (www.palmweb.org). Similarly, a genus level treatment of Sapotaceae16 facilitated production of the World Checklist of Sapotaceae17 which is incorporated into the online Sapotaceae Resource Centre (https://padme.rbge.org.uk/Sapotaceae/data)).
As part of the review workflow, the full synonymy of each taxon concept is carefully checked to make sure the oldest available correct name is accepted for the concept. Sometimes a widely used name was accepted, even though an apparent earlier synonym was found. There are currently some 300 such synonyms indicated as possible earlier names pending further research. If these are confirmed as earlier names following further research it may be appropriate to consider formal rejection of these 300 names, in the interests of nomenclatural stability.
Approaches to family review varied because each plant family tends to have a particular expert community (or sometimes more than one) who collaborate best in different ways. For some families, experts were sent checklists of genera they requested to review, while for other families, such as Myrtaceae18, a workshop was held where all available experts were invited to put together a review strategy. For large families, such as Rubiaceae, experts agreeing to review the whole checklist worked through stacks of printout more than 60 cm high. All these diverse review approaches worked well and much improved the basic checklist. Once the review was completed, the family was added to the WCSP website and thereafter updated via the update workflow below.
The Geographic workflow in detail
The geographic workflow started in 1995, when data were first imported electronically into the WCVP database from the IK database at RBG, Kew. Data entry via this workflow is continuing and is expected to be completed by mid 2021.
This workflow primarily focuses on adding the geographic data from published Floras and regional checklists. Such publications differ in geographic scope from individual protected areas to continental works published over decades. Over the years, the geographic workflow checked first Europe, then Africa, Southern America, Northern America, Asia, Subantarctic, Pacific and is currently finishing the floras of India and Australasia for the families in review. Geographic distribution information was captured using the standard codes at the level of Botanical Country (level 3) of the World Geographical Scheme for Recording Plant Distributions6 (hereafter WGSRPD).
In addition to the geographic distribution information that was added for accepted taxa, synonymy and missing infraspecific names were also added from those publications in order to speed up the A-Z workflow. Lifeform19, and climate zones data (see Standards Used below) for accepted species are also added at this stage, although this data is currently published only for families included in WCSP due to the constraints of current data platforms. When the geographical codes added to a record were deemed to be complete or nearly so, the geography was also added in words, which could be very specific for local endemics or very general for widespread species. The wording of the text would, as far as possible, use the same wording as used in the WGSRPD or a combination thereof. So, a species occurring in BZE (Northeast Brazil) and BZL (Southeast Brazil) would be reported to occur in E. Brazil (Eastern Brazil).
The Update workflow in detail
The update workflow started in 1988, at the same time as the A-Z workflow and will continue as long as WCVP is maintained. The update workflow comprises three parts, weekly updates to the WCVP data available online, incorporation of user feedback and annual import of names added to IPNI in the previous year.
Every day new scientific insights are published and once a week all new journals and books that arrive in RG’s institute are screened and new data incorporated into WCVP. This was first done in the Belgian Botanic Garden library and from 1994 in the library of the Royal Botanic Gardens, Kew. There is also a proliferation of new online journals and eBooks, many of which come to our attention only if authors notify us of their publications. Automation of this literature review process has not been attempted to date due to: (i) the challenges inherent in detecting new synonymy or genuine nomenclatural corrections, as opposed to newly published names which are clearly indicated in compliance with the ICN; (ii) the need for a single process to ensure systematic coverage of the scientific literature; (iii) resource limitations.
The second source of updates comes from the daily stream of emails from users. Some 2,000 emails are received annually, and much improve the data. We aim to address all feedback within two weeks, although some queries requiring further discussion and library consultation may take longer and often involve discussions with the person sending the feedback. We also get requests to review particular genera from experts to whom we send data for review and then amend the database accordingly.
The third source of updates is names data downloaded from IPNI. Early in each calendar year, the scientific names added to IPNI in the previous year are imported manually to WCVP. They are then edited by adding taxonomic status and geography to each record in line with other workflows. In parallel, work is currently ongoing to reconcile all the names stored in the IPNI database with those stored in WCVP so eventually both datasets can share the same permanent IPNI identifiers.
Updates from the above sources become available to WCVP and POWO users on a weekly basis when the names data accessible from the WCVP web portal are updated. The full data download files are refreshed less frequently (currently every few months) because this requires a manual process, pending development of new infrastructure, including an Application Programming Interface.
Standards used
From the outset of compilation work internationally agreed standards have been used to standardise the data. Originally, the database followed the fields proposed by the International Transfer Format for Botanic Garden Records20. This has proven to be important when migrating data to new IT systems and exchanging data with partners. Some of the fields have, over time, become more atomised but the information distributed across them is largely unchanged.
For nomenclatural terms and abbreviations and of course for nomenclatural practice in general, we follow the ICN5
Most of the other standards used to standardize data in the published WCVP dataset are recognised by Biodiversity Information Standards (www.tdwg.org):
For the authors of plant names, we use Authors of Plant Names21 now maintained by IPNI. This standard is widely used and obligatory in many scientific journals.
For journals, the second edition of Botanico-Periodicum-Huntianum (BPH-2) is used22.
For books published until 1945, the second edition of Taxonomic Literature (TL-2)23 is used.
For publications not in TL-2 and for books published after 1945, we follow the standard forms from the IPNI Publication Database which is continuously maintained.
For the additional data in WCVP, not included in the published dataset, the following standards are applied:
For the geographical data we use World Geographical Scheme for Recording Plant Distributions3 with some minor changes for countries that have recently changed name, e.g. Swaziland for which we now use Eswatini.
For the life form data, we follow the system originally proposed by Raunkiær19
Climate zones: Alpine & Arctic, Temperate, Subtropical, Desert, Seasonally Dry Tropical and Wet Tropical used as consistent terminology to summarize the published habitat information from the resources used to construct each species concept.
Taxon acceptance and species concepts
The basic rule of species acceptance in WCVP is very simple; we follow the latest published species concept unless experts advise us otherwise. Of course, anyone familiar with plant taxonomy will immediately realise that taxon acceptance is rarely that straightforward. It is however very important to make a distinction between acceptance in the different taxonomic ranks represented in WCVP (Family, Genus, Species, Infraspecifics). WCVP is primarily a list of species concepts. Taxa at other ranks are not the primary focus, not least because there will always be alternative classifications for stable species concepts. However, since full synonymy is provided, users can easily find the correct name if they prefer to use different generic or infraspecific concepts.
Although there is a pervasive impression that taxonomy is ever-changing and that alternative taxonomies are commonplace, this not our overall experience24. This perception may have some truth at generic level but from our experience there are very few current alternative species concepts supported by multiple scientists. Even at generic level alternative taxonomies are perhaps less problematic than is generally perceived, as shown for example by Vorontsova & Simon who suggest that up to 90% of names will remain unchanged when implementing a monophyletic classification for grasses25. Overall, there is striking consensus at species level, especially as for some groups there are very few if any active taxonomists. Internet searches may sometimes give the impression that multiple species concepts are accepted at the same time, but of course this is merely because older data are neither removed nor updated. It is therefore very important when using online resources to check the date on which a species concept was last updated or which published taxonomy is followed, because even a suppressed name such as Solanum ferox L. can still be found as seemingly accepted online.
Species acceptance in WCVP should be seen as a process rather than a one-off decision to which we adhere no matter what. As explained above under workflows, different publications are used to add the geography and create the species concept and they may not be screened in chronological order. In principle, during compilation we follow the latest published taxonomy and prioritise global accounts over local ones. These two principles are generally sufficient to provide species concepts for the vast majority of names. For the minority of cases, for which no recent taxonomic treatment exists and different current Floras adopt apparently different species concepts, then the situation is examined more closely: we try to find published peer-reviewed papers that include a phylogenetic treatment of the taxon, even if the paper lacks a formal taxonomic component, or we contact experts in the group to request resolution. Where uncertainty remains, then we generally default to retaining the existing taxon concepts rather than merging them without sufficient scientific evidence. All the initial species concepts adopted during collation then undergo the expert review process which will confirm or refine them.
For flowering plant families we follow APG IV1 and for conifers and ferns we follow Plants of the World2 including some recently published minor changes and additions26, for example. For genera we primarily follow global classifications where published (e.g. Legumes of the World27 and updates for the genera of Fabaceae, then partial generic classifications if such exist and Plants of the World2 for genera of which no recent published classification exists.) The generic classifications are also fine-tuned during the review process which is led by specialists in the relevant groups who may have more current, sometimes unpublished data to hand. Infraspecific taxa are accepted in a similar way as species concepts, they do however have the additional complication that for a large part of botanical history, most cultivars were given scientific names. As WCVP only records naturally evolved taxa, names applying to these mutations or human selections are synonymised under the species to which these mutations or cultivars belong. The epithets may be available under the International Code of Nomenclature for Cultivated Plants28, and appropriate cultivar names should be used as set out under that code.
Alternative taxonomies
Botanists, in particular, ask the question if WCVP shows alternative taxonomies. Although this is perceived as being a major issue, we have never found this an issue in the review process or in general use. First, we should emphasize that WCVP is primarily a list of published species concepts and that currently most disagreements are about genera (See also Taxon Acceptance and Species Concepts above). WCVP lists all synonyms and therefore users are, of course, free to use a name in a different genus for the WCVP species concept. For genera we normally follow a published account that involved most of the experts of that group. For example, WCVP follows Genera Orchidacearum29 and subsequent volumes for the generic concepts in the family Orchidaceae with minor changes being made subsequently through discussions and feedback from the authors. The main advantage of following a particular account is that the generic circumscriptions are consistent and based on shared scientific evidence.
WCVP reflects alternative taxonomies in the references cited for each record, which are available through the links on the WCVP website to POWO. It became possible from 2003 onwards to add references for each name and each geographical record. Currently a total of 9,145 publications have been used and cited. When taxonomic changes are made to WCVP, a reference is added so users can see the publications or communications on which this change was based. It is important to make clear that (i) such references are only added to names or synonyms explicitly cited in the publication added and (ii) that the protologue (the work in which the name was originally published) is also a reference and this is included for each name. As a result, for some taxonomic decisions, the reference to the taxonomic work which provides the evidence for the decision may not appear in the record of each name affected by that decision, but only in a linked name record.
Although, over time, many species concepts have changed, in the here and now there are few competing species concepts where there is genuine disagreement with scientific evidence. While it may still be desirable to show current alternative taxonomies, we consider citing references to the competing view as the most objective and practical way to do this.
International collaboration
As noted above, WCVP relies on collaborators around the world. 155 reviewers from 22 countries have been directly involved in expert review of the data for completed families and many others are currently reviewing data. WCVP also has a close relationship with several monographic resources in addition to the family level checklists mentioned above, including Grassbase (www.kew.org/data/grasses-db/index.htm), The Zingiberaceae Resource Centre (https://padme.rbge.org.uk/ZRC/), Cate Araceae (http://cate-araceae.myspecies.info/) and Palmweb (www.palmweb.org), and the Leguminosae30. WCVP also collaborates with floristic initiatives such as the Catálogo de plantas e fungos do Brasil8, Euro + Med Plantbase (http://ww2.bgbm.org/EuroPlusMed/), and World Flora Online13. Collaboration with horticultural data providers is strong too, including the International Daffodil Register (https://apps.rhs.org.uk/horticulturaldatabase/daffodilregister/daffsearch.asp) and the Classified List and International Orchid Register (https://apps.rhs.org.uk/horticulturaldatabase/orchidregister/orchidregister.asp).
WCVP has contributed data to the Catalogue of Life (CoL) and now provides 35% of vascular plant CoL content31. With increasing collaboration between CoL and GBIF in the CoL+ project6 and support of the World Flora on-line community7, CoL+ is likely to become the central hub for access to community-supported consensus taxonomic species lists covering all life. WCVP will provide its data through these initiatives, and will both work with TENs and provide taxon concept data for taxa not covered by any TEN. WCVP is already a baseline resource for TENs for certain plant groups (e.g. palms, legumes) and a source of update information for other TENs. In the case of the palm family, the WFO TEN has been closely involved since the compilation phase of WCVP and WCVP contributes the palm taxonomic data to WFO. The legume community is actively editing and commenting on current WCVP content. For other families e.g. Zingiberaceae, the TEN and WCVP run in parallel and data is frequently exchanged between the TEN and the WCVP editor. Thus the nature of the relationships vary, and in many cases they are still evolving, but clearly have the potential to be mutually beneficial and synergistic, with feedback from TENs helping to update WCVP records. WCVP downloads and website can assist any TEN in the task of routine curation and monitoring the addition of new names. WCVP welcomes collaboration with any TEN. It is envisaged that, eventually, TENs will cover all vascular plant groups and consensus content will flow from TENs through WFO to GBIF and CoL+. However, at the moment only 25% of vascular plant species are covered by the 29 TENs. Hence, the WCVP is a vital resource for updating and supporting the developing TENs network to achieve their vision.
Principles for creating a single authoritative list of the world’s species
A recent paper presented ten principles that can underpin a governance framework for species lists32. Although the origins of WCVP predate this publication by decades, these principles have also underpinned the creation and governance of WCVP. We present a summary in Table 2.
Source: Ecology - nature.com
