Taxonomy

A specimen in Arctos can be identified to a name, or to a taxonomic concept. The vast majority of identifications in Arctos use names.

Names

Names (scientific_name) are just strings, of the form Genus or Genus species or Genus species inf. infra. In other databases these are known as ‘canonical names’. They have no authors associated with them. Creating them is easy, via the GUI or via a bulk loader.

Note that the taxon_name table uses scientific_name as its primary key.

Taxonomy search options in Arctos

Taxonomic search can be performed:

  1. via the Record Search GUI (at https://arctos.database.museum/search.cfm),
  2. via ‘Find Records in Arctos with…’ buttons on the Name page (https://arctos.database.museum/name/),
  3. via writing a Query string for https://arctos.database.museum/search.cfm?, or
  4. (via the SQL interface)

The behavior of options 1-3 (above) is documented below, ordered from more specific searching to less specific:

Entering names into Arctos

The name associated with an identification should be the name on the label, not the synonym of the name (see this issue).

Classifications

For each name there may be several classifications that match the scientific_name; in this context ‘classification’ means a set of non-hierarchical terms (e.g., display_name or author_text) and hierarchical classes (e.g., order, family) for each name. Each classification exists in a ‘source (e.g., ‘Arctos Plants’, ‘WoRMS’). A collection (e.g., UAM:Herb`) ranks the sources that should be used for their classification (in ‘Manage Collection’). At ALA we use three: ‘Arctos Plants’ > ‘WoRMS’ > ‘Arctos’ (a legacy source). If there is no classification for a name in the first source, then the second is searched, and so on. The risk of this fallback is that a name in a secondary source will be a homonym (same name, different meaning), and the family, order and even kingdom may appear incorrect for the specimen. Eventually it will be better to have all names in ‘Arctos Plants’ and deactivate the fallbacks.

Currently (2024-02-29) there are 804 names used in ALA identifications that do not have a classification in these three sources (see SQL, and this issue).

It is not enforced in the database that there is only one classification per source per name, and if there are more, the hierarchical terms get merged, possibly even for homonyms (see this issue). So it is a good idea to make sure that the name usages for ALA have a single classification per source per name (on 2024-02-29 this is true, SQL).

A proposal for a new Classification source

[in progress]

The problem with existing classifications:

A problem with the current taxonomy model in Arctos:

A proposed new classification (‘Alaska plants’) will:

How specimens are searched for:

The code is at https://github.com/ArctosDB/arctos-dev/blob/main/3.2.9/search.cfm

From the taxonomy search page, these are the inputs:

Best practice: Add verbatim identification to get the Author string.

Complete transcription: https://arctos.database.museum/guid/UAM:Herb:108266

Update: https://handbook.arctosdb.org/how_to/How_to_Search_for_Specimens_with_Identification_and_Taxonomy.html

Taxonomic concepts

(to follow)