25.02.2016 14:49

5* publishing in taxonomy: Formicine phylogeny by Ward et al., 2016


Ward et al.’s recent publication on formicine taxonomy is implementing the results from their phylogenetic analyses in taxonomy, that is listing all the new combinations of names, synonymies or resurrection of old names. These results are based on very advanced DNA technologies and large number of base pairs, and the assumption that more bp provide more robust results makes them confident that taxonomic decisions need be made.

How close does this publication get to Tim Berners-Lee 5*, the analogous standard to phylogenetic studies in publishing? How does this publication live up to advanced taxonomic publishing implemented at the Biodiversity Data Journal?

It is an open access article with a DOI and a Zoobank URN, that is registered at Zoobank. The article is available in a single format as PDF, which is 1* in the above scale.

Bibliographic references include for recent articles CrossRef DOIs, but they do not include any DOIs for older articles that are available (eg using Refindit to find them).

Citation is mixed. Recent articles are cited. However, all the old taxonomic publications cited are not listed in the bibliographic references. With that a link to the legacy publication is cut, and in some cases, only the author, not even the publication year is provided. This is exactly what renders our huge corpus of biodiversity literature useless - not even the taxonomists themselves care about it.

Names are not linked to external resources such as Zoobank, or Hymenoptera Name Server for ants, that provide unique identifiers and resolution services for names. There is a reference to antcat where catalogue information is available.

Taxonomic treatments are not delimited. They cannot be cited, nor are the treatment citations linked to the respective treatments.

Cited type specimens are not linked to online digital copies, even though antweb exists as one of taxonomies most formidbable source of digital images of specimens. Though this is not an issue here, it might be argued that revisions like this should look at type specimens, and to make during this process digital copies of them accessible.

The article is completely unstructured and really made for human consumption.

What makes this article different from a printed pre-cyberspace article is that it is a PDF that can easily be distributed online. Otherwise, it needs a human to understand and additional resources to find the full references for the micro-citation, and from there more time to find the reference online. It also needs additional work to enter all the nomenclatorial changes in the respective databases. 

Though this kind of publication is sadly the most widespread anachronistic way of communicating taxonomic results, here is its converted version. It has all the names, treatments, treatment citations, and bibliographic references tagged. If possible, treatment citations are linked to existing treatments.

 The missing citation of the original taxonomic treatment of the taxonomic name in this graph is due to very implicit style of publishing nomenclatorial acts.

The names and new combinations thereof are imported by GBIF, and the treatments will be linked to from NCBI, Wikidata or HOL, as explained in an earlier news item.

 If we are concerned as much about taxonomy as on phylogenetic analyses, taxonomists should take care to avoid publish 1* and rather consider the goal to produce 5* data, to avoid others to reconstruct data they extracted from databases. Creating silos is a disservice to science.