I was taking a look at the data in the Backbone dataset. Looking at Chamaecyparis obtusa, specifically, I noticed something different between this set and Catalogue of Life. In CoL, almost all infraspecific entries were classed as synonyms of the root species Chamaecyparis obtusa.
i.e. from their textree format:
Chamaecyparis obtusa (Siebold & Zucc.) Endl. [species] {ID=TJJH}
=Chamaecyparis acuta Beissn. [species] {ID=C6N35}
=Chamaecyparis andelyensis Gordon [species] {ID=C6N37}
=Chamaecyparis breviramea Maxim. [species] {ID=TJJ8}
=Chamaecyparis keteleri Standish ex. Parl. [species] {ID=C6N3L}
etc.
However, within GBIF, almost all of the Chamaecyparis obtusa infraspecific entries are said to be synonyms of the autonymic variety Chamaecyparis obtusa var. obtusa. This is fine, and even arguably preferable, but I haven’t found it to be consistent.
I checked an alternative to see if this was how things were classified in GBIF, with Cedrus deodara. With Cedrus deodara, there isn’t an autonymic variety, but instead an autonymic form (Cedrud deodara f. deodara), and most infraspecific entries of Cedrus deodara are not said to be synonymous with the autonymic form, but rather the root species.
Is there rhyme or reason to these decisions, or just varies between species?