Heads up: When you commit this change, identifications will be re-assigned to the output taxa based on their atlases. If some or all of the outputs lack atlases, or if atlases overlap for a given record, identifications will be replaced with identifications of the nearest common ancestor taxon, Coliadinae

Taxon split aaaaaa 48px Taxonomic Split 92975 (Draft)

Reverted - Please do not commit!! (see discussion below)
unknown
Added by aleturkmen on June 06, 2021 07:34
split into

Comments

Thumb

@loarie please look into this taxon change - it seems Eurema old has been split into Abaeis and the old Eurema. It might have gone wrong....

Posted by jakob about 2 months ago (Flag)
Thumb

I'm not sure of the cause, but a small number of African observations were incorrectly changed to Abaeis along with this change. (it looks like some Asian observations were affected as well)

Posted by cabintom about 2 months ago (Flag)
Thumb

There are also African observations that went from Eurema to Coliadinae that should have remained as Eurema.

Posted by cabintom about 2 months ago (Flag)
Thumb

happy to take a look - can you add links with some of these specific issues?

Posted by loarie about 2 months ago (Flag)
Thumb

I'm confused by how this taxon change was set up - would it not require Eurema (old) split into Eurema (new) and Abaeis?

Posted by jakob about 2 months ago (Flag)
Thumb
Posted by cabintom about 2 months ago (Flag)
Thumb

Its ok, and usually preferable, to use a split input as an output - read https://www.inaturalist.org/blog/40417-using-a-taxon-split-input-as-an-output
but I'm not totally sure what the point would be / how it would be have if only 1 of the two outputs was atlased as in this scenario

What I think this split would do is replace all ID's of Eurema and replace them with IDs of the common ancestor of Eurema and Abaeis. If thats not what's happening, definitely a bug we need to investigate

Whether or not this split was a good idea is a different question. Given that its not using atlases to replace ID's of Eurema with Abaeis, it definitely has large costs of replacing Id's with Eurema with the common ancestor of Eurema and Abaeis. To get a sense for whether the costs of existing ID's of Eurema that imply too broad a taxon hanging around warrants a split, I would have checked to better understand how many existing IDs of Eurema we're talking about and whether they could be manually addressed. I'm not sure if that was done

Posted by loarie about 2 months ago (Flag)
Thumb

Obviously I'm struggling to keep up with iNat's development!

Posted by jakob about 2 months ago (Flag)
Thumb

This seems like a messy outcome, even though it's not typically "approved" to do so, I'd have much rather made the new genus heading and just manually shuffled all the relevant species independently. That's if atlases were not possible to allow the system to continue applying genus IDs rather than subfamily ones.

Posted by silversea_starsong about 2 months ago (Flag)
Thumb

@silversea_starsong, I would have
1) created new Abaeis species
2) swapped the relevant Eurema species into the new Abaeis species
3) probably stopped there, and let the community manually replace/outvote existing Eurema IDs that meant sensu lato

can you explain what you mean by "made the new genus heading and just manually shuffled all the relevant species independently"? Why would you even bother making a new genus?

Posted by loarie about 2 months ago (Flag)
Thumb

@loarie
You just explained what I meant by that. I would have:

1) created Abaeis if it had not already existed
2) swapped each relevant Eurema independently (or all at once, can iNat do that these days?) into Abaeis
3) as you described for step 3

Posted by silversea_starsong about 2 months ago (Flag)
Thumb

@loarie The genus Abaeis is restricted to the Americas. Large (to the order of hundreds at least) numbers of Eurema from Asia and Africa have now been converted to Coliadinae. Since the genus Abaeis already existed (example: Abaeis nicippe), it is not clear what has been achieved by this change. Can it be undone?

Posted by salmanabdulali about 2 months ago (Flag)
Thumb

The added trouble is that, down the road, the African Eurema would likely need to be swapped to Terias.

Posted by cabintom about 2 months ago (Flag)
Thumb

@salmanabdulali, yes if Abaeis had an atlas before committing then Old World IDs of Eurema would have been left in place rather than being replaced with IDs of Coliadinae.

@aleturkmen do the issues associated with this swap make sense to you? Committing big splits like this with thousands of IDs is very distructive and alot can go wrong so its best to mention some other curators before committing to make sure things are looking good. I'm always happy to take a look.

I'll go ahead and revert this now.

Posted by loarie about 2 months ago (Flag)
Thumb

ok this is reverted

Posted by loarie about 2 months ago (Flag)
Thumb

Thanks.

Posted by salmanabdulali about 2 months ago (Flag)

Add a Comment

Sign In or Sign Up to add comments