Create a taxon swap draft (1 input, 1 output), then:
only one input allowed for a swap”Just curious if you have experimented with trying to save it between those two steps? Maybe it doesn’t “know” it has been changed to a merge until it is saved that way?
I haven’t, but one also shouldn’t need to do that. 
I’m having the same issue, but going the other way. I must have multiple input taxa for my “merge” which is supposed to be a swap now.
This is the first time I’ve used the taxon drop feature, but this seems like a bug to me. Normally, curators can add a taxon change by selecting Taxon Changes on a taxon page and then selecting New taxon change. This autopopulates the input taxon with the taxon from the page you just navigated from. It also defaults to Swap, but other options can be selected. If you select Drop, the input field is cleared, but it is still there if you go back to swap. If you reenter the taxon and try to save, you get this error:
This does not occur if you start from scratch with a blank taxon change (from the taxon changes page).
I’m guessing this happens because the autopopulation confuses the system when the taxon change becomes a drop. There shouldn’t be an output, but because the default is a swap, there was an output to begin with (even though it was blank). Somehow that gets carried over to the drop and, of course, the input and output are the same because there is only one taxon in a drop.
May be related to:
https://forum.inaturalist.org/t/cannot-switch-from-taxon-swap-to-a-different-type-of-taxon-change/5407
https://forum.inaturalist.org/t/unexpected-behavior-when-creating-taxon-change/10609
I suspect they stem from the same issue, so I merged it here and made the topic title a bit broader.
