This is a bug fix release.
In version 2.38 I added a feature to allow single filter transforms if you were able to provide the color of the target. For instance, if you had observations only in V but could provide an estimate of the target's BV color then it would be possible to transform that V observation. The BV color could be provided in two fashions:
1 - Include a B observation so that TA could create an estimate of the target's BV color by doing at least one 2 filter, BV, transform.
2 - Provide the color in a comment like so: "#BVCOLOR= x.xxx
The bug is that in case 2, TA would not have aquired the comp star B reference magnitudes that it would need to do the computation. So the result would be based on a spurious value.
If you used this second method of doing single filter transforms, then you should go back an re-examine those results and re-run the processing with this new version of TA.
The error would not be to subtle; you are likely to have been puzzled by unreasonable results. That is how this bug was found: a user brought the puzzle to me.
I will work with HQ to scan the AID for submitted observations that might have used this feature and I will contact the submitters.
Note that you can get the latest version of TA from https://www.aavso.org/transform-applier or by clicking the comment on TA's front panel which will advise you that a new version is available.
Sorry about the problem,
George