Apparently WolframAlpha thinks that (2 degC) / (1 degC) = 2. For example, an ideal-gas has a volume-change of V_after / V_before = T_after / T_before (where V is volume and T is temperature), so WolframAlpha would say that a temperature-increase from 1 degC to 2 degC would double volume. Or that an ideal-gas going to -1 degC would occupy negative-volume.

@MichaelE2 thanks for the help with that NMaximize bug -- nothing like digging through your own code for bugs for hours, only to discover it isn't yours!

I suppose I should rerun my package's test suite whenever I upgrade MMA

You're welcome. :)