How are missing/ambiguous data handled?

[wzbillings]

In the documentation for dist.hamming(), it says "The default is handle missing data as in pml". However, the pml function documentation doesn't appear to explain the default way that missing data are handled. I took a look at the code in phylo.R but don't understand it enough to answer my own question.

I see that the default argument is exclude = "none", and when I use phangorn::dist.hamming(), and I get different answers than I do from a naive hamming distance method (e.g. stringdist::stringdist(..., method = "hamming"), which implies that the default is to handle ambiguous states and gaps in some way.

If you have time, I would appreciate a brief explanation of the default way that missing data are handled. In particular, is there a difference in the way ambiguous residues and gaps are handled?