o/u

[Davius Sanctex]

[Frances] A number of people have compiled extensive lists of attested u for
/o:, o/ and found no predictable pattern.  It doesn't correlate with length
or with stress.
------
[Davius] What about a statistical pattern, it seems that spelling < u > is
more frequent following labial or labio-velar consonants /p, kW/. Another
question, what about the alternances /i, e/ in Nahuatl loanwords to Spanish?
Is this a problem of the same type?