On pg. 427 in note b to the table listing the Romance reflexes of the paradigm of sum I write that Old Portuguese has som or sõ. But this is not quite right. First, my usage is to refer to the language of the oldest texts not as Old Portuguese but as Galego-Portuguese and second, my friend and esteemed teacher Rip Cohen provides the following correction:
In thirteenth century manuscripts of Galego-Portuguese lyric (compositions spanning the period c. 1220 to c. 1300), the regular form of the first person persent indicative of the verb to be (seer < sedēre) is sõo < *sono (cf. It. sono). The nasalization represented by the til on the õ is a vestige of the final -m of sum. The form soon also appears, but less often, in those manuscripts, and it may represent a different pronunciation, but at any rate it is still bisyllabic. The forms son and sõ, which appear in early sixteenth century copies of thirteenth and fourteenth century poetry must nearly always be corrected to sõo or soon on metrical grounds. And we sometimes find soo where the nasal must be added: sõo. To my knowledge, the form sou does not appear until well after 1350.
Thank you, Rip!