How did Mapos Buang develop a uvular series?

[Ketsuban]

Mapos Buang seems to be the only language in the world with a prenasalised uvular stop. As best I can tell it's just a statistical accident that prenasalised stops and uvular series happen not to coincide areally, but because Oceanic languages tend not to have a uvular series I'm curious to know how Mapos Buang got one. Wikipedia has a phonemic inventory for Proto-Huon Gulf, but the citation is a book I would have to spend money on and given its scope is the entire Oceanic language family it probably isn't going to go into the minute depth needed to answer my question anyway (especially since it seems like the bulk of field work on Mapos Buang happened after the book was published - Bruce Hooley's dictionary is dated 2006).

[bradrn]

[…] but because Oceanic languages tend not to have a uvular series I'm curious to know how Mapos Buang got one.

Proto-Oceanic had uvular *q, which seems like a sufficiently likely source to me. Most of the Oceanic languages ended up getting rid of it, but clearly Mapos Buang didn’t. (I think quite a few of the more basal Oceanic languages of New Guinea still have uvulars.)

EDIT: Consulting Blust’s The Austronesian Languages, apparently this guess isn’t entirely correct. It appears that no Oceanic languages at all have retained *q as /q/. Still, the protophoneme appears to have been present, and *q→*ʁ→/ⁿɢ/ is plausible enough.

[Tropylium]

I know nothing about Mapos Buang, but there is an /ⁿɢ/ (and /ⁿqʰ/) also in Hmong.

[Nortaneous]

Ross's book is online here. His *q is intended as glottal, but can be reflected as velar or uvular in Huon Gulf languages (if I'm reading the chart correctly, which I doubt). POc *q *ʀ correspond to PAN *q *ʀ, which may have had different values.

Mapos Buang seems to be the only language in the world with a prenasalised uvular stop.

Japhug and Caodeng Rgyalrong, Bai, and Serer, in addition to Hmong. Probably elsewhere in Hmong-Mien, Sino-Tibetan, maybe Kra-Dai. I'd be surprised if no Tibetic variety has /ɴɢ/.

[Man in Space]

Mapos Buang seems to be the only language in the world with a prenasalised uvular stop.

Japhug and Caodeng Rgyalrong, Bai, and Serer, in addition to Hmong. Probably elsewhere in Hmong-Mien, Sino-Tibetan, maybe Kra-Dai. I'd be surprised if no Tibetic variety has /ɴɢ/.

Taa/!Xóõ seems to have a voiced one (or at least the voiced uvular stop is default prenasalized) if I’m reading Wikipedia correctly.

[Darren]

There's *r → nd in Mapos Buang's neighbours (pretty much the whole Markham branch of Huon Gulf). So I guess *ʀ → ᶰɢ makes sense in context.