Zompist Bboard Again

I've been reading up on changes to ejective consonants across the world's languages, and I think it's quite interesting how stable they seem to be - in Semitic the original ejective series have become pharyngealised (also probably in Berber?), and in other Afroasiatic languages there appears to a change of at least some consonants to implosives (though I'm not sure how good a model Semitic is for 'regular' sound changes). I am also aware in the Nakh subgroup of Northeast Caucasian, Chechen and Ingush lenite stem-internal ejectives to voiced plosives, but I've been struggling to find parallels in other families.

Which led me to Athabaskan (and Na-Dene more generally, though I've not found a convincing PND reconstruction as of yet) - in this family, there is a very well attested three-way contrast between tenuis, aspirated and ejective. From what I know of the family (which is admittedly rather superficial), I don't know of any individual language that deviates from this three way contrast - some have shifted the places of articulation around, but all ND languages that I'm aware of have preserved tenuis, aspirated and ejective/glottalised rather perfectly (I believe some may have introduced voiced stops via loanwords, but that's outside the scope of my question). I was curious if there are any counterexamples to this apparently well preserved trait.

(Also fwiw, I'm not interested in a glottalic theory debate for PIE)

I'm not 100% sure but I don't think there are any languages that have lost the three-way constrast. Ejectives seem to be quite stable in other Pacific Northwest Coast language families, such as Wakashan and Salishan. For instance, Proto-Wakashan is reconstructed with a voiced-voiceless-ejective distinction, with the voiced-voiceless series merging in Southern Wakashan, while the ejectives are preserved unchanged in the Northern and Southern branches.

I dunno about Athabaskan and Na-Dene, but something very comparable happens with Algic and Algonquian: Wiyot and Yurok each preserve two out of three stop series (plain, aspirated, ejective), but Algonquian itself merges them all into one stop series.

You might also look at languages spoken around the Great Plains, eg. Crow in Siouan, because as far as I can recall, there are multiple language families with multiple consonant series that exhibit interesting sound changes and behaviour around that particular geographical area.

I have nothing to say about the topic, but the page you link to is interesting. For real! Thank you!