The Drouin Institute site is down, and has been almost all weekend, making it very difficult to move forward on the "great aunt/uncle" project (which I'm nearing 10% done).
So, I've been poking around the "lower branches" of the tree, following up on hints, when I came across a PRDH record that seems to connect "dead end" 6th great-grandfather Pierre-Marie Lambert (bef 1714-1789) to his parents (who were already on the tree): Michel Lambert dit Champagne (1680-1733) and Louise Grenier.
Louise was already a known blood relative (7th great-grand aunt) because her brother Joseph was a 7th-great grandfather. But now she's a 7th great-grandmother too!
So this brings in eleven more multi-great-grandparents (which means I have to process them so that I can add the great-aunts/uncles to the current project), and brings the number of identified Canadian direct ancestors to just over 400.
The "evidence" comes from the "Programme de recherche en démographie historique" (generally just PRDH, or The Research Program in Historical Demography), which is a reconstruction of the population of Québec from the Drouin records into a large database; in fact this is the base of the "LaFrance" service (which has extended the PRDH).
If the PRDH record is correct, the one of the few "holes" in the Québec family tree has been plugged.
But I'm still not 100% convinced since I can't (yet) follow HOW this association was made. It SEEMS correct; the only other Lamberts having children around Lotbinière in the early 1700s are pretty much accounted for; any other Pierres seem to have married other people and there's documentation for those marriages. I can't quite figure out where the "before" birth years come from; I've noticed this on other PRDH records compared to Tanguay who just leaves the birth year blank where there's no baptismal record known, or - sometimes - will estimate it from the age given at marriage or death (though I've found that those can be off by 5 years or more!).
Since the PRDH is hosted at the same site as the LaFrance, BOTH are down, and so finding out more about the PRDH will have to wait. If the data are good (or good enough) it would definitely help with other families in the "Lotbinière gap".