In my opinion, the significance of Ada- and for that matter things like OS/2 and mainframe OSes- has to be weighted by its overall significance in industry.
and
By analogy, AutoCAD and CATIA don't get many column-inches in the sort of things that most of us read, and they have minimal presence on Sourceforge and Github. But from the POV of industry as a whole, they're /very/ significant.
If you ask, indeed domain specific languages are often excluded from the stats as less substitutable. But even with general purpose languages, in my experience people talk (and probably respond in questionaries) more what they'd like or
plan to use than what they are
actually using, specially when the latter is considered dull.
Contrary to the eighties where my view was somewhat limited, I do have some recollections from the late nineties client-server era when people talked about hardcore C++, while they were crafting apps in Delphi (or, more likely, VB). Also during the dot com boom, people were talking about all kinds of fancy new language and framework , but if you really asked what they used in production, it was some cobbled together templating system. Quite often homemade.
I sometimes heard echoes of that era in the javascript world, specially say 5-10 years ago.
C was pretty much dominant by '87. Grey Matter easily had a dozen affordable compilers in its catalogue, and library bundles were typically sold tailored for a specific compiler and (in the case of x86) memory model.
But the late eigthies and early nineties were the dos heyday, and I doubt the bulk of the dos era software was in plain C? But I spent that era mostly in an non IT academic bubble in the NL, so that might have skewed the picture.