I'm sure you probably won't see this as I'm posting 7 years after first raised this issue, and the "solution" that I'm gonna mention usually gets a lot of flack from people who don't subscribe to this idea but I'll offer it up anyway.
Czerny was most likely using the metronome in a way that is described as the whole beat metronome practice. So basically this means that the beat/pulse of the music (Schlag in German and Tactus in Italian) is actually made up as binary unit containing 2 smaller subdivisions. So in the metronome number of 1/4 note = 138 as per your example, the 1/4 note represents the Schlag (beat/pulse) and the number 138 actually represents the number of binary subdivisions of the Schlag i.e. the 1/8 note. So in modern metronome reading Czerny's metronome mark would read more like 1/4 note = 69 or 1/8 = 138.
But supposing you don't care for that hypothesis of how to read the metronome, the main way in which in the 18th century composers determined tempo before the invention of the metronome was to use the tempo ordinario. Now it's a little complicated because generally speaking the tempo ordinario is 60 crochets per minute. Some authors have it faster than this (e.g. 75) and some say that it is 60 per minute but that applies to a moderato rather than allegro (which would be a little faster). Either way the tempo ordinario is the sort standard middle tempo to based everything around in relation to.
So, more specifically, if you have a 4/4 (or common time) allegro, with 2 harmonic changes per bar, and the smallest structural/fundamental notes is 16th notes (so ignoring ornaments and such), you can apply the tempo ordinario. You can also ajust accordingly depending on the character of the piece (i.e. if it's more solemn and gentle you can slow down to as far as 50 that I've seen on the odd occassion) or even go as high as 84 or 92 if it's very lively. Although that's usually reflected in the tempo word in extreme cases).
Thus your intuition that 1/4 = 55 or 65 is a good tempo is actually pretty bang on.
It's also something that I believe is true of all of czerny's metronomisation for things like mozart, beethoven, his own works, as well as pretty much all 19th century metronome marks, which if they were taken literally are either nonsensical or impossible. (e.g. Czerny's Op.299 School of Virtuosity)
If you do end up reading this, I hope it helps.
Dan