Figure 1. Tulerpeton parts from Lebedev and Coates 1995 here colorized and newly reconstructed. Note the well-preserved scales, studied in detail by Mondéjar-Fernandez et al. 2014.
The scales of Tulerpeton
were made of compact bone lacking enamel, dentine and isopedine layers found in more basal sarcopterygians (Mondéjar-Fernandez et al. 2014). This was the first step in a process that ultimately also removed bone from scales leaving only keratin scales arising from the epidermis.
Several times later
bone reappeared in the dermis producing osteoderms.
Earlier the Devonian tetrapod Tulerpeton heretically nested as the oldest known of the basalmost amniotes, the last common ancestors of all living reptiles, birds and mammals.
References
Mondéjar-Fernandez J, Clément G and Sanchez S 2014. New insights into the scales of the Devonian tetrapods Tulerpeton curtum Lebedeve, 1984. Journal of Vertebrate Paleontology 34:1454-1459.
The Pterosaur Heresies 
Do you think epidermal scales are homologous with dermal scales at all?
Now that’s funny. Have I been banned here? :-D
Sorry, just a glitch.
Epidermal scales seem to come and go, and some scales appear to be derived from feathers, so let’s not generalize here, but stay with the data and stay specific.
Oh, so you think different epidermal scales aren’t homologous with each other… fair enough. Do you think any of them are homologous with dermal scales? If so, I’d be genuinely interested in why you think so; this is not a rhetorical question.
I have not examined the question and it does not appear in the current matrix. The process of checking that out would be to see if any taxa or clades that have the one trait nest with others that have the second trait without a demonstrable gap. Unfortunately, scales don’t always fossilize, so the gap may be real or not real and thus the question may stay up there like a spinning plate for some time.