[Arran]

Kirk,

I don't know how to sort this bad boy out.  Looking at the Model 979 schematic, there are differences.  On the good side, that schematic shows 2 caps (both 20 uF), but there are tube line-up differences.  The converter is a 6K8 (instead of a 6A7), the IF is a 6K6 (instead of 6D6), the rectifier is a 25Z6 (instead of a 25Z5), and the ballast is a BM43B instead of a BM49C.

According to my tube substitution data, the 6K8 and 6A7 are not subs for one another.  The 6K6 and the 6D6 are similar though not direct substitutes.  The rectifier tubes seem to be direct replacements.
I don't believe that  the ballast tube difference is very significant (maybe someone down the line swapped a '49C for the original '43B).

In any event, if you want to go with that schematic, both caps are shown as 20 uF caps, but I honestly don't think the Model 979 is the right model to use.  Too many differences.

 A 6K8 is a mixer-oscillator/converter tube though, but a triode-hexode rather then a pentagrid type, they could use the 6K8 with some electrical changes, other then the fact the 6K8 uses an octal socket of course. A 6k6 is definitely NOT a substitute for a 6D6, a 6K6 is a power output tube, basically an octal based version of a type #41. I believe you may have been thinking of a 6K7 rather then a 6K6, either that or there was a misprint in your tube substitution manual where they meant to print 6K7 or 6K7G, a type #78 tube is a substitute for a 6D6. A 25Z6 was the octal based successor to a 25Z5, actually it's a more robust tube then a 25Z5, both were intended to be used in a voltage doubler power supply circuit though 99 out of 100 radio models that used them had the two diodes wired in parallel.
Regards
Arran