Actually, looking at my (pretty blurry) copy of the schematic, I see that C38 doesn’t appear to be part of the PSU supply chain, but is simply a smoothing capacitor for the two screen grids in the RF and converter stages. Nevertheless the voltages you quoted are definitely wrong, especially the AC measurements.
But if you look at the schematic you’ll see that the chassis is not connected to DC ground – the DC ground , which you need to connect the -ve meter probe to, is indicated by a “comb” icon, and the chassis is indicated by an upside-down Xmas tree icon. The most easily identified point in the DC ground line is probably the battery negative terminal, which should have a yellow wire. At least, it was yellow in 1948! 🙂
If you make these voltage measurements again, using that point as the negative, you’ll probably get much more sensible values.