In the original wiring of this Mitchell halfpiper, there was half of a paper capacitor connected to the neutral side.  It looks like at one time, it may have been connected across the mains for power factor correction.  Unfortunately I couldn't get the values off it.  I was wondering what the appropriate values for a capacitor to correct for power factor on this particular setup.

There are two choke ballasts running a pair of F20T12 lamps.  Supply is 120 volts 60 Hz.  The ballasts draw .35 amps each (.7 amps total) and have a power factor of about .52.

Iballast*PF/(Freq*2*Pi*V)=4uF per ballast, so 8uF total.
It should be rated for direct mains connection for PFC use (see UL listing) and should have discharge resistor.

Thanks for the info! 

I don't have any great need to improve PF on this particular fixture, but it's the only reason I can think of to have had that cap in there in the first place.

I did a quick search, and the only 8uF lighting caps I could find were the large oil filled type.  I did find compact 8uF film caps, but they are the type used in motor start circuits, like in ceiling fan motors.

You could use any close value, from 6-10 uF will work.

I might give it a try with a 10uF before I buy anything special for this.  I do have both oil and film PFC caps from HID ballast kits.

Could the capacitor be connected on the neutral in series with ONE of the lamps ? If yes then it might be a phase shifting capacitor to reduce overall flicker from the fixture

Could the capacitor be connected on the neutral in series with ONE of the lamps ? If yes then it might be a phase shifting capacitor to reduce overall flicker from the fixture

Yes, but then it's value should be different:
C=Ilamp/(Vmains*sqrt(1-PF^2)*4*Pi*Freq) = 5.4uF
Here the value is quite sensitive (check the current, calculation above neglect saturation of the ballast core, but would be good starting point) and should be rated for at least 2x the mains voltage, so for 120V it should be >280VAC.