25uF, 250v AC, connected across the mains input. Your capacitor looks it may be original to the lantern, the ballast looks newer. There should be a date code on the ballast - either 2 or 4 characters possibly starting with LH XX if 4 characters are used. The capacitor has "7091" on it, ususally it is year and week - that won't work - so perhaps its is week 09 of 1971. If it is, that again seems a bit too new for the lantern which I would have thought is no newer than mid 1960s.
I know it's essential as it does say so on the ballast.
The area which says the voltage has worn off the capacitor in my MA60. Took a while to find out that it was 20µf.
Incidentally my GR150 135w gear came with a 250v 6µf capacitor. That one is wired in series on the ''neutral'' lamp supply between the ballast and the ignitor. I have messaged Urbis Saturn Land not to use the gear till I get a supply of 400v 8µf capacitors in.
The one in my MA50 has failed. That one is rated at 400v. I admit a 250v GEC 8µf capacitor from the York haul works fine in it and i've had it running with that in for a while. It may only exceed 250v at startup as I doubt a 135w SOX lamp needs more than 240v when running normally.
They don't but a 180w lamp does indeed run higher than the supply:
135w: 0.9 A 185 V 20,200 lm 134.7 lm/W
180w: 0.9 A 265 V 30,000 lm 150.0 lm/W
For safety though i'll use a 400v capacitor in the MA50. The MA60 works perfectly well but I could do with a spare for that too as it's PF is in the 0.69-0.74 range which may indicate a failing capacitor.
A word of warning. It is not the running voltage of the lamp which is the issue. For 135/180W SOX with a series connected capacitor, the ballast and the capacitor form a resonant circuit to strike the lamp. This is very much higher than the running voltage. The voltage across the lamp terminals will be around 1KV to get it to strike. The voltage rating of the capacitor shows it has around 400V across it, the ballast will have a similar voltage across it. These voltages are out of phase and the therefore the resultant voltage vector (across the lamp) will be greater than the sum of the individual ballast and capacitor voltages.
Not only is the start up voltage likely to exceed the voltage rating of the capacitor, but also the running voltage will be. At very least when it eventually fails, it may just got short circuit. At worst it will explode. You have been warned.