I had a long reply with explanation almost ready to post and power went out. I'll just do short version in case it goes out again.
With green, red and yellow wires disconnected from line and handset on-hook, measure resistance of green to yellow. If resistance starts relatively low and rapidly climps to open; then connect yellow wire to green and phone should ring. If there is continuity after a couple seconds of the meter being connected, the phone was setup for a party line; connect a capacitor to yellow, other lead of cap to green. Cap should be 250V; mylar, paper or poly. Value probably between 0.1uF to 1uF. I can't tell you the value, but it needs to make a 20Hz resonant circuit with the ringer coil.
For example to find value experimentaly try 0.22uF, then 0.47uF, then 0.22uF parallel with 0.47uF, etc.
In the photo, it looks like a later model 500 (like C or D), so ringer equivelence might be okay. (model number on the bottom)
Bye, (before power goes again)
(while I was waiting for reboot, typing again and getting coffee, rshayes posted; so it looks like I duplicated info. At one time "G" may have meant "ground", but electrical use increased (like 1930s) the hum would have become too great. Correct about party lines and yellow. The internal cap at "A" and "K" wouldn't work (alone) for the different party line ringing frequencies.)
(power still holding, so I went and found the schmatic I had link for the first time around -
http://www.bellsystemmemorial.com/image ... ematic.gif
is for 500C and 500D models.
If any one is curious, the model number for a standard Ma Bell DTMF phone is 2500.)