The valve did open when the motor started, but it STAYED open even after
the motor was up to speed! Yes, the switch was open. WTF??!!

When I drew a schematic to visualize the situation, the problem was
obvious. The start winding was disconnected from the line but still
connected to the solenoid. It was acting like the secondary of a
transformer and generating enough voltage to hold the solenoid open!

If anybody figured this out without having experienced it, I am
impressed! It's something that I never would have anticipated.

Bob

BTW - "fixed" this problem with a relay that separated the winding and
the solenoid, but the motor still didn't start (reliably)! So I
connected it to 220 and it seems fine.