If your case is one of these, take note from this image:
This image is our BEST GUESS, based on messages from members at GeigerCounterEnthusiasts at yahoogroups.com group and physical testing with a working M3FD19 photomultiplier using the test circuit shown above.
The photomultiplier was tested using the diagram shown above, under a 8W fluorescent desktop light at 60cm. Using 105 volts as high voltage, the anode current was just 1 uA, 4 uA with 200 volts and 14 uA with 350 volts. Just for comparing, a BPW34 photodiode on the same position gives 2uA in photovoltaic mode. The photomultiplier seems to response much better at blue light that at red one (as expected). Using a small white led flashlight , it was very easy to get anode currents over 50uA more than one meter away. Maximum voltage applied was 600 volts, and seemed to work well. Because the lack of datasheet, maximum operating voltage is unknown.
The M3FD19 photomultiplier has only 3 stages / dynodes, so gain is small. I think it can be considered as a good blue enhanced device, maybe useful (if high voltage is kept low) for daylight experiments, or even optical communications at night. I doubt it has enough gain to be usable on scintillation applications, except maybe with high output light level plastic scintillators.
It is possible the M3FD19 photomultiplier has the same pinout as M3FS19 and M3FQS19, but it is not confirmed!
My thanks to toto4g for his ideas to get this pinout.