Reading a paper with photoelectric UBV sequences led me to this citation:
https://ui.adsabs.harvard.edu/abs/1964AJ.....69..251M/abstract
...which seems to have been done by a teacher at a prep school in Connecticut. He was obviously a careful worker, finding strong temperature effects in the photomultiplier tube set-up he worked with (an expected problem, as he describes). Most revealing is Figure 3, showing residuals in the photometry from stars measured on two or more nights as function of spectral type (i.e. star-color). Look at the vertical scale: the scatter can't be any worse than 0.005 mag rms, darn good for working from a poor photometric site and variably unstable electronics from 60 years ago --- OK, bright stars, but everything else tending to increase the scatter. I'll bet the referee thought it was a wonderful heuristic exercise. It (again) puts the lie to millimag photometry being impossible. Observers nowadays might do well to try to reproduce these results with new hardware.
\Brian
Thanks Brian. I'll read every word and even between the lines to see if J.R. McCullough can teach me better technique. Got a bug in my bonnet to do more consistent photometry.
Ray