Michael and Bob- I thank you guys for your suggestions. I think that I have confused matters by not providing all the data to back up my assertions. In hopes that it will help, here is a bit more.... I have attached a couple more plots. These are instantaneous frequency measurements, made for the same pair of FTS1050A oscillators (including the notable #7), made with the same TSC5110A DMTD instrument, made about a week apart. You can see a number of small frequency jumps in the B plot.The MADEV for the "B" data is about twice that of the "A" data over much of the tau range. I have performed around 15 tests with #7. What I observe is that the behavior varies from test to test. I see this quirky behavior with #7 and with none of the other 16 FTS1050A and two FTS1150A oscillators in my lab. I haven't shown you all of the data so I can understand some skepticism. I hope that this pair of measurements will be more convincing. Same oscillator pair. Same measurement device. One week apart. To be fair, I have in the past had trouble with a TSC5110A DMTD. Its internal 32 MHz oscillator occasionally went dramatically off-frequency, messing up the measurements.These measurements were made with a different TSC5110A unit. I have also observed the bad behavior of #7 using a TSC5120A., which uses a different design. So... I don't think we can ascribe the results to the instrument. Jim