'Double modes' seen in some PI mode bands are not from test mass resonances, in agreement with model conclusions. Follow up on this .

I've looked at frequency shifting of the peaks seen in the 15060 - 15090 Hz range during 11 hours of the current lock.

• Four predictably follow the drumhead modes of the four test masses; these are the respective ~15070 Hz resonant modes (blue, green, orange, red)
• Four show no following drift, indicating they're not tied to mechanical modes of the optics which would increase as the optic heats up (steel, grey, light pink, mauve)
• One has similar behavior/timing of a resonant mode but shifts down in frequency; aliased down from an upward shifting resonance at 50451 Hz (similar to suggestion by Slawek here)? (pink).
• The last peak is questionable. It does shift over the 11 hours, but doesn't seem to be following one of the test mass drumheads and the Q is ~15 K as opposed to test masses ~20 M (lavender) 

Attached plots:

• spectrum showing the peaks at beginning of lock and 11 hours into it & plot of frequency drifts. Dots over peaks correlate to color mentioned above and color of line in next plot
• plot showing change of frequency, first two hours of lock and 9th and 10th hour of lock. Colors match those mentioned above and peaks in spectrum. I've also included the known ~8160 Hz drumhead modes of the test masses and matched the 15070 Hz mode that follows to that color.