I had this happen to me once many years ago with a homebrew 2mtr transverter with a design problem.
I used a crystal frequency of a couple of megacycles in one of the stages, that inadvertently mixed in the final stage to output two frequencies. Found the problem when a friend asked me which of the two frequencies I wanted him to transmit on. Had a lotta stages with a lotta crystals....very bad design.
I suspect the fellow was using a home brew rig with a 455 if. A 455 oscilator is getting into the final and mixing with his desired signal.
This is the general idea...am sure somebody out there can explain it better.
I saw a station a couple of years ago in a major DX pileup, that had Eight (8) separate carriers spread over 3 KHz calling the DXpedition. I couldn't tell if he was doing it on purpose to maximize his chances of getting through, or if he was running Audio keyed CW from a popular multi-digital program and was overdriving it with his compressor on and generating multiple spurs.
He finally worked the station. Then a few minutes later, I heard the same signal with a different call calling again. It may have been a club station or husband/wife.
Ken - NM9P
FOLLOW UP: The station op (A member of many large DXPeditions btw) told me that in his hurry to get on 80M CW for the DX station that he was on the wrong antenna and not tuned properly for that particular antenna. He further stated that 80M CW was a seldom used band at his qth. Nothing further to add but it definitely was not what is normally displayed on a panadapter by an overdriving or audio distortion situation
. In the above SSB view that started this conversation, Dave doesn't mention the same SSB signal also shows at 14054.5 and 14057.XX although dropping in intensity like the CW signal I noted this AM.
I need to ask the station what level firmware he is using and other particulars.