"Stronger light at the far end might sense when the sensor is a long CCD bar."
Yes, I wondered about that CCD thing, but don't remember it being a big problem up to about 2k cells.
I'm quite sure it's a light tube ATM, with less dots at the bright end and many more dots at the far end. Now I've had a proper look in daylight the dots are all small rectangles roughly 0.4mm x 0.6mm.
I think the missing dots/rectangles at the bright end are probably correct, they just haven't got the radomness quite 100% correct, it might depend on the exact way a photograph surface reflects.
"Are the missing dots less bright than the others or are they completely missing?"
The dots appear to be slightly raised rectangles, I thought the missing ones might be just painted black but even in daylight with a magnifing glass they appear to be completely missing.
"I assume you tried looking at it from different angles to rule out that it only 'looks' that way from the angle you're viewing it at?"
Yes, I have to view it on the side where the very thin CCD line is and 25-50 degrees off vertical to see three bright rows, not much light goes vertically up.
It's switched off in these 2 images, it's just bright daylight showing the bright dots and the missing left hand ones.