special blue light

something i forgot to mention in the Montreal post:

standing on the balcony outside AR's apartment, you can watch the alley four floors down. across the alley is what looks like an old office building - occupied, i mean, just old. it's nighttime, dark with moonlight and streetlights. the lights are mostly out in the building across the alley.

on the ground floor, through a window, seemingly under a desk, you can see a faint blue glow. it seems to come and go, but in fact it's constant. it's a blue LED at the front of a computer tower, or some other device. it's just dim enough that if you foveate it, you see blackness, but if you look away just a degree or more, it pops into view.

it's easy to see this effect with stars at night, or with any dim detail when you're dark adapted. and we all know that the fovea is free of blue cones, though i think this light was not so blue that it wouldn't be stimulating green cones, if they were sensitive enough. the trick is that they were not sensitive enough, but the periphery was. but the answer can't be that it was rods mediating the peripheral seeing, partly because the color sensation was plain, and partly because we were going in and out of the apartment, there were other lights all round, so the rods shouldn't have been especially useful.

so the explanation must be the blue cones and the insensitivity of the green cones. the light was dim enough to be invisible to green cones, which if it were brighter would probably  be stimulated (though maybe it was blue and pure enough that this latter point isn't even true; however, a quick google suggests that standard blue LEDs have peak spectral power at around 540nm, which is within the M-cone tuning width).

anyways, i thought this was interesting. it took me a couple of minutes to convince my companions that the blue light wasn't actually slowly flashing on and off, that it was all in their behavior. that was the best part.