- Joined
- Mar 8, 2004
- Messages
- 4,883
dwhatley;113038 said:
I don't have time to respond to this in detail, but in brief: the receptor in the eye doesn't know the difference between reflected, transmitted, or emitted photons... when a photon hits the rhodopsin in the receptor cell (rods and cones in mammals, rhabdomeres in cephs) it's just the wavelength that is perceived. Because humans have 3 cones, with chemically different forms of rhodopsin, that respond to wavelengths in roughly red, green, and blue, we get some combination of how many red-ish, green-ish, and blue-ish photons are coming from what we're looking at. (It's generally safe to ignore rods in these discussions, for reasons I won't go into). Since most cephs only have one type of rhodopsin, they can't perceive color in the same way we can: they can't say "this is brighter in blue and dimmer in green" because they only get one signal that mashes it all together, like a black-and-white TV. No matter what kinds of filters and such you put in front of it, you can't figure out what colors were in a black and white movie... if you're Ted Turner, you can *guess* what colors should be there, and "colorize" it, and certainly cephs might, in some sense, be doing that, but there's just not enough information to perceive color directly.
The polarization is something quite different: in addition to having the wavelength, each photon also has a polarization direction. Humans' rods and cones respond equally well to light polarized in any direction, so we don't notice polarization at all naturally. If we wear polarized glasses and look at a polarized light source, like a puddle or ocean in late afternoon, we can indirectly benefit from most of the glare being polarized so it can be filtered out by the glasses. Cephs, though, while they only have one "color" type of rhodopsin visual pigment, have the villi "fingers" of their rhabdomeres organized in little squares, some of which are oriented vertically, some horizontally, that act like tiny localized polarizing filters, so by comparing adjacent squares, the octopus or cuttle can tell if there is a difference in polarization direction. But this doesn't convey any color information: the polarization is one type of information in the photon, completely independent of the wavelength of the photon that (in the aggregate) we see as color.
Cuttles might well "know" in the Ted Turner sense that when they see a certain pattern of polarization, that can be identified as a seaweed leaf, and that this means they should try to be more green, but that would be more of an educated guess based on the polarization than because they can see the green-ness directly.
I believe that many cephs can also actively control the polarization of light that reflects off of them, either to camouflage themselves relative to other critters that can see polarization (like some arthropods) or to signal each other via the polarization without ruining their ability to stay hidden from us inferior vertebrates who don't see the polarization. I know this is somehow done through the iridophores, which reflect polarized light predominantly in one direction, I think through some sort of stacked-plate arrangement that acts sort of like a polarizing filter. I'm not sure how, but I believe they can rotate this stack of plates under active muscular control, similar to the chromatophores. But this doesn't modulate color directly (although it does mess with the iridescent "sheen" colors that we see a lot in squids a bit, but it still depends on the incoming light that's reflected.)
I have a half-baked article that summarizes what I've read on ceph vision around somewhere, I should probably finish writing that up, and go into this in more detail. Color perception even in humans is a big mess, because different people using it for different things all invent their own ways of describing, explaining, or quantifying it, that usually don't capture what's really going on. I've studied it both from the computer graphics direction and from the primate vision side, and to some extent both of those groups pay attention to what the painters, photographers, filmmakers, TV makers, chemists, printers, and psychophysics people use to describe color, so I'm rather up to my eyeballs in this sort of thing... but humans are horribly biased by being completely immersed in a visual world that's really very different from a lot of other animals: stomatopods have a lot more visual pigments than we do, I think more than 11 in some species, so the "tri-color" business is more of an artifact anyway, plus the rhodopsins' response curves have strange overlap, where there's much more overlap between red and green than blue, but blue response per photon is rather different, and our early visual processing in the retina and early visual path in the brain (lateral geniculate nucleus, IIRC) do things like compare colors to neighboring regions rather than absolutely, leading to many of the kooky color illusions that can fool us.
How long would the detailed post be?!
