Actually, finding, or avoiding, the light is not quite that simple. It appears you also have to spin round and round as you go http://www.nature.com/nature/journal/v456/n7220/full/nature07590.html
Cool.
I was thinking of benthic proto-chordates, and didn’t even consider pelagic organisms.
The story I read once - most animals have eyes on both sides of their heads - i.e. horses - so they can see predators approaching all around. The ones that relly need to concentrate and know what’s near them and exactly where, find binocular / stereo vision useful. You see this in dogs or even moreso in cats, who have to jump and grab things. Or monkeys, and hence humans, who have to judge how far away that next branch is when they jump for it. So stereo evolves from panavision for animals where it is an advantage.
The evolution of a light-sensitive patch from simple to directional (in a hole) to protected and multi-point, to maeuverable, is an obvious evolution. There are still some primitive examples where it is not necessary.
As a note, evolution loves to do the accidental doubling-up trick. A lumpy simple organism may accidentally double-up its genes and become a segmented double organism. If this is an advantage (twice the limbs) it may happen again- this is why so many life forms are repeating segmented. Then, each segment may change and specialize for advantage. A lobster is multiple segments where some limbs are just little fins, others are big claws, others are legs; a millipede is mostly repeating leg segments. Our ribs are the same concept.
Our eyes have 2 types of light sensors, rods and cones. Rods are low-ight sensitive and give black-white ability (hence, no colors when it’s twilight) While the cones have color sensitivity. The original cones doubled and then again so we have 3 types of cones, red green blue sensitive. SOme people acidentally don’t develop one or more set (color blindness); some people, I read, accidentally develop a fourth set. not sure what the result there is, better color sensitivity? So in that regard, color sensitivity may still be evolving.
However, are color cues a survival / better breeding trait compared to earlier times? (“This meat looks green… better not eat…”) If not, it’s just irrelevant genetic drift.
Good color vision is an important survival trait for monkeys who need to be able to spot the fruit amongst the leaves. For humans, perhaps not so much, but we got to keep it and we find uses for it.
Most primates (including humans) are trichromats (3 cone types), whereas other mammals are, at best, dichromats (just two cone types) and thus cannot distinguish colors that look quite different to non-color-blind humans.
Because women have 2 X chromosomes they carry two copies of genes for cone cell receptor proteins. In their retina,some of their cone cells random express the gene from from one X chromosome and some from the other. Normally this does not make any difference, because the genes are the same, but if one of them has mutated just enough to shift the spectral sensitivity of the cones that use it, she will have 4 cone types (although the new type, I think, is less sensitive overall). I believe there is evidence that this does happen, and that it it does allow these women to distinguish colors that look the same to most of us (I might be able to dig up a site if anyone cares). I doubt whether it is a selective advantage, though, especially as it means that some of their sons will get the mutated gene and not the normal one, leaving them with a form of color blindness. The mothers (and some of their daughters) gain a bit of color sensitivity, but the sons lose out.