Is it true that in the dark, humans only see black and white?
If it’s truly dark, humans see nothing.
The cells in your retina which detect light over the entire visual spectrum are easier to trigger than the cells which only trigger for certain colors. So you need brighter light to distinguish color than merely to see the intensity of light. So yes, when there’s barely enough light to see by, you’ll be seeing in black and white.
What is the word for the colors we see when we close our eyes ?
Andrew is right, AFAIK but I thought I’d add a bit. The two types of cell are rods and cones. The cones are the ones that can detect colour, in particular there are 3 types of cones, that can each detect either red, green or blue components of light. The rods can word in low light conditions, but only in black and white, hence you can only see in black at white at light.
This website gives loads more details on this:
http://hyperphysics.phy-astr.gsu.edu/hbase/vision/rodcone.html, but there are probably a lot more sites out there, and that was one of the first I found using google.
Phosphenes, IIRC.
In low light circumstances, we see using our rods, rather than with the color-sensing cones. The type of vision we have with rods is called scotopic vision, while with the cones we have photopic vision. The peak sensitivity with cones is about 555 nm (green), tailing off in a nearly Gaussian (“bell-shaped curve” pattern)fashion. Scotopic vision peaks a little more in the blue (which may explain why things look sort of “blue” in low light), and still has that Gaussian bell shape.
Pressing on your eyelids gives you sensations of colot from the cones “firing”, and I think Engywook is correct in calling them phosphenes. There is another effect where you can see orange “tails” from small light sources in an otherwise dark room, but that’s due to a different cause. It’s in Walker’s book The Flying Circus of Physics. It’s not to be confused with the halos and coronas you see if there are defects in your eye. There are a lot of things that can cause color or light effects in a darkened room.
See Walker’s book. Or a good book on the Eye.
Actually, if I remember correctly, there are only two kinds of cones: One which distinguishes red from yellow, and one which distinguishes green from blue. Together with the rods which measure absolute intensity, this gives us the three degrees of freedom of color.
Also, if you want to prove to yourself that you can see–but not in color–in dim light, find some things that are identical except for color; for example pieces of construction paper or gumballs (I swear I’m not in 4th grade). Go into a darkened room and mix up said objects. Under the right levels of light, you’ll be able to see the objects just fine, but won’t be able to name their color. Sometimes you’d swear you can tell the color of objects in a room even when it’s so dark you almost can’t see them, but this usually is because you already know what color they are. That’s why you have to have gumballs or some other identical objects to prove it to yourself.
silverfish, it’s supposed to be “work” & “black and white at night”, I guess. That link is worth to be fixed (the URL parser doesn’t like trailing commas): The Rods and Cones of the Human Eye
There have been shown to be four different photoreceptors in our eyes, one rod- and three cone-types. Spectral Sensitivity curves of rods and cones The output of the rods saturates at high illumination levels and doesn’t contribute to color vision. However, there follows a lot of processing, and then a model of one brightness- and two color-channels describes it quite nicely.
You know, that sounds like a threat. Please do use the flat side. 
The locations of rods and cones also has a lot to do with it. The fovea is a tiny area of the retina where the optical system of the eye (pricipally cornea and lens) are focussed. The fovea consists almost exclusively of cones, very highly packed. As you move further out, the concentration of cones drops off dramatically.
One effect of this, which requires conscious effort to confirm, is that peripheral vision is not in colour. When you think you are seeing objects outside the point of focus, you are really seeing a black-and whie image, and you only think you see the colours because your brain is filling in the missing pieces.
There’s lots of other interesting things that happen at night, like seeing stars better out of the corner of your eyes than by looking directly - again because the focus of the eye lacks rods.
I’ve heard this said quite often, but I believe it to be false; I conducted an experiment:
I took a box of my daughter’s felt-tip pens in assorted colours, I mixed them up and put the box behind my back, staring directly ahead at a fixed point on the wall, I rummaged in the box and took out a pen at random. With arm outstretched, I moved the pen slowly forward from behind my field of view, stopping at about 90[sup]o[/sup] (straight out to the side), all the time being most careful to keep my gaze fixed straight ahead. I was consistently able to discern the difference between colours.
I’d agree that peripheral colour perception is vastly impaired (it was not so easy to tell the difference between brick red and scarlet, or between lilac and mauve, but it simply isn’t true to say that colour perception is absent
Try it for yourself.