Understanding and applying color depth options in PS CS5

Question

I was reviewing a very helpful tutorial about bit depth. (Here's the link, in case anyone is interested: http://www.cambridgeincolour.com/tutorials/bit-depth.htm).
I'd like to get a better handle on how PS CS5's choices (Image>Mode>8,16 or 32) actually affect my work.  I think I may be having trouble relating how settings in the .psd file I'm working on affect a final image that might be output as, say, a .png file.  Here are some of the sources of my confusion:
If I open a new .psd document, I find that 8-bit RGB color is the default. Given that 8-bit is VGA, and it takes 24-bit to display True Color, I was surprised, first, that PS nicely displays a piece that I started in a 3D modeling program--a piece that has some photographic texturing of animal fur and horn (a bull).  Those are fairly subtle gradations.  Moreover, it displays it on a transparent background. It seems unlikely that I'm looking at 256 colors.
If I were to upload a .png file to some printing companies (e.g., CafePress), they want PNG24, not PNG8, and, in fact, my 8-bit RGB .psd file can be exported as PNG24.  So, I gasp in wonderment at how my 8-bit .psd file can export to PNG24.  I understand the simple math to bits per channel and bits per pixel, but it doesn't translate to an understanding of how bit depth actually imposes limitations or opens possibilities when creating and outputting images.
Any help would be much appreciated.  Thanks.
Doc_Pit
http://www.cambridgeincolour.com/tutorials/bit-depth.htm

Answer 1

8 bits PER CHANNEL.
16 bits PER CHANNEL.
32 bits PER CHANNEL.
24 bits is meaningless, but usually intended to mean RGB (3 channels) of 8 bits per channel.
PNG 24 is more correctly called PNG RGB 8 bits per channel.
PNG 32 is RGB plus transparency at 8 bits per channel.
(PNG also supports indexed color, and 16 bit per channel modes).