Years and years ago, SLR cameras used to use the split prism method of focusing. This was before autofocusing was a standard feature on cameras. In order to get a lens in focus, the photographer had to look through their SLR with the lens attached and turn the focusing ring accordingly. When the image in the center lined up perfectly, it meant that the scene and subject int he center was in focus.
Today, this method isn’t use as much except if you manage to get your hands on a matte or split prism focusing screen for your DSLR. They work exceptionally well with manual focus lenses, but modern screens and viewfinders have become so good that it’s arguable that you don’t even need them anymore. This method also wasn’t always the most accurate when shooting wide open. Additionally, many photographers back then loved to focus and recompose. So it was almost useless to use this method unless you were stopped down quite a bit or were incredibly careful.
All of this worked through bouncing the image from the mirror to the prism which then showed what the lens saw to the photographer who peered through the viewfinder. This method is still used with some medium format cameras that use manual focusing lenses. But generally, it’s over and done with. Fujifilm, however, brought this method back in a digital version with their X100s camera. However, their simulation is absolutely nothing compared to the real thing.
The video after the jump demonstrates this is better detail, check it out.