To get the 3D image, you have to go 'cross-eyed' so that your left eye is looking at the right image and your right eye is looking at the left image. Once you have this down, you'll see one proper image in the middle, with two partial images on either side of the middle image. Now the trick is to get the middle image into focus and once you do, you'll see it's in 3D.
When viewing a stereo pair this way, I think the image's relief is reversed. Valleys were ridges and vise versa.
Edit: when you're going cross-eyed, you keep going more so until a reference point in the photo (the background ramp in the second photo for example) becomes one. It's at this time when you know it's one image. It really can give you a minor headache after though, and who knows maybe your eyes will get stuck.