Birds’ “knees” point the other way. Take a look at deer, their hind legs also have “knees” going the other way. I guess whenever an animal needed to have its knees point the other way, it’d use its angle. There’s definitely a benefit even if you walk forward.
But it seems there is something fundamentally hard, genetically, to re-engineer the walking fish’s innovations. Vertebrates share a lot of features among their bones.
Then again, if you’re a human and you try to walk on all-fours, your true knees (meaning elbows) will turn the other way! Are humans the only vertebrates whose knees ever [came close to] rotating since the aforementioned stinking fish?
I should clarify this a bit. A bat’s knees and ankles flex the same way relative to the leg as a normal mammal’s. It’s the hip joint that’s changed, so that the leg as a whole is reversed.
The direction a joint flexes is rather irrelevant to the direction of walking. After all, the elbow and the knee flex in opposite directions, yet the animal moves forward. And as has been mentioned, the joints you mention are not knees at all, but ankles (more or less).
Look at the diagram of the cat skeleton Colibri gave in his second post. The joint between the humerus and the forearm bones (radius and ulna) points the same way my elbows do if I try to walk on all fours.
Not all. The tarsier has an extremely elongated hind foot, and climbs using mainly the toes alone.
But the joints are consistent to both forelegs and back legs, it’s only the proportions that are different. Aren’t they? The elbow is technically opposite to the knee, but its equivalent in a quadruped is the ankle, which bends the same way.
Er… Maybe I’m proving your point by arguing the opposite.
Surely the equivalent of the ankle is the wrist, no?
Colibri said the joints of the legs can be any direction, and his example was that the foreleg joint, the elbow, is in reverse to the back leg knee.
But effectively the actual joint directional equivalent on each leg, is the foreleg’s elbow to the back leg’s ankle. The forward motion is achieved by having the two sets of limbs effectively jointed in the same ‘<’ formation.
Yes, they aren’t the equivalent in parallel joint comparison, but they are equivalent in relative placement from the ground up.
Because of intelligent design, duh.
The Aristocrats!
Not at all. This depends entirely on the leg proportions of the animal. Some have elbows and knees at the same level, others don’t. Differences in the direction of flexion of a limb at a particular point above the ground are achieved by lengthening or shortening the limb bones, not by changing the direction of flexion of the joint itself.
People run backwards marathons. Did you know that? Because I think those people are screwed up. 
It’s not so much the same level, though that does play into it, as that they do the same effective work; which is, by bending in the same direction they achieve the most efficient stride. For that reason I consider those joints as being the equivalent of each other.
But I accept that you know more about animal physiology than me, so I will bow to your superior knowledge.
I think to understand the mechanics you need to actually look at a photo of the animals in question. In this pronghorn, the elbows and the knees are at the same level (the joints closest to the body), as are the wrists and ankles. Note that the joint in the middle of the fore leg is the wrist, and the joint in the middle of the hind leg is the ankle. In both cases the joints at the same level bend in opposite directions. The most efficient stride is achieved by having the corresponding leg segments approximately the same length; it has little to do with which direction the joints bend.
The pronghorn’s legs display adaptation for fast running. Most of the heavy muscle mass for moving the legs is up near the body, in the thighs and shoulders. The lower leg segments are thin and elongated in order to lighten them so they can be moved rapidly with a minimum of effort.
Anyway, on a ‘selective advantage’ level of ‘why don’t humans tiptoe when cats do?’, I presume the reason is that cats don’t need to spend any time balancing on only one foot, therefore can get away with a small point of contact (since there’s always two or three other feet on the ground).
Humans, by contrast, need both toes and heels on the ground in order to stay balanced, since often there’s no other limb on the ground.
Somewhere, at sometime, there must have been a mad scientist who tried this.
Just for the record, the appropriate technical terms are:
Plantigrade: Walks on “full foot” (in human sense) with both heel and toe touching ground. Forefoot is equivalent to full hand. (May adopt a different gait when running.) Examples include humans and many other primates, ursids, proboscideans.
Digitigrade: Animal’s “foot” is ball of foot and toes, fingers and fleshy pad wher they join palm. Ankle and wrist constitute distalmost joints of legs.
Unguligrade: Animal runs on (a) a hoof at the end of the third finger and toe, all others being lost or reduced to splints (equids), or (b) a double (cloven) hoof at the end of two conjoined fingers or toes (most artiodactyls), same caveat.
