Yes, strength of bone goes as the cross section L^2 (L is linear size) where as weight goes as L^3. This is why elephants need relatively fat legs and birds are fine with thin.
Muscle strength can be more complicated. The NUMBER of muscle fibers goes as L^2, but the total strength should scale somewhat as the length of the fibers since there are more motor proteins over the muscle fiber length. So this is more than enough (perhaps) to explain the size discrepancy of sprinters with the small-is-better observation. Overall however, all other thing being equal, strength per unit mass for more static, endurance-type activities should scale as gamma = L^2/3, so smaller is usually better for endurance activities. For more dynamic activity the effective exponent goes up, gamma >2/3 (in other words penalty for being bigger is less due to compensating factors). An argument can be made that in the dynamics case (sprinting, jumping) the exponent is close to unity (no size bias), and this explains the often noticed interesting factoid that a human can jump as high as a mouse can jump as high as a flea. For falling damage you totally have to go back to gamma = 2/3 which is why you can drop your hamster off the roof an it will hardly notice.
Of course if you are a endurance-based hunter (or lunch money stealer) you have to be big enough to bring down you prey, or like dogs/wolves, have enough friends to help you out.
The type of scaling stuff was first explicitly mentioned in terms of mathematical laws by Galileo in his "Two New Science", to be distinguished from "Two World Systems" – the one that got him in hot water with the Vatican douchebags. I used to teach that the first day of my freshman physics course.