This has been bothering me lately. My proof-skills are rusty (and were never great to begin with). I dimly recall having seen this (or something related to it) in a math course I took a while ago, but I can't remember any details now.
Obviously, if $a < b$ and $a$ and $b$ are reals >= 1 (or equal to 0), then $a^2 < b^2$, and hence $\sqrt a < \sqrt b$ (since the sqrt of a, b are in the same domain and squaring them would preserve the inequality and we know a must be < b to begin with).
What's not obvious to me is if this holds for real numbers in the range 0-1.
And what about the general case -- do inequalities on arbitrary real numbers hold under exponentiation? (Assuming the exponentiation yields defined results -- no square roots of negative numbers, for example.)