Consider

Take the log (or natural log, or log to your favorite base, whatever) of both sides:

What does this do for you? Well the equation is now in the form

where

So what you do is a linear regression with the log(L) values being your x values and the log(T) values being the y values. Then the slope m of the regression will be n, and the intercept b will be the log(k).

Edit: For reference I get that and with an , so the fit for is excellent. This is not quite the same as Soroban's answer (which is easier to get) but is probably more accurate. (Sorry Soroban, I'm not trying to pick on your answer! )

-Dan