In the theory of numbers, which is still very imperfect, we can place our highest hopes in observations; they will lead us continually to new properties which we shall endeavor to prove afterwards.

In: G. Polya, Induction and Analogy in Mathematics (Volume 1)