I don't think mathematics was ever considered a science. See the definition of the 'liberal arts', which includes science and math as separate entities:
No, they aren't. A mathematical proof is by definition unfalsifiable. If an error is found in a proof (as with Wiles' proof of Fermat's Last Theorem), that it isn't a proof ... yet. And mathematical proofs aren't empirical, another requirement for science.
Scientific theories are falsifiable in perpetuity because the possibility always exists for new empirical evidence to show up that falsifies an existing theory. This possibility doesn't exist for mathematical proofs.
http://en.wikipedia.org/wiki/Liberal_arts_education