A Note on squares in arithmetic progressions, II
We show that the number of squares in an arithmetic progression of length is at most , for certain absolute positive constants , . This improves the previous result of Bombieri, Granville and Pintz [1], where one had the exponent in place of our . The proof uses the same ideas as in [1], but introduces a substantial simplification by working only with elliptic curves rather than curves of genus as in [1].