There are now four known cases of multiply imaged QSOs, one with a detected foreground object at roughly half the affine distance to the QSO: 0957+561 (17 mag, z=1.4, separation =6″), 1115+080 (17 mag, z=1.7, sl=1.8″, s2=2.3″), 2345+007 (19 mag, z=2.1, s=7″) and 1635+267 (19 mag z=2, s=5″). In addition, 1548+115 (19 mag, z=1.9) is a probable lens event with a foreground QSO, but no secondary image has been found. Perhaps 500 candidate QSOs have been surveyed optically for multiple images by all observers. 0957+561 is the only catalogued QSO shown to be multiple. Of the remaining 1548 QSOs currently catalogued, any secondary image is masked by atmospheric scattering of the QSO light. Typically, this sets detection limits of ≳ 3 mag fainter and < 2 arcsec separation from the bright component, for any secondary image. Objective prism and grism surveys look directly for multiple QSOs with identical emission lines and have surveyed 1500 QSOs. The remaining three lensed QSOs come from these more efficient surveys. Although the exciting search for multiply imaged QSOs has only begun, sufficient data already exist to test two hypotheses: (A) QSOs are intrinsically luminous and occasionally are multiply imaged through a chance alignment with a foreground galaxy of sufficient mass gradient and (B) all QSOs are the result of gravitational lens magnification of a distant Seyfert nucleus by foreground galaxy(s). I will first address hypothesis B, then A. I assume that mass (seen and unseen) clusters with galaxies and/or clusters of galaxies.