Experimental and theoretical evidence is presented which suggests that two distinct forms of light-adapted bacteriorhodopsin may exist. We propose that these two forms have characteristic photocycles with significantly different primary quantum yields. INDO-PSDCI molecular orbital procedures and semiempirical molecular dynamics simulations predict that one ground state geometry of bR undergoes photochemistry with a primary quantum yield, Φ1, of ~ 0.27, and that a second ground state geometry, with a slightly displaced counterion, yields Φ1 ~ 0.74. This theoretical model is supported by the observation that literature measurements of Φ1 tend to fall into one of two categories- those that observe Φ1 ~ 0.33 or below, and those that observe Φ1 ~ 0.6 or above. The observation that all photostationary state measurements of the primary quantum yield give values near 0.3, and all direct measurements of the quantum yield result in values near 0.6, suggests that photochemical back reactions may select the bacteriorhodopsin conformation with the lower quantum yield. The two photocycles may have developed as a natural biological requirement that the bacterium have the capacity to adjust the efficiency of the photocycle in relation to the intensity of light and/or membrane electrochemical gradient.