The focus of this paper is on the established observation that the bacteriorhodopsin (BR) photocycle responds to the level of actinic light by altering the proportions of two forms of the M intermediate. The first form of M, called M-fast or MF, decays to the O intermediate. In contrast, the second form of M, called M-slow or MS, decays directly to the ground state, and its decay rate is slower than that of MF. Any proposed scheme for the BR photocycle must account for this light-dependent phenomenon. Several papers have attempted to explain the observation on the basis of photocooperativity, or on the basis of heterogeneous populations. In this paper, we test previously proposed cooperative models with experimental data, and find those models to be inadequate. We show that two new models, one purely cooperative, the other purely heterogeneous, can both fit the data, hence such modelling will not resolve the mechanism. Taking into account the demonstration of heterogeneity, the trimer structure of BR, and certain experimental evidence in favor of cooperativity, it appears likely that both heterogeneity and cooperativity are involved in the adaptation of the BR photocycle to different levels of actinic light.