Following a brief review of the light-driven reactions in photosynthetic membranes, two questions are addressed. (1) Why is the first charge separation reaction in photosynthetic reaction centers so fast; and (2) given what we know about the contemporary structure and function of reaction centers, can we develop a simple model for a much more primitive reaction center? It is proposed that the primary charge separation step in reaction centers is optimized to be ultra-fast principally in order to compete with detrapping into the antenna complex, rather than to compete with radiative and non-radiative losses in the special pair. This leads to a notion of kinetic perfection analogous to that developed for enzymes which operate under diffusion-limited conditions, but elaborated to permit even more 'perfect' function. This hypothesis is testable by changing components in photosynthetic membranes and subjecting them to selective pressures. We speculate that the reaction center is far too complex to have served as an early functional unit, and consider possible roles for the iron-quinone part of the reaction center as a very primitive photosynthetic unit. It is suggested that this working end later became associated with primitive antenna complexes, which then evolved into the elaborate structure we find today.The role of photosynthesis in the origin of life has been a topic of speculation for many years. It is evident that photosynthetic function is ancient and central. As a person who does not work in the field of evolution. I am not very familiar with much of the speculation that precedes this paper. Proposals and speculation by others are likely based on much firmer ground, and therefore I apologize in advance if some of these ideas have already been suggested by others. In this chapter I take the liberty to speculate on how a structure as complex as the contemporary photosynthetic reaction center (RC) could have evolved from more primitive units, and why it retains some of its remarkable properties and seemingly unnecessary components. Both subjects lead to specific predictions and testable hypotheses.