The lowest photon flux density of photosynthetically active radiation at which O2-evolving marine photolithotrophs appear to be able to grow is some 10 nmol photon m−2 s−1, while marine non-O2-evolvers can grow at 4 nmol photon m−2 s−1, in both cases with the photon flux density averaged over the 24 hour L:D cycle. Constraints on the ability to grow at very low fluxes of photosynthetically active radiation fall into three categories. Category one includes essential processes whose efficiency is independent of the rate of energy input, but whose catalysts show phylogenetic variation leading to different energy costs for a given process in different taxa, e.g. light-harvesting complexes, RUBISCO and probably in the sensitivity of PsII to photodamage. The second category comprises essential processes whose efficiency decreases with decreasing energy input rate as a result of back-reactions independent of the energy input rate, e.g. charge recombination following charge separation by PsII and short-circuit H+ fluxes across the thylakoid membrane which decrease the fraction of pumped H+ which can be used in adenosine diphosphate phosphorylation. Category two also includes that component of protein turnover which cannot be related to replacement of polypeptides which were incorrectly assembled following uncorrected errors of transcription or translation, or which were damaged by processes whose rate increases with increasing energy input rate such as photodamage to PsII. The third category includes only O2-dependent damage to the D1 protein of PsII whose rate increases with a decreasing incident flux of photosynthetically active radiation. Processes in categories two and three are most likely to impose the lower limit on the photon flux density which can support photolithotrophic growth. The available literature, mainly on organisms which are not adapted to growth at very low photon flux densities, suggests that three major limitations (charge recombination in PsII, H+ leakage and slippage, and protein turnover) can individually impose lower limits in excess of 20 nmol photon m−2 s−1 on photolithotrophic growth. Furthermore, these three limitations are interactive, so that considering all three processes acting in series leads to an even higher predicted lower photon flux density limit for photolithotrophic growth.