Sverdrup's critical depth hypothesis, which has had an almost canonical status in biological oceanography, has recently been challenged as a universal explanation for the formation of oceanic spring blooms, and several alternative hypotheses have been proposed. Arguments pro and contra alternative explanations have so far relied on theoretical considerations and purely observational data. In this paper, we propose that mesocosm experiments with natural plankton communities could make important contributions to the resolution of the issue. We first briefly review the foundations of the critical depth concept and derive an approximate relationship that relates optically scaled critical depth (="critical optical depth", i.e. the product of the light attenuation coefficient and the critical depth) to light-dependent phytoplankton production in the mixed surface layer. We describe how this relationship can be used to scale experimental mesocosms such that they reproduce ambient light conditions of natural water columns from the surface down to the critical depth and beyond. We illustrate the power of the approach with a mesocosm study in which we experimentally controlled the onset of the spring bloom of a lake plankton community through the manipulation of optically scaled mixed-layer depth. This experiment may be the first experimental demonstration of the critical depth principle acting on a natural plankton community. Compensation light intensity (=minimum average mixed-layer light intensity required to trigger a bloom of the ambient plankton community) could be constrained to be somewhat above 3.2 moles PAR m(-2) d(-1), corresponding to a critical optical depth of 10.5. We compare these numbers to estimates from marine systems and end with a discussion of how experiments could be designed to (i) more accurately determine the critical depth in a given system and (ii) resolve among competing hypotheses for vernal bloom onset.