Simultaneous in situ observations of floc size, waves, and currents in a continental shelf bottom boundary layer do not support generally accepted functional relationships between turbulence and floc size in the sea. In September and October 1996 and January 1997, two tripods were deployed in 70 m of water on the continental shelf south of Woods Hole, Massachusetts. On one a camera photographed particles in suspension 1.2 m above the bottom that had equivalent circular diameters larger than 250 μm, and on the other, three horizontally displaced acoustic current meters measured flow velocity 0.35 m above the bottom. The tripods were separated by ~150 m. Typically, maximal floc diameter stayed relatively constant, around 1 mm, and it showed a dependence on turbulence parameters that was significantly weaker than that predicted by any model that assumes that turbulence-induced stresses limit floc size. Occasionally, when waves and currents generated intense near-bed turbulence, flocs were destroyed. These precipitous decreases in maximal floc size also were not predicted by conventional models. The correlation in time between episodes of floc destruction and elevated combined wave-current stresses provides the first quantitative support for the hypothesis that floc size throughout bottom boundary layers can be controlled by breakup in the intensely sheared near-bed region. These observations demand a reassessment of the forces limiting floc size in the sea, and they indicate the potential for significant simplifying assumptions in models of floc dynamics. ¿ 2001 American Geophysical Union