The first theoretical predictions for the behavior of magnetic particles in water were that sedimentary magnetizations would be fully aligned with the ambient field, yet redeposition experiments showed a strong (and quasi-linear) dependence on the external field. This empirically observed linearity has served as the fundamental assumption of sedimentary paleointensity studies for decades. We present redeposition experiments which suggest instead that the relationship between depositional remanence (DRM) and applied field may frequently be curved for magnetic fields in the range of the Earth's. Numerical simulations using a flocculation model can explain the redeposition data and suggest that DRM will be significantly non-linear when the flocs are small (several microns). There is a strong dependence of floc size on salinity particularly in low salinity environments. Floc size has a profound influence on the efficiency of DRM, hence low salinity environment may give results with poor reproducibility. The size of the floc in which magnetic particles are embedded is not accounted for in current methods of normalization, yet is the most important parameter. On the bright side, however, it now seems possible to quantitatively explain paleointensity in sedimentary systems opening the door to absolute paleointensity estimates from sediments whose key parameters of floc size distribution and settling times can be constrained.