The mosaic model for thin molecular organic microcrystals does not adequately account for the electron diffraction intensities from them; nor is two-beam dynamical theory adequate. A bent-crystal model for diffraction proposed by Cowley [Acta Cryst. (1961), 14, 920-927] explains well the observed electron diffraction intensities from orthorhombic paraffin microcrystals. In addition, previous verifications of n-beam dynamical electron scattering from these crystals (by fit of phase-grating structure-factor moduli to observed data and by the nonexistence of kinematical data at small electron wavelengths) are confirmed further by a good correspondence between calculated and observed n-beam data for increasing crystal thickness (i.e. monolayers of increasing-chain-length paraffins) and the change of continuously excited reciprocal-row intensities as the crystal is rotated about the axis. The better model for thin molecular organic microcrystals is therefore that of an elastically deformed perfect-crystal foil. Although n-beam dynamical effects can be demonstrated, the major limitation to crystal-structure analysis using intensity data from thin organic crystals is caused by the elastic bendings. This is particularly true if the crystals are solution grown and the longest unit-cell direction is therefore parallel to the incident beam. Epitaxial crystal growth, which minimizes bend effects to intensity data by forcing a shorter unit-cell axis to parallel the incident beam, effectively removes this apparent coherence restriction.