Crystallising large, flexible molecules, which are becoming more common in pharmaceutical development, often presents significant challenges for chemists and particle scientists. These difficulties are sometimes attributed to the flexibility of the molecule, and the existence of multiple conformers in solution. Structurally related impurities, frequently present when crystallising these materials, can also impact on growth and habit, and both these aspects are considered in this thesis. This work considers two pharmaceutical compounds, a relatively small but nonetheless flexible molecule, ethenzamide, and a precursor of Amprenavir, a much larger molecule. Both compounds typically grow as thin needles in a wide variety of solvents, and effort was required to grow suitable crystals for structure determination. Ethenzamide has an unusual structure, the amide group being out-of-plane relative to the ring, while in all known co-crystals of the compound, including three new co-crystal structures determined in this work, it has a planar structure with an intramolecular hydrogen bond not seen in the single component crystal. Theoretical structure generation calculations suggest a second polymorph with a planar conformation may exist, though a screen has not found any further solid phases. ab initio work suggests the planar conformation is the stable arrangement in vacuo. Several structures for the Amprenavir intermediate have been determined, as an ethanol solvate, a methanol solvate and a hydrate. A phase diagram has been measured in the industrial solvent mix, and the nucleation and growth properties of this molecule, both pure and in the presence of several structurally related impurities, have been measured. The Cambridge Structural Database has been searched for similar structures, and conformational searches have been carried out for both molecules, using vacuum phase ab initio energy calculations. Infrared spectroscopy has been used to investigate solution phase structure. These theoretical and practical studies will try to relate conformational properties to crystallisation behaviour.