Laboratory testing of railway ballast poses practical difficulties because the particle size is often too large for most standard apparatus. There are therefore advantages in developing a scaled material whose behavior is representative of the full size material. A first stage in validating such an approach is to investigate whether the particle shape is affected by the change in scale. This paper sets out methods for evaluating form and roundness (aspects of shape) and proposes a new measure for evaluating roundness, termed ellipseness. These methods are then applied to a crushed rock railway ballast over a range of particle sizes. Statistical analysis demonstrates a measurable variation in the distributions of form and roundness with particle size over a range of sieve intervals, although the differences are slight and do not necessarily rule out the use of a scaled material for investigating the factors influencing macro mechanical behavior.