Manipulative hand movements involve coordinated movements of the fingers to manipulate an object within the hand, and are classified as either simultaneous or sequential. Simultaneous hand movements are characterized by single coordinated patterns of digit movements, while sequential hand movements involve sequences of such patterns. Here, we investigate the extent of the coordination among 15 hand-joints during simultaneous hand movements, and demonstrate that it leads to a concise representation that facilitates movement recognition. Principal component analysis (PCA), performed in the 15-dimensional (15-D) joint-space, indicates that the first principal-component captures more than 98% of the variability in individual hand movements. Consequently, the first principal direction provides a 15-D feature-vector that describes the underlying-coordination and can be used for automatic recognition. We evaluated this recognition strategy on a set of nine simultaneous hand-movements using a database of six users, each performing six sessions. A dedicated classifier for each user resulted in recognition rates of 97.0 +/- 4.7% during testing, while a single generic classifier achieved 95.2 +/- 2.5% recognition rates. We conclude that the suggested feature-vector captures the invariant structure of simultaneous hand-movements, facilitates their recognition, and may provide insight into motor planning.