Plasticity size effects have been the focus of strong recent interest due to the miniaturization of devices such as actuators and different medical apparatuses. Devices with a sub-micrometre dimension have a high market potential and therefore a thorough understanding of their mechanical properties is required, both for fundamental reasons and to increase the reliability of such devices. The development of new sample preparation techniques and the continuous improvement of testing methods for the characterization of samples with a submillimetre dimension, have allowed to gain new insights in the underlying phenomena governing the plastic deformation at the scale of a few micrometres or less. Nevertheless, the precise phenomena remain elusive and some data are debatable, notably due to constraints of the used testing methods and the damage present in samples produced through focused ion beam milling, which is the most used fabrication process for specimens in the micrometre scale. This thesis has two distinctive objectives. On one hand, we develop a novel process for the production of samples having a submillimetre dimension based on casting. This process consists in pressure infiltration and directional solidification of a liquid metal in monocrystalline moulds containing submillimetre cavities. On the other hand, we use the wires prepared by this novel microcasting process to study dislocational plastic deformation at small length scales, in tension, in order to contribute to our understanding of plasticity size effects and the underlying plasticity mechanisms at the micro scale. Monocrystalline 99.99% purity aluminium wires, with a random crystallographic axis orientation and a diameter between 6 um and 110 um were fabricated using the developed microcasting process. These wires have a surface roughness on the order of 10 nm and are between 1 mm and 3 mm long. Furthermore, they are monocrystalline and contain a low density of dislocations. Results from tensile tests indicate that the critical resolved shear stress increases with decreasing wire diameter below 20 um. The observed increase is inversely proportional to the wire diameter, indicating that the plastic deformation is governed by the dislocation source length. Critical resolved shear stress values calculated from a model based on the stress required to activate a single arm and the additional stress necessary to produce slip step on both sides of the oxide, follow the same trend than the experimental data but with lower stress values. Unlike bulk aluminium, the wires exhibit a highly intermittent flow stress corresponding to the abrupt emission of discrete dislocation avalanches. The complementary cumulative size distribution function of events of size greater than 250 nm follows an exponential distribution. This suggests that the plastic deformation of the wires is governed by a random event by which source activation is stopped, the probability of this event remaining roughly constant as plastic deformation progresses. Variations in the exponent of the exponential distribution of the avalanche sizes shows that the probability that the event by which a given dislocation emission event is stopped depends on the monocrystalline wire axis orientation but not on its diameter. This suggests that the blockage of an emitting dislocation sources is not due to obstacles in its glide plane or along the dislocation length, but to unpinning of its blocked segment.