The development of improved analytical techniques has facilitated the application of short-lived isotopes to the study of magmatic processes, and resulted in a renewed interest in a number of other chronometers. Two approaches have been used to determine the time scales of magmatic processes. Isotopic dating provides absolute ages for the growth of mineral phases. This usually involves analyses of mineral separates such that the textural relations of the individual grains are difficult to establish. An exception is zircon, which can be analysed in situ. The second approach is to use relative chronometry based on major, trace element and isotope profiles in crystals that may have been modified by diffusion. These yield information on how long crystals were at a particular temperature, without indicating when this occurred. The ages are obtained on individual crystals, and so age distributions can be determined on different crystals from the same whole rock. The ages of crystals and the liquid, as represented by the groundmass, in an igneous rock can be different, and in a number of cases it has been shown that even the larger, and therefore typically older crystals formed after the fractional crystallisation responsible for the whole rock composition. One implication is that the processes of magma differentiation responsible for whole rock compositions may not necessarily be inferred from the compositional record of the larger crystals. Different approaches are therefore used to investigate the crystallisation history and the differentiation of magmatic suites. Crystallisation rates are ∼10−10–10−11 cm/s, whereas differentiation to high-silica magmas may take up to 2×105 years. The ages of crystals at the time of eruption can range back to 105–106 years, the older ages tend to be in the more evolved rock types, and it can take 105 years for high-silica magmas to be generated at individual volcanic centres. It appears that the generation of such evolved magmas is thermally controlled, for both fractional crystallisation and the generation of crustal melts, and the rates of fractional crystallisation can, for example, be linked to volcanic power outputs. If crystallisation is in response to magma degassing or decompression, it will be fast and there may be too little time for fractional crystallisation to take place.