Comprehensive knowledge of the order-disorder relations in amphiboles is essential to (1) complete understanding of the crystal chemistry of these minerals, and to (2) the use of amphiboles in thermodynamic calculations (e.g., of temperature and pressure of equilibration) where accurate activity models are critical to the accuracy of such treatments. Moreover, such information is essential to our understanding of phase relations, optical and electrical properties and dehydrogenation mechanisms. As a result, more effort has been expended on characterizing site occupancies in amphiboles than in any other group of minerals. Hawthorne (1983a,b) reviewed in detail all work prior to 1983. Here, we will briefly summarize this work, and focus more on what has been learned since then. We will briefly review the common methods of deriving site populations. It is important that everyone who uses site populations has an appreciation of the methods used to derive this information, as a significant fraction of the data in the literature is wrong, and the user has to be in a position to assess the accuracy and precision of the data that they will use. The most comprehensive method of deriving site populations is crystal-Structure REFinement (SREF), as this method senses every atom (in significant amounts) in a crystal. However, this universality has a negative side. Some atomic species cannot be distinguished as they scatter radiation in a very similar way. Different types of radiation (e.g., X-rays, neutrons) can compensate for this drawback: for example, X-ray scattering cannot easily distinguish between Fe and Mn or Fe and Ti (differences in size of these species can be used for this purpose, but this becomes inaccurate to ineffective in more complicated compositions), whereas neutron scattering can distinguish these species. However, there are many instances where such differentiation is not possible. It is here that …