This review paper is intended to be a guideline to novices on how to conduct research on silicate melt inclusions in volcanic environments, which analytical techniques are more suitable to gather the desired data and the major pitfalls scientist may encounter. Silicate melt inclusions (SMIs) are small quantities of silicate melt that are trapped in minerals during their growth or crystallization. They contain liquids formed in equilibrium with their host minerals and therefore record the liquid line of descent of magmatic systems. Upon trapping, SMIs become ideally closed to the surrounding environment, and behave as time capsules, giving important information about processes that originated magmas and the nature of their mantle source. A melt inclusions investigation is composed of several steps: (1) a detailed petrographic study to characterize and select representative SMIs, with the aim of identifying Melt Inclusions Assemblages (MIA), the only reliable tool to assess that SMIs obey Roedder's rules and have not re-equilibrated following entrapment; (2) a careful preparation of samples for re-heating experiments and microanalysis; (3) high temperature studies in order to homogenize the SMIs and be able to perform. chemical analysis; (4) interpretation of SMI5 data, which should always be compared with bulk rock composition and relevant experimentally derived liquid compositions. We suggest that the search for good SMI candidates for study will be achieved by good petrographic analysis of SMIs and detailed petrographic information (size, shape, appearance, position relative to the host and other SMI within the same host). If the goal of a SMI5 study is to determine volatile concentrations, assessing the relative time of trapping among SMIs is of paramount importance, as it allows understanding if the variation in volatile concentrations is consistent with a certain magmatic physical-chemical process. Researchers that choose to work with SMIs face several challenges. SMIs are not always hosted in all samples or they can be difficult to recognize (mostly due to their small size). Additionally, due several processes, such as post-entrapment crystallization, boundary layer and sampling melt heterogeneity at the micron scale, they may not be representative of the original trapped melt composition. (C) 2015 Elsevier B.V. All rights reserved.