For the most part, MI are absent in granitic rocks because their prolonged cooling history allows complete crystallization of the inclusions, erasing any record of their existence (Tuttle 1952). Only in rapidly cooled intrusions will MI remain intact and unaltered. In the ~280-Ma Mount Genis granite in southeastern Sardinia, Frezzotti (1992) found MI (some glassy) within equigranular granites as well as in crystals lining miarolitic cavities. Coexisting brine was interpreted as having formed during the late stages of crystallization of the magma. The MI were analyzed by EPMA, and several of them contained typical high-silica rhyolite glass with Cl concentrations reflecting equilibrium with vapor and hydrosaline melt (Shinohara 1994). Hansteen & Lustenhouwer (1990) examined MI and coexisting aqueous inclusions from the Permian Eikeren-Skrim granite in Norway. The inclusions homogenized at 685-705 °C, were composed of rhyolitic and rhyodacitic glass (not peralkaline) and were trapped at the same time as a saline aqueous fluid. Eadington & Nashar (1978) described and analyzed glass inclusions from topaz-quartz rocks from the New England district of New South Wales in Australia. Many of the MI contained several wt.% F, as well as apparent high H2O concentrations. Weisbrod (1981) reviewed several studies of coexisting fluid and melt inclusions in shallow intrusions and porphyry deposits. Takenouchi & Imai (1975) described MI from a variety of porphyry and volcanic rocks and observed the effects of cooling history on the characteristics of MI.
Occasionally, intrusive rocks are ejected as xenoliths during volcanic eruptions. Studies of such samples have often resulted in discovery of CO2-rich fluids in equilibrium with melt in the environment where the xenoliths grew (Roedder 1965; Belkin et al. 1985; Frezzotti et al. 1991; Belkin and De Vivo 1993). Other xenoliths show evidence for immiscibility between silicate melt and hypersaline fluids (Roedder & Coombs 1967) or entrapment of both phases, though at different times (De Vivo et al. 1993).
Silicate MI contain information on the dissolved volatile concentrations in igneous rocks (Tables 2 and 3). A variety of analytical and thermometric methods can be used to extract information from MI, as regards magmatic volatile concentrations, the compositions of exsolved magmatic fluids and the pressure and temperature conditions under which magmas undergo crystallization. New techniques promise to increase the number of possible uses of MI, lengthening the first half of Table 1 and shortening its last half. For example, given improving microbeam techniques, MI may contribute new data on the stable-isotopic (H, O, S, Cl) composition of non-degassed magmas, and the effect of degassing on the isotopic composition of silicate melt. New microbeam dating methods may allow assessment of the time delay between crystallization (MI formation) and eruption. Combined with the ever-increasing data-set on volatile solubilities, MI may yield more reliable estimates of the depths to magma chambers than are currently available through mineral geobarometers. As such, MI are likely to remain one of the most useful and reliable methods for understanding the behavior of volatile components in igneous systems and will continue to provide insight for economic geologists, volcanologists and other geoscientists.
Discussions over the past 7 years have helped me to clarify my thinking about MI; I thank A.T. Anderson, C.R. Bacon, D.K. Bird, P. Fiske, G.A. Mahood, S. Newman, E. Roedder, H. Shinohara, T. Sisson, J. Stebbins, and P.J. Wallace for their input. The rest of my education about MI has been the legacy of excellent papers by A.T. Anderson, E. Roedder, R. Clocchiatti and others, as well as the interesting samples I've had the good fortune to study. I appreciate reviews of the manuscript and helpful comments by C. Bacon, K. Bargar, H. Belkin, B. Bodnar, J. Hedenquist, H. Shinohara, and S. Simmons.