The SNC meteorites are a class of eight basaltic achondrites that have extremely young crystallization ages (less-than-or-equal-to 1.3 Ga'). Chassigny, a unique SNC meteorite, is a dunite containing Fo68 olivines and rare poikilitic Ca-pyroxenes. It is one of the most primitive SNC meteorites, and thus is most likely to reveal information about the SNC basalt source region. This study presents a detailed examination of partially crystallized melt inclusions in cumulus olivine grains in Chassigny. These trapped melts are argued to be representative samples of the melt that existed when Chassigny crystallized. This melt has been modified, however, by interaction with the host olivine and by closed-system crystallization. The phases inside the melt inclusions include hydrous Ti-rich amphibole (kaersutite), biotite, two pyroxenes, and rhyolitic and alkali feldspar glasses. All phases have been extensively analyzed with an electron microprobe. The new mineralogical information is combined with a complementary experimental study of kaersutite/melt equilibria. This combined data set can be used to formulate a system of linear equations which can then be solved to determine the composition of the originally trapped melt. This calculation reveals that the trapped melt was an FeO-rich and Al2O3-poor basalt. The melt composition is shown to be consistent with the crystallization sequence both inside the melt inclusions and in the rock matrix. The major element chemistry of this liquid closely resembles terrestrial boninite lavas. The new data also allow the intensive conditions (temperature, total pressure, and water fugacity) of Chassigny crystallization to be estimated. Two-pyroxene geothermometry indicates that equilibration temperatures were approximately 1000 +/- 50-degrees-C, although amphilbole does not coexist with melt until T = 960-degrees-C. The trapped liquid initially contained 1.5 wt% dissolved water. Crystallization of anhydrous phases caused water to buildup in the trapped melt. Prior to kaersutite crystallization, the melt must have contained at least 4 wt% dissolved water, suggesting that a minimum of 1.5 kbar total pressure is required. A maximum total pressure cannot be inferred, but all available data are consistent with low pressure (much-less-than 5 kbar) crystallization. Finally, X(H2O) in the fluid in the melt inclusion is estimated to have been at least 0.8 in order to stabilize amphibole without plagioclase, implying that the water fugacity was approximately 1480 bars. The existence of hydrous melts and conditions appropriate for amphibole crystallization suggest that evolved, SiO2-rich lavas exist on the Chassigny parent body (Mars).
