In order to model the composition of the fluids contributing to subduction-related magmas in Costa Rica and Nicaragua, we present new melt inclusion (MI) data for several volcanoes in Central America.  In addition, we have taken advantage of a new technique to calculate magmatic water contents from nominally anhydrous minerals, such as clinopyroxene. The clinopyroxene-predicted water contents have proved useful in filling in data gaps where olivine-hosted MI are unavailable, and volatile data is therefore lacking.  The data do not support the use of Ba/La (in melts) and Sr/Nd (in clinopyroxene) as simple water proxies.  Instead, at least two fluids are at work beneath SE Central America.  The first fluid, dominant in magmas from the southeastern end of the arc (e.g. Costa Rica) has high Cl/H₂O, high d¹⁸O, low Ba/La and Sr/Nd.  It may be derived in part from high-Cl, low-Ba volcaniclastic sediments associated with the Cocos Ridge.  The second potential subduction fluid, dominant in the northwestern end of the arc (e.g. Nicaragua) has low Cl/H₂O, low d¹⁸O, and high Ba/La and Sr/Nd.  It may be derived, in part, from the breakdown of serpentinite in the down-going plate, as well as overlying pelagic sediments.  The potential of combining clinopyroxene-derived H₂O and trace elements with an expanded MI dataset creates a new and powerful tool in igneous petrology and the modeling of the fluids which generate subduction zone magmas.