I’ve discovered an error in the Glass Density Calc spreadsheet, posted on the Tools for Petrologists page. The original version of the spreadsheet, used to calculate the density of a volcanic glass or magma, contained incorrect values for the partial molar volumes of the oxide components and incorrect corresponding dV/dT values. The new version (version 2) has the correct values and has been checked against other density measurements in the literature and found to be within about 10% error.
Please update to the newest version: available here.