The Mogollon-Datil Volcanic Field in southwest New Mexico was active between 40 and 20 Ma. It is situated c. 500 km from the western North American plate margin, and yet the early pre-30 Ma volcanic rocks comprise a high-K calc-alkaline suite ranging from basaltic andesites to dacites. All the samples are characterized by high and variable LILE and LREE contents, and LILE/HFSE ratios, and variable Nd and Sr isotope ratios of 0.51231–0.51250 and 0.7055–0.7065 respectively. In detail, three groups are recognized: Group 1 rocks have the lower SiO2, incompatible element abundances and 87Sr/86Sr ratios; Group 2 preserves the clearest evidence for open system differentiation, and Group 3 exhibits little variation in 87Sr/86Sr and the most coherent major and trace element arrays. The Group 2 rocks have been modelled in terms of simple binary mixing between selected Group 1 basaltic andesites and a higher silica crustal endmember, which appears to have been derived from the mid-lower crust in the presence of residual garnet. The Group 3 rocks, by contrast, are consistent with near closed system differentiation, modelled in three stages and involving Plag-Aug-O1-Mag and subsequently hornblende and K-feldspar. Published experimental data indicate that the differentiation of the parental basaltic magmas took place at elevated pressures of c. 4-5 kbar, and water contents, c. 2wt% H2O. It is argued that the parental magmas to these calc-alkaline suites were generated in a period of lithospheric extension, and that they were derived from within the continental mantle lithosphere, as inferred for mid- to Late Tertiary rocks elsewhere in the western US, with no significant contribution from contemporaneous subduction processes.