Reaction between alkali gabbro magma and limestone produced a 0.2- to 3-m-thick zone of nepheline pyroxenite and a 0.1- to 1-m-thick band of calc-silicate skarn at their contact. The sinuous embayed pyroxenite-skarn contact, the presence of rounded skarn xenoliths in pyroxenite, and textural evidence for growth of calc-silicate skarn by replacement of both marble and solid pyroxenite indicate that reaction involved assimilation of carbonate wall rock by magma and loss of Al and Si to the skarn. The uneven modal distribution of euhedral titanaugite and enveloping nepheline in pyroxenite, the restricted occurrence of nepheline syenite as dikes in pyroxenite and skarn and the mineralogical similarity of nepheline syenite and the leucocratic matrix of pyroxenite suggest that pyroxenite represents an accumulation of titanaugite cemented by an alkali-rich residual magma and that nepheline syenite represents a part of the residual contaminated magma that was squeezed out of the clinopyroxene crystal mush. Limestone assimilation is modeled by reaction of calcite and magmatic plagioclase, which results in resorption of plagioclase, growth of clinopyroxene enriched in Fe, Ti, and Al, and solution of nepheline and wollastonite in residual contaminated magma. The bulk composition of pyroxenite evolved by addition of Ca from dissolved limestone, loss of Al and Si to skarn, and local segregation of solid clinopyroxene and nepheline syenite magma. The predominance of pyroxenite among contaminated rocks and their restriction to a narrow zone along the intrusive contact provide little evidence for the genesis of a significant volume of nepheline syenite magma by limestone assimilation.