Feldspar phenocrysts of silicic volcanic rocks are commonly in Sr-isotopic disequilibrium with groundmass. In some cases the feldspar is more radiogenic, and in others it is less radiogenic. Several explanations have been published previously, but none of these is able to accommodate both senses of disequilibrium. We present a model by which either more- or less-radiogenic feldspar (or even both within a single eruptive unit) can originate. The model requires a magma body open to interaction with biotite- and feldspar-bearing wall rock. Magma is incrementally contaminated as wall rock melts incongruently. Biotite preferentially melts first, followed by feldspar. Such melting behavior, which is supported by both field and experimental studies, first contaminates magma with a relatively radiogenic addition, followed by a less-radiogenic addition. Feldspar phenocrysts lag behind melt (groundmass of volcanic rock) in incorporating the influx of contaminant, thus resulting in Sr-isotopic disequilibrium between the crystals and melt. The sense of disequilibrium recorded in a volcanic rock depends on when eruption quenches the contamination process. This model is testable by isotopic fingerprinting of individual feldspar crystals. For a given set of geologic boundary conditions, specific core-to-rim Sr-isotopic profiles are expectable. Moreover, phenocrysts that nucleate at different times during the contamination process should record different and predictable parts of the history. Initial results of Sr-isotopic fingerprinting of sanidine phenocrysts from the Taylor Creek Rhyolite are consistent with the model. More tests of the model are desirable.Feldspar phenocrysts of silicic volcanic rocks are commonly in Sr-isotopic disequilibrium with groundmass. In some cases the feldspar is more radiogenic, and in others it is less radiogenic. Several explanations have been published previously, but none of these is able to accommodate both senses of disequilibrium. We present a model by which either more- or less-radiogenic feldspar (or even both within a single eruptive unit) can originate. The model requires a magma body open to interaction with biotite- and feldspar-bearing wall rock. Magma is incrementally contaminated as wall rock melts incongruently. Biotite preferentially melts first, followed by feldspar. Such melting behavior, which is supported by both field and experimental studies, first contaminates magma with a relatively radiogenic addition, followed by a less-radiogenic addition. Feldspar phenocrysts lag behind melt (groundmass of volcanic rock) in incorporating the influx of contaminant, thus resulting in Sr-isotopic disequilibrium between the crystals and melt. The sense of disequilibrium recorded in a volcanic rock depends on when eruption quenches the contamination process. This model is testable by isotopic fingerprinting of individual feldspar crystals. For a given set of geologic boundary conditions, specific core-to-rim Sr-isotopic profiles are expectable. Moreover, phenocrysts that nucleate at different times during the contamination process should record different and predictable parts of the history. Initial results of Sr-isotopic fingerprinting of sanidine phenocrysts from the Taylor Creek Rhyolite are consistent with the model. More tests of the model are desirable.