Thellier-type paleointensity experiments were conducted on welded ash matrix or pumice from the 1912 Novarupta (NV) and 1980 Mt. St. Helens (MSH) pyroclastic density currents (PDCs) with the intention of evaluating their suitability for geomagnetic paleointensity studies. PDCs are common worldwide, but can have complicated thermal and alteration histories. We attempt to address the role that emplacement temperature and postemplacement hydrothermal alteration may play in nonideal paleointensity behavior of PDCs. Results demonstrate two types of nonideal behavior: unstable remanence in multidomain (MD) titanomagnetite, and nonideal behavior linked to fumarolic and vapor phase alteration. Emplacement temperature indirectly influences MSH results by controlling the fraction of homogenous MD versus oxyexsolved pseudo-single domain titanomagnetite. NV samples are more directly influenced by vapor phase alteration. The majority of NV samples show distinct two-slope behavior in the natural remanent magnetizationpartial thermal remanent magnetization plots. We interpret this to arise from a (thermo)chemical remanent magnetization associated with vapor phase alteration, and samples with high water content (>0.75% loss on ignition) generate paleointensities that deviate most strongly from the true value. We find that PDCs can be productively used for paleointensity, but thatas with all paleointensity studiescare should be taken in identifying potential postemplacement alteration below the Curie temperature, and that large, welded flows may be more alteration-prone. One advantage in using PDCs is that they typically have greater areal (spatial) exposure than a basalt flow, allowing for more extensive sampling and better assessment of errors and uncertainty.