Alteration of undisturbed igneous material used for argon dating work, is the most common cause of incorrect (low) estimates of the time of crystallization. Identification of alteration has relied on qualitative and subjective (optical) methods. For ⁴⁰Ar/³⁹Ar dating, I introduce a new parameter—the alteration index (A.I.)—to quantitatively assess alteration. This looks to the quantity of ³⁶Ar (atmospheric argon) released in such studies. A non-dimensional parameter is used, relating the ³⁶Ar levels to that of ³⁹Ar for K-rich phases (K-feldspar, biotite, whole-rock basalt), and to ³⁷Ar for Ca-rich phases (plagioclase feldspar and hornblende). Water contains large amounts of dissolved argon derived from the atmosphere. During chemical weathering, ³⁶Ar carried by water is introduced into the silicate phases of rocks. All common alteration minerals contain water; their ³⁶Ar contents are ∼100–1000 times higher than in anhydrous silicate phases. Incipient alteration, undetected by current tests, is unequivocally recognized by the A.I. method. In ⁴⁰Ar/³⁹Ar stepheating studies, the plateau steps (if any) release argon from the least altered sites. The A.I. of plateau steps for fresh, subaerial, material yields the cut-off value for detecting alteration. Partial loss of ⁴⁰Ar* from altered samples may result in statistically acceptable plateaus that underestimate the true crystallization age by ∼2–10%. Many ages are invalid as accurate estimates of the age of crystallization (a) based on statistical analysis of the apparent ages on plateau/isochron plots and/or (b) ages derived from altered phases within the sample. At subduction zones, the hydrated slab cycles substantial quantities of (atmospheric) argon into the mantle. Monitoring ³⁶Ar levels in fresh (mafic and intermediate) rocks should serve as a sensitive tool in elucidating the role of water driven off the subducted slab in triggering magmatism in convergent zone settings.