We examine the problem of the interaction of the vapor plume produced by the Cretaceous/Tertiary (K/T) extinction impact with the overlying atmosphere. This process is important for evaluating models involving the production of acid rain and the distribution of Ir and associated elements in the K/T boundary layer. We use the Zel'dovich and Razier model for the expansion of the vapor plume produced by both asteroidal and cometary impactors, for impactor masses from 10¹⁰ to 10²⁰ kg and impact velocities from 15 to 60 km/sec. This mass range corresponds to diameter ranges of roughly 200 m to 400 km (asteroids) or 500 km (comets). We balance the momentum of the expanding gas against the mass of the gas plus overlying atmosphere to find the mean velocity of gas plus atmosphere; if this velocity exceeds Earth's escape velocity, we assume that both impact-generated gas and atmosphere are lost from the planet. We estimate the amount of atmosphere lost and the amount of the impactor retained by the Earth, the latter under the assumption that the impactor material is concentrated in the outer portion of the vapor plume. Significant loss of shocked atmosphere limits the efficiency of acid-rain production with increasing impact energy, and the amount of projectile material retained allows us to constrain the minimum-size impactor required to deposit the observed Ir. For sufficiently energetic impacts, all of the atmosphere lying above the plane tangent to the Earth at the point of impact and all of the impactor are lost from the Earth. We conclude that the K/T impactor was most likely an asteroid ≥ 10¹⁴ kg in mass; data that suggest that acid rain was an important phenomenon imply that the asteroid was ≤10¹⁸ kg in mass, or else all the shocked atmosphere would have been blown off. Higher energy impacts would presumably also have disastrous consequences for life on Earth, even though they may neither produce as much acid rain nor leave behind a siderophile signature in a boundary layer. Thus the lack of Ir enhancement at a mass extinction horizon does not necessarily rule out an impact as the cause of the extinction. Lastly, the result that sufficiently energetic impacts can result in a net loss of volatiles from the Earth complicates models in which much or all of the Earth's volatile inventory was delivered by the impact of comets.