Crater Lake caldera (8 × 10 km), formed 6845 years B. P. (¹⁴C age) during the climactic eruption of the volcanic edifice known as Mount Mazama, is intermediate in size between small calderas associated with central vent eruptions and large calderas that have ring fracture vent systems. Our quantitative study of lithic fragments in the ejecta confirms the existing model of changes in vent configuration during the climactic eruption of Mount Mazama. Initial activity was from a single vent that produced a rhyodacite pumice fall from a Plinian column. Altered preexisting volcanic rocks are the predominant lithic type in the Plinian deposit, and their extensive hydrothermal alteration is considered as evidence of their relatively deep origin. The Wineglass Welded Tuff lies atop the Plinian deposit and contains a higher proportion of fresh volcanic rocks, suggesting enlargement of the single vent by slumping of its walls. This same vent enlargement caused the Plinian eruption column to collapse and feed valley‐hugging pyroclastic flows that deposited the Wineglass Welded Tuff. When enough material was erupted from the shallow magma chamber that its roof was no longer adequately supported, Mount Mazama collapsed to form the caldera, while highly energetic pyroclastic flows produced the climactic ignimbrite. A lag breccia that represents the proximal facies of the compositionally zoned climactic ignimbrite lies atop the Wineglass Welded Tuff and contains predominantly altered volcanic rocks of deeper origin, accompanied by minor granitoids from the magma chamber walls. Azimuthal differences in lithic component proportions in the lag breccia correlate well with the geology of the caldera walls, indicating that the climactic ignimbrite was ejected by multiple vents along a ring fracture system. Systematic lithic component changes within the lag breccia suggest different quarrying levels that reflect waxing and waning of the discharge rate during the caldera collapse phase of the climactic eruption. Our lithic component analysis demonstrates that calderas that may be too small to experience structural resurgence, such as Crater Lake, nevertheless may form by syneruptive subsidence along ring fractures.