Not all calderas conform to the currently favored model, in which a cylindrical block subsides as in cauldrons of deeply eroded volcanoes. Some calderas are downsagged structures, a fact deduced mainly from the backtilting of ignimbrite sheets assumed originally to have possessed an outwardly directed slope. Few calderas have postcaldera vent rings; postcaldera vents more commonly lie on a transverse line, occupy a central position, or are scattered, indicating that any possible caldera ring faults were not sufficiently significant lines of weakness to guide uprising magmas. About as many volcanoes have vent rings outside the caldera as inside and vent rings are known also where caldera faults are lacking. These and other lines of evidence suggest that vent rings on intermediate or silicic volcanoes are more likely to be situated on cone sheet-type fractures, caused by upwardly directed magmatic pressure and related to updoming, than on ring faults related to caldera subsidence. The prevailing view of caldera formation is a catastrophic one, but some calderas grow incrementally in response to a succession of moderate-sized eruptions. On some volcanoes having a small caldera, the present caldera may be merely the youngest of a succession generated by relatively small-magnitude events. Attention is drawn to the fact that cauldrons and ring-dike complexes are a feature of intracontinental cratonic volcanoes. It is significant that the anomalously large Tertiary calderas of the Basin and Range Province, which conform more closely to the currently favored caldera model than do the calderas in epicontinental settings, also cut Precambrian crust.