The vault over the crossing at Pisa cathedral is elliptical in plan and has an exposed exterior profile. These characteristics, unique in 11th-century Tuscany, have caused scholars to doubt that the dome could have been built as part of the original cathedral fabric, which was largely complete by 1118. Yet Sanpaolesi's recent discussion of the problem, based on his study of the cupola during its restoration in 1957, presents substantial evidence in support of an early date (1090-1100) for the squinches and tambour below the cupola proper, and he argues convincingly that the cupola itself must share this date. Acceptance of his conclusions, however, does not necessitate assuming that the cupola was foreseen when the cathedral was planned in 1063, or that its original appearance conformed to what we see today. Indeed, I will argue that the unusual shape of the cupola is the inevitable consequence of a decision to vault the crossing, taken after the foundations of the cathedral were already laid. Further, I will present evidence which suggests that the exterior profile of the dome was not exposed originally, and has only been exposed since the late 14th century. Finally, I will show that in its original form the cupola was closely related to contemporary Lombard structures.