The structural merit of this scheme lies in what it has derived from the forms and adjustments of the Pantheon. Its weakness consists in the increased elevation, lifting the dome away from its abutment to such an extent that it may be questioned whether it could have been made safe without chains. The thrusts of a hemispherical dome are vastly more powerful than those of a vault of pointed outline, like the dome of Florence, but if properly abutted, as in the Pantheon,[43] it is perfectly safe, and makes a better ceiling than a pointed vault. In reducing the efficiency of his abutment by raising the springing of the dome so high, the architect ought to have diminished the force of its thrust in a corresponding degree by giving it a pointed form. This would have made it more safe, but it would have been inconsistent with the classic Roman models to which he was striving to conform. As for Bramante’s intended architectural treatment of the rest of the building we have, as before remarked, no precise information. It appears, however, most probable that he meant to cover the arms of the cross with barrel vaulting on massive square piers and arches, with a single order of pilasters such as Alberti had used in St. Andrea at Mantua, and such as Michael Angelo actually employed, though in a way peculiar to himself, and probably unlike that in which Bramante would have done it. For Bramante would, I think, have followed Alberti’s example in raising the order on pedestals--the great scale of St. Peter’s especially calling for such treatment. Bramante would have realized that a single order large enough to rest directly on the pavement and allow the entablature to pass over the crowns of the arches of the great arcades, would dwarf the apparent scale of the whole interior, as Michael Angelo’s order actually does. But whatever his intention was as to these details, Bramante died before they could be carried out, and we are left in the dark as to what the church as a whole would have been had he lived to complete it. To the work of his numerous successors prior to the appointment of Michael Angelo, we need give no attention because their labours did not materially affect the final result. Their work was largely on paper only, and the building as it now exists is substantially Michael Angelo’s design, based on that of Bramante, but with extensive, and damaging, additions by subsequent architects. [Illustration: FIG. 26.--The model.]