[20] It may be thought that this would condemn the use of metal clamps in masonry, such as were inserted in the walls of the Parthenon, or the wooden ties that were, in some cases, used in parts of Gothic buildings. But there is a wide difference between such use of clamps and ties, and the binding chains of the great domes of the Renaissance. In the Greek and Gothic work the masonry forms are favourable to stability independently of the clamps and ties. These were inserted either for security against unusual dangers, as from earthquakes, or for temporary security against rupture while the work was in progress, before the interaction of the parts of the system was fully established; but a dome without abutment violates the constant conditions of stability. [21] _Discrizione e Studj dell’Insigne Fabbrica di S. Maria del Fiore_, Florence, 1733, p. xxi. [22] _Memorie degli Architette_, etc., Florence, 1785, vol. 1, p. 190. [23] Fontana, Nelli, Cecchini, and others. [24] These ruptures were first observed in the year 1693 (Nelli, _op. cit._, p. 13), and it was then advised by the architect Carlo Fontana to add a new chain of iron. Nelli, however, argued that the fissures had not arisen from thrust, but were due to a slight yielding of the foundations, and he urged that no chain be added, but that a bit of marble be dove-tailed into the vault across the opening, in order that any further movement might be detected by the breaking of this marble. For three years no further sign of disturbance was noticed, but a slight earthquake in 1697 broke a portion of the masonry of the outer face of the dome opposite the fissure across which the marble had been placed. It appears, however, to have been concluded that there was still no danger from thrust, and no new chain was inserted. Cecchini (_Opinione intorno lo Stato della gran Cupola del Duomo di Firenze_, published together with Nelli’s _Discorsi_, etc., p. 82) speaks of several cracks in both the inner and the outer shells of the vault, and also in the supporting piers, even down to the ground. But he agrees with Nelli in attributing these to movements of the foundations from which he concludes that no further danger is to be apprehended, and he affirms that the structure is entirely safe. [25] Cf. Nelli, _op. cit._, p. 73. [26] The thrusts of a hemispherical dome are, in some degree, restrained by the binding of its continuous courses of masonry under compression, but this is not enough for security, as experience has shown; and in a polygonal dome, like Brunelleschi’s, there is no such binding force, because there are no continuous circles of masonry.