science, the materials and principles of earlier ages. X. But whether this be granted me or not, the fact is, that every idea respecting size, proportion, decoration, or construction, on which we are at present in the habit of acting or judging, depends on presupposition of such materials: and as I both feel myself unable to escape the influence of these prejudices, and believe that my readers will be equally so, it may be perhaps permitted to me to assume that true architecture does not admit iron as a constructive material,[7] and that such works as the cast-iron central spire of Rouen Cathedral, or the iron roofs and pillars of our railway stations, and of some of our churches, are not architecture at all. Yet it is evident that metals may, and sometimes must, enter into the construction to a certain extent, as nails in wooden architecture, and therefore as legitimately rivets and solderings in stone; neither can we well deny to the Gothic architect the power of supporting statues, pinnacles, or traceries by iron bars; and if we grant this I do not see how we can help allowing Brunelleschi his iron chain around the dome of Florence, or the builders of Salisbury their elaborate iron binding of the central tower.[8] If, however, we would not fall into the old sophistry of the grains of corn and the heap, we must find a rule which may enable us to stop somewhere. This rule is, I think, that metals may be used as a _cement_ but not as a _support_. For as cements of other kinds are often so strong that the stones may easier be broken than separated, and the wall becomes a solid mass without for that reason losing the character of architecture, there is no reason why, when a nation has obtained the knowledge and practice of iron work, metal rods or rivets should not be used in the place of cement, and establish the same or a greater strength and adherence, without in any wise inducing departure from the types and system of architecture before established; nor does it make any difference except as to sightliness, whether the metal bands or rods so employed, be in the body of the wall or on its exterior, or set as stays and cross-bands; so only that the use of them be always and distinctly one which might be superseded by mere strength of cement; as for instance if a pinnacle or mullion be propped or tied by an iron band, it is evident that the iron only prevents the separation of the stones by lateral force, which the cement would have done, had it been strong enough. But the moment that the iron in the least degree takes the place of the stone, and acts by its resistance to crushing, and bears superincumbent weight, or if it acts by its own weight as a counterpoise, and so supersedes the use of pinnacles or buttresses in resisting a lateral thrust, or if, in the form of a rod or girder, it is used to do what wooden beams would have done as well, that instant the building ceases, so far as such applications of metal extend, to be true architecture.