The stone roof-plate, enriched with mouldings and foliage, and, perhaps, supported on sculptured corbels, becomes the crowning horizontal feature; and the parapet--the defence of the workmen engaged on the roofs--is pierced into tracery, or forms a miniature arcade, giving delicacy and lightness of effect to the generally massive structure; while the bell-tower, raised high to make its voice heard from afar, becomes the culminating ornament of the whole exterior. So completely was it the recognised principle of the architecture to render the useful and constructive parts sources of decoration, that, where any deliberate decoration was made use of, it was often formed of imitations of constructional features, such as window tracery, arcades, gables, pinnacles, columns, etc. I am not prepared to say that this is in itself to be applauded; indeed, I think it ought, at the least, to be kept within moderate limits; but it nevertheless owed its origin to the firm hold which the principle of rendering construction the leading source of decoration had upon the architects. Being accustomed to decorate construction, they got into the habit of using constructive forms as decorations. My illustrations have hitherto, perhaps, for the most part, been taken from churches, but the same principle of common sense applies equally to secular structures. Each is treated in a manner suited to its class and purpose. Those classes and purposes differ, as a matter of course, in a majority of cases, from their correlatives at the present day, as they did in different periods of the Middle Ages themselves, and in the different countries of Europe, at any given period; so that the mere fact of such differences existing is no argument against any lesson we may learn from them. I presume, for example, that no great analogy can be established between a Roman villa and one of the nineteenth century in England, and not much between an Italian Renaissance palace of the fifteenth century and a London mansion of the nineteenth. Even in Germany and in France at the present day the houses differ greatly from those in England. The question of the _rationale_ of a style is rather whether it is so flexible and so essentially founded on common sense and reason that it will readily shape itself to meet practical demands, however varied they may be. Now, it is scarcely possible for a building of the Middle Ages and one for a kindred purpose at the present day to differ more widely in their requirements than did different buildings of the same age; and if the most varied demands of one period are equally met by a given style, why should we fear that the same style would fail to meet variations proceeding from a change of manners and habits?