"There is no article," says the learned Ware, "in the whole compass of the architect's employment, that is more important, or more worthy of distinct consideration, than the roof; and there is this satisfaction for the mind of the man of genius in that profession, that there is no part in which is greater room for improvement"

The suggestions above quoted, although made in the year 1756, remained quite unheeded till near the close of the last century, when Mr. Peter Nicholson made public the germ of an invention, which has, in process of time, brought about as great a revolution in the art of carpentry as the introduction of the arch did in that of masonry. The leading feature of the invention is the substitution of iron rods for wooden king and queen posts. The design by Mr. Nicholson was published in 1797; but, as late as 1828, Mr. Tredgold says, in his excellent Treatise on Carpentry, "It has been proposed to let the ends of the principal rafters abut against each other, and to suspend the king-posts by straps of iron; but a piece of good carpentry should depend as little on straps as possible." From the tenor of his remarks, it is reasonable to suppose that few, if any, successful experiments had been made; for he afterwards refers the reader to a design in his work, where the rafters abut against each other, and the beam is suspended by planks bolted to their sides. "This method," he adds, "is perhaps the best in use." A valuable standard work, entitled "Treatise on Architecture, Building, etc," was published in Edinburgh in 1844. On page 154 of the work, attention is called to the suggestion of Mr. Nicholson, made forty-seven years before. The writer (Dr. Thomas Young) says, "There is a very ingenious project offered to the public by Mr. Nicholson ('Carpenter's Assistant,' p. 68). He proposes iron rods for king-posts, queen-posts, and all other situations where beams perform the office of ties. . . . We abound in iron; but we must send abroad for building-timber. This is, therefore, a valuable project. At the same time, however, let us not overrate its value." From the foregoing, it appears, that, up to a late day, but little advance had been made; the old methods of construction being looked upon with more favor than the new.

At what time, or by whom, the idea was first practically carried out in this country, is uncertain. The burden of evidence, however, indicates, that, although first published by Mr. Asher Benjamin, he was indebted for the suggestion to Mr. Charles G. Hall, now of Roxbury, Mass. Mr. Hall, an Englishman by birth, and an able architect and engineer, arrived in America in 1823. He soon became associated in business with an eminent architect of that day, Mr. Alexander Parris. Under the direction of these gentlemen, many large and important buildings were erected, in the roofs of most, if not all, of which, the principle under consideration was employed. Being thus freely used, it soon commended itself to the judgment of other architects, who in turn adopted it; and the work of Mr. Benjamin was no sooner published than a reform commenced, which has steadily advanced, until its great value and economy are universally acknowledged.

On the plates of this work pertaining to roofs are designs calculated for various spans, and of such rise, or pitch, as will accommodate them to any style of building; each having been so designed that timber may be used of such dimensions as will properly support a covering of heavy slates.

Of the inclination, or pitch, of the several roofs, but little need be said, since designs for buildings are so varied, that an attempt to illustrate them all would-only encumber the work. A few suggestions will be made, which, together with those amendments naturally presenting themselves in particular cases, will give all the information required.

The pitch of any roof, covered with shingles or slates, should never be less than one-fourth the width of the entire span; for, if it be less, rain and snow will, in severe storms, be driven through the crevices. If the design of the building demands less inclination, a covering of tin, copper, lead, or something of like nature, should be used; in which case, any rise above a twenty-fourth of the whole span will be all that is required. The extent of the span will, however, to a certain degree, govern the inclination and form of the roof, in order to give strength to the truss. If the span is great, and a low roof is desired, it is best to truncate it, as shown on Plate IX., Fig. 1. Where slating is used, the boards should be matched, and planed to a uniform thickness; for if the joints are left open, as may be allowed in shingling, the passage of air through the openings carries with it rain or dry-snow, when, in ordinary storms, it would exhibit no sign of defect.