perpendicular style would not, in many cases, lead us to prefer wood to stone, for the internal covering, especially where the form of the depressed arch is predominant. Almost every pattern of timber roof which antiquity presents to our notice, has been applied to buildings of that age; and with those who profess an adherence to principles, this should be a powerful argument in favour of the perpendicular style. The examples of it are most numerous, from the cathedral to the humble village church; and it is certainly the most manageable of all the styles.

We now come to the modern trussed roof. It might be interesting to trace the progress which this kind of roof made, from its origin to the present time, when it is exhibited, in every conceivable form, and of every available material, especially in railroad architecture. But a few examples are all which will be required to illustrate our subject. A question may here be asked, which involves principles belonging to architecture generally, and may, therefore, very well affect particulars; namely, whether it be right to exclude from church architecture, every useful discovery or invention, whether in the materials or construction, or in the method of applying them to the purposes of building, which the experience of three centuries may have brought about.

And this suggests another, somewhat more practical-whether it be not as objectionable even on the ground of architectural propriety, in an eminently constructive department, such as that of the roof, to try how much useless timber we can put into a building by way of ornament, as to endeavour to ascertain the smallest quantity that will be sufficient for the purpose, and build accordingly. The question may be put in a variety of ways, but whatever be the proper answer, it cannot be denied that the system lately introduced, of copying the old high pitched open roof, requires that we should increase the solidity of our walls, merely to make them bear a heavier description of roof! If this be really necessary, must it not serve as an argument against the use of a style that requires such a sacrifice ?*

But to return to the tie-beamed roof. The principle of the flattened timber roof is here brought into operation, only with this difference, that in the place of a solid beam in one length, we have a beam formed of many pieces, and the ingenuity of the builder is exercised in placing these pieces to most advantage. It becomes in fact a large piece of frame-work. The whole beam thus constructed These are made to rest upon the

is technically termed a truss-beam.

* It has been said that the average increase of the population of the whole kingdom, is one thousand a day! This is unhappily not so far from the truth as some will perhaps suppose. But set down the increase at only half that, what are our means to meet the demand for more churches? We recommend this to the serious consideration of those who profess, and whose real object, we doubt not, is to build churches to the honour and glory of Almighty God. But see further, Christian Remembrancer, vol. v. page 89, and vol. iv. page 265. Article, "Styles of Church Architecture."

walls at intervals, and so to bear up the covering laid upon them. The comparative strength and lightness of this kind of roof is enormous, and when contrasted with the old open roof, the difference of weight between them appears almost incredible. Take, for example, one section of Westminster hall and one of the theatre at Birmingham; the span of the latter, clear, between walls, is eighty feet. Compare the weight of the material of the latter with that of the former, and it is as nothing; not to mention the light upright walls of the one and the bulging buttressed abutments of the other. It is not here recommended, God forbid! to substitute in all cases, and in every particular, the former for the latter; but only, that by bringing these matters before the notice of amateurs, they may see what an amazing mechanical power they are spurning from that architecture, which ought to be, were it possible, perfect in every point of view. The truss-beam can be accommodated to any pitch, from the highest, to one which is not much steeper than the flattened roof of the fifteenth century. In favour of the old roofs, it may, however, be observed (as regards their constructive features, and not taking the appearance of either into account,) that they are better able to contend against neglect and other casualties, than the modern trussed roofs. The former owe their strength to the solidity and profusion of their parts; the latter to the form and position of theirs, on which every thing must depend. But as the one could be repaired as easily as the other, the old roof has no very decided superiority over the modern in this respect. Again; in churches with aisles, the thrust of the old high-pitched roof can be partially brought down within a few feet of the aisle roofs, which contribute greatly to increase the abutment required, and render a lighter wall sufficient. But these are small gains for the cost. No provision is made against the contingencies of defective masonry, settlements, insufficient abutments, and the like. The walls, instead of being held together as they would be by the modern truss-beam, are in continual danger of being pushed outwards, should any part of the building, which helps to form the abutment, fail. This is not the case with the old flattened roof, which, if constructed of sound well-seasoned oak, of sufficient scantling, would bid fair to last as long as the materials of which it is formed are capable of lasting. For small churches having aisles, it might still be found a convenient form of roof; and in the hands of an architect of genius and rcal taste, might serve to exemplify the capabilities of the later Tudor style. In large churches, the truss-beam might take the place of the massive oak-beam, and the inclination be thereby made sufficient to enable slate to be used instead of lead. Much would depend upon the taste and skill of the architect, in making his roof harmonize with the rest of the building. In nine churches out of ten which are now in progress, it becomes almost impracticable to cover the nave or the chancel with the old flattened roof: in these instances, therefore, the truss-beam might take the place of the old solid beam, and the attention of the architect would then be directed towards making it both effective

and efficient. Although we do not profess to be designers of roofs, or indeed of any other portions of an edifice, whether ecclesiastical or domestic, we may still venture to suggest the principles upon which, as we think, they ought to be constructed. They should be what they appear to be, and not imitations of other kinds of roofs. The constructive features should be fully developed, and made the vehicles of decoration. Every piece should be clearly seen to do its work; and therefore no appended spandrils, or other curved pieces, should be admitted, except in entire subordination to the main features of the frame-work, and that only by way of ornamental filling up, should such decoration be thought advisable. Of course, the architect will keep in view the character of the building upon which he is engaged-that it is a church, and not a rail-road station; and it would be for him to consider well which material would be the more suitable, wood or iron, for ties and posts (king or queen posts). For many reasons we should give the preference to wood for such purposes; but as iron for such minor purposes as bolts, plates, stirrups, &c. is now so generally adopted, (in preference to the old mortise and tenon,) it might possibly be made in these the means of decoration, as it used to be in the hinges and fastenings of doors. In that case it would appear as iron, and not be painted that it may resemble the wood-work, as is usually done. Perhaps iron might be found available in many ways, both in the constructive and decorative features of the roof. As far as the appearance of such roofs is concerned, the main difficulty to be overcome seems to us to be, the extreme tenuity and general absence of massiveness in their several parts, as contrasted with those of the old styles. When we consider that a roof of this description would be something considerably less than half the weight of the old untied roof, it is evident that it must assume a new character. The old roof abounded in timber, and rose from the side as well as from the top of the walls which carried it. The hammer-beams and arches are often of vast proportions, and the interval between them and the principals are usually filled up with tracery or panelling. All this gives to this kind of roof an appearance of massiveness which the trussed roof could not have. To carry out the principle of lightness, the interval between the truss-beams must be short; but if the appearance of a stronger frame be desired, as it probably would, then the purlins might be trussed from the bottom of the queen posts, without the intervention of subordinate principals. Again, as the tie-beam would have nothing but itself to carry, it might be considerably reduced in bulk, in the interval between the queen posts. Even an iron rod of sufficient strength might do in that position, and then what remained would have the appearance of hammer-beams. In short, the ways and means of constructing such roofs, would be found to be at least not more limited than in any of the old styles of roofing. We do not, however, go so far as to affirm, that all the hints we have given might be advantageously acted upon; but having an opinion of our own as to the practicability of these roofs for churches, we have ventured to give it in detail, in

the hope that when the rage for imitating the performances of the mediæval builders is somewhat abated, there may be found those who will be willing to make the experiment.

It will be said, perhaps, of the diagrams given at the end of this article, the models of the roof here recommended, that they are those of ceiled roofs, which were never intended to be seen from below, the builder's sole aim being to construct a strong, not a beautiful roof. Granted so far. But towards this most desirable end, one step is already gained in the mechanical power therein exhibited. Taken by themselves, the forms are not otherwise than pleasing; the constructive features are prominent, and the intent of every part is soon perceived. This never fails to produce a satisfactory impression on the eye of the beholder. But let us give a little further examination to the objection. In some of the earliest wood roofs that remain, the frame-timbers are carefully concealed by boarding. That of the nave of Peterborough cathedral is a wellknown Norman example; and the old compass roof, such as we may suppose to have been in very general use in the thirteenth century, was usually boarded underneath the frame, and the construction of the roof thereby concealed. Examples of this may be seen in the choir of the chapel and in the library of Merton College, Oxford, in the old church at Yarmouth, and in some others. But these were the parents of the open roof of a later age. The picturesque effect of a frame so constructed was observed in the course of time: the boarding was omitted, the arrangement of the parts by degrees more and more accommodated to architectural effect, and by the close of the next century the style became thoroughly elaborated in the stupendous roof of Westminster Hall. Certainly the resemblance between roofs of that style and the old compass roof, is not by any means a perfect one; but that the latter was the parent of the former will not be denied: and why should not the same success attend the adoption of a more scientific method of construction? The objection we think sufficiently answered in the success which followed upon the removal of the boarded ceilings of the compass roof. Perhaps we shall here be met with another objection to our theory, namely, that the kind of roof we are advocating has been already tried, but without any success; and we shall be referred to our own account of the matter at the commencement of this article. Now, it is not exactly to the purpose, to say that because a scheme has been tried by one set of men and failed, the same fate must necessarily attend it in the hands of another. If they have failed it is because in endeavouring to accommodate the trussed roof to pointed architecture, they have rather sought to imitate the features of the old open roof, than to develop in Gothic characters the principle upon which the modern roof is constructed.

It must be confessed, however, that to mature such a scheme, or even to make a good beginning at the present time, is almost hopeless. The tide of fashion sets just the other way, and our architects are now fully occupied in calculations of the weight, not the strength of timber, and the capability of brick clerestories to resist lateral

pressure. They will not therefore like to go back again to their former work, and reconsider it, with a view to its improvement. But we cannot tell for how short a period the taste which now prevails may last. Just now, the peculiar merits of the masonry of the thirteenth, and the carpentry of the fifteenth, centuries, engross all our attention by-and-by we shall, perhaps, become tired of that occupation, and look for faults. To assert, or to attempt to prove, that the mediaval architecture is defective in principle is not our object. Indeed, many of the productions of that era prove the contrary; but there are instances, and those not few, which lead us to think it probable that the recognised principles of construction were not always fully understood by those who ought to have been guided by them. In most cases, where failure has taken place in the construction, it must be attributed either to this, or to their system of building; very possibly to both. The length of time which many of their works have stood, bear witness to the solidity of the workmanship, and the skill of the builders; and that some failures should have occurred in the lapse of centuries is not to be wondered at. But, upon examination, these failures are found to be almost always in one and the same direction: the side walls are thrust outwards; and this has taken place more or less in every building constructed on the principle of counterpoise, from the stone vault to the untied open roof of timber. Nor is this otherwise than what might be expected: a building always "settles" after, or at the time of, completion: struts and braces are compressed, and ties placed in a state of tension. Even masonry and brick-work are not exempt from the effects of "settling," however little their form may change in consequence. Now, it is only in bridges that the system of counterpoise is fully carried out, and the nearest approach to that in ecclesiastical architecture, is where the clerestory is dispensed with, and the aisle roofs are of equal height with that of the nave. When these are vaulted, as they are in the Temple church, the whole may be compared to a bridge of three arches ;* only there is this important difference between the two, that the one finds a firm resistless abutment in the earth or bank which terminates it at either end; the other has not this advantage, and in practice it cannot have even an equivalent. In general the architects are content to provide against the thrust, so far as to prevent the walls being pushed down; but they do not appear to have gone much beyond that in the means they took to form their abutments. The Temple church is a case in point, and an exemplification of what has just been stated. Such buildings cannot be otherwise than affected by settling, in all parts where the system of counterpoise is not perfect. The same may be said of walls carrying wood roofs that are not tied. Where there is a clerestory, the aisle roofs push the walls inwards, while the nave roof pushes them outwards; and according as the lines of contact are near or far apart, in such proportion will the wall be affected in settling:

*See British Critic, 1842. vol. xxxi. p. 461. Article, "Open Roofs." But we differ with the writer on some rather important points.