Archaeological evidence above ground: The construction of the house

The Construction of the House

It is now necessary to give a brief general account of the construction of the building. Except in the Georgian block and the eighteenth-century extension of the south range, which were built with brick walls carrying timber floors and roof, the walls were carried up in brick to the level of the first floor, which they supported. Above this level the south, west and middle ranges (and presumably the north range) had originally been constructed of oak timbers, with post and panel walls to the first floor, carrying timber roofs. The roof of the west range had principals spaced at intervals of from 14 feet to 14 feet 6 inches, the trusses being of the simple queen-post type (fig. 7). As has already been noticed, the roof of the south range was arch-braced (fig. 5), the first five principals from the east end being spaced at ten-foot and the next two at twelve-foot intervals, with a final bay of nine feet at the west end. On the face of it there would appear to have been a difference of builder, if not of date, between these two roofs, each of which spanned roughly 18 feet 9 inches.

Where it had remained undisturbed the brickwork above ground was remarkably uniform in character, and was, therefore, of little help in dating the various parts of the structure, though other evidence suggested that it was not all of the same date. The bricks measured generally from 8¹/8 to 9 inches in length, 3⁷/8 to 4³/8 inches in breadth, and 1⁷/8 to 2³/8 inches in thickness. A peculiarity of some of them was a slightly sunken margin, varying in width from a quarter to half-an-inch, all round their top faces, so that the joints showed wider on the face than their normal thickness in the wall. All walls were built in English bond rising eight courses to 21 inches, or more, up to 22 inches. The mortar was everywhere of soft lime with scarcely any recognizable admixture of sand. (fn. 1) Wall thicknesses varied from four bricks in the back wall of the Georgian block (which, containing as it did the bricked-up four-centred arch already noticed, was obviously of earlier date than the rest of the block) to two bricks on the courtyard sides of the south, west and middle ranges. The outer walls of the south and west ranges were two and a half bricks thick. Remains of a moulded plinth 4½ inches, or two courses, in depth were found in various places on the walls bounding Court I, and a similar one, at a higher level, could be traced in Court II. In the angle between the Georgian and south ranges a second chamfered plinth occurred below the first.

Brickwork details of interest uncovered in the course of demolition included two fourlight mullioned and transomed windows, one on either side of the middle range (fig. 3); two four-centred moulded arches, one at the south and the other at the north end of the passage joining Courts I and II (Plates 25a, 25b), that at the north end having a three-light mullioned window above it (fig. 18, page 72); and, finally, the elaborately moulded arch in the back of the Georgian range (fig. 2, page 8).

The four-light windows had equal mullions, each built up of single bricks, divided equally by transoms formed of bricks on edge. Mullions, heads and jambs were hollowsplayed, and transoms weathered above with a straight splay and hollow-splayed beneath, both inside and out. Sills were weathered outside and square within. Two double-mitred bricks were built into each mullion to form the junctions with the transoms, which were carried on wrought iron bars let into sinkings on their undersides. Both windows were sunk within outer splayed jambs returned at the heads and sills to give three-inch plain margins all round. The three-light window was similarly constructed, but lacked transoms and outer jambs (fig. 3, page 12, fig. 18, page 72).

No specially moulded bricks were used in any of these windows, which were formed entirely with bricks similar to those used in the walling generally, the mouldings, mitres and weatherings all having been cut, in situ, with admirable precision. The same was true of the passage arches which were identical and consisted each of two half-brick rings, the upper one plain, the lower one cleanly and accurately cut to ovolo and cavetto mouldings separated by a fillet (Plate 25a, 25b). These mouldings were continued on the jambs, but the manner in which they were finished at the bottom could not be ascertained, as the lower parts had been hacked when the rendering of Roman cement was applied. In contrast with the passage arches the much larger arch in the back wall of the Georgian range appeared rough. It was made up of axed bricks of varying profiles, only the simplest hollow chamfer being repeated in building up the series of mouldings (fig. 2, page 8). The fact that parts of one of the jambs were found, at the time of demolition, to be covered with a thin coating of white plaster consisting almost entirely of lime and hair, may indicate that the arch was left rough because it was from the first intended to be plastered—a conclusion that is supported by its position in the original building as disclosed by excavation (see page 30).

Though not strictly matters of structural method, two unusual items of brickwork must be recorded. The first was the existence in Room 27 of three small arched niches 1 foot 9 inches wide by 2 feet high, two of them each 9 inches deep, on the south wall, and the third, 1 foot deep, on the west wall, to the south of a blocked door with a four-centred arched head (S on plan on Plate 5). The arches were plain axed half-brick rings, slightly pointed at the top. The sills of those on the south wall were about 2 feet above the top of the dwarf wall discussed below (Plate 25e). The one on the west wall was two courses higher. Their purpose is unknown. The other item was a small extent of painted brickwork which had been covered over in the building of the porch in the south-east corner of Court I. The whole area had been coloured Venetian red and the joints lined-in thinly with white. This was the only example of such imitative painting to be found, though there were vestiges of internal decorative painting in one or two places, besides the striking wall painting discovered in Room 14 and described below.

Immediately below first-floor level the brick walls were capped with oak plates measuring roughly 8 inches in width by 7½ inches in depth, over which the primary (cross) binders were halved and dovetailed. Normally the binders were of greater depth than the plates, so that their undersides projected below the beds of the plates. Those in the south range were 11½ inches square in section, being moulded on their exposed undersides, and having 1 inch rebates on each upper edge into which the original elm floor boards had fitted, leaving their upper surfaces level with the exposed tops of the binders. The secondary binders spanning centrally between the primary binders were also moulded on their undersides and mitred to the primary binders. They were shorter in depth than the latter by the 1 inch thickness of the elm floor boards, which were laid over them. The oak floor joists, which spanned from the secondary binders to the wall plates, averaged about 7 inches in width by 5 in depth, and were set with their tops level with the tops of the secondary binders (Plate 17a). The construction of the floors of the west and middle ranges was similar, but the binders were unmoulded (Plate 17b).

The walls of the upper storey were constructed with oak storey posts 8 or 9 inches square in section standing on, and sunk into, the wall plates. The storey posts in their turn carried the eaves plates of the same size and material, the panels thus formed being filled in with vertical timber studding roughly 5 or 6 inches thick and varying in width from 4 to 9 inches. Some of the storey posts were braced at the bottom with curved struts measuring roughly 8 inches by 1½ inches in section.

In the east wall of the west range, on the upper storey, were the remains of two two-light oak mullioned windows (Y, Y on Plate 7) fitted between storey posts set two feet apart. Each window was made up, apart from the storey posts, of four pieces of timber, two forming the head, one for the sill and one for the mullion. Externally the mullion, which measured 10 inches front to back by 7 inches wide, was set with its front edge in the plane of the outer faces of the storey posts, and had a square nosing flanked by ovolo mouldings on either side. The internal nosing, which was chamfered, was set back from the plane of the inner faces of the storey posts, and was flanked by ogees separated from hollow chamfers by fillets. Both externally and internally the responding mouldings were worked on the storey posts, which acted as jambs. The external ovolo and internal hollow chamfer were carried round the four-centred arched heads of the lights, these being formed on one cross piece let into the underside of the upper head and having triangular sinkings in the spandrels of the arches internally. On the inside, in front of this arched head the outer mouldings of the mullions continued upwards, to be returned horizontally on the underside of the upper, structural, head which was tenoned into the storey posts, as was the sill. The mullion was stub-tenoned into the sill, which was plain, with square arrises at the top, and rounded ones below (fig. 4, Plate 24a).

The roof of the west range had trusses of the queen post type. The principal rafters, which were no larger than the intermediate ones and in the same plane, were from 6 to 8 inches wide by about 6 inches deep, reduced to rather less than half this depth above the purlins. These, measuring 6 inches by 4½, rested in a notch formed between rafters and collars, and were stiffened by simple arched struts springing from the principals. Most of the collars, which varied in section, but averaged about 9½ inches by 6, were formed of bent timbers giving them a lift in the centre. The one nearest the north wall of the middle range was straight. Some of the principals rested on and were fixed with a bridle joint to tie beams averaging about 8 inches by 8 in section. Others were bridle-tenoned to the plates and halved and pegged to the tie beams, which thus acted more as ceiling binders than as tie beams proper. All rafters were halved and pegged at the top. Vertical studs, varying in number and in size, and housed into the collars, occupied the space between the tie beams and the collars, and also the triangular space above the collars.

Most interesting was the arch-braced collar roof over the original portion of the south range (Plates 18, 19a, 19b, 19c). Seen from below this had the form of a four-centred arched wagon vault, with ribs, which obscured its essentially simple construction. Basically, however, it was a collar roof having 9 inch by 7 inch principal rafters and 12 inch by 8 inch collars, with 8 inch by 7 inch braced purlins let into the principals just above the collars. The feet of the principal and common rafters (the latter measuring 4 inches by 5) rested on heavy, 10½ inches by 10, eaves plates supported beneath the principals by the storey posts, the plates being joined over the posts. The collars, which had a slight upward lift in the middle on both upper and lower faces, were tenoned into the principals. Beneath them, and likewise tenoned into the principals, were lower collars which formed the middle sections of moulded arched braces. These were built up of five pieces to form prominent four-centred timber ribs, their undersides reduced in width by means of cavetto mouldings on each edge and their sides enriched with planted bolection mouldings and beads. Between each archbraced principal less prominent minor ribs, 4 inches square in section, followed the same four-centred curve. The minor ribs were housed, at the apex of the vault, into a 5 inch by 4½ inch ridge rib, and divided, at about one-third of their lengths from the springing, by minor purlins, measuring 6 inches by 6 inches, into which their upper and lower lengths were tenoned. The feet of the lower lengths were carried on 7½ inch by 6 inch minor plates, half-notched to the storey posts at a level below that of the main structural plate fig. 5).

The jointing of the various members merits description. The principal rafters were mitred at the apex and joined by a tenon cut off square at the end to meet the mitre about 3 inches below the apex, thus ensuring a covering for the end grain of the timber. Both collars were tenoned into the principals, and the upper one was stub-tenoned into the lower at the centre. The ridge rib was tenoned into the lower collar, and had wedge-shaped notches along its sides for the housing of the minor arched ribs. Great ingenuity was displayed in the jointing of the five pieces of the arched brace. The piece which formed the lower collar had its mouldings worked on the solid, and was scarfed and tenoned to the pieces on either side, the bolection moulding stopping short of the scarfing in such a way that the planted bolection mouldings could overlap the centre piece and cover the joint. The remaining pieces were likewise scarfed and tenoned. Finally, what was perhaps the most interesting example of carpentry in the whole roof occurred at the junction of the storey posts and the main plates. The plates were scarfed and double-tenoned where their separate lengths joined over the storey posts, and the latter were tenoned into them in such a way that their mortices, in the bottoms of the plates, came immediately below the central tenons of the scarfed tenon joints. The storey posts had, in addition, open mortices at the top of their inner faces to receive the tenon of the arched brace, and, below this on either side, a notch to receive the halved ends of the minor plates.

The above description refers only to the kinds of joint met with in the construction of this roof, and takes no account of minor variations in execution which were in fact a feature of the work, suggesting that each of the craftsmen employed on it had his own special way of doing any particular detail. The outstanding example of such variations in execution was perhaps to be observed in the jointing of the heads and jambs of the door frames R, W and X, and of the one found built into the west wall of the Georgian block. In this last the jamb was tenoned into the head, and the mouldings were part mitred, part scribed, at the angles (fig. 6), whereas in W the head was tenoned into the jamb and the mouldings returned by means of a remarkable mitre, undercut on the jamb, and involving some complicated scribing in consequence (Plate 23c). Other variations occurred in R and X, creating the impression that the carpenter's mitre was here still in the experimental stage of its development. Indeed it has been impossible to study such construction without becoming aware that the craftsmen who built Brooke House were men who exercised their own initiative, some within the general terms of a style that was passing, and others within those of a style that for them was as yet only vaguely defined.

It very soon became evident, during the course of demolition, that a radical change had been made in the structure of the building at an early period. Beneath the floor of the west range the foundations of a wall ran parallel with the east wall of the range and almost touching it on the inside for the whole length of Court I. Similar foundations were found on the north side of the south range and on both sides of the middle range in its original state (i.e. before the building of the kitchen offices on its north side (fn. 2) ) indicating that here, as well as in the west range, the courtyard walls had at some time been rebuilt immediately outside their former lines. Furthermore, there was evidence, in the form of a storey post still remaining in situ in the south-west corner of Court I, that the original walls had been of timber standing on a brick foundation or ground wall. Since the bottom of this storey post, which was L-shaped in plan, stopped short about 3 inches above the top of the foundation wall on the east side of the west range, and the top course of bricks in this foundation was finished with a bull-nosed brick on its outer edge, it seems more than likely that the post had been left intact at the time of the alteration, which involved a slight raising of the level of the floor. In this case the 3 inch gap below the storey post would have been occupied by the halved end of the sole plate on which the timber walls were built. Moreover, the eastward facing arm of the L-shaped storey post was in line with the brick ground wall discovered beneath the floor of the northern, eighteenthcentury, portion of the south block, suggesting that the original north wall of the south range had been similarly constructed. Incidentally the foundation of the new north wall broke forward in two deep bays into what, presumably, had been the southern end of Court I.

The rebuilding, in brick, of the ground-floor walls of Court I, outside the lines of the original walls, could have been effected with little disturbance of the superstructure if the upper storey had been jettied out beyond the face of the lower one, as indeed was normal in walls constructed entirely of timber. The new wall could then have supported the ends of the beams. That this was the case originally on both sides of the middle range and on the south side, at least, of the south range was demonstrated by the discovery, in the course of the demolition, of binders halved and dovetailed below as if to fit over a plate about 18 inches from the ends, which were shaped. Examples of these occurred towards the west end of the north side of the middle range, and at the western end of the south wall of the south range. Another with a dovetailed tenon occurring immediately above the first ground wall remained in position on the south side of the middle range.

There was no evidence that a similar enlargement of the ground floor had taken place on the south side of the south range or the west side of the west range, or that these walls had ever, in the lower storey, been constructed of anything but brick. Mixed construction, with one wall wholly or partly of brick and another wholly of timber, is not unknown in houses of the late fifteenth or early sixteenth century: and in the case of Brooke House it is reasonable to suppose that the greater protection of brick walls, as compared with timber, may have been considered desirable on the outer faces of the building. On the other hand there were clear signs that the first floor of the west range of Court I had actually been reduced in width along its west side, and that a similar reduction had been made at the eastern end of the south side of the south range.

It will be remembered that, externally, the west front had no jetty at the level of the first floor. The construction at this point was not visible until the house came to be demolished, when it was found that the ends of the beams, which, presumably, had once formed a jetty had, at some time, been sawn back in situ behind the face of the wall below (Plate 19d). The storey posts and studding of the upper part of the wall were here, as elsewhere, standing on the wall plate above the brick wall of the lower storey: but, if the floor beams had once projected beyond the face of the wall such an arrangement would have been nonsensical. The inference is that the upper storey had originally been jettied out and its wall carried, in the normal way, on a timber sill fixed above the ends of the projecting beams (fig. 7). Why the alteration should have been made is not clear. The only plausible reasons that can be advanced are either that the ends of the beams had rotted, or that they had been burned. In the first case, since there are reasons for thinking that the alteration was made within a hundred years, at most, of the building of the house, and very probably much less, it is necessary to assume an unusually rapid decay. With regard to the second, it must be recorded that no signs of fire were found in any of the remaining timbers.

Although the enlargement of the ground floor could have been made without undue disturbance of the superstructure, the reduction of the first floor could not, since the timbers of the upper storey carried the roof. The arch-braced roof of the south range was, in fact, clearly designed to fit the altered span, and may be assumed, therefore, to have been wholly remodelled if not completely renewed at the time the alterations were made. It is impossible to say whether or not that on the west range was renewed as well. There were no signs of alterations to the timbers, but it must be observed that the only alterations involved in reducing the span of a simple roof of this type would be the cutting of new mortises in the tie beams, and the shortening of the rafters by cutting off their heads; both of which operations could have been carried out without leaving any trace of a change. If, therefore, this roof was not renewed but merely altered (as seems, perhaps, the more likely case) it is nevertheless very probable that the alteration was made at the time of the changes in the south range, since the ridges of the two ranges were on a level. The question of the dating of these alterations will be discussed later (page 70).