Appendix I. The Distortion of the West Front of Peterborough Cathedral.

 

The west front is remarkable not only for its unique architectural design, which has been much discussed, but also for the virtuosity of its construction.  Built between about 1200 and 1238, the two triangular piers were perhaps the most massive individual structures of their era.  The splays of the three portals are decorated with thirty-six detached shafts, each shaft being fifty-two feet in length, consisting of four rods of Alwalton ‘marble’ six inches in diameter and averaging thirteen feet in length.  Each rod is attached to the main pier by small and delicately formed ring-shafts (Fig. 263).  On the back of the piers and on the west wall behind the great arcade and on the staircase towers and gables more than two hundred more detached colonettes are displayed, many only four inches in diameter and in length between four and sixteen feet.  The total length of Alwalton stone rods displayed is about four thousand feet or 1200 m.

      The quarrying and the transportation of these fragile rods from the quarry by water to the cathedral would have been a major work of skilful organisation, while consideration of the difficulties involved in the fashioning, polishing and raising of them into position at the site as the piers increased in height can only fill one with wonder at the final achievement.

      The construction of the arches above the piers was an equally remarkable achievement.  Each of the two side openings is 26 feet wide, while the full width of each arch is 43 feet.  The whole structure is not only the most massive but also one of the earliest two-centred arch constructions in England, for the choir at Canterbury was still under construction when the Peterborough arcade was begun, and was not dedicated until 1220.   The arches of the Canterbury choir arcade measure only 16 feet from centre to centre as against the 43 feet at Peterborough, and only 28 feet in total height compared with the 78 feet of the Peterborough façade arcade.

     But there is one other aspect of the Peterborough front that distinguishes it from most great medieval churches.  As it rises above ground level it bulges outwards until at the string-course it is more than two feet from the vertical at the centre (Fig. 264).

     Worthy and painstaking attempts have been made to explain this distortion, most notably by Bush and Stallard (1996).  Unfortunately, although they provide valuable information, their conclusion that the bulge was a deliberate part of the architect’s design in order to create an optical illusion is mistaken.

      They describe how during a violent storm in 1895 the falling from the façade of some masonry, including a whole pinnacle, led to the close examination of the whole façade structure by the distinguished architect John Loughborough Pearson.  He reported that

the detached clustered columns, with the string-course above them, leans outwards to the extent of fully two feet, and that the three gable ends have an inclination in the same direction of about six inches (Bush, Stallard 1996, 4).     

 

 He added that “surprisingly little movement had taken place between the groining”.  This latter remark is taken to mean that the front had not moved recently and must also mean that the vaulting behind the front was constructed after the front had distorted and had been restrained.

      Pearson realised that the distortion had begun while the front was still under construction:

     The three gables lean over at a less angle than the pillars below them, and some of the work in connection with these gables inside the roof is nearly perpendicular.  I am very much disposed to think that the pillars began to settle at a very early period, even perhaps before the gables were erected and certainly before the groining of the arcade was put in.  It must have been at some early period that the four great ties were introduced which pass through the wall for the purpose of hanging the west front to the square towers behind; these ties are still in a perfect state and show clearly that since their introduction little or no further movement has taken place (Bush & Stallard  1996, 4).

    

Pearson found everything else “perfectly upright”.  He inserted four extra tie-bars, and a year later that found nothing had changed, the falling of the masonry being attributed to “the effects of time, the want of timely repair and…the action of the weather”.

     The outward bulge of the upper structure is not a simple curve but, as is seen in the sectional plan, probably made by Pearson (Fig. 265), (reproduced by Bush and Stallard 1996, 10,11), the distortion has three straight sections, in their words:

changeing direction above each of the triangular piers  Neither is there any sag along the length of the passage such as one would expect to find above pillars that had sunk (Bush and Stallard 1996, 6). 

  

     James Irvine, who was Pearson’s Clerk of Works, and who assembled a vast collection of papers concerning the cathedral’s fabric, was of the opinion that there had been a change of design above the string-course probably as a result of the bulging of the Front.  He then produced a design for the gables as he believed they had been originally intended.  His opinions have been much derided and his attempt to have his ideas published in The Builder were sharply refused by the editor.  Furthermore, his suggestion that the Perpendicular Porch had been inserted in order to prevent further leaning was totally mistaken, as Reilly (1997) and Bush and Stallard (1996) have observed, for any further expansion of the bulge would have cracked the piers or demolished the Porch.  We are indebted to Ruskin and Morris and their fellow members of the Society for the Protection of Ancient Buildings for preventing the then Dean and Chapter from allowing Irvine to reconstruct the gables according to his own theories.  He was allowed, however, to rebuild the northern gable, using numbered stones, and thus the front remains unaltered in appearance.

     For all this information we must acknowledge the value of Bush and Stallard’s investigation.  Nevertheless one must dissent from the conclusion at which they arrived.  They are of the opinion that the distortion was deliberately designed and created by the architect of 1200 in order to create an optical illusion.  Their conviction is that this forward distortion of the geometry was intended to create the effect of an upward curve of the string-course, the purpose of which was to correct the effect of sagging that a perfectly level course might have.  The inspiration for this idea would have to have been the entasis found in Greek and Roman architecture.

     The arguments against this idea are several.  The medieval architects had no knowledge of Greek architecture.  Greece was a very foreign country at that time, being within the realm of the Eastern Church centred on Constantinople, and fiercely opposed to the Western Church centred on Rome.  The two had separated in the ninth century with a mutual animosity that has lasted virtually to the present day.  The sacking of Constantinople in 1202 by the armies of the Crusade was a central event.  But even if an architect had been in such a crusade he would never have encountered a Greek temple.  If he had done so it would have meant nothing to him.  If he had ever heard of entasis, which is highly unlikely, in fact impossible at that time, he would have no means of measuring it.  The entasis on the east and west facades of the Parthenon is 2.61 inches, and 4.39 inches on the north and south facades, hardly comparable with the 26 inches at Peterborough.  In any case the Greek entasis in the horizontals is always upward, never forward as at Peterborough.

     If the medieval architect had nevertheless wished to create such an optical illusion he would certainly have curved his string-course upward, a simple and stable arrangement, instead of forward to create an unstable structure requiring the insertion of massive tie-bars.  As for optical illusions, they belong to the Renaissance, and in any case the eyes of Heaven would never have been deceived by an optical illusion.

     The true cause of the distortion at Peterborough lies in the fundamental weakness of the structural design of the front, which was created in the very early experimental years of the two-centred arch in England.  It also needs to be borne in mind that medieval architects, even centuries later, in W.C.Leedie’s words, “were incapable of even the simplest structural calculation” (Leedy 1980, 23).

     A row of arches depends for its stability on the equal balance of horizontal thrust of arch against arch.  It has to be appreciated that stone masonry is not solid like concrete.  In its fragmented nature it is fundamentally flexible, particularly in the months immediately after its building when the lime mortar is slowly hardening.  The degree of lateral thrust created by an arch depends on its width, its shape, and the weight of its own structural mass plus the mass of superstructure that it supports.  In most arcades the arches are all the same shape and size and support equal weights of masonry.  Thus any arcade will constitute a balance of stresses and thus a stable structure.  But in the Peterborough front the central arch is different in shape and size and supports less weight than the side arches.  The thrust that it exerts towards the north and south arches is much less than the thrust that those larger heavier arches exert towards the centre.  Thus the extra force that they apply forces the central arch either to bend upwards, which would entail lifting the weight bearing down upon it, or, much more easily, to bend forward where there is no opposing force whatever.

     The difference in thrust between the arches can be only roughly estimated, but as the arcs of the wider arches subtend an angle of 80 degrees while the arcs of the central arch subtends an angle of only 65 degrees, there is an imbalance of about 5 to 4 in the mass of the arches alone.  Furthermore, the central arch supports a superstructure only 30 feet in width, while the larger arches each support 40 feet, an imbalance of 4 to 3.  An added factor exacerbating the situation arises from the difference in shape between the arches.  A narrow arch produces a total thrust that is directed at an angle more towards the vertical than the thrust of arches of a wider proportion. Consequently a smaller proportion of the total thrust is directed horizontally from the central arch than from the side arches.  The total effect is that each side of the central arch produces a horizontal thrust only slightly more than half of the opposing thrust of each of the larger side arches.

     What may be found surprising is the very small lateral expansion of the side arches that will produce a forward expansion in the central arch of 26 inches.  A close approximation can easily be calculated using the Pythagorean theorem:

      Thus an expansion of only half an inch from each of the 43foot wide side arches would cause the central arch to move forward 24 inches, and that is exactly what happened. 

      It is hardly possible to contemplate the moment when the distortion of the great arcade was first discovered without envisaging the horror and utter despair of all those involved in the construction, from bishop and architect to lime-burner and water-carrier as they envisage the whole wonderful structure crashing to the ground.  Their sheer terror as they all waited for the fashioning and insertion of the great tie bars can hardly be imagined.  Perhaps easier to contemplate is the certain thanksgiving and celebration that must have followed the successful salvation of the great enterprise.

     The bulging of the upper levels had no effect whatever on the straightness at ground level, which remains perfect, and the distortion in no way bears upon the geometricality of the architecture.