Monday, 30 January 2017

Sehzade Mosque, Istanbul

I have left my hometown Istanbul over ten years ago. Occasionally, I find myself longing for the breath-taking sight of its undulating sea of domes and sharply rising minarets. When I started writing this blog, I knew I would one day write about its monuments. 


Figure 1 – A view of the Sehzade mosque from east (date unknown, retrieved from [1])

In November last year, I visited Sehzade Mosque, one of Sinan’s countless masterpieces. It was built between the years 1544-1548, in the memory of Sehzade (Prince) Mehmed, Sultan Suleyman’s much beloved son. The architectural customs of the empire would have dictated Mehmed’s mausoleum to be located in Bursa, the empire’s previous capital. It was customary for the deceased prince’s funds to be used to finance the building of a mausoleum, and associated charitable and religions institutions. However, breaking with tradition, Sultan Suleyman supported the building of a grand complex in the capital Istanbul, equal in rank to imperial mosques built in the name of his forebearers (Fatih, Bayezid and Selim). The complex was thus accorded with architectural features reserved to imperial mosques. These included two elegantly decorated minarets with two galleries and a wonderfully elaborate structural plan [2] (see Figure 1). No other monument built for an Ottoman prince thereafter was accorded such privileges, and in my opinion, no other mosque thereafter executed with such decorative brilliance.

This monument was only made possible with careful planning and engineering. For the remainder of this post, I will focus on a few engineering aspects that I have noted during my visit and research.


Figure 2 – A schematic of the Sehzade mosque with an illustration of the flow of loads. Retrieved from [3]

The rational structural plan

The design of the Sehzade Mosque is underpinned by a rational structural plan that reflects the flow of forces. This is illustrated by a schematic of the mosque in Figure 2. The mosque is placed on a north-south axis, and a 19m diameter semi-circular dome carves an unimpeded central space. The brick dome sits on the crown of four thick stone arches. The central dome exerts lateral pressure which is transmitted to these arches, and to the brick pendentives. While the pendentives have little lateral load carrying capacity, diagonally placed flying buttresses and weighing towers, enhance the lateral load resistance in this direction (see Figure 3). The lateral loads which are transmitted to the arch and pendentive are then distributed downwards. A portion of these loads is transmitted via the piers to its foundations. The remaining portion (in the N-S and E-W directions) is transmitted to four semi-domes, which then transfer the lateral thrust to the perimeter buttress walls. From here, the forces flow to the foundations from the chunky buttress walls, which are skilfully hidden behind a domical arcade (see Figure 1). It is reasonable to think that the structure was constructed bottom up, starting with perimeter buttress walls and ending with the central dome. This rational arrangement of structural elements results in a graceful and dynamic exterior (see Figure 3) and a unified interior space with a gradually receding ceiling (see Figure 4).
While evaluating the plan and dimensions, it is useful to quantify the distribution of lateral loads in the system. The maximum thrust that can be carried by the side domes under self-weight is proportional to the square of the diameter of the dome. Therefore due to the 1:2.5 diameter ratio between central and higher tier side domes, each side dome can only carry a maximum of 1/6th of the lateral thrust exerted by the central dome. This demonstrates that the piers under the arches carry a significant portion of the lateral load, while the side domes provide stability against foundation movements. This explains why the piers are massive, with a 4.5m thickness.


Figure 3 – A photo of the Sehzade mosque. Note the diagonally placed flying buttresses on the drum of the mosque and the weighing towers placed in front. (Retrieved from [8])

Further information can be obtained by investigating the side cross-section view of the central semi-circular dome of the Sehzade mosque in Figure 5. When viewed from outside (e.g. Figure 3), it appears as if the area under the dome is an upright drum, with window openings to let the light in. However, seen from inside, the windows openings appear curved (see Figure 4). Looking closely at the plane section in Figure 5, it becomes obvious that at angle of embrace of approx. 110 degrees, the dome section becomes thicker, to become flat on the outside. According to simple membrane analyses [5], below this is the height where meridional cracks would appear due to emerging tensile hoop stresses for solid domes. However, having a thicker section here is likely to have made the dome more resilient towards this sort of damage.


Figure 4 – A photo of the interior of Sehzade mosque, looking north towards the mihrab.


Figure 5 – A cross section to-scale drawing of the Sehzade mosque (Retrieved from [4])


The elusive earthquake resistance

The Sehzade mosque has been stable for centuries, and is unlikely to experience large static movements which may threaten its stability. However, dynamic loading is a major threat, since Istanbul is in a zone of seismic activity. During its 500 year lifetime, Sehzade mosque has survived several major earthquakes, including the 1766 and 1894 earthquakes, which have caused damage in nearby historic buildings (the dome of Fatih mosque collapsed in 1766).
It is difficult to ascertain if the Sehzade mosque (or other similar mosques) will survive potentially more intense future earthquakes. We have limited information (or visual record) of the damage sustained by historic mosques during earthquakes. However, previous experience and research have highlighted that that these structures are particularly vulnerable. Their first translation and torsional vibration modes of mosques are typically in the high frequency range (0.2-0.5s in both directions), which will result in an amplification of the applied dynamic forces on the structure.
From previous observations of earthquake damage, several critical elements can be identified:
  1.        Foundations under the piers which carry lateral loads: A significant portion of the induced lateral loads due to self-weight are transmitted to the base of the piers. Base shears and moments due to lateral earthquake loading would similarly be expected to concentrate in this area. Potential settlements here during dynamic loading can be detrimental to the stability of the dome.
  2.        Pendentives: The arches of the mosque represent the two load carrying frames in the orthogonal directions. Due to earthquake directionality effects, ground motion in orthogonal directions can be different, leading to differential movement and tensile stresses in the pendentive area that connects the orthogonal arches. Damage in this area could lead to support relief for the dome and partial collapse. Such damage has been observed in the past for the dome of Hagia Sophia.
  3.       Dome windows:  Due to the presence of window openings on the dome, the lateral loading needs to be transmitted through the reduced solid sections of the dome in between windows. These represent likely locations for strain concentration, and may get damaged due to lateral loads.

Nonetheless, our understanding of the dynamic behaviour of these structures remain rudimentary. In order to ensure that these structures survive for another 500 years, we need to have an improved understanding of their dynamic behaviour. One way to achieve this would be to use new sensing techniques to measure the dynamic response of these structures, to advance our understanding of their behaviour. 

Acoustic characteristics

An imam situated in the south (mihrab) area of the mosque leads prayer in mosques. Occasionally, the muezzin, who would be located in the southeast corner of the Sehzade mosque, reads prayers aloud. The congregation responds to the prayers recited by the imam and muezzin. Therefore, optimising the acoustics of mosques, especially from the perspective of speech intelligibility, was an important engineering problem in Sinan’s period. 

Within this context, it is desirable to achieve a homogenous dissipation of sound energy inside the mosque, to prevent directional sound effects. It is equally important to ensure sufficiently quick absorption of sound by the structure, to ensure that consecutive sounds are not mixed. The domed structure of the mosque presented a challenging environment to achieve these objectives. The curved surface of the dome reflects the sound energy in different directions with little dissipation and much time delay, resulting in the sound to dissipate over a long time. This causes a long reverberation time, and it is not acoustically desirable.

There were numerous established acoustic solutions to deal with the aforementioned issues. For instance, using plasters with strong absorption characteristics to cover the dome surface enabled a more diffuse sound field. In a recent investigation of the nearby Suleymaniye mosque, it was concluded that the restoration of dome decorations with cement based plasters, has led to increased reverberation times. The original lime based plaster absorbed the sound more effectively, decreasing the reverberation times.


Figure 6 – (left) Schematic for a Helmholtz (cavity) resonator from Sultan Ahmed mosque and (right) a photo of a filled-in cavity (Photos retrieved from [7])

Another solution for acoustic improvement is Helmholtz resonators. This is a time-honoured method of absorbing a narrow frequency band from broadband sound waves. It is achieved in a fascinatingly simple manner. A cavity is opened on highly sound-reflective surfaces inside the mosque. From inside the mosque, these cavities look like simple cylindirical holes with diameters up to 10cm. However, inside the hole, the narrow cavity neck gets wider, and can protrude 0.5m deep into the dome, providing a large cavity volume (see Figure 6 for similar resonators from another mosque). The incoming sound waves at a particular frequency are affected by the springiness of the air inside the cavity. The incoming sound tries to squeeze the air in the cavity, but the air resists, this interaction causes a range of sound waves to behave like a mass on a spring, whose motion is dissipated (absorbed) over time. By changing the dimensions of the cavity neck and backing volume, the absorption frequency band can be modulated.

During a restoration of Sehzade mosque during the 1990s, 144 cavity (Helmholtz) resonators were found [7]. Most of these were located on the dome. The original sizes of these resonators were such that they were designed to filter low frequency noise, smaller than 250 Hz. Unfortunately, many were filled in during previous restoration works (see Figure 6), and were not functioning as intended. Since measurements had demonstrated that reverberation time for low frequencies is unusually high inside the mosque, it was suggested that the destructive restoration works have affected the mosque acoustics negatively. However, the influence of this erroneous intervention may not be felt strongly by congregations today, imams and muezzins often use microphones to lead prayer.

Despite these small changes, Sehzade Mosque is a well-preserved and wonderful monument, that makes one wonder what life in Istanbul in the 16th century would have been like, under that sea of domes and sharply rising minarets.

REFERENCES
[2] Neciopglu, Gulru. The Age of Sinan. Reaktion Books, 2011.
[3] Karaesmen, E., et al. "Seismic behaviour of old masonry structures." Proceedings of the tenth World Conference on earthquake engineering. 1992.
[4] Kuban, Doğan, and Cemal Emden. Osmanlı mimarisi. Yapi-Endustri Merkezi, 2007.
[5] Heyman, Jacques. The stone skeleton: structural engineering of masonry architecture. Cambridge University Press, 1997
[6] Gül, Zühre Sü, and Mehmet Çalışkan. "A DISCUSSION ON THE ACOUSTICS OF SÜLEYMANİYE MOSQUE FOR ITS ORIGINAL STATE."
[7] Kayili, Mutbul. "Acoustic solutions in classic ottoman architecture." Foundation for Science technology and Civilization (2005).


Sunday, 8 January 2017

2017: Turn and face the strange / ch-ch-ch-Changes

Happy new year to all my friends! Marie and I celebrated the new year in Cambridge together with Berk, Eda and Ishan. 


















With 2017, came new resolutions and hopes, as well as some changes to this blog. Last year, I aimed to write weekly posts but managed only three posts in four months, equalling to approximately 0.2 posts per week. To remedy that, I will now aim to write one post every month. Hence, I am making a counterintuitive case of proposing lower expectations for promoting productivity. Let’s see how I fare.

I have also decided to write on a broader array of subjects. For a little while now, I have been working on optical measurement systems (using laser and RGB camera based systems), and the related fundamental physics intrigue me. Venturing into these subjects will allow me to occasionally write about another interest, cinema. As with the previous posts on structures, I will try to focus on historical and mechanical aspects.

Oh, and also, I have been silently on Twitter for much of 2016. It has been surprising to find so much original and interesting content there! I have decided therefore to join in and share interesting things that I come across. On Twitter, there will be less mulling over as opposed to blog posts, and this way, I hope can share some of my excitement of discovering something new. You can find my tweets at @physionotrace

But before all that, I need to go back to a visit that I made in 2016, to fulfil another resolution. 

Sunday, 23 October 2016

Leper Chapel, Barnwell

The Leper Chapel of Barnwell is one of the most unspoilt examples of Norman architecture in the UK [1]. It is a building constructed of a broad nave and a narrow chancel. Today, it sits quietly by the busy Newmarket Road. You can visit the chapel on a single day in September when the Stourbridge Fair takes place (see Figures 1 and 2).
















Figure 1 – View of the chapel from the east, with Stourbridge Fair stalls in front.


Figure 2 – Plan view of the chapel [2]

The chapel’s history is inextricably linked with that of the fair. The chapel dates to the middle of the 12th century, where it was constructed as a part of a hospital for a colony of lepers. The lepers were expected to lead a religious life, strictly observing the rituals. They also worked the land and begged for alms. Despite this, their sources of income were meagre, and the rental of the hospital land was necessary. This led to the beginnings of the Stourbridge Fair, which in a few centuries would be described as ‘the biggest fair in Europe’.
With the decline of leprosy in the latter part of the 13th century, the wholesale rental of the chapel and its surrounding lands became common practice. By the middle of the 18th century, the chapel ceased to be a place of worship. Throughout the year, it was used as a storage place for stalls of the fair. During the fair, it was used as a drinking booth. However, the continued commercial use meant that the chapel was kept in good condition.


Figure 3 – A view of the chapel from the northeast, drawn by Cotman [3]

.
Figure 4 – A drawing of the interior of the chapel, looking towards the chancel arch [3]

The decline of the Stourbridge fair in the 18th century initiated a period of neglect for the chapel. Prints from the early 19th century (see Figures 3 and 4) document the dilapidated building [3]. In these prints, we can see chimneys rising on the nave and chancel roofs as well as a new door in the chancel. Pearce [4] relates these aspects to the various uses of the building during the period. In particular, the new door may have facilitated moving the fair stalls in and out. This door may have also allowed separate use of the chancel and nave for inhabitation.

Figure 5 – The western wall of the chapel which was constructed in the 19th century by G. G. Scott



Figure 6 –Western and southern walls of the chapel


Figure 7 – The medieval northern nave window and the western wall window of Scott. It is difficult to tell which one is medieval.

Today, the chapel is in good condition. This was possible due to the restoration works carried out in the 19th and 20th centuries, in part by George Gilbert Scott. Scott had a notorious preference for ‘restoring the whole church to the purest style of which traces remain’ [6]. Nonetheless, his reconstructed western wall is in perfect harmony with the rest of the church (see Figures 5 and 6). The west wall maintains the continuity of fine details such as the sawtooth patterned string course, the billet frieze and the shafted quoins. His central window, with its elaborate nook shafts and chevron mouldings, is a precise replica of the medieval northern window of the nave (see Figure 7). However, it is possible to distinguish the western wall of the nave from the southern and northern walls. The cobblestone that was used in the western wall is large and tightly packed and the mortar has a more cementitious appearance.

Figure 8 – The heavily modified eastern wall of the chapel

Not all of the restoration and rehabilitation efforts were as reverent as Scott’s. The east chancel wall preserves a detailed record of these less successful efforts (Figure 8). Cotman’s prints show a medieval lancet window here (see Figure 3), although today we find an unremarkable rectangular window. Just above this window is the billet frieze. Surprisingly, the level of the frieze on this wall is lower than on the northern and southern chancel walls, disturbing the continuity. The light cream coloured stones, which are above the existing window, may be indicating the original position of the lancet window, above which the billet frieze may have originally stood, in continuity with other walls.


Figure 9 – Various construction details of the eastern chancel wall

A closer examination of the upper part of the chancel eastern wall reveals further information (Figure 9). Here, different types of building material indicate phases of structural interventions. Above the billet frieze, stones appear to trace an arch. These stones may have been the remainders of the vault which originally covered the chancel, and which collapsed at the turn of the 14th century. Above this curved trace, the wall is of rubble construction. The triangular edges of this construction, where the wall meets the timber roof, is delineated. Interestingly, this triangular trace is marked twice, first by limestone, second by bricks. Does this indicate that the structure was roofed several times?

Figure 10 – Chancel arch, viewed from west. Photo is from [5].

Figure 11 – The vault responds in the chancel


Figure 12 – The nave roof


Figure 13 – The wall plate and post, placed inappropriately in front of a medieval nave window

On the other side of the eastern chancel wall, inside the chapel, we see two stubby and elegant columns, which are diagonally placed (Figure 11). These are the vault responds [5], which predated the timber roof that exists today. The current roof is reportedly from c. 1400, though as discussed above, there are some signs of later alterations. The roof consists of principal rafters, which are supported by wall plates and posts (Figure 12). The specific placing of posts on the nave wall is disruptive and structurally inefficient. In particular, it is hard to understand why wall plates and posts were placed right in the middle of a medieval window, causing its partial demolition (Figure 13). 

REFERENCES
[1] Bradley, Simon, and Nikolaus Pevsner. Cambridgeshire. The Buildings of England. Yale University Press, 2014
The website holds wonderful information on various ecclesiastical buildings.
[3] Cotman, J. S.  Antiquities of St. Mary's Chapel at Stourbridge, 1819.
[4] Pearce, Barry. A Brief History of the Leper Chapel, 2016. Pearce’s research on the chapel was presented with information boards during the 2016 Stourbridge Fair.
[5] Retrieved from the website: http://www.crsbi.ac.uk/site/1100/. The website features detailed information and high quality photos of Romanesque architectural details. 
[6] https://en.wikipedia.org/wiki/Victorian_restoration

Sunday, 2 October 2016

St Mary Magdalene, Ickleton




In St. Mary Magdalene, the parish church of Ickleton, there is a unique opportunity to see what churches must have looked like in the Medieval Ages [1]. However, there is nothing to signify the extraordinary interior of the church as you approach it from the outside.

Figure 1 - St Mary Magdalene viewed from south

Figure 2 - St Mary Magdalene viewed from west

The church nave is flanked on either side by aisles, which appear to have been constructed in different times. Notable is the central Norman door, which dates from the 13th century. It is also noteworthy that the southern aisle is considerably wider than the northern aisle. When we go inside, we find out that this section houses the organ.

Figure 3 - St Mary Magdalene broach spire viewed

Another interesting feature is the 14th century broach spire (reconstructed later in the 20th century). The features on it are quite eye-catching. There is a very unusual external bell, placed towards the top, at the end of a cantilever beam and covered by a canopy. The bell is reportedly 0.7m in diameter and 250 kg in weight. Only a single bell exists today, but in the past another bell was present symmetrically [2]. Considering that the spire skeleton is made of wood, it is likely that this additional weight from the bell is providing significant stability against wind loads. From simple calculations, it does seem that the bell weight is significant in comparison to the weight of the spire skeleton itself. This may explain why the bell was placed so high up in the mast externally. Its weight may have provided potential stability issues arising in the spire top, which is the most vulnerable area of the spire against high winds [3].























Figure 4 - An illustration of the St Mary Magdalene spire internal skeleton

 A weathervane is visible at the tip of the wooden spire mast, which extends all the way to the top of the transept tower. Here it is scarf jointed to wooden beams, which bear on to the tower. It wasn’t possible to see the construction of this spire but an illustration from the pamphlet provided inside the church reveals its internal construction.

Figure 5 - The northern doorway of the porch. Note the crack on the right

Figure 6 - The crown  of the porch viaduct. Note the tell tales.

Figure 7 - Eastern side of the porch viaduct, showing the tierceron ribs. Note the carved embossings. 

To enter the church, you pass through a 14th century porch. The porch is a covered tierceron vault with beautifully carved bosses of a placid Christ and an agonising virgin Mary. There are some clear signs of settlement on the northeastern corner of the vault (Figure 5), which has resulted in crack opening at the crown along the diagonal ribs (Figure 6). Also there is cracking where the tierceron ribs meet on the eastern side (Figure 7). It seems as if the response mechanisms which form on the ribs govern the settlement response, and this is quite distinct from the behaviour of a cross vault where Sabouret cracks could form [4].



Figure 8 -The nave of the church, looking west. 


Figure 9 -The Roman nave columns. 


Figure 10 -12th century wall painting depicting St Stephen's martyrdom

Finally, the glorious interior (see Figure 8, retrieved from Wikipedia). The nave is adorned with a beautiful arcade of semi-circular arches resting on cushion capitals of 12th and 13th century Norman and Roman columns (see Figure 9). According to the pamphlet, the arches are constructed of alternating blocks of Roman tiles and Barnack stone. I could not identify any particular reasons for this construction practice, except that it may have a created a strong visual effect. However, considering that the arch was typically covered with decorations (see Figure 10), achieving this visual effect is unlikely to have been the objective of masons. It is also noteworthy that in some of the arches, the clay tiles are not present, and the arch is entirely constructed of stone. The alternating red and white colour on the skin of these arches (see Figure 9), is given by the use of different mortars. The red mortar features crushed brick, from where it derives its pinkish hue.


  Figure 11 -12th century wall painting depicting the flagellation of Christ

This brings us to the 12th century cycle of wall paintings in the northern nave wall, above the arches.These describe a Passion cycle, with scenes from the Last Supper, The Betrayal, The Flagellation (see Figure 11) and Christ Carrying the Cross. Such paintings are extremely rare, and they were revealed only after an arsonist tried to burn the church down on 24 August 1979. It is fascinating that these paintings have survived. According to the pamphlet, this is partially attributable to the technique of ‘true fresco’, where the paint is applied when the plaster is still wet. This ensured durability, but required quick applications of paint and limited the colour palette (note for instance, lack of the colour blue in Figure 11). Therefore another layer of painting was applied once the plaster was dry (Fresco-secco technique  [5])  to complete the finer details (e.g. faces), but this layer has not survived.

REFERENCES
[1] Bradley, Simon, and Nikolaus Pevsner. Cambridgeshire. The Buildings of England. Yale University Press, 2014
[2] https://en.wikipedia.org/wiki/Ickleton
[3] Heyman, Jacques. The stone skeleton: structural engineering of masonry architecture. Cambridge University Press, 1997.
[4] Theodossopoulos, Dimitris. "Structural design of high gothic vaulting systems in England." International Journal of Architectural Heritage 2.1 (2008): 1-24.
[5] https://en.wikipedia.org/wiki/Fresco-secco 

Sunday, 28 August 2016

St Andrew's Church, Chesterton

This Grade-1 listed church is in the village of Chesterton, a suburb of Cambridge on the north bank of the Cam. The church is close to the town centre, yet due to its isolation from all other architectural highlights of the town, I had not run across it before.

The church has an interesting history that extends to the Norman times [1]. Information on the early churches on this site are scarce, however, the history of the church since the 13th century is well-documented. Some building work was carried out in the 13th century but the majority of the church that stands today was constructed in the 14th and 15th centuries in Gothic Decorated and Perpendicular styles with coarse stone and rubble. The nearby Chesterton Tower, which survives to date, was also constructed during this period to accommodate the priest. Some restoration and extensions of the church were done in 19th and 20th centuries.

I have highlighted below, with pictures, some intriguing architectural and engineering features from St Andrew’s:

The western tower and spire: 

A tall spire tops the 14th century tower of the church. The tower has a tapering structure, highlighted by the four levels of projected and downward sloping string courses. These string courses may have had practical uses; they might have been used for intermittent support of the scaffold tower during construction as well as playing a role in the drainage system of the tower. Up to 2/3rd of its height, the tower is supported by four offset angle buttresses, which are constructed in three steps. The placement of eastern buttresses is curious and highlights the later construction of the embattled aisles and porch. Two-light bell-openings are visible beneath the embattlements on the tower.


For me, the biggest curiosity here is the octagonal stone spire. The spire is almost as tall as the tower, however, its base covers only half to two-thirds of the tower footprint. This particular type of spire is called a needle spire, and according to [2] it is useful as it provides a pathway on top of the tower where scaffolding could be erected for repairs. The handlebars on top of the battlements suggest that this may have been the case during the restoration of the tower in 1968. Due to the needle spire arrangement, structural elements which transmit the load from the spire to the tower are necessary. [3] suggests this could be achieved with squinch arches. Indeed, squinch arches which spring from the level of the bell-opening arch may provide a circular footprint ( with the help of ring beams?) in which the octagonal spire could be seated.



The Doom: The oldest element in the church is the 13th century chancel arch, above which traces of a magnificent medieval painting remains. This is one of the few examples in the country as many similar paintings were destroyed during the Reformation. Over the exposed dome painting, there remains what I suspect are traces of later (and more austere) decorations. Traditionally, the Doom (or Last Judgement) paintings were strategically placed on eye-catching locations within the church to inspire (or intimidate) the congregation. The current painting is from the 15th century and extends into the clerestory. On its West and East sides it contrasts Heaven and Hell. On the western side elegant angels and straight and pale figures inhabit the New Jerusalem with its domes and towers. Underneath, a Tudor rose stands, stating the King as the head of the Church. On the eastern side, distorted demons are dragging common people, priests and kings out of their graves with much enjoyment. Nobody is spared. The thistle below is a liturgical symbol of sin and alludes to the crowning of Christ with thorns.






The arm rests: Beautiful carving of armrests of the pews with poppyheads, mythical beasts and mischievous standing figures is a secret pleasure of this church. The carving is playful and creative, the mythical beasts show a great variety (alongside raging wyverns lie obedient talbots) and look whichever way they please. According to [4], the standing figures provide useful evidence for dating the pews within ten years. It appears that fashion was changing quickly, even in the 1430s.


References
[1] https://www.standrews-chesterton.org/
See the official church website for well-curated information on the history of the church. Check the wonderful section ‘History of St. Andrews in 100 objects’.
[2] https://architecturaltravels.wordpress.com/2012/07/09/spire/
[3] Heyman, Jacques. The stone skeleton: structural engineering of masonry architecture. Cambridge University Press, 1997.
[4] Bradley, Simon, and Nikolaus Pevsner. Cambridgeshire. The Buildings of England. Yale University Press, 2014