Climate change is driving international heritage to reconsider the conservation strategies applied to traditional historic buildings. To date, the preferred repair method for rubble stone buildings in the UK has been pointing, and the style employed is the heritage joint: a recessed joint that highlights the profile of the stone with aggregates exposed. This paper challenges that conservation approach, by demonstrating that a bare stone method leaves historic fabric in Scotland vulnerable to water penetration, particularly in the north and west regions where wind-driven rain predominates. In addition, we argue that the heritage joint lacks historic authenticity, and that the accurate presentation of architecture is related to the original functionality of the building system. By turning to the archaeological record, we demonstrate that pre-industrial Scottish vernacular buildings were originally completely lime-coated, and that this was an effective way of managing moisture. By the 18th century, changing tastes expressed a preference for bare stone and ultimately this led to finishes being stripped and the joints between stones being pointed in many different styles, few of which are employed within the heritage sector. The strong winds experienced in storm conditions create external/internal pressure differences on windward elevations. The rain that accompanies wind is attracted into the internal space in these conditions where wall widths diminish, and where there is no external finish. In an innovative experiment, the actualistic pressure differences observed in storm conditions are created using a purpose-built airtight chamber and these pressure differences are accompanied by simulated rainfall. The results of the experiment show that pressure difference is a significant contributor to water penetration at the window embrasure. Crucially, the experiment demonstrates that completely lime coating rubble stone buildings moderates this ingress, while the heritage joint is the least effective barrier to water. The baseline for conserving and improving historic rubble building fabric is considering complete lime finishes as vital tools in building resilience.
The heritage industry recognises the significant challenges involved in adapting traditional pre-industrial rubble buildings to achieve lower thermal conductivity [1–3], improve internal comfort [4] and reduce energy use and running costs [5]. However, far less attention has been paid to whether the original design, functionality and presentation of rubble stone buildings have been accurately understood. This paper examines the impact and subsequent loss of external lime finishes and asks how their reintroduction would moderate water penetration in storm conditions when external/internal pressure differences create an attractive environment for water ingress. While the study is centred on Scotland’s architecture and climate, the results are applicable to other regions where vernacular buildings have been stripped of finishes and where extreme weather events are becoming increasingly frequent.
In response to the energy and climate crisis, Historic Environment Scotland commissioned a PhD research programme at the University of Stirling. As part of this work, a nationwide survey of 311 sites was undertaken to establish the spatial and temporal extent of external lime finishes in Scotland [6]. The survey characterised finish types, thickness and evolution, and was underpinned by detailed desktop analysis of archive material, early photographs, prints, paintings and rare but crucial documentary sources. First published in 2019 and updated in 2022 to include evidence for Scotland’s varying pointing styles, the results were compelling. Survivals of complete finishes were recorded from the Early Mediaeval period (300–1100 CE in Scotland) through to the 20th century in more isolated areas.
Fieldwork observations from this study, together with ongoing research undertaken by Historic Environment Scotland as part of its High Level Masonry Programme, provide evidence for Scotland’s diverse range of complete finishes, pointing methods, changes in building style and the resultant structural flaws. These findings form the baseline for the experimental work presented here. The research demonstrated that the thickness of harl reduced following the introduction of Palladianism in the 18th century [7], and that the use of complete finishes diminished over time. By the late 17th century, ashlar stone had become the preference of the elite, while by the late 18th century, Romanticism promoted the use of split and sawn stone in rustic styles [8,9]. These shifts enabled the development of Scotland’s varied pointing traditions [10], in direct contrast to the universal application of finishes prior to the Long Eighteenth Century (c. 1688–c.1850).
Scotland was found to have 16 distinct pointing styles, with the overwhelming commonalities being lining-out: the pressing of mortar while still green, using a purpose-made tool. All the examples were at least flush with the face of the wall and in most examples completely covered the joint (Fig. 1). No evidence was identified in the archaeological record for the now ubiquitous ‘heritage joint’ (Gerard Lynch, personal communication). This method involves filling open joints with new lime mortar and then recessing them with a stiff brush, clearly defining the stone edges and exposing aggregate to produce an artificially aged appearance (Fig. 2). The heritage joint emerged in the early 20th century as a technique for the conservation of ruins and was later adopted for roofed buildings [11], a practice that has continued since.
Some examples of the 16 Scottish lime pointing styles c. 18th–19th century.
Four photographs of stone walls, showing various pointing styles. One shows very wide mortar joints with just the high points one the stone visible, another shows fine joints with slate shims in and the other two show mortar joints that have been lined out with a purpose made tool while mortar was still soft.
Inverkeithing Town House c. 18th century, an example of the ‘heritage joint’.
Two photographs. One shows a stone building and the other shows a close up of a stone wall, with recessed mortar joints which have been brushed back to expose the aggregates.
Further important observations emerged from the fieldwork. Many buildings were found to have been harled or finished during construction [12], a practice that likely aided carbonation – the absorption of atmospheric carbon dioxide to form a durable mortar – demonstrated by the long-term survival of finishes in exposed locations. Less robust limewash was predominantly observed in sheltered areas. Limewash was also seen to extend over harled surfaces and fold around chamfered and roll-moulded details into door and window openings, creating seamless external surfaces.
The fieldwork concluded that there has been a widespread misunderstanding of both the functional and presentational value of rubble stone buildings in Scotland. Prior to the Long Eighteenth Century, all masonry surfaces were coated with harl or plaster. This finding challenges an embedded historiography that has directly informed practical conservation approaches [13,14].
The aim of this paper is to highlight an inherent flaw in the design and construction of Long Eighteenth-Century residential buildings and to examine the role of lime finishes in mitigating the resultant water penetration. In doing so, it demonstrates the inappropriateness of the heritage joint as a conservation methodology. The focus is on the reduced wall width below the windowsill, a repeated construction detail driven by the pursuit of classical symmetry. This change in practice created a vulnerable transverse joint beneath the sill and within the supporting masonry below, which, as the archaeology demonstrates, was managed by lime coats. These coats were crucial because during storm conditions, the velocity of wind is curtailed abruptly at the wall face, generating a pressure difference between the external and internal environments, and the difference in pressure creates an attractive internal environment for moisture in adverse conditions.
To investigate this mechanism, a purpose-built wall with a window opening – normally used for building-craft training at the Scottish Lime Centre Trust (SLCT), Charlestown, Fife, KY11 3EN (National Grid Reference: NT 06478 83813) – was employed. A timber airtight chamber was constructed within the window embrasure. External and internal pressure differences were replicated by extracting air from the chamber while applying simulated rainfall representative of storm conditions. Water ingress was quantified at 10 Pa pressure-difference increments. In response to the resulting inundation, a range of Scottish lime finishes was applied and their moderating influence assessed, before being directly compared with the performance of the heritage joint.
Wall morphology – the reduction in wall width and the resultant points of vulnerability
In tandem with the evolution of finishes, the survey noted that with the arrival of neo-classism in Scotland in the late 17th century, wall widths generally reduced from approximately 1.5 m to 600 mm (Fig. 3). In key areas, such as windowsills and lintels, masonry could be as narrow in section as 150 mm (Fig. 4). It was also noted that the number and size of windows increased throughout the period of the Long Eighteenth Century, sometimes accounting for approximately one third of the total volume of the elevation. This is problematic, particularly where multiple window openings existed on two separate but adjacent corners. By combining the visualisations of the section and plan it is possible to understand the profile of walls as a series of substantial piers of approximately 600 mm width, triangulated at the corners and linked by insubstantial ties approximately 150–200 mm wide under the windows. This design was universal within the context of domestic architecture.
Schematic plan showing a typical wall of the Long Eighteenth Century.
A diagram showing the plan of the walls of a building, with window openings. There are large stone blocks shown on the outer and inner faces, with rubble in the middle. The illustration highlights how the masonry reduces in thickness below the window making this area vulnerable.
Schematic section of a typical wall of the Long Eighteenth Century.
a diagram showing the section of a wall in a building, with a window and the stones that support the window, including a sill and padstone. The illustration highlights how the masonry reduces in thickness below the window making this area vulnerable.
A common observation during the survey was cracking of the external envelope around the windows of historic buildings that often extended from ground level up through the masonry, sills to crowsteps, skews and eaves tabling (Fig. 5). This phenomenon frequently occurred when raised margins with very fine joints and little tolerance for adjustment were introduced in the late 17th century (Fig. 6). Masons were required to place the sill on padstones located to the left and right of the window opening, to ensure that the sill was placed and remained level and in the exact plane of the wall necessary to accommodate the margins. The padstones might have been as small as a single piece of slate or a larger profiled section of sandstone (Fig. 7). The accumulated compressive load of the margins was therefore transferred to the padstones. In this scenario, the under-sill joint and the masonry below the sill was not under compression, because the rubble stone was slotted in from the face of the wall, rather than built from above. The under-sill area was especially vulnerable to water inundation because of the greater discharge from window glass. Both the sill and lintel have a transverse joint, being any joint in a single skin of masonry without a core and inner leaf. This is problematic because British Standard (BS) BS5628-3:2005 – a code of practice for the use of masonry, specifically focusing on materials and components, design and workmanship – makes clear that in no circumstances may a wall of 190 mm or less be left exposed, even in the most benign of conditions. Working within the code’s parameters, single-skin masonry would be regarded as an unsuitable construction without a coating. In short, the heritage industry expects once-coated, often deteriorated fabric in exposed conditions to perform like modern masonry with vapour barriers and render.
Examples of the cracking that occurs in and around the under-sill area, extending from ground level to crowsteps.
two photographs of a building, showing cracking patterns under and between windows.
Left: An example of a raised margin and the relationship between the sill and the lintel. Right: Examples of the fineness of the joints between each of the stones that form the margin.
three photographs. One shows the raised margin of a window surround, which casts a shadow onto the wall and the others show the very narrow joints of the stone of the margin.
Examples of how the sill was supported on padstones.
three photographs of walls under window sills, showing shallower and less regular stones in this area compared to the rest of the surrounding wall. The images illustrate how masons used padstones on each side of the window and placed the sill on these, filling in underneath retrospectively.Wind-driven rain and the impact of internal/external pressure differences
Scotland has some of the highest rainfall and wind speeds in the UK [15]. Spells of intense rain with a driving component are becoming more common and the relationship between the width of masonry units and a deteriorating climate have become crucial. Assessments have been made that examine the impact of wind-driven rain (WDR) upon building fabric [16–18]. The presumption in this work and others is that WDR is the agency that facilitates water penetration, as well as the capillarity of different stone and joint types [19]. This paper accepts the importance of WDR and capillarity. However, wind also creates pressure differences between the outer face of the windward elevation and the interior space, and that pressure difference between the two environments is also an impactful contributor to inundation within the context of non-airtight historic structures. Wind is the corollary of pressure difference (PD), and the experiment below compares the capacity of lime-harled and uncovered masonry surfaces to resist air and water penetration in a variable air pressure environment where wall thickness diminishes.
Furthermore, air distributed around a building may have differences in pressure. These differences are related to three primary factors: the geometry of the structure, the angularity of any additional features, and the height of a building and the roof pitch. Wind speed and its direction also influence air pressure differences. However, when wind direction is in line with a specific elevation, the pressure at that point is defined by:
P=Kpov o22,
where P is the elevation pressure point in Pa, K is the drag coefficient, vo is the wind velocity in metres per second and po the force of wind velocity in kilograms per square metre [20].
The air flow that envelopes a structure, rather than confronting a vertical face, does so as a dynamic flow with static pressure [21,22]. However, when low pressure air with an accompanying wind confronts the vertical face of the building, the wind velocity is curtailed, and stagnation pressure is created as the wind and its inherent kinetic energy [23] decelerates. The stagnation point denotes the height on a building elevation at which the wind is at its greatest velocity when it contacts the vertical face and divides around the sides and roof. In addition, complex building morphology – shape, height, sculpture, castellated parapet and roof pitch, for instance – determines the local pressure coefficient at any given point. The net pressure varies significantly around the building, from positive pressure on the windward side, to negative pressure on the leeward side. It is significant that the stagnation pressure point is in the upper area, because the evidence of water penetration in the upper area is ipso facto likely to be apparent lower down the elevation due to the forces of gravity. Identifying sources of penetration can be difficult, but single-skin, diminished masonry areas are obvious starting points in any assessment.
Air infiltrates and exits the historic building envelope through several places, mainly the roof, eaves, doors and windows [24]. Cracks in masonry may form additional ‘air leakage paths’ [25,26]. These cracks may be small – almost too small to be seen – but these can be the source of inundation in storm conditions where external pressure is greater than internal pressure (Fig. 8). Figure 8 illustrates a notional building facing southwest, with an inline directional wind, and a 45° pitched roof. The use of the illustration reflects the model used in the experiment below and is not described with a scale. Within the illustration, increasing positive pressure is indicated by orange, and stagnation pressure is indicated by red. Variable pressure on the roof is dependent on angle and indicated by yellow and green. Negative pressure on the leeward side and internal space is indicated by blue. Positive external and negative internal pressure differences are at their greatest in storm conditions when air infiltration on the windward elevation is less than the sum of air exfiltration on the side and leeward elevations [27–32]. During storms Ciara (8–9 February 2020) and Aiden (1–2 November 2020), gusting pressure differences of between c. 90 and 280 Pa were recorded within a roofed structure in Stirlingshire (Meek, personal observations), with wind speeds of between 70 and 80 mph and rainfall 100–177 mm in a 24-h period [33].
Schematic notional building facing southwest, describing pressure differences relative to changes in the building geometry.
a schematic drawing using colours to show pressure difference around a building. The building is shown on the right hand side and the air is coloured blue inside and blue and turquoise on the leeward side, indicating lower pressure. On the windward side the colours are red, orange and yellow, indicating higher pressure, with the highest in front of the wall face, lessening as it goes over the roofline.The aim of the experiment
The aim of the experiment was three-fold: first, to replicate actualistic environments where pressure is exerted on the windward elevation, and where there is uncontrolled air leakage from all other elevations and the roof combined; second, to critically evaluate the prevailing conservation practice of using recessed pointing with exposed aggregates – commonly referred to as the ‘heritage joint’; and third, the results of the evaluation were compared with traditional Scottish lime coats. To test the impact of different pointing and coating methods on moisture resistance, we conducted a controlled experiment simulating storm-driven water ingress. Specifically, we measured water permeability across three conditions: (a) a 465 × 0.065 mm under-sill crack, (b) an intact under-sill finished with high-calcium hot-mixed mortar and (c) natural hydraulic lime mortars.
Methodology
An experimental sealed chamber was constructed off site, then installed within a simulated window embrasure constructed of Dunaverig Old Red Sandstone wall measuring height (H) 2.3 m × width (W) 0.6 m at the SLCT. The total wall length was 3.6 m. A pressure difference across the window embrasure was created by extracting air from the chamber to create specific internal/external pressure differentials in accordance with the British Standard BS EN 13829-2001 ‘Thermal performance of buildings – determination of air permeability of buildings – fan pressurization method’. The method was scaled down and substituted with an industrial dust extractor, which recreated the pressure differences defined in the British Standard above. The construction and installation of the chamber, and all the building and associated lime work, were undertaken by the first author, who is suitably qualified and experienced. During construction, the side, floor and roof panels of the chamber were glued and screwed together. Prior to installation, the masonry was tooled flush internally to accommodate the chamber. The open face of the chamber was scribed to the profile of flattened masonry. Discrepancies between the open face of the chamber and masonry were subsequently filled with polyurethane expanding foam to ensure an airtight fit between the wall and the chamber (Fig. 9).
Left: The rear, ‘interior’ of the wall located within the window embrasure. Right: The ‘exterior’ of the wall.
One photograph showing an upright rectangular box representing the room interior and another photograph showing a stone wall with window in it, and the wall is wet below the window.Creating pressure difference, monitoring and pressure regulation
The current Building Regulations (approved document Part L1A 2010) recognise the phenomenon of air leakage within the context of new building and requires that a new dwelling should achieve an air leakage of 5 m3/hm2 or less, using a method known as the ‘blower door test’ (BDT). This tests that a new building is constructed to a recognised standard of air exchanges and must comply with British Standard BS EN 13829:2001 ‘Thermal performance of buildings – determination of air permeability of buildings – fan pressurisation method’. This is the currently accepted UK maximum pressure difference using the BDT as 50 Pa. Recognising the impact of more extreme weather events, the United States employs the American Society for Testing Materials (ASTM) standard for testing air leakage and has increased the BDT value to 75 Pa. In this experiment we downscaled the baseline value to 70 Pa. This value is considerably under the real pressure difference of 290 Pa recorded during Storm Aiden (1–2 November 2020, previously mentioned).
The downscaled BDT pressure difference was created using an SIP 1.5 industrial dust extractor, collection volume 183 m/h, motor: 230 V~50 Hz 1100 W (SIP Industrial Products, Shepshed, Loughborough, UK). This was permanently housed on one of the sides of the chamber (rear left) and fenced off for the duration of the experiments. The pressure was monitored with a calibrated digital manometer (Comark C9500, Fluke Precision Measurements Ltd, Norwich, UK), to assess actual pressure differences between the external environment and the interior of the chamber. The manometer expressed pressure differentials at 10 Pa intervals. The pressure was regulated within the chamber using a sliding vent that reduced or increased the pressure difference, dependant on the differences required, and the prevailing atmospheric conditions within the SLCT workshop. The chamber was provided with a glazed port at the rear for in-trial observations and a demountable side panel for access between pressurised runs (Fig. 10).
Left: The rear of the chamber with glazed viewing port, sliding vent. Right: Access to the chamber via demountable side panel.
Two photographs. One shows the inside of the chamber, with two long rectangular sections at the top and two shorter sections visible at the bottom. In the top left section is a glazed viewing port and in the right section is a circular metal vent. The other photograph shows a rectangular opening it the wooden chamber.‘Window’ design and construction
The simulated window embrasure on the ‘windward’ side of the experimental sealed chamber was closed with a timber frame covered in ply and then a secondary sheet of damp proof membrane. The panel was carefully scribed into the window opening to ensure a tight fit. The objective of covering the panel with membrane was to mimic the discharge of water from glass in an intense storm scenario. The window area was H 1.675 × W 0.9 m and the area under the sill was H 0.46 × W 1.25 m, with a wall width under the sill of approximately 0.190 mm.
‘Rainwater’ delivery system
Increased intense storms with high volumes or rainfall have been recorded over the last decade, with the highest rainfall in January 2015 peaking at 341 mm at the Honister Pass, Cumbria (Met Office). The objective of the rainfall simulation was to provide a realistic volume of water. An independent hose system, using an off-the-shelf garden irrigation system from the tap to the window, recreated a high rainfall episode. The water from the tap was restricted to 1.5 bar with the volume of water regulated at 2.78 L/min, which is comparable with Storm Desmond in 2015.
Water capture
Water was captured on the ‘inward’ face of the wall in a fixed trough made of galvanised steel with stopped ends and fitted to the floor of the chamber at 60 mm from the face of the wall. Water retention pads with a capacity of 340 mL, made from a combination of paper pulp and a propriety ‘super-absorbent’ polymer, were placed in the trough. Pads were weighed prior to each run and the combined weight of ‘dry’ pads was deducted from the combined weight of the pads/mat with water. The pads were placed and retrieved from the trough via the demountable side panel between each pressured run and subsequently weighed.
Construction mortar
Two separate experiments were run using different mortar mixtures, applied sequentially and employing the same wall. The mortar used in Experiment 1 replicated the construction mortar undertaken by the SLCT. It was made with Melville Gates (Angle Park, Ladybank, Cupar, Fife, Scotland) coarse sand and Natural Hydraulic Lime (NHL) to reflect the local hydraulic limes historically quarried and manufactured at Limekilns (Fife, Scotland). The hydraulic lime used in the experiment was manufactured by St. Astier and classified as compressive strength 2 MPa. The mortar used in Experiment 2 was a hot mixed mortar (HMM), meaning the sand and aggregates were combined with quicklime and sufficient water to produce an exothermic reaction. Melville Gates sand and Shap high calcium kibbled quicklime at a rate of 3 parts sand and aggregates to 1 part quicklime were combined by volume. On slaking, this produced a mortar with a binder ratio of c. 2:1 that closely resembles the high binder to sand aggregate ratios found in historic buildings [34].
The experimental stages were conducted sequentially. Experiment 1 (NHL mortar): at each stage – pointing, sneck harl (meaning a partial harl dashed only over the joints), harl and limewash – a three-month interval was left between the applications. Staff at the SLCT undertook the curing regime in accordance with best practice. The SLCT is a recognised lime training provider. The HMM experiment intervals were extended to six months, which better reflects the longer time required for a high calcium mortar to carbonate.
In both experiments, the term ‘open slot’ refers to the experimental wall with a simulated constructional crack, formed by sawing through the under-sill joint with a hacksaw blade < 1 mm, formed by filing down the blade. The term ‘closed slot’ refers to a wall with a simulated crack sealed, and the wall pointed employing the heritage joint.
Harl and sneck harl applications
Harl was applied in two coats at c. 10 mm, with sufficient time between applications for the mortar to ‘firm up’, but remain ‘leathery’, meaning that it was no longer a semi-liquid but still malleable enough to mould. The sneck harl was applied as a single coat of c. 10 mm thickness (Fig. 11).
Top left: Under-sill area with harl. Right: harl and limewash. Lower central: Sneck harl application. Lower left: The heritage joint application.
Four photos showing firstly the under-sill area with harl, the second shows harl and limewash, the third shows sneck harl application and in the fourth photo the heritage joint application.The experiments
The experiments examined water penetration through single-skin masonry below a window in two scenarios: the first being a wall constructed and finished with NHL mortars, which was then contrasted with the second, HMM alternative. Timings for individual runs were relative to the wall condition: open slot, closed slot, sneck harled, harled, and harled and limewashed. The open slot was tested at 2-min intervals, alternating between 0 and 10 Pa, to 0–70 Pa with six repeats for each 10 Pa increment. The closed slot was tested at 30-min alternating pressure differences. Sneck harled, harled and harled, and limewashed scenarios were tested at 60-min alternating pressure differences.
Statistical analysis
The impact of air pressure difference and wall covering treatment on water ingress was statistically analysed in the R Program for Statistical Computing version 4.3.2 (R Core Team, 2022). A generalised linear model (GLM) was used, assuming a gamma error distribution (GLM function in the stats package). We were not interested in the interaction between air pressure and treatment, and this was therefore excluded from the model, giving a final model (water ingress ~ treatment + pressure difference). Post hoc pairwise comparisons between treatments were conducted using Tukey’s honestly significant difference (HSD) test (lsmeans function in the lsmeans package).
Results
Increasing air pressure also increased the volume of water penetrating the wall in a statistically significant way (GLM: X21,250 = 5053, p < 0.001). Differences in water penetration were noted at low pressure differences with an open slot. In the NHL trial at 0 Pa difference, the rate of penetration was around 120 mL/min, increasing to around 380 mL/min at 70 Pa PD. In the HMM trial, water increased from 120 mL/min to 460 mL/min over the same pressure range. In both the NHL and HMM scenarios, the impact of pressure difference at 70 Pa is significant enough to be visualised through the viewing port. The presence and type of surface finish had a highly significant influence on the volume of water penetration (GLM: X27,250 = 110, p < 0.001), and the amount of water penetration was statistically significantly different between all lime treatments applied (Tukey’s HSD: p < 0.05). Water inundation was progressively reduced when the differing lime treatments were applied (Fig. 11).
With the slot filled, it was no longer possible to visually observe the increased flow of water penetration, even at 70 Pa, but the weight of the pads indicated that increased volumes were related to PD. By accessing the chamber during PD runs and observing the wall at floor level, it was possible to see water bubbling in the perpendicular joints at 70 Pa. This significant observation correlates with the difference in compression between bedding and perpendicular joints.
In the NHL pointing trial the penetration rate at 0 Pa difference was 7 mL/min, increasing to 12 mL/min at 70 Pa PD (Fig. 12). In the NHL sneck harl trial at 0 Pa difference, the rate of penetration was 4 mL/min, increasing to 7 mL/min at 70 Pa, an increase of 3 mL/min. In the HMM trial at 0 Pa, the rate of penetration was 2 mL/min, whilst at 70 Pa this increased to 7 mL/min, an increase of 5 mL/min (Figs. 13 and 14). In the NHL harl trial at 0 Pa difference, the rate of penetration was 2 mL/min, increasing to 3 mL/min at 70 PA PD. This compares to the HMM trial of 3 mL/min versus 5 mL/min for the same pressure change. In the NHL harl and limewash trial, as well as the HMM trial, there was no measurable increase in water ingress across the full pressure range. When the pressure difference was raised to 120 Pa (HMM), there was a significant exponential rise in inundation, which was particularly noticeable during the sneck harl model, rising from 2 mL/min to 14 mL/min. Full harl showed a lower increase from 3 mL/min to 5 mL/min (Fig. 15).
Water entering the chamber with NHL and HMM in an open slot scenario. Contrasting harl scenarios are also shown for comparison. Lines indicate linear regressions. The shaded areas represent the standard error of the mean. Lines indicate linear regressions.
A graph showing the difference between water entering the chamber, at various pressure differences, with contrasting harl scenarios shown for comparison. HMM open slot lets in the most water, followed by NHL open slot, with both of these rising dramatically when exposed to increasing pressure. With all of the other scenarios, such as harl, snack harl, harl and limewash, letting in very little water at all pressures.
Water ingress in an NHL wall with an open slot relative to pressure difference and the moderating response of NHL mortars.
A graph showing water ingress relative to pressure difference in a wall pointed with NHL 2, which shows it rising exponentially as the pressure difference increases. This is compared to sneck harl, harl and harl and limewash scenarios where the water ingress remains minimal and stable, even as the pressure increases.
Water ingress for different surface covers and across pressure differential. Data are identical to those in Fig. 4, but scaled to resolve low water ingress values with the varying treatments.
A graph showing the same information as figure 13, but enlarged for clarity to illustrate how varying treatments progressively help reduce the volume of water entering the chamber under increasing pressure. The open slot allows the most water in, while full harl and limewash only let in trace amounts of water.
Water ingress values for NHL and HMM treatments compare water ingress between the two mortar types. The HMM Sneck harl data series displays an exponential increase.
A graph comparing the effectiveness of the two mortar types, NHL and HMM. It illustrates that both mortar applications are effective at reducing water ingress, but there is a marginal improvement with NHL. The trend is reversed when full HMM and harl and limewash are applied, and no water entered the chamber. The graph also shows an exponential increase in water inundation when pressure difference was raised beyond the 70 Pa baseline.Discussion
The anomaly in the greater volume of water entering the chamber with HMM harl compared with NHL harl is counter-intuitive, given the current understanding of the benefits of hot-mixed mortars [35], but can be explained by the observations made before and during the pressurised runs. High calcium, high binder content mortars are prone to shrinkage as they cure [36], and if the mortar remains ‘plastic’ the process of shrinkage continues. The HMM mortars showed visible shrinkage cracks at all stages: building, sneck harl and harl.
In the NHL trial, shrinkage cracking was not visible and only became apparent when bubbling water at the perpendicular joints was seen during the pressurised trial from within the chamber. This stopped when the sneck harl was applied. The bubbling remained evident with the HMM at all stages until the harl was limewashed, at a pressure difference of 70 Pa and above, after which no measurable water entered the chamber. HMM shrinkage can be viewed as a contributor to water penetration within this context, which was resolved only when limewash filled the cracks (Fig. 16). The relatively poor performing NHL sneck harl may be attributed to changes in the build quality of under-sill masonry. The experiment inherited the wall, but taking down and rebuilding the under-sill area for the HMM trials was undertaken by the first author.
HMM shrinkage cracks indicated by the red arrows.
A photograph of a mortar joint with several micro-cracks marked up with red arrows
During the open slot stage of the experiment, two fans were introduced to replicate the impact of WDR. The fans were placed on the ‘outside’ of the wall driving ‘rain’ back into the window opening, sill and under-sill masonry. This had no impact on the volume of water entering the chamber – only when a pressure difference was introduced was there any increase.
Another significant observation within the study was the condition of harl and limewash when the pressure testing regime was concluded. While the HMM mortar and wash remained largely complete, they were damaged and friable. Had this been exposed and not protected, it is unlikely to have survived extreme wind or frost damage. Conversely the NHL mortars remained robust and without damage with only slight degradation to the limewash.
As this paper demonstrates, there are vulnerabilities in reducing wall widths between windows which are clearly visible, often from ground level and travelling up through the structure. Given that these represent movement, it is reasonable to assume that there will be smaller invisible cracking taking place. The water bubbles observed within the chamber are illustrative of the micro-cracking expected when buildings make imperceptible shifts.
Conclusions
The fieldwork on which this experiment is predicated demonstrates that there has been an underestimation of the extent of lime finishes in Scotland with no previous understanding of how they evolve into the various pointing styles. While the study has focused on Scottish traditional buildings, vernacular buildings anywhere in the UK that have been stripped of their finishes are vulnerable, particularly where wind and rainfall are high. Moreover, the experimental work demonstrates that anywhere in the world that experiences high winds and rain would benefit from the research. We created realistic external/internal pressure differences at well below those real pressure differences experienced in storm conditions, namely storms Aiden and Ciara. The key findings of the study are:
External/internal pressure difference is a major contributor to water inundation anywhere where wall widths diminish and where cracking has occurred, however small and irrespective of mortar type. There was a simple correlation between the volume of water entering the chamber and increased pressure difference. This was particularly marked in pressure differences of more than 30 Pa and exponential when exceeding the 70 Pa baseline employed for the bulk of the experiments. Innovative experimental testing therefore reveals that extreme weather amplifies pressure differentials that drive moisture into masonry.
The impact of pressure difference can be regulated by successively increasing the depth of coverage of lime coats, the most effective cover being full harl and limewash, the least effective the heritage joint, irrespective of binder type.
Notwithstanding the statement above, the NHL mortars proved to be significantly more durable and less prone to shrinkage at all stages. The experiment also revealed that the initial tamping of the HMM is an inadequate measure to stop longer-term shrinkage, because the mortars are not ‘fixed’ until fully carbonated.
To date there is little to compel building owners, local authorities, national amenity societies and statutory advisors to recoat historically lime finished buildings, even when the evidence from archaeological and desktop studies is compelling. There are no lime training centres in Scotland teaching anything other than the heritage joint. This experimental work brings together the cultural and physical evidence helping to inform policy which is necessary because there is a widespread misunderstanding of the role of finishes – lime finishes ensure that rubble buildings are seamless and less vulnerable to the impacts of WDR and pressure differences.
Open data and materials availability statement
All data generated or analysed during this study are included in this published article.
Declarations and conflicts of interestResearch ethics statement
Not applicable to this article.
Consent for publication statement
The authors declare that research participants’ informed consent to publication of findings – including photos, videos and any personal or identifiable information – was secured prior to publication.
Conflicts of interest statement
The authors declare no conflicts of interest with this article.
Artificial intelligence declaration
No AI tools were used in the writing or development of the manuscript.
ReferencesRyeCScottCThe SPAB research report 1. U-value reportLondonThe Society for the Protection of Ancient Buildings2010BergerJKadochBEstimation of the thermal properties of an historic building wall by combining modal identification method and optimal experiment designBuild Environ [online]2020185107065Available from:10.1016/j.buildenv.2020.107065Rodríguez-EstebanACamino-OleaSLlorente-AlvarezACabeza-PrietoASaez-PerezPBrick of the historical heritage: comparative analysis of the thermal conductivity, density and moistureMater Sci Eng [online]20211203032042Available from:10.1088/1757-899X/1203/3/032042Martínez-MolinaATort-AusinaIChoSVivancosJLEnergy efficiency and thermal comfort in historic buildings: a reviewRenew Sustain Energy Rev [online]2016617085Available from:10.1016/j.rser.2016.03.018WiseFGillichAPalmerPRetrofit information challenges and potential solutions: perspectives of households, retrofit professionals and local policy makers in the United KingdomEnergy Res Soc Sci [online]2025119103866Available from:10.1016/j.erss.2024.103866AddymanTMeekTTechnical paper 31: Historic external lime finishes in ScotlandEdinburghHistoric Environment Scotland20191240Gonzalez-LongoCThe architectural innovations of Mr James Smith of Whitehill (c. 1645–1731) within the European contextLoweryJHummLMackecknieAThe architecture of Scotland, 1660–1750 [online]EdinburghEdinburgh University Press2020191212Available from:10.3366/edinburgh/9781474455268.003.0011MacinnesR‘Rubblemania’: ethic and aesthetic in Scottish architectureJ Des Hist [online]199693137151Available from:10.1093/jdh/9.3.137McKeanCThe Scottish Chateau: the Country House of Renaissance ScotlandStroudSutton Publishing2001242258MeekTTechnical paper 33: Masonry pointing and joint finishingEdinburghHistoric Environment Scotland2022DaveyAThe care and conservation of the Georgian house: a manualAbingdon-on-Thames, OxfordArchitectural Press1978AdderleyPMeekTHarl-as-you-go! Integration of mortars and the implications for robustness and longevity in an exposed environmentJ Build Limes Forum201926820MeekTThe practical and cultural implications of a bare stone paradigm in heritage conservationConstr Hist2020647MeekTImplications of a bare stone paradigm in heritage conservationThe Magazine of the Architectural Heritage Society of ScotlandSpring20241014Met OfficeWhere are the windiest parts of the UK? [online] 1981–2010Available from: https://weather.metoffice.gov.uk/learn-about/weather/types-of-weather/wind/windiest-place-in-ukLaycockEAWoodCUnderstanding and controlling the ingress of driven rain through exposed, solid wall masonry structuresGeol Soc Spec Publ [online]2014391175191Available from:10.1144/SP391.1OrrSAVilesHCharacterisation of building exposure to wind-driven rain in the UK and evaluation of current standardsJ Wind Eng Ind Aerodyn [online]20181808897Available from:10.1016/j.jweia.2018.07.013OrrSAYoungMStelfoxDCurranJVilesHWind-driven rain and future risk to built heritage in the United Kingdom: novel metrics for characterising rain spellsSci Total Environ [online]2018640–64110981111Available from:10.1016/j.scitotenv.2018.05.354FusadeLVilesHA comparison of standard and realistic curing conditions of natural hydraulic lime repointing mortar for damp masonry: impact on laboratory evaluationJ Cult Herit [online]2019378293Available from:10.1016/j.culher.2018.11.011HuifenZFuhuaYQianZResearch on the impact of wind angles on the residential building energy consumptionMath Probl Eng [online]20142014115Available from:10.1155/2014/794650StephensonDGouldDBritish regional geology: the Grampian Highlands4th ed.Reprint 2007Keyworth, NottinghamBritish Geological Survey1995LinQIshidaYTanakaHMochidaAYangQTamuraYLarge eddy simulations of strong wind mechanisms at pedestrian level around square-section buildings with same aspect ratios and different sizesBuild Environ [online]2023243110680Available from:10.1016/j.buildenv.2023.110680MinzJKleidonAImbergerMBadgerJEvaluating the impact of kinetic energy removal by wind turbines on the technical wind energy potential of the German BightMeteorol Z [online]2025345777Available from:10.1127/metz/2025/1215Martín-GarínAMillán-GarcíaJAHidalgo-BetanzosJMHernández-MinguillónRJBaïriAAirtightness analysis of the built heritage–field measurements of nineteenth century buildings through blower door testsEnergies [online]202013246727Available from:10.3390/en13246727QuirouetteRThe difference between a vapour barrier and an air barrierBuilding Practice Note 54 [online]OttawaNational Research Council Canada1985Available from:10.4224/20338315QuirouetteROn-site exterior wall monitoring methods for air leakage, condensation and rain penetration control problemsConference paper to the Research Division, Canada Mortgage and Housing CorporationGovernment of Canada Publications1999p. 38VickeryBJInternal pressures and interactions with the building envelopeJ Wind Eng Ind Aerodyn [online]1994531–2125144Available from:10.1016/0167-6105(94)90022-1HolmesJDWind loading of structuresInternal pressuresLondonSpon Press/Taylor Francis e-Library 20042001128137SharmaRNThe fundamentals of building internal pressure dynamics induced through a dominant openingProceedings of the eighth Asia-Pacific conference on wind engineering, 10–14 December 2013 [online]Chennai, IndiaThe Eighth Asia-Pacific Conference on Wind Engineering2013269289Available from:10.3850/978-981-07-8012-8_INVI-03Al-MenyawiYMConcrete block masonry construction to resist severe winds [online]. PhD ThesisTexas Tech University2001Accessed 4 December 2025Available from: https://ttu-ir.tdl.org/server/api/core/bitstreams/94242d0c-0cdf-4338-b4e9-bf3b5e310f46/contentHumphreysMTGingerJDHendersonDJInternal pressures in a full-scale test enclosure with windward wall openingsJ Wind Eng Ind Aerodyn [online]20191892118124Available from:10.1016/j.jweia.2019.03.024WangWLVan de LindtJWQuantifying the effect of improved school and residential building codes for tornadoes in community resilienceResilient Cities Struct [online]2022116579Available from:10.1016/j.rcns.2022.04.001Met OfficeStorm Ciara [online]2019Accessed 4 December 2025Available from: https://www.metoffice.gov.uk/api/assets/file/2020_02_storm_ciarapdf?prefix=assetsLynchGThe myth in the mixLondonThe Building Conservation Directory2007WigginsDTechnical Paper 27Hot-mixed lime mortars: microstructure and functional performanceEdinburghHistoric Environment Scotland2018912Arizzi ACultroneGThe water transfer properties and drying shrinkage of aerial lime-based mortars: an assessment of their quality as repair rendering materialsEnviron Earth Sci [online]20147116991710Available from:10.1007/s12665-013-2574-x