experimental earth structures, renders and...

1

RESEARCH REPORT: EXPERIMENTAL EARTH STRUCTURES, RENDERS AND PLASTERS

RESE

AR

CH

REPO

RT

EX

PER

IMEN

TA

L E

AR

TH

STR

UC

TU

RES, R

EN

DER

S AN

D P

LA

STER

S

EXPERIMENTAL EARTH STRUCTURES, RENDERS

AND PLASTERS


EXPERIMENTAL EARTH STRUCTURES,

RENDERS AND PLASTERS

byTom Morton and

Rebecca Little

Published byHistoric Scotland

ISBN978-1-84917-203-5

© Crown CopyrightEdinburgh 2015


Historic ScotlandLongmore House Salisbury Place Edinburgh EH9 1SH Tel: 0131 668 8600 Email: [email protected]

While every care has been taken on the preparation of this publication, Historic Scotland specifically excludes any liability for errors, omissions or otherwise arising from its contents and readers must satisfy themselves as to the principles and practices described.


CONTENTS

FOREWORD iACKNOWLEDGEMENTS ii

1. SUMMARY 1

2. INTRODUCTION 3

3. THE TEST PROGRAMME 5

3.1 Internal Tests: Stanley Mills 53.2 External Tests: Culzean 53.3 External Tests: Battleby 53.4 External Tests: Fort George 6

4. OBSERVED DECAY 7 MECHANISMS

4.1 Shrinkage 74.2 Surface Erosion 84.3 Sacrificial Erosion 94.4 Freeze/ Thaw Cycles 104.5 Organic Growth 114.6 Delamination 114.7 Dampness 12

5. CONSTRUCTION TECHNIQUES 14

5.1 External Walls 145.1.1 Mudwall 145.1.2 Claywall 155.1.3 Rammed Earth 165.1.4 Stone with Earth Mortar: 175.1.5 Clay and Bool 185.1.6 Earth Brick with Earth Mortar 185.1.7 Turf 19

5.2 Panels 195.2.1 Earth Daub on Stake and Rice 205.2.2 Earth Daub on Horizontal Rails 205.2.3 Earth Daub on Horizontal Rope 205.2.4 Earth Daub on Kebber and Mott 21

5.3 Coatings 215.3.1 Earth Renders and Plasters 215.3.2 Lime Renders and Plasters 235.3.3 Cement Render 24

5.4 Washes, Paints and Distempers 245.4.1 Limewashes and Distempers 245.4.2 Dung Paint 255.4.3 Oil Paint 255.4.4 Tallow Paint 25

6. MATERIALS 26

6.1 Earth 266.2 Straw 306.3 Flax 306.4 Hay 306.5 Hair 316.6 Dung 316.7 Manure 316.8 Lime 316.9 Wood Shavings 316.10 Seaweed 316.11 Ash 326.12 Oil 326.13 Tallow 326.14 Whey 326.15 Lanolin 326.16 Urine 326.17 Blood 33

7. GUIDANCE FOR 34 CONSERVATION PRACTICE

7.1 Sourcing and Specifying Materials 347.2 Soil types and Methods of Analysis 347.3 Skills and Working Methods 35

7.3.1 Contractor Skills 357.3.2 Mechanisation 357.3.3 Souring 357.3.4 Programming and Protection 35

8. CONCLUSIONS 36

8.1 The Test Programme 368.2 The Scottish Traditions of Earth 36

Building8.3 Materials 368.4 Durability 368.5 The Construction Process 368.6 Repair and Conservation 368.7 Recommendations for Further 37

Research

9. GLOSSARY 38

APPENDICES 39


i

FOREWORD

Scotland has a strong tradition of earth building. As a readily-available material, earth was used extensively for vernacular construction across the country until the use of stone, brick and lime mortars became more popular on the 19th century. Examples of such structures survive throughout the country, illustrating the many types of earth construction and their regional diversity. Today it is not uncommon for earth structures to be unexpectedly revealed beneath subsequent lime or cement render coatings where they had not been previously recognised, often discovered during building works.

Although earth has proven itself to be a robust material, there is still limited information available about the conservation and maintenance of these structures. Historic Scotland commissioned this research project in order to gain a better understanding of historic construction techniques and decay mechanisms, and ultimately provide guidance for maintenance and repair. In addition, it is hoped that a better understanding and appreciation of the material and its role in our heritage may lead to new and innovative ways to incorporate earth in new construction.

Despite the decline in the use of earth in Scotland from the 19th century, it is estimated that over a third of the world population presently lives in dwellings of earth construction. With the current emphasis on reducing carbon emissions and increasing sustainability there has been a revival of interest in the use of earth as a construction material. Its widespread availability and local sourcing, plus its ability to be relatively easily worked with relatively little processing mean earth is generally considered to be a sustainable building material (although like all natural resources it must be responsibly sourced). When appropriately used and adequately maintained, earth can provide strong and durable structures capable of withstanding the harsh Scottish climate.

The research for this publication took place between 1996 and 2004 at a number of test sites in different part of Scotland where the authors set up a number of purpose-built test walls replicating types of construction observed in traditional buildings. Over the eight years, observations were made on a range of performance criteria including deterioration when exposed to weather and the behaviour of earth alongside other materials such as lime and cement renders. This report carefully documents the compositions of earth materials used for the test walls, the construction types and techniques, and detailed observations of their performance over time.

As with other traditional materials in Scotland, the widespread use of more modern construction materials and techniques has led to a gap in the knowledge and expertise required to work with earth structures. Historic Scotland hopes that the dissemination of the results of this research will help to provide a better understanding of earth as a building material and assist the conservation and maintenance of one of Scotland’s most important, yet poorly understood, vernacular materials. As our understanding grows, there will hopefully be a resurrection of the skills required to successfully create new earth structures and to maintain our existing earth-built heritage.

Dr Ewan Hyslop Head of Sustainability, Research & Technical Education, Historic Scotland May 2015


ii

ACKNOWLEDGEMENTS

We would like to acknowledge the time and effort of the many people who contributed to this testing programme, including the many practical volunteers who assisted in the test constructions, staff at the National Trust of Scotland, Culzean Castle; Scottish Natural Heritage, Battleby; Historic Scotland, Fort George; and the Errol Brick Company.

We would like to thank Bob and Jo Little for their support during the writing of this report.

Comments on the historical context of earth construction techniques in this document relies on the work of Bruce Walker and Chris McGregor, but the interpretation put on this data is the authors and does not necessarily reflect the views of Walker and McGregor.

Where commercial companies or products are referred to in this document, the information is provided in good faith, but the inclusion of any particular firm or product does not imply endorsement by the authors or Historic Scotland.


1

1. SUMMARY

A wide-ranging programme of more than two hundred and thirty tests, carried out over seven years, has explored and assessed the traditional Scottish techniques of building with earth, gaining a better understanding of historic construction techniques, the processes of decay that affect surviving earth structures and the appropriate use and specification of materials for repairs and new work.

The programme demonstrated that the commonly found traditional techniques are ones that are practical to build with readily available materials, and, if properly constructed, prove durable in the Scottish climate. Less common regional techniques, developed out of locally sourced building materials, were also successful, but required specialist skills. Techniques from outside the tradition were more difficult to build successfully and often proved less durable.

A range of subsoils were demonstrated to be successful as the basis for construction materials, with grading proving the most important factor in their selection. Expansive clay content was not demonstrated to effect durability. Six distinct decay mechanisms were observed, including shrinkage, surface erosion, sacrificial erosion, freeze/thaw cycles and delamination.

When other materials were combined with earth, their degree of compatibility had a significant effect on durability. Earth mixed with more dense materials was more vulnerable to decay than monolithic earth construction techniques. Lime coatings were demonstrated to have fundamentally different properties from earth backgrounds, raising questions over current practice.

Successful earth construction was demonstrated to require skill and experience in the selection, preparation and application of materials. There is a small skills base within Scotland, capable of such work, and this should be fostered through use in appropriate projects.

Earth construction is a field with considerable potential for development in conservation knowledge and practical applications. This programme highlighted four areas that would merit further technical research:

• The tests were based on a knowledge of traditional practice that is recognised to be incomplete. There would be benefit in assessing the results of these tests in the context of current practice and historical precedent within the U.K. and internationally, to gain further technical understanding.

• The tests indicated that earth renders may have significant potential as a technically appropriate coating material, though the complex factors influencing a successful render are poorly understood and merit further research.

• The tests indicated that lime renders can have significant technical deficiencies as coatings for earth materials. These are currently widely used in such circumstances and further research would help to develop appropriate best practice guidance.

• The tests highlighted a number of issues relating to the practical repair and conservation of weathered earth materials that would benefit from further research to develop appropriate techniques for application in the field.

This report, the photographic archive and the test walls that are still standing, are a valuable information resource of long-term relevance which can inform practical conservation projects and future research into earth materials.

2


Fig. 1: MAP of Scotland showing the three site locations

LEWIS

ISLAY

Soun

d of Ju

ra

Firth

of Cl

yde

ORKNEY

Fort George

Culzean Castle

SHETLAND

Kirkwall

Lerwick

Stornoway

Inverness

Aberdeen

St Andrews

Melrose

Dumfries

Stranraer

Ayr

Glasgow

Dundee

Perth

Edinburgh

Moray Firth

Firth of Forth

MULL

Stirling

Oban

Battleby Centre

3


2. INTRODUCTION

This report presents the findings of an experimental research programme to test the performance of earth building materials, which was carried out by Historic Scotland between 1996 and 2004.

The programme was initiated following a renewed awareness of the importance of earth materials in traditional construction, as described in TAN 6: Earth Structures and Construction in Scotland (1996). It was a specific response to the recognition that the complete cessation of the traditional use of earth materials in Scotland by the early 20th century had left a knowledge and skills gap that limited the technical understanding of surviving earth structures and hindered the development and implementation of appropriate conservation practice.

A wide programme of field tests was established to replicate traditional construction techniques using a representative range of bioregional materials. These included:

• twenty-nine internal test walls of panel construction, comprising a range of earth daub on timber armature and rope techniques

• seventy-eight internal tests of coatings, comprising a range of earth and lime techniques

• thirty internal tests of washes, comprising a range of lime, oil, dung, tallow and other materials

• twenty-four external test walls of monolithic construction, comprising a range of mudwall, rammed earth and claywall techniques

• thirteen external test walls of masonry construction, comprising a range of stone with earth mortar, earth brick with earth mortar and turf block techniques

• fourteen external test walls of panel construction, comprising a range of earth daub on stake and rice, horizontal rails and ropes and kebber and motte techniques

• twenty-nine external tests of applied coatings, comprising a range of clay renders, lime renders, harls and stuccos

• eleven external tests of washes, comprising a range of lime, oil and dung materials

The tests comprised small sample walls and panels and all the tests were monitored and assessed for their performance during construction and initial durability. The external tests were monitored and assessed for their durability over a period of up to seven years. These tests were carried out at three external sites, chosen to represent a typical range of climatic conditions (Fig. 1).

The programme was based on the sparse archival sources of information and rare surviving physical examples of traditional technique. These are often incomplete or difficult to interpret and it was recognised that the lack of certainty over some traditional techniques and materials, together with the lack of contractor experience and the limited physical size of the tests, would lead to some variation in the accuracy of representing traditional construction. However, it was intended that the diversity of tests would provide the data necessary to develop a greater technical understanding of materials and techniques, and that the programme as a whole would, in itself, develop construction skills, both in individuals and in the conservation community as a whole.

This report presents the results and conclusions from this research, interpreting the test data in cognisance of the limiting factors described above, and taking into account ongoing experience in the conservation of earth buildings and other related research that has occurred during the course of the programme.

The report includes a limited presentation of the results of individual tests as Test Data Sheets (Appendix A). These are intended to act as a gateway to access the extensive data record of the programme, so that its detailed experience of individual materials and techniques may inform individual conservation projects or other related research.

The associated archive of monitoring photographs and inspection reports is available for view, by arrangement, in the library of Historic Environment Scotland, Longmore House, Salisbury Place, Edinburgh, EH9 1SH.

Two of the three principal external tests sites are still standing at the time of writing this report, and will be subject to quinquennial inspection. These present a valuable ongoing resource of information on material, technique and weathering, which may also be accessed, by arrangement.

It is intended that the information contained in this report and its associated archives will assist in the understanding of surviving earth structures, of how they were built and the decay mechanisms that act upon them, and suggest appropriate approaches for their conservation and restoration.

4


Table 1: TEST SITE CLIMATE DATA

Site: Culzean Battleby Fort George

Altitude: 40m 50m 5m

Distance inland: 300m 41km 200m

Rainfall: Annual average 901mm 738mm 593mm

Min. monthly average 47mm 43mm 33mm

Max. monthly average 100mm 82mm 75mm

Daily bright sunshine: Annual average 4h 3.6h 3.5h

Min. monthly average 1.3h 1.1h 1.2h

Max. monthly average 6.8h 6h 5.6h

Temperature: Average daily max. 12°C 12.2°C 11.6°C

Average daily min. 5.4°C 4.5°C 4.7°C

Absolute max. 29.8°C 30°C 30.6°C

Absolute min. -11.7°C -18.9°C -16.7°C

Days with snow lying: Annual average 5.9 15.1 16.5

Min. monthly average 0 0 0

Max. monthly average 2.3 5.7 5.6

5


3. THE TEST PROGRAMME

An initial programme of tests was carried out at one internal site, focusing on experimentation with types of internal partition construction. Tests were carried out on a wide range of test panels, with different armatures, daubs, coatings and finishes. Some testing of earth floors was also undertaken. These tests were assessed for their workability, with initial shrinkage and durability monitored for several months. These tests informed the selection of materials and techniques used on the external tests.

The main programme of tests was subsequently carried out at three external sites, selected to provide a range of regional climatic conditions, typical of Scotland’s northern European, maritime environment. At each site, walls and renders made using identical materials and processes provided a control mechanism, allowing the relative effects of climate to be isolated. A wide variety of regional soils and techniques were tested alongside the control examples, forming the majority of the programme. Details of all the test panels, including specification of construction materials are given in the Appendix (page 41). Construction methods are described in Chapter 5, and further details of materials are given in Chapter 6.

Apart from repairs to minor initial frost and mechanical damage, maintenance was generally not carried out. There was some growth of vegetation around the tests and accumulation of decayed material. There was also evidence of minor vandalism on one site, with the early loss of some vulnerable tests.

The external tests were inspected and photographed once or twice a year for six years. At one site the tests were removed and this process was monitored as part of the programme.

3.1 Internal Tests: Stanley Mills

Fig. 2: Stanley Mills test site

The test site at Stanley Mills (Perthshire, NO 114 328), was inside a large disused rural mill building. A total of twenty-nine test panels were built in February 1996 and dismantled in the summer of 1997.

Seven types of armature were tested including stake and rice, horizontal rails, horizontal ropes, vertical spars, rails and ropes and rope lattice. Four types of earth and two types of fibre were used in the daub tests. Fifty-four types of earth renders and twenty-four types of lime renders were tested. Thirty types of washes were tested. Two types of earth were tested as floors.

3.2 External Tests: Culzean

Fig. 3: Culzean test site

The test site in the grounds of Culzean Castle (Ayrshire, NS 230 098), owned by the National Trust for Scotland, had the highest rainfall of the three sites, but was least prone to frosts. The test site was in a sheltered and shaded woodland situation, which would modify the standard climate data given in the table below in respect of sunshine and wind speed.

Ten external test walls and three test panels were built in June 1996 and eight test coatings and two test washes were applied in June 1997. The tests were still standing in March 2004.

3.3 External Tests: Battleby

Fig. 4: Battleby test site

6


The test site in the grounds of The Battleby Centre (Perthshire, NO 086 291), regional headquarters of Scottish Natural Heritage, is an inland, open rural situation with moderate exposure and most prone to frosts.

Twelve external test walls and four test panels were built in June 1996 and nine test coatings and two test washes were applied in June 1997. The tests were dismantled in 2003.

3.4 External Tests: Fort George

Fig. 5: Fort George test site

The test site at Historic Scotland’s yard at Fort George (Inverness-shire, NH 774 567) is coastal, and severely exposed.

Eleven external test walls and two test panels were built in June 1996 and ten test coatings and four test washes were applied in June 1997. The tests were still standing in March 2004.

7


4. OBSERVED DECAY MECHANISMS

A variety of common decay mechanisms were observed during the programme. Their combined affect on individual tests can be charted by following the relevant series in the detailed photographic archive. This chapter considers the individual mechanisms in isolation.

In analysing the results, the type and rate of decay observed on an individual test wall was relative to the nature of the test programme as a whole and numerous specific factors influencing individual tests, including quality of construction. In general, the rates of decay that the tests produced were accelerated in comparison to what would be experienced by a standing building.

The generalised descriptions of individual decay mechanisms given below take these variable factors into account and are presented so as to inform assessments of buildings and structures. However, in referring to the data and photographs of individual tests, it is important to recognise that these do not replicate exactly what would happen to earth structures outside the test conditions, for the following reasons:

• The small size of the tests meant that individual areas of weakness, such as corners of walls and edges of panels, would have a disproportionate effect on the rate of decay of the wall as a whole.

• Some of the walls were built with a greater batter than is found in traditional building, in order to achieve stability during construction, which created a more vulnerable surface.

• The roof protection was often less than in traditional construction, with some walls projecting beyond the roof line and some of the panels not receiving roofs until a year after they were built.

• Decay mechanisms act differently on a building than they do on the test walls and panels. Buildings are exposed only on the outside, whereas the tests were exposed on all sides, leading to accelerated erosion.

• Walls in buildings, particularly when heated or inhabited, have much more complex internal and surface patterns of moisture and thermal movement, which affects their durability, particularly in relation to coatings.

• Building walls are also liable to abrasion through use, rising damp, splash back and flooding, none of which were tested as part of this programme.

• Finally, apart from some initial minor repairs, the tests were not given any maintenance, which has a critical effect on long-term durability, particularly of the washes and paints.

However, it should also be recognised that, for the same reasons, the tests results may reflect more accurately decay mechanisms as they relate to incomplete structures and monuments.

Fig. 6: Initial drying shrinkage

4.1 ShrinkageAll the tests displayed the effects of initial drying shrinkage to varying degrees (Fig. 6). In the worst cases, cracks developed through the depth of the test wall and large sections of material fell away, but often there was only minor and superficial surface crazing.

The interface between lifts in monolithic constructions, and to denser materials in masonry constructions, could also be a point of weakness. If the head of the lower lift had dried, the upper mix could subsequently shrink away as it dried, leading to a break in bond, vulnerable to erosion. This is unlikely to occur in practice, where time between lifts tends to be short, but it can affect repairs.

Similarly, earth mortars could shrink away from masonry, though this effect was reduced if the masonry was dipped in water before use.

Generally the denser walls experienced the least shrinkage and the thinner panels experienced the most. Other factors influencing the degree of shrinkage included expansiveness of clay, proportion of aggregate, proportion

8


and type of fibre, moisture content of the mix, solar orientation, wind exposure, surface texture and, in the panels, type of armature.

The types of clay were not analysed and so it is not possible to determine what effect these had on the tests, though one of the earths is believed to have had significant expansive clay content.

It was generally observed that higher proportions of aggregate resulted in reduced shrinkage and conversely, that higher proportions of fines would lead to increased cracking. An uneven mix would lead to localised cracking, while well-graded mixes performed well.

It was generally observed that higher proportions of fibre resulted in reduced shrinkage, with 50%, by loose volume, the maximum workable proportion of fibre that could be added. The flax fibre had much less impact on shrinkage than the other various straws, which all had similar levels of effectiveness.

It was generally observed that higher moisture contents in mixes resulted in increased shrinkage, though this did not seem to be a significant factor for earth plasters. Controlling the amount of added water to the minimum necessary to produce a mix with good working qualities requires experience, but, at least for walls, it is an important factor in determining the final quality of the work and its long-term durability. Difficulty in achieving such control was a particular problem in shuttered walls constructed during sporadically wet weather, where rainwater gathered in the shutter and the drying effects of the sun and wind were inhibited.

Souring, that is allowing the mix to mature some days prior to use, proved very effective in reducing the amount of water required to produce a workable mix (section 7.3.3).

Solar orientation was a significant factor in the thin panel tests, where uneven drying resulted in greater cracking on the shaded side than on the sunny side. There was a similar effect on the renders and washes, but to a lesser degree. The use of shutters, kept in place after construction, had a similar shading effect, exacerbated by the lack of surface air movement, which inhibited evaporation.

Wind exposure could be a factor in warm, dry conditions, where it would accelerate surface drying on the windward side. This could lead to rapid drying and some cracking.

It was generally observed that rougher surface textures resulted in reduced shrinkage, the bigger surface area with higher porosity facilitating evaporation. The daubs with large fibres, such as wood shavings, had a large surface area and performed well in this respect. With smooth surfaces, a surface tension could develop which promoted cracking. Although the shuttered tests tended to produce a smooth surface, this did not generally cause cracking, as the rate of drying was slow and even.

Roughness was only beneficial if it occurred on the surface of a homogenous material. When it was caused by uneven mixing, the results were poorer.

Most daubs were polished with a trowel the day after application, tightening the surface, but some had to be pressed back for several days to inhibit cracking. The scoring of daubs in a large diamond pattern proved surprisingly effective in reducing shrinkage considering that the increased surface area was negligible. The cracks were also less obvious as they were concealed within the score lines. The surfaces of historic internal daubs do not generally appear highly worked and it seems likely that the initial mix and application were well chosen and did not require a large amount of aftercare.

Significant crack patterns commonly developed in the thin panels, which directly related to the underlying armature. The critical factor was the rigidity of the background. In the timber and willow armatures, the size of the cracks directly related to the size of the armature material, while the flexibility of the rope backgrounds allowed the daub to shrink as one block, with minimal surface cracking, but greater edge cracks.

4.2 Surface Erosion

Fig. 7: Surface erosion

All the external tests experienced erosion of the surface to varying degree (Fig. 7). The factors influencing the degree of erosion included microclimatic exposure, roof protection, slope of surface, material grading, surface coatings and repair materials.

The exposure and microclimate varied on the test sites, leading to site-specific erosion patterns, corresponding to prevailing intensities of wind-driven rain.

At Culzean, the SE elevation was consistently the most eroded, though this was only light compared to the other sites. There was generally very little weathering on the other faces.

At Battleby, the SW and SE elevations were consistently the worst eroded, the NE had minor erosion and the NW was virtually unweathered.

9


At Fort George, the NW and SW had the worst surface erosion, the NE seemed to suffer most from frost damage and the SE was most sheltered.

The degree of protection to elevations afforded by the roof is directly related to the size of the overhang. Generally, on exposed elevations long-term protection was given at an angle of 45 degrees from the roof edge. Where exposure was mild, a much greater degree of protection was observed. Where the wall projected beyond the line of the roof, there was significantly accelerated erosion below the drip line. Where the roof shed rainwater unevenly, specific severe erosion patterns could be related to routes of channelled water.

Some of the walls were built with a greater batter than is normally observed in traditional structures, for a variety of reasons related to their construction processes. In these situations, a significant slope would encourage accelerated surface erosion. Similar acceleration can occur in buildings where walls slump or suffer structural movement.

Well-graded materials were much better able to resist surface erosion than poorly graded materials. In well-graded mixes, it was commonly observed that there would be an initial washing away of surface fines, which filled in initial minor shrinkage cracks. This created a tight, even and stable surface with exposed aggregate and fibres, which suffered only very gradual erosion. Where the material was unevenly mixed, local erosion patterns would become established, resulting in an uneven surface. Where the material was poorly graded, a stable surface would not become established and erosion would be progressive.

Render coatings prevented surface erosion, while they remained in place. Washes provided degrees of enhanced surface durability, which varied from extremely resilient to completely ineffective.

Where sacrificial erosion became established at the interface with another material, a friable edge could become established on the earth surface, facilitating surface erosion.

Where the mix was of good quality, initial erosion patterns stabilised to a tight surface after a year and this created a sound, but textured surface which was ideal for the application of renders. Fresh surfaces, by contrast, could be smooth, with a lot of surface fines and this inhibited adhesion.

4.3 Sacrificial Erosion

Fig. 8: Sacrificial erosion, large scale

Fig.9: Sacrificial erosion, small scale

10


A recurring decay pattern, observed at a variety of scales, was the sacrificial erosion of earth where it abutted another material.

This effect was most dramatic on more vulnerable walls, where the base of the earth wall could be significantly undermined immediately above the masonry plinth (Fig. 8). Even in situations where the material otherwise performed well, erosion could become very severe in this location, threatening the stability of the wall. The relatively dense plinth inhibited the downward migration of moisture and, as a result, the earth could remain damp for extended periods, leaving it more vulnerable to erosion and frost damage.

The walls that did not have masonry plinths, stone with earth mortar, turf block and kebber and mott, also suffered most decay at the base. However, this was attributable primarily to the increased exposure to wind driven rain and the sacrificial erosion above the plinths on the other walls could, by comparison, be demonstrated to be a distinct and separate decay mechanism, exacerbated by the increased exposure.

Plinths are a traditional device protecting against rising damp and splash back from roof rainwater. In these tests, most of the plinths were constructed of dense concrete block, which will have focused this decay mechanism. However, in the walls that had less dense, sandstone plinths, their more porous nature led to greater retention of moisture, facilitating decay by other means. In a wide variety of historic earth structures, the interface to the plinth is an area where decay if often observed.

The same pattern appeared on a much smaller scale at the junctions between masonry and earth mortar (Fig. 9), though this was less with air-dried earth bricks than with stone, and was mitigated by dampening of the masonry immediately prior to use.

Where a poor bond was achieved between lifts in monolithic constructions due to differential shrinkage, the interface between lifts could demonstrate a weak sacrificial erosion pattern.

Surface repairs using denser, less vapour permeable materials, such as cement, accelerated local erosion of the earth material around the repair. Dampness was retained in the adjacent earth, making it more prone to erosion, which developed progressively as cracks and pockets between the different materials trapped further moisture and a friable, decaying edge became established on the earth.

4.4 Freeze/Thaw Cycles

Fig. 10: Freeze/thaw decay

All external tests were exposed to freeze/thaw cycles, which caused severe local decay in certain specific conditions (Fig. 10). Dampness and orientation were the critical factors. Where the earth was sufficiently damp, ice crystallised at low temperatures, expanding to form cracks and fissures. As the ice melted, the structure of the earth material broke down along these lines of weakness, developing a crumbly open texture, vulnerable to further decay.

Lower parts of walls on all sites suffered frost damage in the first winter, some six months after construction, indicating a potential need for initial targeted protection. Local repairs were subsequently carried out the following summer, usually in the same materials. Thereafter, significant decay only occurred at low levels on NE elevations in exposed situations. This was largely confined to the Fort George site.

Two types of walls were more affected than others. On the stone with earth mortar, the decay focused on the earth joints and this was probably related to the reduced ability of the wall to dry out, due to the denser facing material. The largest monolithic walls also seemed to have more frost damage. This may have been partly related to their larger mass and the fact that they were built on porous sandstone plinths. It is thought that some of these walls may have taken up to three years to dry out in their core.

11


The critical factor in vulnerability to this decay mechanism was high levels of retained dampness in the earth material. NE elevations were least exposed to the drying effects of the sun and the earth could remain locally damp for extended periods, especially at low level, where some tests walls were never observed to become fully dry.

Although the prevailing weather at the Fort George site came from the SW, this site suffered occasionally severe wind driven rain from the NE. This had a disproportionate effect, as drying on this elevation was very protracted, leaving the wall vulnerable to frost attack.

Grading can have be a factor in frost damage. Sandier mixes with more voids can be more able to absorb the expansion of ice crystals than mixes with more fines.

Surface coatings were extremely effective in reducing retained moisture and therefore inhibiting frost damage. However, softer coatings could themselves be vulnerable in the worst conditions.

4.5 Organic Growth

Fig 11: Organic growth

A variety of organic growths developed on the tests, though these did not generally appear to cause significant decay in the short-term (Fig. 11). Prolonged high moisture levels and the presence of organic additives were the critical factors.

A number of the tests developed mould growths in the initial period after construction. This particularly affected the internal panels, which were less exposed to the drying affects of the sun and wind. As the tests dried out, the moulds disappeared and they did not cause any decay. The dung additives seemed to encourage mould growth, but the animal fats had no apparent affect.

A number of tests contained cereal grains within the fibres, which sprouted in the initial period after construction. As the test walls dried, growth was not sustained and the plants died. No significant decay resulted. This occurred principally in the mixes containing the old type of organic wheat straw. The other straws were produced by modern threshing methods which are much more efficient in removing the grain. The flax and hay did not contain seeds.

Some of the turf constructions continued to grow on exposed faces for a period after construction. These also died after the initial drying out and did not result in deterioration of the wall.

Grasses and other plants colonised some tests. This only occurred when significant decay had already taken place, creating damp open surfaces where windblown seeds could germinate. These plant growths could create a secondary decay mechanism, through root growth and raised moisture levels.

Fungal growths became established on some of the timber armatures, but only after the daub had eroded and only where the timbers were in direct contact with the ground. This was a secondary decay mechanism that would eventually have resulted in the complete decomposition of the armature.

4.6 Delamination

Fig. 12: Delamination

A number of the tests experienced delamination in a variety of forms.

12


All coatings that failed to establish a strong mechanical key would detach in sections at the interface to the background, when weakened by other decay mechanisms. Factors contributing to an inadequate key included friable, dusty and smooth background surfaces.

Lime and cement coatings could detach from the earthen background while retaining integrity as a sheet (Fig. 12). This did not happen with earthen coatings. Although the lime coatings were microporous, they were less so than earth backgrounds, which readily absorbed and desorbed moisture. A relative built up of moisture could develop at the interface, leading to decay of the earth surface and loss of mechanical bond.

The consistent occurrence of vertical cracks at the corners of such coatings was an indication of differential movement between lime coatings and earth backgrounds. The lime coatings would have experienced some initial shrinkage, but significant cracks did not develop at this stage. Traditional buildings tend to have more rounded corners than these tests, spreading differential movement stresses, but they can still display cracking in this location.

The spectacular failure of the cement render was an extreme example of the same process, where the rate of decay was accelerated by the relative impermeability and rigidity of the coating.

A similar decay process was observed on the walls coated with an oil wash. Here the oil impregnated earth surface remained a robust skin, but the earth behind deteriorated by the same mechanism and eventually the skin detached.

In all these cases, when the coating detached, the exposed background earth surface was a decayed one, loose, friable and vulnerable to rapid surface erosion.

4.7 Dampness

Fig. 13: Dampness

The test programme demonstrated that the relationship of earth materials to moisture is a complex one, with subtle interactions affecting their durability and performance (Fig. 13).

Well-graded mixes minimise voids within the material and this is heightened when the material is compacted during the construction process. Such mixes were demonstrated to be highly durable in conditions of strong wind driven rain and freeze/thaw cycles. These materials absorbed less moisture than poorly graded mixes and were readily able to desorb absorbed moisture at the first opportunity. This ability to pass through normal wetting and drying cycles without physical change is fundamental to the long-term durability of exposed earth materials.

The tests did not demonstrate that an apparently expansive clay content had a significant effect on durability, which was surprising. Expansive clays are not commonly found in Scottish soils, but their use in construction would generally be avoided, except as waterproofing layers. It may be that the clay content in this case was not truly expansive and that its effect was minimised by good grading. X-ray diffraction, which would have determined the type of clay, was not carried out.

Similarly, the render mixes that had expansive fibres, such as wood shavings, did not exhibit reduced performance. It is thought that the type of fibres used in this test were not prone to significant swelling, and experience outside of the test programme has demonstrated that expansive fibres can significantly accelerate decay.

There seemed to be a recovery threshold for water-based erosion of earthen materials. This was passed when the fabric was significantly degraded by water or frost action and was not able to recover strength or adhesion on drying. This threshold was distinct from the normal wetting and drying cycles that only caused superficial damage.

The most important mechanism contributing to reaching this threshold was the leaching out of the clay content form the earth. Without the binding action of clay, earth materials lost their coherence and were unable to resist other decay processes. This weakening process is evident in the surviving historic buildings and is impossible to reverse and difficult to effectively repair.

The initial period of drying out was a critical time of vulnerability for all earth materials. While surface erosion commonly established a stable surface by redistributing surface fines into shrinkage cracks, the initial dampness of the material made it vulnerable to more severe decay during this period. This was particularly evident in the tests where the wall face comprised a combination of masonry and earth. Sacrificial erosion of the earth could occur with a local severity that was much greater than would have been the case if the earth had been cured.

13


This vulnerability would suggest that some form of protection may be appropriate for the first year after construction. However, where renders or pointing are to be applied, this initial erosion can achieve the basis for a good mechanical key. Traditional practice would commonly apply coatings or lime pointing at least a year after construction.

There was also significant seasonal variation in the condition of the walls, related to the level of moisture in the earth material. The walls were observed to visibly recover in the summer, between spring and autumn inspections. This dynamic effect should be considered when inspecting existing earth structures.

14


5. CONSTRUCTION TECHNIQUES

All of the earth building techniques that are commonly found in Scottish traditional construction were tested in the programme, along with a few other relevant techniques. Variations of technique and material were tested in the different climatic conditions and their method and durability assessed.

Following limited frost and mechanical damage in the first year, some tests were given local repairs and these were monitored. Otherwise, there was no maintenance or intervention in the progression of decay mechanisms.

The following descriptions give a summary of the assessment for each construction technique, based on the test results. Reference can be made to the test data sheets in Appendix A for more information on individual tests. These contain cross-references to the photographic archive and inspection reports.

5.1 External Walls A total of thirty-seven test walls were constructed at three external sites. Tests were carried out both with, and without, applied coatings and finishes.

The mudwalls and turfwalls were 1m long x 0.6m wide x 1m high, and the shuttered walls were 1m. long x 0.45m wide x 1m high, with the long sides facing SW and NE. All were built off masonry plinths, 300mm high, except for the stone with earth mortar and turfwall, which were built directly off the ground. All walls had pitched roofs, giving approximately a minimum of 25mm projection beyond the wall face, though there was considerable variation.

The shutters, used in some wall constructions, were made from 18mm marine plywood, with bolts and braces, as required.

The tests demonstrated that a wide range of subsoils is generally suitable for use in earth wall construction. Earths containing expansive clays, if adequately tempered to achieve an appropriately graded mix, could perform well.

The type of straw additive was generally demonstrated not to be significant. Flax proved a difficult additive to mix by foot, though it has since been mixed easily by machine. Compared to straw, the relatively fine flax fibres provided good reinforcement, but produced a less good surface with more cracking.

Mixing by mechanical means proved significantly more efficient than mixing by foot, for all but the smallest quantities of work.

5.1.1 Mudwall

Fig. 14: Standard mudwall

Also known as cob, this technique involved the manual formation of a monolithic wall in lifts, from a damp earth, usually mixed with straw (Fig. 14). The face was formed by dressing off or building against shutters. There were twelve test walls. The technique is described in TAN 6, 5.02, i-iii, 5.03, ii.

The standard form of mudwall construction proved reasonably efficient and practicable in Scottish conditions, though protection against rain is advisable to maintain continuity and control the moisture content of the mix. These were the most easily constructed walls, relatively simple to produce and quick to erect, reflecting their widespread use in traditional construction.

Fig. 15: Shuttered mudwall

15


The use of shuttering could achieve a faster build, if the shuttering process was well designed (Fig. 15). The shutters were not struck immediately as it was thought the walls might slump, and this significantly impeded drying. The shuttering needed to be robust to avoid bulging during construction. Oiled, smooth-faced shutters were effective in avoiding loss of surface when the shutters were struck. Some shutters were left in place for up to two weeks after construction and these walls could have shrunk back before the shutters were struck. The shuttered walls did not require dressing and weathered impressively well, slightly better than the standard mudwall construction. As with all shuttered constructions, it was very important to avoid rain collecting in the shutter.

Fig. 16: Cumbrian Mudwall

The Cumbrian mudwall technique (Appendix A.1.4), of alternating thin layers of earth and layers of straw, proved hard work (Fig. 16). The straw absorbed moisture from the earth, helping it firm up quickly and allowing a faster rate of construction. On advice regarding traditional practice, the test did not add any water to the earth, which, although beaten, was very stiff, difficult to work, and required a greater surface slope on the wall.

This technique may have been more successful with the controlled addition of water to the mix of this particular earth, possibly in advance and soured. This could potentially have allowed a very simple and continuous construction process.

Generally, these walls proved durable, with rates of decay relating to site exposure. Surface erosion typically stabilised after the first year and was minimal in the shuttered versions. There was minor frost damage in the first winter and some local undermining near the base thereafter.

Repairs in cement proved inappropriate, resulting in accelerated erosion.

All the walls provided a good key for renders, with the shuttered versions having a markedly rougher surface than, for example, the rammed earth.

5.1.2 Claywall

Fig. 17: Claywall

In this technique (Fig. 17), shutters were filled with a combination of an earth/straw mix and randomly arranged stones (TAN 6, 5.03,iv). Similar to shuttered mudwall, it used up poor quality building stone. There were three test walls.

16


This proved a very quick form of construction. In comparison to the clay and bool walls, which had placed stones, these walls had less structural strength due to the random distribution of the stones.

As with the other shuttered walls, this technique required a fairly plastic mix and was most successful if mortar was thrown with some force into the shutters rather than placed and tamped into position.

Decay patterns were light and similar to those for shuttered mudwall, except that, where stones were exposed or near to the surface, local pockets of erosion could develop.

5.1.3 Rammed Earth

Fig. 18: Compaction of thin layer within shutters for rammed earth construction

Also known as pise, this technique involved the compaction of earth in thin layers between shutters (Fig. 18), by manual or mechanical means, sometimes with a masonry face (TAN 6, 5.03, i, v-vii.) There were nine test walls.

Control of moisture was very important for this technique, with careful processing of the earth required to achieve the necessary even, dryish, crumbly, fine-textured loam. Stronger shuttering was required than in other tests because of the lateral pressure exerted during the ramming process. The one advantage to this technique was that the shuttering could be struck immediately, as the mix was not sticky as were the wetter mudwall mixes.

This technique proved very problematic in practice. The soils that were used tended to be heavy and clay rich, requiring considerable processing prior to use. Ingress of rain during construction was a particular problem and despite the care taken in processing, the quality of the results was quite variable. Even with the most appropriate soil type, it was found that there was a narrow range of working conditions in which rammed earth of an acceptable quality could be achieved.

Rates of construction were much slower than with the shuttered mudwall, though they would have been quicker if earth could have been kept at consistent moisture content. The various masonry-faced rammed earth

tests were by far the slowest techniques of all the tests. Consistent quality was difficult to achieve as the face could not be seen during construction and it was felt that the enhanced durability was unlikely to merit the increased labour.

These difficulties in construction caused a variable quality of finished work, which led directly to accelerated erosion. The best work weathered very lightly, but the poor work suffered severe local erosion in exposed conditions. For this reason, although it was denser, overall this technique was not as effective as mudwall.

Fig. 19: Rammed earth with corner lime reinforcement

Lime reinforcement of the corners gave some protection, but led to local decay at the junction of materials (Fig. 19).

17


Fig. 20: Rammed earth with banded lime reinforcement

Banded lime reinforcement (Fig. 20) produced local erosion of the earth, and its ability to improve the key of lime renders remained untested.

Fig. 21: Cement ‘stabilised’ rammed earth

The test of ‘stabilisation’ with a cement additive proved highly durable (Fig. 21). In practice, the appropriate application of this technique in conservation would need very careful consideration of material compatibility.

The difficulties in production of rammed earth and its associated variable durability relate directly to typical Scottish climatic conditions and soil characteristics. This may partly explain why rammed earth never became a significant traditional technique in Scotland.

5.1.4 Stone with Earth Mortar

Fig. 22: Stone with earth mortar

This technique (Fig. 22) involved building a traditional masonry wall, with stones laid in earth mortar (TAN 6, 5.02,vii). There were three test walls.

This proved a simple and effective technique, which worked best when any large aggregate was removed form the earth and the mortar was mixed to a smooth paste to facilitate bonding in thin mortar beds. Achieving the correct consistency, as with all techniques, required some skill and experience.

The earth mortar joints weathered back in relation to their exposure. Where expansive clays were used, the thin mortar beds revealed extensive cracking, which facilitated the establishment of other decay mechanisms.

This type of wall is still very common throughout Scotland, though it is often concealed behind lime pointing. Based on these tests, pointing would not be necessary with a good mix, in a sheltered location. Otherwise, the time before pointing is necessary will vary with the mix and exposure.

18


5.1.5 Clay and Bool

Fig. 23: Clay and bool

This technique (Fig. 23) involved laying courses of round stones in earth mortar against shuttering, to form a masonry face. There were three tests of this construction.

This technique was found to require a greater degree of skill to construct effectively than other forms of masonry. With the bools set against shuttering, the face of the wall is built unseen and it is not possible to view the bond between courses during construction. As the stones are rounded, skill and experience are needed to achieve a reasonable quality of build.

The rounded shape of the bools created a larger area of clay exposed on the wall face and joints that were more vulnerable to decay, than with the other stone-faced techniques. Weathering was therefore faster and with a poor earth mix, stones could become detached in locally exposed areas.

Traditionally, this was a regional technique responding to the abundant local availability of a particular type of stone. As such, specialist skills would have existed locally. In repairing or replicating this technique, it would be important to recognise the level of skill needed and have appropriate training and sample quality control.

Traditional sources indicate that it could take up to five years before walls were pointed with lime and limewashed, but this always seems to have been the traditional finish. The tests confirmed that this technique is unlikely to prove durable without such secondary protection.

5.1.6 Earth Block with Earth Mortar

Fig. 24: Earth brick with earth mortar

Also known as adobe, this technique involved the construction of a masonry wall from air-dried earth bricks, laid in earth mortar (Fig. 24). There were three tests of this construction.

This proved a simple and effective technique in construction, with good durability.

In contrast with tests that had stone or fired brick facing masonry, the compatibility of materials in this technique did not create any local weakness where decay could focus. Indeed, in time it became difficult to differentiate the bricks from the mortar as a stable weathered surface developed.

The fabrication of air-dried blocks requires some effort and planning in advance, for production in controlled, off-site conditions. For these reasons this was not a technique that flourished in Scottish traditional practice, though it was used regionally in England.

This technique has been used in the repair of earth buildings in Scotland, where the quality control, flexibility and minimal shrinkage of the blocks have proved very useful qualities. While these tests give reassurance about the durability of such repairs, they also suggested that more research is needed into the composition of blocks.

19


A variety of commercially produced earth bricks and mortar are currently available. While there are limits on their suitability as an appropriate repair material for external walls, they have wider applicability internal in traditional buildings, where their moisture regulating qualities, adaptability and reversibility are of significant benefit.

5.1.7 Turf

Fig. 25: Turf block

These tests included a number of techniques using turf - surface vegetation, usually grass, with varying amounts of root systems and topsoils.

The turf block technique (Fig. 25) involved laying shaped blocks of turf and rooted topsoil in courses, without mortar, to form a masonry wall. (TAN 6, 3.02)

This proved an effective and durable form of construction, if appropriate material was sourced. The turf was simple to prepare, but had to be used within twenty-four hours, to avoid the turfs becoming inflexible as they dried out. During construction, the diagonal joints tightened as the blocks compressed under the weight of the following layers, bonding the fibres together and creating a very solid wall. Achieving a good quality face required a degree of specialist skill.

Fig. 26: Turf between hurdles

Another technique (Fig. 26) tested involved laying thinner turfs between wattle hurdles (TAN 6, 3.04). This proved a rapid construction method. The stake and rice hurdles

provided a temporary face and, as they degraded, the turf stabilised and continuity of the walls integrity was maintained.

Only one turf block wall achieved good durability, while the others collapsed early in the programme. It is thought that the successful test had an excellent material for that particular technique, while the others were less appropriate combinations of technique and material. As there is great variety in traditional turf construction, these limited tests cannot be considered a conclusive assessment.

Combinations of topsoils and grass that will make good quality blocks require careful sourcing. Historical sources indicate significant land degradation could result from sourcing this material and an appropriate environmental impact assessment should be made for any proposed application of this technique.

5.2 PanelsFourteen external tests of panel constructions were conducted, following twenty-nine preliminary internal tests on a wider range of panels. The panels were tested both with, and without, a variety of coatings.

The panels were 1m long x 1m high x 100mm deep. Apart from the earth daub on kebber and mott tests, they were built on 300mm high plinths. Initially, no roof was provided for the external panels and this represented an extremely severe weathering test.

Good workable daubs were achieved with minimal effort by pre-mixing several days in advance. A wide variety of daub mixes could be applied to these armatures, but control of moisture content and grading were important. The most successful mixes were easily moulded in the hand, not so wet as to slump in the frame and not too dry to stick to the frame.

Varying amounts of straw did not make a significant difference during application, though 50%, by loose volume as an additive, seemed to be the maximum workable proportion. The daubs with flax fibre showed more cracking than those with straw. This attributable to their finer fibres, which provided less reinforcement and made it difficult to achieve an even distribution in the mix.

Due to their thinness, the panels showed greater shrinkage cracking than the wall tests and the cracking patterns reflected the arrangement of the underlying armature. Overall shrinkage varied up to 2%.

Deficiencies in the earth type and mix generally showed a greater effect than on the wall tests.

Generally, the framed panels proved quick and simple to construct, but in the long-term they would require a greater input of labour due to maintenance and aftercare requirements.

20


5.2.1 Earth Daub on Stake and Rice

Fig. 27: Earth daub on stake and rice

Also known as wattle and daub, this technique involves applying an earth mortar onto an armature of thin ‘rice’ woven between vertical stakes (Fig. 27). There were four external tests and fifteen internal tests of this method, with variations of technique, mix, finish and climate.

A fairly open weave of rice, comprising fine pieces of hazel and willow with gaps of 10-20mm between, allowed good penetration of the daub. In a metre wide frame, rice of 10-20mm diameter was workable. Thicker stems were difficult to weave, while thinner stems did not support the daub.

While the daubs could be applied from one side, with a shutter forming the other face, it was more effective to press the daub in from both sides, achieving a firm bond in the centre.

Shrinkage cracking developed as a pattern of fine cracks in both vertical and horizontal directions, following the underlying armature.

This technique only proved durable at the sheltered site.

5.2.2 Earth Daub on Horizontal Rails

Fig. 28: Earth daub on horizontal rails

This technique involves the application of earth mortar onto an armature of parallel timber rails, fixed between posts (Fig. 28). There were three external tests and six internal tests of this method, with variations of technique, mix, finish and climate.

This technique worked with a wider variation of sizes and spacings within the armature, than was the case with the stake and rice. Shuttering was generally required to one side of the panel, as a support for the daub, which had to be left in place for several days as the daub dried.

It is possible to work without shuttering if long ‘sausages’ of straw soaked in wet daub are wrapped around the horizontal rails and allowed to dry before a second application of daub. However, this method seemed extremely laborious and unnecessarily complicated in comparison with the shuttered method.

This technique proved slightly less durable than the stake and rice panels, with larger shrinkage cracks reflecting the armature arrangement.

5.2.3 Earth Daub on Horizontal Rope

Fig. 29: Earth daub on horizontal rope

21


This technique involves the application of earth mortar onto an armature of parallel or latticed ropes, fixed between posts (Fig. 29). There were four internal tests of this method, with variations of technique, mix and finish.

This worked in a similar manner to the daub on horizontal rails, except that there was less cracking within the panel, as the rope was better able to absorb shrinkage, transferring shrinkage into larger cracks at the edges. Different types of straw or flax rope did not appear to alter performance.

5.2.4 Earth Daub on Kebber and Mott

Fig. 30: Earth daub on kebber and mott

This technique involves the application of earth mortar onto a palisade of spaced, parallel timber posts, driven into the ground (Fig. 30). There were three external tests of this method, with variations of technique, mix, finish and climate.

This method worked best with a drier mortar, as the lack of horizontal members meant it was prone to slumping. This was one of the most problematic techniques as the tests were extremely vulnerable to rain damage during, and following, construction. Rainwater tended to track down the poles, breaking the earth bond. All panels required some early repairs. The base, where the panel met the ground, was vulnerable to damp related decay.

This technique did not prove durable, except at the sheltered site. Large cracks developed corresponding to the armature and the daub progressively eroded in exposed conditions.

5.3 CoatingsTwenty-seven external tests of varied coatings were conducted, following seventy-eight preliminary internal tests on a range of mostly plaster coatings. The panels were tested on a variety of backgrounds and both with, and without, applied finishes.

The coatings covered half of the external test walls, both sides of half of the external test panels and a quarter of the internal test panels. The number of layers varied and overall thicknesses ranged from 15-25mm.

A variety of factors were investigated for their effects on key. Different mechanical scoring patterns proved equally effective. Weathered surfaces achieved a better bond than unweathered ones and rougher surfaces gave a better key than smooth ones.

5.3.1 Earth Renders and Plasters

Fig. 31: Earth render

Eighteen external tests of earth renders were conducted, following sixty preliminary internal tests on a wider range of earth renders and plasters (Fig. 31). A wide variety of mixes, earth types, backgrounds, additives and finishes were tested.

While a range of earths renders proved successful, good grading and fibre content was found to be the most important factors, with additives capable of further enhancing good performance.

Mix, grading and finish could have a significant effect on workability and drying patterns, while different backgrounds was only a minor factor.

22


There were no obvious differences in the finished results containing different amounts of water, though mixes that were too dry did not adhere well.

Various multi-layered applications were tested with good results and proved appropriate for mixes prone to shrinkage, but with a good mix, a single application seemed a more efficient method than building up several thin layers to the same thickness. It was possible to apply earth render as a harl, that is by throwing it on, but the adhesion and finish achieved were less even than with the lime harls.

Renders could be given a smooth finish with metal and wooden floats. Excessive working of the surface seemed to loosen the bond between render and background and could separate the clay binder from the fines in the mix. Hand finishes were less prone to this problem. Renders with rough, open surfaces were less prone to shrinkage cracks than those with smooth, even ones. Textures were introduced by scoring, scratching and indenting with a variety of tools.

Generally, all soils were processed to a finer grading for renders than they were for wall and daub mixes. This removed larger aggregate and stones, which would have reduced workability and weakened the render. Clay-rich mortars, with no added fibre or aggregate, adhered well initially, but pulled away from the background during drying due to excessive shrinkage.

Mixes without fibre, which were tempered with various types and quantities of aggregate, did not fatten up when matured and had poor adhesion during application. As well as being difficult to work, they developed extensive cracking and became too brittle to polish soon after application.

Better results were possible where several layers of such mixes were applied, allowing cracks to develop between applications and working with increasingly lean mixes towards the outer surface of the render. A very fine polished finish was achieved when adding wood and coal ash in place of half of the normal quantity of sand. The surface of this render had a pale bloom with dark flecks and no cracking.

The inclusion of fibre proved extremely beneficial for shrinkage control and reinforcement, but it was necessary to cut fibres shorter than for building mixes. With straw and hay additives, souring significantly improved mortar workability and allowed greater quantities of straw to be absorbed, further reducing shrinkage and cracking.

With flax additives, the benefits of souring were much less significant, and although general durability performance was good, it had poorer working qualities. Hair additives allowed more cracking than straw or flax, but these mixes still performed better than those without any additive.

There is a wide range of potential additives suggested by traditional sources as improving the performance of earth renders. In the external tests the additives that were tested with earth renders included wood shavings, dung, manure, oil, tallow and quicklime.

The addition of wood shavings had a very positive effect on shrinkage reducing cracking to a minimum and proving durable (Appendix A: A.5.12, A.5.13). However, it should be noted that a recent building project used a similar coating but experienced much poorer performance. This was attributed to the fact that the quality and size of wood shavings was more varied than in the test mix.

The addition of dung significantly improved the working qualities of earth mixes and moderately improved their durability (Appendix A: A.5.8, A.5.9, A.5.10).

The addition of manure significantly improved the working qualities of earth mixes, but did not affect their durability (Appendix A: A.5.11).

The addition of oil reduced adhesion of earth mixes but significantly improved their durability (Appendix A: A.5.16). This was quite different from the effect of oil as a coating on earth render.

The addition of tallow was difficult to achieve, but significantly improved their durability (Appendix A: A.5.17).

The addition of quicklime marginally improved the preparation of earth mixes, but reduced their durability.

The application of limewash on to earth mixes moderately improved their durability in the short term (Appendix A: A.5.18).

The addition of urine was of no apparent benefit.

Overall, the research suggested that earth plasters and renders can be effective protective and finishing coatings for earthen walls. If they achieve a good mechanical key, they do not suffer the incompatibility problems of other finishes relating to differential movement and vapour permeability.

An earth render can have a good durability, especially when combined with specific additives and given a limewash finish. However, the tests, and other recent experience, suggest that the earth grading, additive specification and method of application create quite narrow margins where an earth render mix will be successful and this is an area that would merit further research.

23


5.3.2 Lime Renders and Plasters

Fig. 32: Lime render

Nine external tests of harls and lime stuccos were conducted, following nineteen preliminary internal tests on a wider range of lime coatings (Fig. 32). Mixes were prepared with both lime putty and hot slaked quicklime, a variety of additives was also tested. Most were finished with limewash. Limewash was also tested on its own.

The quicklime mortars were slaked on site and stored for some time before application. This means of preparation proved advantageous as there were many additives and mixing methods that benefited from the application of heat. The mixes were not applied hot as it was thought that unslaked lime might reduce performance. Subsequent applications in practice have demonstrated that mixes used immediately as ‘hot lime’ mixes can be very successful.

As well as hot applications, the programme did not test hydraulic mixes, different methods of application, or applications where the wetness of the wall varied.

As a binder, lime can be used alone or mixed with clay. Lime-earth combinations are common in England, but rarely found in Scotland. In contrast, lime based coatings are very common on Scottish earth buildings and these include harls, stuccos, plasters and limewashes.

There is a wide range of potential additives suggested by traditional sources as improving the performance of lime renders. In the external tests the additives that were tested with lime renders included dung, seaweed, whey, tallow and oil.

The addition of seaweed to a hot lime mortar produced a strong pungent odour of iodine. The dried seaweed softened to produce a mix with normal working properties and slightly improved durability (Appendix A: A.6.4).

The addition of whey produced a rich sticky mortar, with an attractive brownish-pink colour when cured. This proved robust against surface erosion for a considerable time, but eventually loss of bond to the earth background led to complete delamination of the render (Appendix A: A.6.5).

The addition of oil did not affect the working qualities of the mortar, slightly reduced surface erosion but did not inhibit the renders failure (Appendix A A.6.6).

The stucco seemed to adhere better than the renders, but durability was comparably poor.

As a group, the lime coatings did not perform well, with rapid loss of key leading to complete failure. Means of preparation and application did not make a significant difference to overall performance. Some additives slowed the decay process to a small degree. Although there was no maintenance of limewashes, this was not a significant factor.

This poor performance was surprising, as lime renders are commonly held to be appropriate finishes for earth buildings and are found within traditional practice. While the nature of the test conditions would have accelerated the rates of decay, they demonstrated that there was a fundamental incompatibility between lime coatings and earth backgrounds in external situations, even where soft limes are used. This is related to differential vapour porosity and movement.

Lime renders that have been applied in practice in recent years to historic and new earth buildings in Scotland do not demonstrate rates of failure comparable with these tests. However, they have experienced corner cracking and some localised delamination. The results of these tests raise fundamental questions about the use of lime renders on earth buildings and this is an area where further research is merited.

24


5.3.3 Cement Render

Fig. 33: Cement render

Ordinary Portland Cement was tested as a coating to mudwall and performed extremely badly (Fig. 33). The lack of vapour permeability led to the rapid deterioration of the background material and subsequent collapse of the test wall.

Cement renders are known to have been widely applied to traditional earth buildings in modern times. This test confirms that they are inappropriate for this purpose and their removal would generally be of benefit to their earth backgrounds.

5.4 Washes, Paints and DistempersTen external tests of various types of liquid coatings were conducted, following a wider range of twenty-nine preliminary internal tests. The materials were applied to various backgrounds as washes, paints and distempers.

The results were generally poor, though durability was compromised by the lack of regular re-application that is often traditional practice. The differences between individual test results suggested that some materials could significantly improve durability.

5.4.1 Limewashes and Distempers

Fig. 34: Limewash

Five external tests of limewash were conducted, following twenty-six preliminary internal tests on a wider range of limewashes and distempers (Fig. 34). A variety of mixes, lime types, backgrounds, and additives were tested.

Limewash could be applied directly to both fresh or weathered earth mortar as long as the surface to be painted was damp and dust free. The wash proved particularly effective for filling small cracks and hollows and it could be worked into the earth background to produce a harder finish than earth had on its own. Such applications worked better on damp, fresh earth, than on a mature earth surface.

Freshly slaked limewash produced a brighter finish, with better coverage, than the same consistency and application of wash made from diluted putty.

There is a wide range of potential additives suggested by traditional sources as improving the performance of limewashes. In these tests the additives that were tested with limewashes included sand, whey, oil, tallow, lanolin, ash and earth pigments.

The addition of fine sand to the wash was beneficial over very rough surfaces by reducing cracking, evening out irregularities and improving the key.

25


The addition of whey, slaked with lime, produced glossy finishes with poor adhesion and reduced durability (Appendix A: A.7.8).

The addition of boiled linseed oil, slaked with lime, produced attractive glossy finishes with improved adhesion and reduced durability (Appendix A: A.7.7).

The addition of tallow, slaked with lime, produced attractive glossy finishes with improved adhesion. Tallow proved too thick to be added to cold limewash.

The addition of lanolin, slaked with lime, produced a darker, more porous looking wash that was much less prone to dusting (Appendix A: A.7.9). Lanolin could not be added to cold limewash. Durability could not be assessed as the background collapsed early on for other reasons.

The addition of natural earth pigments made limewashes that behaved more like slurry washes, giving good coverage over rough surfaces due to the presence of fines.

The addition of ash, soot and charcoal had the same effect.

Generally, the limewashes did not perform well, mostly lacking durability. Additives, other than of fines, did not produce a significant improvement. Results were variable though, with one good result on the exposed site (Appendix A: A.7.6) suggesting that further research is needed.

It should be noted that there was no maintenance of limewashes. The regular, cyclical application of limewash is an essential part of traditional practice. These tests indicate that with its thin, sacrificial nature, limewash seems to avoid the incompatibility problems that were demonstrated to significantly effect lime renders.

5.4.2 Dung Paint

Fig. 35: Dung paint

Two external tests of dung paint were conducted, following one preliminary internal test (Fig. 35). The dung was tested without additive and was sufficiently liquid in its natural condition not to require the addition of water.

Fresh cow sharn was tested as a wash on earth daub and proved completely ineffective, having poor adhesion and negligible durability.

5.4.3 Oil Paint

Fig. 36: Oil paint

Two external tests of oil paint were conducted (Fig. 36), following two preliminary internal tests. No additives or dilution were tested.

Boiled linseed oil was tested as paint, brushed onto earth daub. It improved initial durability but in the long term proved damaging. The oil impregnated the earth surface, creating a robust skin, resilient to surface erosion. However, this surface also had poor vapour permeability. The earth background gradually deteriorated and eventually the skin detached, exposing an open and vulnerable earth surface.

5.4.4 Tallow PaintOne internal test of tallow paint was conducted, without additive or dilution.

Liquid tallow was tested as paint, brushed onto dry earth render and was completely ineffective. It was not absorbed and left a skin of hard fat on the surface.

26


6. MATERIALS

A variety of materials were tested in varying combinations as suggested by historical precedent, archival sources or contemporary practice.

While there was significant variety in the ease of workability of these combinations of materials, the compatibility of mixed materials proved a critical factor in determining long-term durability.

6.1 EarthSix types of earth were tested as the primary material for all the tests in the programme, apart from the lime coating and some of the washes and paints. One type, the Gallowflat Red, formed a control mix, used for walls, daub and render at all three external sites. Some tests were carried out with the earth unadulterated, but in most cases the earth was added to in a variety of ways.

The performance of the soils varied with the technique and exposure, though their qualities were tempered, by mixing with varying proportions of sand and grave, to compensate for their assessed characteristics, in an attempt to achieve a balance of particle sizes.

Rich, silty earth, such as the two Gallowflat types, were easy to mix, but were less sticky than the other soils. Their high proportion of fines made them vulnerable to surface

erosion and leaching of the clay binder. In external applications, such soils may require additional protection. However their fine texture and good workability would make a good basis for internal plasters.

Stony soils, such as the Drongan and Carlisle, can be too stony for machine processing and would require considerable processing for use in fine applications. However, their coarse grading is more durable externally and the weathered aggregate provides an excellent mechanical key for surface coatings.

Soils with an apparently expansive clay content, such as the Westerfolds, did not show any reduced durability in comparison to other soils. This was surprising and may be attributable to the tempering. The inclusion of a very coarse gravel, from the mixing area, may have improved durability. Detailed analysis of clay types was not carried out and this may sometimes be necessary to explain the performance of some, more complex, clay contents.

There is potential to temper a wide range of soils. Rich, fine soils can be tempered with clean, coarse sand and gravel, while dirtier, coarser soils can benefit from the fines contained in unwashed sands.

Table 2 Red Gallowflat Clay

Characteristics: A silty clay loam, stored for some time prior to use. An estuarine high raised beach deposit of silts and clays.

Source: The Errol Brick Company clay pit at Gallowflat (Perthshire, NO 215 210). A deposit of Harviestoun Series, Carbrook Association, Soil Survey of Scotland Sheet 48/49.

Field Analysis: Not carried out, as suppliers analysis was available.

Lab Analysis: Grading: 27% clay, 54% silt, 9% sand, 10% gravel Supplier information at time of tests

31% clay, 69% silt and sand, 0% gravel post test analysis

Liquid Limit: 43% post test analysis

Plastic Limit: 27% post test analysis

Plasticity Index: 16% post test analysis

Linear Shrinkage: 11% post test analysis

27


Table 2 Red Gallowflat Clay

Mixes tested 3:3:1, earth : sand : gravel as mudwall, 5 tests as claywall, 1 testas rammed earth, 1 testas clay and bool mortar, 1 testas earth daub, 4 test

1:1, earth: sand as mortar, 3 tests

5:3:2, earth: sand : gravel with flax, as air-dried bricks, 1 test

2:4:3, earth: sand : hay as render, 4 tests

2:4:3, earth: sand : hay, + 10% dung as render, 3 tests

2:4:3, earth: sand : hay + 0.2% oil as render, 1 test

2:4:3, earth: sand : hay + 0.2% tallow as render, 1 test

2:4:3, earth : sand: seaweed as render, 1 test

2:2:4:3, earth: sand : hay : ash as render, 1 test

1:1:2, earth : sand : wood shavings as render, 2 tests

1:4, earth : sand, +20% manure as render, 1 test

Comments: The difference between the two laboratory analyses indicates the degree of variation that naturally occurs in this soil. The Brown Gallowflat clay is a nearby deposit of slightly different characteristics. These deposits have variable amounts of sand and silt present and are used for making bricks. Their good workability has established them in use as a repair material. There are many traditional earth buildings in the local area, which would have used similar earth materials.

This was a rich, even material, whose high silt content made it vulnerable to surface erosion.

Table 3 Brown Gallowflat Clay

Characteristics: A silty clay loam, stored for some time prior to use. An estuarine high raised beach deposit of silts and clays.

Source: The Errol Brick Company clay pit at Gallowflat (Perthshire, NO 215 210). A deposit of Harviestoun Series, Carbrook Association, Soil Survey of Scotland Sheet 48/49.

Field Analysis: Not carried out.

Grading: 24% clay, 76% silt and sand, 0% gravel post test analysis





Mixes tested earth as dug as rammed earth, 2 tests

3:3:1, earth : sand :gravel as rammed earth, 1 test

9:1, earth: cement as rammed earth, 1 test

Comments: This nearby deposit was siltier and sandier than the Gallowflat Red, but had broadly similar characteristics. These deposits have variable amounts of sand and silt present and are used for making bricks. Their good workability has established them in use as a repair material. There are many traditional earth buildings in the local area, which would have used similar earth materials.

This was a rich, even material, whose high silt content made it vulnerable to surface erosion.

28


Table 4 Drongan Clay

Characteristics: A boulder clay containing a large coarse fraction of stone and gravel. Freshly dug, larges stones were broken up or removed. A till derived mainly from Upper Coal Measures marls.

Source: An open cast coal mine, Drongan (Ayrshire, NS 438 190). A deposit of Drongan Series, Drongan Association, Soil Survey of Scotland Sheet 13/14.

Field Analysis: Grading: 43% clay, 22% silt, 17% sand, 18% gravel Jar sedimentation test

Others: The other tests indicated a more balanced composition across the whole particulate range.

Lab. Analysis: Grading: 10% clay, 32% silt, 36% sand, 22% gravel post test analysis





Mixes tested 10:2:1, earth : sand : gravel as mudwall, 2 testsas claywall, 1 testas clay and bool mortar, 1 testas earth daub, 2 tests

earth as dug as rammed earth, 1 test

5:3, earth : sand as rammed earth, 1 testas mortar, 1 test

10:3, earth: flax as air-dried blocks, 1 test

2:1, earth : hay as render, 1 test

Comments: The field grading test gave an inaccurate result. This is thought to have been because the clay fraction was very sticky and bound larger particles during the test, creating a falsely high reading. The other field tests gave an indication of this and test mixes were developed on the basis of the broader assessment. In the tests, these mixes worked well, with working qualities consistent with this view and the post test analysis.

While surface colour changes on some of the tests indicated a possible leaching of the clay content, it was felt that this was the most durable of the soils. This was the local soil tested at the most sheltered site, and it would have been interesting to have tested it in more exposed conditions. This was a rough, dirty material, too stony to be used in small machines.

29


Table 5 Carlisle Clay

Characteristics: A boulder clay with a large coarse fraction.

Source: A development site on the outskirts of Carlisle.

Field Analysis: Grading: 61% clay, 6% sand, 33% gravel

Lab Analysis: Not carried out.

Mixes tested: as dug as Cumbrian mudwall, 1 test

2:1, earth : hay as render, 1 test

Comments: Although it has a high clay content, the rough character of this soil is well proven to give a durable construction material, by the numerous surviving buildings around Carlisle, which provide a more exposed test than the Culzean site provided. Examples in the filed show that, while the fines will suffer from surface erosion, the large coarse faction is important in establishing a durable surface, and helps to give a good key for protective coatings.

Table 6 Westerfolds Clay

Characteristics: A rich, very even, blue-grey clay, locally used for puddling. A deposit of Lucastrine clay.

Source: Agricultural land at Westerfolds Farm, by Elgin (Inverness-shire, NJ 195 679). A deposit of Kintrae Series, Duffus Association, Soil Survey of Scotland Sheet 85/95.

Field Analysis: The sedimentation test did not settle, but indicated an expansive clay content of at least 50%, with a very high concentration of fines.

Lab Analysis: Grading: 55% clay, 44% silt, 5% sand

estimated from post test analysis





Mixes tested: 3:6:1, earth : sand : gravel as mudwall, 2 testsas claywall, 1 testas clay and bool mortar, 1 testas earth daub, 3 tests

1:2, earth: sand as mortar, 2 tests

5:6:4, earth: sand: gravel with flax, as air-dried blocks, 1 test

2:4:3, earth: sand : hay as render, 1 test

Comments: The field analysis, and the fact that the material is used by local plumbers for puddling, indicated that the earth had an expansive clay content. Locally, buildings do not seem to use this material for construction, but the tests did not indicate a poorer durability link to clay content in comparison to other types of earth.

The laboratory tests were affected by a poor quality of material available for sampling. The grading result is an approximation, extracted from the test results. The other tests are more accurate, with the results for liquid limit and plastic limit perhaps a little below their true value.

30


Table 7 Kinnoir Clay

Characteristics: A grey-green clay, of uneven composition. This earth was used as it was granier and more crumbly than the Westerfolds clay. A deposit of recent alluvium.

Source: The site of a former brickworks, Corse of Kinnoir, Huntly (Aberdeenshire, NJ 564 426). A deposit of Alluvium Association, Soil Survey of Scotland Sheet 86.

Field Analysis: Grading: 68% clay, 13% silt, 8% sand, 11% gravel Jar sedimentation test

Others: Other field tests suggested a more even balance between the clay and silt fractions.

Lab Analysis: Grading: 17% clay, 37% silt, 34% sand, 12% gravel post test analysis


Plastic Limit: 27%, may be almost considered non-plastic post test analysis



Mixes tested: earth as dug as rammed earth, 1 testas turf blocks, 1 test

1:2, earth: sand as rammed earth, 1 test

2:1, earth: hay as render

Comments: The field grading test gave an inaccurate result. This is thought to have been because the earth was of uneven composition, and the sample may not have been typical. The other field tests gave an indication of this and test mixes were developed on the basis of the broader assessment. This was not felt to be a good soil, however the site was the source for exceptionally good turf construction materials.

Soil maps indicate this is a recent deposition of alluvium. The properties of such soils are very variable and hard to predict from map information.

6.2 StrawFour types of straw were tested as an additive to mudwalls, earth mortars and earth daubs. Types of straw that were tested included an old type of organic wheat straw, and modern non-organic wheat, barley and oat straws.

The straw was chopped to different lengths, depending on the technique. All straws benefited from souring.

The addition of straw proved very beneficial in reducing shrinkage and absorbing moisture. Performance generally increased in line with proportion, but maximum proportion that maintained workability was 50%, by loose volume.

The properties of the modern straws varied, barley straw was soft while wheat and oat straw were more brittle. These differences were probably more pronounced in the past.

Organic straw proved tougher and less absorbent than non-organic straw. Modern non-organic straws have often have a high nitrate content and this is believed to make them more brittle and less durable. These tests did not find a difference in durability over the six to seven year length of the programme, however questions remain over long-term performance.

6.3 FlaxThe flax was sourced as fibres, with the hurds having naturally retted. These proved difficult to mix manually, as the fibres tended to clump together. This led to an uneven surface, which accelerated erosion, by creating areas of weakness. Subsequently more even mixes have been achieved by mechanical means.

The fibres did not have the beneficial absorptive qualities of the various straws, though the fineness of the fibres gave good reinforcement. This fineness was also an advantage in forming air-dried earth blocks.

Flax did not benefit from souring.

6.4 HayOne type of hay was tested as an additive to earth renders. This performed well, with its fine texture and lack of seeds proving appropriate qualities.

31


6.5 HairHorse hair was tested as an additive to earth render. This proved slow and difficult to mix evenly, with the hair having to be teased in, to achieve an even distribution.

6.6 DungFresh cow sharn was tested as an additive in earth render, lime render, and was used on its own as a render and as a paint.

In the earth mixes, the dung improved workability, producing a rich, sticky mortar, requiring less added water, which was lovely to work with and did not smell. It had good adhesion and reduced cracking on drying when combined with other fibres, but increased cracking when other fibres were not present.

As a render on its own it was not very effective, with significantly less durability than when mixed with earth. As a wash it was completely ineffective.

The stabilising properties of dung as an additive, improvement of adhesion, binding, plasticity and durability, is linked to the fibre content and other factors. When combined with earth there can be chemical reactions that lead to the formation of a binding gel. It is these chemical reactions that make dung more useful as an additive than when used on its own. Further research into this would be beneficial.

6.7 Manure Horse manure was tested as an additive to earth render. It did not improve mixing or adhesion, as dung did, but the introduction of many small fibres did have a positive effect by reducing shrinkage. Manure appears to lack the more beneficial chemical reactions with earth that are obtained with dung.

6.8 LimeQuicklime was tested as an additive in earth mixes, as a binder in coatings and mortar, and, with water, as a wash. Lime putty was tested as a binder in coatings and mortar, and, with water, as a wash. Only non-hydraulic limes were tested.

When combining earth mixes with lime, it proved beneficial to add quicklime rather than putty, as the slaking process helped break down the heavy clay soils. In these internal tests, there was no other apparent benefit to the addition of lime in earth mixes and overall performance was poorer. The mixes had lower vapour permeability and their surfaces were more brittle, with finer cracking patters. The mixes were not tested externally where other properties may have been apparent.

In coatings, mortars and washes, quicklime generally performed better than lime putty. The ‘hot lime’ technique facilitated the inclusion of animal fat additives which could not be included cold. The benefits of hot lime applications, apparent in subsequent practice, were not tested.

Limited lime pointing on the clay and bool was of limited success, due to the rounded nature of the masonry.

As a render, lime generally proved ineffective. It was a robust defence against surface erosion, but rapidly lost its key, cracked at corners and completely delaminated from the earth backgrounds. This process derives from the different vapour porosities and movement characteristics of lime and clay-bound materials.

The poor performance of lime coatings in these tests significantly exaggerated weaknesses in performance found in the field. However, the tests suggest a fundamental incompatibility between lime coatings and earth backgrounds, which would merit further research.

As a wash, lime proved only moderately effective, but it was recognised that the long–term performance of limewash could be good, if it is given the routine periodic application of fresh coats, as part of a maintenance regime.

6.9 Wood ShavingsFine, consistently graded hardwood shavings, from a planer, were tested as an additive to earth renders. They had a significantly beneficial effect in controlling shrinkage and reducing cracking, related to their reinforcing properties and increased surface area. The open-textured render proved very durable.

However, experience in practice suggests a more complex performance that cautions against indiscriminate use. The performance of a similar render, on a new mudwall building near one of the test sites, was initially good. However, the render became badly affected by wind-driven rain and frost, and did not prove an effective coating in the long term. The ‘shavings’ in this case were more varied, containing both softwood sawdust and larger particles. The testing of bricks containing softwood wood shavings, by the Errol Brick Company, has shown that they have poor durability in externally exposed situations, though they work well in internal applications.

These poor performances are related to the absorbent nature of wood as compared to other fibres, such as straw. Wood fibres can expand significantly on absorbing water, breaking up the material in a manner similar to frost. It may be that the shavings that were tested as part of this programme were not very absorbent and that the renders were better able to accommodate such expansion due to their thinness. In any case, their performance was certainly impressive and suggests that further research would be worthwhile.

6.10 SeaweedDried, chopped seaweed was tested as an additive to an earth render and a hot lime harl. In both cases it proved ineffective, producing workable mixes, but with reduced durability. It was ineffective as a fibre, and although it performed best in the tests of additives in lime, this was not thought to be significant.

32


Seaweed is thought to be able to improve binding, waterproofing and act as a gelling agent. It is known to have been used as an additive in ancient and traditional practice in Mediterranean countries and it is believed that the test methodology did not accurately replicate this precedent. Further research would be merited.

6.11 AshSieved, hardwood ash was tested as an additive to earth render. This improved durability and it is thought to relate to its fine grading and possibly a more complex, chemical reaction.

6.12 OilBoiled linseed oil was tested as an additive in earth render and lime harl, as well as being used, on its own, as a paint.

As an additive, oil did not affect workability of the earth render mixes, but made them darker and harder when dry. It significantly improved durability without apparently affecting the earth background (Appendix: A.5.16).

It did not affect workability of the lime render mixes and increased its durability against surface erosion, though the coating itself proved inappropriate (Appendix: A.6.6).

Limewashes slaked with small amounts of linseed oil produced lovely glossy finishes that adhered well to the earth walls, but durability was not affected (Appendix A: A.7.7).

As a paint, oil improved surface durability, but, in the long-term, significantly accelerated the decay of the earth background (Appendix A: A.7.4)

All fats and oils improve waterproofing, adhesion and durability. The tests demonstrate that oil is effective as an additive to other mixes, but that its reduction of surface vapour permeability make it inappropriate for use as a paint.

6.13 TallowLiquid tallow was tested as an additive to earth and lime render, as well as being used, on its own, as a paint.

As an additive to the earth mix, the tallow was difficult to combine cold and had to be trodden and beaten into the mix. However it significantly improved durability (Appendix A: A.5.17).

As an additive to lime render and wash, the tallow was added during the slaking process and this proved a simple and effective means of including the additive. No difference in performance was found when used on internal tests.

As an additive to limewash, it gave a glossy finish and improved adhesion. There was no obvious affect on durability, although, as a repeated application, there might be problems due to reduced vapour permeability.

As a wash it was completely ineffective, failing to be absorbed.

All fats and oils improve waterproofing, adhesion and durability. The tests demonstrated the importance of choosing an appropriate form of additive for the particular mix. Tallow was only effective as an additive to hot lime mixes, where it improved the performance of a coating that itself proved inappropriate.

6.14 WheyFresh buttermilk was tested as an additive to earth render, lime render and limewash.

As an additive to ‘hot lime’ render, it produced a rich, sticky mortar with an attractive colour and improved durability against surface erosion. The render seemed more flexible than other lime renders (Appendix A: A.6.5).

Limewashes slaked with small amounts of buttermilk produced an attractive colour, but the wash was frothy, full of bubbles and difficult to apply. The wash flaked off and did not prove durable (Appendix A: A.7.8).

The stabilising properties of whey, waterproofing, increased hardness and shiny surface, are also found when using other milk products such as skimmed milk, casein and cream cheese. These additives are generally used in lime-based mixes and rely on the formation of calcium caseinate, which is hard and relatively insoluble.

6.15 LanolinLanolin was tested as an additive to limewash and did not seem to significantly affect performance (Appendix A: A.7.9).

6.16 UrineFresh horses urine was tested as an additive to earth plasters. The addition of horse urine seemed to aid the mixing of the earth mixes, requiring less water. There was no apparent difference during application and drying, apart from the overpowering and unpleasant smell. There was no apparent difference in durability.

Urine, whether fresh or stale, animal or human, is held to encourage the dispersion of fines during mixing, reducing shrinkage, promoting drying and increasing hardness. The tests indicated that, in practice, these improvements are marginal and that the unpleasantness of using the material greatly outweighs any benefits in improved performance.

33


6.17 BloodFresh ox blood was tested as an additive to earth render. It did not affect workability and darkened the render when polished.

Blood was probably particularly used on high status buildings, on earth-lime floors, where its properties of waterproofing, increased hardness and ability to achieve a polished surface, would have been advantageous.

34


7. GUIDANCE FOR CONSERVATION PRACTICE

7.1 Sourcing and Specifying Materials A range of contemporary sources for earth was used in the tests, including working agricultural land, former and working clay pits and overburden from open cast mining. These types of sources often supply materials from a greater depth than was the case in traditional practice. As a result, the earths they provide tend to be cleaner, more uniform and have higher clay content than those found in traditional structures. This means that such contemporary sources often require significant tempering to achieve appropriate grading and tend to produce a very even appearance.

Only one of the six types of earth sourced for the tests could be described as replicating the common traditional characteristics of being usable as dug and producing a wall with some variation of colour and texture. While such soil sources may be closer to traditional practice, they require more careful quality control during preparation of the material.

When re-using material sourced from an existing structure to effect a repair, it is important to recognise that the material may have changed from its original composition. In particular, fines, including the clay binder, may have been eroded, or leached out, and should be reinstated to achieve an effective and compatible material. Laboratory testing is an appropriate means of establishing whether this is required.

Careful consideration should also be given to the appropriate moisture content of the repair material, in order to control differential shrinkage between the repair material and the original fabric. This may lead to the decision not to replicate the original construction technique. The use of dried bricks with wet mortar, both made of the selected repair material, can be an effective means of controlling shrinkage in some situations.

This project demonstrated that one-off manufacture of bricks for an individual project needs to be well planned, with adequate time allowed for drying. Commercially produced, air-dried, clay/fibre bricks are available from a number of UK and European sources and these provide potential repair materials with tested characteristics. However, the properties of these materials vary and some are not recommended by their manufacturers for use in externally exposed situations. Care should always be taken to ensure that appropriately compatible repair materials are selected.

A similar careful selection should be made of fibres to be used in repairs. It may be appropriate to use a finer, stronger fibre, such as flax, to achieve a well-bonded repair, rather than replicate the original straw. A view should also be taken on whether to specify organic or non-organic material. While the tests did not detect

a significant difference in short-term durability, there remain concerns in the long-term over the use of modern straws. A precautionary approach would lead to the use of organic straw, which is in limited supply and seasonally produced.

In all repair situations, it is important to design and specify a programme of repairs as a holistic approach to the structure after carrying out a thorough inspection and developing a full understanding of the original construction techniques and all subsequent changes to the building or structure. In developing this understanding, useful reference could be made to the techniques and weathering patterns described in the archive and visible in the standing test walls associated with this project.

7.2 Soil Types and Methods of AnalysisThe project demonstrated that a good range of Scottish sub-soils are appropriate as the basis for an earth building mix. Expansive clays are relatively rare and, in the tests that were conducted with them, they generally demonstrated adequate performance. However, a precautionary approach would avoid use of such clays, except as waterproofing layers.

Most lowland sub-soils could potentially be used as the basis for a building mix, but to achieve an appropriately graded material, many would require a level of processing and modification that would render the process inefficient. In order to avoid the expense, effort and time involved in significantly altering a soil to achieve a workable mix, it is important to begin with a soil that is close to the desired grading.

Generally, a sandy loam, without expansive clay content, with few stones and an even texture, will be an appropriate base subsoil for most purposes. In practice, to achieve this will often require the minor addition of sand, riddling out or crushing of stones, or blending to achieve evenness.

It is possible to identify the location of appropriate subsoils by reference to the Soil Survey of Scotland maps, 1:63,360 and 1:50,000 series, together with the associated Soil Map Unit Descriptions, which give information on individual soil characteristics. These are a useful tool, in combination with local knowledge and surviving earth structures, to locate possible sources of appropriate soils in most lowland areas.

The James Hutton Institute, previously known as The Macaulay Institute for Soil and Research, Aberdeen has considerable expertise in the distribution and characteristics of subsoils and may give useful guidance on possible sources for appropriate soils.

35


These methods should give a general indication of sources of appropriate soils, but it is important to carry out tests on the proposed material to confirm its specific characteristics. This project found that methods of field analysis are usually a sufficiently accurate indicator of soil characteristics. For the purposes of this research, field tests were carried out on most soils to establish colour, texture, workability (plasticity, compaction and cohesion), shrinkage, sedimentation, and organic content. However, in some cases the results were misleading and required careful interpretation.

In situations where the field tests are inconclusive or where a more precisely specified material is required, laboratory testing will provide more accurate results. For the purposes of this research, most of the soils were subject to laboratory testing for grading (wet sieving, dry sieving, sedimentation) and expansiveness (plastic limit, liquid limit, linear shrinkage to BS 1377). These were carried out at Bath University Department of Architecture and Civil Engineering, which has considerable experience of analysing soils for use in construction.

7.3 Skills and Working Methods7.3.1 Contractor SkillsThe project demonstrated that it is possible to use unskilled and inexperienced labour to successfully construct earth walls. However, this did result in a variable quality of work, with some reduced durability. Achieving a good and consistent quality of materials preparation and application requires a contractor with skill and experience. These traditional skills have begun to be re-established, through the family of recent earth building projects in Scotland, which have included applied research, experimental and demonstration structures, conservation, and new buildings.

The number of specialist skilled contractors and consultants in this field is currently small and it is important that their skills and experience are used to achieve a good quality of work in projects involving earth materials and to foster the development of the skills base.

The Scottish Lime Centre has established a capability for training and technical support in relation to modern mass-produced clay plasters. Though these materials have important differences from traditional clay plasters and renders, the skills involved are broadly similar.

7.3.2 MechanisationBuilding with earth is a labour intensive activity. This is a significant factor behind its decline as the construction industry ’modernised’ during the 19th century. The project demonstrated that manual mixing and application can achieve a better mix in certain circumstances, but is usually inefficient for all but the smallest of works. Generally, mechanisation, using a variety of standard and purpose-made equipment, achieves an efficient and consistent process of construction and should be used wherever practicable.

In this programme tractors and JCBs were used to mix and prepare earth materials, crushing larger stones. A roll pan mixer was used for the preparation of some renders. Elsewhere, a vertical pugging mill has been very successful in preparing a variety of earth materials.

7.3.3 SouringThe maturing of mixes over several days was generally demonstrated to be of significant benefit, by allowing materials to be combined more easily, with less added water, to produce mixes that were sticky and malleable at the time of application. This effect is thought to relate to a re-ordering of the structural patterns within the clay minerals, brought about by their electro-magnetic attraction (Minke, 2000).

The benefits of pre-mixing were particularly beneficial for the very stiff, more clay-rich soils. With such soils, mixing can require a higher proportion of water to be added, which then evaporates during the souring period.

7.3.4 Programming and ProtectionOne of the characteristics of working with materials bound with clay, which set by drying, is the importance of controlling the construction programme and conditions so that there is an appropriate timing of tasks and provision of climate protection within a continuous construction programme. This is necessary to achieve a good quality of build, a good bond between lifts, coatings and washes, and appropriate curing, all of which are important factors in the final quality of construction.

Damping down of background earth materials proved an important preparatory task for the application of renders, as it reduced initial suction. Most earth renders needed pressing back to remove cracks, a day after application. However, the best mixes did not require this re-working.

Covering to protect from over-rapid drying out by wind and sun, was only carried out on the lime coatings. The earth renders did not require such protection, as the work was carried out in the spring. Initial frost protection was also not necessary, due to the time of year the work was carried out.

36


8. CONCLUSIONS

8.1 The Test ProgrammeThe programme was successful in providing a valuable insight into traditional techniques of construction whose knowledge and skills base has disappeared. It has provided a useful ongoing resource of information on weathering and durability with relevance to conservation practice.

The programme established a skills and knowledge base among contractors and consultants, which has grown in applications outside the research programme.

In the context of ongoing practice and other research conducted during its seventh year, the programme established several key areas were further research would be valuable.

8.2 The Scottish Traditions of Earth Building

Perhaps unsurprisingly, the techniques that performed best in this test programme were the ones that are found most frequently in traditional Scottish construction. The characteristics of these successful techniques are that they:

• require a low level of processing of typical Scottish sub-soils and associated materials (taking into account that traditionally soils and materials were often sourced in a different way from contemporary practice, do not require specialist equipment)

• can be accomplished with a variable quality of locally sourced materials

• are reasonably tolerant of inclement working conditions

• can achieve a durable construction, requiring minimal maintenance in Scottish climatic conditions.

Regional techniques, using particular local materials required specialist skills, which would historically have existed locally. Techniques from outside the tradition were generally inappropriate to climatic conditions or required significantly more effort.

8.3 MaterialsWhile a wide range of materials can potentially be used successfully in earth construction and the basic principles are simple, the best quality work can only be achieved within quite narrow parameters of materials specification and construction technique.

Of primary importance is grading of the earth. Well-graded materials can be remarkably durable, while poorly graded mixes will never perform well. The second important criterion is compatibility of materials. This affects additives, facing materials, coatings and finishes. The programme produced unexpectedly good and bad

results in a number of tests and these indicated that some of the interactions between materials are complex and not fully understood.

8.4 DurabilityEarth materials can be robust and durable in exposed conditions. There are a series of common decay mechanisms that can be avoided by good design and quality of work. Exposure to wind-driven rain, roof protection and solar orientation are the most important criteria in determining durability for good quality materials.

The first year, when the material is drying, is a period of particular vulnerability to decay. Because of this, targeted frost protection could be appropriate in the first winter. High quality materials can achieve a stable and durable surface after this period. With the passing of at least a year, surface erosion prepares the earth surface to achieve a good key for coatings or lime pointing.

Homogenous wall types proved more durable than composite types. Constructions which had a surface of earth and other, denser, materials, proved to be vulnerable to progressive decay at the interface of these materials. The earth would decay sacrificially with the result that the wall as a whole was less durable than well made monolithic earth walls.

8.5 The Construction ProcessEarth materials are safe and simple to use in construction, but they require a level of skill that is not always appreciated. Control of materials preparation, application and aftercare are very important in determining long-term durability. These factors are of heightened importance in the repair of existing earth buildings because of the complexity of achieving compatibility to existing earth materials.

A small skills base exists in this field, which should be used and developed to ensure a capability for appropriate repair to earth buildings and structures.

8.6 Repair and ConservationThe need for consistent high quality in the specification, preparation and application of materials is especially important in work to existing earth structures, because of the vulnerability to incompatible materials and potential for rapid decay highlighted in this research.

In determining an appropriate programme of repairs, it is critical to have a thorough understanding of the decay mechanisms that have, and are, acting on an existing earth structure.

37


Earth materials have the potential to be durable and compatible repair and conservation materials. This potential extends beyond applications to existing earth structures. Their malleability, reversibility and hygroscopic interaction with water vapour, make them appropriate for work to other materials and structures.

The need for routine maintenance should be recognised in earth structures, particularly those with lime finishes.

8.7 Recommendations for Further Research

A deeper understanding of some of the results of this programme, and how they could influence practice, would be gained by a comparative study with current and traditional practice in other parts of the U.K and Europe. This would broaden the context for the test results by relating them to a wider body of practical experience, and surviving historical examples, than is available in Scotland.

There are significant potential benefits in further research to follow up the best results for earth renders. The tests revealed that these have unexpected potential as appropriate finishes, though the criteria for successful applications are not understood. The potential benefit of further research is heightened by the concerns, raised in the tests, over the suitability of less vapour permeable finishes on earth materials, which are currently extensively used.

It would be beneficial to undertake further research into the long-term the compatibility of lime finishes on earth materials. The tests raised significant questions over this combination of materials, which is currently often considered best practice in conservation situations.

Further research would be merited to develop applied techniques in the conservation of weathered earth structures. The tests indicated a number of issues including mechanical key and leaching of clay binders, where compatibility and repair technique in achieving effective conservation lacks a developed understanding and methodology.

38


9. GLOSSARY

BOSS Where a layer had detached from the surface to which it was applied leaving a void between.

CLAY For grading purposes, this is defined as particle sizes less than 0.002mm.

CLAY AND BOOL A masonry faced construction technique using rounded stones set in clay mortar against shuttering.

CLAYWALL A shuttered construction technique of clay mix and randomly distributed stones.

DAUB An earth mix applied to an armature.

DUNG Animal excrement, not containing fibres.

EARTH For the purposes of construction, this usually means a subsoil obtained below the organic horizon.

GRAVEL For grading purposes, this is defined as particle sizes between 2 and 60mm.

HARL A protective coating applied externally to a wall surface, thrown by trowel to achieve a finished surface.

KEBBER AND MOTT A technique of daub applied to a palisade of vertical posts, driven directly into the ground.

LIFT A vertical layer of construction, usually representing a days work.

MANURE Animal excrement, containing hay fibres.

MUD Earth in a liquid form.

MUDWALL A monolithic construction technique of earth with straw in vertical lifts. Also known as cob.

PAINT A finishing application of liquid, undiluted.

PLASTER A protective coating applied internally to a wall surface, either by trowel or hand.

RENDER A protective coating applied externally to a wall surface, either by trowel or hand and usually given a smooth floated surface.

SAND For grading purposes, this is defined as particle sizes between 0.06 and 2mm.

SILT For grading purposes, this is defined as particle sizes between 0.002 and 0.06mm.

STAKE AND RICE A form of wattlework, with fine vegetation woven horizontally between thicker vertical posts.

TURF Surface vegetation, including their roots and soil bound to them.

WASH A finishing application of liquid, diluted with water.

39


APPENDIX A: EXTERNAL TEST SUMMARY DATA SHEETS

A.1 Mudwall & Claywall Walls 41A.1.1 Control Mudwall: Culzean 41A.1.2 Control Mudwall: Battleby 42A.1.3 Control Mudwall: Fort George 43A.1.4 Cumbrian Mudwall: Culzean 44A.1.5 Local Mudwall: Culzean 45A.1.6 Local Mudwall: Fort George 46A.1.7 Mudwall with Flax: Battleby 47A.1.8 Mudwall with Organic Wheat 48

Straw: BattlebyA.1.9 Mudwall with Cement Repairs: 49

BattlebyA.1.10 Shuttered Mudwall: Culzean 50A.1.11 Shuttered Mudwall: Battleby 51A.1.12 Shuttered Mudwall: Fort George 52A.1.13 Claywall: Culzean 53A.1.14 Claywall: Battleby 54A.1.15 Claywall: Fort George 55

A.2 Rammed Earth Walls 56A.2.1 Rammed Earth: Culzean 56A.2.2 Rammed Earth: Battleby 57A.2.3 Rammed Earth: Fort George 58A.2.4 Rammed Earth with Banded Lime 59

Reinforcement: BattlebyA.2.5 Rammed Earth with Corner Lime 60

Reinforcement: BattlebyA.2.6 Rammed Earth with Cement 61

Additive: BattlebyA.2.7 Stone-faced Rammed Earth: 62

Culzean A.2.8 Stone-faced Rammed Earth: 63

Battleby A.2.9 Brick-faced Rammed Earth: 64

Fort George

A.3 Masonry Walls 65A.3.1 Stone with Earth Mortar: Culzean 65A.3.2 Stone with Earth Mortar: Battleby 66A.3.3 Stone with Earth Mortar: 67

Fort George A.3.4 Clay and Bool: Culzean 68A.3.5 Clay and Bool: Battleby 69A.3.6 Clay and Bool: Fort George 70A.3.7 Earth and Flax Block with Earth 71

Mortar: CulzeanA.3.8 Earth and Flax Block with Earth 72

Mortar: BattlebyA.3.9 Earth and Flax Block with Earth 73

Mortar: Fort GeorgeA.3.10 Turf Block: Battleby 74A.3.11 Turf Block: Fort George 75A.3.12 Turf between Hurdles: Battleby 76A.3.13 Peat Block: Fort George 77

A.4 Panels 78A.4.1 Earth Daub on Stake and Rice: 78

CulzeanA.4.2 Earth Daub on Stake and Rice: 79

BattlebyA.4.3 Earth Daub on Stake and Rice: 80

BattlebyA.4.4 Earth Daub on Stake and Rice: 81

Fort GeorgeA.4.5 Earth Daub on Horizontal Rails: 82

CulzeanA.4.6 Earth Daub on Horizontal Rails: 83

BattlebyA.4.7 Earth Daub on Horizontal Rails: 84

Fort GeorgeA.4.8 Earth Daub on Kebber and Mott: 85

CulzeanA.4.9 Earth Daub on Kebber and Mott: 86

BattlebyA.4.10 Earth Daub on Kebber and Mott: 87

Fort George

40


A.5 Earth Coatings 88A.5.1 Control Earth Render: Culzean 88A.5.2 Control Earth Render: Battleby 89A.5.3 Control Earth Render: Fort George 90A.5.4 Cumbrian Earth Render: Culzean 91A.5.5 Local Earth Render: Culzean 92A.5.6 Local Earth Render: Fort George 93A.5.7 Earth Render onto Masonry: 94

BattlebyA.5.8 Earth and Dung Render: Culzean 95A.5.9 Earth and Dung Render: Battleby 96A.5.10 Earth and Dung Render: 97

Fort George A.5.11 Earth and Manure Render: 98

Fort GeorgeA.5.12 Earth and Wood Shavings Render: 99

CulzeanA.5.13 Earth and Wood Shavings 100

Render: BattlebyA.5.14 Earth and Seaweed 101

Render: Battleby A.5.15 Earth and Ash Render: Battleby 102A.5.16 Earth and Oil Render: 103

Fort George A.5.17 Earth and Tallow Render: 104

Fort George A.5.18 Local Earth Render and 105

Limewash: Fort George

A.6 Lime & Cement Coatings 106A.6.1 Lime Harl: Culzean 106A.6.2 Lime Harl: Culzean 107A.6.3 Lime Harl: Battleby 108A.6.4 Lime and Seaweed Harl: Battleby 109A.6.5 Lime and Whey Harl: Fort George 110A.6.6 Lime and Oil Harl: Fort George 111A.6.7 Lime Stucco: Culzean 112A.6.8 Lime Stucco: Battleby 113A.6.9 Lime Stucco: Fort George 114A.6.10 Cement Render: Battleby 115

A.7 Washes and Paints 116A.7.1 Dung Paint: Culzean 116A.7.2 Dung Paint: Battleby 117A.7.3 Oil Paint: Culzean 118A.7.4 Oil Paint: Battleby 119A.7.5 Limewash: Culzean 120A.7.6 Limewash: Fort George 121A.7.7 Limewash with Oil: Fort George 122A.7.8 Limewash with Whey: Culzean 123A.7.9 Limewash with Lanolin: 124

Fort George

41


APPENDIX A: EXTERNAL TEST DATA SHEETS

A.1 Mudwall and Claywall Walls

A.1.1 Control Mudwall: Culzean

Summer 1996 SW Autumn 2002 NW

Mix: 3:3:1, by volume, earth : sand : gravel, as 10:3, earth mix : straw, + 5-10% water.

Material: Earth: Type: Red Gallowflat clay, a silty clay loam.

Grading: 31% clay, 59% silt, 10% sand.

Sand: Locally sourced, sharp concreting sand.

Gravel: Locally sourced pea gravel, <6mm.

Straw: Old type, organic wheat straw, chopped to max. 300mm lengths.

Precedent: A common form of earth construction in Scotland, as described in TAN 6: 5.02(i) Mudwall. Similar materials were used during the repair of the Cottown Schoolhouse, St. Maddoes, Perthshire.

Method: The mortar was prepared to a stiff mix on the ground, by trampling by foot. The walls were built in two lifts of 500mm, several weeks apart, with the upper lift constructed in wet conditions, leading to poor compaction. The wall face was dressed with a fork and spade a couple of days after construction. The walls had a batter of 50mm over their height, greater than traditional practice in order to achieve stability at the wall head.

Coating: An earth render was applied to the SE half of the wall (A.5.1).

Chronology: Built: May 1996.

Coating Applied: May 1997.

Inspected: Sept. ‘96, April, Aug. & Nov. ‘97, March & Nov. ‘98, March & Nov. ‘99, April & Nov. 2000, Autumn 2001, Autumn 2002.

Current Status: Still standing in March 2004.

Exposure: Sheltered and shaded.

Weathering: The wall remained in good condition with a tight surface. Local erosion developed where the wall projected beyond the roof drip line, creating small gullies on the NW elevation. With light weathering, the surface of the wall gave poor adhesion for render.

Comments: The light weathering is consistent with field observations of historic structures on sheltered sites.

Archive ref: 8.2

42


A.1.2 Control Mudwall: Battleby

Summer 1996 NE Autumn 2002 NE

Mix: 3:3:1, by volume, earth : sand : gravel, as 10:3, mix : straw, + 5-10% water.







Method: The mortar was prepared to a stiff mix on the ground by trampling by foot and use of a tractor. The walls were built in two lifts of 500mm, several weeks apart, with the joint left with a rough key to promote binding between the layers. The wall face was dressed with a fork and spade a few days after construction.




Inspected: Aug. ‘96, March ‘97, July ‘98, Feb. March & Nov. ‘99, April & Nov. 2000, Autumn 2001, Autumn 2002.

Current Status: Dismantled, Spring 2003.

Exposure: Moderate, prevailing SW.

Weathering: The wall generally remained in good condition with constructional tool marks still visible and only local weathering. This was concentrated towards the base, which was least protected by the roof. Here the material remained damp and became loose and friable, with up to 40mm loss. On the most exposed SW side there was some initial washing out of surface fines, exposing the aggregate, but the surface stabilised. On the shaded NE side these was minor seasonal growth of algae near the base.

Deconstruction: The wall fell in solid pieces, requiring a mechanical digger to dismantle.

Comments: The localised weathering is consistent with field observations of historic structures.

Archive ref: 5.1

43


A.1.3 Control Mudwall: Fort George

Summer 1996 NW Autumn 2002 NW








Method: The mortar was prepared to a stiff mix on the ground, by trampling by foot, two days prior to use. The walls were built in two lifts of 500mm, several weeks apart, with the upper lift constructed in wet conditions, leading to poor compaction. The wall face was dressed with a fork and spade a couple of days after construction.

Coating: An earth render was applied to the NW half of the wall (A.5.3).

Chronology: Built: June 1996.


Inspected: Aug. ‘96, Jan, March & June ’97, June ’98, Jan, March & Nov. ‘99, April & Nov. 2000, Autumn 2001, Autumn 2002.


Exposure: Exposed, prevailing W, occasionally severe NE.

Weathering: The wall remained in reasonably good condition. There was some initial washing out of surface fines, which exposed the aggregate, but the surface stabilised to a tight surface, particularly on the most exposed SW elevation. Progressive surface erosion continued, concentrated towards the base, which was least protected by the roof, with the loss of at least 50mm of material. The SW elevation was least eroded, with deep cracking on the SE and damage from water runs on the NE and NW elevations.

Comments: This site is more exposed than any of the locations of observed historic mudwall. This wall compared well with the other mudwalls on the same site (A.1.6, A.1.12). It is thought that the SW face may have received some sheltering effect form the other test walls.

Archive ref: 11.1

44


A.1.4 Cumbrian Mudwall: Culzean


Mix: Earth as dug, with straw in alternate layers, no added water.

Material: Earth: Type: Cumbrian clay, a boulder clay with a large coarse fraction.

Grading: 61% clay and silt, 6% sand, 33% gravel.

Straw: Modern, non-organic barley straw, chopped to max. 300mm lengths.

Precedent: This mudwall mix and technique were suggested and carried out by officers from Carlisle District Council Planning Department, as representing traditional local practice in that area.

Method: The mortar was ground underfoot on the ground, achieving a stiff but plastic consistency in a process that required considerable effort. The wall was built in lifts of 80 -100mm, in a continuous process, with damp straw laid in layers between lifts. The dryness of the mix allowed a rapid construction but required a significant batter to the walls. The straw was trimmed with a billhook.

Coating: An earth render using the same earth was applied to the SE half of the wall (A.5.4).



Inspected: Sept. ‘96, April, Aug. ‘97, Nov. ‘98, March & Nov. ‘99, April & Nov. 2000, Autumn 2001, Autumn 2002.



Weathering: The wall remained in generally good condition, with only minor erosion at lower levels, where some light washing of fines tightened loose and cracked surfaces.

Comments: The mix may have had better working qualities if water had been added. In practice, gauging the correct amount and timing of added water requires experience in order to optimise workability without impeding drying.

Archive ref: 8.1

45


A.1.5 Local Mudwall: Culzean

Summer 1996 SW Autumn 2002 SW


Material: Earth: Type: Drongan clay, a boulder clay containing a large coarse fraction of stone and gravel.

Grading: 43% clay, 22% silt, 17% sand, 18% gravel.



Straw: Modern, non-organic wheat straw, chopped to max. 300mm lengths.

Precedent: A common form of earth construction in Scotland, as described in TAN 6: 5.02(i) Mudwall. This earth resembles that used at the clay bigging at Burns Cottage, Ayr.

Method: The mortar was prepared to a stiff mix on the ground, by compaction by a JCB, which broke up the larger stones. The walls were built in two lifts of 500mm, several weeks apart, with the joint left with a rough key to promote binding between the layers. The second mix was of a poorer quality than the first. The wall face was dressed with a fork and spade a couple of days after construction. The walls had a batter of 50mm over their height, greater than traditional practice in order to achieve stability at the wall head.

Coating: An earth render using the same local material was applied to the SE half of the wall (A.5.5).






Weathering: This wall showed the most weathering of all the mudwalls on this sheltered site, though it still compared favourably with those at the other, more exposed, sites. Weathering was concentrated towards the lower areas on the SW, where the earth turned a dark brown colour compared with the normal ochre. A whitish bloom occurred in the most sheltered areas.

Comments: The higher degree of weathering may be attributed to the high proportion of fines in the earth, its position within the group, the shape of the wall and poor roof protection.

Archive ref: 8.3

46


A.1.6 Local Mudwall: Fort George

Summer 1996 SE Autumn 2002 SE

Mix: 3:6:1, by volume, earth : sand : gravel, as 10:3 mix : straw, + 5-10% water.

Material: Earth: Type: Westerfold’s clay, a rich blue-grey clay, locally used for puddling.

Grading: The sedimentation test indicated an expansive clay content of at least 50%, with a very high concentration of fines.


Gravel: Locally sourced gravel, <20mm.


Precedent: A common form of earth construction in Scotland, as described in TAN 6: 5.02(i) Mudwall.

Method: The mortar was prepared to a stiff mix on the ground, by compaction by a JCB. The walls were built in two lifts of 500mm, several weeks apart, with the joint left with a rough key to promote binding between the layers. Both mixes were rather wet, with the second being built in rain. The wall face was dressed with a fork and spade a couple of days after construction.

Coating: An earth render was applied to the NW half of the wall (A.1.6).



Inspected: Aug. ‘96, March & June ’97, June ’98, Jan, March & Nov. ‘99, April & Nov. 2000, Autumn 2001, Autumn 2002.



Weathering: There was severe initial erosion of the surface fines, but this stabilised and there was less cracking than on the control wall at the same site. Small local areas of erosion developed where clay lumps had not been thoroughly mixed with the aggregate. These weathering patterns continued with most erosion at lower levels on the SE, SW and NE elevations, with up to 60mm loss. The SE elevation was less weathered with tool marks still evident. This wall faired much better than the control mudwall at the same site.

Comments: The wall performed surprisingly well, if the field test indications were correct in indicating and expansive clay content, and this can be attributed to the well-tempered mix.

Archive ref: 11.5

47


A.1.7 Mudwall with Flax: Battleby

Summer 1996 NW Winter 2001 NW Autumn 2002 NW

Mix: 3:3:1, by volume, earth : sand : gravel, as 10:3, mix : flax, + 5-10% water.





Flax: Sourced from a local farm. The flax had naturally retted, separating the fine fibres from the hurds. Only the fibres were used, uncut.

Precedent: There is no known specific example of the use of flax in place of straw, though flax was more available historically than it is at present.

Method: The mortar was prepared to a stiff mix by trampling by foot on the ground. This required considerably more effort than with straw mudwall and worked best when the mortar was mixed before the flax was added. The flax fibres were teased apart and added by hand. The walls were built in two lifts of 500mm, several weeks apart, with the joint left with a rough key to promote binding between the layers. The second mix was of a better quality than the first. The wall face was dressed with a fork and spade a few days after construction.

Coating: An earth and seaweed render was applied to the SE half of the wall (A.5.14.).






Weathering: For six years, the wall remained in generally good condition with weathering focused on the lower parts of the SW and SE elevations, where there was 40 –50mm loss. There was some minor wear at the lift joint and weathering to local surface bulges. In the winter of 2002 the wall lost its roof, leading to rapid erosion, creating fissures on the wall head and loose surfaces and areas of collapse affecting all elevations. Despite this, tool marks were still evident in some areas.

Deconstruction: The wall was easily dismantled, as it had partially collapsed following loss of its roof in 2002.

Comments: Poor mixing led to increased weathering. Good quality mixes has since been produced by a pug mill.

Archive ref: 5.2

48


A.1.8 Mudwall with Organic Wheat Straw: Battleby







Straw: Lower lift: Modern organic wheat straw, chopped to max. 300mm lengths.

Upper lifts: Old type, organic wheat straw, chopped to max. 300mm lengths.


Method: The mortar was prepared to a stiff mix on the ground by trampling by foot and use of a tractor. The walls were built in three lifts of 400, 300 and 300mm, several weeks apart, with the joint left with a rough key to promote binding between the layers. The wall face was dressed with a fork and spade a few days after construction. The first mix was carried out in wet conditions and was poorly consolidated. The upper lift and had an exaggerated batter. The roof gave reduced protection.

Coating: An earth and dung render was applied to the SE half of the wall (A.5.9).






Weathering: The wall remained generally sound, although the bottom lift became heavily eroded as it projected beyond the roof drip line. Here the fines leached out of the surface, leaving a crumbly and open texture. The erosion patterns accelerated after three years, especially on the NE elevation, where up to 100mm was lost. Plants colonised a damp area at the base to the north.


Comments: Poor roof protection led to significant decay, which resembled decay patterns observed at The Schoolhouse, Cottown. There was no discernable difference attributable to the type of straw.

Archive ref: 5.4

49


A.1.9 Mudwall with Cement Repairs: Battleby

Summer 1998 SW Summer 2003 SW







Precedent: Many surviving mudwall buildings have had cement repairs and cement renders applied to them, and these are believed to be inappropriate.

Method: The mortar was prepared to a stiff mix on the ground by trampling by foot and use of a tractor. The walls were built in two lifts of 500mm, several weeks apart, with the joint left with a rough key to promote binding between the layers. The wall face was dressed with a fork and spade a few days after construction. As a test, a section of good wall was cut away and repaired with stone and cement.

Coating: A cement render was applied to the SE half of the wall (A.6.10).




Current Status: Collapsed, Spring 2003.


Weathering: Extreme weathering resulted in fracturing and delamination of the earth, with loss up to 80mm in three years. The wall retained moisture, particularly in the lower section and the earth developed an open, crumbly texture. The earth decayed sacrificially to the cement repairs. Surface breakdown continued, leaving the lower part extremely undermined, friable and damp, with up to 100mm loss and colonisation by plants. While the upper section remained much drier, it also degraded to a lesser extent.

Deconstruction: The wall fell down without assistance in Spring 2003.

Comments: The severe decay of the wall is attributable directly to the cement render.

Archive ref: 5.3

50


A.1.10 Shuttered Mudwall: Culzean

Summer 1996 SE Autumn 2002 NW

Mix: 10:2:1, by volume, earth : sand : gravel, as 10:3, mix : straw, with 5-10% water.






Precedent: There are archival references that indicate the use of shuttered mudwall construction, (TAN 6, 5.02(i)).

Method: The mortar was prepared to a stiff mix by compaction by a JCB, which broke up the larger stones. The walls were built by throwing and light tamping by hand into plywood shuttering, in two lifts of 500mm, several weeks apart, with the joint left with a rough key to promote binding between the layers. Through bolts, laid in sand, were used to prevent the formwork bulging. When the mudwall had shrunk away from the shutter, it was removed, leaving a finished face. There were a few holes in the face where tamping was uneven.

Coating: An earth and wood shavings render using the same local material was applied to the SE half of the wall (A.5.12).






Weathering: There was very little weathering of surfaces, with only small areas of erosion related to abrasion or deficient roof cover.

Comments: The light level of weathering can be attributed to the sheltered climate of the site and the vertical nature of the shuttered walls, which retained less moisture and had better roof protection than the un-shuttered walls, which had a surface batter.

Archive ref: 8.4

51


A.1.11 Shuttered Mudwall: Battleby

Summer 1996 NW Autumn 2002 NW Autumn 2002 SE








Method: The mortar was prepared to a stiff mix by trampling by foot on the ground. The walls were built by throwing and light tamping by hand into plywood shuttering, in two lifts of 500mm, several weeks apart, with the joint left with a rough key to promote binding between the layers. Through bolts, laid in sand, were used to prevent the formwork bulging. When the mudwall had shrunk away from the shutter, it was removed, leaving a finished face. There were a few holes in the face where tamping was uneven. Minor early frost damage was repaired with earth mortar.

Coating: An earth and wood shavings render was applied to the SE half of the wall (A.5.13).



Inspected: Aug. ‘96, March ‘97, July ‘98, Feb, March & Nov. ‘99, April & Nov. 2000, Autumn 2001, Autumn 2002.



Weathering: This was the least weathered wall on this site, with only superficial erosion generally apparent. Where render was lost, the underlying mudwall eroded rapidly, especially in the lower SE elevation, with up to 30mm loss.


Comments: The light level of weathering can be attributed in part to the vertical nature of the shuttered walls, which retained less moisture and had better roof protection than the un-shuttered walls, which had a surface batter. The local loss of earth was linked to local exposure.

Archive ref: 5.6

52


A.1.12 Shuttered Mudwall: Fort George

Summer 1996 SE Autumn 2002 SE


Material: Earth: Type: Westerfolds clay, a rich blue-grey clay, locally used for puddling.






Method: The mortar was prepared to a stiff mix by trampling by foot on the ground. The walls were built by hand with light tamping into plywood shuttering, in two lifts of 500mm, three weeks apart, with the joint left with a rough key to promote binding between the layers. Through bolts, laid in sand, were used to prevent the formwork bulging. When the mudwall had shrunk away from the shutter, it was removed, leaving a finished face. There were a few holes in the face where tamping was uneven.

Coating: An earth and oil render was applied to the NW half of the wall (A.5.16).






Weathering: There was differential shrinkage and weathering between the two lifts. The first lift remained sharp, but erosion developed around the base, with the loss of up to 30mm of material. The second lift continued to shrink after the shuttering was struck and developed more cracks. This lift was coarser, less durable, and developed significant local loss.

Comments: Generally this wall performed very well. The light level of weathering can be attributed in part to the vertical nature of the shuttered walls, which retained less moisture and had better roof protection than the un-shuttered walls, which had a surface batter. The difference between lifts can be attributed to poor mixing and compaction in the second lift.

Archive ref: 11.3

53


A.1.13 Claywall: Culzean







Straw: Modern, non-organic wheat straw, chopped to 300-450mm lengths.

Stone: Field clearance boulders, approximately 150mm dia.

Precedent: A common traditional technique, which used up inferior building stone, as described in TAN 6, 5.03 (iv).

Method: The mortar was prepared to a stiff mix by compaction by a JCB, which broke up the larger stones. The wall was built in two lifts of 500mm, several weeks apart, as a mudwall mix with randomly placed stones laid between ply formwork, stiffened with through bolts, laid in sand. The shutters were removed immediately on completion. Both lifts were of variable quality.

Coating: An earth and dung render was applied to the SE half of the wall (A.5.8).






Weathering: The wall remained in very good condition, with tight surfaces, crisp edges and only small areas of local erosion. The SE elevation showed some loss of surface fines and green algal growth.

Comments: The light level of weathering can be attributed to the sheltered climate of the site and the vertical nature of the shuttered walls, which retained less moisture and had better roof protection than the un-shuttered walls, which had a surface batter.

Archive ref: 8.8

54


A.1.14 Claywall: Battleby

Summer 1996 NW Winter 2001 NW Autumn 2002 NW






Straw: Modern, non-organic wheat straw, chopped to 300–450mm lengths.

Stone: Large mixed rubble.


Method: The mortar was prepared to a stiff mix by compaction by a JCB. The wall was built in two lifts of 500mm, several weeks apart, as a mudwall mix with randomly placed stones laid between ply formwork, stiffened with through bolts, laid in sand. The shutters were removed immediately on completion. The second lift was of lesser quality. A number of cracks soon developed, especially in the upper lift.

Coating: A lime and seaweed harl was applied to the SE half of the wall (A.6.4).




Current Status: Fell down, Winter 2001-2.


Weathering: The wall remained in very good condition, with only superficial weathering, for five years. Following loss of its roof, erosion was very rapid, leading to collapse and complete loss within a year.

Deconstruction: The wall eroded completely prior to dismantling.

Comments: Although the vertical nature of the shuttered walls, would have tended to reduce decay, the breakdown of the wall seems to have been significantly more rapid than that of mudwall in the same circumstances. This is attributable to the composite nature of the material, where the random interfaces between earth and stone presented vulnerable joints to the face, which could channel water into the wall.

Archive ref: 5.11

55


A.1.15 Claywall: Fort George







Straw: Modern, non-organic wheat straw, chopped to 300-450mm lengths.

Stone: Large mixed rubble.


Method: The mortar was prepared to a stiff mix by compaction by a JCB. The wall was built in two lifts of 500mm, several weeks apart, as a mudwall mix with randomly placed stones laid between ply formwork, stiffened with through bolts, laid in sand. The shutters were removed immediately on completion. There were surface voids and initial cracking where stone sat at the face.

Coating: An earth and tallow render was applied to the NW half of the wall (A.5.17).






Weathering: The wall remained in generally good condition, with erosion patterns similar to those of the shuttered mudwall at the same site (A.1.12), with stones revealed as the earth eroded. Most erosion developed adjacent to stones and at corners, with up to 90mm loss. The upper lift became coarser and more pitted than the lower. Local erosion was associated with areas of unmixed earth and there was a friable area at low level on the NE elevation.

Comments: This is thought to have been an unrepresentative mix, which inadvertently gained a very coarse gravel fraction form the mixing site.

Archive ref: 11.7

56


A.2 Rammed Earth Walls

A.2.1 Rammed Earth: Culzean


Mix: Earth as dug.

Material: Earth: Type: Drongan clay, a boulder clay containing a large coarse fraction of stone and gravel. Fresh dug, the drier, crumblier earth from the top of the pile was selected.





Precedent: Also known as Pise, there are archival sources referring to this technique being introduced form France, as described in TAN 6, 5.03(i).

Method: The earth was chopped, dried, sieved and compacted with a manual rammer between ply formwork stiffened with through bolts laid in sand. On the first lift, construction had to stop due to heavy rain. The shutters were removed immediately on completion and although oiled, they did not detach cleanly from the earth face. Shortly after, a vertical crack appeared and a section of wall broke away. Repairs were made during the second lift, using the same mixture but with added hay.

Coating: A lime stucco was applied to the SE half of the wall (A.6.7).



Inspected: Sept. ‘96, April, Aug. & Nov. ‘97, March & Nov. ‘98, March & Nov. ‘99, April& Nov. 2000, Autumn 2001, Autumn 2002.



Weathering: There was considerable initial superficial surface cracking, but no significant deterioration or erosion of the surface, with only a few small water runs developing. The repairs were well keyed and difficult to differentiate from the original material, other than by their fibrous content. There was extensive development of surface green algal growth.

Comments: The defects in the first lift could be attributed to the earth being of too high clay content and used too moist. This technique proved difficult.

Archive ref: 8.6

57


A.2.2 Rammed Earth: Battleby

Summer 1996 NE Spring 2003 NE

Mix: 3:3:1, by volume, earth : sand : gravel.





Precedent: Also known as Pise, there are archival sources referring to this technique being introduced form France, as described in TAN 6, 5.03(i).

Method: The earth was chopped, dried and sieved and compacted with a manual rammer, in ply formwork stiffened with braces and through bolts. The first lift attempted was completed in a day and shutters struck immediately. A large crack soon appeared down the centre of the wall, attributed to movement of the foundations. Apart from this, the material had worked well and the wall was taken down and material prepared for re-use. The second attempted first lift used a slightly drier mix and the formwork was left in place for three days. However a similar large vertical crack again appeared. The crack was repaired before the second lift commenced. This was carried out in the same manner, though compaction was more varied. No vertical crack appeared in the upper lift.

Coating: A lime stucco was applied to the SE half of the wall (A.6.8).






Weathering: Generally the wall remained in good condition. There was considerable initial fine surface cracking, concentrated towards the base and corners, though loss of the initial surface in 1999 revealed a tighter face. Surface erosion continued at a higher rate than other shuttered walls on this site, particularly on the SE elevation, creating damp, crumbly and delaminated surfaces with up to 30mm loss. The upper lift developed a rougher texture than the lower.


Comments: The cause of the cracks in the first lifts was movement of the foundations due to the pressure exerted during compaction of the earth. The other defects could be attributed to the earth having too high a clay and silt content and being too moist. This technique proved difficult.

Archive ref: 5.7

58


A.2.3 Rammed Earth: Fort George


Mix: Earth, as dug.

Material: Earth: Type: Kinnoir clay, a grey-green clay, used as it was grainier and more crumbly than the Westerfolds clay.

Grading: 68% clay, 13% silt, 8% sand, 11 % gravel (jar sedimentation test). Other field tests suggested a more even balance between the clay and silt fraction.

Precedent: Clay and turf buildings are found in they vicinity of the earth source.

Method: The earth was chopped, dried, sieved and rammed with a manual rammer. The ply formwork was oiled and stiffened with through bolts laid in sand. The first lift went well, though one corner was damaged on removing the shuttering, two days after ramming. The second lift occurred during heavy rain showers, which collected in the shutter and made the mix too damp to ram properly. The completed wall had a variable finish with a great many surface voids, attributable to the wet conditions. Minor repairs were undertaken prior to rendering.

Coating: A lime stucco was applied to the NW half of the wall (A.6.9).



Inspected: Aug. ‘96, March & June ’97, June ’98, Jan, March & Nov. ‘99, March & Nov. 2000, Autumn 2001, Autumn 2002.


Exposure: Very exposed, prevailing W, occasionally severe NE.

Weathering: The wall surface quickly became cracked and uneven, with the main areas of damage at the corners and lift joints, areas where compaction is likely to have been less. This increased to severe erosion in three years leaving a honeycomb texture of hollows and cracks. Decay at this time was focused on the NE side, indicating frost damage to clay rich parts of the mix. Erosion continued over the next three years, with up to 120mm loss on all faces, except the SE, which remained relatively intact. Over 75% of the surface was loose and friable. This was probably the most weathered wall still standing.

Deconstruction: The wall proved surprisingly solid in the centre during demolition.

Comments: The sustained high rate of loss relates to the earth composition and variable build quality.

Archive ref: 11.4

59


A.2.4 Rammed Earth with Banded Lime Reinforcement: Battleby


Mix: Earth mortar: Earth, as dug.

Lime mortar: 1:3, by volume, lime putty : sand.

Material: Earth: Type: Brown Gallowflat clay, a silty clay loam.

Grading: 54% silt, 27% clay, 9% sand, 10% gravel.


Lime: Shap lime putty (non-hydraulic).

Precedent: This wall was based on a description of a wall constructed by Scots masons for the Czar of Russia in the late eighteenth century. It is thought that the lime bands were designed to improve the key of lime finishes.

Method: The earth was chopped, dried and sieved and compacted with a manual rammer, into ply formwork stiffened with through bolts laid in sand. Bands of lime mortar, 50mm high and 100mm deep, were run to the perimeter at the base of each 70mm earth layer. The first lift was completed with a good finish achieved. An attempt to build the second lift four weeks later in wet weather was abandoned when the wall became unstable and the mix became too wet to achieve compaction. The shutters were removed some days later when the wall had dried out.

Coating: No finish was applied.


Inspected: July ‘98, Feb., March & Nov. ‘99, April & Nov. 2000, Autumn 2001, Autumn 2002.



Weathering: Generally remained in good condition, especially the upper section, which had tight surfaces and sharp corners. The lower section was more weathered, with loss of up to 40mm at the base. On exposed faces, the earth eroded sacrificially, leaving the lime layers proud of the earth surface. Cracks developed at the joint of these materials, leading to further erosion. The wall supported algal growth, especially on the lower NE elevation.

Deconstruction: The wall fell in very solid pieces, requiring a mechanical digger to dismantle.

Comments: The wall weathered better than the other lime reinforced wall (A.2.5).

Archive ref: 5.18

60


A.2.5 Rammed Earth with Corner Lime Reinforcement: Battleby


Mix: Earth mortar: Earth, as dug.

Lime mortar: 1:3, by volume, lime putty : sand.




Lime: Shap lime putty (non-hydraulic).

Precedent: This form of construction is used in contemporary French practice, and is thought to have developed from traditional techniques.

Method: The earth was chopped, dried and sieved and compacted with a manual rammer, into oiled ply formwork, stiffened with through bolts. Vertical bands of lime mortar were trowelled into each corner prior to ramming. The first lift was completed with the loss of some material on removing the shutters. The second lift was carried out four weeks later. The lime was visible in the corners but did not form a defined fillet.



Inspected: July ‘98, Feb., March & Nov. ‘99, April & Nov. 2000, Autumn 2001, Autumn 2002.



Weathering: Generally, the wall remained in good condition, especially the upper section, which had tight surfaces and sharp corners. The lower section was more weathered, with loss of up to 50mm earth and 30mm lime/earth at the base. The lime appeared to slow the rate of earth erosion, though the earth did erode sacrificially, leaving the lime layers proud of the earth surface. The base of the wall remained damp and supported algal growth. In common with the wall at Culzean (A.2.1) and the other lime reinforced wall (A.2.4), this wall did not suffer the significant cracking and delamination that occurred to the rammed earth walls (A.2.2, A2.3). The surface was also tighter and firmer.


Comments: The lack of cracking and delamination was mainly attributed to the dry construction conditions and type of earth.

Archive ref: 5.17

61


A.2.6 Rammed Earth with Cement Additive: Battleby


Mix: 9:1, Earth : cement.



Cement: Ordinary Portland Cement.

Precedent: This is a modern technique of ‘stabilised’ earth construction.

Method: The earth was chopped, dried, sieved and compacted with a manual rammer, into ply formwork stiffened with through bolts laid in sand. An even mix of the earth and cement was difficult to achieve manually. The threaded rods were impossible to remove. The finished wall was very solid and free of cracks.






Weathering: The wall showed very little apparent weathering, other than a small patch near the SW base and slight damage at the corners.


Comments: This wall should be considered as a weak concrete wall rather than an earth wall, as the inclusion of cement in the mix fundamentally changes its characteristics.

Archive ref: 5.8

62


A.2.7 Stone-faced Rammed Earth: Culzean


Mix: Rammed earth: 5:3, by volume, earth : sand.

Earth Mortar 1:2, by volume, earth : sand.

Material: Earth: Type: Drongan clay, a boulder clay containing a large coarse fraction of stone and gravel. Fresh dug, the drier, crumblier earth from the top of the pile was selected.



Stone: Small pieces of freshly cut local sandstone rubble.

Precedent: A variation of the standard rammed earth technique, as described in TAN 6, 5.03 (v).

Method: The earth was chopped, dried, sieved. The stones were laid in earth mortar to the shutter face and earth compacted in behind with a manual rammer. The ply formwork was oiled and stiffened with through bolts laid in sand. The second lift was affected by rain, which gathered in the formwork. The force of earth being rammed into the core caused the mortar to squeeze out from the stone beds and onto their face, leaving a surface resembling plaster.

Coating: A lime harl was applied to the SE half of the wall (A.6.1).






Weathering: The wall remained in very good condition, showing virtually no weathering.

Comments: The sheltered climate protected the vulnerable earth joints.

Archive ref: 8.5

63


A.2.8 Stone-faced Rammed Earth: Battleby


Mix: Rammed earth: 3:3:1, by volume, earth (brown) : sand : gravel.

Earth Mortar 1:1, by volume, earth (red) : sand.



Earth: Type: Red Gallowflat clay, a silty clay loam.




Stone: Broken sandstone flagstone rubble.

Precedent: A variation of the standard rammed earth technique, as described in TAN 6, 5.03 (v).

Method: The earth was chopped, dried, sieved and compacted with a manual rammer, in ply formwork stiffened with braces and through bolts, laid in sand. The masonry was laid to the wall face and earth packed in behind as the wall core and carried into the stone beds. The first lift shuttering was struck a week after completion. The second lift was less tightly built.




Inspected: Aug. ‘96, March ‘97, July ‘98, Feb., March & Nov. ‘99, April & Nov. 2000, Autumn 2001, Autumn 2002.



Weathering: Generally the wall remained in good condition, but with significant local erosion of mortar joints towards the base. The wall initially developed horizontal cracks, at approximately 50mm intervals, and these were more pronounced on the upper lift, but they tightened up as the surface fines weathered. The lower vertical mortar joints weathered up to 60mm and horizontal joints up to 40mm.

Deconstruction: The wall fell in solid pieces, requiring a mechanical digger to dismantle, but broke up easier than the monolithic walls.

Comments: The cause of the cracks in the first lifts was movement of the foundations due to the pressure exerted during compaction of the earth. The other defects could be attributed to the earth having too high a clay and silt content and being too moist. This technique proved difficult.

Archive ref: 5.5

64


A.2.9 Brick-faced Rammed Earth: Fort George


Mix: Rammed Earth: Kinnoir earth, as dug.

Brickwork Mortar: 1:2, Westerfolds earth : sand.

Material: Earth: Type: Kinnoir clay, a grey-green clay, this earth was used as it was grainer and more crumbly than the Westerfolds clay.


Earth: Type: Westerfold’s clay. A rich blue-grey clay, locally used for puddling.


Sand: Local sharp concreting sand.

Brick: Recycled handmade bricks, locally sourced.

Precedent: Many brick faced earth buildings exist around Errol and elsewhere, though these are usually mudwall and the facing a later alteration. As described in TAN6, 5.03(vi).

Method: The bricks were laid in 10mm beds of Westerfolds earth mortar around the inside of the shuttering with Kinnoir earth rammed in behind with a manual rammer as the wall core. It was felt that the Westerfolds earth was too clay rich to use as rammed earth. The Kinnoir earth was chopped, dried and sieved. The ply formwork was oiled and stiffened with through bolts laid in sand. Both lifts went well and took half the time of the stone-faced rammed earth. The bricks formed better corners and it was much easier to ram the core to a regular face.

Coating: An limewash with oil was applied to the NW half of the wall (A.7.7).




Current Status: Still standing, March 2004.


Weathering: The wall suffered significant erosion of mortar, though this seemed to stabilise. There was some initial cracking in the mortar beds, but these tightened as the surface fines weathered away. A crack, which developed in the NW brickwork, also closed. The mortar joints eroded, with up to 40mm loss in three years and up to 50mm in seven years, with the majority of the joints losing 20-30mm and the SE elevation retaining some full joints.

Comments: Such a loss of earth mortar would have traditionally have been repaired with a lime pointing.

Archive ref: 11.8

65


A.3 Masonry

A.3.1 Stone with Earth Mortar: Culzean


Mix: 5:3, by volume, earth : sand.

Material: Earth: Type: Drongan clay, a boulder clay containing a large coarse fraction of stone and gravel. Fresh dug, the drier, crumblier earth from the top of the pile was selected. Stones over 20mm were removed.



Stone: Locally sourced, sandstone rubble off-cuts.

Precedent: A common form of earth construction in Scotland, as described in TAN 6, 5.02(vii).

Method: A traditional random rubble wall was built with mortar beds of 10-15mm, combined with small pinnings to achieve a tight finish. The stones were wetted before use and the mortar was left with a rough texture. The wall was built directly off the ground, with no plinth.

Coating: A whey limewash was applied to the SE half of the wall (A.7.8).






Weathering: The wall generally remained in good condition, though there was erosion and dampness of the mortar towards the base, with up to 30mm loss on the NE elevation, and 10mm on the SE and SW. Vertical joints weathered more than horizontal joints. The weathered mortar developed a dull brown colour, while the unweathered mortar retained its original ochre colour.

Comments: The change of colour indicated a leaching out of mineral fines.

Archive ref: 8.12

66


A.3.2 Stone with Earth Mortar: Battleby


Mix: 1:1, earth : sand, as mortar.




Stone: Local sandstone rubble.


Method: A traditional random rubble wall was built with mortar beds of 10-15mm, combined with small pinnings to achieve a tight finish. The stones were wetted before use and the mortar was left with a rough texture. The wall was built directly off the ground, with no plinth.

Coating: An earth and ash render was applied to the SE half of the wall (A.5.15).






Weathering: The wall generally remained in very good condition, though the mortar tended to erode back from the face, leaving vulnerable gaps in the joints. This was only significant on the lower sections of the NE and SE elevations, where there was up to 50mm loss. The SW elevation was still flush in places and here there were fine cracks in the mortar. The lower sections showed more weathering and on the NW elevation this area remained damp. Algae and plant colonisation occurred on the lower SE and NE elevations, but did not seem to accelerate decay.

Deconstruction: The wall fell in solid pieces, requiring a mechanical digger to dismantle, but broke up easier than the monolithic walls.

Comments: The mortar was relatively rich. Frost damage was worse in the joints at the base because this area did not dry out effectively.

Archive ref: 5.16

67


A.3.3 Stone with Earth Mortar: Fort George


Mix: 1:2, earth : sand, as mortar.




Stone: Local sandstone rubble.


Method: The mortar was mixed by foot to a fairly wet mix. A traditional random rubble wall was built over two days, with the stones wetted before use. The mortar was finished flush, with small pinnings used to achieve a tight finish. The mortar was left with a rough texture. The wall was built directly off the ground, with no plinth.

Coating: An limewash with lanolin was applied to the NW half of the wall (A.7.9).




Current Status: Partially collapsed, March 2004.


Weathering: There was considerable initial shrinkage of the mortar, which left cracks between the mortar and the stones, especially in the vertical joints. The mortar weathered up to 20mm in three years. In the lower 300mm, the mortar remained damp and became friable, with some frost damage. The wall lost its roof in winter 98-99 and this led to accelerated erosion and partial collapse. Plants colonised the core and the clay binder leached out of the mortar in this area. The remaining wall appears to be poorly consolidated, but still standing.

Comments: The decayed wall was very comparable with many standing ruins in Scotland. The repair of this type of decayed structure is problematic because so much of the clay binder can have leached out.

Archive ref: 11.2

68


A.3.4 Clay and Bool: Culzean


Mix: 10:2:1, by volume, earth : sand : gravel as 10:3, mix : straw, with 5-10% water.




Straw: Modern, non-organic wheat straw.

Stone: Locally sourced, rounded field stones, approximately 150mm dia.

Precedent: This technique was popular on the Moray coast, where there are large banks of regularly sized, sea-washed stones. Described in TAN 6: 5.03(iii). Traditional sources indicate it could take up to five years for the earth mortar to weather back sufficiently to allow lime pointing to be carried out, prior to the walls being limewashed.

Method: The stones were laid in courses in earth mortar to the wall face, against oiled ply formwork, stiffened with through bolts laid in sand. Earth was packed in behind as the wall core. The first lift was not well compacted and there was some rain damage to the head. The second lift was of variable quality. Poor consolidation created a rough, open surface, when the shuttering was struck. Repairs were carried out with earth mortar.

Coating: No applied finish.


Repaired: May 1997.




Weathering: The wall generally remained in good condition, but with significant localised weathering concentrated at the base and corners. The repairs tended to loose their key to the background material, with 50% loose within three years. Most cracking occurred in the background material.

Comments: Although the wall experienced more local damage than the other shuttered walls on this site, it should be noted that the wall was poorly consolidated during construction and survived extremely well considering the poor initial surface. The varied quality of construction was not more strongly revealed in weathering patterns because of the sheltered nature of the site.

Archive ref: 8.9

69


A.3.5 Clay and Bool: Battleby

Summer 1996 SW Spring 2003 SW

Mix: 3:3:1, by volume, earth : sand : gravel as 10:3, mix : straw, with 5-10% water.




Straw: Lower 5 courses: ‘Random Rivet’ wheat straw.

Upper 2 courses: Flax straw.

Stone: Rounded field stones, approximately 150mm long x 100mm wide.


Method: The stones were laid in courses in earth mortar to the wall face, against oiled ply formwork, stiffened with through bolts laid in sand. Earth was packed in behind as the wall core before the next mortar bed was laid on top. The first lift comprised four courses. The second lift comprised three courses. When shuttering was struck there was some loss of mortar at the corners. Repairs with earth mortar were carried out.



Repaired: May 1997.




Weathering: The wall generally remained in good condition, with weathering concentrated on the corners and lower SW and SE elevations, where the mortar suffered surface cracking and delamination, with up to 80mm loss. The repairs appeared to weather less than the base material, though they tended to loose their key, with 30% boss within three years.

Deconstruction: The wall fell in solid pieces, requiring a mechanical digger to dismantle, but broke up more easily than the monolithic walls.

Comments: The stones were difficult to bond, as the face was not visible during construction. Experience would allow a consistent pattern and quality to be achieved.

Archive ref: 5.12

70


A.3.6 Clay and Bool: Fort George



Material: Earth: Type: Westerfolds clay, a rich blue-grey clay, locally used for puddling.



Straw: Old style, organic wheat straw, chopped.

Stone: Rounded beach stones, approximately 100-150mm dia.


Method: The stones were laid in courses in earth mortar to the wall face, against oiled ply formwork, stiffened with through bolts, laid in sand. Earth was packed in behind as the wall core before the next mortar bed was laid on top. The first lift progressed well, but the second lift was affected by heavy rain and finished quickly. When the shuttering was struck there was some loss of mortar at the corners. Repairs with lime mortar were carried out the following year.





Current Status: Still standing, March 2004.


Weathering: There was some initial horizontal cracking along the bed lines, which later tightened. The wall generally remained in good condition in the centre of the elevations, with some areas remaining almost unweathered. The corners eroded badly with progressive loss of up to 100mm of mortar. The stones in this wall were more vulnerable to loosing their key than at other sites and as erosion progressed these fell out. The repairs weather less than the base material and tended to loose their key, with 50%loss and 20% boss within three years.

Comments: The decay indicated a clay rich mortar or poor grading.

Archive ref: 11.9

71


A.3.7 Earth and Flax Block with Earth Mortar: Culzean


Mix: Bricks: Earth as dug, as 10:3, earth : flax, in 440 x 220 x 100 blocks.

Mortar: 5:3, by volume, earth : sand.




Flax: Sourced from a farm. The flax had naturally retted, separating the fine fibres from the hurds. Only the fibres were used, uncut.

Precedent: Traditional in regions of England, but not known to have been historically used in Scotland. Similar materials were used in the repair of the Cottown Schoolhouse (ref. future TAN?).

Method: The blocks were hand made in wooden moulds and air dried for three months prior to use. The mortar was mixed fairly wet, by foot. The blockwork was built up in one day in courses with mortar beds of 10–15mm, with the blocks dipped in water prior to use. If the stony mortar was too firm, the blocks cracked as they were tamped into place.







Weathering: Generally this wall remained in very good condition. There was some slight cracking of the bricks in exposed areas and some weathering of surface fines. Clumps of unmixed flax seemed to cause weak areas prone to erosion.

Comments: Although the flax fibre was fine, it was felt that chopped straw would have performed better, giving a more even mix.

Archive ref: 8.7

72


A.3.8 Earth and Flax Block with Earth Mortar: Battleby


Mix: Bricks: 5:3:2, earth: sand : gravel, as 10:3, mix : flax , in 440 x 220 x 100 blocks.

Mortar: 1:1, by volume, earth : sand.




Flax: Sourced from a local farm. The flax had naturally retted, separating the fine fibres from the hurds. Only the fibres were used, uncut.

Precedent: Traditional in regions of England, but not known to have been historically used in Scotland. Similar materials were used in the repair of the Cottown Schoolhouse.

Method: The blocks were hand made in wooden moulds and air dried for three months prior to use. The mortar was mixed fairly wet, by foot. The blockwork was built up in one day in courses with mortar beds of 10 –15mm, with the blocks dipped in water prior to use.







Weathering: There was some cracking and spalling of the surface of the blocks on exposed elevations, with significant loss of up to 30mm on the SE and SW elevations, near the base where the surface remained damp. By contrast the NW elevation was virtually unweathered. The surface of the earth bricks seemed less resistant to decay through wetting and drying cycles than comparable mudwall mixes. There was no visible difference in weathering between the blocks and mortar.

Deconstruction: The wall fell in very solid pieces, with the blockwork remaining well bonded.

Comments: The mix was not well graded, perhaps too rich in fines. The flax did not add resilience, unless it was well mixed with the earth, but this was difficult to achieve and the benefits were localised as a result. Although the flax fibre was fine, it was felt that chopped straw would have performed better, giving a more even mix.

Archive ref: 5.10

73


A.3.9 Earth and Flax Block with Earth Mortar: Fort George


Mix: Bricks: 5:6:4, earth : sand : gravel, as 10:3, earth : flax , in 440 x 220 x 100 blocks.

Mortar: 1:2, earth : sand, as mortar.




Flax: Sourced from a farm. The flax had naturally retted, separating the fine fibres from the hurds. Only the fibres were used, uncut.

Precedent: Traditional in regions of England, but not known to have been historically used in Scotland. Similar materials were used in the repair of the Cottown Schoolhouse.

Method: The blocks were hand made in wooden moulds and air dried for three months prior to use. The mortar was mixed fairly wet, by foot. The blockwork was built up in one day in courses with mortar beds of 10 –15mm, with the blocks dipped in water prior to use.

Coating: A lime and oil harl was applied to the NW half of the wall (A.6.6).






Weathering: A dense network of fine to medium cracks initially developed in both bricks and mortar. Severe surface weathering of fines on the SW and NE faces tightened these cracks and obscured the delineation between bricks and mortar. While severe weathering continued causing up to 90mm of loss on the east corner, the wall remained reasonably intact. Newly exposed material suffered cracking and scaling, while weathered surfaces developed a tight, but very pitted surface, with flax fibres standing proud.

Comments: The shrinkage cracking may indicate a too rich mix. Although the flax fibre was fine, it was felt that chopped straw would have performed better, giving a more even mix.

Archive ref: 11.10

74


A.3.10 Turf Block: Battleby


Material: Turf Blocks: Size: 600mm l. x 400mm w. x 120mm h.

Shapes : One type with inclined short sides.One type with inclined long sides.

Source: Surface layer cut from near the site of a former brickworks at Pitfour, Perthshire. The earth was clay rich, but stony and the turf was well established with a dense root system.

Cutting: The area was grided out and the turf sizes set out by a series of vertical and inclined cuts in the ground. A traditional flaughter spade was used to cut the blocks and at an even depth. The stony nature of the soil made it difficult to cut and shape the blocks.

Precedent: This construction was based on the turf gable surviving at Corse Croft, Kinnoir, Huntly, Aberdeenshire, which was the subject of conservation repairs in 1996. All indications are that the turf was traditionally used while still fresh and during damp weather. As described in TAN 6, 3.04.

Method: The wall was constructed in a Garden Wall bond, that is, with courses comprising a row of headers backed by a row of stretchers, alternating with courses comprising a row of stretchers backed by a row of headers. The turfs were laid grass side down. The wall was constructed within twenty-four hours of the turfs being cut, but they had started to dry out and this made compaction difficult to achieve. Due to the stoniness of the soil, the turfs required more care in handling. The wall settled during construction, tightening the joints. The wall was given a clay and turf capping instead of the standard waterproof roof.



Inspected: July ‘98, Feb., March & Nov. ‘99, March & Nov. 2000.

Current Status: Fully collapsed, 2000.


Weathering: The wall rapidly degraded, with partial collapse within two years and full collapse in three.

Deconstruction: Not recorded.

Comments: The wall was lost without an adequate record of decay and collapse. Though the clay and turf capping was probably inferior to the standard type, the rate of decay was probably principally related to the soil character.

Archive ref: 5.19

75


A.3.11 Turf Block: Fort George

Summer 1996 NE Winter 2001 NE

Material: Turf Blocks: Size: 600mm l. x 400mm w. x 120mm h.


Source: Surface layer cut from the edge of a former clay pit at Kinnoir, Aberdeenshire. The earth was clay rich and the turf was well established with a dense root system.

Cutting: The area was grided and the turf sizes set out by a series of vertical and inclined cuts in the ground. A traditional flaughter spade was used to cut the blocks and at an even depth. The stony nature of the soil made it difficult to cut and shape the blocks.

Precedent: This construction was based on the turf gable surviving at Corse Croft, Kinnoir, Huntly, Aberdeenshire, which was the subject of conservation repairs in 1996. All indications are that the turf was traditionally used while still fresh and during damp weather. As described in TAN 6, 3.04.

Method: The wall was constructed in a Garden Wall bond, that is, with courses comprising a row of headers backed by a row of stretchers, alternating with courses comprising a row of stretchers backed by a row of headers. The turfs were laid grass side down. The wall was constructed within twenty-four hours of the turfs being cut, but they had started to dry out and this made compaction difficult to achieve. The wall was given a standard roof.

Coating: An earth render and limewash was applied to the NW half of the wall (A.5.18).


Inspected: Aug. ‘96, March & June ’97, June ’98, Jan, March & Nov. ’99, April & Nov. 2000, Autumn 2001, Autumn 2002.



Weathering: Initially, a number of cracks developed in vertical joints and there was a limited amount of grass growth on the lower SW elevation. Thereafter the wall remained in excellent condition, with sharp edges remaining and only superficial weathering at corners and on high points. Widespread surface colonisation by lichen and algae developed, especially near the base on shaded sides. This was the least weathered wall on this site.

Comments: Weathering patterns follow those observed on Corse Croft, a rare surviving example of turf construction. It is thought that the source used is particularly good for this technique and that this may be partly why Corse Croft has survived. The results of this test should therefore not be directly applied to other sources for turf construction.

Archive ref: 11.6

76


A.3.12 Turfs between Hurdles: Battleby


Material: Turfs: Size: 500mm l. x 300mm w. x 80mm h.

Shapes : One type, with all perpendicular sides.

Source: Surface layer cut from the site. The earth was clay rich, but slightly stony and the turf was reasonably well established with a not very dense root system.

Cutting: The area was grided and the turf sizes set out by a series of vertical cuts in the ground. A traditional flaughter spade was used to cut the blocks and at an even depth.

Precedent: This construction was based on archive descriptions, but it is thought that the wall was thinner than would have been the case in traditional practice. It is thought that traditionally the hurdles would give temporary support to the wall until the turf had settled into a stable condition. As described in TAN6, 3.04.

Method: The wall was constructed in alternating courses of turfs laid grass side down and turfs laid grass side up, between light stake and rice hurdles, which formed the face of the wall. Although not strong, the hurdles allowed the use of smaller, more friable turfs and the wall was built more quickly than the turf wall without hurdles. The wall settled during construction, tightening the joints. The wall was not given a cap.



Inspected: July ‘98, Feb., March & Nov. ‘99, April & Nov. 2000.

Current Status: Fully collapsed, 2000.


Weathering: Both sections developed living grass on top, but little on the faces, apart from near the base where conditions remained more moist. The section with hurdles gained some shelter from these and retained more vegetation, while the other section achieved a stable surface where the courses tightened but remained distinct. The wattled section has shrunk vertically by about 25% at time of removal.

Deconstruction: The walls were damaged, apparently by vandalism and were removed. The turves were easily disassembles and had not grown together.

Comments: It is thought that, had they not been damaged by vandalism, the walls would have remained stable and weathered well. With a roof, they would have been drier and performed differently. The standard coursed turf had little bond strength and was easily broken up by vandals, while the wattles section was afforded protection by the framing, making it overall more durable.

Archive ref: 5.19

77


A.3.13 Peat Block: Fort George

Material: Peat Blocks: Size: 600mm l. x 400mm w. x 120mm h.


Source: Surface layer cut from the edge of a peat bank at Inverbroom, by Ullapool.

Cutting: The material was cut from a working bank. A traditional peat cutting spade was used to cut the blocks and at an even depth.

Precedent: A traditional technique in the north-west, as described in TAN 6, 3.02.

Method: The wall was constructed in a Garden Wall bond, that is, with courses comprising a row of headers backed by a row of stretchers, alternating with courses comprising a row of stretchers backed by a row of headers. The blocks were laid grass side down. The wall was constructed within twenty-four hours of the blocks being cut, but they had started to dry out and this made compaction difficult to achieve. The wall was given a standard roof.



Inspected: Aug. ‘96.

Current Status: Removed by summer ‘97.


Weathering: The wall did not survive beyond the first season and weathering patterns were not recorded.

Comments: It is thought weathering may have been partnered by accidental damage to cause the early loss of this wall.

Archive ref: 11.6

78


A.4 Panels

A.4.1 Earth Daub on Stake and Rice: Culzean


Daub Mix: 10:2:1, by volume, earth : sand : gravel, as 10:3, mix : straw, with 5-10% water.

Material: Earth: Type: Drongan clay, a boulder clay containing a large coarse fraction of stone and gravel. Fresh dug, the larger stones were broken up by machine compaction.




Straw: NE panel: Modern barley straw, chopped to 150mm.

SW panel: Old type, organic wheat straw, chopped to 150mm.

Armature: A double panel of ‘rice’ comprising willow woven with 10-20mm gaps, between vertical ‘stakes’ of hazel at approximately 230mm centres, fixed to a timber frame.

Method: The mortar was prepared to a soft mix by compaction by foot. The straw was added to the mix immediately prior to application. The wheat straw produced a stiffer mix, which was easier to finish by hand though rougher when floated. The mix was pressed into place, in one coat working from both sides, to a floated surface on the SW elevation and a scored surface on the NE elevation.

Finish: Limewash, applied a year later:

NE panel: Made from lime putty.

SW panel: Made from slaked lime.

Precedent: A historically common form of earth construction in Scotland, as described in TAN 6, 5.01(i).


Finish Applied: May 1997.




Weathering: Both panels developed cracks, up to 5mm wide in the barley straw and up to 3mm in the wheat straw. The panel remained in generally good condition, with only about 5% loss of daub. These eroded in a few small areas at the edges, where the fines washed out when a break occurred in the limewash layer. The putty wash appeared to be slightly more durable than the slaked wash. There was no discernable difference attributable to the straw types.

Comments: The straw in the base coat would have worked at 150mm lengths and the top coat at 50mm.

Archive ref: 8.10

79


A.4.2 Earth Daub on Stake and Rice: Battleby

Summer 1996 NE Summer 1998 NE

Daub Mix: 3:3:1, by volume, earth : sand : gravel, as 10:3, mix : straw, + 5-10% water.





Straw: NE panel: Flax fibres, chopped to 150mm lengths.

SW panel: Old type, organic wheat straw, chopped to 150mm.


Method: The mortar was prepared to a soft mix by compaction by foot. The straw was added to the mix immediately prior to application. The wheat straw produced a stiffer mix, which was easier to finish by hand though rougher when floated. The mix was pressed into place, in one coat working from both sides, to a floated surface on the SW elevation and a scored surface on the NE elevation.







Inspected: Aug. ‘96, March ‘97, July ‘98, Feb.



Weathering: Both panels developed uniform cracks, following the horizontal and vertical pattern of armature, though the cracking was worse in the flax fibre mix. The scratched NE side showed more distress than the smooth SW side. Otherwise, the panel remained in generally good condition for three years.

Deconstruction: The wall was dismantled in 1999, with the cause of final collapse unknown, but vandalism suspected.

Comments: The premature loss of the panel prevented a complete assessment.

Archive ref: 5.13

80


A.4.3 Earth Daub on Stake and Rice: Battleby







Straw: NE panel: Modern, non-organic barley straw, chopped to 150mm.

SW panel: Modern oat straw, chopped to 150mm.


Method: The mortar was prepared to a soft mix by compaction by foot. The straw was added to the mix immediately prior to application. The wheat straw produced a stiffer mix, which was easier to finish by hand though rougher when floated. The mix was pressed into place, in one coat working from both sides, to a floated surface on the SW elevation and a scored surface on the NE elevation. This panel has a rougher surface than that of panel ref. A.4.2, with exposed straw standing proud of the surface. The daub had a tight finish.







Inspected: Aug. ‘96, March ‘97, July ‘98, Feb.



Weathering: Both panels were in very good condition after the first year when the finish was applied. The west panel suffered badly during the next year, with 60% loss by 1998, but vandalism is suspected to have been a contributing factor. The other panel, barley straw and lime putty had 25% loss appearing as a more normal weathering pattern due to wind-driven rain. In both cases, once decay had begun and rain was able to penetrate the shallow core, decay was progressive and relatively rapid.

Deconstruction: The wall was dismantled in 1999, with the cause of final collapse unknown, but vandalism suspected.

Comments: These modern straws could be added in larger amounts than the Old organic type used in the other panel (3.4.2) as they are softer, less springy and more absorbant. The result seems to have been fewer cracks and a better finish.

Archive ref: 5.15

81


A.4.4 Earth Daub on Stake and Rice: Fort George


Daub Mix: 3:6:1, by volume, earth : sand : gravel, as 10:3 mix : straw, + 5-10% water.






SW panel: Modern oat straw, chopped to 150mm.


Method: The mortar was prepared to a soft mix by compaction by foot. The straw was added to the mix immediately prior to application. The wheat straw produced a stiffer mix, which was easier to finish by hand though rougher when floated. The mix was pressed into place, in one coat working from both sides, to a floated surface on the SW elevation and a scored surface on the NE elevation. This panel has a rougher surface than that of panel ref. A.4.2, with exposed straw standing proud of the surface. The daub had a tight finish.

Finish: NE panel: An Earth and Dung Render was applied a year after construction (A.5.10).

SW panel: Earth and Manure Render was applied a year after construction (A.5.11).




Inspected: Aug. ‘96, March & June ’97, June ’98, Jan, March & Nov. ‘99, April & Nov. 2000, Autumn 2001.

Current Status: Fully eroded, 2001.


Weathering: The head of the panels soon developed a 10mm crack. The daub suffered severe progressive surface weathering, with 10% loss in three years, concentrated at the top of the panel and on the NE side. The majority was still in place after four years, but after a further year all timber and daub had eroded. The earth/dung render was more resilient to the SW, while the earth/manure was more resilient to the NE.

Deconstruction: The wall did not survive to deconstruction.

Comments: Considering the exposure and nature of the panel, the test proved fairly resilient, with an ability for the surface to self-heal through movement of fines.

Archive ref: 11.12

82


A.4.5 Earth Daub on Horizontal Rails: Culzean







Straw: NE panel: Modern oat straw, chopped to 150mm.

SW panel: Wheat.

Armature: A double panel of hazel rails, 20-30mm dia, at approximately 200mm centres, fixed to a timber frame.

Method: The mortar was prepared to a soft mix by compaction by foot. The straw was added to the mix immediately prior to application. One side was shuttered and the daub was pressed in from the other side. The exposed face was scored in a diamond pattern. The wheat straw worked well. As the oat straw was more brittle, it had to be chopped to shorter lengths and required more effort to mix. The oat straw mix was difficult to apply as the straw broke up during application.

Finish: NE panel: Dung paint (A.7.1).

SW panel: Oil paint (A.7.3).

Precedent: A historically common form of earth construction in Scotland, as described in TAN 6, 5.01(iii).






Weathering: The daub remained intact and showed almost no signs of weathering where the coatings had failed.

Comments: In this sheltered site, this technique performed well.

Archive ref: 8.11

83


A.4.6 Earth Daub on Horizontal Rails: Battleby








SW panel: Wheat, chopped to 150mm.


Method: The mortar was prepared to a soft mix by compaction by foot. The straw was added to the mix immediately prior to application. One side was shuttered and the daub was pressed in from the other side. The exposed face was scored in a diamond pattern. The wheat straw worked well. As the oat straw was more brittle, it had to be chopped to shorter lengths and required more effort to mix. The oat straw mix was difficult to apply as the straw broke up during application.

Finish: NE panel: Dung wash (A.7.2).

SW panel: Oil wash (A.7.4).




Inspected: Aug. ‘96, March ‘97, July ‘98, Feb., March & Nov. ’99, April & Nov. 2000.

Current Status: The wall was dismantled following its collapse between 2001 and 2002.


Weathering: The daub experienced accelerating erosion to the lower half of the panel, with delamination and cracking through the thickness of the daub. There was 30% loss in three years and 75% loss in five. The remaining daub, in the upper section, was heavily delaminated and cracked.

Deconstruction: No record.

Comments: Due to its premature loss, the assessment of this test was incomplete.

Archive ref: 5.14

84


A.4.7 Earth Daub on Horizontal Rails: Fort George


Daub Mix: 3:6:1, by volume, earth : sand : gravel, as 10:3 mix : straw, + 5-10% water.







Method: The mortar was prepared to a soft mix by compaction by foot. The straw was added to the mix immediately prior to application. The panel was shuttered on the SW elevation, with the mix thrown in from the open side, working from the base up, achieving a shuttered finish on the SW and a scored finish on the NE. The shuttering was left in place for three weeks. Shrinkage cracks appeared equally on both sides corresponding with the timber armature.

Finish: NE panel: No applied finish.

SW panel: A lime and whey harl was applied a year after construction (A.6.5).




Inspected: Aug. ‘96, March & June ’97, June ’98, Jan, March & Nov. ‘99, April & Nov. 2000, Autumn 2001.



Weathering: The exposed daub suffered heavy surface erosion. The horizontal layers of exposed straw on the SW elevation contrasted with the more random network of fibres on the NE. Most erosion took place in the fifth year, with the armature becoming exposed on the SW elevation of the unrendered panel.


Comments: The technique did not prove durable in these exposed conditions.

Archive ref: 11.13

85


A.4.8 Earth Daub on Kebber and Mott: Culzean







Straw: Old type, organic wheat straw, chopped to 150mm.

Armature: A vertical palisade of ‘cabers’, unstripped green hazel poles, 50-80mm dia, at 120mm centres, driven 80mm into soft ground, with a shallow drainage trench around the base.

Method: The mortar was prepared to a soft mix by compaction by foot. The straw was added to the mix immediately prior to application. The mix worked reasonably well and was finished by hand.

Finish: A limewash was applied to both sides of the panel the following year (A.7.5).

Precedent: A historically common form of earth construction in Scotland, as described in TAN 6, 5.01(ii).






Weathering: Lack of roof protection led to partial collapse during the first few months, and subsequent local repairs. The remaining daub had a rough texture with lots of cracking. The daub eroded at weak points in the limewash and erosion of friable daub undermined the base. The base area became damp and friable with colonisation by plants, but overall the daub remained mostly intact.

Comments: This panel compared favourably with the one at Battleby, mainly due to its sheltered site.

Archive ref: 8.13

86


A.4.9 Earth Daub on Kebber and Mott: Battleby

Summer 1996 SW Summer 1998 SW Winter 2001 SW








Method: The mortar was prepared to a soft mix by compaction by foot. The straw was added to the mix immediately prior to application. The initial application was too wet and slumped. This was patched a few days later. The daub was finished by hand to a rough surface.

Finish: No applied finish.




Inspected: Aug. ‘96, March ‘97, July ‘98, Feb., March & Nov. ’99, April & Nov. 2000.

Current Status: Dismantled, 2001.


Weathering: Lack of roof protection led to significant erosion in the first few months. The remaining daub suffered from cracking and delamination, with 40% loss within three years. In the next two years there was no more loss, but the surface became more delaminated and fibrous, with little of the original surface remaining. Fungal growth became established on the exposed base of the cabers.

Deconstruction: The panel fully eroded prior to deconstruction.

Comments: This technique did not proved durable at this moderately exposed site.

Archive ref: 5.9

87


A.4.10 Earth Daub on Kebber and Mott: Fort George

Summer 1996 NE Sprring 1999 NE

Daub Mix: 3:6:1, by volume, earth : sand : gravel, as 10:3 mix : straw, with 5-10% water.







Method: The mortar was prepared to a soft mix by compaction by foot. The straw was added to the mix immediately prior to application. Heavy showers meant that the mix was very wet, making it difficult to work. There was little initial cracking.

Finish: NW half: A limewash was applied a year after construction (A.7.6).

SE half: No applied finish.

Precedent: Documented to have been problematic durability as external walls.



Inspected: Aug. ‘96, March & June ’97, June ’98, Jan, March & Nov. ‘99, April & Nov. 2000.



Weathering: Vertical cracks developed on the NE side, corresponding with the armature. With lack of roof protection, erosion quickly developed at the top of the panel. The remaining daub eroded more steadily, with 20% left after three years, which was well keyed to the timber, with a fibrous surface and damp near the base. The daub fully eroded after four years, with fungal growth established at the base of the cabers.


Comments: This technique eroded fastest on this site, due to the exposure.

Archive ref: 11.11

88


A.5 Earth Renders

A.5.1 Control Earth Render: Culzean

Spring 1997 NE Autumn 2002 NE

Mix: 2:4:3, by volume, earth : sand : hay.




Hay: Modern hay, chopped to 20mm, sourced as animal bedding.

Background: Control Mudwall (A.1.1).

Precedent: A historically common form of earth construction in Scotland, as described in TAN 6, 7.

Method: The render was prepared wet and matured for four days prior to application. The base wall surface was dusty and crumbly and was dampened to a slurry prior to the application of the render, which was thrown on and smoothed by hand in a single 25mm coat, feathered in at the base. The render was polished with a trowel the following day. Adhesion was not as good as at the other sites.

Chronology: Background built: May 1996.


Inspected: Aug. ‘97, Nov. ‘98, March & Nov. ‘99, April & Nov. 2000, Autumn 2001, Autumn 2002.



Weathering: The render generally remained in good condition, with minor erosion from exposed areas and some loss of adhesion, mostly on the NE elevation. Overall 5% of the render was lost, 20% boss and 75% remained well adhered. The remaining render had a tight surface of exposed fibres and aggregate, with very few cracks.

Comments: The areas of poor adhesion were due to the inferior base surface, which could be attributable to lack of weathering on a sheltered site. Damp and frost are also thought to have affected the wall.

Archive ref: 8.2

89


A.5.2 Control Earth Render: Battleby









Method: The render was prepared wet and matured for three days prior to application. The base wall surface was sound and dampened prior to the application of the render, which was thrown on and smoothed by hand as a single 25mm coat, feathered in at the base. The render was polished with a trowel the following day. The process was straightforward and adhesion was good.



Inspected: July ‘98, Feb. March & Nov. ‘99, April & Nov. 2000, Autumn 2001, Autumn 2002.



Weathering: The render progressively eroded from exposed areas. On the exposed SW face, almost all the render was lost, apart from that immediately under the roof. Overall 40% of the render was lost, 20% boss and 40% well adhered. The remaining render had a tight surface of exposed fibres and aggregate, with very few cracks.

Deconstruction: The surviving render remained adhered to the wall during dismantling.

Comments: Resilience was good outside the area most exposed to wind-driven rain.

Archive ref: 5.1

90


A.5.3 Control Earth Render: Fort George









Method: The render was prepared wet and matured for three days prior to application. The base wall surface was sound and dampened prior to the application of the render, which was thrown on and smoothed by hand as a single 25mm coat, feathered in at the base. The render was polished with a trowel the following day. Although the material was rather wet, adhesion was good.

Chronology: Background built: June 1996.


Inspected: June ’97, June ’98, March & Nov. ‘99, April & Nov. 2000, Autumn 2001, Autumn 2002.



Weathering: During drying out, the render became boss in places, with fine cracks appearing on the surface. This gradually increased to 50% boss, with severe loss of surface fines. Breakdown on the key occurred near the base of the SE and NW elevations. The render progressively eroded, but least occurred on the SW elevation. Overall 50% of the render was lost, 30% boss and 20% well adhered. The remaining render had a tight and robust surface of exposed fibres. At eroding edges the render appeared loose and friable.

Comments: Weakness in the render, possibly due to an over wet application, was exacerbated by severe exposure.

Archive ref: 11.1

91


A.5.4 Carlisle Earth Render: Culzean


Mix: 2:1, by volume, earth : hay.

Material: Earth: Type: Cumbrian clay, a boulder clay with a large coarse fraction.

Grading: 61% clay and silt, 6% sand, 33% gravel.


Background: Cumbrian Mudwall of the same earth (A.1.4).


Method: The render was prepared wet and matured for two days prior to application. The base wall surface was sound and dampened prior to the application of the render, which was thrown on and smoothed by hand as a single 25mm coat, feathered in at the base. The render was polished with a trowel the following day. The material was very rich and sticky, adhesion was good and it dried to a hard, shiny surface with very little dust. There was considerable cracking as it dried and the surface was re-worked at intervals for two days.



Inspected: Aug. & Nov. ‘97, March & Nov. ‘98, March & Nov. ‘99, March & Nov. 2000, Autumn 2001, Autumn 2002.



Weathering: The surface initially became heavily cracked and 30% boss. Minor erosion in exposed areas and significant loss of adhesion developed, mostly on the NE elevation. Overall, 5% of the render was lost, 40% boss and 55% remained well adhered. The remaining render at upper levels had a tight surface with brush marks still visible, though at lower levels it was damp and loose.

Comments: The render was clay rich, and would have suffered less shrinkage if tempered with sand and aggregate.

Archive ref: 8.1

92


A.5.5 Local Earth Render: Culzean



Material: Earth: Type: Drongan clay, a boulder clay containing a large coarse fraction.



Background: Local Mudwall of the same earth (A.1.5).


Method: The render was prepared wet and matured for two days prior to application. The base wall surface was sound and dampened prior to the application of the render, which was thrown on and smoothed by hand as a single 25mm coat, feathered in at the base. The render was polished with a trowel the following day. The material was rich and had good plasticity. This was the contractors preferred mix for workability. There was cracking as the render dried and the surface was re-worked at intervals.






Weathering: Initially, the render developed a dense network of cracks, followed by considerable loss of surface fines, though these helped to fill some of the cracks. Complete erosion progressively occurred in areas projecting beyond the roof line. Overall, 60% of the render was lost, 25% boss and 15% remained well adhered. The remaining render at upper levels had a tight surface, though at lower levels it was damp and friable.

Comments: The render was clay rich, and would have suffered less shrinkage if tempered with sand and aggregate.

Archive ref: 8.3

93


A.5.6 Local Earth Render: Fort George

Summer 1997 NE Autumn 2002 SW


Material: Earth: Type: Westerfold’s clay, a rich blue-grey clay, Locally used for puddling.




Background: Local Mudwall of the same earth (A.1.6).


Method: The render was prepared wet and matured for three days prior to application. The base wall surface was sound and dampened prior to the application of the render, which was thrown on and smoothed by hand as a single 25mm coat, feathered in at the base. The render was polished with a trowel the following day. The material was stiff to use but applied well and seemed to achieve a good key.






Weathering: The render developed some initial cracks and there was loss of surface fines, though these helped to fill some of the cracks. After three years, 20% of the render had become boss. Thereafter, strong weathering progressively affected the NE and NW elevations. Overall, 40% of the render was lost, 30% boss and 30% remained well adhered. The render on the SW elevation retained a firm, tight finish.

Comments: The render performed better than might have been expected from an apparently expansive clay content.

Archive ref: 11.5

94


A.5.7 Earth Render onto Masonry: Battleby







Background: Stone-faced Rammed Earth (A.2.8).


Method: The render was prepared wet and matured for three days prior to application. The base wall surface was sound and dampened prior to the application of the render, which was thrown on and smoothed by hand as a single 25mm coat, feathered in at the base. The render was polished with a trowel the following day. The process was straightforward and adhesion was good.



Inspected: July ‘98, Feb, March & Nov ‘99, April & Nov. 2000, Autumn 2001, Autumn 2002.



Weathering: The render progressively eroded from exposed areas over the first four years, with the SW elevation worst affected. Thereafter erosion stabilised. Overall 75% of the render was lost, 15% boss and 10% well adhered. The remaining render was located immediately under the roof and retained a tight surface with only superficial damage.


Comments: There was more erosion than on the comparable render onto mudwall, indicating reduced compatibility.

Archive ref: 5.5

95


A.5.8 Earth and Dung Render: Culzean

Spring 1998 NE Autumn 2002 NE Autumn 2002 NW

Mix: 2:4:3, by volume, earth : sand : hay, in 10:1, earth mix : dung.





Dung: Fresh cow sharn.

Background: Claywall (A.1.13).

Precedent: A variant of an historically common form of earth construction in Scotland, as described in TAN 6, 7.

Method: The render was prepared wet and matured for three days prior to application. The dung mixed well with the mortar, adding plasticity and stickiness. The base wall surface was sound and dampened prior to the application of the render, which was thrown on and smoothed to a rough polish by hand in a single 25mm coat, feathered in at the base. The render was polished with a trowel the following day. Adhesion was not as good as at the other sites.






Weathering: The render initially developed a network of fine cracks, up to 2mm wide, 15% became boss and there was the development of green and white algal growth at lower levels. These stabilised after three years and the algal growths declined. There was progressive light surface weathering. Deterioration generally was worst on the SE elevation.

Comments: In a sheltered location, the test performed very well. Compared to the control, it may have had slightly less erosion, but in common with the other sites, dung seems to have encouraged algae growth.

Archive ref: 8.8

96


A.5.9 Earth and Dung Render: Battleby

Autumn 2002 SW Autumn 2002 NE

Mix: 2:4:3, by volume, earth : sand : hay, in a 10:1, earth mix : dung.






Background: Mudwall with Organic Wheat Straw (A.1.8).


Method: The render was prepared wet and matured for three days prior to application. The dung mixed well with the mortar, adding plasticity and stickiness. The base wall surface was sound and dampened prior to the application of the render, which was thrown on and smoothed to a rough polish by hand in a single 25mm coat, feathered in at the base. The render was polished with a trowel the following day. This had the best working qualities of the earth renders.






Weathering: The render suffered some erosion, concentrated in the lower section, which was exposed beyond the roof drip line. Overall 30% of the render was lost, 40% boss and 30% remained well adhered. On the SW face the render remained intact and became only 10% boss. Initial cracking was filled by washed fines on the SW but not the NE and SE elevations. The remaining render had a tight surface, with handprints still visible. The eroding edge was loose. The render developed a green coating, presumably algae, to a much greater extent than on the control render or background mudwall.


Comments: The render performed well, but failed where exposed beyond the roof drip line. Orientation was noticeable in the loss, with the SW more exposed to the drying effects of sun and wind than the NE and to an extent, the SE.

Archive ref: 5.4

97


A.5.10 Earth and Dung Render: Fort George


Mix: 2:4:3, by volume, earth : sand : hay, in a 10:1, earth mix : dung.






Background: Earth Daub on Stake and Rice (A.4.4).


Method: The render was prepared wet and matured for six days prior to application. The render was prepared wet and matured for six days prior to application. The inclusions of dung allowed the earth to break down much more easily and a very sticky mix was produced with only a small amount of water required to produce a smooth paste. The base wall surface was sound and dampened prior to the application of the render. On the SW elevation it was thrown on and then floated to approximately 10mm thickness. On the NE face the render was trowelled on and floated to a scratched surface. Both appeared to adhere well.



Inspected: June ’97, June ’98, March & Nov. ‘99, April & Nov. 2000.

Current Status: Fully eroded by 2001.


Weathering: The render initially performed well, remaining well adhered and with no loss or boss in three years. There was significant surface weathering, though the fines filled initial shrinkage cracks to produce a tight surface with a fine texture, due to the exposure of many small fibres. There was minor frost damage on the NE elevation. The SW, thrown, face performed better than the trowelled. Thereafter, there was rapid loss and all the render was lost by the fourth year.

Comments: The dung was an excellent plasticiser and binder and the render performed well.

Archive ref: 11.12

98


A.5.11 Earth and Manure Render: Fort George


Mix: 1:4, by volume, earth : sand, in a 5:1, earth mix : manure.




Manure: Fresh horse sharn.

Background: Earth Daub on Stake and Rice (A.4.4).


Method: The render was prepared wet and matured for six days prior to application. The inclusions of manure allowed the earth to break down much more easily and a very sticky mix was produced with only a small amount of water required to produce a smooth paste, though the mix was much leaner than the dung mix (A.5.10). The base wall surface was sound and dampened prior to the application of the render. On the SW elevation the render was trowelled on in a 10mm coat with a scratched surface. On the NE face it was thrown onto a floated surface and then floated to approximately 10mm thickness. Both appeared to adhere well.



Inspected: June ’97, June ’98, March & Nov. ‘99, April & Nov. 2000.

Current Status: Fully eroded by 2001.


Weathering: The render rapidly eroded, with 90%loss of material on the SW face and 5% loss on the NE within three years, with no areas boss. There was significant surface weathering, with the fines filled initial shrinkage cracks to produce a tight surface with a fine texture due to the exposure of many small fibres. The NE, thrown, face performed better than the SW towelled. Thereafter, there was rapid loss and all the remaining render was lost by the fourth year.

Comments: The manure produced a leaner mix than that with the dung (A.5.10), with more water required to produce a paste. The render was also less durable.

Archive ref: 11.12

99


A.5.12 Earth and Wood Shavings Render: Culzean

Summer 1997 NE Autumn 2002 SE Autumn 2002 SW

Mix: 1:1:2, by volume, earth : sand : wood shavings.




Wood shavings: Fine shavings from a sawmill planer.

Background: Shuttered Mudwall (A.1.10).

Precedent: A variation of an historically common form of earth construction in Scotland, as described in TAN 6, 7.

Method: The render was prepared wet and matured for two days prior to application. The base wall surface was sound and dampened prior to the application of the render, which was thrown on and smoothed by hand in a single 25mm coat, feathered in at the base. The mix was crumbly and harder to apply than the control mix. The render was polished with a trowel the following day and showed no immediate signs of cracking.






Weathering: Some fine cracks did develop, but adhesion remained good. Overall, none of the render was lost, 5% boss and 95% remained well adhered. The cracks tightened as fines washed out of the surface to reveal the wood shavings in a fluffy, open texture, less defined than the same render at the more exposed Battleby example.

Comments: The render performed very well.

Archive ref: 8.4

100


A.5.13 Earth and Wood Shavings Render: Battleby

Summer 1998 SW Spring 2003 SE

Mix: 1:1:2, by volume, earth : sand : wood shavings.




Wood shavings: Fine shavings from a sawmill planer.


Precedent: A variation of an historically common form of earth construction in Scotland, as described in TAN 6, 7.

Method: The render was prepared wet and matured for three days prior to application. The base wall surface was sound and dampened prior to the application of the render, which was thrown on and smoothed by hand as a single 25mm coat, feathered in at the base. The render was polished with a trowel the following day. The material was crumbly and difficult to apply, but coverage was good and adhesion was good if applied with care. The surface had an open texture when completed.






Weathering: For two years, the render remained in place, with a good bond and some loss of surface fines, particularly to the SW. Thereafter, there was significant loss on the SW and SE elevations. Overall, 45% of the render was lost, 25% boss and 30% remained well adhered. The surviving render had a very open, feathery texture of exposed wood fibres.


Comments: The feathery, open texture may have contributed to durability.

Archive ref: 5.6

101


A.5.14 Earth and Seaweed Render: Battleby


Mix: 2:4:3, by volume, earth : sand : seaweed.




Seaweed: Dry tangle stems, cut to 10-20mm flakes using a brushwood chopper.

Background: Mudwall with Flax (A.1.7).

Precedent: There is historic and contemporary use of milled seaweed and seaweed leaves in Mediterranean areas.

Method: The render was prepared wet and matured for three days prior to application. The base wall surface was sound and dampened prior to the application of the render, which was thrown on and smoothed to a rough polish by hand, as a single 25mm coat, feathered in at the base. The chopped seaweed did not mix well with the mortar, reduced its workability and did not improve its initial strength.




Current Status: Loss complete, 2002.


Weathering: The render suffered progressive erosion. After two years, the render was heavily eroded and 10% boss. After four years, 50% of the render was lost, 50% was boss with none well adhered. The surviving render was extremely eroded. In the winter of 2002, the wall lost its roof, leading to rapid decay of the background wall and complete loss of the surviving render.


Comments: It is thought that the material did not replicate historic mixes accurately and these may merit further investigation.

Archive ref: 5.2

102


A.5.15 Earth and Ash Render: Battleby

Summer 1999 SE Summer 2002 SE

Mix: 2:2:3:2, by volume, earth: sand : hay : ash.





Ash: Finely sieved, hardwood ash.

Background: Stone with earth mortar (A.3.2).


Method: The render was prepared wet and matured for three days prior to application. The dung mixed well with the mortar, adding plasticity and stickiness. The base wall surface was sound and dampened prior to the application of the render, which was thrown on and smoothed to a rough polish by hand as a single 25mm coat, feathered in at the base. The render was polished with a trowel the following day. The material made a good washable mortar and adhered well to the earth mortar, less well to the stone.



Inspected: July ‘98, Feb. March & Nov. ‘99, April & Nov. 2000, Autumn 2001, Autumn exposed 2002.



Weathering: The render gradually thinned, while maintaining a good bond and in two years 10% had been lost. After five years, 70% of the render was lost, 15% boss and 15% remained well adhered. Erosion was concentrated on the SE and NE elevations, while most remained on the SW elevation.


Comments: The extra fines in the ash may have contributed to durability. Decay patterns related to orientation, with the sun and wind on the SW face having a beneficial drying effect.

Archive ref: 5.16

103


A.5.16 Earth and Oil Render: Fort George


Mix: 2:4:3, by volume, earth : sand : hay, as 500:1, mix : oil.





Oil: Boiled linseed oil.


Precedent: A variant on an historically common form of earth construction in Scotland, as described in TAN 6, 7.

Method: The render was prepared wet and divided into 5 l. portions, to which 10 ml. Of boiled linseed oil was added, and the render left to mature for four days. The base wall surface was sound and dampened prior to the application of the render, which was thrown on and smoothed by hand, as a single 25mm coat, feathered in at the base. The render did not seem to adhere well to the wall. The render was polished with a trowel the following day.






Weathering: The render weathered well. Overall, none of the render was lost, 10% boss and 90% remained well adhered. There was some loss of surface fines exposing the fibres and this was greatest on the NW and least on the NE elevation.

Comments: The render performed well in an exposed situation, maintaining a good key to the vertical, monolithic background material.

Archive ref: 11.3

104


A.5.17 Earth and Tallow Render: Fort George

Summer 1998 SW Spring 2001 SE Autumn 2002 NE

Mix: 2:4:3, by volume, earth : sand : hay, as 500:1, mix : tallow.

Material: Earth: Type: Red Gallowflat clay, silty clay loam.




Tallow: Commercial liquid animal fat.

Background: Claywall (A.1.15).


Method: The render was prepared wet and divided into 5 l. portions, to which 10 ml. Of tallow was added, and the render left to mature for four days. The base wall surface was sound and dampened prior to the application of the render, which was thrown on and smoothed by hand, as a single 25mm coat, feathered in at the base. The render was polished with a trowel the following day.






Weathering: The render kept good adhesion, though there was considerable loss of surface fines, especially at the corners, which left a very fibrous surface, in a manner typical of earth renders at the site. Overall, 5% of the render was lost, none was boss and 95% remained well adhered.

Comments: The render performed well.

Archive ref: 11.7

105


A.5.18 Local Earth Render and Limewash: Fort George



Material: Earth: Type: Kinnoir clay, a grey-green clay.




Background: Turf Block wall (A.3.11).

Finish: Limewash, 5 coats.

Precedent: A variant on an historically common form of earth construction in Scotland, as described in TAN 6, 7.

Method: The render was prepared wet and matured for one day prior to application. The base wall surface was sound and dampened prior to the application of the render, which was thrown on and smoothed by hand in a single 25mm coat. The render had very poor adhesion and had to be worked into the turf to achieve a bond. The render was polished with a trowel the following day. Five coats of limewash were applied over two days, while the render was still damp.






Weathering: The render gradually thinned back from high points rather than eroding in sections. Within two years 30% of the render was lost and 20% boss. After five years, 70% of the render was lost, none boss and 30% remained well adhered. The remaining render was in low points, mostly on the NE elevation and very tight to the background.

Comments: The render performed better than might have been expected from an apparently silt rich earth.

Archive ref: 11.6

106


A.6 Lime and Cement Coatings

A.6.1 Lime Harl: Culzean


Mix: 1:3, by volume, lime : sand.

Finish: Limewash, five coats.

Material: Lime: Kirkby-Stephen lime putty, slightly hydraulic.


Limewash: Prepared from Shap quicklime, used 5 day old.

Background: Stone-faced rammed earth (A.2.7).

Precedent: A common traditional finish.

Method: The base wall surface was sound and dampened with limewash prior to the application of the render. The first coat was trowelled on to approximately 7mm, leaving a rough texture. The top coat was thrown on as a fine harl at approximately 5mm,and finished with limewash while still ‘green’.






Weathering: This was the least weathered render on this site. Overall, none of the render was lost, 5% boss and 95% remained well adhered. There was slight evidence of surface weathering on the SE elevation.

Comments: The stone proved a more compatible background than the earth.

Archive ref: 8.5

107


A.6.2 Lime Harl: Culzean

Summer 1998 SE Autumn 2002 SW Autumn 2002 NW






Background: Earth and Flax Block with Earth Mortar (A.3.7).


Method: The base wall surface was sound and dampened with limewash prior to the application of the render. The first coat was trowelled on to approximately 7mm, leaving a rough texture. The top coat was thrown on as a fine harl at approximately 5mm, and finished with limewash while still ‘green’.






Weathering: Though superficially in good condition, with minor weathering on the SW and NE elevations, the render lost almost all adhesion. Vertical cracks developed at the corners, more pronounced towards the base as the render lost its key, becoming 50% boss within two years. Overall, none of the render was lost, but 90% was boss and only 10% remained well adhered.

Comments: This test demonstrated the failure of lime renders onto earth, even on this sheltered site.

Archive ref: 8.7

108


A.6.3 Lime Harl: Battleby

Summer 1998 SW Summer 2002 SE

Mix: First Coat: 1:3 by volume, lime putty : sand.

Second Coat: 1:4, by volume, quicklime : sand.


Material: Lime Putty: Kirkby-Stephen lime putty, slightly hydraulic.

Quicklime Shap quicklime.





Method: The base wall surface was sound and dampened with limewash prior to the application of the render. The first coat was trowelled on to approximately 7mm, leaving a rough texture. The top coat was thrown on as a fine harl at approximately 5mm,and finished with limewash while still ‘green’.





Exposure: Moderated, prevailing SW.

Weathering: The render weathered badly. Deep vertical cracks developed at the corners and the render lost its key, with 30% loss and 80% becoming boss within two years. Overall, 65% of the render was lost and the remaining 35% was boss and extensively cracked.


Comments: This test demonstrated the incompatibility of lime render on earth materials.

Archive ref: 5.10

109


A.6.4 Lime and Seaweed Harl: Battleby

Summer 1998 NE Winter 2001 SE

Mix: 1:4, by volume, quicklime: sand, in water, + seaweed 7kg.


Quicklime Shap quicklime.


Seaweed: Sourced as dry tangle stems, finely chopped cut to 10-20mm flakes using a brushwood chopper.


Background: Claywall, (A.1.14).

Precedent: A variant of a common traditional finish.

Method: The mix was slaked on site and matured for fifteen days before use. The slaking dissolved some of the seaweed, leaving a rather lumpy mix. The base wall surface was sound and dampened prior to the application of the first coat, which was trowelled on by hand to approximately 7mm, leaving a rough key. The second coat was cast on the following day, to produce a fine harl finish at approximately 5mm and finished with limewash while still ‘green’. The mortar had working qualities similar to the other lime harls.



Inspected: July ‘98, Feb, March & Nov ‘99, April & Nov. 2000, Autumn 2001.

Current Status: Wall collapsed, winter 2001-2.


Weathering: There was a loss of key, with large cracks developing at the corners and the render 50% boss within two years. Overall, in the four years before the collapse of the wall, 40% of the render was lost and the remaining 60% was boss. The harl was only slightly weathered on the surface. This render appeared to perform the best of the comparable lime renders.


Comments: This appeared to perform the best of the lime renders onto earth, perhaps because it is more flexible.

Archive ref: 5.11

110


A.6.5 Lime and Whey Harl: Fort George

Summer 1998 SW Winter 2001 SW

Mix: 2:8:4:1, by volume, quicklime: sand: water: whey.


Material: Lime: Shap quicklime.



Whey: Clearish liquid, a byproduct of cheese-making from a local dairy.

Background: Earth Daub on Ladder (A.4.7).


Method: The mix was slaked on site and matured for nine days before use. The base wall surface was sound and dampened prior to the application of the first coat, which was trowelled on by hand to approximately 7mm, leaving a rough key. The mortar was diluted with water to achieve a throwing consistency for the second coat, which was cast on the following day to approximately 5mm, to produce a fine harl finish. The material made a good, rich, workable mortar.



Inspected: June ’97, June ’98, March & Nov. ‘99, April & Nov. 2000, Autumn 2001.



Weathering: Initial good performance was followed by rapid failure. After two years there was no loss and only 5 % boss, with some surface scaling and hairline cracks on the SW elevation. In the third year the render on the SW elevation fell away in a single sheet. The remaining harl had completely eroded by the end of that year.

Deconstruction: The render had been lost before dismantling.

Comments: The exposure of the site accelerated the rate of decay.

Archive ref: 11.13

111


A.6.6 Lime and Oil Harl: Fort George

Summer 1998 SW Autumn 2002 SE

Mix: 57:229:121:1, by volume, quicklime: sand: water: oil.








Method: The mix was slaked on site and matured for fifteen days before use. The base wall surface was sound and dampened prior to the application of the first coat, which was trowelled on by hand to approximately 7mm, leaving a rough key. The second coat was cast on the following day to approximately 5mm, to produce a fine harl finish. The mortar was very effective to work with and resulted in a good surface.






Weathering: Scaling and cracking developed on all faces, followed by accelerated decay, with 30% loss in two years. The earth background appeared to be friable adjacent to the edge of the harl and there were many clay fines washed over the harl surface. Overall, 50% of the render was lost, 45% was boss and 5% remained well adhered. Loss was concentrated on the SW elevation, with most surviving on the NE.

Comments: This seemed to perform better than the other lime harls on this site.

Archive ref: 11.10

112


A.6.7 Lime Stucco: Culzean




Material: Lime: Shap quicklime, slaked to a powder.



Background: Rammed earth (A.2.1).

Precedent: The recipe was taken from Book of Farm Buildings, Stevens and Burn, 1851.

Method: The lime was slow to slake on site, due to the leanness of the mix, but left for a few days it proved sticky and malleable. The base wall surface was sound and dampened prior to the application of the first coat, which was trowelled on by hand to 7mm, leaving a rough key. The lean mix seemed to suit the porosity of the background and there was virtually not shrinkage in the first coat. The second coat was brushed on as a 5mm slurry, to resemble a fine harl. Five coats of limewash were applied over the following two days while the mortar was still ‘green’.






Weathering: The render weathered badly. In two years, 70 % became boss and deep vertical cracks had developed at the corners getting wider towards the base. Overall, 75% of the render was lost and the remaining 25% was boss. The only surviving render was on the SW elevation.

Comments: This method of preparing lime is not recommended as it is unnecessarily hazardous and proved ineffective.

Archive ref: 8.6

113


A.6.8 Lime Stucco: Battleby


Mix: 1:6:3 by volume, lime : sand : water.







Method: The lime was slow to slake on site, due to the leanness of the mix, but left for a few days it proved sticky and malleable. The base wall surface was sound and dampened prior to the application of the first coat, which was trowelled on by hand to 7mm, leaving a rough key. The lean mix seemed to suit the porosity of the background and there was virtually not shrinkage in the first coat. The second coat was brushed on as a 5mm slurry, to resemble a fine harl. Five coats of limewash were applied over the following two days while the mortar was still ‘green’.






Weathering: The render weathered badly. In two years, 70 % became boss and deep vertical cracks had developed at the corners getting wider towards the base. Overall, 75% of the render was lost and the remaining 25% was boss. The only surviving render was on the SW elevation.


Comments: The decay did not seem to be slowed by the vertical nature of the background.

Archive ref: 5.7

114


A.6.9 Lime Stucco: Fort George









Method: The lime was slaked on site and four days later mixed with the sand. This mixture stiffened before being knocked up, and was applied to the wall eight days later. The base wall surface was sound and dampened prior to the application of the first coat, which was trowelled on by hand, leaving a rough key. The second coat was applied in the same manner the following day, but was floated and limewashed whilst still damp. The mortar seemed lean but was easy to apply.



Inspected: June ’97, June ’98, March & Nov. ‘99, April& Nov. 2000, Autumn 2001, Autumn 2002.



Weathering: The render weathered very badly. In two years, all the render was completely lost.

Deconstruction: The render had been lost before dismantling.

Comments: A dramatically poor result for a lime coating.

Archive ref: 11.4

115


A.6.10 Cement Render: Battleby


Mix: 1:6, by volume, cement : sand.

Material: Cement: Ordinary Portland Cement.


Background: Mudwall with Cement Repairs (A.1.9).

Precedent: Many surviving mudwall buildings have had cement repairs and cement renders applied, which are believed to be inappropriate.

Method: The base wall surface was sound and dampened prior to the application of the render. The render was prepared wet and immediately trowelled onto the wall in a single coat, 5-10mm thick.




Current Status: Fell over, Winter 2002/3.


Weathering: 40% of the render was lost and the remaining 60% became boss within three years.

Deconstruction: The wall failed prior to dismantling.

Comments: The render completely failed and led to the collapse of the wall.

Archive ref: 5.3

116


A.7 Washes and Paints

A.7.1 Dung Paint: Culzean


Mix: Dung, as sourced.

Material: Dung: Fresh cow sharn.

Background: Earth Daub on Ladder, NW panel (A.4.5).

Precedent: This material is known to have been used as a top dressing to clay thatch, and as a top coat to earth plasters, from the seventeenth century.

Method: The base wall surface was sound and dampened prior to the application of the wash, which was painted onto the wall in three coats on a hot day. The material applied well, but cracked badly on drying. Worms appeared on the surface.



Inspected: Aug. & Nov. ‘97, March & Nov. ‘98, March & Nov. ‘99.

Current Status: All eroded, March 2000.


Weathering: Within two years, all the coating on the NE elevation and 60% of the coating on the SW elevation was lost, with the remainder loose from the wall. By the third year all was gone.

Comments: The test did not perform well on a sheltered site.

Archive ref: 8.11

117


A.7.2 Dung Paint: Battleby

Summer 1998 NE Spring 1999 NE Autumn 2000 NE

Mix: Dung, as sourced.

Material: Dung: Fresh cow sharn.

Background: Earth Daub on Ladder, NW panel (A.4.6).

Precedent: This material is known to have been used as a top dressing to clay thatch, and as a top coat to earth plasters, from the seventeenth century.

Method: The base wall surface was sound and dampened prior to the application of the wash, which was painted onto the wall in three coats on a hot day. The material applied well, but cracked badly on drying.



Inspected: July ‘98, Feb, March & Nov ‘99, April & Nov. 2000.

Current Status: Wall dismantled, 2001.


Weathering: The dung weathered rapidly, with 80% lost within two years, the remainder being small fragments in sheltered hollows. By the third year 5% remained, as tiny fragments in sheltered hollows.


Comments: The test did not perform well.

Archive ref: 5.14

118


A.7.3 Oil Paint: Culzean


Mix: Oil, as sourced.

Material: Oil: Boiled linseed oil.

Background: Earth Daub on Ladder, SE panel (A.4.5).

Precedent: This is used as a contemporary finish, particularly to rammed earth. It is not recorded in Scotland as a traditional finish to entirely earthen walls, but is known to have been used as a preliminary coating to earth mortared stone walls, prior to the application of oil paints. It would have been considered a relatively expensive coating.

Method: The base wall surface was sound and not dampened prior to the application of the oil, which was painted onto the wall in three coats on a hot day. The first coat was fully absorbed into the daub and by the end of the third coat the surface was completely impermeable.



Inspected: Aug. ‘97, Nov. ‘98, March & Nov. ’99, April & Nov. 2000, Autumn 2001, Autumn 2002.

Current Status: Still intact, March 2004.


Weathering: The surface remained tight and hard, with minimal weathering.

Comments: This may not replicate performance in buildings. These have more complex patterns of movement of moisture through walls, which could be impeded by the reduction of vapour permeability by the use oil coatings.

Archive ref: 8.11

119


A.7.4 Oil Paint: Battleby

Summer 1998 NE Spring 1999 NE Autumn 2000 NE

Mix: Oil, as sourced.

Material: Oil: Boiled linseed oil.

Background: Earth Daub on Ladder, SE panel (A.4.6).

Precedent: This is used as a contemporary finish, particularly to rammed earth. It is not recorded in Scotland as a traditional finish to entirely earthen walls, but is known to have been used as a preliminary coating to earth mortared stone walls, prior to the application of oil paints. It would have been considered a relatively expensive coating.

Method: The base wall surface was sound and not dampened prior to the application of the oil, which was painted onto the wall in three coats on a hot day. The first coat was fully absorbed into the daub and by the end of the third coat the surface was completely impermeable.



Inspected: July ‘98, Feb, March & Nov ‘99, April & Nov. 2000.

Current Status: Wall dismantled, 2001.


Weathering: The surface remained tight and hard, with minimal weathering for two years. Then, although the oil impregnated surface remained sound as a hard skin, fracturing and delamination occurred in the background daub leaving it extremely friable.


Comments: This may be indicative of performance in buildings, which have more complex patterns of moisture movement through walls that can be impeded by the reduction of vapour permeability by the use oil coatings.

Archive ref: 5.14

120


A.7.5 Limewash: Culzean


Mix: 1:2 by volume, lime : sand.

Material: Lime: Prepared from Shap quicklime, used 5 days old.

Sand: Very fine sand.

Background: Earth Daub on Kebber and Mott, NE side, (A.4.8).


Method: The base wall surface was sound and dampened prior to the application of the wash, which was brushed on in five coats over two days. The wash seemed lean but was easy to apply.



Inspected: Aug. ‘97, Nov. ‘98, March & Nov. ’99, April & Nov. 2000, Autumn 2001, Autumn 2002.

Current Status: Still intact, March 2004.


Weathering: The limewash became extremely cracked and delaminated. Overall, 10% of the wash was lost and the remaining 90% was flaking off extensively on both sides.

Comments: The poor durability is thought to relate to the damp site and abrasion from vegetation.

Archive ref: 8.13

121


A.7.6 Limewash: Fort George


Mix: 1:2 by volume, lime : sand.

Material: Lime: Prepared from Shap quicklime, used 2 weeks old.

Sand: Very fine sand.

Background: Earth Daub on Kebber and Mott, NW side, (A.4.10).


Method: The base wall surface, comprising original material and recent daub repairs, was sound and dampened prior to the application of the first coat, which was brushed on in five coats over two days. The wash worked well, closing shrinkage cracks and the junction between the repairs and original material.



Inspected: Aug. & Nov. ‘97, March & Nov. ‘98, March & Nov. ’99, April & Nov. 2000.

Current Status: Panel eroded, 2000.


Weathering: The limewash eroded at high points in the daub, but only 5% was lost within two years. The background dub had fully eroded in three years.

Comments: This test performed well, in context.

Archive ref: 11.11

122


A.7.7 Limewash with Oil: Fort George


Mix: 250:1, limewash : oil.

Material: Limewash: Prepared from Shap quicklime, used 5 day old.


Background: Brick-faced Rammed Earth, (A.2.9).


Method: The lime was slaked on site and oil added while still warm. It was applied a few days later. The base wall surface was sound and dampened prior to the application of the wash, which was brushed on by hand in five coats.


Wash Applied: May 1997.

Inspected: June ’97, June ’98, March & Nov. ‘99, March & Nov. 2000, Autumn 2001, Autumn 2002.

Current Status: Still partially extant, March 2004.


Weathering: The wash eroded quickly, resulting in a 90% loss in two years and 95% loss in five. The wash eroded off the earth mortar quicker than off the bricks and survived just below the roof.

Comments: The material was considered ineffective.

Archive ref: 11.8

123


A.7.8 Limewash with Whey: Culzean


Mix: 7:35:4, quicklime : water : whey, by volume, equivalent to 10:1, limewash : whey.

Material: Limewash: Prepared from Shap quicklime, used 5 day old.

Whey: Buttermilk, from a local dairy.

Background: Stone with Earth Mortar (A.3.1).


Method: The materials were slaked together on site, with a very slow but hot reaction producing a thick, butter-coloured limewash. The base wall surface was sound and dampened prior to the application of the wash in five coats over two days. Although the colour was attractive, the wash was very frothy and full of bubbles, which made it difficult to apply. The material did not adhere well.




Current Status: Still partially intact, March 2004.


Weathering: The wash thinned and flaked in exposed areas, resulting in a 60% loss in two years and 75% loss in five.

Comments: The material was considered ineffective.

Archive ref: 8.12

124


A.7.9 Limewash with Lanolin: Fort George

Summer 1997 SW Autumn 2002 SE

Mix: 50:250:1, by volume, quicklime : water : lanolin.

Material: Quicklime: Prepared from Shap quicklime, used 5 day old.

Lanolin: Pure lanolin, sourced from a chemist.

Background: Stone with Earth Mortar (A.3.3).


Method: The limewash was slaked on site and matured for seven days. The base wall surface was sound and was saturated with water spray prior to the application of the first coat of wash, and dampened before each subsequent coat. Five coats were applied with a soft brush over two days.


Wash Applied: May 1997.

Inspected: June ’97, June ’98, March & Nov. ’99, March & Nov. 2000, Autumn 2001, Autumn 2002.

Current Status: Still partially intact, March 2004.

Exposure: Exposed, prevailing SW, occasionally severe NE.

Weathering: The limewash gradually weathered in response to exposure. In 1999 the wall partly collapsed and was roofless. Some stones, apparently the more porous ones, retained good limewash cover to the end.

Comments: The material did not seem different in working for the added lanolin, but it proved more durable than the limewash with oil A.7.7 at the same site.

Archive ref: 11.2

125


experimental earth structures, renders and...

Documents