NATIONAL PARK SERVICE Ruins Stabilization in the Southwestern United States Publications in Archeology 10

EROSION FROM CAPILLARY WATER

Capillary water is a term used to designate moisture in walls and fill and is considered as "ordinary water which occurs in small voids so that the surface tension of the water becomes an important factor in determining its behavior" (Plummer and Dore, 1940, p. 59). Use of this term will distinguish it from hygroscopic moisture which surrounds and is closely associated with the individual soil grains and cannot be evaporated by air drying. Gravitational water, on the other hand, is that occurring in sufficiently large amounts to behave according to usual hydraulic laws. Percolation down through fill against a structure might be considered as gravitational water but, for the purpose of handling in ruins stabilization, there is no difference between percolation and capillary water since it rapidly becomes the latter.

This section deals with moisture entering masonry walls either by upward movement of capillary moisture or the outward movement through walls of moisture entrapped in fill areas. Capillary moisture acts either upon the aboriginal soil mortar in the masonry, or upon the softer varieties of sandstone and is one of the major factors in the disintegration of ruins, particularly in excavated sites. It is also one of the most troublesome conditions to control.

The effect of erosion caused by moisture at the base or lower portions of a wall becomes rapidly cumulative. Erosion of narrow but long sections results in severe damage to or loss of higher areas by settling or collapse. The surface disintegration of a masonry wall is caused by the movement of water through it. The damage occurs at the point where the moisture leaves the masonry and comes in contact with the air. Even soft sandstone which is continually damp below ground surface remains in fair condition so long as it is not exposed to alternate cycles of wetting and drying. Where one side of a wall remains damp from contained fill, the damage occurs not on that side but on the opposite side at the point where entrapped moisture comes in contact with the air.

The area of efflorescence and decay may be at some distance above ground level—at the limit of capillary rise from moist soil (fig. 15). It may be toward the upper part of a retained damp fill, or in a protected location where there has been surface absorption from melting snow. For example, the moisture content of a silty-loam fill from within a drained area at Pueblo Bonito was found to be 16 percent in a month when precipitation was less than 1/2 inch (.42). Similar erosion of face stone may be observed in cave sites even though the walls are based on bedrock. Moisture is derived from the drip of melting snow, rainfall, or seepage on the cave floor (fig. 16). Where masonry is constructed of a durable stone, or there is thick bedding of the mortar, this efflorescence first becomes noticeable and more pronounced in the mortar than in the stone.

FIGURE 15. Damage caused by outward movement of moisture from fill behind kiva wall. The most severe disintegration is at the limit of capillary rise, just below the concrete capping. (Clan Kiva, West Ruin, Aztec Ruins National Monument, N. M.)

FIGURE 16. Damage as the result of rising moisture from melting snow. Note that decay is confined to the limit of capillary rise in the mortar column. (Cliff dwelling in Walnut Canyon National Monument, Ariz.)

At open sites in more northern latitudes, most of the capillary water in soils or structures is derived from melting snow, despite the fact that precipitation from rainfall may be heavier during other seasons of the year. There is little surface movement of this moisture; it is absorbed into the soil or wall immediately upon melting. For this reason, it is extremely difficult to control by ditching, tiling, dry-barreling or other drainage measures.

EVALUATION OF METHODS

Methods employed in the past to prevent moisture from reaching and being absorbed by walls have included 1) surface drainage, 2) subsurface drainage by means of tile and gravel backfills, 3) construction of concrete curtain walls to cut off movement of water, and 4) sealing backs or subsurface portions of walls with impervious coatings.

An evaluation has been made of 21 fill-retaining walls which had been repaired by the above methods, individually and in combination. These walls retain from 1/2 to 1-1/2 stories of fill behind them. The evaluation was confined to walls which were in the most hazardous condition, and which were considered to present the greatest problems, i.e., walls which were of soft and poorly cemented sandstones, usually without bedding planes and with little resistance to weathering. Samples of deteriorating stone removed from Pueblo Bonito and oven dried to a constant weight show moisture content of from 4 to 26 percent. Thirty percent is the theoretical limit for sandstones. Only walls where erosion from capillary moisture took effect near the base or center of the wall were considered. Areas that required only capping were excluded.

Of these 21 fill-retaining walls, 13 remained in good to excellent condition after a period of 10 to 15 years. Eight showed moisture absorption in varying degrees, from small damp spots that appeared during the winter to four examples where it has been necessary to replace disintegrated stone. Eleven of the 13 walls in good condition were so located that it was necessary to provide drainage lines through the walls to the exterior. The other two were in graded areas. Seven were treated after repair by the application of a seal coating on the reverse or fill side. Eight of the 11 drains were still operative. Three had become filled with drifted sand.

Walls that remained in the very best condition were those six which were sealed off from the moisture in the retained fill by a coating of impervious material, and where the drainage remained operative. It is this combination of factors—prevention of moisture absorption from the banked fill, and removal of surface moisture on the face side—which gives the best results. Of the remaining seven walls in good condition, one had been seal coated on the reverse side but the drainage plugged, as it was in two other instances. Drainage remained open in two, and the last two of the 13 had required grading only. All of these walls are in good condition, but they do not show the absolute freedom from moisture as do the six where seal coating was combined with drainage maintenance.

Of the remaining eight walls, four with large damp areas and four requiring repair, only two have been provided with surface drains; both became inoperative. None had been seal coated. Thus, the eight poorest examples lacked both drainage and seal coating. Six of these eight walls were not provided with drainage to the outside because of location (deep rooms enclosed on all four sides by two or three rooms of fill). Of these six, four were provided with a measure of subsurface drainage—backfilling against the base of the wall with gravel and then broken rock. This proved to be inefficient. Considerable wind-borne material has drifted over this gravel backfill since it was laid.

A similar evaluation was made of 19 walls (not retaining fill) constructed of and repaired with the same class of soft stone as in the 21 tested which retained fill. Fifteen remained in good condition. Four showed large wet areas where failure of drainage lines had impounded water (figs. 17 and 18).

FIGURE 17. Basal erosion from capillary moisture is particularly damaging in the interiors of excavated sites where drainage is poor. (West Ruin, Aztec Ruins National Monument, Ariz.)

FIGURE 18. The severe undermining and basal erosion of a wall resulting from capillary moisture. If left untreated, such walls soon collapse. (West Ruin, Aztec Ruins National Monument, N. M.)

In the testing and evaluation of methods for the control of erosion by capillary water, experimental walls were constructed using various techniques and mortar materials. The tests were conducted on square areas that could be filled with wet soil and kept damp in accelerated weathering. They have shown the following results:

1. In the treatment of stone, the soft friable specimens laid in the wall were treated with several commercial masonry waterproofings. (These were types for below-ground masonry and are not to be confused with above-ground repellents employing silicones.) Clear types did not prove effective; heavier, opaque types such as "Aquella" painted on all but the exposed surface proved effective in preventing moisture absorption into the facing stone. Such treatment is time-consuming, and the opaque types are difficult to use because the treatment must be withheld from exposed portions of individual stones.

2. A test for moisture resistance was made with concrete mortar containing an emulsified asphalt "hydropel." Care was taken to completely embed the stone (except for the exposed face) in this treated mortar, and to avoid stone-to-stone contact. A similar method was employed in setting a patch in a continuously damp ruin wall. In the test and in the ruin there was no moisture absorption through the mortar into the stone. The treated mortar is more resistant than plain concrete, and there appears to be no movement of moisture along the line of bond between stone and mortar.

3. In stone laid in soil-bitumen mortar, failure became evident first at the base, the area of greatest moisture penetration. The stone became damp and there was progressive disintegration of the surface. Despite the condition of the stone, the soil-bitumen mortar remained hard and dry. It is evident that penetration of moisture was due to lack of bond between the mortar and the stone.

4. In stone set in ordinary concrete mortar there was some moisture penetration; stone at the base became damp and there was some disintegration of the surface. The mortar also became damp; it was evident that moisture had moved through the mortar into the stone.

In summary, the results of field use and accelerated weathering tests indicate that maximum control of percolating moisture from entrapped fill and the rise of capillary moisture is achieved only by a combination of several procedures: 1) sealing of the wall as effectively as possible from the moisture-bearing soil, 2) the maintenance of proper drainage, and 3) the employment of a moisture-resistant mortar. Specially treated mortars were the most resistant. Ordinary concrete mortar permitted the passage of some moisture, and soil-bitumen mortar was unsatisfactory. No tests were made on soil-cement mortars, but work subsequent to the weathering test indicates that thin mortars of soil-cement would be no more resistant to the passage of moisture than was the soil-bitumen, and that such mortars also would prove unsatisfactory for this specialized application.

CONTROL OF CAPILLARY MOISTURE

The use either of a concrete mortar whose water resistant qualities have been increased by the addition of an emulsified asphalt similar to "hydropel," or of special waterproof cements, is recommended (Plastic, Medusa, or equal). In either case the manufacturer's recommendations should be followed. In the case of "hydropel" 1-1/2 gallons per sack of cement are used.

Waterproofing

For a workable mixture, the water or liquid-cement ratio should be kept at not more than six gallons per sack of cement. This liquid is to include the "hydropel" or other emulsified asphalt. In all probability, the liquid-cement ratio can be reduced below the 6-gallon limit due to increased workability obtained with the asphalt admixture. The asphalt is added at the time of mixing.

There is a great variation in type and gradation of sands. It is advisable therefore to make trial mixes to determine accurately the sand-cement ratio requirements. In many areas, mortar sand, as opposed to concrete sand, can be obtained from commercial sources. The ratio will probably range from 1 to 2-1/2 for the more finely graded. Sand should be clean and free from inclusions of dirt or organic impurities. Mortar should be thoroughly mixed. Thorough mixing improves the plasticity and workability; less mixing water is required to obtain proper working consistency when the mixing time is increased.

+Placement

In laying out stabilization for a wall deteriorating from moisture erosion, consideration must be given to any proposed changes in the fill level (fig. 19). Where the eroded wall also retains fill, the treated mortar should extend somewhat higher than the eroded area to insure against possible increases in the capillary rise. Thus, it is better to include too much rather than too little mortar in the area stabilized. While this might violate the precept of keeping the replacement of prehistoric or original material at a minimum, it is desirable that the replacement be made all at one time rather than over a period of years, as moisture works its way around too small a patch.

FIGURE 19. Field control of moisture erosion. Wall depicted in section at left before treatment; following stabilization at right.

If breakdown of the lower wall areas has proceeded over any period of time, possible faults in the surface above these areas will also have to be taken into consideration. Cracks, bulged areas, and separation of courses due to settling are apt to occur. Furthermore, where erosion has cut into the wall some distance it may be possible to reface the surface using thin stone or other original material without removing the deteriorated area. This is not a profitable undertaking. It will not permit the use of a heavy mortar bed, and little support is given the upper levels of the structure since a thin patch cannot be tied securely to the original work.

Defective masonry must be removed in narrow vertical sections. Some experience is required to determine the amount that can safely be taken out at one time. This is dependent upon the general type of construction and condition of the structure. It is seldom advisable to use jacks and blocking to remove large sections at one time. Removal of stone is best accomplished by cutting the mortar with a cape chisel or other sharp tool, and prying rather than attempting to break or knock material out with a hammer. Since vibrations are easily transmitted, bond in adjacent sections is apt to be broken.

Before setting stone, dampen it thoroughly. This helps remove the film of dirt which interferes with the bond and, particularly in hot weather, the absorbed moisture aids in setting the mortar. Do not smooth the mortar bed unduly but leave it slightly furrowed so that it will meet irregularities in the stone. The stone can be tapped into place with the handle of a mason's hammer.

On finely surfaced walls, the joints between facing stone are often filled with small spalls. Some masons have a tendency to drive the spalls into the joint after the larger stone has been set. Do not permit this. It will raise the larger stone out of the mortar bed and spoil the bond. If spalls of any size are used, they must be set in the mortar first and the upper stone set in place over them. This procedure is time consuming, since it is difficult to clean the extruded mortar from small spalls.

Treatment of Oblique Surfaces

The most frequent result of moisture erosion in the upper levels of heavy walls retaining fill is loss of the facing on the exposed side. This results in an irregular and oblique surface. There are two alternatives in stabilizing these areas: 1) either the missing face of the wall can be rebuilt and sealed off from the damp fill, or 2) the broken slope can be taken down and relaid in a moisture-resistant mortar, and soil-cement pointed for appearance. Rebuilding the missing face is more time consuming and expensive, but it is sometimes advisable where the wall is in very bad condition or the additional strength is needed to buttress adjoining walls.

Relaying the oblique surface in a moisture-resistant cement mortar is more effective from the point of general appearance (fig. 20). The only drawback to this method is that large areas of the waterproofed mortar will be exposed. It is difficult, if not impossible, to color cement mortars containing emulsified asphalt. Special waterproof cements, however, can be colored.

FIGURE 20. The application of waterproofed cement mortar to an oblique wall.

Where the extent of exposed mortar is not too great, the appearance can be greatly improved by holding the concrete back 1/2 to 3/4ths inch from the final surface, and pointing to this depth with stiff soil mortar. The appearance of the soil pointing will be much more satisfactory than that which can be obtained with concrete mortars. The soil pointing will require replacement after a few years, but it is an easy job which can be accomplished by unskilled labor and with little supervision. This technique was first used extensively by Earl Morris at Mesa Verde during 1934, though he undoubtedly developed it earlier (Hamilton and Kittredge, 1935, pp. 1-14). It is a particularly valuable method of retaining appearance while employing durable and moisture-resistant materials in the body of the structure. It can be used in many applications where soil, soft lime, or other distinctive mortars are widely exposed.

In resetting stone on an oblique face it is best to lay the work out in vertical strips rather than in horizontal sections. This aids in maintaining the slope, and it is easier to match the pattern of the adjacent surface. The same method as that used in setting repairs in the face is employed—each exposed face stone is well bedded in the treated concrete mortar. This will mean that the entire vertical section must be taken down and relaid, rather than merely cleaning out the soil mortar and grouting the interstices with mortar. Depth of mortar will vary with the height of the wall and the size of stone used, but it is suggested that a thickness of 6 inches be taken as a minimum.

The point of greatest weakness in such an installation is at the bottom where the slope joins the remaining vertical face (if any) of the wall. At this point the vertical face must be taken down and the upper three to four courses reset in the water-resistant mortar. Extra care must be taken at this point to prevent accumulated water from the slope entering the lower wall.

Seal Coats

While the employment of water-resistant mortars combined with surface drainage will aid materially in the reduction of moisture-absorption in many situations, this procedure should always be carried out in connection with sealing off the reverse side of the wall wherever possible. To be effective, the sealed area must extend well below the level of the exposed and eroded area. Difficulties encountered include work on irregular surfaces, breaks or holes in the surface to be sealed, and adjoining or abutting structures.

Extensive sealing-off of wall areas is more often required in excavated sites than in unexcavated areas due to the more porous nature of the fill in the former. The application of an impervious seal coat for the reverse covered side of a wall will spoil it for future exhibition. This is seldom, if ever, a consideration in excavated sites. Extensive areas were sealed off in the mission structures of San Isidro and San Buenaventura at Gran Quivira. In both instances no future excavations were contemplated.

The purpose of sealing off portions of a wall is to prevent the entrance of moisture into the wall from enclosed fill or from higher levels outside the site where the structure is partially subterranean. The procedure is much the same as that employed in making basements waterproof. The area against the wall must be excavated, and any abutting walls or portions of structures removed.

Prehistoric walls with soil mortar do not present a sufficiently sound surface on which to work. When the wall is cleaned of loose and friable material, it should be plastered with a sound cement plaster, one or two coats as required. Care must be taken to cure this properly so that no cracks develop. After the plaster has cured, it is treated with applications of hot asphalt or an approved asphalt foundation paint.

When the area against a sealed surface is backfilled, some effort must be made to compact the fill near original density, and to provide surface drainage. In a recent project (1970) at Tumacacori National Monument where the walls of the Mission Church were successfully damp-proofed, a sealant and a liner were employed as follows: 1) exterior foundations were trenched; 2) the rubble stone masonry foundation walls were grouted and pargeted; 3) after curing, the pargeted surfaces were waterproofed by a trowelled-on membrane at a 60 mil thickness, using Thiokol 411-H Elastometric membrane. This is a trade name for a two-component, chemically-curing liquid polysulfide-based waterproofing compound, specifically formulated for manual mixing and application by troweling, brushing, or squeegee. The compound is easily applied without heating, and cures at an ambient temperature of about 40° F. to form a highly flexible, seamless, rubbery membrane impervious to water. While the Thiokol was still damp, a 20 mil. Polyvinyl Chloride (PVC) liner was laid across the bottom of the trench and up the Thiokol-covered wall. The PVC was used in the 24-foot width. Two feet of PVC rested on the bottom of the trench, 4 feet were adhered against the wall, and the remaining 18 feet were rolled up on temporary supports while the trench was backfilled. The PVC was then sealed to the plastered foundation with a 3-inch-wide band of PVC-to-concrete adhesive applied above the Thiokol. All overlapping sections of PVC were sealed to each other with PVC-to-PVC adhesive. A 20-foot area from the wall was then graded with a front end loader to provide drainage away from the wall. The remaining 18 feet of PVC was unrolled and spread over the graded area and backfilled with local soil. The PVC is buried about 6 inches below grade at the wall, and 18 inches where the graded area joins the ditch leading surface water away from the Mission.

At Fort Bowie, the clayey, impervious soil and standing water required other procedures engineered to local conditions to combat dampness and capillarity in unprotected adobe wall remnants. These systems have functioned very well since installation in 1965. Briefly noted, the foundations were pargeted and treated with asphalt emulsion. Before backfilling the trenches, a drainage system was installed by laying perforated asphalt pipe covered with gravel. The pipes were pitched to inconspicuous outlets beyond the structures. In another instance, a long interior room was drained by laying a polyethylene liner several inches below grade over the entire "floor," and a perforated pipe laid along the center of the long axis of the room after it had been graded or pitched slightly to the center and to one end. The pipe extends beneath one wall and exits at an inconspicuous outlet. The floor-liner method of halting capillary action was used successfully in open sites at Mesa Verde and other areas where bedrock is close to the surface, and where other related drainage measures are not feasible.

RELATED DRAINAGE

Discussed here are drainage problems relating to removal of surface water from areas treated as above, and also the removal of surface water from all areas within a site. Drainage over large, relatively level areas of a site is seldom a problem except, perhaps, in large plazas. In most Pueblo sites, surface water within rooms is absorbed into the soil and presents little difficulty. Problems are encountered where adjacent rooms or small areas are at widely varying levels, and where there is either surface washing from one level to another, or moisture penetration through a wall.

Dry-Barreling

The use of dry-barrels is suggested for enclosed areas where drainage lines to an exterior point would be difficult to construct and maintain. The purpose of a dry-barrel is to quickly drain water falling into a room to a level below that at which it can cause damage by percolation into adjoining areas.

A dry-barrel is usually constructed at the center of a room or confined space (fig. 21). A circular hole 3 to 4 feet in diameter is dug to a depth which is at least 3 feet below the lowest level in adjoining rooms. The hole is then backfilled with random stone or gravel with fairly well graded material at the top to prevent the entrance of large amounts of silt or fine material. The surrounding surface is graded to the sides of the dry-barrel.

FIGURE 21. Construction of a dry-barrel (in section). Note that it has been carried well below the adjacent floor levels.

Surface Drains

It has become apparent after years of experience that surface drainage presents some unusual problems in maintenance that are not encountered in ordinary structures. Surface drainage systems are often employed in sites or portions of sites that do not receive regular maintenance. Change of personnel in areas results in total lack of information on the nature and location of drains. Drainage lines often become rodent burrows, and interpretive personnel will frequently cover the open ends of drains to hide them, a practice that usually results in complete stoppage. With this in mind, very careful consideration must be given to laying out any drainage system. Small individual drains running to open areas are far better than large systems draining several rooms and areas together. The larger the system, or the more rooms that are drained together, the greater is the chance of damage when one section of the system becomes inoperative. Tile drains with vertical sections and elbows connected to horizontal runs are to be avoided. Soil in most sites is loose; there is always an abundance of blowing sand or cinder, and such drains become easily clogged.

The preferred drain is one with a small concrete settling box covered with a grate, the tile line extending from the side of the settling box (fig. 22). Tile lines through walls are preferred to "weep holes," in spite of the fact that their appearance is not as satisfactory. Weep holes tend to admit water into the masonry unless very well constructed. While the necessities of appearance would seem to dictate concealed or very unobtrusive lines emptying from a small area, annual inspection of numerous sites has demonstrated that such lines are simply not maintained, and that an obtrusive line which empties surface water from an area provides far better protection than a concealed line which has become inoperative through accumulations of sand or cinders.

FIGURE 22. A settling type inlet for drain lines.

REPAIR OF WALL BREAKS

The discussion here will deal with the stabilization of masonry walls containing ordinary structural failures—loose areas, holes entirely through the wall, or holes which extend only part way through as are often found in heavy structures of veneered construction.

At least 50 percent of the repair in the Sout west will be done with stone of some type. That used in setting a patch should match stone in the original wall as closely as possible. There is little chance that the requirement that a patch be discernible upon close inspection will not be met, since it is practically impossible to construct one that cannot be located by expert inspection.

Stone from the ground in the vicinity of a site, as opposed to that saved from an excavation, would ordinarily seem to meet the requirements. However, stone which has been exposed on the ground for centuries is sometimes not the best. If it is porous, e.g., a soft sandstone, it is no doubt partially disintegrated. Where walls must be repaired with faced stone of the softer varieties, it is usually a major problem to locate useable pieces. In some cases, replacement of pecked or dimpled stone will require an exhorbitant amount of time to properly shape and surface the individual stones (fig. 23). Porous stone often absorbs salts from the soil in which it lies. After the stone is set in a wall and has become damp and then dried a few times, these salts will leach out and mark the patch as a white area.

FIGURE 23. An example of soft, pecked and dimpled stone in random areas over wall surface. Replacement of the stone requires extreme care if the character of the wall is not to be changed. (Penasco Blanco Ruin, Chaco Canyon National Monument, N. M.)

Small spalls are also difficult to find and are rarely, if ever, saved from an excavation. Time can be saved if suitable rock and spalls are quarried. Useable quarries can often be located close to the work, and a single truckload of spalls will save much time and effort searching for individual pieces.

Patching Simple Wall Breaks

Determining the cause of a break in the first place will be of some help in planning repairs and to prevent a recurrence. This is particularly true when the forces that result in the break are still active and must be counteracted before the patch will be effective.

Two of the most common types of simple breaks are those caused by breakdown or removal of some wood member, rotting, excessive weight, or vandalism; and disintegration of soil mortar where it forms a major part of the wall.

Breaks often become enlarged over the years so that it is difficult to determine just how they start. This poses a problem if plans are made to fill the hole completely with repair masonry. One should always ask whether such holes should be filled in solid, or whether there was once a door way, small opening, or a beam socket.

Doorways were often partially filled during the time of occupancy. Close examination of the masonry remaining in the bottom of the break will sometimes provide evidence of an opening. Removing a few stones may reveal the side or jamb of a door. Changes in masonry style are another indication of filled openings. Within a room, examination of an opposite wall will sometimes serve to indicate the position of beam sockets.

In placing a patch, care must be taken to clean the masonry surface surrounding the break (fig. 24). All loose stone and unsound mortar should be removed. This might necessitate excavating around it to a depth of several feet. In large sites, if the bottom of the break cannot be reached for one reason or another, the base of the patch should be laid in reinforced concrete to form as good a footing as possible. There are only two essential requirements, i.e., the patch should be laid in the same type of stone and in the same pattern as the adjoining wall, and it must be tight. Excessive shrinkage is apt to cause the most trouble in soil-mortar patches. If this is likely to be a problem, it is best to use cement or masonry mortar. The stone placed in the patch must be clean and damp. A film of dirt prevents bonding with any kind of mortar. The masonry adjacent to the patch must also be clean and damp for the same reason. Whisk brooms can be used for dampening the original masonry; the fine spray of a back pack pump gives excellent results.

FIGURE 24. Wall break in a partially excavated site. A patch must be designed to prevent further flow of water and fill at the top, and some excavation will be required to base patch on sound masonry. (East Ruin, Aztec Ruins National Monument, N.M.)

In a great number of cases there will be as much mortar exposed on the wall face as there is stone. In anything but soil mortar, slick trowel marks will show as long as the mortar lasts. The best way to obliterate tool marks is with a whisk broom or by hand, before the mortar has begun to set; if left they will certainly spoil the appearance of the job.

In cases where colored soil-cement or concrete mortars are used, the matter of troweling is also important. Troweling of colored mortars results in a thin film of the color on the surface and also produces a reflective surface. It must be remembered, particularly with colored mortars, to rake or scratch thin mortar joints and brush large areas after the mortar has dried, but before it has set hard. The most effective method of retaining good appearance is to point the surface with soil mortar.

Doorways and Large Openings

Breaks in doorways and other large openings usually take the form of an inverted triangle, the apex of which is at the bottom and the greatest breadth higher in the wall (fig. 25). In repairing such openings, one is tempted to replace only the missing stone. But unless the material is of large size and the wall exceptionally clean, this is apt to be a poor investment in labor and materials. The bond between stone and mortar in prehistoric walls is weak. At best, the replaced stone will not bond well with the original work, due largely to the amount of soil mortar exposed in the interior of the wall. This results in a patch of one solid unit, an inverted triangle, and in too little bond with the remainder of the wall. When such patches fail they do not break up, but separate in one piece from the structure. It is therefore better to remove sufficient original material to reset it in a patch, the base of which is somewhat larger than the base of the original break, and which will be stable in itself and without dependence upon the bond with the original work.

FIGURE 25. Repair at the position of the arrows will be based on a very small area in comparison with the total size of break. Every effort should be made to tie such patches securely to the existing wall. (Pueblo del Arroyo Ruin, Room 14, north wall, Chaco Canyon National Monument, N. M.)

Slumped and Filled Areas

An effective method of handling small slumped or loose areas was developed at Wupatki National Monument by A. E. Buchenberg. Here small vertical sections of the slumped areas were cleaned of soil mortar; cement or soil-cement mortar was packed between the stone. This makes an authentic appearing repair, provided the color and texture of the replaced mortar is not too conspicuous. To be effective, the grout must extend from 1/4th to 1/3rd of the way through the wall. This is particularly true with soil-cement mortars. Grouting in an area of slumped stone keeps a ruin looking like a ruin instead of an unfinished building, and it is far cheaper in labor and material than the alternative of replacing the entire section.

Vertical Surfaces

Vertical wall areas containing an excess of soil mortar at the surface often become badly weathered, while the interior of the wall remains sound. Large wall sections can be saved and future repairs extensively reduced by grouting between the exposed surface stone (fig. 26). The same cautions must be exercised here as noted above. Loose soil mortar must be removed and sufficient depth obtained so that there will be some bond between stone and mortar.

FIGURE 26. Deep grouting with soil-cement on the badly weathered face of this cobblestone wall will not change its appearance, but will halt the rapid rate of erosion. (East Ruin, Aztec Ruins National Monument, N. M.)

Respalling

Where masonry walls containing a large proportion of spalls or chinking on the surface are subject to extreme weathering in exposed locations, surface repair may become a time-consuming project. As the spalling is lost to weathering, the soil mortar between individual stones is then exposed. It weathers far more rapidly than did the spalling, and decay of the wall is accelerated. Loose and softened soil mortar must be scraped from between the individual stones, the space then partially filled with cement mortar and the spalls tapped into place in the mortar. Large walls require considerable equipment and an experienced crew (fig. 27).

FIGURE 27. Scaffolding in place at completion of respalling and grouting large wall surfaces. (Wukoki Ruin, Wupatki National Monument, Ariz.)

Break Where Weight is a Factor

Considered here are partially displaced or shattered areas which still furnish some measure of support, but which must be removed in order to effect repairs. In general, the same precautions pertaining to other stabilization apply with added force: repair work must be laid in a high strength concrete or masonry cement mortar and based on sound foundation; the edges of the adjoining original work must be clean, sound and dampened; and all replaced stone must be firmly seated in the mortar.

Temporary Support

A fairly common situation encountered involves heavy walls (figs. 28, 29, and 30) in which one side has been broken away due to excess weight or shifting of the upper structure. The problem involves support of the sound part of the structure while removing and replacing the damaged area. This can usually be accomplished with the use of jacks and timber blocking. It might involve cutting a narrow vertical section in the displaced portion in which to place the supports. Since construction and soundness of walls varies from room to room in a single site, and no two situations are identical, only general suggestions can be given for temporary support of walls during stabilization as follows:

FIGURE 28. The temporary support of a heavy wall with timber. The arch at lower right was constructed in the 1920's as a repair measure. The first step in stabilization involved the placing of a beam through a hole adjacent to the arch. Not of aboriginal origin, the arch was subsequently removed. (Room 1, Chettro Kettle Ruin, Chaco Canyon National Monument, N. M.)

FIGURE 29. A second beam resting on jacks is placed through the wall depicted in figure 28. Arrows indicate extent of steel plates above the beam. The area of settling was reduced with upward pressure.

FIGURE 30. Jacks and beams provide support of wall depicted in figures 28 and 29 while arch section is removed. The break is filled with repair masonry.

1. Only screw-type jacks should be employed. Do not try to use ratchet jacks.

2. Solid foundations must be provided. Never place a jack on top of an upright timber; use crib-work to reach the desired height.

3. Neither the base nor the head of the jack should bear directly against the masonry. Both must be separated from it by means of steel plates (or wood blocking in temporary installations) and cement grout.

The use of plates greatly increases the bearing surface and permits support of more than one stone. A layer of cement grout at proportions of 1-to-1 will take up inequalities in the stone or between several stones and provide an even distribution of force. Due to the irregular nature of building stone, it is usually impossible to seat a jack squarely against it anyway. In some cases, especially in small areas, it will be impossible to support breaks by extending beams through the wall, and it will be necessary to place jacks within the break itself (fig. 31). In either veneered or single-width walls, examination will indicate the approximate centerline of masonry to be supported. The jack must be placed as nearly as possible in this line. A variation toward the center of the wall will tip the above masonry downward and out when force is applied. Conversely, if the jack is too far out, pressure will act toward the center of the wall and tend to kick the jack out of place.

FIGURE 31. Placement of jack in line with shattered facing near the bottom of a 14-foot wall. The jack provided support while the broken area at the right was being replaced. (Chettro Kettle Ruin, Chaco Canyon National Monument, N. M.)

Once the jack is placed with plates and grout, it can be run up with light pressure until the plates are evenly forced against the masonry and some grout is extruded. With the grout conforming to the shape of the supported masonry, allow sufficient time (at least 24 hours) for it to gain an initial set before more pressure is applied. Always take your time when setting jacks. If more than one is to be used, set them on succeeding days. While temporary jacks are in place, they should be tested frequently. The grout will shrink slightly as it sets; if wood blocking is used it will be somewhat compressed. Cribwork resting on soil will settle. The jacks must be taken up to compensate for all this shrinkage.

Permanent Support

Due to limitations of space or the nature of the wall break, it will sometimes be more satisfactory to install jacks permanently. This is also a quick method of providing support when a section of masonry has been taken down. The requirements here are much the same as for temporary installations, except that wood blocking cannot be used. Once the jack has been set for permanent installations, masonry can be built around it and up to the level of the screw. Before placing a full load on the jack, wait until both the grout and masonry have set.

The amount of pressure one should exert with jacks in both permanent and temporary installations is somewhat of a problem. Masonry weighs about 120 pounds per cubic foot. A column of masonry 12 feet high, 1.5 feet thick, and 2 feet wide will weigh a little over 2 tons. Since this column has some support of its own, it is doubtful that more than half the weight, or 1 ton, will come completely on the jack. A jack with a screw 1 inch in diameter has a capacity of 5 tons; one with a screw 2 inches in diameter has a 20-ton capacity. This might give some indication of how much weight is being taken up. Probably the best procedure is to take the jack up a slight amount each day and carefully examine the wall above for signs of movement.

Somewhat the same procedure as outlined above for small areas can be employed in supporting large sections of wall where the entire width must be removed. This is also a useful method of support where large horizontal timbers are to be replaced with cast members (fig. 32). In supporting large and heavy sections, it is best to run large timbers or steel beams entirely through the wall, supporting the ends with jacks. All precautions listed above for other temporary support should be followed in regard to placement, grouting between the supports and the masonry, cribwork bases for the jacks and frequent compensation for settlement.

FIGURE 32. A temporary support. Six steel beams buttress the upper half of the wall while a decayed horizontal timber is replaced. The timber is partly removed here. With care, steel beams can be run through pole sockets or other natural openings without damage to the wall. (Chettro Kettle Ruin, Chaco Canyon National Monument, N. M.)

Working conditions are often made safer by external bracing against bulged or shattered areas. Jacks with timber set in sleeves are much more satisfactory than timbers alone, as the use of jacks allows some force to be-applied to the surface. Timbers alone will not take up much stress until there has been some movement of the structure against them. To be effective, pressure must be applied at the point of greatest movement or distortion.

REALIGNMENT

Realignment of distorted, out-of-plumb walls is seldom practicable unless the wall is in otherwise good condition (figs. 33, 34, and 35). If the wall has to be realigned and is in poor condition, it must first be strongly repaired with a soil mortar. Do not repair the wall with cement or stabilized soil mortars. Considerable time, care, and equipment are required to realign a wall of any size. The general requirements are, first, that the wall must be completely formed on both sides and, secondly, that it be thoroughly damp if it is not to be cracked further or distorted in some other directions. The steps taken should include:

FIGURE 33. Beginning realignment. Cables have been placed through the base of the wall to be tied to the jacks. Note bolt, lower right, for holding frame against wall. (Pueblo Bonito, Chaco Canyon National Monument, N. M.)

FIGURE 34. Realignment of wall depicted in figure 33. The framework is in place on both sides and is tied together. Frame lined with asphalt paper, space between frame and wall packed with sand. Note that the five screw jacks in sleeves bear against a 4 x 4 timber to distribute pressure.

FIGURE 35. The wall depicted in figures 33 and 34 after pressure had been applied for three weeks by means of jacks in sleeves.

1. Before any work is done, run small holes through the base of the wall to tie the two sides of the form together with wire or thin bolts.

2. Construct heavy forms of 2-inch lumber large enough to completely cover the area to be moved. Line the inside of these forms with asphalt building paper.

3. Set up the forms; leave a space of between 2 to 3 inches between each form and the surface of the wall. Tie the forms both top and bottom.

4. Pack this space with wet sand. Be sure there are no voids. Cover the top of the wall also. Begin to keep this sand continually wet. This requires time and patience. Do not flood the wall in order to hurry the process, as this will wash out the soil mortar instead of dampening it. Allow from 3 weeks to a month to dampen a wall 2 feet thick.

5. While the wall is being dampened, set dead-men or other supports against which the jacks can be placed. Timbers in sleeves are necessary to keep the jacks in alignment. Mine jacks may also be effective.

6. After the wall is damp, apply pressure with the jacks gradually. Mark the top of the form and set stakes or other points so that movement in the wall can be measured. It will require from 2 to 3 weeks to straighten a wall. Do not try to hurry it. There will be some compression in the forms and timbers, and settlement of the dead-men. Additional blocking behind the jacks will be required.

7. After the wall is straight, allow it to remain in the forms for another week or two to start drying while it is still supported.

REPAIRING WALL FEATURES

Crushed, sagging, or rotted wood members will often be found in place over open or filled doorways, windows, or other openings. Little can be done for low and narrow walls except to remove the timbers and replace them, either with wood or cast members. Rebuild the wall where necessary. Where wood is the replacing material, some reinforcing effort must be made. In short spans, steel plates may be set above the lintels. In longer spans the wall above the wood must be set in reinforced concrete mortar. In any case, the steps are the same (fig. 36):

FIGURE 36. The placement of reinforcing in concrete mortar over pole lintels.

1. Masonry above the lintel must be removed for the entire width of the opening, plus at least 6 additional inches beyond each end of the new lintel. In thick walls one side may be completed first and the other side after this has set.

2. Place the new lintel above the opening after the sides have been repaired; build up the masonry at the ends until it is slightly above the level of the new lintel.

3. If plates are used, lay them over the wood but make certain the ends rest on the masonry at the end and not on the wood. If a reinforced section is used, run a thin layer of concrete mortar over the wood. Embed reinforcing steel in this, likewise making sure the steel and concrete extend beyond the ends of the wood. Complete filling the space above the lintel with masonry set in concrete mortar. The amount of reinforcing and height of the concrete band will be determined by the span of the opening.

Support With Wood Members in Place

If cracked or decayed timbers are found above an opening which has been partially or completely filled in prehistoric times, it is best to leave the outside poles in place and base the repair masonry on the fill, carrying it up through the center of the wall. The exterior poles or lintels are left for appearance only. The steps in stabilizing such an area are:

1. Repair the prehistoric fill in the doorway up to the level of the wood which is to be left in place. Make a small hole at one side of the opening above the outside timber deep enough to expose the interior lintels.

2. Cut out sections of the interior lintels. The length to be removed will depend on the span of the opening and the condition of the wall above. One-third to one-half of the total length may be removed at one time. In large openings, or where the wall is particularly unstable, one-fifth or one-fourth is sufficient. Lintels are most easily removed by a combination of drilling a row of holes across them and sawing between these holes, rather than chopping or chiseling them out. Pounding on them is apt to bring down most of the wall in fragile areas.

3. With one-third or so of the lintel out, build up a vertical section of repair masonry from the patched fill below to the original wall above.

4. After this has set, repeat the process of cutting out a section of the interior lintels and bring the repair up behind the outside lintel.

Beam and Pole Sockets

A great deal of prehistoric construction seems to have followed the following course: (1) the walls were raised to ceiling height and the vigas laid; (2) these were covered with small poles or savinos and on top went the matting, split cedar, bark, or whatever was handy; (3) this in turn was covered with soil for the next floor; and (4) after this ceiling and floor combination was laid, the walls were raised for the next story.

This sequence, with the ceiling and floor finished before another story was added, often resulted in 1) the upper story set back; 2) the upper story overhanging the first, or a combination of the two, i.e., set back at one end of the room and overhanging at the other; and 3) a weakened strip where the ceiling material and dirt floor extended into the wall. Quite often there is no continuous contact of the masonry vertically from one room to that above—rather there is a strip of soil that was the floor, or partly rotted vegetal material that was the ceiling, separating the two.

In repairing these strips, whatever remains of the ceiling material will have to be removed if it has not entirely rotted out (as is most likely). However, the location of the pole or savino sockets should always be preserved. The location of the sockets can usually be found while cleaning out the back part of the strip, or at least enough of them to indicate the location and spacing of the remainder.

The most satisfactory method of repairing pole or savino sockets is to cut several stub poles that are the same size as the originals. Oil them with a light motor oil. After the location of each socket is determined, set a stub in place of the original and make the patch around the stub. The stub need not be set in the wall the entire depth of the original, and the space behind the stub should be well filled with mortar. When the patch is completed and the mortar set, the stub can be removed. The hole is left in the exact shape of the stub and the stabilized mortar will hold the stone in place at the top of the hole.

The same method can be used in repairing large beam sockets, since they were seldom if ever finished off with a slab on the top, or made so that they would stand without the support of the viga. In rebuilding big viga sockets around a stub viga, it is usually best to use concrete and leave the stub in place some 24 hours to allow the concrete to gain an initial set.

Cast Concrete Members

Wood is the least durable material used in the repair of ruins. It may be treated with preservatives and protected from weight in the wall above by integral members which carry the load. However, large timbers are difficult to treat thoroughly unless facilities are present for complete immersion. Furthermore, the construction of integral members in small horizontal sections is time consuming. Integral members made in conjunction with the replacement of wood beams and lintels cause some disturbance to the adjoining masonry. The use of cast lintels and poles eliminates many of the above difficulties. They permit the replacement of wood with little disturbance; they provide a strong, rigid and very durable member. Cast lintels take the same surface and texture as the wood from which cast. With a little care they may be stained to faithfully reproduce the original (fig. 37).

FIGURE 37. The cast concrete lintels have the same appearance as the originals from which they were cast. (Chettro Kettle Ruin, Chaco Canyon National Monument, N. M.)

Excellent instructions for making casts may be found in Chapter VII of the National Park Service publication entitled Field Manual for Museums. The data given below assume some familiarity with those instructions and with casting in general. They are given only as supplemental information, and apply only to the casting of concrete structural members in a three-part mold.

When several pole lintels are to be cast it is preferable to make a lumber form to hold the mold as long as the longest piece to be cast, and some 4 to 6 inches wider and deeper than the diameter of the largest piece. This form will hold the plaster mold around the original wood lintel, and will also hold this mold while the concrete piece is being cast. The form should either be hinged so that the sides swing away without interfering with the mold, or they may be fastened together with light bolts and wing nuts for ease of assembly.

In using plaster molds the piece to be cast must be dry and clean. Deep cracks and holes should be filled to within 1/8th inch of the surface with plasteline or similar material. The piece should then be well oiled with a light motor oil or one designed for use on concrete forms.

Close the lumber form and place 1 or 2 inches of fine damp sand in the bottom. Place the wood lintel on this sand and add more sand until the upper third of the pole is exposed. At this point provision must be made to key the parts of the mold together so they can be replaced in exact alignment after the pole is removed. Rounded depressions may be made in the sand and the plaster flowing into them will act as keys for the adjoining side of the mold. A more accurate method involves placing a bolt toward each end of the cast, thread down in the sand, and with the head protruding enough so that it will be set firmly in the plaster. This permits the parts of the mold to be bolted together and prevents movement in any direction. The bolts must be long enough to extend entirely through the sand.

Most field men are familiar with mixing plaster of paris and detailed instructions may be found in Chapter VII of the Field Manual for Museums. Enough plaster should be made for the first pour to cover the exposed part of the lintel and fill the wooden form to the top. Care should be taken not to entrap air bubbles in the plaster. When the pour is made, smooth off the top as the mold will be turned over and this part becomes the bottom on which the remainder of the cast will rest. If the mold is to be a large one such as for casting timbers or vigas, it is best to reinforce with strips of burlap dipped in plaster, with small dry branches, pieces of discarded reinforcing steel, etc.

After the plaster has set, pull back the sides of the form. Turn the form over gently and remove the sand. One-third of the pole will be set in plaster with the other two-thirds exposed. Shellac the edges of the plaster where the keys have been made. Oil this edge and the exposed part of the pole. Place the part just made with the pole attached in the bottom of the form (pole up) and reassemble the sides and ends. Place a strip of 2-inch lumber lengthwise on the center of the pole to separate the two remaining parts of the mold. Fill any inequalities between the strip and the pole with plasteline.

The second pour of plaster is made on both sides of the longitudinal strip, level with the top of the form, and completes the three parts of the mold. When the plaster has set, break the form and remove the three parts of the mold from the pole. Some care must be exercised in taking these various pieces apart.

When the parts of the mold are completely dry, shellac them thoroughly. Reassemble the parts back inside the wooden form, making sure that each fits correctly against the other. (There will be a 2-inch opening, lengthwise, between the pieces along the top.) Tighten the form so that there is no movement.

Oil the interior of the mold. Cut the correct length of reinforcing steel of 1 or 5/8ths-inch diameter. Mix a grout of sand and cement at 1-to-1 plus enough concrete with fine aggregate to fill the mold. The most difficult part of making a cast is placing the grout so that all inequalities in the surface of the mold will be filled, as this work must be done through the 2-inch opening. Coat the interior with the grout, place part of the concrete in the mold and embed the reinforcing in the center of the cast. Complete filling the mold with grout and cement.

There are various ways in which the cast poles may be colored. One of the most satisfactory techniques involves the following procedures. When the grout and concrete are prepared, a small amount of mortar color is added as a ground color. It must not make the mix darker than the lightest part of the wood lintel which is being copied. This serves as a base or background for the following stain. After the concrete is removed from the plaster mold, it should be cured in a warm, damp atmosphere for seven days and then let dry out slowly or it is apt to flake. After the curing period, scrub the concrete with zinc sulphate and let dry. It is now ready to stain. Ordinary penetrating wood stain (not varnish stain) is recommended by the Portland Cement Association. Start with a light stain and do not apply it evenly. Follow the directions for rubbing the stain off after a few minutes. Continue to stain the piece using darker colors until the desired effect is obtained.

The suggestions above have been given for making the mold of plaster of paris, a material fairly easy to work and one that sets rapidly. Experienced workers may wish to employ any of the various latex molding compounds if they are making large runs of one particular pole lintel.

Once the technique of simple casting has been mastered there is no end to the possibilities. Lintels may be cast with withes in place, duplicating the actual withes with which they were tied together in the wall. Particular logs which have been badly cracked or partially rotted may be cast in exact replica, and these placed back in position in the structure (fig. 38).

FIGURE 38. A horizontal cast log replaces the original timber which had rotted on the interior and in the embedded portions. (Chettro Kettle Ruin, Chaco Canyon National Monument, N.M.)

INTEGRAL STRUCTURAL MEMBERS

An integral structural member is an internal, reinforced concrete beam placed within the repaired portion of a structure to give added strength. It may be employed 1) in arch form to stabilize holes either part way or entirely through walls, 2) for both vertical and horizontal support, 3) to give added strength over horizontal timbers embedded in the wall, and 4) as bracing for overhanging sections.

Reinforced Arch Sections

With arch sections, large irregular holes may be stabilized "as is." The hole is still present and it has the same appearance and irregularity as before. When reinforced, it will be slightly smaller than the original, but will not develop further.

Full Width Sections

Almost any size or shape of hole can be stabilized by the following method, but little time or money is saved by using it for very small holes. To start a reinforced arch in a large break entirely through the wall, first clean the break to sound masonry and lay at least two courses in concrete mortar across the bottom of the break. More may be necessary in very large holes, or additional courses may be built up on the sides of the hole to form a good base on which to rest the steel which will follow.

Next, two pieces of light reinforcing steel 3/8ths or 1/2 inch thick are bent to fit within the outlines of the break. One piece goes on each side of the wall and just far enough inside the opening to allow a covering face of masonry. The two ends of each piece will rest on the masonry just laid across the bottom of the opening (fig. 39).

FIGURE 39 Integral members starting a reinforced arch section, rods have been bent to fit the contours of the break and are wired in place. The workman is laying the courses across the bottom of the break in concrete mortar to provide a solid base for the reinforced section. Above this level, masonry on the perimeter of the hole will be laid in concrete mortar which will also surround the steel rods. Masonry in the center of the hole will be laid dry or with a minimum of soil mortar; its only function will be to support masonry laid in cement until the latter has set. This dry masonry will then be removed. (Pueblo Alto, Chaco Canyon National Monument, N. M.)

When the steel is in place, it should be securely wired and tied together with additional bracing where required. The purpose now is to build up a thin band or lining of masonry set in concrete mortar around the reinforcing steel. The repair masonry is laid in horizontal courses across the extent of the break, but only that part which encloses the steel reinforcing is laid in concrete mortar. The portion of each course which does not enclose the reinforcing is laid dry, but it must be laid solidly because it will have to support, without shifting, the projecting stone in the vertical irregularities as well as all of the top of the patch until the concrete mortar is set. If the stone is very irregular, it is well to use some soil mortar in the center to give more stability. This procedure—concrete mortar between the stone on the outside next to the steel and dry masonry or soil mortar in the center—is followed until the top is reached. The top courses will also be set in concrete.

After the hole is filled, the concrete mortar area must be kept damp until the concrete has set. The dry masonry in the center is then removed and the hole is left in the same irregular shape (but slightly smaller) as originally, and protected from further breakdown by the reinforced band (fig. 40).

FIGURE 40. Integral member; a large reinforced arch section. The size of the break through the wall precluded filling the center with dry masonry as support, and the timber framework was substituted. The heavily reinforced band is socketed in the cliff wall at the left side, and based on bedrock at the right. The integral member has just been finished and the mortar has not yet dried. After the mortar has set the timber framework, the dry masonry at the right and that on top of the horizontal plank were removed. The large hole with its original outline remained. (Gila Cliff Dwellings National Monument, N. M)

Exactly the same type of reinforced arch sections may be constructed in thick walls where only one side of the masonry is broken out. Ordinarily, only one piece of reinforcing is required.

Vertical and Horizontal Members

The principal application of these members is to tie together sections of separated and out-of-plumb walls, to reinforce walls against which there is pressure of retained fill, to re-lay overhanging and precarious sections, and to provide support over horizontal logs (figs. 41, 42, and 43). Since conditions and structures vary so greatly, it is difficult to anticipate all possible applications. In general, the requirements are:

FIGURE 41. Integral members; steel in place for a narrow vertical section. The steel must be tied in securely to the cross wall at the top to provide support for the overhanging area. Some support will be required for the vertical strip of masonry as it is being laid, and until the concrete mortar has set. (Chettro Kettle Ruin, Chaco Canyon National Monument.)

FIGURE 42. Reinforced integral members are important in the stabilization of spirelike structures such as this at the Tower Kiva, Kin Ya-a Ruin, Chaco Canyon National Monument, N. M. Considerable strength must be added without serious or objectionable alterations of the original profile.

FIGURE 43. The placement of reinforcing in the third story portion of the structure shown in figure 42. The purpose is to reinforce heavily the vertical capping which will be placed up exposed ends of the wall to prevent further erosion, and to tie this part of the wall together.

1. Vertical sections must be based on adequate foundations, and the top should always be well tied into a stable part of the structure by means of longitudinal rods.

2. Care must be taken to achieve as good a bond or tie as possible between the member and the supported wall. Sufficient patching should be done adjacent to the member to prevent any separation from the wall.

3. In arch sections, in overhanging areas, and wherever steel will provide a major part of the support, care must be taken to adequately support the member by framework, dry masonry or the like until the concrete has gained full strength. Correct curing of the concrete in these thin bands is essential.

CAPPING

Capping is considered here as all repairs which extend to and include the tops of walls, and all work of resetting or otherwise altering the upper courses in masonry, brick or adobe construction to provide additional strength or protection against weathering. Capping is intended to tie the upper courses of masonry together, and may employ light reinforcement where walls are out-of-plumb or there are slight overhanging areas. Capping is intended to provide a water-resistant layer at the top of the wall, and to provide additional strength where walls are subjected to unguided visitor traffic.

In brick or adobe construction, primary reliance on the effectiveness of the cap must be placed upon the substitute masonry units used—stabilized adobes made for the purpose, new or treated brickwork. In setting cappings in stone masonry, strength and waterproofing will depend upon the substitute mortar employed. Standard construction practices require the use of continuous metal flashings in the upper courses of exposed walls tops, parapets, cornices and the like. This requirement can be met and the method employed where regular units laid in courses, brick and adobe walls, are being repaired. Standard architectural requirements should be studied and followed in this respect.

The use of metal flashings is impractical for sloping and random courses of masonry walls, particularly where there are great variations in elevation in a single wall. Here capping will be a compromise between sound, watertight construction, appearance and authenticity. Workers should investigate the possibilities of using fabric, plastic, or rubber flashing. Sheet lead might also prove workable.

Evaluation

The most durable capping on masonry structures of the Southwest, at least from the point of service, are some of the 1917-1918 cement caps made at Aztec by Earl Morris, and the 1920-1921 cement cappings laid at Chettro Kettle by Sam Huddelson for the School of American Research. Most soil-bitumen caps laid on various other Chaco ruins in the period from 1937 to 1940 also are quite durable. These three capping projects employed the same technique, in that sufficient depth of stone was relaid in the mortar to provide some measure of tensile strength—an average depth of a foot or more. Each contained embedded stone laid in the same manner as the original work, and in about the same proportion of stone to mortar. Examination of these and other examples indicates that, to be successful, a capping must have sufficient strength and thickness to be a complete unit itself, and will be fastened to the top of the wall rather than a thin sheet or membrane laid over the wall to serve merely as waterproofing.

Field Control of Visitor Traffic

Over three million visitors toured National Park Service archeological and historical parks and monuments in the Southwest during 1972. The greater proportion of these visitors made unattended trips through the historic and prehistoric ruins. The day is past when ruins stabilization, particularly capping, is directed only to protection from the elements. Prime consideration must also be given areas receiving heavy use by the ever-increasing numbers of visitors.

Thus, for the capping of masonry walls, only the best cement mortar made with clean sharp concrete or mortar sand should be used. In many areas this graded sand must be trucked in from commercial sources. It is expensive, but it is also more economical than restabilization following a job using inferior materials.

At extremely isolated sites where materials must be transported long distances by pack animal, soil-cement made with local soils may be substituted for sand-cement mortar. However, it should be substituted only as a last resort since soil-cement, while resistant to weathering, does not usually produce a durable bond against foot traffic, the climbing up of protruding stone, and similar abuse.

1. In capping of any wall, sufficient loose and decayed material must be removed to insure a good bond with sound masonry (of stone, brick or adobe) below (figs. 44 and 45). Where regular units are employed in coursed walls, a sound, even surface must be reached or built up on which to place the flashing. In most cases, sufficient courses should be removed to provide the desired depth of capping without the addition of new units. Where the upper levels of a patch are to form a cap, new stone or other units must be employed (fig. 46).

FIGURE 44. In many cases such as that depicted above it will be difficult to draw a distinction between general wall repairs and capping. This wall will have to be carried up to the level of the outside fill to prevent further washing over the top. Whatever type of mortar is used for such wall repairs, the top courses must be set in a cement mortar to provide as durable and waterproof a cap as possible. (Pueblo del Arroyo, Chaco Canyon National Monument, N.M.)

FIGURE 45. Patching of this hole at Pueblo del Arroyo will be combined with a short section of horizontal capping plus vertical capping for the exposed and weathered vertical rise of wall.

FIGURE 46. A good deal of repair masonry must be laid to fill breaks and support undercut areas in this wall at Pueblo del Arroyo. Some support must also be provided for the embedded timbers. Soil or masonry cement mortars may be used for the major part of this work. The upper courses however must be laid in concrete mortar as capping to halt further washing from the top. In this example, damage was started by surface water from adjacent, unexcavated fill and every effort should be made to prevent erosion.

2. If possible, the work should be laid out so that each wall will be capped as a unit. It is not advisable to run capping across wall junctures. It makes for poor appearance; there is also apt to be some movement at this point, and if capping is carried across the juncture it will break at some weak area.

3. The capping of wide walls should be sloped or tilted to drain water away from wall junctures for a short distance, and laid so that moisture does not pond on the tops. Weak points of capping are at the outside bottom edges where the cap comes into contact with the wall face below. Capping should not be feathered out at these edges, but should retain their full cross-section. These are the points at which runoff enters the wall below, and every effort should be made to produce tight mortar joints with the underlying wall face.

4. Any capping will be a comparatively thin and weak part of the structure. Every effort must be made to produce a durable job. All stone employed should be clean and free of soil. It should be damp when placed in the mortar to insure good bond, and to prevent too rapid drying of the mortar. Control must be maintained over the mixing to prevent excessive amounts of sand or water. Lastly, some provision must be made to insure proper damp curing.

5. Treatment of capping with silicone water-proofing is recommended, particularly those caps employing adobes or stabilized adobes and similar units.

CONTROL OF EFFLORESCENCE

Efflorescence, the appearance of whitish "alkali" deposits on wall surfaces, is a problem closely related to several others covered to this point, including the use of cement mortars, the rise of capillary moisture in walls, and the repair and capping of these walls. It is a condition which is occurring with increasing frequency in the Southwest, particularly in stone masonry structures.

In brief, efflorescence is the result of water soluble salts in masonry brought in a water solution to the surface. The salts remain following evaporation of the water. Two conditions are required before efflorescence will occur: 1) water soluble salts present in the newer masonry components, or in the older components onto which the new work has been added; and 2) presence of moisture to dissolve and carry the salts to the surface. The moisture may be hygroscopic, capillary, or hydraulic in nature. The soluble salts which cause efflorescence may be present in varying amounts in sand, cement, lime, admixtures, mixing water, and masonry units, both new and old.

The primary steps one must take to control efflorescence include the careful selection and use of materials, and the employment of workmanship and design practices which prevent the entrance of water into masonry work. Stone, water, and sand in the vicinity of some sites undergoing stabilization are highly suspect with regard to the presence of deleterious salts. As mentioned earlier in this chapter under "Repair of Wall Breaks," time can be saved if suitable rock and spalls are quarried. The sand should be washed. The mixing water for masonry mortar must be clean and should be pure enough to drink. The following practices will aid in the elimination of efflorescence: 1) in all sizable masonry work that is one stone deep or more, fill head and bed joints completely with mortar; where walls are more than one stone wide, fill compound wall core with rubble and spalls completely embedded in mortar; 2) design capping deep enough to shed water, laying cap stones in such manner that they cannot trap water; 3) cover the top of completed masonry work to keep out rain; and 4) where fill lies behind the wall, seal the back side as described under section entitled "Control of Capillary Moisture."

Where salt deposits have become highly objectionable, attempts should be made to curb the source of moisture, by brushing, and with the use of solvents. A light sand blasting should be made only as a last resort, followed by treatment with a hydrocarbon-based silicone (Daracone or equal).

ROOFING AND ROOF SUPPORTS

In the past, a variety of methods have been used for protecting existing prehistoric ceilings and roofs in situ. They range from the elaborate evaporation-pan type roof installed at Aztec Ruins in the early 1920's to simple, temporary structures. Due to the very nature of the problem—intact ceilings often a story or two below the upper limits of the walls—protective coverings have been difficult to construct, and few have been unqualified successes. Protective coverings of a permanent nature have seldom been constructed near the tops of walls, a story or more above the ceilings, for the following reasons: 1) the difficulties of roofing irregular wall tops, 2) the unsightly appearance of roofs protruding above the masonry, 3) the openings between the ceiling level and the wall top which must be closed or sealed off, and 4) the aboriginal ceilings which often require some support from the protective covering.

Despite their imperfection, the early evaporation-pan roofs installed at Aztec were soundly designed and constructed. They were of concrete slab construction, made to hold water until it evaporated. Such a design obviates the need for drainage lines and downspouts and does not mar the appearance of a site. The roofs functioned quite well during the summer and fall rains; they had ample capacity for seasonal rainfall. During the winters, however, they accumulated great depths of melting snow and ice, and the water level rose above the limits of the pan. At this level, water worked through the masonry and entered the ceiling material below. In time, the concrete of the pans subjected to constant moisture and alternate freezing and thawing began to crack and the surface to disintegrate.

As a remedy, the evaporation pans were remodeled into roofs with drains. Hidden tile drains were installed in the heavy walls and the concrete slab was covered with a built-up type roof covering, finished off with pea gravel or chips. This solution has been only partly successful because the drains frequently became clogged. Most drains are situated at the north side of the site and became stopped with alternately freezing and thawing snow and ice during the winter months. Some of the more troublesome drains were later converted to larger size, opening directly to the exterior of the wall. The result is less attractive, but provides a watertight roofing.

From this and similar experience with difficult roofs evolved the coverings designed for and installed over 14 prehistoric ceilings in the East Ruin of Aztec during 1957 (figs. 47 and 48). These were in an unexcavated site where the only previous protective measure taken was the temporary bracing of cracked and broken timbers.

FIGURE 47. Temporary support of cracked and broken roof beams by means of jacks and timber props. Approximately 6 feet of overburden lay on this ceiling. The temporary supports were in place for several years before the overburden could be removed and a protective roofing constructed. (East Ruin, Aztec Ruins National Monument, N.M.)

FIGURE 48. The ceiling shown in figure 47 after removal of the overburden and construction of a slab roof which protected the ceiling and permitted support of broken beams. The bolts are fastened in the slab above, and were taken up enough to transfer most of the weight from the old ceiling beams to the new roof.

Requirements and dimensions for a protective covering will vary from room to room. However, the major limiting factor is the strength and condition of the walls which support the new roof. In many instances, such as the heavy-walled construction of Pueblo III, stabilized walls are sufficiently strong to support a permanent slab-type roof. Where the weight is excessive for existing walls, the same type of slab roof can be constructed, supported by upright members, i.e., steel I-beam supports rising alongside the walls and based on concrete slabs at floor level. The vertical columns will support the slab while lateral stability is provided by the walls. The vertical supports must pierce the aboriginal ceiling and they must do so at points where they do not interfere with its important members, but where they also furnish the required support.

Slab Roofing

Since ceiling areas vary considerably in size and span, and the condition of the walls also varies, the following are general suggestions only on the construction of concrete slab-type coverings for protection and support.

Lightweight concrete made with commercial lightweight aggregates has extensive use in the construction of modern roof decks. Every consideration should be given to its use when designing slab-type coverings. Weights of concrete made with lightweight aggregate run from 25 to 50 pounds per cubic foot as opposed to an average of 150 pounds per cubic foot for concrete made with sand and crushed rock aggregate. This permits coverings where the weight of heavier aggregates would be prohibitive, and also permits the use of much lighter members below the slab. When sand is added to the lightweight aggregate, a medium weight concrete results; a wide range of strengths may be obtained by varying the proportion of sand to lightweight aggregate. In general, lightweight slab construction should be based on corrugated metal such as Corruform or Tufcor rather than upon paper-backed metal lath or ribbed metal lath. The thickness of ordinary lightweight concrete decks varies from 1-1/2 to 3 inches.

The general design of slab-type roofings are shown in figures 49, 50, and 51. No part of the new roofing may be supported by the aboriginal ceiling either during or after construction. This requires that the construction be self-supporting during the time that the concrete is being poured. Ordinarily, concrete forms for floors and decks are supported during pouring by vertical members from below, and the supports and forms are removed after the concrete has set. However, this is not possible in the placement of slabs over existing roofs where the space between the new slab and the old ceiling is usually less than a foot, and there is no means of access between the two.

FIGURE 49. Typical roof covering and support employed at Aztec Ruins National Monument.

FIGURE 50. Details of roof supports (in section) at Aztec Ruins National Monument.

FIGURE 51. Construction of a slab covering. The numbered members are I-beams run through heavy masonry walls. The longitudinal members are 2 x 4's inset in the I-beams as purlins. Corrugated metal decking will be nailed to the wood as a base for a concrete slab. A covered vent will be provided through the concrete. Drainage will be provided by tile line through the wall at the right. (East Ruin, Aztec Ruins National Monument, N.M.)

The basic members of the new roofing should be steel I-beams. Their construction permits the wood members of the roofing, the rafters or purlins, to be inset in the sides of the beams. Thus, there is no wood separating the concrete from the main supporting members. Wood left in this position would rot and allow the roof to settle. Dimensions of the I-beams are determined by the span, the placement of beams and the type of slab, and the normal or lightweight concrete used. Roofs with a span of 6 to 8 feet normally require 4-inch beams of 7.7 weight, on 4-foot centers. The I-beams should be run through the full width of the wall and should rest on a concrete pad if possible.

Likewise, the number and placement of the wood members of the roofing is determined by the type and thickness of the slab to be poured. They should be ample to support the wet concrete and be well nailed and braced.

While paperbacked metal lath can be used under lightweight concrete, it is preferable to use one of the standard corrugated metal deckings as a covering for the structural members and a base for the concrete; it is light in weight, and easy to handle and cut.

Whatever concrete is used should contain reinforcing, the amounts depending somewhat upon size and weight of the sections of the ceiling below. For example, in the roofs with beams on 4-foot centers, No. 4 rebar was placed on 8-inch centers at right angles to the I-beams. Number 3 rebar was used at right angles again, the short dimension of the slab, on 18-inch centers. Concrete slabs will usually be about 3 inches thick and the reinforcing should be above the center, supported on "chairs" of steel or similar material. All reinforcing should be securely tied. It is preferable to purchase the reinforcing cut to length and the ends bent if desired.

Ceiling Supports

If the prehistoric ceiling has been subjected to heavy overburden or has been wet, it will probably require support from the protective slab, a provision which must be made when the slab is first being laid out. Support may consist of bolts which extend up through the center of broken or cracked beams or by pairs of bolts, one on each side of the beam to support a hanger under the beam. Bolts should have ample threads (threaded rod is preferable to standard bolts) so that they can be taken up sufficiently to support most of the weight on the beam. In normal concrete, large cut washers are sufficient to anchor the bolts; in lightweight concrete the upper ends of the bolts or rods should be tied in to the reinforcing. Ventilation must be provided between the old ceiling and the new roof to prevent accumulation of excessive moisture from condensation. Most prehistoric ceilings are not tight enough to prevent some movement of air. When clearing debris or covering preparatory to constructing a slab, sufficient material should be removed to permit some percolation of air through the ceiling. Provision should also be made for ventilation through the top. This is best accomplished by including a vent of 4-inch pipe in the slab, the ends covered with screen or turned down.

Finishing and Flashing

One of the most troublesome problems encountered in the construction and maintenance of roofs set deeply within structures is that of flashing. It is impossible to place metal flashing in the usual irregular ashlar of a ruin wall without tearing out a wide band of the masonry and relaying it with the flashing in place—a dangerous procedure in deep rooms, impossible in narrow walls and always weakening to the structure.

To avoid the difficulties of inserting metal flashing, well-formed cant strips should be employed. They do not always provide the absolute freedom from water seepage as a correctly employed metal flashing, but they are often the only possible alternative. In placing cant strips, either in new roofs or in repairing existing slab roofs, the wall where the strip is to lay should always be cleaned and the joints well pointed with cement mortar some distance above the top of the strip. This strip can then be waterproofed with hot asphalt, pitch or whatever material is used on the remainder of the roof.

In pouring the slab, every effort should be made to produce as dense and watertight concrete as is possible. The water-cement ratio of normal concrete should be strictly controlled. When poured, the material should be well-rodded and the top screened, i.e., leveled out to grade to a true, even surface. Whether of normal or lightweight type, it must be damp cured for no less than three days and in hot weather protected from excessive drying for a longer period.

The use of either type of concrete will require a waterproofing surface. Either kind must be thoroughly dry before the surface is applied. Lightweight concrete should be covered with a built-up roof, surfaced with chips or pea gravel. While normal concrete is employed, it also must be waterproofed. A built-up roof can also be used.

Where it is desirable to cover the slab with soil for improved appearance, the slab may be water proofed by an application of hot pitch mopped on. Pitch does not deteriorate as rapidly under a soil covering as the usual asphalts used for built-up roofings. If a soil covering is to be used, the wall surface should be repointed with concrete to at least the depth of the soil, and this area also waterproofed by an application of pitch.

SETTLEMENT AND SUBGRADE

As mentioned in this book, one of the common structural faults encountered in prehistoric sites involves the loose and unconsolidated fill upon which they were built. True foundations or footings are not present in prehistoric sites. Walls were usually laid on existing ground surfaces with some attempts at leveling and occasional terracing on sloped areas. In some cases a trench was prepared for the first several courses of masonry. Inadequate foundation conditions include walls or parts of walls which rest on materials containing organic matter, on low density materials such as loose deposits of silt or sand, and on talus deposits, spoil piles, and dumps.

By contrast, the post-1540 A.D. historical structures of stone masonry or adobe bricks were built on true, prepared foundations. Unless they were located on favorably drained locations, however, many were not waterproof because the foundation stones were laid either in soil mortar or porous lime mortar.

It is surprising that wall failures resulting from settlement are not more common. In the Pueblo III multi-storied structures of massive masonry, vertical cracks and slumped or severely leaning segments are frequently noted. These are the result of stresses and bearing loads of heavier walls. Shear stresses from unequal settlement are often indicated by cracked or broken masonry stones.

In modern construction of foundations, a careful evaluation of the subgrade soil is necessary, requiring tests which serve to define the character of the material and the quantitative properties which define specific performance characteristics such as shear strength, pore-water strength, capillary stress, compressibility, and permeability. Foundations for rigid structures are usually evaluated on the basis of bearing capacity which involves both the shearing strength of the soil and its consolidating characteristics. Therefore, the conditions for which adequacy of a foundation should be established include bearing capacity, stability, settlement, and permeability. Field and laboratory tests required to measure these characteristics demand specialized equipment and highly experienced investigators.

Soils containing large amounts of silt and clay are rather unstable. These materials exhibit marked changes in physical properties with change of water content. A hard, dry clay may be suitable as a foundation for heavy loads, but may turn into a quagmire when wet. Many of the fine soils shrink on drying and expand on wetting which may adversely affect structures founded on them.

A frequent occurrence in valley fills of the Southwestern United States is the percolation of water through soils. This results in the gradual removal of soil particles either by solution or mechanical movement of particles, introducing disturbances in the overburden indicated by piping and sinkholes. Where such pipes and sink holes are sufficiently close to prehistoric or historic sites to threaten stability of the structures, they should be stripped to sufficient depths and filled with impervious compacted soil.

The fine grained valley fill-soils in the populous archeological areas such as Chaco Canyon contained sufficient gradation, uniformity, and in-place density of silts, clays, and sands at the time the large sites were constructed to provide a proper amount of compactability and compressibility for bearing the weighty architectural loads. Exceptions, of course, resulted in the occasional slumped and leaning walls involving massive masonry. Today in Chaco Canyon some of the ruin areas are covered with several feet of aeolian silt and sand which are unsuitable for foundation material. Piping and sinkholes in the valley floor near the central arroyo pose threats both to modern and prehistoric construction.

UNDERPINNING AND SHORING

The most frequently used and simplest method of improving foundation stability and halting settlement is to excavate beneath the lower limits of the wall, remove unsuitable material, and spread the base by installing a widened footing of masonry so that unit pressures are acceptable. If large wall areas are involved, short sections of underpinning must be installed and allowed to cure before moving to the adjacent section, thereby not leaving too much wall unsupported at one time.

Concrete piers or foundations may be installed by the above skip-tunnel method. Both horizontally and vertically placed reinforcing rods should be installed before the concrete is poured. Any shrinkage between a cured foundation and the wall it supports must be tightly grouted before trenches are backfilled, tamped, and the surface soil graded to drain away from the wall and concealed footing.

In modern construction it should be realized that, unless a structure rests on solid rock foundation, some settlement cannot be avoided. Where it is not feasible to found a structure on solid rock either by placing it directly on bedrock or by means of poles, piers, caissons or walls, the structure must be capable of withstanding some vertical movement in the foundation. As suggested for underpinning prehistoric structures, the usual method of reducing movement along the foundation to within acceptable limits is by spreading the base or footings so that unit pressures are sufficiently small.

Wherever historic and prehistoric structural foundations have been investigated, it has been shown that a firm, uniform foundation soil, free from loose sand, low density silt, soft clay, isolated rock masses, or cultural refuse, proved adequate for most structures with moderate loadings.

SUB-SOIL INTRUSION GROUT

While a broad guide to the character and gradation of soils can be accomplished by personnel with little training, the determination of soil engineering properties requires considerable skill and insight if reliable information is to be developed. Therefore, in complex problems involving cause of foundation failures, and for providing engineering-designed remedial measures for settlement of walls, it is suggested that the field man call upon the services of a soils and foundation expert. Moreover, foundation problems in general are not susceptible to exact engineering analysis, even with careful soil sampling.

A valuable illustration of how a serious foundation problem was solved at Canyon de Chelly National Monument by the National Park Service, in cooperation with a large, private firm skilled in soil stabilization, concrete, and cement grouting, is briefly outlined below. The solutions have wide application to historic structures elsewhere, particularly those which may be severely cracked due to settlement of foundations.

White House Ruin, a four-story stone masonry pueblo dating from ca. 1050-1300 A. D. was built on valley fill against a sheer canyon wall of a stream which flows part of the year. The water table fluctuates seasonally, but is always consistently high. The present sandy stream bed is 50 yards in front of the ruin. During the dry season, water can be reached at from 3 to 4 feet below the surface sand. Before prehistoric occupation of the canyon, the stream channel flowed at various times along and at the base of the cliff. Later, sediments and alluvial fill built up against the cliff and it was along these strips of fill that the prehistoric Indians built their homes. Stabilization of the ruin was accomplished in 1941-42 and again in 1956 by the National Park Service. Work was confined to minor repairing and patching of walls which suffered defects partly from weathering, but increasingly more from visitor impact. By 1958, however, large vertical and diagonal cracks, as well as wall separations, were rapidly developing throughout the structure.

A thorough inspection of the site was made by National Park Service archeologists and the chief engineer of a large private firm whose speciality is foundations. In the opinion of the engineer, seasonal fluctuations in the water table were removing fines from the upper soil strata. This downward migration of fines is very common in valley configurations and results in reduced density and bearing value of soil immediately beneath structures founded on the valley floor. The engineer suggested that it is usually possible to improve these values by intrusion-grout injection. The problem was further complicated by the fact that some walls of the ruin rest on refuse and earlier occupation debris.

In the best judgement of the assembled experts, two courses of action were considered for attempting to save White House Ruin:

1. The most positive method, and by far the most expensive, involved the use of concrete cast-in-place or mixed-in-place piles placed adjacent to the failing foundation or lowest limit of the wall and tied to it by means of needle beams beneath the wall. To be effective, the entire length of the wall must be supported by a grade or bearing beam which, in this case, would have to be installed by tedious hand methods. Because of the expense involved and the considerable disturbance of underlying occupational debris, pile underpinning was ruled out: National Park Service budgets simply could not bear the cost.

2. The second method of foundation improvement consists of injecting intrusion grout into the unstable soil. This relatively economical process, described below, depends very largely for its success on the skill and experience of the job superintendent and the reliability of the firm. It is a technique which has been applied in literally hundreds of commercial situations with complete success in most cases.

The second method was chosen for White House Ruin and was successfully performed in October 1958. Since that date, no further settlement of walls has been detected. In a subsequent stabilization job, undertaken shortly after the substrate was consolidated by grouting, all cracks and defects in the superstructure were corrected not only to make the ruin stable again, but also to allow us to observe any further settlement should it occur.

The intrusion process of sub-soil stabilization is accomplished by placing small diameter pipes up to 30 feet in length into the soil to be stabilized. Intrusion grout is forced through these pipes to fill voids and to consolidate the material. By repeating this process the required number of times, the soil is strengthened by the material placed and a raise can be effected by the hydraulic pressure exerted while placing the intrusion into the soil. The lateral distance between inserted pipes and the amount of stabilization required varies with soil conditions and loading specifications.

The intrusion grout used at White House Ruin was a cement base slurry which included the patented admixtures, Intrusion Aid and Alfesil. Intrusion Aid provides delayed setting, suspension of solids, low water-cement ratio and compensates for normal setting shrinkage. Alfesil is a premium grade of finely divided siliceous material which contributes to the penetrability of the grout. Alfesil is pozzolanic material composed essentially of compounds of silica, aluminum, and iron. It possesses the property of combining with lime that is liberated during the process of hydration of portland cement to form additional insoluble cementitious and strength producing compounds.

Exceptional care was exercised during grouting operations since hydraulic rise was to be avoided. The grout pipes were driven into the underlying material adjacent to and under the walls throughout the ruin to a depth of 20 feet. Grade stakes were established adjacent to the walls at strategic points, and a careful check maintained to avoid all but the faintest movement of substructural material. The grout was pumped into the soil under controlled pressure as the grout pipes were slowly withdrawn, creating lenses and columns of hardened grout within the foundation material to increase its bearing capacity.

It is emphasized that sub-soil intrusion grouting beneath irreplaceable ruins is not a do-it-yourself project. Every foundation presents a distinct grouting problem, depending on the composition and nature of the substrate. The method requires considerable specialized, highly mobile equipment permitting economical placement in normally inaccessible locations. Equipment includes an air compressor, paving breaker and driver head, air-operated grout mixer, water batches, and grout pumps, together with insert pipes, grout, air and water hoses, level, gauges, fittings and various hand tools. The job superintendent, crew chief, and laborers must be highly skilled and experienced.

Tests were conducted throughout the White House Ruin project to obtain reliable information on the increase in soil density as result of grouting, at the same time permitting no hydraulic rise which, if not gauged properly, would have lifted walls out of the ground. Since 1958, the sub-soil stabilization at White House Ruin has halted the settlement of the substrate and the subsequent cracking, separating, and shearing of prehistoric masonry.

MONOLITHIC STRUCTURES

The history of the preservation of monolithic soil structures such as occur at Casa Grande, detailed elsewhere in this book, has been one of experimentation, frustration and unsatisfactory results. The early use of burned brick in repairing and supporting large undercut sections of the "Big House" in Compound A has proven eminently satisfactory. Much less satisfactory have been the preservation attempts on low and exposed walls in the remainder of Compound A, Compound B, and the Clan House. While waterproofing and hardening solutions have been tried, the major effort directed toward preservation has been with thin coatings or plasters of soil-cement, soil-cement made with caliche, cement stuccos, and a covering or capping of formed adobes. Caliche is a crusted calcium carbonate formed on certain soils in dry regions.

None of these have proven to be the ultimate, one-application panacea. All the various coatings or plasters have had to be renewed. However, it should be pointed out that they were not all complete failures. Although they must be replaced, these coatings have protected and saved many prehistoric walls. The earliest of the coverings were of thin cement stucco; they were followed by the use of soil-cement at a fairly stiff consistency, and tinted to match the original work. They were plastered over all exposed surfaces. The weakness of these applications, soil-cements and cement stuccos, was the absolute lack of bond between the prehistoric wall and the "plaster" as well as the tendency of the plaster to absorb moisture and thus aggravate the lack of bond. The appearance of such surface coverings was undesirable.

The addition of woven wire and other reinforcing material in the plaster has lengthened the life of some of these surface applications of soil-cement. Recently developed water repellent. applied to the surface indicate that they also will materially prolong the useful life of the repair.

Since long experience at Casa Grande has demonstrated that it is impossible to make a thin covering of either cement, stucco, or soil-cement which will adhere to the original wall or will not crack and remain waterproof, a slightly different approach has been tried recently on exposed walls of Compound B.

The thought behind this approach is that a heavier, thicker section of soil-cement or caliche-cement would not be dependent upon the prehistoric wall for bond and support but would, if thick enough, be self-supporting. Given the present state of knowledge, it is evident that we will not be able to develop and thoroughly test any preservative or hardening solution for exposed soil walls in sufficient time to abandon other tried, though partial, solutions. It is for this reason that formed soil-cement or caliche-cement coatings have been used recently at Casa Grande. With modifications, they may be used in other, similar situations. They represent the latest step in the evolution of plaster-like coverings, and it is possible that they may prove to be the final solution.

Formed repairs to badly weathered walls of caliche at Casa Grande (exposed since the excavation was completed in 1908) are more natural appearing than are the thinner, plastered coverings. The thinner coverings followed the undulating, irregular, undercut convolutions of the wall. The former repairs are somewhat more regular and more nearly match the formed appearance of the original wall surface.

Following is a brief description of the latest wall surfacing work at Casa Grande.

In all caliche walls, i.e., those of high lime soil content encountered in Compound B, weathering had greatly reduced the original width above the surface of the ground. Both sides of the wall were trenched until the original width was reached at depths up to 18 inches. Forms of heavy plywood, reinforced with 2x4's, were then placed at the edge of the wall rising vertically from the uneroded subsurface line (fig. 52). The space between the remaining wall surface and the forms was wide enough so that the caliche-cement would be sufficiently thick to stand by itself, and not be dependent on the aboriginal wall for support. General requirements for placing concrete forms were followed; they were oiled, and securely braced.

FIGURE 52. Placement of forms for stabilization of a wall with caliche-based soil cement. The forms are braced at the sides and tied at the top. (Compound B, Casa Grande Ruins National Monument, Ariz.)

Trial mixes of several local soils were made. A caliche obtained from commercial sources was selected for use as most nearly matching the prehistoric wall and providing the hardest product. A mix of 18 percent cement with a small amount of mortar color was used. A portable, drum-type mixer was used.

The prehistoric wall was dampened and the mix placed between it and the form (fig. 53) well tamped in place to fill undercut spots. It was placed at a depth sufficient to cover the rise of the wall face and was then sloped to conform to the irregular, sloping surface of the wall top. After curing, the forms were removed and when completely dry the entire surface was treated with silicone water repellent (fig. 54). This use of heavier, self-supporting coverings of native materials parallels the already proven use of heavy sections of soil-cement in the repair of pit houses and for making facings in cuts in soft fill (figs. 9 and 10).

FIGURE 53. The start of placing caliche-cement mix over the wall depicted in figure 52. Note that the forms follow the irregular line of the wall.

FIGURE 54. The application of a silicone water repellent. It is applied in a coarse stream, sufficient to produce a rundown of from 6 to 8 inches. (Casa Grande Ruins National Monument, Ariz.)

ADOBE STRUCTURES

This subject is closely related to monolithic soil structures in that the raw materials are the same. Only the method of construction differs. The word "adobe" appears to have come from the Spanish, adobar, to plaster, traceable through Arabic to Egyptian hieroglyphic for "brick." Hence the term has come to mean "sun dried brick." Adobe is, therefore, merely soil or earth—usually a combination of sand, clay, and silt mixed with water used in the manufacture of a brick in a mold. Bricks made from soil in this manner are also called "adobes." The best adobe soil is a coarse grained, well-graded earth (see Chapter 3). It is not to be confused with the so-called adobe clay or gumbo found in some strata which heaves and expands when wet, and shrinks badly when drying, forming large cracks (Neubauer, n. d.). This material is identical to the soil or earth mortar employed by the prehistoric Indians in laying stone masonry, in pit house construction, and in the pisé de terre construction of Casa Grande where the puddled or stiff mud was laid in thick courses.

Other writers have shown that adobe or earth brick construction has a long history, going back to B. C. 7000 in the Near East; to B. C. 5000 in Anatolia, Crete, Egypt, and Indus Valley; and to B. C. 3000 in the Chicama Valley of Peru in the New World (Lumpkins, 1971; Steen, 1971). While adobe construction was known in Peru and Mexico before the arrival of the Spaniards in the 1500's, nevertheless it is they who are generally credited with having introduced the form-molded earth brick to the Southwestern United States.

Soil has very little strength compared to other materials used for building a structure. Moreover, compared to the maximum soil strength that may be found, there is a great variation both from soil to soil and within a given soil type depending on how it was deposited or placed. Since moisture is the most influential factor affecting the properties of a soil, and is the principal agent. subject to change from natural causes, it is not surprising that adobe construction is confined largely to arid and semi-arid regions of the world. In the United States, adobe construction is found principally in the Southwest and West in areas of low rainfall, where average annual precipitation is 20 inches or less.

Recognizing the severe limitation of adobe construction, its successful longevity depends, as one writer so colorfully put it, on three things: an adobe building should be given a pair of rubbers, a hat, and sometimes a raincoat (Hubbel, 1943, p. 23). Thus to preserve and maintain an adobe structure successfully, it must be well drained and the foundation must be waterproof to prevent the entrance of either hydraulic or capillary moisture. The structure must be properly roofed, and the exterior walls should be waterproofed either with a material mixed with the bricks or protected by a plaster or stucco surface. If any of the three essential protective elements are missing, the structure is vulnerable to deterioration which may range from extremely rapid (virtually overnight in the event of a cloud burst) to very slow. The two most important elements are roofing and foundation. If both are sound and waterproof, an adobe structure will last indefinitely, and the rate of erosion of the exterior walls will be greatly minimized.

Upon acquisition by the National Park Service of such historic sites as Fort Bowie, Fort Davis, Fort Union, Pecos, Tumacacori, and the early 20th century Spanish-American residences in Big Bend National Park, nearly all of the adobe structures were in ruins. Walls were standing to their original height only in those rare instances where a building was maintained until (or a few years prior to) the date of acquisition by the National Park Service. All of the structures at Fort Bowie, Fort Union, and Pecos were in ruinous condition. Many were reduced either to low mounds or mere foundations.

It should be mentioned that by 1900 the roof of the nave at Tumacacori was missing, the result of a combination of vandalism and erosion. The sidewalls, which were constructed of adobe bricks, had begun to deteriorate very severely. If Frank Pinkley, then "The Boss" of the Southwestern National Monuments Organization, had not installed a roof over the nave and performed other repairs in 1920, there would be very little left of this 18th and 19th century mission other than a mound of "melted" adobe.

Modified Restorations

In view of the above comments, the most effective method of preserving an adobe building involves a partial or modified restoration, especially applicable where sufficient architectural remains are present, and assuming that adequate documentary and archeo-historical data are available. Figures 55 through 59 show modified restorations at Fort Davis National Historic Site where it was possible to reconstruct walls partially to the extent necessary to support new roofs and to arrest deterioration. Authentic roofs and porches were constructed, based on good evidence. House interiors were left. largely as found. This treatment accomplished the dual purpose of preserving the original fabric of the structures and creating an effect, so far as external appearances are concerned, similar to the historic period.

FIGURE 55. One of four 2-story adobe structures, designated HB-14, at Fort Davis National Historic Site. The building was probably used as a residence for important non-commissioned officers. Front view in 1964 before repairs were undertaken by the National Park Service.

FIGURE 56. Stabilization and modified restoration in process at HB-14, shown in figure 55. Soil-cement adobes were used for repairing walls. A concealed integral bond beam was installed around the building between the first and second story levels. Concrete bearing beams were installed at roof line to support restored roof.

FIGURE 57. The completed stabilization and modified restoration of building HB-14, Fort Davis National Historic Site. Compare with figures 55 and 56.

FIGURE 58. Building HB-15, a non-commissioned officers quarters at Fort Davis National Historic Site, before combined stabilization and modified restoration work was accomplished by the National Park Service.

FIGURE 59. Building HB-15 and its outbuilding, a kitchen, following modified restoration by the National Park Service. Soil-cement adobes simulating original adobes were used as modular units of repair. Compare with figure 58.

This construction restored the exteriors of two 2-story buildings to their original appearance. The tops of the walls, weakened by erosion, were rebuilt. and capped with reinforced concrete-bearing beams to support new roofs. Where exterior faces of adobe walls were severely weathered, a masonry veneer was added to rebuild the wall section to its original thickness using soil-cement units colored and textured to match the existing adobes. New wood porches were added to the exterior entrances, and doors and windows restored to the openings to secure the buildings. Thus the structures are interpreted as part of the historic scene. A self-guiding trail provides access and supplies necessary for interpretation.

The visitor center at Fort Davis is an adaptive restoration. The 1870 enlisted men's barracks, which was a roofless, ruinous shell in 1964, was left virtually intact and a modern veneer of soil-cement bricks built around it. Some sections of original adobe walls were pared down so that the veneer would make up the original width. Bearing beams and integral support were provided for the modern though authentic roof. The restored portion of the exterior was faithfully preserved in both form and detail, while the interior was converted to modern, functional use (figs. 60 and 61). A small, glass-enclosed framework on the interior of the restored barracks shows an original portion of adobe wall for viewing by visitors.

FIGURE 60. The 1870 enlisted men's barracks (in right foreground), Fort Davis National Historic Site. The photo was taken in 1960, before National Park Service status. Historic documents, drawings, and photographs provided ample evidence on which to base a restoration.

FIGURE 61. A view of the Fort Davis barracks after restoration. The ruinous shell seen in figure 60 was veneered with hand molded soil cement bricks, 4" x 4" x 12". They were made with one part portland cement (untinted), three parts unwashed sand, and three parts native soil. The bricks laid in masonry mortar with rough joints, set with spalls, were wire-brushed. Original walls were pared down where required to accommodate the veneer brick and to retain original wall dimensions and thickness. Bond beams were installed at roof line to support the restored, hipped, shingled roof. The soil cement brick walls were painted with white cement paint, although originally they were whitewashed. The building is used functionally today as the Visitor Center. Cost of the restoration: $129,000 or $12.90 per square foot.

In most cases, however, both policy and economy dictate that historic, ruined adobe structures will also effectively serve their purpose if they are not restored, but are stabilized to maintain a ruined scene. In other words, preserve them "as is," and permit them to retain an atmosphere of abandonment.

In preserving an adobe structure, or the ruinous wall remnants of a structure, regardless of whether the end result will be partial restoration or a stabilized site, a choice must be made with respect to methods or materials to be used. In the present state-of-the-art, the hard reality is that there are only two basic choices: either stabilized or unstabilized adobe, used either as mortar or in brick form. It has been implied earlier in this work that an all-purpose chemical surface spray which will preserve adobe indefinitely has not yet been developed. For more than 40 years the National Park Service has experimented with chemical sprays which will both harden and weatherproof adobe. Many of these experiments were conducted by Charlie R. Steen, beginning at Casa Grande Ruins National Monument. None was successful. However, from a combination of Steen's experiments and those of the Ruins Stabilization Unit, plus the modification and adaptation of developments in the commercial field, two valuable lessons were learned: 1) certain silicones are excellent for waterproofing hard surfaces such as soil-cement, and 2) the use of additives in manufacturing simulated adobes greatly increases their longevity.

The use of either plain or stabilized adobe modular units should be thoroughly planned and tested for any given area. Usually, soil for historic-adobe buildings was obtained within a short distance of those buildings. Thus, at Fort Bowie National Historic Site at least one or two of the original borrow pits were located by the ruins stabilization crew in advance of proposed work. Various soils from the vicinity should be tested as described in Chapter 3, and either experimental adobe bricks, soil-cement bricks, or soil bricks with chemical additives should be made and tested for matching color and texture, and for weathering and durability. A year or more lead time is not too long for selecting and testing materials to insure satisfactory results on any project.

Manufacture of Adobe Brick

The manufacture of adobe and stabilized adobe is briefly outlined below together with some of the formulas which have been used successfully. It cannot be emphasized too strongly that suitable building blocks for any given area depend on exhaustive tests with local soils and various formulas followed by accelerated weathering tests and systematic observations similar to those described in Appendix 1. Hence, the discussion below must be considered as a guideline only. Local conditions may vary to the extent that some of the formulas cited may have to be radically changed and, in some instances, may not be suitable at all.

Adobe and stabilized-earth blocks are molded in two ways: with a machine press, or by casting the plastic-soil material in forms by hand.

Several earth-block presses are on the market. One of these devices is the CINVA-RAM block press. Evolved by the Inter-American Housing Center (CINVA) in Bogota, Columbia, South America, the all steel press consists of a mold box in which slightly moist earth or earth with cement or chemical additives is compressed by a hand-operated piston and lever system. The press is portable, weighing about 130 pounds. The three basic operations in making the compressed blocks are: (1) loading the mold box, (2) compressing the mix, and (3) ejecting the finished product. The compressed block can be removed immediately from the press without the use of a pallet. The press is made by Metalibec Ltd., Apartado Aereo 11798, Bogota, Columbia, a subsidiary of International Basic Economy Corporation in New York. Bellows-Valvair, a division of IBEC, 200 W. Exchange Street, Akron, Ohio 44309, are distributors of the CINVA-RAM press for the United States and Canada. Their selling price for the press in 1967 was $175, F. O. B. their warehouse, Tallmadge, Ohio.

The blocks made from these devices have two advantages over cast blocks. The press-made blocks are more uniform in size and shape. They are usually stronger—as much as several times as strong. But the presses make only one block at a time, and production is slower than casting. The National Bureau of Standards tested pressurized blocks using a laboratory-machine press which produced a block of similar quality to the CINVA-RAM block. The mix employed was a soil containing 46 percent sand and 46 percent silt and clay, with 8 percent portland cement. Blocks withstood pressures up to 800 pounds per square inch compressive strength (U. S. National Bureau of Standards; Building Materials Structures Report BMS 78). This compares with 100 p. s. i. for hand-molded adobes. and about 400 p. s. i. for hand-molded soil-cement adobes of the same formula used in the block press.

Press-made blocks are difficult to antique. Uniformity, and the use of structurally modular units are desirable in modern construction, but in historic adobe construction the original blocks which must be matched are usually weathered and rounded along exposed faces. To achieve this effect, it is easier to make a slumped block in a hand form and to texture the surface which will be exposed with a brush or other tool before the block has completely "set."

Hand forms for molding bricks may be made to mold any number of bricks. At Fort Bowie National Historic Site, the Ruins Stabilization crew made a form out of 1x6 lumber which molded six adobes at a time, in the most common size of 4x8x16. Lining the form with sheet aluminum made it easier to use.

After selecting the soil, sample blocks should be made and permitted to dry. If the sample warps or cracks upon drying, there is too much clay or loam in the soil and sand must be added to make a satisfactory building block. If the sample block crumbles, there is too much sand in the soil, and clay should be added. A distinguished Southwestern architect and authority on adobe (Lumpkins, 1971, p. 9) presents the following mechanical characteristics for determining whether a soil will make good adobes:

1. It is easy to mix with shovel or hoe when water is added. The soil should part free of the tool and should not gum or stick.

2. It should not fall apart when turned with a shovel into a small 8-inch-high mound.

3. It should slip free of the mold when the frame is lifted.

4. When left to dry it should not. curl, warp or crack.

5. When dry the block should be easily moved, with little tendency to break or chip off at the corners.

6. Light to moderate rains of a 10 to 15 minute period should show little erosion or washing of the adobe brick.

Aside from the fact that soil for adobe bricks should be well graded with proper amounts of clay and sand, it should also be relatively free of humus. Soils containing alkali, caliche, or gypsum should not be used.

As Lumpkins (1971, p. 9) further points out, adobes that are not thoroughly dry should not be placed in a wall. He suggests that "to test for dryness, the block should be broken in two. If the outer edges are lighter than the core area, don't use; allow more time for drying. In the summer most soils will dry ready for use in from 30 to 60 days after the bricks are stacked."

An empirical field test for strength of cured adobe bricks is as follows: drop an adobe from waist height to the ground; if the adobe does not break, it is strong enough.

While manual labor is, of course, the historical method of brickmaking (which involved mixing the mud by hand-hoe in a pit), modern power-operated mixers are the choice today because of high labor costs. A hoe-type plaster mixer also does a more thorough job. Once the proper formula for either adobe or stabilized adobe has been determined, the ingredients should be carefully measured for each mix. Next to composition of the soil, the essential control for adobe brick is in the moisture content. The allowable limits range from 16 to 20 percent by weight. If moisture content is excessive, the mixture will be too thin and the block will slump excessively when the form is removed, and there will be serious shrinkage cracks in the dry block. Adobe bricks are cured by air-drying, while soil-cement simulated adobes must be damp cured for a week. To conserve water, the stabilized soil-cement adobes should be covered with damp burlap which, in turn, is covered with polyethylene sheeting. On the driest of days, usually only one or two wettings of the burlap is necessary.

Most adobe experts today do not advocate the admixture of straw or other vegetal fibers to the adobe mix. Although straw mixed with the soil may reduce excessive cracking, it also provides available channels through the block for the immediate conduction of water, increasing the effects of capillary action.

To compute the number of adobe bricks required per square foot of wall of any thickness, obtain the number of square inches in the face of one brick, including the thickness of the mortar joints and divide 144 by the result. For example, assume that a brick 16 inches long, 3 inches high and 8 inches wide is being used, and that both the vertical and horizontal joints will be 1/2 inch thick. A brick 16 inches long, plus 1/2 inch for the vertical, or end mortar joint, makes a total length of 16-1/2 inches. A brick 3 inches high, plus 1/2 inch for the horizontal or bed-mortar joint, makes the total height of each brick course 3-1/2 inches. Then 16-1/2 x 3-1/2 inches = 57-3/4 square inches on the face. Dividing 144 by 57-3/4 = 2.49 or 2-1/2 bricks per square foot of an 8-inch wide wall (one brick thick). If the wall is 16 to 18 inches (2 bricks) thick, 2 x 2-1/2 = five bricks per square foot. It may be mentioned that considerable stabilization requires veneering which may involve a brick only 3 to 4 inches wide rather than 8 inches, in which case a divider may be placed in the form, thus cutting the space in half lengthwise. While the number of bricks required per square foot of wall area remains the same, twice as many bricks can be made with the same amount of mix.

Similarly, by computing the number of cubic inches of mortar surrounding one brick when laid in the wall, dividing by 1,728, and multiplying the answer by the number of brick to be used, the number of cubic feet of mortar required to lay the brick may be obtained.

The mortar for laying blocks should be of the same material as the block itself. There are interesting examples of differential weathering where dissimilar mortar was used. In several walls at Fort Bowie, adobes were laid in lime mortar. It is probable that these same walls were originally covered with lime plaster.

Turning now to a discussion of stabilized adobe, it should be understood at the outset that there is disagreement among experts as to the relative merits of using stabilized adobe between plain adobe block construction, and the stabilized block where used for purposes of either restoration or stabilization. The addition of some of the stabilizers used today will change the color and appearance of the adobe block. If this objection can be overcome by surface texturing and integral tinting, or if the block work is to be covered by other materials such as plaster or paint, then the use of a durable, weatherproof block is highly desirable. It should be pointed out, however, that even though the original source of soil may be used on a given structure, it is very difficult to make plain adobes in any quantity where color and texture matches exactly those in the historic wall into which they are being integrated. Given enough time for experimentation with local materials in advance of a project, it is usually possible to come up with a suitable stabilized block. At any rate, current consensus in the adobe areas of the National Park Service is that, for reasons of policy and economy, stabilized block work which will need less repair or replacement in the future should be the objective.

Stabilization of Adobe Brick

There are several methods for stabilizing adobe brick, and only those which are commonly used or have been tested and found effective in varying degrees will be mentioned. The addition of a stabilizer in no way affects the choice of soil or sand-clay proportions previously outlined.

Two methods for stabilizing adobe bricks have already been mentioned. One is with the use of cement and mortar color admixtures, and the other is with the addition of an emulsified asphalt or bituminous stabilizer especially designed for the purpose. Instructions for making soil-cement and bituminous-treated adobes are identical to the methods outlined for the mortar in Chapter 3, under the sections entitled "Soil-Cement Mortar" and "Soil-Bitumen Mortar," with the exception that about 5 to 10 percent less water is required. If a CINVA-RAM or other block press is used, water for the mix must be held to an absolute minimum, and as with all other mixes described, all ingredients must be thoroughly pulverized and mixed. Bricks made with block presses also require 4 to 6 percent less cement because of compaction. When using a block press, a simple test may be made to determine the correct amount of moisture by squeezing hard a ball of the soil mix in the palm of the hand. If the ball can be broken in two without crumbling and without leaving any moisture on the hand, the moisture content is correct. Obviously, the tendency is to use too much water. During the mixing of material to be block-pressed, the mix appears just slightly moist. If any water is visible, the mix is too wet.

Some representative formulas used on stabilization jobs in Southwestern adobe areas are listed below along with notes on coloration. At Fort Bowie, the mix consisted of six parts local, well-graded soil, one part portland cement, and mortar color. Three coloring mixes were evolved in order to achieve the several color shades of the bricks. For a buff color, one-fourth of a one pound coffee can of Desert Tan mortar color, and one can of Sunset Orange was premixed dry with each sack of portland cement. To produce a reddish adobe, 25 percent of the normal brownish soil was replaced with a reddish clay, using the same mortar color combination in the stock mix of cement as before. Gray adobe walls were matched by using one-fourth can of Desert Tan added to the stock mix of one bag of portland cement. Sunset Orange and Desert Tan are products of the Tamms Mortar Colors, Inc., 1222 Ardmore Ave., Itasca, Ill. 60143 (Morris, 1967, pp. 1,2).

At Pecos, in the 1969 stabilization of the 18th-century mission church, the following formula was used: four parts local soil screened through 1/4-inch mesh, two parts concrete sand, and one part stock mix. The stock mix consisted of 1 cup of No. 6803 Frank D. Davis Brown Mortar Color mixed into one bag of portland cement. The cement color was specially formulated for the National Park Service by a private firm, Frank D. Davis Company, Cement Colors, 3285 East 26th St., Los Angeles, Calif., based on the soil sample sent to the firm in 1968.

The formula for soil-cement currently used for stabilization and repair work at Fort Union National Monument is listed as follows: 1-1/2 cubic feet of concrete sand, 1-1/2 cubic feet of local soil screened through 1/4-inch mesh, 1 pound burnt sienna dry-mortar color, 1/4 pound raw umber dry-mortar color, and 6 pounds portland cement. These quantities make up one batch of mix.

At Fort Davis National Historic Site the formula currently in use is: three parts concrete sand, three parts local soil screened in 1/4-inch mesh, and one part stock mix. The stock mix is one cup of Tamms Sunset Orange mixed with one sack of portland cement.

From the above formulas it will be seen that for sites in the Southwestern United States generally, about 43 percent by volume each of sand and soil, and 14 percent by volume of portland cement (including various mortar colors) has given the best results. Only through trial and error, and by advance, systematic experimentation with local materials, will the project supervisor determine the optimum mix for a given area.

An unusual adaptation of soil-cement is currently employed on extremely eroded walls at Fort Bowie. Soil-cement in fairly heavy amounts, one to three inches in thickness, is applied to moistened walls. The surface is left rough, coinciding with the texture, the uneven surface, and outline of the original wall. The surface of the soil-cement is then scored with a tool in such manner that the scored lines coincide closely with the covered vertical and horizontal mortar joints. These scored joints are then pointed with a lighter colored soil-cement mortar matching the original. The finished product is a simulated, weathered adobe wall surface, and is a much quicker, more economical method than veneering and capping with stabilized adobes. Treated with a silicone (Daracone), the effective life by this method is 20 years or longer. Similar soil-cement veneers, both formed or hand-applied in Compounds A and B at Casa Grande, have been in place for 17 years. The color there had become somewhat objectionable, due not to the original mix, but to recent patches. The surface was given a bonded mud mortar wash in 1972 which can be repeated at several-year intervals as required.

Aside from making stabilized adobe with cement or bitumuls, there is now a move to use various new chemicals and additives with water and soil to make a strong, weather-resistant brick which has the same color and texture as a plain adobe brick. Only two of the more promising chemicals will be mentioned.

One of the newer soil stabilizers in the plastic resin class is Soil Seal Concentrate, developed in the commercial field as a patented spray to halt erosion along soil or sand banks. It is manufactured by the Soil Seal Corporation, 6311 Rutland Ave., Riverside, Calif. The cost in 1972 was $3.84 per gallon. The firm's literature describes it as a formulation of "balanced copolymers," apparently a methacrylate compound which is formulated as a plastic emulsion miscible in water. Recommended use varies from 5 to 15 per cent Soil Seal mixed with water. Example: a 10 percent solution would be mixed with nine parts water, one part Soil Seal.

Steen reports that in about 1965, Soil Seal was "mixed into adobe and a house was built of the bricks made. The house has stood in Southern California since then. It is unplastered and reportedly shows no sign of deterioration or erosion" (Steen, n. d. p. 8). Both Steen and the staff at Pecos National Monument experimented extensively with Soil Seal in 1969-1970. The following formulas are now being used successfully at Pecos in making stabilized adobes for repair work on the mission and convent. For interior, non-exposed work, such as the partly restored kiva, the following mix is used: three parts sand, three parts soil, and Soil Seal and water mixed in the ratio of 1 to 15 or 20. For exterior, exposed adobes, the Soil-Seal-water ratio is 1 to 10.

Some of these adobes have been in place about two years. During periods of heavy precipitation they become quite soaked. Nevertheless, they display remarkable cohesiveness, dry out rapidly, and have good resiliency during freeze-thaw periods (personal communication, Gary Matlock, supervisory archeologist). A value judgment indicates that it is too early to determine accurately the projected effective life of the plasticized resin in these adobes.

Another chemical which has come into prominence and is undergoing rather intensive experimentation in California and New Mexico is Acryl-60, an acrylic liquid polymer for curing and patching concrete and masonry (Steen, 1970, p.6). Steen also reports that this chemical has been successfully used as a spray to harden and waterproof adobe walls in the California State Park System. Manufactured and sold by Standard Drywall Products, Inc., New Eagle, Pennsylvania, 15067, Acryl-60 sold at $4.68 per gallon in 1972.

Simultaneously with their experiments on Soil Seal Concentrate, Steen and the Pecos staff concluded successful test blocks with Acryl-60 as an additive in making adobes. A section of veneer and capping of adobes now in place toward the top of the north wall of the transept of the 18th-century mission church were made with the following mix: three parts sand, three parts soil, and Acryl-60 and water in the ratio of 1 to 15.

A stronger, more durable brick is being used in the California State Parks (Johnson, n.d.) employing Acryl-60 as an additive to soil-cement with the following formula: one part cement, eight to ten parts soil, and 1 quart of Acryl-60 to each cubic foot of mix by volume. Add this material to the mixing water (one part Acryl-60 to three parts water). Mortar color is added as required.

From the above discussion it will be seen that a perfect solution to the vexing problem of preserving ruined adobe structures has not been discovered. Most of the methods described are long term maintenance or holding actions. If they had not been implemented, however, few of the treated walls described would remain today. While advances in chemical and coating technology can greatly assist in inhibiting deterioration of adobe, it is doubtful that they will be successful in halting it altogether. Solidification and induration of adobe-in-place by electro-chemical means, by infra-red baking, and by methods yet to be discovered, are certainly not beyond the realm of the possible. Highly sophisticated and technical experiments conducted through expensive interdisciplinary research may be required to solve this problem completely.