ARLINGTON HOUSE Historic Structures Report Phase II

PLASTER STUDY

Denver Service Center

ARLINGTON HOUSE

PLASTER ANALYSIS

On November 3, 1980 ten samples of plaster removed from the interior walls of the north wing of Arlington House were sent to the Denver Service Center as part of the fabric analysis conducted in conjunction with the Historic Structure Report. Chemical analysis of the samples was conducted by Greg Koehn, private contractor, utilizing the testing methodology developed by the North Atlantic Regional Preservation Center of the National Park Service in Boston, Massachusetts. Interpretation of the results and report-writing was undertaken by David Arbogast, historical architect, of the Denver Service Center. The purpose of the analysis was to determine approximate dates and types of plastering used in the house. In some locations as many as three sets of brown and skim coats were applied to the walls with paint finishes located on intermediate skim coats, as well as the most recent skim coat. The brown and skim coats were separately analyzed, where possible, because of the considerable differences in composition. Thus, a total of eighteen samples were analyzed.

The samples were located and numbered by Lee Arnest, historical architect, of the Denver Service Center, and Mrs. Mullins, Site Curator. The sample numbers and locations are as follows:

Sample 1A, the most recent skim coat from the Custis chamber, revealed a mixture of eleven parts of sand, by volume, to seven parts of lime, by volume. The relatively high proportion of lime to sand is typical of nineteenth century plaster skim coats. The use of gypsum plaster did not attain general popularity until the twentieth century, thus placing this plaster as possibly dating from the historic period.

Sample 1B. the brown coat associated with the skim coat of sample 1A. displayed a typical nineteenth century mix of roughly five parts of sand to each part of lime, by volume, with horsehair used as a binder. This is a relatively typical nineteenth century plaster mix for brown and scratch coats.

Sample 2A, found below the most recent plaster of the Custis Chamber, contained, in its skim coat, a slightly higher ratio of lime to sand than that found in sample 1A. but not significantly higher to indicate any difference in original specifications between the two. It is possible, because of their similarity, that they may have been applied by the same plaster or specified by the same builder.

Sample 2B, the brown coat associated with the skim coat of sample 2A, exhibited a formulation of approximately three parts of sand, by volume, to each part of lime, with horsehair used as a binder. Again, this is a common type of plaster used for scratch and brown coats throughout the nineteenth century.

Sample 3A, also from the Custis Chamber, showed, through analysis, a mixture roughly comprising five parts of sand to two parts of lime, by volume, with horsehair used as a binder. Its mixture and location in relation to sample 3B single it out as probably being a brown coat applied during the nineteenth century. It is also possible that it may have been a very early skim coat, lacking the high lime content and whiteness of the later skim coats.

Sample 3B, the plaster coat beneath sample 3A, appears to have been a scratch coat of plaster applied during the nineteenth century. Its analysis revealed a mixture of roughly twenty parts of sand, by volume, to three parts of lime, with horsehair used as a binder. The extremely high sand content accounts for the softness of the plaster. Such plaster could only have been used as a scratch coat.

Sample 4, located also in the Custis chamber, revealed a mix with a slightly higher lime content than that seen in sample 3A. For all intents and purposes, however, they are the same plaster with the only exception being that sample 4 lacks horsehair as a binder.

Sample 5A, the skim coat from the chimney breast of the inner hall, (north wing), displayed a mixture of roughly five parts of sand to four parts of lime, by volume, with horsehair used as a binder. With the exceptions of the horsehair content, this plaster is a relatively typical mix for the nineteenth century.

Sample 5B, the brown or scratch coat beneath sample 5A, contained roughly thirteen parts of sand to two parts of lime, by volume, with horsehair used as a binder. This mix is approximately identical to that found in sample 3B.

Sample 6, also from the west wall of the inner hall, (north wing), exhibited a formulation of approximately two parts of sand to each part of lime by volume with horsehair used as a binder. A typical nineteenth century brown or scratch coat mix, this plaster may well date from the original construction of the house.

Sample 7, the only sample removed from the school and sewing room, showed, through analysis, a mixture, by volume, of approximately five parts of sand to each part of lime with horsehair as a binder. This provides a relatively soft mix suitable as a brown or a scratch coat.

Sample 8, located on the south wall of the inner hall, (north wing), appears to have been a standard nineteenth century brown or scratch coat, having approximately six parts of sand to every part of lime, by volume, with horsehair used as a binder. This mix is within the range seen for brown and scratch coats.

Sample 9A, a skim coat in the inner hall, (north wing) revealed a typical mixture of approximately four parts of sand, by volume, to three parts of lime with no horsehair used. This plaster is similar to many of the other skim coats analyzed.

Sample 9B, the brown or scratch coat associated with sample 9A, displayed a composition of equal parts of sand and lime without horsehair. This mix is quite unusual for a brown coat with such a high lime content. The sample analyzed may have been an anomaly.

Sample 10A, also a skim coat from the inner hall, (north wing) contained approximately four parts of sand, by volume, to each part of lime without horsehair. This mix contained an unusually high proportion of sand to lime for a skim coat, not unlike sample 3A.

Sample 10B, the brown or scratch coat associated with sample 10A, exhibited a formulation of five parts of sand to each part of lime, by volume, using horsehair as a binder. This mix is similar to 1B and sample 7.

Sample 11A, yet another skim coat from the inner hall, (north wing), showed, through analysis, a mixture which was highly unusual, containing approximately thirty parts of lime, by volume, to each part of sand. The small sample size, a product of thin layer, leads to statistical problems. It is also possible that the sample may have been an anomaly out of an otherwise consistent plaster having considerably less lime per sand.

Sample 11B, the brown coat beneath sample 11A, provided a much more typical mixture when analyzed. Its formula was almost the same as that seen in samples 1B, 7, and 10B.

When plotted as a statistical chart, the plaster samples provide an interesting range. The chart shows a clear difference between mixes used generally for skim coats and those used for brown coats. The chart is on the following page.

From the above chart it can be seen that the brown and scratch coats of plaster tend to be grouped near the upper right side of the chart, indicating a generally higher proportion of sand to lime than the skim coats. In addition, the horsehair tends to be associated with those samples at the right upper side, indicating that the skim coats did not employ horsehair as a binding agent as a rule.

In all cases, the sands and fines resulting from the analysis were virtually identical, indicating similar material sources for all of the samples. The fines can be seen as being dirt associated with the original sand and not as an additive, as is the case with pigmented or Portland cement mortars.

These samples provide a reasonable indication of the ages of the various layers of plaster in the north wing. Comparison of these results with written historical records is suggested. Further analysis might also be conducted with samples from the center block and the south wing, as well as the mortar used in the brick masonry. Often similar, if not identical, mixtures were used in mortar and plaster alike.