R.L. Patty

Bulletin No.


Document Type



Department of Agricultural Engineering


Perhaps the most important question in connection with pisé or rammed earth type of wall construction is the identification of favorable soils for this purpose. One who is inexperienced in this type of construction invariably will choose the wrong kind of soil for it. As a stiff clay soil will become hard when dry it is naturally supposed that a clay soil would be favorable for rammed earth. This is not the case. In fact, only a small amount of clay is desirable while plenty of.sand or mineral aggregate is excellent. A test of the soil for sand content will often be sufficient for determining whether a soil is satisfactory or not, but not always. All sandy soils have been found favorable but probably 20 per cent of favorable soils will have a low sand content and will therefore not be identified by the simple test for sand. These are soils which contain a high silt content and are comparatively low in clay. In no instance has a soil with a high clay content been found favorable for the purpose. In an effort to find a reliable method for accurately identifying favorable soils for rammed earth construction the South Dakota Experiment Station built test walls for determining the resistance of various types of soils to weathering. The soil used in these walls was then analyzed for total sand, total colloids, total clay, and total silt. After standing for four to five years the quality of the walls was rated or graded in per cent. Comparisons were then made of the resistance of the various walls to weathering and the total sand in each; the resistance of the walls to weathering and the total colloids in each; the resistance to weathering and the total clay in each, and the resistance to weathering and the total silt in each. A fairly close relationship was found between the resistance to weathering and the total sand in the soil but there were exceptions as indicated above. A very close relationship was found between the resistance of the walls to weathering and the amount of colloids in the soil. In fact since no appreciable exceptions were found to this relationship the results of the study seem to justify the statement that any natural soil can be identified definitely as favorable or unfavorable for pisé construction if its colloidal content is known. In view of the fact that the total colloids in a soil can now be determined in less than an hour's time after the sample is dried to constant weight, this is a very significant fact.


rammed earth walls, pise, soil colloids



Publication Date









South Dakota Experiment Station, South Dakota State College of Agriculture and Mechanical Arts