532 The Modern Cottage: Experiments in Pise. [Sept., 



Organic components are bad, being particularly liable to shrinking 

 in drying, and they weaken coherence in the earth as a whole. 

 Top or vegetable soils are, therefore, generally unsuitable. 



An all-important point is the amount of water present in 

 the earth at the time of actual use. This factor is so dependent 

 on the composition and character of the soil to be used that a 

 statement of a percentage can only be taken as a very rough 

 guide. In general it will be found that the water present 

 should be between 7 per cent, and 14 per cent, of the weight of 

 the dried earth. In illustration of this factor of water content 

 it may be stated that a clay-gravel-sand mixture gave the best 

 results with as much as 15 per cent, of water ; while a chalk 

 loam mixture with 13 per cent, of water shrank excessively on 

 drying, but worked satisfactorily with 7 per cent. The tendency 

 of inexperience is to work with too damp a soil. As long as the 

 rammer will consolidate without pulverising the material it may 

 be assumed to be wet enough. Should it " pug " under the 

 rammer the earth is certainly too wet and will shrink and crack 

 when drying out. A good average sample of the earth should 

 be taken for testing, sufficient to make a number of test blocks ; 

 and each block should be made up at a different stage of the 

 earth's drying, so that the most satisfactory water content may 

 be determined by the tests. 



The percentage of moisture is most easily found by 

 weighing the samples as made, and again when they are quite 

 dry ; the difference in weight, taken as a percentage of the dry 

 weight, giving the figure required. Exact measurements of the 

 length of the sample should also be made, both at the time of 

 consolidation and when dry. The percentage of shrinkage 

 should in no case exceed 2^ per cent, to 3 per cent., and can 

 generally be kept under 2 per cent when the water content is 

 low. 



(b) Shuttering. — The shuttering required should be capable of 

 resisting the considerable thrust of the ramming; ensuring a 

 true face to the wall ; be easily and speedily placed in position 

 and moved as the work proceeds without support from the 

 ground; and be economical in cost. The Ministry's technical 

 officers have devised a set of shuttering rather different from 

 that designed by Mr. Williams-Ellis and used for the Merrow 

 Bungalow. This appears to meet the above requirements, has 

 proved satisfactory in use, and is a pattern which has been 

 adopted by many private owners who are beginning to build in 

 pise de terre. 



(c) Best Form of Rammer. — Iron, with a smooth surface, is a more 

 satisfactory material for a rammer than wood. Two forms of 

 rammer, each weighing about 7 lb., are desirable, one flat, the 

 other shaped like a heart, tapering in thickness to the bottom, 

 which should have a blunt edge. 



(p) External Rendering. — Various materials for rendering have been 

 tried on pieces of experimental wall at Amesbury, but no definite 



