H. A. Tempany 313 



soil can readily be moulded but at the same time does not adhere unduly 

 to the hands or instruments used in the course of the preparation of 

 the sample. Methods of standardising plasticity in the case of certain 

 soils were described, consisting in the addition of measured quantities 

 of water to the soil accompanied by kneading, as in the case of the 

 titration of flour to ascertain the quantity of water required to prepare 

 a standard dough, and also by means of the determination of the weight 

 required to crush a block of soil when at the correct consistency. 



The method of measurement consisted in preparing bricks of the 

 kneaded soil of standard dimensions, the insertion therein of two marks 

 at a known distance apart and the measurement of the distance between 

 the marks after the bricks had dried. The difference between the 

 original and the final measurement was expressed as a percentage of 

 the original distance between the two marks and gives the coefficient of 

 linear shrinkage in the case of the sample under examination. 



Comparison of the results of such measurements with the condition 

 of growing cacao on a number of soils in Dominica and St Lucia showed 

 that over a range of 34 examples a very fairly close correlation could 

 be observed, and that under the conditions of climate prevailing in 

 these islands, when the shrinkage of the surface soil exceeded 10 per 

 cent, and that of the subsoil exceeded 12 per cent, the conditions were 

 unfavourable 16 cacao production. 



In relation to the actual magnitude of the effect observed, over a 

 large range of soil samples from different parts of the West Indies, the 

 linear shrinkage has been found to vary from 2-0 per cent, in the case 

 of a very open sandy soil to 16 per cent, in the case of an exceptionally 

 heavy clay soil. 



In a note appended to the paper in question the probable relationship 



between gel formation in the case of the soil colloids and the shrinkage 



observed, was briefly discussed by the present writer and attention 



directed to the approximate interdependence existing between the 



magnitude of the shrinkage observed in any case and the proportion 



of particles less than -01 mm. in diameter present in the sample, 



,. percentage' of particles less than -01 mm. ,■.-,■ , 



the ratio / , . . ^ ^ exhibitmg a rough 



per cent, shrinkage observed 



degree of constancy. 



Since the publication of the original paper the method has been 



employed fairly extensively by experiment station workers in the 



West Indies as a simple and convenient means of gaining a certain 



degree of information concerning the physical character of soils and 



