1. The attraction between spheres, due to surface tension forces, in a lenticular drop between them is calculated for spheres in contact and at increasing separations up to the point of rupture of the drop. Hence the work of separation of the spheres is calculated.

2. Experimental measurements confirm the validity of these calculations, down to very small drop sizes, where it is likely that the failure is in the experimental method. As predicted by Fisher, the force decreases with increasing drop size, while the work of separation increases. Since, however, it is shown that the smaller lenticels are more easily ruptured, no discrimination is provided between the differing explanations of Haines and Fisher of measurements made with the Atterberg apparatus for measurement of soil cohesion.

An experimental verification of the validity of Fisher's calculation of the pressure deficiency inside the drop is also given.

3. It is pointed out that these results depend on the geometry of the system; types of contact other than that of spheres will show different behaviour. Hence generalizations about the behaviour of a real soil, based on an idealized soil of packed spheres, may lead to erroneous conclusions.