1887 .] Mr A. C. Elliott on Formula for Retaining Wall. 87 
Rankine uses the term granular mass to indicate that the earth 
is assumed to have no tenacity or cohesion. Some kinds of material 
might fairly claim to he so described, but in ordinary practice it is 
quite common to meet with earth which will stand with a vertical 
face for a considerable time ; but since the action of time and 
weather will inevitably result in the material ultimately assuming 
a slope more or less constant for that particular kind, it is not only 
prudent but necessary to allow for the almost total failure of tenacity 
or cohesive strength with lapse of time. In the retaining wall 
problem the effect of assuming any degree of tenacity being clearly 
operative in reducing the resultant force representing the mutual 
action of the earth mass and the wall, Rankine makes no scruple 
to discarding tenacity altogether. 
It may be remarked that Coulomb’s method implicitly takes 
account of some degree of tenacity. Nearly all direct experiments 
have shown considerable divergence between the actual overturning 
moment of the earth pressure on the wall, and that calculated by 
Coulomb’s or Rankine’s method (employing the accepted methods 
for determining the principal constant), the divergence very com- 
monly amounting to upwards of 50 per cent, in excess of the 
observed value. The discrepancy is, in the' author’s opinion, in 
great part clearly due to the ignoring in the mathematical investi- 
gation of the effect of tenacity. On the ground that no satisfactory 
allowance could be made on account of a quantity which is at once 
a function of time and weather, Rankine, as has been already re- 
marked, expressly rejects the tenacity from consideration, so that, in 
the case of Rankine’s formula at least, one ought not to be surprised 
if the calculated should exceed the actual overturning moment of 
the earth pressure to a considerable extent. 
But granting this, or at any rate taking leave of it, there still 
remains the experimental evidence that Coulomb’s method in its 
complete form, though much more unsatisfactory from a physical 
point of view, gives better results than Rankine’s. Attention has 
consequently been redirected to Rankine’s method with the object 
if possible of removing this anomaly ; and, accordingly, it has been 
pointed out that Rankine simply applies the conditions of equi- 
librium, obtaining at a point in the interior of the earth mass, to a 
point situated in the surface of separation, between the wall and 
