A. Wade—Formation of ‘Dreikante’. 397 
(See Text-fig. 3.) This is probably explained by a zone of com- 
pression in the air at this point causing most of the grains to rise a little 
before they would otherwise do so. They apparently glide upwards 
over this zone ina path which is convex towards the wind. The action 
of the wind drives them forward against the face of the stone from 
which they rebound, describing what appear to be parabolic curves. 
The large dreikante shown in Plate XXXII is an excellent example, 
showing beautiful wind-cut curves. 
It seemed to me that from a measurement of these curves one might 
be able to deduce some law with regard to the action taking place. 
I attempted the task, but though I was able to prove that they were 
parabolic, I was unable to obtain any good result, probably. owing to 
the inaccuracy of my methods in transferring the curves on the faces 
of the stones to paper. This proved to be a task requiring very 
delicate instruments. There seemed, however, to be some similarity 
between the curves from the faces of different specimens, and it should 
be possible, I think, to obtain from them a measure of the velocity of 
the wind and the energy expended in erosion. The presence of stones 
and pebbles near to one another cause various modifications of the 
simple case. Side currents, eddies from other stones, all produce these 
effects, and these can be traced on the dreikante under consideration 
when examined in situ. 
But how are we to explain the Brazil-nut shape? The cutting 
away of one face only is not enough to do that. The cutting away, 
however, continues until the pebble stands on a very narrow base in 
almost unstable equilibrium. A little extra wind and it is blown 
over, and the cutting commences again on a fresh side. Evidences of 
this are quite common on the desert. Sometimes at night the ratthng 
of the pebbles moving under the action of a high wind is very con- 
siderable. ‘The five-faced pebble shown in plan in Text-fig. 4 could 
not be explained by the earlier theory. It was originally a partly 
developed dreikante which has been twisted at right angles to its 
original direction and the abrasive action continued. 
Fic. 4. 
Most of the well-developed dreikante show a kind of spiral twist 
in the direction of their long axis. This appears to show that the 
path of a particle in a wind eddy is not a very simple curve. There 
are a good many curiosities in the formation of dreikante which are 
