is,3 Smith: Tropical Geology and Engineering 231 
(103 miles) an hour and the roads are swept literally by a 
natural cyclone blower and all loose material is removed as ef- 
fectually as if a vacuum cleaner had gone over them. So we must 
consider not only the work of the deluge of rain but the com- 
bined effect of water and driving wind. Road maintenance from 
these and other causes is a very serious problem, in the Phil- 
ippines at least, and any criticism of our highways should be 
tempered by a consideration of the difficulties we have to over- 
come. 
Two chief sources of trouble in concrete mixtures in the Phil- 
ippines are found ; first, the fineness and generally poor 
quality of the sand, and second, lack of strength of the crushed 
rock. Both Reibling and King, of the Bureau of Science, have 
repeatedly called attention to these weaknesses in local concrete, 
as revealed by their long series of actual tests made on thou- 
sands of samples. The ultimate cause for this condition of 
things is found in the mineralogy and geology of the raw mate- 
rials employed. Not only is the mineral composition of the sand 
commonly used here unsatisfactory, but also the extremely 
weathered condition of the materials is a matter of menace. 
Now it is agreed that the practical tests are of prime importance 
in the proving of such materials, but the examination of the sand 
and of thin sections of the rock is also of very great value. Such 
examination will reveal the cause for the failure ; and, of course, 
this is really the most important thing of all, if we want to 
avoid future trouble of this kind. 
The so-called “sand” oftenest used here is sand only by 
courtesy. The microscope reveals little or no quartz, the feld- 
spar present is often in a far-gone condition of decay, and the 
predominant minerals are often dark-colored hornblendes, py- 
roxene, olivines, etc. To one accustomed to the clear, white, 
sharp Ottawa sand used in standard tests, the local sand appears 
more like dirt. 
Reverting to the subject of landslides, a thorough acquaintance 
with tropical conditions will reveal at once the futility of at- 
tempting to control slides like those of the Culebra cut at Pan- 
ama. Retaining walls, drainage, etc., are all makeshifts. The 
only practical and final solution of the problem is to increase 
the sliding until the walls are brought to the angle of repose 
where they will no longer slide. . According to MacDonald, geo- 
logist of the Canal Commission, the most important types of 
slides were structural breaks and deformations and for this type 
of slide there was only one remedy that had utilitarian value 
