106 J. F, CAMPBELL ON GLACIAL PERIODS. 
All exposed rock-surfaces are discoloured and partially decomposed 
to a considerable depth by the action of frost, heat, rain, and air. 
These weathered surfaces are scaling off and crumbling. The hill- 
sides are strewed with large blocks of gneiss, which are weathering, 
scaling, and crumbling, like the solid rocks, in situ, where they fell, 
or rolled, or were slowly weathered out of softer rock. The scales, 
which weather off and crumble, decompose till crystals of felspar, 
quartz, and mica, &c. separate and break up. The felspar turns to 
kaolin, the quartz to sand, the colouring-matter of mica, hornblende, 
&e. turns to rust. Rains sort and pack these decomposed minerals, 
wash them down hill, and bury gneiss blocks in sheets of kaolin, in 
sheets of gravel, in coloured clays, and in “‘laterite.” The mate- 
rials of the decomposed gneiss, packed in sheets on the hill-sides, 
cover the slopes. ‘The result is that some of the higher grounds are 
like glacial work. As in Ceylon, I am satisfied that these hill forms 
are due to weathering wm situ. The lower slopes, where rains have 
space to gather into larger streams, are deeply furrowed by long V- 
shaped gorges, in whose steep sides the structure of the gneiss is well 
seen. These greater furrows clearly are records of “ pluvial erosion.” 
At Newera Ellya, in Ceylon, and at Ootacomund in the Nilghiris, 
artificial lakes have been made by building walls across water-fur- 
rows. ‘The plan of a lake so made is very different from that of a 
mountain-tarn in a glaciated country. The water-line is angular, 
not curved. I could find nothing glacial between Madras and the 
highest point in Southern India which is within sight of the western 
sea. Two experts well acquainted with the country, the Curator of 
the Madras Museum and the officer who engineered most of the hill- 
roads, confirmed my opinion. No marks of glacial action in this 
region are known to them. J heard much of glacial action in the 
country. I could find none; butI can understand mistaking “late- 
rite” for “‘ boulder-clay,” and some weathered surfaces for ‘“ roches 
moutonnées.” 
23. On the 8th, 9th, and 10th of March, 1877, I crossed from 
Madras to Bombay. The Madras side is a plam. Where the 
hills begin, steep slopes are furrowed by rains and streams. The 
structure of the rock was clearly seen in conical hills and in long 
ridges. JI took the rock to be gneiss. ‘The upper plateau is an un- 
dulating country with isolated hills of granite in all stages of wea- 
thering. The last stage is a cairn, or ridge of loose stones, or, where 
these have been crumbled to dust, a low dome or boss of granite. 
At first sight many of these forms are like glacial work; but after 
passing many it isclearly seen that the work is “ weathering in situ.” 
The dome shape is the structure of the rock. The surface is not 
worn by the grinding of ice ; for it often passes beneath upper layers 
in all stages of weathering, from cracked surfaces to isolated stones. 
The isolated stones are not “‘erratics”; they are but remnants of an 
upper layer weathered away by rains. About midway between the 
coasts, near the branch railway to Hyderabad, trap is reached, and 
the form of the country changes. The Ghat on the west side gives 
a section of about 2000 feet of beds of trap. Above and below the 
