THE AUSTRALIAN NATURALIST. 1 iT 
phytic growths on these rocks, and the presence of large trees 
growing in the debris, show that in normal years erosion almost 
ceases. In the great floods, which impress themselves on human 
memory from their very rareness, the trees will be uprooted and 
the ferns, et¢., scoured away. It is at such times that the ob- 
structions are most effectively eroded, though the converse action 
may happen, too, as the obstruction may serve to accumulate 
a great pile of debris of all kinds. 
Continuing upstream, we come to a very wild gully with great 
dark caverns on the right and on the left, and we can see the 
stream at work undercutting the sandstone walls. We have 
reached a soft shaly bed in the sandstone, and the wildness of 
the scene is a testimony to the activity of the erosion. 
We soon come to the “Cathedral,” where the narrow gully 
with precipitous walls gives an effect which strongly suggests 
a Gothic nave, especially as the cliffs overhang on both sides, 
owing to the superior resisting power of a hard bed of rock 
which caps the cliffs. The track leaves the stream bed at the 
entrance to the Cathedral and climbs up a small tributary valley 
until it reaches the level of a soft stratum which has been much 
eroded. This shaly bed is easily traced for the next half-mile, 
as the track continues at this horizon although the river is LOOft. 
below. ‘This section is often referred to by the misleading title 
—the Underground River. The narrow chasm is cut through 
alternate strata of hard and soft rock, the harder beds remaining 
bare and little altered since the time when the gorge was cut; 
whilst the softer beds are fretted into a series of long caves, in 
which many ferns grow with luxuriant freshness. 
The method whereby the water was able to cut a deep gorge 
through this resistant rock is clearly seen, for the hard walls 
of the underground river are a continuous series of giant pot- 
holes, which make by their summation a long narrow gorge. 
It will be observed that in the places where there are three 
terraces of hard rock, the potholes in each terrace are inde- 
pendent, showing that each layer was cut through independently 
by a small waterfall, which having cut a cylindrical hole through 
one hard stratum rapidly excavated the softer bed below, so that 
the hard rock was undereut and eventually fell. In this way 
the waterfall cut back and started to make a new pothole. 
This action can be seen in progress at a small fall near Neate’s 
Glen; but the consecutive cutting of the hard beds can be best 
realised as we walk up the underground river. At first there 
