a22 MR. P. F, KENDALL ON A SYSTEM OF. [Aug. 1go2? 
flow behind the spur is of considerable magnitude. This channel 
probably commenced to be eroded at a little above 825 feet, and 
continued to cut down almost to the 800-foot contour, though the 
valley now called the Peat Holes is so deeply filled with peat as 
to conceal the 800-foot contour where it passes. A small projection 
of the hillside known as Middle Heads, near the Ewe-Crag-Slack 
intake, is trenched by a small parallel series of marginal channels. 
The upper one is a little above the 825-foot contour, the next 
between 800 and 825 feet, and the lowest, perhaps the best marked, 
runs a little above the 775-foot contour. These belong to the 
aligned sequence falling to Ewe-Crag Slack; and it is interesting 
to observe that the single large channel at the Peat Holes is the 
equivalent in time of these three small overflows. The lowest 
overflow is at a higher altitude than Ewe-Crag Slack. A period 
of somewhat rapid retreat of the ice-margin seems to have ensued, 
and this area was, by means of a series of marginal channels, 
placed in communication with-a great direct overflow at Stonegate, 
4 miles to the eastward. Before considering this, I will briefly 
mention a temporary direct overflow that possesses some features of 
especial interest. 
On the Eskdale side of the watershed near Doubting Castle, just 
1 mile east of Ewe Crag, a small patch of gravel occurs exactly 
on the 725-foot contour. This gravel is well-bedded, and contains 
foreign stones, such as porphyrite, granite, flint, and basalt. Its 
position and contents alike render it probable that it has washed. 
over the watershed, and it is seen that it lies in a small valley which 
rises to a depression in the watershed on Girrick Moor, only 760 feet 
above sea-level. The actual passage through is a broad swampy 
tract marked at Danby Peat-Pits. J have probed the peat at many 
places up through the depression, and found about 8 feet of peat 
resting on gravel. On the outer face of the hills, a thin covering 
of Drift extends quite to the watershed. The explanation of these 
phenomena appears to be that, at the maximum extension of the ice, 
one lobe of its very much indented front reached for a short time 
far enough to impound water in a small recess of the hills. The 
overflow passed over into Eskdale, and the gravel which it carried 
was arrested at the level of Lake Eskdale (725 feet), the exact — 
height at which the Murk-Mire-Moor overflow com- 
menced its action. This would indicate that the maximum 
extension of the ice at Girrick synchronized with its farthest advance 
into the Murk-Esk Valley : such coincidences are, I conceive, neither 
necessary, nor usual (see fig. 15, p. 520). 
The Stonegate Direct Overflow and its Related Phenomena. 
I now come to one of the most complicated and difficuit 
problems that has yet engaged my attention. I have already 
stated that the Cleveland watershed was overridden by ice from 
the seaward end as far as Stonegate, and the overriding mass 
not only pushed for 13 miles up the Esk Valley and formed a 
