JOWETT AND MUFF : GLACIATION OF BRADFORD, ETC. 243 
Before leaving the subject of overflow channels, an anomalous 
dry valle}^ exhibiting several of the characteristics of the overflow 
channels, should be mentioned. This valley is situated on the 
south side of the River Worth, a mile and a half due west of 
Haworth, and a few hundred yards west-north- west of Stan- 
bury. It differs from the normal overflow channel in that it 
does not trench through a watershed, but occurs near the bottom 
of a large valley not many feet above the modern stream. It 
is a curved valley cut into the slope of the hill, and it possesses 
steep sides with a broad floor sloping to the east. The altitude 
of its floor at the intake is 705 feet above O.D. (see PL XIX., 
Fig. 2). There is not the slightest trace in the main valley of any 
drift dam by which the River Worth might have been obstructed 
and forced to cut a new channel for itself, as the River Aire 
was compelled to do by the Nab Wood moraine (p. 203). The 
valley has many characters in common with Prof. Kendall's 
type of " in-and-out " channels, and the only suggestion 
that we can offer as to its origin is that it was caused by 
the water flowing round a lobe of ice (or perhaps a mass of ice 
isolated hy irregular melting) which lay upon the north slope 
of the Worth Valley, and obstructed the water flowing from 
the higher parts of the vaUey at this place. 
We have no knowledge as to the manner in which a glacier 
would disappear from a country consisting of a series of rather 
narrow valleys separated by high ridges, which directly cross 
the path of the ice. The complex series of overflow channels 
described above show clearly that the retreat in Airedale was 
intermittent, and that during its progress slight readvances 
took place. Within the area which was once ice-covered, there 
is scarcely sufficient evidence to allow of the marking of the 
position of the ice-edge throughout the length of the 
dale during the various phases of retreat. The positions of 
the overflow channels are determined as much by the shape 
of the ridge through which they are cut as by the nature of 
the retreat of the ice. A favourably- situated gap like that of 
Sugden End (p. 234) persisted, and was deeply excavated during 
the cutting of several valleys at different heights on the ridges 
to the east and west of it. The highest dry gaps, however. 
