Mr. J. Ball on the Structure of Glaciers. 495 
rangement of dirt and debris in irregular curved streaks, much 
narrower than the “ dirt-bands,’ and not regularly disposed 
across the whole width of the glacier. These probably corre- 
sponded with the phenomena observed on the Rhone Glacier and 
that of Grindelwald by Messrs. Tyndall and Huxley, and subse- 
quently explained in their paper in the Philosophical Transactions. 
4, Dirt-Bands of M. Agassiz.—The language of M. Agassiz 
and his friends, in treating of the appearances which they sup- 
posed identical with the dirt-bands of Professor Forbes, is not 
always so clear and consistent as to enable the readers of the 
‘ Systéme Glaciaire’ to recognize with certainty the phenomena . 
which they describe. This is evidently due to the want of ac- 
curate principles of physical science on the part of the author 
and his fellow-labourers. But, as a record of observations, it 
seems to me that thew work contains much information that 
cannot properly be neglected, and amongst other facts, the 
system of dirt-lines, which Agassiz regards as evidence of 
the stratification of the glacier, deserves, I think, further care- 
ful examination and discussion. I have frequently remarked 
the recurrence at pretty uniform intervals of broader and more 
strongly marked blue bands in the veined structure, sometimes 
accompanied by the appearance of a thin stratum of dirt-frag- 
-ments which may be traced along the walls of intersecting cre- 
vasses, and into the substance of the glacier. These are the 
lines which M. Agassiz has identified with the dirt-bands; and 
to these, if I am not mistaken, most of the statements in his 
writings apply. But at times it would appear as if the observa- 
tions recorded had reference to two completely distinct pheeno- 
mena, of less general occurrence. One of these is the existence 
of lines indicating closed crevasses, of which the lips often retain 
for a considerable time enough of dirt to mark their presence on 
the glacier. When seen from a moderate height above the gla- 
cier, where their direction can be compared with that of existing 
crevasses, these are easily recognized, but on its surface they 
may readily be confounded with the class of dirt-lines now under 
consideration. The other phenomenon, which has, as I believe, 
been a source of confusion, is peculiar to the upper region of 
great glaciers. The annual fall of snow, and the corresponding 
destruction of it by heat and evaporation, are, as is well known, 
very variable in amount. New beds of névé, and even small 
glaciers, are sometimes formed on the lower mountains, and dis- 
appear again in subsequent seasons. The same thing happens 
on the surface of great glaciers. After a layer of dirt has been 
_ deposited on a part of the glacier which has been exposed for 
one, or perhaps for several successive seasons, it may easily 
_ become covered over by beds of névé, which may finally, if they 
