552 Proceedings of the Royal Society 
of a mile in average breadth. A rainy day followed, and then four 
days of uninterrupted dry weather, during which the stream 
returned nearly to the same state in volume and appearance as 
after the moderate flood already described. There was this differ- 
ence, however, even in its composition ; nitrate of silver feebly 
indicated chlorides, and acetate of lead also feebly indicated car- 
bonates. The difference wss probably owing to a material differ- 
ence in the direction and force of the wind. On the former occasion 
the wind blew from the north-east, with no great violence, over 
about 90 miles of land ; but on the latter occasion it blew with 
fury from west to south-west over Loch Fyne at distances varying 
from 18 to 15 miles only. In the latter case sea-spray must have 
been swept up into the air and carried far by the storm. In the 
former less would be raised into the atmosphere, and much would 
be deposited again in passing over 90 miles of land. In 1845 I 
found chlorides distinctly indicated by a white cloudiness, when 
nitrate of silver was added to rain-water collected on the top of 
Goat-Fell in Arran, towards the close of a violent four days’ south- 
westerly gale, attended with frequent heavy rain, the sea in the 
direction of the wind being 12 miles distant, and 2800 feet 
below. 
The facts now stated, which I have often corroborated by less 
minute observation of other streams in the mica-slate district of 
Loch Long, Loch Goil, and Loch Lomond, will convey some idea 
of the constitution of these waters in three conditions, viz., after 
high floods, moderate floods, and dry weather. To complete the 
series, it is an object of interest to add their condition after very 
prolonged drought. In that case the streamlets, except those fed 
by small upland “ tarns,” will come at last to convey only the 
water proceeding from springs ; and many not so supplied will dry 
up altogether. For the composition of those which continue to 
run we may look to the springs themselves which feed them, because 
in their then very low state, running chiefly over rocks and stones, 
their waters will contract little additional impregnation in their 
course downwards. I have examined several springs in the mica- 
slate district under consideration. They have generally presented 
rather more saline constituents than the streams in their ordinary 
state, and invariably no colour appreciable by any of the ocular 
