112 PROCEEDINGS OF THE GEOLOGICAL SOCIETY. [DcC. 5, 



roid f . The two former do not give me the heights of the heads of 

 the valleys, but only those of the village nearest to its head, and this 

 is usually rather below the level of the head of the valley. Another 

 source of error arises from the valleys being always more or less 

 choked up at their upper ends by detritus brought down from the 

 mountains by the torrents, and of which we cannot see the depth. 

 Judging from what I could observe of such deposits, I think it likely 

 that they do not often exceed 200 or 300 feet in such situations. 

 Fortunately these two errors are in opposite directions, and therefore 

 tend to neutralize one another. From these and other causes I con- 

 sider that the altitudes here given can never be relied on within 100 

 feet ; but they probably are all within 200 feet of the truth. 



The highest valleys which contain glaciers do not admit of so near 

 an approach to accuracy : we have no means of knowing the depth 

 of the ice of the glacier, and can therefore only form a rough guess 

 of the level to which the head of the valley has been excavated. 



The Table No. I. (p. 123) exhibits the heights above the sea of the 

 heads of a few valleys which I measured, and of a number of villages 

 which stand close to the heads of their respective valleys or just 

 below a spot where the valley exhibits a sudden change of level. 

 Where several names occur on one line, those places are on the same 

 line of drainage : where the village stands considerably below the 

 head of the valley, it is marked thus *, to show that the height there 

 given is below that of the head of the valley. All the measures are 

 reduced to English feet. 



The following levels may be reduced from the Table No. I. by 

 throwing together those heights which approximate most closely. 



As I have no accurate measurements of the heads of the highest 

 valleys, owing to their being filled with ice, they are not inserted in 

 the Table ; but I may repeat that the branches of the Mer de Glace 

 end at above 9000 feet, and probably correspond at their highest 

 level with the highest line of erosion observed. 



1. The Piedmontese Val Ferret ends at 7660 feet : I have not the 

 exact height of the head of the AUee Blanche, but have reason to 

 believe it to be about the same. This height nearly corresponds to 

 that of the '* second line of erosion," noted at 7500 feet. 



t The aneroid barometer is a treacherous guide, owing to the frequency with 

 which the hand is displaced by a jolt, and it cannot be trusted alone : but it will 

 answer all the purposes of a geologist in a tour where he has frequent opportuni- 

 ties of comparing it with a mercury barometer, or of checking the accuracy of its 

 index in passing spots of which the heights are known. As there is hardly a hill 

 or a village in Switzerland of which the height above the sea is not published, 

 these opportunities are offered to the tourist several times every day; and, if he 

 finds that no displacement of the hand has taken place, he may rely on his inter- 

 mediate observations to within 10 feet. The contemporary changes of the atmo- 

 spheric pressure may be traced from the meteorological observations at Geneva and 

 at the Great Saint Bernard, which are published monthly in the " Bibliotheque 

 Universelle de Geneve," and from the recorded observations of Prof. Wolf of 

 Berne, which, though not published, are, through that gentleman's kindnes§, 

 equally accessible. 



