Himavava Mountains. 235 
-uncertainty* has been occasioned, small it is true, but still greater than 
need lave biegn. My first idea was to attach to the wooden rods, thin 
iron or brass slips, either of an. equal length or something shorter—hy 
means of which, and a thermometer, it might be seen whether and how 
much the wood had been affected in length. The modified plan was to 
‘construct a machine, which I afterwards distinguished by the name of a 
comparator, and by means of which, I thought the changes which the wood 
* might undergo, would be detected with as much certainty as those in metal, 
by means of the thermometer. In forming this judgment I overlooked 
however a very essential difference—the homogeneity of the metal, and the 
want of that quality in the wood, which circumstance causes so much un- 
certainty, that judging from experience I would say, that no two pieces of 
wood will lengthen and contract in the same manner and degree for any 
length of time. Fig. 7, (Plate II.) gives a view uf this comparator, as 
finished, and figs. 8 and 9, explain certain parts referred to, in the follow- 
ing description. | 
13. Ir consists of a frame of wood, supported on four legs, strengthened 
by cross pieces, so that in lifting, no alteration of figure takes place.. To 
this frame is screwed a wooden piece eight feet in length, and of the same 
thickness and breadth of the measuring rods, represented by figs. 8 and 9. 
To it is attached, about an inch above it, a brass cylindrical rod of the same 
length, by brass rings which screw into it. To the last ring marked a, the 
a 
* Apour two feet. 
VOL. XIV. 30 
