CHAP. xir. TEE QUE BR AD A OF GUALLABAMBA. 219 



slopes are not steep, the abruptness of its cliffs bordering the 

 quebrada can hardly be exceeded ; and there is nothing elsewhere 

 in the neighbourhood of Equatorial America equalling the grandeur 

 of this profound earthquake fissure.^ Just where the ground 

 commenced to fall steeply, I halted to examine the barometers, 

 for the purpose of determining the depth of the ravine, and for 

 reading the aneroids against the mercurial ; as this was a favour- 

 able occasion for comparing the indications of the two classes 

 of barometers. 



No reference having been made to the aneroids since p. 72, 

 it may be supposed that they were put aside, and were neglected. 

 This was not the case. Systematic comparison of the barometers 

 was part of my daily routine, though regarded almost as waste 

 of time ; for it was difficult to see what advantage could be 

 derived from employing instruments which all read lower than 

 the truth, and differed to a large extent one from another. 

 The comjDarisons which were made since we left Ohimborazo 

 shewed that the index-errors of all the aneroids remained nearly 

 constant at any given pressure ; and had a tendency to augment 

 while ascending (that is, with pressure diminishing) and to 

 lessen whilst descending (pressure increasing).^ At Quito, on 

 the 20th of March, the mean error of the whole amounted to 

 — 1*009 inches ; that is to say, the mean of the whole of the 

 aneroids indicated a pressure more than an inch too lo'w at 

 Quito. If this mean had been employed for determination 

 of altitude, in conjunction with the Guayaquil observations, it 

 would have made the height of the Capital above the level of 

 the sea about 1,400 feet in excess of the truth. 



Although it appeared to me that these aneroids were worth- 

 less for determination of elevation above the level of the sea, 

 I had already remarked that their indications often accorded 



1 A few remarks upon this and other quebradas are made in a later Chapter, 



2 See the Table at the end of Appendix C. giving the mean error of the 

 aneroids. 



