701 



duration of the stationary state near the limits 

 is almost null, as at Cum ana, the changes are 

 announced when the limit is attained, by a 

 change in the convexity of the column of mer- 

 cury. The variations appear to be independent 

 of those of the temperature and the seasons. If 

 the mercury was descending from 2 U till 4 h , or 

 rising from 4 h till ll h , a violent storm, an earth- 

 quake, showers, and the most impetuous winds, 

 would not alter its movement ; which nothing 

 appears to determine but the real time, or the 

 position of the sun. The regularity of the va- 

 riations was constant in the rainy season, both 

 in the thick forests of the Atabapo, and on the 

 table-land of Pasto (1600 toises) and of Mex- 

 ico. When the duration of the stationary state 

 was prolonged, it was most frequently at 4 h in 

 the afternoon, and from 4 h till 9 h in the morn- 

 ing. At Lima, the maximum of the evening 

 oscillated from 9i h to lU h . The observations 

 I made at 4 h in the morning are, unfortunately, 

 the least numerous. The only place where 

 during the course of my voyage I remarked a 

 great deviation, is the town of Quito, situated 

 in a narrow valley, and close to the volcano of 

 Pichincha. I could only make observations in 

 this valley during the months of January, Fe- 

 bruary, and March, where the maximum, in 

 very variable and rainy weather, was rather 

 near noon, than at 9 h in the morning, and where 



