128 EXCURSIONS IN MADEIRA 



the cavern at Praya Formosa, close to the sea, about three miles 

 west of Funchal, and upwards of thirty feet deep, I found a 

 difference of only 3° F. and in that of St. Roque, about 1000 feet 

 above the sea, north of Funchal, and nearly sixty feet deep, a 

 difference of only 4° F. 



There is an extraordinary difference, however, between the tem- 

 perature of the wells and that of the air of Funchal ; the former 

 (Mr. Lundie's, Mr. Young's, and Mr. Serle's, all upwards of 

 twenty feet deep and in the open air) being 58° when the latter 

 was 69° — but this is explained by recollecting, that these weUs are 

 supplied by streams which descend from heights of 3800 feet 

 behind the town, where there would be a corresponding difference 

 in the mean temperature ; for that of the spring near the ]\Iount 

 Church, (enclosed at the expense of Consul Murray) and about 

 1900 feet above the wells in question, was 58", the air within 

 being 62°, by an observation which I made in October. The 

 observations on the Peak of Ruivo gave eighty-nine toizes to a 

 centigrade degree, for the decrease of temperature ; those on the 

 brink of the Coural, at the point of view from which the drawing 

 was made, ninety-five toises ; those at the Mount Church, ninety- 

 eight toises; the two latter results are probably in excess, from 

 the elevations being backed by greater height, and not at all 

 insulated. I have calculated in toises, merely because the results 

 of De Humboldt's and De Saussure's observations are given in this 

 measure. When at Arieiro, a cold north wind came on to blow 

 suddenly, and lowered the thermometer so considerably for the 

 time, that I could make no conclusions ; and the locality of 

 Mr. Veitch's quinta is such, from the torrents adjoining it on each 

 side, that it is much colder, as a particular spot, than the country 

 about it, and at the same height above the sea. November seems 

 to be the month whose mean temperature is the closest approxi- 

 mation to that of the year in these regions. 



