METEOROLOGY. 213 



Tables No. 2 a and 2 b contain half-hourly observations on the 

 daily periodical variation of the barometer for 31 days, made with a 

 view to determine the precise time of maximums and miniraums and 

 ;the amount of daily am|:)litu(le. With regard to the latter, if we take 

 the mean of every six days in successit>n, beginning with the 10th of 

 'May, we get the following mean amplitudes: .OfiO, .070, .068, .066, 

 '.064; showing a gradual rise and fall in the numbers. The greatest 

 jraean amplitude is for the period from 16th to 2l8t May, so that even 

 ithese additional numbers are still in accordance with the view taken 

 iwithregaid to the amount of amplitude tor the different periods of 

 ithe .vear, allided to in my last letter. All the half hourly observa- 

 [tions u[) to June 9, inclusive, hitherto made by me on the subject of 

 i periodical variation, whicli are for 56 days, prove for the occurrence of 

 I the a. m. maximum, the average time to be at IQh. lO^m. a. m., and for 

 Ithe J), ra. minimum 4/i. Z\^m. p. m., which seems to agree pretty well 

 with the time of daily maximums and minimums found in other parts 

 of tlie globe. 



Table No. 3 is to exhibit the cumber of hours of rain during the 

 diftt'rent times of the day for each month from July, 1856, to May, 

 1857, recapitulated from tables No. 6 e and/. The vertical distance 

 of the curve a b c from the base a d gives us the mean value of dura- 

 tion of rain for any given time of the day between 6 a. m. and 6 p. m. 

 This curve is the expression of the mean for eight months from July, 

 1856, to February, 1857, and is laid down according to the mean 

 numbers directly above it. It demonstrates very plainly that in the 

 moining between 6 and 7 there was no rain; but with the advance 

 of the day the rain augmented and reached its maximum between 2 

 and 3 p. m., whence it gradually abated towards evening. During 

 the night it very seldom rains. Mr. Boussingault's observations, 

 which he made in another part of South America near Marraato, prove 

 that at that jjjace more rain fell at night than during the day ; and he 

 says, in his Rural Economy, "every one in South America allows that 

 it rains principally during the night." Now this is in direct opposi- 

 tion with my observations here, and it shows, therefore, that a ceitain 

 state of the weather, especially with regard to rain, may sometimes be 

 limited to small districts only. 



From table No. 3 we also see that the month of February, which 

 is commonly considered to be one of the dryest of the ; ear, and pro- 

 perly belonging to the very centre of the dry season, has been the 

 wettest mouth of the year, with the exception of May. The dryest 

 months were March and April. 



By a glance at the tables No. 6 a to 6 /, we may have a ready 

 survey over the dry and wet months of the year and the distribution 

 of rain in general. Here we tind that the limits of the dry and the 

 rainy seasons are not very distinct, and from May, 1854, till the end 

 of 1855, a period of 20 months, we find no well defined dry season, 

 the mduth of February, 1855, being the dryest. But with New Year's 

 Day, 1856, there commences a dry season which lasts for five months, 

 the longest and dryest the colonists ever enjoyed. And it was in this 

 extraordinary dry period that the loose layer of half decomposed vege- 



