METEOROLOGY. 185 



where it is tied and where it was in contact with the surrounding 

 brass tube, I found a spot of one-eighth of an inch diameter, as if cor- 

 roded by some kind of acid In the centre of this spot was a hole one- 

 sixteenth inch diameter. The corroded rim around the hole was very 

 smooth and viscid, similar to partly dissolved india rubber. After 

 sewing up the whole and giving it a coat of glue, to prevent the mer- 

 cury from leaking out, 1 filled the glass tube again as cautiously as pos- 

 sible, to prevent the formation of air bubbles. In this I succeeded pretty 

 well, and, with the exception of one minute portion of air, which 

 escaped into the vacuum, the latter seemed to be complete. The 

 mercury then showed but a small difference (yo g ^^ yJo parts of an 

 inch lower) compared with its former state. Hoping to succeed still 

 better the second time, I tried my hand once more at it, but did not 

 succeed so well this time, as some moisture had settled in the glass 

 tube. The mercury is now at least one-tenth of an inch lower than it 

 ought to be. 



The barometrical observations made with this instrument since the 

 1st November, 1856, can, of course, not be considered as normal, and 

 can be used only with a view to institute comparisons among them- 

 selves. 



I feel this defect the more acutely as I hoped to measure a number 

 of mountains and other localities, and to complete a twelve months' 

 register, to find out the mean height of the barometrical column for 

 the different months of the year. Up to the 1st November I found 

 the mean height greatest in July. Hitherto I have measured only 

 the pass over the mountains on the road from the colony to Victoria. 

 On this spot the barometer was 23.334 at 7A. 30m. a. m., September 9, 

 with the thermometer at 61°. 



In the diagrams on sheet No. 1,1 have laid down, in a graphical 

 manner, the hourly and half-hourly rise and fall of the barometer from 

 6 a. m. till 9 p. m. for 12 days. We can see here, at once, the greater 

 amplitude of the daily periodical variations in October compared with 

 that of June ; also that the hours of maximum of the different days 

 in October are not far apart from each other and near to 10^ a. m., 

 and the hours of minimum not far from 4 p. m.; while, on the con- 

 trary, in June, the hours of maximum, as well as those of minimum, 

 are much more scattered, and therefore not so regular. 



On sheet No. 2 are the half-hourly observations laid down for 24 

 hours, from 4 a. m., October 7, till 4 a. m. next day. Here we 

 observe that, in the morning, the maximum, as well as the minimum, 

 is somewhat higlier than the maximum and the minimum in the 

 evening. This seems to be a general rule with all the daily periodi- 

 cal variations. 



On sheet No. 3 the daily mean barometer heights from June 10 to 

 October 30 are put down and connected by straight lines to denote the 

 course of the barometer from day to day throughout the several 

 months. A kind of periodical rising and sinking is observable here, 

 alternately taking place in periods of 4 or 5 days, at least for June, 

 July, August and September. 



Oq sheet No. 4 is to be found a comparison of the mean monthly 

 barometer heights of Colonia Tovar with those of St. Louis, Mo., 



