29S CLIMATE OF
On the Guiana side of the Amazon, in the islands of Mexiana and Marajo, the seasons are more strongly marked than even higher up the river. In the dry season, for about three months, no rain ever falls ; and in the wet it is almost continual.
But it is in the country about the falls of the Rio Negro that the most curious modification of the seasons occur. Here the regular tropical dry season has almost disappeared, and a constant alternation of showers and sunshine occurs, almost all the year round. In the months of June, July, August, and September, when the Amazonian summer is in all its glory, we have here only a little finer weather about June, and then rain again as much as ever ; till, in January or February, when the wet season in the Amazon commences, there is generally here a month or two of fine warm weather. It is then that the river, which has been very slowly falling since July, empties rapidly, and in March is generally at its lowest ebb. In the beginning of April it suddenly begins to rise, and by the end of May has risen twenty feet, and then continues slowly rising till July, when it reaches its highest point, and begins to fall with the Amazon. The district of the greatest quantity of rain, or rather of the greatest number of rainy days, seems to be very limited, extending only from a little below the falls of Sao Gabriel to Marabitanas at the con- fines of Brazil, where the Pirapoco and Cocoi mountains, and the Serra of Tunuhy, seem to form a separation from the Venezuela district, where there is a more regular summer in the months of December, January, and February.
The water of the Rio Negro in the month of September did not vary in temperature more than two degrees. I unfor- tunately lost my thermometers, or had intended making a regular series of observations on the waters of the higher parts of the rivers I ascended.
The extreme variation of the barometer at Para" for three years was only three-tenths of an inch (see diagram, Plate XIL). The mean height, with all the necessary corrections, would seem to be almost exactly thirty inches ; I have, how- ever, already given my reasons for believing that there is a considerable difference in the pressure of the atmosphere in the interior of the country. In the month of May some very cold days are said to occur annually on the Upper Amazon
