sample square of the whole series, and a few explanatory words will disentangle this web of figures. The radii extending from the inner to the outer circle inclose sixteen points of the compass—as north, north-northeast, northeast, etc. Every two concentric circles contain the data for each season. The problem being, then, to compare the relative prevalence of the same wind in different months, it is done as follows: suppose it a northerly wind; looking at the figures between
the two radii opening toward the top and between the outer and second circles, we see that, of periods of eight hours each, there were 32 in December, 21 in January, and 29 in February; the figures between the same radii and the second and third circles show that there were 41 periods in March, 33 in April, and 6 in May; and similarly, for each wind between every two radii. To compare different winds for the same month, say December, we look at the first figure to the left in each space between the outer and second circles, and find that, of periods of eight hours each, the wind was 32 times from the north, 29 times from north-northeast, 56 from northeast, and so on round the compass.
The figures 416, 385, and 408, in the upper right-hand corner, denote the total number of observations in December, January, and February, respectively; and similarly for the other months in the other corners. The figures in the center express the periods of calm in the several months.
Though this arrangement is compact and ingenious, still, when we come to make the comparison that is the real object of the chart, viz., the relative frequency of different winds in several adjoining squares, we find the task a little irksome.
Second, the Thermal Charts. These show the temperature of the