through the centre of which runs a narrower belt of still lower pressure, towards which the trade-winds blow. Since these belts of high pressure can only be maintained by air flowing in upon them in the upper regions of the atmosphere, it is towards these belts of high pressure that the upper currents of the air flow. The southern belt of high pressure lies nearly parallel to the equator, and is of nearly uniform breadth throughout; but the belt north of the equator has a very irregular outline, and great difference in its breadth and in its inclination to the equator,—these irregularities being due to the unequal distribution of land and water in the northern hemisphere.

113. Regions of Low Pressure.-Considered in a broad sense, there are only two regions of low pressure—one round each pole, bounded by, or contained within, the belts of high pressure just described. The most remarkable of these, in so far as it is known, is the region of low pressure surrounding the south pole. In reference to this singular depression, the registers of upwards of 100,000 observations have been examined, and the mean of them taken for the different latitudes, with the following result::

It is to be regretted that as the longitudes were not taken into account in taking those means, the geographical distribution of this anomalous depression cannot yet be accurately defined. The depression round the north pole is divided into two distinct centres, at each of which a diminution of pressure, still further below the average pressure, prevails. These two centres are the north part of the Atlantic and the north part of the Pacific Oceans. Their relation to each other, and as parts

of the great north polar depression, would appear with greater distinctness if drawn on a polar projection of the northern hemisphere.

114. There is a smaller area of diminished pressure in Hindostan, which is entirely due to the low summer pressure of that region during the south-west monsoon.

115. The whole of the above depressions may be regarded as due to the presence of an excessive amount of moisture in the atmosphere. The influence of high temperatures in lowering the mean annual pressure over any portion of the earth's surface is slight in comparison with the depressing influence of the vapour of the atmosphere,-almost the only instances being the slight depression in Central Asia, caused entirely by the summer depression, and the lower pressure which prevails over part of Africa and of Southern Asia. It may therefore be concluded that THE CHIEF DISTURBING INFLUENCES AT WORK IN THE ATMOSPHERE ARE THE FORCES CALLED INTO PLAY BY ITS

AQUEOUS VAPOUR-thus giving to this element a paramount claim on our regard in studying winds, storms, and other atmospheric changes.

CHAPTER IV.

TEMPERATURE, HOW OBSERVED AND CALCULATED.

116. THE temperature of the air is ascertained by the Thermometer, fig. 12, which consists of a small closed glass

-130

-120

-110

-100

-90

ន

-80

R 8 8 9

-50

40

30

tube, having a bulb at one end, and partially filled with mercury or spirit of wine. Of these fluids mercury is the best, owing to its uniform expansion by heat; the quickness with which, from its low specific heat, it indicates changes of temperature; and the great range of its fluidity. A spirit thermometer must be used when the temperature falls below -37°.9, the point at which mercury freezes, as determined by Dr Balfour Stewart. Spirit thermometers are also of

great use for registering the greatest cold.

117. The following points in the construction of thermometers must be attended to. The mercury should be pure and dry, and boiled so as to expel the air; the bore of the tube should be equal throughout; the bulb should be large in proportion to the bore, so that the degrees may be large and easily read; and no air should be left in the top of the tube. It is also most desirable that thermometers should be filled and hermetically sealed for at least a year before the scale is engraven on them. The reason is, that the fibres of the glass take some time to assume their permanent position; and since, in this transition state, the

Fig. 12.

atmosphere pressing on the exterior surface of the bulb constantly tends to push the mercury further up the vacuum above the column, the result is that the bulb becomes permanently contracted in size, and the column stands higher on the scale. Nothing is more common than to find such thermometers, after being some time in use, to read 0°.5 or 1°.0 higher than they did when previously compared with the standard. Owing to the changes to which they are subject, thermometers should occasionally be compared with a standard thermometer, or have their freezing-point tested by plunging them in melting ice.

118. Division of Thermometer Scales.-Before the indications of different thermometers can be compared, it is necessary that there be two fixed points on their scales, each of which indicates precisely the same temperature. The points which have been chosen are the temperature at which water freezes, and the temperature at which it boils when the barometric pressure is 29.905 inches reduced to 32°. In both cases distilled water must be used; for if the water contain salts or other impurities, it will freeze and boil at different temperatures. If the pressure exceed 29.905 inches, the temperature of the boiling-point will be higher, but if less it will be lower, the proportion being one degree of Fahrenheit for every 0.589 inch of pressure at moderate heights. Hence, when the barometer is 28.000, water will boil at 208.6 instead of 212°. At Santa Fé de Bogota, which is 8727 feet high, and atmospheric pressure is 22.061 inches, water boils at 197°.46; in Mexico, at 7000 feet high, it boils at 200°; and at Quito, 9000 feet, it boils at 194°.

119. Advantage is taken of this circumstance to measure roughly the heights of mountains. The temperature at which water boils is observed, from which the pressure of the air is deduced, and compared with the pressure observed at the same time at some neighbouring place, the height of which is known. From the difference of the two pressures the height is ascertained. An observation should be made at a lower level where the height is known before ascending the moun

tain whose height is to be measured, and another observation in descending, so that errors arising from fluctuation in the pressure may be compensated for. The exact hour of the day of each observation should also be noted, so that the correction for daily variation may be applied.

120. The space on the scale between these two points has been divided in different ways. FAHRENHEIT, a native of Dantzic, fixed the zero-point at the greatest cold then known to have occurred, in Iceland, supposing that lower temperatures would seldom require examination. The space from freezing to boiling he divided into 180 equal parts; and since his zero-point is 32 of these parts below freezing, the freezing-point of water is 32°, and the boiling-point 212°. This is the scale in common use in England and America, and its practical advantages over other scales are these: (1) The degrees are smaller than in the other scales, and hence greater exactness in observing is attained; and (2) as the temperature rarely falls below zero, the minus sign is seldom required,-an advantage of some value in summing up and printing tables of temperature.

121. Celsius, a professor at Upsal, divided the scale between the two fixed points into 100 parts, the freezing-point being zero. This thermometer is generally called Centigrade, from the division of its scale into 100 parts. It is used in France and some other Continental countries, and is extremely convenient for purposes of scientific inquiry.

122. In Reaumur's thermometer the same space is divided into 80 parts, the freezing-point being the zero of the scale. It is in use in Germany and Russia.

123. It is often required to convert temperatures expressed in the Centigrade or Reaumur's scale into Fahrenheit's scale, and vice versa. Since the space between the two standard points is divided in Fahrenheit's thermometer into 180 parts, in Centigrade into 100 parts, and in Reaumur's into 80, the proportions are

Hence to convert Centigrade degrees to Fahrenheit, we multiply them by 9 and divide by 5, and add 32° to the result,