In practice also the amount of cloud can be represented by a numerical estimate of the relative portions of visible sky covered by cloud and free from cloud. Thus an estimate is formed by inspection of the number of tenths of the whole sky which is covered by cloud. Various artifices may be adopted in order to obtain a satisfactory estimate and make it easy for different observers to agree in their estimates. Dividing the sky into quadrants and estimating the proportions of cloud in each quadrant is the most usual mode of procedure. Visibility is another quantity for which in recent years numerical expression has been sought. The enumeration is arrived at by selecting a series of objects at successive distances and noting the furthest of these objects which is visible. In that way the following numerical scheme has been evolved to cover the whole range of vision from dense fog to very transparent air:

The introduction of these numerical estimates enables the results to be summarised by numerical mean values and not exclusively by the number of occurrences, and in this way, although no instrument may be used, the representation differs from those of the typical non-instrumental observations which are represented by a simple count of the number of occurrences without regard to the intensity of any one of them.

Instrumental observations

The instruments which by international agreement form the ordinary equipment of a climatological station are the mercury-barometer, the thermometer also of mercury, the maximum thermometer, the minimum thermometer, the wet bulb or a hair hygrometer, the rain-gauge. They can be supplemented by a barograph for obtaining a continuous record of pressure, a thermograph which treats temperature in like manner, by an anemometer for recording the direction and velocity of the wind, by a sunshine-recorder for indicating the duration of sunshine, a pyrheliometer for measuring the thermal intensity of sunshine, a pyrgeometer for measuring the loss of heat by radiation from the earth, a nephoscope for finding the direction of motion of clouds, and many others; but for the present representation of Mediterranean climates we limit our attention to the more common instrumentsthe barometer, thermometers and rain-gauge-in addition to the noninstrumental observations.

In respect of the units employed for the representation of these quantities the world is still divided into two unequal parts. One part uses British units, the mercury inch at 32° F in latitude 45°, the Fahrenheit degree and the

inch; and the other part uses the millimetre of mercury at o° C in latitude 45°, the centigrade scale and the millimetre.

The practice of giving pressure in terms of a hypothetical inch or millimetre at the freezing-point of water in latitude 45°, which was introduced into international practice only at the beginning of the current century, has seriously changed the aspect of the question of the choice of units for the expression of pressure. At a meeting of the British Association many years ago we heard the best-known British authority on units and physical constants urge that the correction of readings of the barometer for latitude was undesirable, that what was wanted in meteorology was not pressure in pressureunits but something equivalent to the height of the atmosphere at the place of observation, i.e. the "barometric height." But international agreement has decided otherwise. By that agreement the last obstacle was removed from the expression of pressure in the manner appropriate to the expression of a distributed force, namely the thrust in absolute units of force upon a unit of area.

We have accordingly selected units for the expression of these physical measurements of the atmosphere on the same ground that electricians and magneticians have chosen the units appropriate to their branch of science, namely those which are systematic and on that account the most convenient in the long run for computations involving more than one physical quantity. Those units are:

For the barometric pressure or for vapour-pressure 1000 C.G.S. units of force, called dynes, per square centimetre. This unit we call a millibar. A pressure of 1000 millibars is our standard atmospheric pressure to which other measurements are reduced when that kind of reduction is required. That is equivalent to 750-076 millimetres or 29.5306 inches of mercury at the freezing-point of water in latitude 45°. To indicate a measure of pressure in millibars in formulae or tables we use the symbol "mb."

For temperature-the tercentesimal scale of centigrade degrees measured from 273° C below the freezing-point of water. For all practical purposes the scale agrees with the absolute thermodynamic scale, but there is an actual difference which requires elaborate experiments for its determination; the value accepted at the present time for the difference between the absolute temperature and the tercentesimal temperature amounts to about a tenth of a degree. In order to indicate a measure of temperature on the tercentesimal scale in formulae or tables we use the symbol "t," or preferably "tt" to avoid confusion with the symbol for time.

For the depth of rainfall we use the millimetre with its symbol "mm.” For the velocity of the wind we use the metre per second or possibly on occasions the kilometre per hour. The symbols for these measurements are "m/s" and "km/hr."

We will not interrupt the narrative at this stage to give in detail the reasons for this selection, beyond saying that pressure and temperature as used in meteorology require much greater precision than is attached by common. usage to a quotation of a reading of an ordinary domestic weather-glass and its attached thermometer. The readings used for meteorological purposes have passed through processes of correction and reduction and require some acknowledged evidence of the fact. The difference between a hall-barometer and a scientific instrument is not large but it is vital for meteorology. Our chief reason is however the necessity in meteorology for systematic units. Let anyone who wishes to convince himself of the necessity make a small collection of meteorological data and try the systematic computation of such meteorological quantities as the density of air, or of the water-vapour which it contains, the gradient-wind, the radiation from black earth, the energy of sunshine and the potential temperature. We can promise him that his only feeling after a week of computation will be a lament that the angular velocity of the earth's rotation cannot be represented by some better expression than 72.92 × 10-6 radians per second.

CLIMATIC SUMMARIES FOR MEDITERRANEAN STATIONS

With this preface we may pass on to consider the climatological tables in which an attempt is made to use the available material to represent the climate as concisely as possible with reference to the following elements: the direction of the wind at different hours of the day and in different seasons, and the number of gales or the mean wind-velocity; the frequency of fog and the amount of water-vapour in the air; temperature and its ranges; the normal values and variability of rainfall; and the number of days of precipitation of any sort as well as of hail ▲, snow, thunderstorms R and dust-storms. In the two cases where periodic irrigation by river forms part of the economic conditions the normal heights of the river in the several months have been given. The most conspicuous features of the common climate is the difference between summer and winter, not only in respect of solar radiation and temperature, but also in respect of rainfall and in the inverse sense the amount of water-vapour in the air.

The stations selected are:

Helwan for Cairo and the ancient Memphis, at the head of the Delta of the Nile. The level of the river has a normal range of 5.1 metres between May-June and September-October, with an almost rain-free record for the seven months May to November and really appreciable rainfall ● only in January; its yearly total is 34 mm (an inch and a quarter); the temperature is always above the freezing-point, reaching 315 t (108° F) in June and above 300 t (81° F) on every day of the four months from June to September. Winds markedly Northerly in the summer when the etesian winds blow in Greece, and tending to be Easterly in the winter, but never calm. A dry atmosphere especially in the middle of the day when humidity ranges from

20 per cent. in early summer to 44 per cent. in winter. The record shows 14 dust-storms a year, chiefly in the early part of the year.

At Aswan, not far from the site of the ancient Thebes, no rain is measured

at the present day.

Babylon, now included in Iraq, the Chaldaean capital on the river Euphrates, which ranges in level through 3.5 metres between an autumn lowness in September-October and a spring flood in April1. It has three times as much rainfall as Helwan, and a much greater range of temperature-which reaches the freezing-point in three months of the year and 320 t (117° F) in July and August-less water-vapour, fewer dust-storms, if the standard of observation is the same, less wind, and chiefly from North and West, especially in the summer months. The rainfall is a winter-supply, amounting to 109 mm in the year (4 inches). The four months June to September are rainless.

Jerusalem lies on a high plateau 748 metres above sea-level and about 1000 metres above the gorge of the lower Jordan and Dead Sea. It has normally a good supply of rainfall during the six months November to April, 649 mm in the year (26 inches) but a rainless summer (June to September). Temperature may go as high as 312 t (102° F) and as low as 267 t (21° F). There are normally three days of snow in the year and it may be very heavy. A thickness of 29 inches was measured after a storm in February 1920 (see Meteorological Magazine, October 1920, p. 200). Winds are chiefly from the West in the summer and more variable in the winter. The air is not so dry as in Helwan or Babylon.

Beirut on the Syrian coast, where a rainfall which amounts to 906 mm (36 inches) in the year may occur in any month, but in the summer, June to August, falls on less than one day a year. Temperature is milder. It has reached the freezing-point in December but its normal daily minimum in that month is 12tt above the freezing-point. The winds are generally South or West in the early morning, changing to North or West in the middle of the day.

Candia on the North coast of Crete near to the site of the ancient Cnossus with 535 mm of rainfall per annum (21 inches). Being on a Mediterranean island the rainfall is more evenly distributed throughout the year, though it is still markedly characteristic of the winter season. As in the case of Beirut one-fifth of the year's supply falls in December. The temperature has ranged from just above the freezing-point to the high figure of 319 t (115° F). The winds as recorded are North or West in the summer and South in the winter but their average force is very light and there are many calms-that however may be a matter of exposure.

1 It is noteworthy that, according to Sir Norman Lockyer, the Egyptian temples were oriented to the sunrise at the summer solstice which was approximately coincident with the rise of the Nile at Memphis; according to modern reckoning the flood begins nine days earlier at the site of Thebes.

The temples of Assyria on the other hand were oriented for sunrise at the equinoxes, and in spring at Babylon that would be coincident with the rise of the Euphrates. (The Dawn of Astronomy, pp. 230, 240.)

Fig. 3. Normal distribution of the rainfall of January, April, July and October over the regions of the ancient world.

The coasts are shown in black outline: the shaded areas indicate the regions above 2000 metres (6562 ft.) in height. The red lines are lines of equal rainfall (isohyets) drawn for 25 mm, 50 mm, 100 mm, 200 mm, 300 mm, 400 mm. One line of 5 mm is suggested for April.