The object of this paper is to give a general idea of the meteorological conditions prevailing on the Rhodesian tableland. Observations have been taken in Bulawayo since February 1897, but the observations from January 1904 to January 1906 only have been used for the present paper, as these only have been controlled by recording instruments. The data from the recording sheets have not been tabulated, and I am not able to give at present more than a summary of the results likely to be gathered from them.

A mere glance at the Anemometer traces shows a large predominance of wind in the East quarter, the main winds of the year ranging from E. to S.E., E. and E.S.E. having a prevailing character. The records seem to show one kind of disturbance only, which is the diurnal rotation of the wind.. This rotation is very frequent here, and, as far as we are concerned, of great importance. Other disturbances are so rare that they cannot have any appreciable effect on the main results.

Taking, therefore, the anemometer records as a guide, I find that we have two main weather periods, one of winds in the East quarter, and one of daily veering winds. There is often a transition period from one to the other when the wind veers from E. to N. or N.E. during the hot hours of the day.

In the East wind periods the main wind seems to be E.S.E.; in fact, the vane remains for days together parallel to the streets of Bulawayo, which run 14 degrees S. of E. South-east winds also occur, but the wind hardly ever remains in a more southerly direction for any length of time. There is often a veering of the vane from E. or E.S.E. to a point more south during the day with a motion back to its original position after a few hours. The strength of the wind ranges from one to four on the Beaufort Scale, stronger winds are rare and calms also; out of 1009 hours of calm registered in the two years under consideration, only 48 fell in these periods or on periods of East wind with some North in the middle of the day. Most of those calms were of short duration, the longest being one of seven and one of eight hours.

ones.

The periods of daily veering winds are shorter than the preceding There were on the whole 84 such periods, giving a total of 234 days out of 731. In these periods the wind is generally Easterly during night; after sunrise, some time between 7 and 10 a.m., the vane turns North; in the afternoon it is West, and some time late in the afternoon or at night we find it South; and, during night, it completes the rotation to East. This is, of course, the main tendency of the wind on those days. The passage from one quarter to the other is cften very rapid, and calms are frequently taking the place of one quarter, especially in the passage from West to South. The rotation is practically always diurnal; on three occasions only, did it take two days to complete it, and now and then it took place in the space of a few hours. Strong winds are rare. Practically all the

hours of calm, 961 out of 1009, occurred in these periods. Besides these rotations, I have noted about half-a-dozen clear movements from East through South to South-West and back, or right round through South in the course of a day.

The veerings of the vane from E. or E.S.E. to N.E. or N. in the middle of the day and back in the afternoon are not rare. The behaviour of the barometer, thermometer, and hygrometer traces induce me to think that they ought to be classed in the complete rotation periods through North.

The importance for us of these diurnal rotation periods comes from the fact that over 80 per cent. of the rains in these two years fell in this weather type.

From January, 1904, to January, 1906, the rainfall has been 31.61ins. distributed in the following way :

4.89ins. in the East wind periods.

4.41ins. in the periods of East wind with several hours of N.E. or N. in the day.

21.78ins. in the daily rotation periods.

.53ins. on one rotation of the wind through South.

This gives 70 per cent. for the veering wind periods alone, and 84 per cent. if we consider them with those in which there was a notable amount of N. or N.E. wind.

A more accurate discussion will very likely show that the amount of rain due to the East wind periods is even less than 15 per cent., as I find that a fair portion of the 4.89ins. fell in the passage from the veering periods to the East wind periods, and ought to be considered as part of the rains of the preceding day.

There is a marked difference between the two sets of rains; the east rains are mostly drizzles, amounting at the highest to one or two-tenths of an inch in 24 hours. They are generally less than onetenth. Occasionally they turn into a slow, steady rain. On only one occasion was there a record of more than one inch, 1.28ins., in one of the east wind periods, and most of this fell in the first day of it.

The rains in the other periods are almost all heavy rains connected with thunderstorms. Nearly all the thunderstorms noted in the meteorological register are on days when the wind was veering, a few on the days when N. or N.E. was very prominent, two on the first day of some east wind periods. Thunder and lightning without storm or rain occur very frequently, very rarely with pure east winds. Slow, steady rains are rare with veering winds, and are almost always the clearing rains after a thunderstorm. The fact that most of our rains come from thunderstorms explains why they have a very local character and why stations very little distant from one another and in practically the same geographical position show rather marked differences in the rain (v.g., Jan., 1906, rainfall at the Observatory 10.11ins.; at the Waterworks, two miles away, 7.25ins.).

This seems also to confirm the general idea that the Rhodesian rains are due to the North and West winds. This question of the origin of the rains in Rhodesia is not, however, easily settled. Mr.

*

Sutton and Mr. Stewart strongly oppose the prevalent theory which maintains that the moisture-laden winds from the Indian Ocean are the primary source of the South African rains. "Our rain," writes Mr. Sutton (for Cape Colony) "originates on the Equator, being carried hither in the upper atmospheric currents flowing from West and North. The rain begins on the East Coast when the moist upper stratum meets the lower bodies of air, damp with the moisture from the Indian Ocean. Then it gradually works back from the East coast as the Eastern air becomes heavier with vapour.

The idea that the rains are due to a mixture of the upper and lower currents seems to me to explain, partly, at least, what takes place in South-Western Rhodesia. Rain occurs here mostly when the conditions for such a mixture are fulfilled, i.e., when there is a barometric depression and an increase of temperature during the periods of North or veering winds. There is then a double reason for an upward movement of the lower air. Judging from the cloud observations taken at 8 a.m. and 8 p.m. during the last two years, I think that the shell of the atmosphere affected by the lower currents must be rather shallow, and that a small change in the temperature and pressure will, on that account, force very rapidly the lower current into the upper one. The higher and middle clouds move, we may say, invariably in the same direction, which is Eastwards from a point perhaps slightly North of West. This direction is, as far as I can judge, constant the whole year round.

As I shall show later, the mean surface wind is almost exactly opposite, blowing from a point between E. and E.S.E. This, however, leaves open the question whether the moisture comes originally from the upper West or North-West current or from the lower easterly one. I incline to believe that a great part, the greater part, perhaps, of the moisture comes direct from the Indian Ocean by the East and South-East winds. These, as a rule, it is true, give us clear weather, but there are several indications which point out that they must be rich in moisture which is to be condensed and brought down as rain. There is, first, the fact that the mean relative humidity is generally higher in the East wind periods than in the others, and that, even in the rainy season, when there is little or no rain in the former, whilst there is often a large amount in the latter. It would, however, be advisable to study the dew point and the absolute humidity in connection with the relative humidity to settle satisfactorily the question whether the East winds are really more moist than the others. Another fact also is that a drop in the temperature readily brings forth clouds on East wind days. There is here a particular kind of cold weather for which the natives have a special name; they call it "Amakassa." It is exclusively connected with East to South-East winds. On these occasions we have very low, fast-moving cumulus clouds of ragged appearance. They often coalesce into a cumulostratus, covering the whole sky and invariably move in the same direction as the surface wind. They

"An introduction to the study of S.A. Rainfall." J. R. Sutton. Trans. of S.A. Phil. Soc. Vol. xv. P.I. "Meteorology in South Africa," Ch. Stewart.

rarely give an appreciable rain, though the air is unpleasantly damp. Drizzles, lasting two or three days, and yielding anything from zero to one-tenth of an inch per day is the most they give us. During that time the Hygrometer remains very high, and evaporation below the average even with high wind. This kind of weather is a noted feature of the Rhodesian climate; no data are, however, available to decide whether it could be traced right to the Indian Ocean or whether it might not be due to rains further South or South-East. These reasons would make me hesitate to accept on the whole for Rhodesia Mr. Sutton's theory of the Cape rains. *

It will be of interest to compare the other records with the Anemometer sheets. The Barometer, Thermometer, and Hygrometer show a distinct periodic movement, and a comparison with the winds shows a very close synchronism with the wind periods.

To demonstrate this roughly, I have drawn up the following table of means for each wind period :— †

Mean of Mercurial Barometer reading at 8 a.m. reduced to 32° Fahr.:

read off a Richard Hair Hygrometer sheet.

For all these means I have followed the wind periods except now and then for the hygrometer. I found there that the humidity curves have a different appearance with the different winds, but that the change from one period to another does not always coincide exactly with the changes in the other elements.

There is sometimes a difference of 24 hours, sometimes, on the contrary, the changes in the Hygrometer follow very closely the wind variations. In cases where I found only one day's difference in the periods, I put the day of transition now in one and now in the other period, according to the shape of the curve. The mean relative humidity, calculated in this way, is, of course, liable to be very erroneous in the rainy season; but I do not want absolute figures, I only want to find the way in which the quantity varies. In the dry season from the beginning of April to the end of October, the mean relative humidity does not differ very much from the mean of the highest and lowest readings of the day.

The general result of this tabulation may be stated as follows: A period of easterly winds corresponds to a high barometer, low temperature (both air and sun temperature) and a high Relative Humidity. A period of veering wind to a low barometer a high temperature and a low Relative Humidity.

Out of 84 periods in which there was a daily veering of the winds round the compass 65 fall between two high barometers, 3 between two low barometers, and the others on a rising or falling barometer.

* At Boroma, near Tete, on the Portuguese Zambezi, latitude 16° North East of us, I find that half the rainfall of 1891-92 is attributed to S.E. windsI have not yet seen the later rainfall returns.

+ See Appendix.

The fluctuation of the barometer is very small at Bulawayo (Lat. 20° 9′). The mean of the extreme ranges for eight years gives only .222 of an inch. The difference between two consecutive days in the two years under consideration was nearly always under .1, the highest was .158.

The correspondence in even this summary handling of the figures, between barometer and wind changes, points out the importance of relatively small pressure variations in these latitudes.

The air temperature results show on the whole the truth of the statement that a low temperature corresponds to East winds; there are more inversions, however, than in the pressure figures, but the Black Bulb temperatures show a remarkable correlation. Out of 70 periods of daily veering winds which occurred from April, 1904, to November 9th, 1905, during which time the Black Bulb readings were taken, 54 show a higher temperature than in the East Winds that preceded or followed. The differences are sometimes very large. especially in the rainy season. Clouds covering the sky in the middle of the day may partly account for this great difference, but in the rainy months clouds appear after 10 a.m. nearly every day, so that the chances may be said to be equally distributed. But the same sequence of high and low temperatures, though with smaller differences, is observed in the clearer months, for instance in July and August, when the cloudiness comes down to I.I.

The variations of the Relative Humidity are also very interesting. Even in the rainy season, with a large amount of rain in the low barometer periods, there is a marked lower Relative Humidity than on high barometer and East wind periods. The temperature variations partly account for that, but not completely, I think. The East Wind must be more saturated with moisture, although a large amount of it has often been deposited on the high lands of the Eastern coast. Το settle this point satisfactorily, the dew point and absolute humidity figures would be required. It might be said also that the greater relative humidity corresponding then to the East winds may be considered as the effect of the rains that fell in the preceding days, especially as the rains setting in from some other direction with thunderstorms, often continue as slow rains with winds ranging from E. to S. But the same sequence of high relative humidities with easterly winds and low ones with other winds is prevailing with much more marked lifferences in the dry season when no rain falls. The Relative Humidity curves show a peculiarity worth mentioning. During East Winds the amplitude is very great, ranging often from 10 per cent. or lower to 90 per cent. or over, whilst during the other periods the daily curve is very contracted round the mean of the day, the maximum remaining considerably lower than the maximum during the East Winds.

Special mention ought to be made of the periods in which the wind passes from E.S. E. or E. to N. E. or N. in the middle of the day and back, usually some time in the afternoon before sunset. This might correspond to the law which Dr. Sprung* gives in his

* Lehrbuch der Meteorologie. 1885 p. 345.