NOVEMBER I, 1906] 
IVA ORE 
19 
The following summary of the annual rainfall in the 
Upper Nile basin is taken from Sir William Willcocks’s 
““ Nile in 1904.’’ In the catchment basin of the Victoria 
and Alber; Lakes, the mean annual rainfall may be taken 
as 50 inches, with large fluctuations between good and 
bad years; over the Albert Nile region it is about 4o inches, 
with severe droughts occasionally and excessive rain in 
some years. In the catchment basin of the Sobat River it 
probably averages 40 inches, and in that of the Bahr-el- 
Gazal region 30 inches. The rainfall over the Abyssinian 
plateau may be taken as 50 inches, and in the lower reaches 
of the Blue Nile and Atbara 30 inches. These are un- 
doubtedly rough estimates, but, so far as can be judged 
from the exact data given for a number of individual 
stations in Captain Lyons’s work, they are approximately 
correct values. They also indicate that the mean annual 
rainfall over the Upper Nile basin differs little from 
40 inches. This is a somewhat remarkable result, as it 
agrees closely with the average rainfall in India, which, 
according to Blanford, is 42 inches. 
The rainfall in the equatorial lake belt resembles in its 
seasonal distribution that of Ceylon, and that of the inter- 
mediate region (the Soudan and Abyssinia) that of western 
India. In western India, as in the East African monsoon 
region, the cool and dry season is rainless, with clear 
skies and light to moderate land winds. The rains in 
each agree in period, in the comparative suddenness of the 
change from the dry to the wet season, in the occurrence 
of almost daily heavy rainfall, and also in the rapid with- 
drawal of the humid currents at the end of the season. 
The meteorological data indicate clearly that the rainfall 
in both areas is due to the rapid extension of the south- 
east trade winds northwards from the equatorial belt at 
the same critical epoch, and probably under the same 
general conditions. There is one very important difference. 
The monsoon current in the Nile basin does not extend 
beyond lat. 16° to 18° N., being bounded to the north, not 
by a range of mountains, but by an area of permanent 
low pressure during the season, due to thermal actions. 
It curves rapidly from south to west, and is hence deter- 
mined directly to the western face of the Abyssinian 
plateau and mountain masses, which in their highest points 
attain an elevation of 15,000 feet. The Bombay current 
in India extends as far northwards as the East Punjab 
(lat. 30° to 35° N.), where its further progress is barred 
by the Himalayas. The Abyssinian plateau exhausts the 
humid current much more completely than the West Ghats, 
as the rainfall at Massowa and other towns on the Red 
Sea to the east of the plateau is practically nil. 
It would be interesting to determine whether the humid 
current is converted into a vertical movement over the 
plateau or whether it continues to march eastwards, and 
perhaps to contribute to the monsoon rainfall (of the same 
period) in the mountain region of Yemen, in south-west 
Arabia. 
Captain Lyons has devoted considerable attention to the 
question of the variations of the Nile flood; and hence of 
the rainfall in the Nile basin, from year to year. The 
data show that very large variations occasionally occur 
amounting to +35 per cent. of the mean. He infers from 
the data of years that they do not exhibit any regular 
cyclical variation, and hence that they cannot be directly 
correlated with the eleven-year sun-spot period or the thirty- 
five-year Bruckner period. 
It is now, we believe, fully established that Abyssinia, 
India, and Burma, with the Malay Peninsula, receive 
nearly the whole of their rainfall from the same vast 
reservoir and evaporating area, the Indian Ocean and seas, 
and under the same general meteorological conditions, and 
by means of the same general air movement. These facts, 
on the one hand, indicate a probable similarity or 
parallelism of the seasonal variation of rainfall in all 
three areas due to general conditions in the contributing 
Oceanic area, and, on the other, an unequal and unlike 
variation due to variation of local conditions in the 
three large areas of distribution; also as the rainfall of 
the Abyssinian plateau is due to the same branch of the 
monsoon current as that of western India, any parallelism 
of variation is more likely to be exhibited by these two 
regions than by either compared with Burma or north- 
eastern India (dependent on the Bay monsoon current). 
NO. 1931, VOL. 75] 
The actual variation in any one year will hence be due 
to the resultant of the general and of the local conditions. 
It is also probable that the largest variations will be due 
to the general variation over the whole area of supply 
The data furnished by Captain Lyons are, on the whole, 
in full agreement with these inferences. The most remarlkk- 
able case of similarity of seasonal variation is exhibited 
by the data of the past fourteen years. The following 
gives comparative data of the rainfall of India and of the 
Nile floods from 1892 to 1905. The former data are 
obtained from the Indian meteorological publications, and 
the latter from Captain Lyons’s memoir :— 
Ratio of mean Ratio of 
Year actual to normal actual to normal 
rainfall in India Nile flood 
1892 I'I2 1‘20 
1893 E20 i 0'99 
1894 TOG. Gen 1°22 
1895 0°95 Ils 
1896 0°88 1°06 
1897 0°99 0°89 
1898 I‘OI 1‘07 
1899 0°73 0°63 
1900 0°99 0°89 
1901 0°90 0°87 
1902 0°95 0°63 
1903 1°05 0°89 
1904 ? below O'75 
1905 much below normal 0°65 
Period 1892—4 1°16 I"I4 
oA iso58 aC 0°94 1°05 
1899-190 
m0 ee ae O%98) 50s ate 0°78 
It is a noteworthy fact that the Abyssinian rainfall, as 
indicated by the Nile floods, is subject to much larger 
range of variation than the rainfall of India, as might 
perhaps have been anticipated. The data show that from 
1892-4 the rainfall in India and in Abyssinia (assumed to 
be roughly proportional to the total Nile flood) was in 
considerable excess from 1892-4, about normal from 1895-7, 
and more or less in defect from 1898 to 1905. The 
parallelism would have been more exact if the rainfall of 
western India had been given instead of that for the whole 
of India. The 1896 drought in India was due chiefly to 
the weakness of the Bay current, and not of the Arabian 
Sea current. It may be noted that the data for the vari- 
ations of the level of the Victoria Lake agree generally 
with those of the Abyssinian rainfall, as indicated by the 
Nile floods. Thus, according to Captain Lyons, 1892-5 
was a period of high level, 1896-1902 a period of falling 
level, and 1903 a year of rising level. This remarkable 
parallelism, strictly in accordance with the general simple 
inferences stated above, suggests two problems for the con- 
sideration of meteorologists. These are, first, the causes 
of the large variations from year to year of the rain supply 
over the immense land area of India, the Soudan, and 
Abyssinia, and, secondly, the determination of any invari- 
able antecedent conditions which may serve as indications 
and be utilised for forecasting these variations. Captain 
Lyons in the last chapter of his memoir takes up both of 
these problems, but acknowledges that his investigations 
are only in the introductory stage. It is, however, interest- 
ing that his present conclusions on the whole agree with 
those of Indian meteorologists. He shows, for example, that 
pressure in the Egyptian region is below the normal in 
seasons of good Nile floods and wice versd. This is the 
usual relation between pressure and rainfall in India, and 
is also in accordance with theory. Captain Lyons also 
points out that the monsoon variations of pressure are fre- 
quently, if not invariably, the continuation of similar con- 
ditions which have prevailed for some time previously. 
This is also in accordance with Indian experience. He 
also points out that they are probably in some cases re- 
lated to the widely distributed variations of pressure studied 
by Sir Norman Lockyer and Dr. Lockyer, and also to 
the long-period variations in India. The latter are marked 
by or accompany prolonged abnormal variations or anoma- 
lies of the Indo-oceanic air movement. He also considers 
that they are occasionally determined by variations in the 
position and intensity of north-east Atlantic anticyclones. 
This is by no means improbable, but until more is known 
