June i, 1892.] THE TROPIORL A-OBfTOOLTORlST. 
931 
Coolies (our labonrorB) have to be considered. Good 
^' lines" on Eskdale were nearly complete by last 
advices. The directors believe in treating coolies 
well, and that if there be a pressure for labour those 
estates situated, aa the company's are, in a healthy 
district, nnd furnished with comfortable" lines, " will 
be greatly advantaged through having the preference 
before others among coolies. Prices for fine teas, like 
those from the company's properties, keep up very 
well, and are about as high as they were a year ago, 
thoatth the average price of Ceylon tea has fallen. I 
Coffee on St. Leonard's promises to again a fair j 
crop. It has been estimated at 3,000 bushels for 1892 I 
by Mr. Edward Grigson, who at that figure much I 
under-estimated it last year. There were excessive rains [ 
that must have caused some loss in January and Feb- 
ruary. On Liddesdale, in January, the total fall was j 
45J inches in 25 days, ajfainst an average of about 70 
for the year. We have had no crop figures since ; but 
if there was the same margin on the estimate this year I 
as last year, some coffee may be lost and yet leave a 
fair crop. Tea from Eskdale abo promises excep- 
tionally well ; and the general prospects for the year 
current seemed bright and promising, 
Questions wereasked about the aocounts, and re- 
marks made by Messrs. Johnston, Anstruther, Wilson, 
and others. When these were answered, the report 
was adopted. A dividend at the rate of 10 per cent per 
annum was voted for the four and a half mo'iths of 
1891, and £50 for division amon:; the directors for 
past work. Mr. Grieve and Mr. Brooke were respec- 
tively elected and re-elected directors, Mr. F 6. M. 
Grove, A. 0. A., auditor. 
Mr. Grieve, in returning thanks, said that he had 
the highest opinion of the estatt's, of their capabilities, 
and of their prospects, and that he had backed his 
opinions in the large amount of shares he held. He | 
added that the chairman had remarked on fine teas 
keeping np in price, though the average price had j 
fallen. fle(Mr. Grieve, might add in confirmation that 
he had, since he entered the room, a memorandum of 
prices put into his bands showing that bis Eskdale teas 
were selling that week at Id per lb. advance in each 
grade over the prices of the corresponding date last 
year. 
A vote of thanks to the chairman oonoladed the pro- 
ceedings.— JT. and C. Mail, April 22nd. 
THE INFLUENCE OF FORESTS ON 
WATER-SUPPLY. 
Does cultivation and protection of forests cause an 
increase in rainfall ? The reply of Mr. Henry Gannett, 
as published in Scienee, does not tend to confirm the 
generally admitted opinion on this question ; whilst 
the statistics collected by thU scientist have the more 
value, in that they refer to extended trac's in which the 
conditions of the country and the climate, both before 
and after changes in onltnral treatment, are perfectly 
well known. 
His observations extend over — 
I. — An area of prairie lands in the Sttte of Iowa 
in the north of Missouri, in the South of Minnesota, 
IlLiuoio, and partly in Indiana. This area, measuring 
about 163,000 square miles, was formerly entirely covered 
with grass, but during the last 30 years large portions 
of it have been afforested. 
II. — The Slate of Ohio, with an area of about 58,000 
eqnare miles, formerly entirely covered with forests 
ot which at the present not one-tenth fxists.. 
III. — An area of about 18,4C0 square miles situated 
in MaseachussettK, Rhode Island, and Connecticut, 
which wan densely wooded before its colonisation by 
Europeans. After the almost total destruction of 
these forests, about one-half of the area has, since 
1860, been ro-ufforested. 
If the removal ot forests prodnoea a decrease, and 
aff orestation an incro'lBe, in the rainfall, the result of 
observations extending over a long series of years 
shonld show in the first instance an iijorease in the 
rainfall, in the second a doorcase, and in the third a 
decrease up to fijUU, and an incrcaso aflor that date. 
But the statistics collected by Mr. Gannett show 
that in these prairie lands an increase iu the area nnder 
forest has not only not been followed by an increase 
in rainfall, but by an appreciable decrease. In the 
second instance, that of Ohio, a decrease in rainfall 
has indeed been proved, but this decrease ia so insigni- 
ficant that it oannot be seriously advanced as a con- 
oluaive proof of the anfavorabU effect of diaaffores- 
tatiDU. The reaults of statistics collected in the third 
instance, that of Massachusetts, also do not tend to 
confirm in any way the generally accepted theory, for 
up to 1860 it is shown that there was an evident 
inciease in the rainfall over this area, reaching a maxi- 
mum of 28 inches annually. 
Mr Gannett also investigated the question as 
whether the cultivation of land denuded of 
forest growth resnlted in inflnenoinK the rain- 
fall ; but the result of these, investigations proved 
that no increase or decrease had occurred. 
In writing generally on the cansea of atmospheric 
phenomena, we have replied to the often put question 
which forms the title of this article long before Mr. 
Gannett wrote on the subject. In this periodical some 
six years ago we said " that forests do not produce 
rain, but that they pliy the important part of stor- 
ing it up." 
As far as concerns Algeria, we have arranged the 
obfervations registered at various meteorological sta- 
tions in the provinces of Oran and Oenstantine, and 
these (ibsetvations, extending over a period of 25 years 
refer to large areas covered with forest adjacent to' 
others, which are entirely free from forest growth ; 
and whilst the areas are not to be compared with those 
reported on by Mr. Gannett, jet the results of the ob- 
servations are very conclusive. 
The region where the rainfall observations have most 
interest for the forester is bounded on the north, be- 
tween Bulgaria and Lalalle, by the Mediterranean, on 
the east and west by the valleys of Summam and Sez- 
bouse, and on the south by the high plateaux forming 
the water-shed between the sea and the desert of 
Sahara. This tract is in area about 47,000 square miles ; 
and thouRb no regular re-afforestation works are being 
carried out, yet the closure of large extents of forest and 
pasture land against the destructive action of the natives 
may almost be regarded aa having a similar effect. 
In spite, however, of these protective moasnres, many 
thousands of acres have from 1850 to 1875 been burnt 
over, and it is especially in these burnt areas, when 
compared with others sucoeisfully protected, that the 
rainfall statistics have the greatest significance. 
These statistics show the following results : — 
I. — That nearly the same amount of rain fell annu- 
ally before and after removal of forest growth, and 
before and after re-afforestation. 
II. — That totally different effects are produced by the 
annual rainfall before and after removal of forest 
growth, and before and after re afforestation. 
During the summer following the removal of forest 
growth, the spring level begins to fall, and the follow- 
ing year most of th" springs dry up. 
In consequencB of the water-courses cease to be 
permanent and become intermittent, being transformed, 
during actual rainfall, into impetuous torrents, which 
cease to flow during dry weather. 
The valley of Oned-Guebli to the north of the pro- 
vince of Oonstantine furnishes a remarkable instance 
of this. 
This immense valley is divided into two portions by 
the river of the same name, and the western side in- 
cludes the densest forests of this region, whilst the 
eastern is almost entirely denuded. 
During eight years of topographical re.search in 
these mountains, we have invariably remarked that 
during the winter, when heavy rain falls persistently, 
often for weeks at a time, the floods in the water 
courses from the Western or wooded side rise slowly, 
and rarely overflow the banks, and even after tropical 
rain storms, which are frequent, the water remains 
clear. 
On the eastern or denuded side, however, this ia 
not the case. Scarcely has the raiu commenced when 
each small ravino becomes a torrent, which rolls 
