April 1, 1904.J THE TROPICAL AGRICULTURIST. 
663 
was removed anfficiently early to allow the cotton 
plants to branch oat afterwards. The rows ot cotton, 
when the maize was removed, were two feet apart ; 
and I recommend this distance as suitable for the 
American varieties. The plants should be thinned 
out to 15 inches to 21 inches apart in the rows 
in the case of a healthy vigorous crop. It is import- 
ant that the rows of mixed crop should be bullock hoed, 
and when the maize is removed, the space between 
the rows of cotton again intercultnred. The ordinary 
indigenous implement used in a good cotton district 
did excellent work last year. 
In 1903 I artaoged for experiments in Behar in 
sowing and cultivation on the lines referred to in 
the last paragraph. It is 
NECESSARY TO HAVE EXPERT COTTON CHLTIVATOES 
to do the work ; also trained bullocks and the neceseary 
implements. One pair of bullocks dealt with about 
40 acres last year. I can provide this year from Pasa 
as a centre five pairs of bullocks", a trained man in 
charge of each pair, the implements required, and a 
field man from my office to supervise the work 
I should like to deal with 5 or 10 bighas on each 
indigo concern within reach, provided the selected 
areas are grouped near to eiich other, so that one 
pair of liullocka can sow expeditiously about 30 
bighas altogether. The land must be of fair average 
quality and as carefully prepared as for makai, but 
should get no special treatment in the way of manur- 
ing or otherwise. 
THE OBJECT OF THESE TRIALS 
is to determine whether cotton of superior quality 
can be grown profitably on extensive areas uuder 
conditions of cultivation which can be ordinarily 
arranged for. 
I do not believe that the cultivation of cotton 
will extend in Behar unless planters can through 
their tenants arrange to grow the crop under a 
share system or some other system. The diiiicuUy 
in the way of actual cultivation by the planters 
themselves is chitfly on account of losses by theft 
which would be considerable or great, there being 900 
or more people to the square mile. 
An opinion prevails that certain perennial varieties 
of cotton which are known to produce very fine lint 
ciin be profitably cultivated on extensive areas. I 
should advise caution over this scheme. 
THE INSECT ENEMIES OF COTTON 
aie very numerous, and even iu annual crops, where 
the same area is not occupied perhaps oftener than 
once in three years, it is extremely difficult to pre- 
vent considerable damage by boll worm, and judg- 
ing by the damage done by this pest to the produce 
of perennial trees as ordinarily grown, I would expect 
very serious damage to occur in large ulantations. 
Rotation unquestionably reduces risk ot loss. Mr. 
Lefroy, the Entomologist of my Department, has 
taken up the study of insect pests affecting cotton in 
India, and will before next season be able to recom- 
mend practical remedial measures which will keep 
them in check. 
RELATIONSHIP OF WOODS TO 
DOMESTIC WATER SUPPLIES. 
This subject has, for more than twenty years, 
occupied mirch of the attention of Forest Experi- 
mental Stations, especially in Germany, France, Aus- 
tria, and Switzerland ; and in view of its importance 
the conclusions arrived at may be usefully sum- 
matised. 
. It has been asserted, and theoretically the con- 
liention is doubtless correct, that 
MASSES OF WOODIiiND INCREASE THE RAINFAIL. 
The causes of this result are sought for in the 
reduction of temperature associated svith forests, and 
in the greater absolute and relitive humidity of the 
air in woods. But although it may be possible to 
obtain experimental proof by means ot elaborate and 
long-continued observ^itions iu a region where ex- 
tensive afforest.xtion or deforestation is taking place, 
it may at once be said that such tree planting as is 
practically possible in Britain can have no appreciable 
infJueuce on the rainfall. Trees do, however, under 
certain conditions of the atmosphere, condense dew 
on their leaves and branche", and this effect may 
often be seen in the wet state of the eround underneath 
trees on a foggy morning, when the surface elsewhere 
is comparatively dry. 
But the case is materially different where the fate 
of the rain and snow that fall on a tract of v/oodland 
is considered. The foliage, branches and stems ot the 
trees intercept much of the rain and snow so that it 
never reaches the giound at all, the amount so in- 
tercepted usually ranging from 30 to 45 per cent, of 
the total, but much depends on the character of the 
rainfall, and on the species of the tree. In a district 
of heavy annual rainfall a smaller proportion of the 
precipitations is caught by and evaporated from the 
trees then when the rainfall is light. Similarly in 
the case of heavy and long-coutinued rain, as con- 
trasted with gentle showers ; in the latter case, 
in fact, but little of the water reaches the ground 
through the leafy canopy of a dense forest. 
Then again much dep;i:dg on the kind of tree, ever- 
greens intercepting more water throughout a year 
than deciduous trees, and a large proportion of the 
ra'nfall is evaporated from the leaves and branches 
in summer than in winter. 
CONSERVATION OF MOISTURE IN THE SOIL. 
But although less rain-water reaches the soil ot a 
wood than finds its way to the ground in the open 
country, the moisture in the soil is much better con- 
served in the former than in the latter case. Long- 
continued observations have shown that more water 
drains from a wooded area than from one devoid 
of trees. The greater abundance of water in forest 
soil, in spile of the trees intercepting a large pro- 
portion of the rainfall, is due partly to the reduction 
of evaporation owing to the exclusion of the sun's 
rays by the foliage, partly to the air in a forest 
being more humid, and thus better fitted to discourage 
evaporation, and partly to the absoibent and retentive 
character of the decaying vegetable matter that covers 
the ground ot a dense and well-managed wood. 
The lace-work of tree roots, too, that occupy the 
soil of a forest, offers mechanical resistance to the 
rapid surface-flow ot water. It ia also to be noted 
that roots penetrate to great depths, and when they 
die they leave holes through which water readily pene- 
trates from the surface. The friable condition of the 
soil of a wood, too, permits ready percolation ot 
water, whereas in the open country the denser 
character of the surface of the ground is less favour- 
able to the entrance of water. The consequence ia 
that 
STREAMS IN A WOODED COUNTRY 
are not so subject to rapid lises and falls, the flow 
being maintained more eq lably throughout the year. 
Where water supply for domestic or industrial pur- 
poses is concerned, the avoidance of violent freshets 
on the one hand, and scanty flow on the other, is alike 
desirable. Not only may the water of sudden and 
heavy floods ba lost owing to the incapacity of the 
reservoir to contain it, but such floods have also the 
disadvantage of carrying much mud and similar 
material in suspension, and this gradually silts up 
reservoiis, besides entailing increased expenditure in 
filtering. 
It may be pointed out that the water of a reservoir 
surrounded by well stocked woodland is not subjected 
to the same amount of violent agitation during "ales 
as is the case when such sheltering agency is absent. 
The mud and silt deposited on the bottom, and es- 
pecially along the margin, is consequently left compara- 
tively undisturbed, with corresponding advantages in 
the matter ot purity. 
