o2 THE FUTURE OF ARID LANDS 
larly it was reported at the conference in Jerusalem in 1953 
that all the historic, botanical and archaeological evidence pointed 
to little change in climate during the last eighty years or so in 
Israel and India. 
Tixeront points out that droughts do occur and we ought to 
become aware of their statistical probability, and he makes a 
plea that we study all the climatic factors which may lead to the 
development of drought periods, as conversely to flood times. 
In practice in Tunisia information from historical and archaeo- 
logical sources has been used to good effect. The planting of olive 
trees at the same spacing as used by the Romans in dry farming 
is successful today. So, too, wells, cisterns, and irrigation channels 
of Roman origin give a guide to modern siting and use. 
In semi-arid and arid areas transpiration and evaporation, or 
to put them together as Thornthwaite does under the term 
evapotranspiration, constitute the major factor in the recircula- 
tion of rainfall into the air, but it is notably difficult to measure 
satisfactorily because of differences in plant cover. It is becoming 
clearly evident that in some instances an increase in crops or 
pastures or tree cover may severely tax the capacity of ground 
water supplies, and in general a balance has to be struck between 
the needs of crops and the water supply. Lysimeter experiments 
in South Africa indicate that only about 3% of rainfall goes lower 
than the root zone of veldt grasses. Phreatophytes, plants which 
grow their roots down to the water table like alfalfa, are notable 
for their efficient transpiration. Indeed, I have on occasion recom- 
mended the use of alfalfa to drain waterlogged orchards in irri- 
gated country. Worthless phreatophytes may waste millions of 
acre-feet of water in some arid country. 
In order to assess the water requirements of an area, wherever 
it may be, some way of measuring evaporation and transpiration 
must be used, yet strange though it may seem, we have not yet 
achieved exactitude even in reading evaporation pans. Thorn- 
thwaite, who has devoted his scientific life to the study of climate, 
has proposed a method of estimating the water need of a region 
so that it is possible, if the rainfall and the Wee 
are known, to determine how much additional water, if any, 1 
needed by way of irrigation. He defines the total water need as 
