SCIENCE IN MAN'S STRUGGLE ON ARID LANDS 33 
snow, trees that hold this snow in the air, away from the ground, 
actually reduce the water intake of the soil. Yet one cannot easily 
generalize. For example, on a thin sloping soil over cracked lime- 
stone, pine plantings may slow down runoff and give added water 
to the substratum, thus restoring springs or wells in the adjacent 
lowland. On a sloping deep soil of little rain, vigorously growing 
eucalyptus trees can use so much water that springs and wells 
in the lowland dry up. Nor does undergrazed grassland necessarily 
give the best situation. Moderate grazing commonly leads to more 
spreading of the plants and less chance for runoff between the 
bunches of grass. Thus on many soils moderate grazing gives both 
optimum production of livestock and optimum water yield. 
Much interest is developing among engineers in the possibility 
of control of underground water storage in contrast to storage by 
dams. In hot, dry areas evaporation losses from artificial ponds 
are enormous. If the water can be stored underground, these 
losses can be avoided. But such storage is useful only where leaking 
through deep cracks or contamination by salt can be avoided and 
where pumping costs are not excessive. 
Tritium may be a splendid new tool to help in the study of 
water shortage. 
Radioactive hydrogen with an atomic weight of three, or 
tritium, is produced in the upper atmosphere by cosmic ray bom- 
bardment and is brought to the earth as a component of rain 
water. The longer the time the water vapor remains in the air, 
the higher the content of tritium. Hence the normal amounts for 
various regions differ, but they are fairly constant for any one 
place, barring unusual disturbances. Tritium formation ceases 
when the rain reaches the earth and that present in the water 
decays at a known rate: its half-life is about 12 years. Thus by 
comparing the tritium content of a water sample with that of the 
rainfall in the same region, one can estimate the time since the 
water fell as rain. 
Libby and his co-workers (1) have made extensive studies of 
water from wells, springs, rivers, and other sources with interesting 
results. Since the life of tritium is short and its content in rain is 
small, dating by this means is probably limited to periods under 
