Drip irrigation has been advocated as one way of improving both yield and water use efficiency for irrigated cotton. It has even been claimed that drip irrigation can be used to 'crop areas previously considered unsuitable for conventional irrigation methods .... because. ... 'the root zone is maintained at both optimal water and nutrient levels' (Anon,

1988). In this context drip irrigation may be considered a 'non independent' irrigation system where the uniformity of application of water and nutrients by the irrigation system overcomes viability in soil properties. Irrigators have been led to believe that the precision of application of water to the crop from drip irrigation will enable soil and climatic limitations to be to be adequately managed. Field experiment on cotton comparing irrigation treatments were conducted at Warren in the 1999/2000 and 2000/2001 seasons. The control was irrigated at 100% of predicted crop water use (ETcrop), and three treatments that applied 50%, 75% and 125% ETCrop .

Cotton yield was greatest in the 100% treatment followed by the 75% and 125% treatments with lowest yield coming from the 50% treatment for 1999/2000 season. In contrast the 100% and 125% were similar followed by the 75% and 50% for the 2000/2001season. This yield trend, combined with similar soil moisture contents in both the 100% and 125% treatments, suggests that the 100% treatment was supplying adequate water to satisfy crop demand for the 1999/2000 season. Some waterstress was observed in the mid season in 2000/2001 season because of small amounts of underirrigation (0.5 mm Iday).

We found that the irrigation efficiency was limited in both seasons by the large amount of water required to germinate the crop (27% of water applied to the 100% treatment). It is clear that it would be better to use this extra irrigation water to obtain the greater production recorded between the 50% and 100% treatments than to use it only for crop establishment.

We found that gradual development of waterstress (by under-irrigating daily) allowed cotton plants to make osmotic adjustments that allowed the 50% treatment to yield only 17% less than the 100% treatment despite only receiving 31% less water.

Neutron probe readings showed that treatments receiving insufficient water to satisfy crop demand were able to extract water from part of the soilprofile that was not wet by irrigation. Use of this water allowed the 50% and 75% treatments to grow at similar rates to the 100% and 125% treatments for one month in peak growing conditions after the soil was saturated by rain in late December(1999/2000 season) while rain in November of the 2000/2001 season had a similar effect.

Neutron probe readings proved to be sensitive to small changes in applied water and hence are a useful scheduling tool for drip irrigated cotton.