Dry matter production by crops is a function of the rate of establishment, size and duration of the canopy as well as the CO 2 fixation efficiency of that canopy. All of these factors can be affected by plant water deficits to a greater or lesser degree and can therefore be maintained by irrigation. One of the objectives of the drip irrigation research in Mauritius was to identify the soil and plant water conditions which must be maintained to prevent growth depression. The studies reported here, carried out in the dry north of Mauritius, examined the plant-water relationships of estate grown sugar-cane receiving varying amounts of irrigation supplied by drip irrigation. Irrigation supply was based on a Penman potential requirement modified with a crop factor appropriate to sugar-cane (ETc). Four treatments were compared, 1.5, 1.0 or 0.5 ETc as well as a rainfed treatment. Measurements were made of leaf water potential (ψ) and its components, osmotic potential (π) and turgor potential (P) in two cultivars S17 and R570. At the same time net photosynthesis, stomatal conductance and leaf or total extension growth were measured throughout several days in November and December 1986 and 1987. It is at this time of the year that the largest soil water deficits develop. ψ was very close to zero in all irrigated treatments before dawn and only in rain-fed treatments was it significantly lower. During daylight periods differences emerged between irrigated treatments. ψ fell to as low as - 1.7 to 1.8 MPa in rain-fed cane in dry periods and at this point photosynthesis, stomatal conductance and leaf extension growth were very low. Intermediate rates of photosynthesis, stomatal conductance and growth were observed in cane receiving 0.5 ETc. The two highest irrigation treatments were usually similar both in water relations, growth and gas exchange but occasionally the 1.5 ETc treatment had a slightly lower water potential and the physiological variables also had lower rates. It is presumed that this was caused by waterlogging and poor soil aeration which prevailed temporarily. In two studies of the development of water stress, leaf extension rates were much more sensitive to stress than photosynthesis or stomatal conductance. The best index of a requirement for water was therefore the depression of day-time leaf extension, which would even fall in well-irrigated treatments under high evaporative demand conditions. Recovery from water stress was quite rapid with photosynthesis, stomatal conductance and leaf growth recovering in a few days to rates which often exceeded those of unstressed cane.