Different Regulated Deficit Irrigation (RDI) scheduling practices were investigated on an Oxisol (Rhodic Ferrasol) with an effective rooting depth of 900mm located in Nelspruit, Mpumulanga in the summer rainfall area of South Africa. The experimental site receives an annual rainfall of 750mm in the form of thunderstorms. The average daily maximum temperature is 30°C in January (mid summer) and 23°C in July (mid winter). The average minimum daily temperature varies between 18°C in January and 8°C in July. The area is frost free. The Citrus cultivar, cv. Olinda Valencia, on cv. Cairn ruff lemon rootstock with a tree spacing of 7 x 7m was used in the trail. The twenty-six year old trees are still in good condition without greening symptoms and have an average yield of 200 kg/tree. Irrigation water with an excellent quality, classified as a class C1S1 (USDA), was used to irrigate the trees with two micro jets (each covering 270°) placed 30cm from the stem on both sides in the row. The growing season was divided into three phases during which different regulated deficit irrigation treatments were applied. Phase I, the cell division phase, started after full bloom (1st August) and continued to 56 days after full bloom (15 November). Phase two followed phase until 200 days after full bloom (15th March). The fruit development levelled off during phase three which ended when the fruit matured and were harvested. RDI practices were adapted to investigate the effect of maximum stress during one or more of the different growth phases. The soil profile was filled to field water capacity at the beginning of a stress treatment and no irrigation water was applied until the end of the growth phase. Different combinations of stress treatments during the different growth phases were applied and the effect on growth, fruit yield, size and quality were determined. The water content of the non-stress treatments in the different growth phases was kept between field capacity and 50% of readily available water content. In the control treatment no stress was applied in any of the three growth phases. The total seasonal crop water requirement for Citrus in this investigation was 26 m3/tree or 10 100 m3/ha. The average rainfall efficiency was estimated as 33% and rainfall contributed to 30% of the annual crop water use. The stress period during the rapid fruit development stage (phase II) had the greatest detrimental effect on fruit size and yield. Water stress during growth phase I caused a delay in flowering and trees had a tendency to flower following a period of severe stress. It was evident from the results that flowering can be manipulated with severe stress treatments. If trees were only stressed during growth phase three, yield was not drastically reduced as was expected. If water is scarce this will therefore be the safest phase during which to apply stress. The most critical phases in which Citrus trees are most susceptible to stress are during flowering and the rapid fruit growth phase. Both yield and fruit size were reduced due to water stress during these phases.