From 2008 to 2012, under a joint funding by the Institut Français du Caoutchouc (IFC) and its affiliated societies (Michelin, SIPH and Socfinco), a network of large scale experiments has been set up and followed up in 5 agro-industrial plantations to study the influence of a gradient of fertilization of mature rubber trees on latex physiological parameters and rubber yield potential. Experiments were set up on clone PB260 (Socfindo Tanjung Maria and Aek Pamienke plantation in North Sumatra, Indonesia, SAPH-Toupah and SOGB-Grand Bereby plantations in Côte d'Ivoire) and on clone IRCA109 (RENL-Osse River plantation in Nigeria). During the first 3 years (2008/09 to 2010/11), the trees were tapped downward in S/2 d4. During the 4th year (2011/12), the trees were tapped upward in S/4 d4. Each experiment has been designed with a similar protocol: Split plot design with stimulation intensity as the main treatment and fertilizer rate as secondary treatment, with 4 replications. The 3 stimulation intensities (main plots) were INT1 (normal stimulation intensity adapted to the clone, according to Cirad stimulation recommendation grid), INT2 (intensified stimulation: INT1 + 2 stimulations/y) and INT3 (intensified stimulation: INT1 + 4 stimulations/y). The 4 fertilization rates (secondary plots) were IF0 (No fertilization control), IF1 (single dose as the standard Cirad recommendation : 50g N, 40 g P2O5, 90 g K2O per tree per year, applied in a single application 45 to 60 days after complete refoliation), IF2 (double dose : 100 g N, 80 g P2O5, 180 g K2O per tree per year) and IF3 (triple dose: 150 g N, 120 g P2O5, 270 g K2O per tree per year). The 12 main plots (4 replications x 3 stimulation intensities) were fitted to 12 tapping tasks selected for their homogeneity (girth and production) before trial implementation. After 4 years, maximum cumulative production (kg/ha) was always observed with a fertilized treatment (IF2 or IF3) combined with intensified stimulation (INT2 or INT