Fracture of rubber begins at flaws or inclusions where local stress exceeds a critical level. Recently, a study has been undertaken to investigate the effect of inclusions on the fracture property of rubber. The inclusions studied are dirt from guayule rubber of particle size between 18 and 6000/~m. A controlled amount of dirt of definite size has been incorporated in the rubber matrix. It is worth mentioning that these dirt particles remain present even after the extraction of rubber from shrub. Fracture of guayule rubber containing different sizes and amounts of "dirt", which are basically plant tissues in nature, has been reported elsewhere [1]. Thomas [2] and Gent [3] have treated the fracture of elastomers in detail and Greensmith [4] has reported tensile and fatigue failure of natural rubber containing sand and sawdust. He observed a reduction in strength properties with the incorporation of these foreign materials. Gent et al. have investigated the tensile failure of silicone rubber containing a glass bead [5]. In the above investigations, however, the surface morphology around a flaw has not been studied. Fractography of several elastomers under different modes of failure namely tensile, tear, fatigue and abrasion has been reported before by us [6-8]. In all these studies, it has been observed that different modes of failure generate typical fracture surfaces. In this letter, the results of my observations of tensile and fatigue fracture surfaces ofelastomers containing inhomogeneous inclusions are reported. Mix formulations are as follows: natural rubber 100%; zinc oxide 6%; stearic acid 0.5%; mercaptobenzothiazole 0.5%; sulphur 3.5%; dirt 0%, 3.0%. 3% dirt has been added to the sample. Mixing and compounding have been done on an open mill. Hence, there is some breakdown of the dirt, reducing the particle size ranges from 45 to 6000/~m to 45 to 1000/~m. The small amount of dirt does not have any effect on scorch time and optimum cure time. Curing was done at 150°C at the optimum cure time. Dumbbell specimens for tensile and fatigue testing were used. Tensile testing was carried out in an Instron machine (ASTM D 412-75 method) and the fatigue experiment in a Monsanto fatigue-to-failure tester (frequency, 100 cycles min ~ ). The samples were sputter-coated with gold immediately after fracture and the fracture surfaces were examined within 24 h of test. A JEOL SMU3 scanning electron microscope was used for the study. The tensile fracture surface of guayule rubber is shown in Fig. 1. Few tear lines running from one end Figure 1 Tensile fracture surface of guayule rubber containing no external dirt: general fracture surface.