Tearing in rubber is known to initiate from an inherent flaw present in the rubber (Gent 2005). When the rubber is stretched, the local stress in the vicinity of a flaw is intensified. Once the local stress reaches a critical level, the rubber tears by extension of the crack. It has been widely reported that the rate of crack growth in rubber is determined by a characteristic energy per unit area of the fracture surface created, often known as the tearing energy or the strain energy release rate (Thomas 1994). This is defined as

whereW is the total elastic strain energy in a specimen of thickness h measured in the unstrained state and c is the length of a crack. The suffix l denotes that no external work is done at the system boundaries to create new crack surfaces. In the case of a trouser tear test piece in Figure 1, the tearing energy is given by

where F is the applied tearing force, h is the specimen thickness, λ is the extension ratio in the legs, b is the total width of the specimen and W is the elastic stored energy in the legs of the specimen far removed from tear.W is determined from integration of a tensile stress-strain curve at a strain that corresponds to the extension ratio in the legs of the specimen at the point

tearing. The relationship between the rate of tearing and the strain energy release rate is a material characteristic that is independent of test piece geometry (Greensmith and Thomas 1955). In the trouser tear specimen, a tearing force applied to the legs produces tearing at the crack tip. The region of each of the legs is essentially in uniaxial extension with the corresponding extension ratio, λ. The crack length increases by dc which results in an increase in the volume in the region of uniaxial extension in the legs (by 2a · h · dc). The separation l of the clamps is increased by

If the tearing is smooth and continuous the average crack growth rate can be determined from the rate of separation of the clamps, S by the relation

where l is the separation of the clamps of the test machine, t is time and c is the crack length. Figure 2 shows the four different types of measured force versus time response that are typically observed during tearing of rubber samples. Steady tearing, as depicted in Figure 2(a) and (b) can easily be interpreted using the framework described above as the tearing rate is constant and the relationship between strain energy release rate and the crack growth rate can be evaluated to characterize the tearing behaviour. However, for a lot of rubber materials the behaviour can be either stick slip in nature, as shown in Figure 2(c) or even knotty in behaviour as shown in Figure 2(d).