Crack growth rates (r) were measured in pure shear test specimens as a func
tion of strain energy release rate (G) for a non-crystallising SBR elastome
r. Measurements were made as a function of: extent of swelling with Dibutyl
Adipate; carbon black content; and crosslink density. In some cases experi
ments were carried out over a range of temperatures. In most cases the resu
lting G versus r plots showed a clear transition from rough to smooth crack
surface behaviour with increasing crack growth rate, with an intervening s
lip/stick region. In the high speed/steady tear/smooth region the value of
G necessary to drive a crack at a given rate was determined largely by the
magnitude of the visco-elastic losses in the crack tip region, increasing w
ith: decreasing temperature; increasing molar mass between crosslinks; decr
easing extent of swelling; and increasing carbon black content. However G w
as independent of specimen thickness in this region suggesting that crack t
ip effects were minimal. In the low speed/rough region changes in the magni
tude of G with materials and temperature/rate variables could not be explai
ned by changes in visco-elastic loss alone. Furthermore the magnitude of G
increased significantly with increasing specimen thickness. This suggested
that in this region cavitation ahead of the growing crack tip resulting fro
m dilatational stresses determined the crack tip diameter, and hence the ma
gnitude of G. (C) 2000 Kluwer Academic Publishers.