Yearling Friesian steers, fitted with rumen fistulas, were grazed for 11 weeks on wheat or oat stubbles at a stocking rate of 1.2 steers ha--1. Following rain in the third week, half of each stubble area was sprayed with desiccant herbicide to establish weed-free stubbles. The stubbles initially yielded c. 2200 kg DM ha-1 of crop residues, plus up to c. 200 kg DM ha-1 of fallen grain. Green herbage was always <50 kg DM ha-1 on weed-free stubbles, but on weedy plots increased from c. 100-500 kg DM ha-l after the fourth week. On weed-free stubbles, the diet selected contained 80-100% dead plant material, with mean OM digestibility (in vitro) ranging from 0.54 to 0.45, and from 0.63 to 0.52 for wheat and oat stubbles respectively. On weedy plots, at levels of green herbage > 100 kg DM ha-1, cattle diets contained 60-70% green herbage, with a mean OM digestibility of 0.66. Over the whole experiment, estimated mean intakes of digestible organic matter (DOM) varied from 2.0 to 3.8 and from 3.5 to 7.5 kg day-1 on the weed-free and weedy stubbles respectively, and were consistently higher on the oat stubble than on the wheat stubble. The maintenance requirement of DOM (12 g per kg live weight) was acheived with green herbage levels of 10-25 kg DM ha-1. Steers grazing the weed-free stubbles generally maintained weight, while those on the weedy stubbles gained 34-48 kg during the experiment.Mean concentrations of ammonia in the rumen were always < 100 and often <50 mg N l-1 on the weed-free stubbles except at the end of the experiment, while those on the weedy stubbles were generally > 100 mg N l-1. Concentrations of rumen volatile fatty acids (VFA) were generally <80 mmol l-1, and fell to <50 mmol l-1 on the weed-free stubbles in the second half of the experiment.The observations are discussed in terms of the utilization of the dead plant material in crop stubbles by grazing cattle, and likely responses to supplements.