Our objective was to evaluate the effects of 3 extracts (hexane, ether and ethanol) isolated from tarbush on intake of alfalfa pellets by sheep. In each of 2 experiments, 45 ewe lambs received 1 of 5 treatments for 5 d. Treatments were alfalfa pellets (CON) or alfalfa pellets plus ethanol carrier (CAR), hexane extract (HEX), ether extract (ETH) or ethanol extract t(ETOH). Treatments were isolated from tarbush leaves (36 kg) using a sequential extraction of hexane, diethyl ether and 100% ethanol. Extractions (3 kg of leaves in 7 L of solvent, constant mixing, 22 h per solvent) were filtered, solvents removed using a rotary evaporator and resuspended in ethanol. Lambs were offered 640 g of alfalfa pellets (DM basis) and intake was monitored during a 20-min interval each morning. Extracts were applied to alfalfa pellets at the same concentration as in tarbush in Exp. 1, and at 10-fold dilutions in Exp. 2. Lambs were adapted to handling and individual pen feeding for 10 d and maintained and fed alfalfa pellets (4.7% of BW, DM basis) as one group except during 20-min tests. All 3 extracts reduced intake by lambs in Exp. 1 ( P < .0001) compared to CON or CAR. Mean intake during the 20-min period was 361, 393, 204, 212 and 228 g for CON, CAR, HEX, ETH and ETOH, respectively (SEM=28.9). Intake did not differ between CON and CAR, nor among HEX, ETH and ETOH. Mean intake in Exp. 2 was not different (468, 455, 389, 381 and 431 g for CON, CAR, HEX, ETH and ETOH, respectively; SEM =30.5; P =.187). Three fractions containing mixtures of compounds of increasing polarity were equally effective in decreasing intake of alfalfa pellets at concentrations encountered in intact tarbush plants; thus, several compounds are probably involved in the low palatability of this shrub exhibited by livestock.