Maintaining dry matter intake (DMI) during the last 2 weeks of gestation is difficult, and a failure to do so is reported to have negative implications on postcalving DMI (Grummer, 1995) and subsequent milk production (Bertics et al. 1992). DMI at the end of gestation, but excluding the last 1–2 d before parturition, can decrease by up to 30% (Lodge et al. 1975; Bertics et al. 1992; Dewhurst et al. 2000). The extent of the decline varies and may be influenced by a number of factors, such as level of feeding precalving (Lodge et al. 1975; Murphy, 1999; Dewhurst et al. 2000) or amount of non-structural carbohydrates in the precalving diet (Coppock et al. 1972; Forbes, 1995). Most studies were undertaken with cows receiving total mixed ration (TMR) or conserved forage and concentrates. However, this may have been one of the reasons for the witnessed depression. Increasing the proportion of rapidly fermentable carbohydrates or alternatively decreasing the fibre concentration in the precalving diet appears to exacerbate this decline (Coppock et al. 1972; Minor et al. 1998). Research suggests that ruminants may actively select against concentrates as plasma oestrogen concentrations increase, such as happens in the weeks preceding parturition (Forbes, 1995). When hay was offered ad libitum, increasing concentrations of oestrogen (up to 400 μg infused intravenously) had no effect. However, when a complete pelleted diet was offered, a 90-μg infusion caused a lower feed intake, followed by a significant increase after treatment was stopped. The lack of depression in precalving DMI reported by Coppock et al. (1972) and Vazquez-Anon et al. (1994) when forage was the principal component of the diet, supports this.