In experiments based on ruminants' individual dry matter intake (DMI) assessment, several external markers can be used to estimate faecal output when total faeces collection is not possible. However, preparation of the markers to be administered and analytical procedures used for marker content determination are time-consuming thus strongly limiting the number of animals involved in the experiments. In this paper, polyethylene glycol (PEG, molecular weight 6000 da) was tested as a faecal marker. Four trials were conducted on dry, non-lactating ewes kept in digestibility crates that allowed individual measurements. The overall experiment was designed to assess the major factors that could lessen the effectiveness of this method, assuming that the use of grab samples of faeces is sufficient. Trial 1 was designed to test two levels of PEG (20 and 40 g/day) administered in two equal amounts. Trial 2 was designed to test the effect of either a single morning (0800 h) dose (20 g/day) or a twice daily administration (0800 and 1600 h) of the same fractionated dose. Trial 3 was designed to test a 20 g/day dose of PEG administered once daily to ewes fed with hays of different qualities: medium (MH) and low (LH). In trial 4, a lower dose of PEG (10 g/day) was administered once a day to ewes fed with fresh oat-vetch forage. It was demonstrated that PEG could be precisely estimated (average prediction error53.47 g/kg) with near-infrared reflectance spectroscopy (NIRS). On the basis of the four trials, it has been proved that PEG administration (20 and 40 g/day) did not significantly affect the DMI of ewes fed dry diets (trials 1, 2 and 3), whereas there was an unexpected increase of DMI for ewes fed exclusively with green feed (trial 4) without DM digestibility modification. Providing PEG as a single dose (0800 h) or split into two equal parts (0800 and 1600 h) did not alter the estimated DMI. Considering the interest of grab sampling, there were clear variations of PEG in faece