In order to measure the digestibility of feed with an indicator method two trials were realized on the dairy research farm "Karkendamm". The first trial was to test the suitability of TiO2 as an indicator to determine the digestibility in cows. The rate of recovery of the indicator was 98.8% (95.5%-101.9%). The concentration of TiO2 of the rectal samples in the morning (daily or every second day) corresponded very well with the values from total faeces collection. The samples taken in the evening showed lower concentrations. The estimation of digestibility of organic matter led to the same values with the total collection or the indicator method with morning samples. The digestibility derived from evening samples were 1.7-2.3%-units digestibility derived from evening samples were 1.7-2.3%-units lower in comparison to total collection. In the second trial the digestibility derived from the in vitro method "Hohenheimer-Futterwert-Test" and from a digestion trial with 30 cows using the indicator method were compared. The digestibility calculated from the date of the in vivo trial was 9.9%-units lower than the one from the "Hohenheimer-Futterwert-Test". The depression of digestibility was due to the variation in concentrate feeding between cows. If the cows got less than 8 kg concentrates the depression in digestibility was 1.9% when increasing the level of feeding one unit of maintenance requirement. If the cows got more than 8 kg concentrates the depression was 3.1%-units. It could be concluded that the "Hohenheimer-Futterwert-Test" underestimates the digestibility if the level of feeding is increased to a certain extent.