Noninvasive sampling of mammalian hairs for surveying their populations and for providing density estimations is widely applicable in wildlife ecology and management. However, the efficiency of the method may differ depending on the species or local circumstances. We modified a method of hair trapping from free-ranging Eurasian lynx (Lynx lynx) to collect DNA samples to work in a low-density population. We constructed hair traps based on a device developed for Canada lynx and assessed their effectiveness in Białowieża Forest, Poland. We set 153 hair traps baited with beaver (Castor canadensis) castoreum and catnip oil at points previously used by lynx for scent-marking. We conducted the study in 2 consecutive winter and summer seasons during 2003–2004. Lynx rubbed 22–46% of the hair traps in 5 different trapping sessions. Lynx were more likely to rub hair traps set directly at scent-marking points on conspicuous marked objects than when they were set some distance (1–3 m) from the marked objects. Efficiency of hair-trapping sessions increased from 30.1 to 46.4% after selecting the most likely points. The percentage of traps visited and rubbed by lynx was higher in winter than summer in 2 consecutive years (30.1 vs. 22.2% and 46.4 vs. 23.3%, respectively), which may be related to mating behavior. This method proved efficient for monitoring low-density Eurasian lynx populations.