Free-ranging reindeer (Rangifer tarandus L.) eat a diverse diet of graminoids, herbs, shrubs, woody plants, mushroom, lichen and mosses. DNA metabarcoding is an established method in dietary studies and through e.g. non-invasive faecal collections it allows identification of the complete diet. However, there are methodological shortcomings regarding the application of DNA metabarcoding, especially for species with highly diverse diets and complex digestive systems such as the ruminant reindeer. In an effort to reveal the advantages and limitations of DNA metabarcoding, two experiments were conducted during this thesis. In the first experiment, a controlled feeding trial was carried out at UiT - The Arctic University of Norway, three female semi-domesticated reindeer (R. t. tarandus) were fed (i) pelleted reindeer food (RF-80, containing a known mixture of eleven species), (ii) birch (Betula pubescens Ehrh.) and (iii) increasing amounts of lichens (mainly Cladonia stellaris (Opiz) Pouzar and Vezda). The methodological questions were: 1) Does DNA metabarcoding detect all species in a known diet and does it allow identification to species level for all species?; 2) For how long is DNA of a particular food item present in the digestive system?; and 3) How well is relative food biomass represented by output sequences? In the second experiment, samples were collected from different sections of the gastrointestinal (GI) tract of four male semi-domesticated reindeer on natural pasture in Finnmark, Norway to answer: 4) Does DNA detectability change through the GI tract? The controlled feeding experiment detected eleven of the thirteen fed species, and the specific primers used in this experiment detected a higher diversity of species present in the diet than the general eukaryote primer. However, the short specific primer increased the number of wrongfully assigned taxa. DNA of B. pubescens was detected within 12 hours and it was present in the faeces for minimum 26 days after feeding (Question 2). It was possible to see a correlation between the fed biomass and the proportion of DNA sequences of C. stellaris, when using a general eukaryote primer (Question 3). In answer to Question 4, the data from this study indicate that for seed plant, eukaryote and fungal DNA, there is higher species richness in the distal colon than in the rumen, whilst the opposite is found for bryophyte DNA. Higher diversity is found in the cecum and distal colon for seed plant and eukaryote DNA, in the rumen for fungal DNA and possibly also in the rumen for bryophyte DNA. These results imply that non-invasive faecal samples are better to use, compared to invasive rumen samples, when studying species richness of seed plants in the diet or the overall diversity in the diet, regardless of targeted taxonomic group.