Description of the subject. Specific densities of 13 samples of eight vegetable feed ingredients and of blood meal have been established. Objectives. The aim of the study was to use the parameter specific density for classification of a selection of feed materials. The range of specific densities for each of the tested materials could support the identification of ingredients in a compound feed. Ingredients in mixtures could be separated when different specific densities apply to these ingredients. Method. The share of the flotate and of the sediment of each of the tested materials has been established in duplicate along a range of increased specific density of the solvent mixture, from 1.22 g·cm-3 up to 1.52 g·cm-3, initially in four steps. Intermediate densities have been used where appropriate for collecting more detailed ranges. The specific densities which resulted in approximately 100% sediment and in approximately 100% flotate are used as indicators for the range of the specific density of that material. Results. The materials used appeared to show three different ranges of specific density: corn gluten and blood meal (1.12 – 1.32 g·cm-3), corn distillers’ grain, rapeseed kernel meal, soya kernel meal and wheat middlings (1.25 – 1.42 g·cm-3), and palm kernel meal, citrus pulp and beet pulp (1.35 – 1.55 g·cm-3). The specific densities range from those found for vertebrate muscles and insect fragments (below 1.3 g·cm-3) up to the specific density of bone fragments (exceeding 1.62 g·cm-3). Four mixed samples, each existing of two different ingredients at a ratio of 1:1 (w/w) were tested with solutions at different specific densities. The ingredients of each mixed sample appeared to separate in the expected fractions, predominantly in either the flotate or in the sediment. Conclusions. The experiments show that separation based on specific density can assist in the identification of feed ingredients in one or more combinations of sediment and flotate fractions.