In the past few years in Europe grass seed production declines. This tendency is especially pronounced in meadow fescue. Seed shedding and therefore yield losses are the problem in seed production. This can be reduced if seed of higher moisture contents is harvested. The impacts of drying temperatures of 70, 60, 50, 40 and 22°C on changes in dormancy and germination of seed harvested with moisture contents of 45, 35 and 25% were observed in the present study. The analysis was done immediately after seed drying, then three months later and eight months later. Seeds with the moisture content of 45% that were dried at 70°C were not dormant at all after harvest, but seeds were damaged, which resulted in reduced germination. Drying temperatures of 40°C and 50°C resulted in maximum germination of seed harvested with 45% moisture after three months. After eight months the best germination of all seeds was obtained at 22°C and 40°C. The seed ageing test confirmed faster deterioration of seeds harvested with higher moisture contents. Seed harvested with 25% moisture and dried at 22°C is the most suitable seed for longer storage.

This research was conducted to determine the effect of seed rates in mixtures of pea + oat on the green forage and silage protein fractions evaluated by Cornell net carbohydrate and protein system (CNCPS). Experiment was established in autumn of 2012, on October the 20 th and plant samples were taken in spring 2013 at forming the first pods on 2/3 plants of pea at Institute for forage crops, Kruševac, Republic of Serbia, using five different mixture rates of pea and oat crops (100% pea + 0% oat; 0% pea + 100% oat; 25% pea + 75% oat; 50% pea + 50% oat and 75% pea + 25% oat). After harvesting pea:oat mixtures were treated with bacterial inoculant and ensiled in anaerobic jars for 45 days. Green forages and silage samples were analyzed for DM (dry matter), CP (crude protein), primary protein fractions-TP (true protein), NPN (non protein nitrogen), IP (insoluble protein), SolP (soluble protein), NDICP (neutral detergent insoluble crude protein) and ADICP (acid detergent insoluble crude protein) and protein fractions by CNCPS. An analysis of variance found statistically significant differences among mixture rates for all variables, except IP and SolP. Silage from monoculture pea had the highest NPN (696.2 g kg -1 CP) and SolP (713.8 g kg -1 CP), followed by the mixture of pea with oat 75:25 (662.5 and 653.4 g kg -1 CP, respectively). Analyzing the CNCPS protein fractions of pea:oat silages it was found that silage from pea monoculture contained the highest PA fraction (non protein nitrogen, immediately degraded in the rumen) and that PA fraction increased with increasing pea ratios in silages, which was a direct reflection of their high NPN and SolP in green forages and in silages. Because of those facts 25:50 and 50:50 pea:oat mixtures silages could be recommended for ruminant feeding. The investigated bacterial inoculant can increase the TP content, as well as PB 1 (true protein rapidly degraded in the rumen) and PB 3 (slowly degraded True Protein in the rumen, because it is associated with the cell wall) fractions.

Products of metabolism that are directly involved in growth, development and reproduction and also secondary products of metabolism were studied in 16 natural populations of three species of the genus Trifolium (T. repens -four populations, T. alpestre -three populations, T. pannonicum -nine populations), collected from the central Balkans region, i.e. territory of Serbia. Statistical analysis showed the presence of variability in all species of the test material. The best quality of dry matter, estimated by the ratio of crude protein and crude fiber, was achieved in the populations of white clover. In this species, average protein content was 216.4 g kg . It can be concluded that the selection of these populations may lead to creation of different varieties of fodder crops that could give safe animal feed and would be suitable for growing under adverse conditions of mountainous climate.

This paper describes the content of the "Vegetation Database Grassland Vegetation of Serbia" (GIVD ID EU-RS-002). The grassland database is held at the Department of Botany, Faculty of Agriculture, Belgrade, Serbia. It contains all available published relevé material (relevés originate from published sources and grey literature -theses, survey reports) from the territory of Serbia. Presently 4,252 relevés are available in the database which runs offline with TURBOVEG. All registered plant communities from phytosociological classes Molinio-Arrhenetheretea, Festuco-Brometea and Festucetea vaginatae are covered. The database is used for archiving relevé material, for preparation of classifications and other analyses of vegetation of Serbia and neighbouring countries. Sampling periods: Keywords

The aim of this paper is to present a numeric-statistical model by which it is possible to evaluate the quality of the observed grassland upon floristic and vegetation assessment. Thanks to this new methodological approach, the impact of each individual plant species on overall quality of the grassland could be estimated. The main goal was to determine species which significantly determine the pastoral value of the grassland. The quality (pastoral value) of the grassland was calculated using the numerical values of quality index of each individual species of each relevé of the community. For testing this numerical method, the total of 11 relevés of 4 grassland communities of hilly-mountainous area of Mt. Kopaonik was used. Analyzed vegetation includes previously unpublished phytocenological relevés as result of our own field research. All analyzed relevés were obtained using the method of the Swiss-French phytocenological school. The results showed that the best and the worst quality were determined for the ass. Festuco-Brometum erectii, and the ass. Nardetum strictae, respectively. Species that highly contributed to good and bad quality of grassland were Arrhenatherum elatius, Festuca rubra, Dactylis glomerata, Trifolium repens, etc., and Carduus alpestris, Hieracium hoppeanum, Ornithogalum umbelatum, respectively Key words: semi-natural grasslands, numerical index of grassland quality, numerical assessment of grassland quality, pastoral value, cluster analysis, multivariate analysis.