B. Papp szerk.: Studia Botanica Hungarica 38. 2007 (Budapest, 2007)

Csontos, Péter, Tamás, Júlia; Balogh, Lajos: Thousend-seed weight records of species from the flora of Hungary, II. Dicotyledonopsida

ecological databases were established in western Europe (GRIME et al. 1981, THOMPSON et al. 1997, VAN SLÁGEREN 2003) to support these research works, as well as better planning of vegetation restoration projects. Building a database in Hungary on seed size and shape was also initiated (CSONTOS 1998). This database includes data of three different sources: data published in the literature, measurements of seeds received from botanical gardens and measurements of seeds collected in spontaneous populations in Hungary. In the present paper data of

227 dicotyledonous species of the Hungarian flora are published for what thousand-seed weights, based on new measurements, are first reported. MATERIALS AND METHODS Seeds or fruits of species were collected in the field at mature stage. For species with easy-to-harvest seeds several hundreds were collected, otherwise at least 100 seeds were taken. Whenever it was possible, attention was paid to collect pooled seed samples from several mother plants of the same stand. In case of fleshy fruits the seeds were cleaned in the lab and dried at room temperature. The seed samples were placed into labelled paper bags and stored at room temperature until the end of the vegetation period (early November). For measurements hundred-seed lots were prepared (damaged seeds were not considered). Three lots were measured for the majority of species, in some cases two or one. Up to 30 g the accuracy was 0.1 mg (using KERN-4

10 automatic chemical balance), whereas 0.1 g above 30 g (NAGEMA digital weighing machine). Thousand-seed weight of a species was calculated as the average of the seed lots measured. In case of 40 species the seeds were collected from spontaneous populations by colleagues of the Botanical Garden of the Ecological and Botanical Research Institute of the Hungarian Academy of Sciences (Vácrátót). These samples were subjected to the same procedure described above. To illustrate the use of seed weight data in vegetation studies, we made statistical comparisons of average seed weights of pairs of three distinct species groups (1: annuals and biennials, n = 37; 2: non-woody perennials, n = 132 and 3: woody perennials, n = 58) following the proposed method of THOMPSON and RABINOVITZ (1989). Since the logarithm of seed weight data for the third group did not follow Gaussian distribution the non-parametric KruskalWallis test was applied (INSTAT 1997). P < 0.05 was required for significant differences among group medians. RESULTS AND DISCUSSION The Appendix contains of the thousand-seed weight records of

227 dicotyledons of the Hungarian flora, including 36 plant taxa introduced or