Dynamic conservation of forest genetic resources (FGR) means maintaining the genetic diversity of trees within an evolutionary process and allowing generation turnover in the forest. We assessed the network of forests areas managed for the dynamic conservation of FGR (conservation units) across Europe (33 countries). On the basis of information available in the European Information System on FGR (EUFGIS Portal), species distribution maps, and environmental stratification of the continent, we developed ecogeographic indicators, a marginality index, and demographic indicators to assess and monitor forest conservation efforts. The pan-European network has 1967 conservation units, 2737 populations of target trees, and 86 species of target trees. We detected a poor coincidence between FGR conservation and other biodiversity conservation objectives within this network. We identified 2 complementary strategies: a species-oriented strategy in which national conservation networks are specifically designed for key target species and a site-oriented strategy in which multiple-target units include so-called secondary species conserved within a few sites. The network is highly unbalanced in terms of species representation, and 7 key target species are conserved in 60% of the conservation units. We performed specific gap analyses for 11 tree species, including assessment of ecogeographic, demographic, and genetic criteria. For each species, we identified gaps, particularly in the marginal parts of their distribution range, and found multiple redundant conservation units in other areas. The Mediterranean forests and to a lesser extent the boreal forests are underrepresented. Monitoring the conservation efficiency of each unit remains challenging; however, <2% of the conserved populations seem to be at risk of extinction. On the basis of our results, we recommend combining species-oriented and site-oriented strategies.

Dormancy release as influenced by duration of outdoor winter chilling in Florence (Italy) was studied under different photoperiodic and temperature treatments in collected twigs of two European (Ulmus glabra Huds. and Ulmus minor Mill.) and four Asian (Ulmus pumila L., Ulmus parvifolia Jacq., Ulmus macrocarpa Hance and Ulmus villosa Brandis) elm clones. Photoperiod had no effect on dormancy release, and there was no evidence that photoperiod affected bud burst during quiescence in the studied elm clones. Thermal time (day degrees >0 degrees C) to bud burst decreased in all the clones with increasing outdoor chilling. Although all the clones exhibited a rather weak dormancy, they significantly differed from each other. Dormancy was released earlier in the Asian than in the European clones, and the clones could be ranked from the U. pumila clone (very weak and short dormancy) to the U. minor clone (relatively stronger and longer dormancy), the other clones being intermediate. In all the clones except U. minor, the observed decrement in thermal time to bud burst was efficiently explained as an inverse exponential function of the number of chill days < or =5 degrees C received outdoor in autumn and winter. Endodormancy, as measured by the single-node cuttings test, was weak and short in all the clones. The latter result suggests that correlative inhibitions were largely responsible for preventing bud burst during winter in these elm clones.

Ulmus laevisPall. is a broad-leaved deciduous tree with a central and eastern European distribution. We describe the development of six polymorphic microsatellite markers for this species. These markers were also tested for utility in U. americana , U. glabra, U. minor and U. pumila . One additional marker gave ambiguous results in U. laevis but amplified clearly in three other species. In U. laevis , the number of alleles observed per locus ranged from two to nine. Five loci showed polymorphism in at least one of the nontarget species tested.

Reaction norm variation between and within populations of two rare plant species, Vicia pisiformis and V. dumetorum (Fabaceae) A split-family design with progenies grown across a steep temperature gradient was used to examine the patterns of reaction norm variation in two locally rare, thermophilous plant species, Vicia pisiformis and V dumetorum. We predicted (i) a negative association beiween the amount of genetic variation and the magnitude of response across temperatures; (ii) broader ecological tolerance and higher levels of genetic variation for reaction norm parameters in the broadly adapted V dumetorum than in the ecologically restricted Vpisiformis; and (iii) higher performance across environments for central populations than for geographically marginal populations. Patterns of variation in 10 vegetative characters, including measures of overall plant size, provided no evidence of a trade-off between plasticity and genetic variation. Our data indicate greater performance across temperatures and more divergent genotypes in V dumetorum than in V pisiformis. Central populations of V pisiformis performed better than conspecific populations from the marginal (northern) regions.

The amount and pattern of genetic variation was surveyed in two Swedish and three Czech populations of the rare perennial forest plant &pisifotmi~. This species has a mainly easterlycontinental European distribution and has few and small populations in Sweden. It is classified as 'vulnerable' on the Swedish Red Data lit. Seeds from natural populations were collected and grown under controlled conditions in growth chambers. The variation was estimated in growth and fecundity traits and with Random Amplified Polymorphic DNA, RAPD. Low inter-and intra-population variation in R4PD-markers was found using 11 primers, with a similarity index (Wetton) for the families of 0.98. In contrast, multivariate analysis of variance showed significant morphological differences withim and between populations. Also in the univariate ANOVAs, a number of the traits showed si@cant betweenand/or within population differentiation. Cluster analysis for the morphological traits and RAPDs (UPGMA) did not structure the variation of families in accordance with their geographical distance. A Mantel test based on comparisons between Mahalonobis and Jaccard distance for morphological and RAPD data, respectively, did not reveal any significant correlation between the two matrices. It is concluded that if a genetic conservation program is to be applied on Vicia pi$mzi, different sampling strategies are needed to capture morphological vs RAPD variation. This is, to our knowledge, the first investigation that compares RAPD and morphological variation in a threatened plant species. 0 1997 The Linnean Society of London ADDITIONAL