Fræðaþing landbúnaðarins - 16.02.2007, Side 17

15 • Fræðaþing landbúnaðarins 4, 2007
enhancements by dominant or subordinate species.
Elevated CO, may not primarily exert a direct selection pressure but may act mainly
through changes in the physical environment (e.g. increased water availability (Morgan
et al., 2004)) or in the competitive environment of the individuals. Under elevated
C02 changes in the competitive environment could include (i) increased intensity of
competition related to the increased biomass produced or (ii) changed competition due
to altered species composition of the plant stand. Adaptation of plant populations to their
competitive environment as affected by nearest neighbours was observed (Turkington
and Harper, 1979; Turkington, 1989; Liischer and Jacquard, 1991; Luscher et al.,
1992; Expert et al, 1997) as well as for the intensity of competition as affected by the
frequency of disturbance (McNeilly, 1981; 1984).
Further insight into the significance of elevated CO, as a selecting force may be gained
by comparing populations from sites at which C02 had been increased during the
last decades (natural CO, springs) with populations from sites with ambient CO, but
otherwise comparable environmental conditions. Since the environmental factors (e.g.,
soil pH, concentration of sulphur in the soil) in the vicinity of natural C02 springs often
differ considerably from those in the surrounding areas, this approach may be difficult to
implement. An altemative solution would be to grow isolated populations at elevated and
ambient C02 concentration in long-term fumigation experiments (some years) for many
generations and to compare the populations at the end of the experiment (Wieneke et al.,
2004). This would be feasible with plants with short life cycles (e.g., Poa annua).
Analyses of genetic variability to elevated C02
In the Swiss FACE experiment, intraspecific variability in the response to elevated C02
was determined from the phenotypic (yield) response of cloned genotypes (9-14 per
species) from 12 different grassland species. The response of genotypes was examined
in plant communities under field conditions for three years. The overall results for
the species L. perenne, L. multiflorum, Arrhenatherum elatius, Dactylis glomerata,
Festuca pratensis, Holcus lanatus, Trisetum flavescens, Rumex obtusifolius, R. acetosa,
Ranunculus friesianus, T. repens and T. pratense did not reveal statistically significant
intraspecific variability in the response to elevated C02 (Lúscher et al., 1998) when
analysed with standard ANOVA (analyses of variance) techniques. In all the overall
ANOVAs of the different years and the seasons in any one year, intraspecific variability
in the response to C02 was far to be statistically significant (p>0.95). In such overall
ANOVA analyses, significant intraspecific variability in the response to CO, (CO,
x genotype(species) interaction) of some individual species may be masked by the
other species, which contribute only little to the CO, x genotype(species) mean square.
However, no significant intraspecific variability was detected even when tested for
each species individually. Significant CO, x genotype interactions within a species may
occur when (1) genotypic variation at elevated CO, is larger than at ambient CO,, (2)
genotypic variation at elevated CO, is smaller than at ambient C02 or (3) the ranking of
genotypes is different for the CO, levels (Figure 1) (Schmid et al., 1996). However, CO,
x genotype interactions are tested under the assumption that variance is the same in both
C02 treatments and, thus, (1) and (2) would appear only partly in the C02 x genotype