Results

Choice of oviposition spots: Apollo

Oviposition behaviour was observed at 12 different spots, 8 of which had at least one egg for a total of 16 eggs. Oviposition substrates included juniper bushes (5 times), reindeer lichen (2 times) and Iceland moss (1 time). The probability of egg occurrence decreased with increasing proportions of fallen wood (e.g., sticks, bark, dead trees) on the ground within 10 m from the centre of the study plot (p = 0.011). In addition, the cover of < 5 cm field vegetation within 4 m from the centre was also included in the final model due to decreasing the AIC, despite only showing a weak negative relationship with egg occurrence (p = 0.18).

The two variables that together had the highest significance on the occurrence probability of eggs laid by Apollo females: a) Field vegetation < 5 cm (4 m), and b) Fallen wood total (10 m). Hollow circles = individual data points, dashed lines = upper and lower limits of 95% confidence intervals.

Choice of oviposition spots: Large Blue

Comparisons of the probability of oviposition on four different bud developmental stages. Same letters = no significant difference, different letters = significant difference.

Eggs were found at only 3 of the 17 spots where oviposition behaviour occurred. For this reason, data collected at all instances that this behaviour occurred was included in the analysis. The probability of oviposition behaviour increased with increasing cover of host plants (p = 0.07), especially those with buds of developmental stage 2 (< 10 % opened, p = 0.012), within 0.8 m from the centre of the plot, and decreased with increasing < 5 cm field vegetation cover (p = 0.12). Moreover, the behaviour occurred significantly more often on developmental stage 4 (red bud) compared to stage 1 (bud primordium, p = 0.02).

The three habitat variables that together had the highest significance on the probability of oviposition behaviour: a) Host plants (0.8 m), b) Bud stage 2, and c) Field vegetation < 5 cm (4 m).

Olfactory & sunlight orientation

22 Apollo larvae were tested in total. For the olfactory orientation tests, no significant difference in the mean final distance to the host plant clusters could be found whether the larvae were released downwind or upwind from them (p = 0.71).

Mean final distance of the Apollo larvae to the host plant cluster relative to the starting point (0 cm) when released downwind or upwind. Negative values = longer distance, positive values = shorter distance.

For the sunlight orientation tests, 18 larvae had a final position in a sunny spot while only 4 had a final position in a shaded spot. In other words, the proportions of larvae were 0.82 in sunny spots and 0.18 in shaded spots respectively, which were significantly different from the expected proportion of 0.5 (Chisq = 8.91, p = 0.003) if movements had been random.

Proportions of Apollo larvae with a final position in either a sunny or a shaded spot. Dashed lines represent expected proportions of 0.5 in both spot types.

Flight patterns and habitat preference

208 Marsh Fritillaries were studied in total. Out of all the individuals that completely crossed over the edges of the study plot, 22 crossed the grazed edge and 127 crossed the interior edges, which constitutes proportions of 0.15 and 0.85 respectively. The proportion of butterflies that crossed the grazed edge was significantly lower compared to the expected proportion of 0.25 (Chisq = 8.32, p = 0.004).

The number of individuals that returned to the plot after crossing over the edges was 6 for the grazed edge and 47 for the interior edges. This constitutes proportions of 0.11 and 0.89 for grazed and interior edges respectively, which meant that significantly fewer butterflies than expected returned after crossing over the grazed edge (Chisq = 5.29, p = 0.02).

Proportions of Marsh Fritillaries that a) crossed over grazed and interior edges without returning, or b) returned after crossing over grazed and interior edges. Dashed lines represent expected proportions.

The number of butterflies that visited a certain square in the study plot was found to increase with increasing host plant cover (p < 0.0001) and the distance to the grazed edge (p = 0.023). When tested separately, nectar plant cover also showed a significant positive relationship with the number of visiting butterflies (p < 0.001), but this variable did not improve the AIC of the final model.

The two variables that together had a significant effect on the number of Marsh Fritillaries visiting certain squares in the plot: a) Host plants (%), and b) Distance to grazed edge (in squares). Black dots = individual data points, gray areas = 95 % confidence intervals.