Oviposition by mutualistic seed-parasitic pollinators and its effects on annual fitness of single- and multi-flowered host plants

The balance of intimate relationships between plants and seed-eating pollin ators can depend on pollinator behaviour in relation to floral characters, such as flower size and flower number. Here, we examined how pollinator ovi position in relation to these traits affected annual fitness (seed output) of single- and multi-flowered Trollius europaeus along altitudinal gradient s in subarctic Sweden and the French Alps. Small flies (Chiastocheta spp.) pollinate T. europaeus and their larvae feed on developing seeds. Assuming that late flowers in multi-flowered plants attracted flies to the earliest flower on the same plant, we expected more eggs and higher seed predation i n early multiple flowers than in single flowers. Relative seed predation wo uld thereby increase with flower number. Both in Sweden and the Alps, more eggs were placed on large flowers. Early multiple flowers were slightly lar ger than single flowers, and about twice the size of secondary flowers. As a result, and possibly combined with the effects of secondary flowers, earl y multiple flowers attracted more ovipositing flies and experienced relativ ely higher seed predation. However, this did not generally result in higher seed predation of multi-flowered hosts. Multiple flowers had greater seed output than single flowers at all altitudes, also in the high alpine and su barctic sites, where single flowers were more abundant. We hypothesise that the distribution of multiple flowers generally is enforced by environmenta l factors, rather than by fly-host plant interactions, because only very ra rely tin triple-flowered alpine plants) was seed predation disproportionate , and the relationship skewed to the disadvantage of the host. The outcome of the mutualistic interaction was often similar in alpine and subarctic po pulations, but the underlying factors were different. Subarctic flowers had high abortion and low predation rates, while alpine flowers experienced th e reversed situation. The higher fly abundance in the Alps suggests a more intense mutualistic interaction than in Sweden. Despite varying ecological and environmental conditions at these sites, the mutualistic relationship w as generally in balance. However, when it was unbalanced, this could be exp lained by fly behaviour in response to floral traits, and by differences in fly abundance.