Pollination ecology of Tasmanian leatherwood (Eucryphia lucida Eucryphiaceae Labill.) and the impacts of hive honeybees

Tasmanian leatherwood (Eucryphia lucida Labill.) is a tall (up to 30 m) native tree occurring as a canopy co-dominant in Tasmania's cool temperate rainforest. A large proportion (ca. 37%) of E. lucida's total distribution occurs within the Tasmanian Wilderness World Heritage Area. The nectar of E. lucida is used in the production of leatherwood honey and is highly sought after by commercial apiarists. I investigated the impacts of commercially managed honeybees on E. lucida and its native pollinators. Because the type and severity of honeybee impacts are intimately related to the pollination system of the forage species, I also examined aspects of the pollination ecology of E. lucida. Flowers of E. lucida are relatively long-lived (12-13 days) and protandrous, with around 6-7 days of pollen presentation followed by 6 days of stigma receptivity. - However, the degree of overlap between the male and female phases of anthesis depends on the rate at which pollen is removed from anthers by insect visitors (i.e. flowers are facultatively protandrous). E.lucida flowers secrete a relatively dilute nectar (ca. 20% sugar wt/wt) from nectaries at the bases of the stamens. Nectar is secreted continuously, although secretion rates are substantially lower at night. Flowers typically contain small volumes of liquid nectar in the early morning which on warm days is rapidly concentrated through evaporative water loss to > 60% wt/wt. This concentrated nectar is highly attractive to insects and flowers typically receive multiple insect visits over a single day. E. lucida is partially self fertile. Fruit and seed set in bagged flowers which received a superabundance of autogamous self pollen (34% fruit set and 16 % seed set) was relatively low compared to fruit and seed set in un-bagged flowers (80% fruit set and 36% seed set). Stigmas of un-bagged flowers carried large amounts of pollen (estimated at 1700 grains/stigma), and E. lucida flowers do not appear to be pollen limited. Flowers of E. lucida received visits from a broad range of native diurnal insects (dipterans;16 families, coleopterans; 6 families, hymenopterans; 5 families, and lepidopterans; 2 families) and nocturnal insects (tipulid flies, elaterid beetles, blattellid cockroaches, and geometrid and pyralid moths), as well as from the introduced honeybee. E. lucida flowers also supported a range of squatter insects which used the flowers as a semi-permanent refuge (mainly thrips, staphylinid beetles, and spiders). Visitation rates varied enormously between sites, ranging from < 2 to > 25 visits per flower per 10-hour day. Nocturnal visitation rates were < 2 visits per flower per 10-hour night. Large dipterans and large coleopterans appeared to be the most important native pollinators of E. lucida. E. lucida appears to be well adapted for maximising pollination under conditions of temporal and spatial heterogeneity in the native pollinator service. Nectar production is independent of temperature, humidity and local shading, and flowers rapidly accumulate nectar sugar on cold days when insects are inactive. E. lucida flowers do not reabsorb accumulated nectar sugar. In contrast, the rate of anther dehiscence is strongly and positively dependent on temperature above 10°C, so that pollen release is retarded on cold days. The resulting patterns of nectar and pollen release appear to maximise both male and female function in E. lucida flowers under a broad array of weather and pollinator-abundance conditions. I examined the impacts of hive bees at 13 sites, 7 in the vicinity of a commercial apiary and 6 control sites located >2 km from the nearest apiary (hive bees foraged within 2 km of hives during E. lucida flowering). Honeybee activity at flowers was significantly higher near apiary sites compared to control sites, although the mean increase (by a factor of 2.5) was relatively modest. This increase in honeybees resulted in a significant depression in the availability of nectar sugar in flowers around apiaries. Hive honeybees appeared to be excluding feral honeybees from the vicinity of apiary sites. However, there was little evidence that hive bees caused a decline in the visitation rate or abundance of native insects, apparently due to very low numbers of native insects and a superabundance of nectar sugar at some of the sites. However, hive bees may reduce the number of native insects visiting E. lucida flowers at a subset of rainforest sites with abundant native insects and low levels of available nectar sugar. E. lucida flowers were also depleted of pollen more quickly at apiary sites compared to control sites, resulting in a 17% reduction in the standing crop of pollen in male flowers in the vicinity of apiaries. Fruit set tended to be higher near apiaries, although there was no difference in the number of pollen grains on stigmas, fruit dehiscence, fruit weight or seed set between apiary and control sites. Therefore, despite removing pollen more rapidly and reducing the availability of pollen in male flowers, hive bees appeared to have little net impact on the reproductive performance of E. lucida trees.