Kakamega Forest represents the last remnant of

equatorial rainforest in Kenya. Hence, it is of great

importance as last habitat for species specialized on

these forests and their associated ecosystems and, consequently, for East Africa’s biodiversity. The forest has been divided

into several fragments due to anthropogenic habitat

change, which is still continuing today. Even though

large parts of the forest have been designated protected

areas, they are still used for illegal hunting, logging,

and firewood collection.

This thesis focusses the pollination ecology of two

species of Acanthus. A. eminens occupies clearings

and riversides inside the forest, while A. polystachyus

grows in copses and hedgerows of the surrounding

farmland, as well as at the forest edge. Through forest

fragmentation, A. eminens continues to loose suitable

habitats. In addition, the distance of its populations to

the congener is reduced, and the relative abundance

of the species shifts towards A. polystachyus.

As flowering time and floral morphology of the

species are highly similar, it seems likely that habitat

fragmentation causes changes in pollination and

reproduction of the species. In particular, A. eminens

may loose pollinators to A. polystachyus, and receive

more heterospecific pollen. As A. eminens is a common

species of the natural ecosystem, such effects

would indicate that habitat fragmentation causes longterm

changes in ecosystem processes of Kakamega

Forest, threatening its future existence in its natural

state.

In my thesis, I demonstrate that the species flower in

synchrony during winter dry season between October

and February. There are no negative effects of distance

to or habitat availability of the congener on pollination

and reproduction of the species. However, I find plants

in drier habitats and flowers opening later during dry

season to display higher fruit set.

Both species are pollinated by carpenter bees (Xylocopa).

There is no indication for a partitioning of pollinators

through divergent flower morphology, as the

species display highly similar flowers. A. eminens offers

greater quantities of both nectar and pollen, while

the total amount of floral rewards is greater in the

larger individuals and populations of A. polystachyus.

Both species likely represent important resources for

pollinators in their respective habitats. Differences in

pollinator abundance between populations are likely caused by habitat preferences of the pollinators.

Even though the species share most of their pollinators,

interspecific pollen transfer does not affect

their reproduction. This may be explained by the prevalence

of geitonogamy, as bees commonly visit several

flowers on the same inflorescence, plant, and neighbouring,

likely related individuals. Moderate visitation

rates with low pollination success indicate that carpenter

bees are ineffective pollinators, and that the greatest

proportion of pollen is lost between visits. Consequently,

most pollen is transferred between close,

potentially closely related neighbours, reducing the

incidence of heterospecific pollen transfer.

These findings indicate that reproduction of either

species may be limited by pollen quantity or pollen

quality. Hand-pollination experiments show that both

species are limited by pollen quantity but not pollen

quality, as fruit set nearly doubles when supplementary

pollen, regardless of its source, is provided. As

there is no difference in seed set in seed viability between treatments, I

conclude that pollen quality is generally high.

There is convincing evidence that A. eminens and

A. polystachyus are adapted to flowering during dry

season, which offers favorable conditions for fruit and

seed development as well as for seed dispersal and germination.

Differences in humidity are strongly linked

to reproduction in either species, and these effects are

strong enough to mask all putative effects of competition

for pollination and heterospecific pollen transfer.