The superfamiliy of bees, Apiformes, comprises more than 20,000 species. Within the group, the eusocial species like honeybees and bumblebees are receiving increased attention due to their outstanding importance for pollination of many crop and wild plants, their exceptional eusocial lifestyle and complex behavioral repertoire, which makes them an interesting invertebrate model to study mechanisms of sensory perception, learning and memory. In bees and most animals, vision is one of the major senses since almost every living organism and many biological processes depend on light energy. Bees show various forms of vision, e.g. color vision, achromatic vision or polarized vision in order to orientate in space, recognize mating partners, detect suitable nest sites and search for rewarding food sources. To catch photons and convert light energy into electric signals, bees possess compound eyes which consists of thousands of single ommatidia comprising a fixed number of photoreceptors; they are characterized by a specific opsin protein with distinct spectral sensitivity. Different visual demands, e.g. the detection of a single virgin queen by a drone, or the identification and discrimination of flowers during foraging bouts by workers, gave rise to the exceptional sex-specific morphology and physiology of male and female compound eyes in honeybees. Since Karl von Frisch first demonstrated color vision in honeybees more than 100 years ago, much effort has been devoted to gain insight into the molecular, morphological and physiological characteristics of (sex-specific) bee compound eyes and the corresponding photoreceptors. However, to date, almost nothing is known about the underlying mechanisms during pupal development which pattern the retina and give rise to the distinct photoreceptor distribution. Hence, in Chapter 2 and 3 I aimed to better understand the retinal development and photoreceptor determination in the honeybee eye. In a first step, the intrinsic temporal expression pattern of opsins within the retina was evaluated by quantifying opsin mRNA expression levels during the pupal phase of honeybee workers and drones. First results revealed that honeybee workers and drones express three different opsin genes, UVop, BLop and Lop1 during pupal development which give rise to an ultraviolet, blue, and green-light sensitive photoreceptor. Moreover, opsin expression patterns differed between both sexes and the onset of a particular opsin occurred at different time points during retinal development. Immunostainings of the developing honeybee retina in Chapter 2 showed that at the beginning of pupation the retina consist only of a thin hypodermis. However, at this stage all retinal structures are already present. From about mid of pupation, opsin expression levels increase and goes hand in hand with the differentiation of the rhabdoms, suggesting a two-step process in photoreceptor development and differentiation in the honeybee compound eye. In a first step the photoreceptor cells meet its fate during late pupation; in a second step, the quantity of opsin expression in each photoreceptor strongly increase up to the 25-fold shortly after eclosion. To date, the underlying mechanisms leading to different photoreceptor types have been intensively studied in the fruit fly, Drosophila melanogaster, and to some extend in butterflies. Interestingly, the molecular mechanisms seemed to be conserved within insects and e.g. the two transcription factors, spalt and spineless, which have been shown to be essential for photoreceptor determination in flies and butterflies, have been also identified in the honeybee. In chapter 3, I investigated the expression patterns of both transcription factors during pupal development of honeybee workers and showed that spalt is mainly expressed during the first few pupal stages which might correlate with the onset of BLop expression. Further, spineless showed a prominent peak at mid of pupation which might initiates the expression of Lop1. However, whether spalt and spineless are also essential for photoreceptor determination in the honeybee has still to be investigated, e.g. by a knockdown/out of the respective transcription factor during retinal development which leads to a spectral phenotype, e.g. a dichromatic eye. Such spectral phenotypes can then be tested in behavioral experiments in order to test the function of specific photoreceptors for color perception and the entrainment of the circadian clock. In order to evaluate the color discrimination capabilities of bees and the quality of color perception, a reliable behavioral experiment under controlled conditions is a prerequisite. Hence, in chapter 4, I aimed to establish the visual PER paradigm as a suitable method for behaviorally testing color vision in bees. Since PER color vision has considered to be difficult in bees and was not successful in Western honeybees without ablating the bee’s antennae or presenting color stimuli in combination with other cues for several decades, the experimental setup was first established in bumblebees which have been shown to be robust and reliable, e.g. during electrophysiological recordings. Workers and drones of the bufftailed bumblebee, Bombus terrestris were able to associate different monochromatic light stimuli with a sugar reward and succeeded in discriminating a rewarded color stimulus from an unrewarded color stimulus. They were also able to retrieve the learned stimulus after two hours, and workers successfully transferred the learned information to a new behavioral context. In the next step, the experimental setup was adapted to honeybees. In chapter 5, I tested the setup in two medium-sized honeybees, the Eastern honeybee, Apis cerana and the Western honeybee, Apis mellifera. Both honeybee species were able to associate and discriminate between two monochromatic light stimuli, blue and green light, with peak sensitivities of 435 nm and 528 nm. Eastern and Western honeybees also successfully retrieve the learned stimulus after two hours, similar to the bumblebees. Visual conditioning setups and training protocols in my study significantly differed from previous studies using PER conditioning. A crucial feature found to be important for a successful visual PER conditioning is the duration of the conditioned stimulus presentation. In chapter 6, I systematically tested different length of stimuli presentations, since visual PER conditioning in earlier studies tended to be only successful when the conditioned stimulus is presented for more than 10 seconds. In this thesis, intact honeybee workers could successfully discriminate two monochromatic lights when the stimulus was presented 10 s before reward was offered, but failed, when the duration of stimulus presentation was shorter than 4 s. In order to allow a more comparable conditioning, I developed a new setup which includes a shutter, driven by a PC based software program. The revised setup allows a more precise and automatized visual PER conditioning, facilitating performance levels comparable to olfactory conditioning and providing now an excellent method to evaluate visual perception and cognition of bees under constant and controlled conditions in future studies.

I. Timing is a crucial feature in organisms that live within a variable and changing environment. Complex mechanisms to measure time are wide-spread and were shown to exist in many taxa. These mechanisms are expected to provide fitness benefits by enabling organisms to anticipate environmental changes and adapt accordingly. However, very few studies have addressed the adaptive value of proper timing. The objective of this PhD-project was to investigate mechanisms and fitness consequences of timing decisions concerning colony phenology and foraging activity in the honey bee (Apis mellifera), a social insect species with a high degree of social organization and one of the most important pollinators of wild plants and crops. In chapter II, a study is presented that aimed to identify the consequences of disrupted synchrony between colony phenology and the local environment by manipulating the timing of brood onset after hibernation. In a follow-up experiment, the importance of environmental factors for the timing of brood onset was investigated to assess the potential of climate change to disrupt synchronization of colony phenology (Chapter III). Chapter IV aimed to prove for the first time that honey bees can use interval time-place learning to improve foraging activity in a variable environment. Chapter V investigates the fitness benefits of information exchange between nest mates via waggle dance communication about a resource environment that is heterogeneous in space and time. II. In the study presented in chapter II, the importance of the timing of brood onset after hibernation as critical point in honey bee colony phenology in temperate zones was investigated. Honey bee colonies were overwintered at two climatically different sites. By translocating colonies from each site to the other in late winter, timing of brood onset was manipulated and consequently colony phenology was desynchronized with the local environment. Delaying colony phenology in respect to the local environment decreased the capability of colonies to exploit the abundant spring bloom. Early brood onset, on the other hand, increased the loads of the brood parasite Varroa destructor later in the season with negative impact on colony worker population size. This indicates a timing related trade-off and illustrates the importance of investigating effects of climate change on complex multi-trophic systems. It can be concluded that timing of brood onset in honey bees is an important fitness relevant step for colony phenology that is highly sensitive to climatic conditions in late winter. Further, phenology shifts and mismatches driven by climate change can have severe fitness consequences. III. In chapter III, I assess the importance of the environmental factors ambient temperature and photoperiod as well as elapsed time on the timing of brood onset. Twenty-four hibernating honey bee colonies were placed into environmental chambers and allocated to different combinations of two temperature regimes and three different light regimes. Brood onset was identified non-invasively by tracking comb temperature within the winter cluster. The experiment revealed that ambient temperature plays a major role in the timing of brood onset, but the response of honey bee colonies to temperature increases is modified by photoperiod. Further, the data indicate the involvement of an internal clock. I conclude that the timing of brood onset is complex but probably highly susceptible to climate change and especially spells of warm weather in winter. IV. In chapter IV, it was examined if honey bees are capable of interval time-place learning and if this ability improves foraging efficiency in a dynamic resource environment. In a field experiment with artificial feeders, foragers were able to learn time intervals and use this ability to anticipate time periods during which feeders were active. Further, interval time-place learning enabled foragers to increase nectar uptake rates. It was concluded that interval time-place learning can help honey bee foragers to adapt to the complex and variable temporal patterns of floral resource environments. V. The study presented in chapter V identified the importance of the honey bee waggle dance communication for the spatiotemporal coordination of honey bee foraging activity in resource environments that can vary from day to day. Consequences of disrupting the instructional component of honey bee dance communication were investigated in eight temperate zone landscapes with different levels of spatiotemporal complexity. While nectar uptake of colonies was not affected, waggle dance communication significantly benefitted pollen harvest irrespective of landscape complexity. I suggest that this is explained by the fact that honey bees prefer to forage pollen in semi-natural habitats, which provide diverse resource species but are sparse and presumably hard to find in intensively managed agricultural landscapes. I conclude that waggle dance communication helps to ensure a sufficient and diverse pollen diet which is crucial for honey bee colony health. VI. In my PhD-project, I could show that honey bee colonies are able to adapt their activities to a seasonally and daily changing environment, which affects resource uptake, colony development, colony health and ultimately colony fitness. Ongoing global change, however, puts timing in honey bee colonies at risk. Climate change has the potential to cause mismatches with the local resource environment. Intensivation of agricultural management with decreased resource diversity and short resource peaks in spring followed by distinctive gaps increases the probability of mismatches. Even the highly efficient foraging system of honey bees might not ensure a sufficiently diverse and healthy diet in such an environment. The global introduction of the parasitic mite V. destructor and the increased exposure to pesticides in intensively managed landscapes further degrades honey bee colony health. This might lead to reduced cognitive capabilities in workers and impact the communication and social organization in colonies, thereby undermining the ability of honey bee colonies to adapt to their environment.