Callandrena, a subgenus of 80 described species of bees in the genus Andrena (Hymenoptera: Andrenidae) is shown, via phylogenetic analysis of both mitochondrial and nuclear DNA sequence data, to be polyphyletic. The characters previously uniting this group have likely arisen by convergent evolution among unrelated lineages that have independently specialized on flowers of the Asteraceae for pollen consumption. At this time, we cannot definitively state whether there are two or three clades of bees formerly ascribed to Callandrena, nor whether one clade may belong to the European subgenus Chrysandrena as has been proposed, so we do not erect a new subgenus of Andrena at this time. The limits of Callandrena sensu stricto are provisionally delimited based on a single morphological character. The phylogenetic relationships of the two groups relative to other Andrena subgenera are discussed in Chapter 2. In Chapter 3, Clade B, as defined for Callandrena in Chapter 2, is used to investigate the evolution of pollen host choice. Diet breadth is determined by analysis of pollen loads of at least 20 individual females per species. The choice of host and degree of specificity are then mapped onto the molecular phylogeny to investigate the evolution of these traits. Oligolecty appears to be the ancestral state in Andrena; polylecty has evolved several times; and reversals to oligolecty within these lineages have also occurred. Within the oligolectic lineage studied, host shifts were not uncommon. During the course of this study, a number of undescribed species were collected. Five are described in Chapter 4. / text

<p>In recent decades, the development of strategies to prevent or slow the loss of biodiversity has become an important task for ecologists. In most terrestrial ecosystems wild-bees play a key role as pollinators of herbs, shrubs and trees. The scope of this thesis was to study 1) pollinator effectiveness of specialist bees vs. generalist flower-visitors, 2) critical floral resources for wild-bees, and 3) methods to estimate the size of wild-bee populations. The wild-bee species <i>Andrena hattorfiana </i>and <i>A. marginata </i>were used as model species. These two species are specialized on pollen from the plant family Dipsacaceae.</p><p>The bee <i>A. hattorfiana </i>was found to be a frequent visitor but a poor pollinator of its preferred food-plant <i>Knautia arvensis</i>. The female bees exert such a strong preference for pollen-producing inflorescences that they likely have deleterious effects on the plant, harvesting valuable pollen that could have been transferred to conspecific stigmas by other flower-visitors. To explore the relationship between wild-bees and their food-plants, the concept of pollen budget was developed. We quantified pollen production in the food-plant population and pollen consumption of wild-bee nests. A survey of the visitation by all flower-visitor taxa indicated that the degree of utilization (the fraction of the total pollen amount that is harvested and utilized by <i>A. hattorfiana</i>) varied from 12% to 80% among <i>K. arvensis</i> populations (N=26). The bee <i>Andrena marginata</i> utilized 44% of the pollen production in a population of <i>Succisa pratensis</i>. The pollen budget suggests that with an average flower-visitor diversity and abundance, 330 individuals of the food-plant <i>K. arvensis </i>are required to sustain a population of 20 <i>A. hattorfiana </i>♀ (the approximate median natural population size). Based on a study of <i>A. hattorfiana</i>, considerable simplifications were proposed for the commonly used mark-recapture design for measuring wild-bee population size. For this species, population size estimated based on mark-recapture data was strongly correlated with the number of observations per survey-walk. The results suggest that large-scale surveys of solitary bee species can be simplified by performing survey-walks.</p><p>The pollen budget and the method proposed for estimating the size of bee populations have the potential to become valuable tools for monitoring and management of wild-bee populations.</p>

In recent decades, the development of strategies to prevent or slow the loss of biodiversity has become an important task for ecologists. In most terrestrial ecosystems wild-bees play a key role as pollinators of herbs, shrubs and trees. The scope of this thesis was to study 1) pollinator effectiveness of specialist bees vs. generalist flower-visitors, 2) critical floral resources for wild-bees, and 3) methods to estimate the size of wild-bee populations. The wild-bee species Andrena hattorfiana and A. marginata were used as model species. These two species are specialized on pollen from the plant family Dipsacaceae. The bee A. hattorfiana was found to be a frequent visitor but a poor pollinator of its preferred food-plant Knautia arvensis. The female bees exert such a strong preference for pollen-producing inflorescences that they likely have deleterious effects on the plant, harvesting valuable pollen that could have been transferred to conspecific stigmas by other flower-visitors. To explore the relationship between wild-bees and their food-plants, the concept of pollen budget was developed. We quantified pollen production in the food-plant population and pollen consumption of wild-bee nests. A survey of the visitation by all flower-visitor taxa indicated that the degree of utilization (the fraction of the total pollen amount that is harvested and utilized by A. hattorfiana) varied from 12% to 80% among K. arvensis populations (N=26). The bee Andrena marginata utilized 44% of the pollen production in a population of Succisa pratensis. The pollen budget suggests that with an average flower-visitor diversity and abundance, 330 individuals of the food-plant K. arvensis are required to sustain a population of 20 A. hattorfiana ♀ (the approximate median natural population size). Based on a study of A. hattorfiana, considerable simplifications were proposed for the commonly used mark-recapture design for measuring wild-bee population size. For this species, population size estimated based on mark-recapture data was strongly correlated with the number of observations per survey-walk. The results suggest that large-scale surveys of solitary bee species can be simplified by performing survey-walks. The pollen budget and the method proposed for estimating the size of bee populations have the potential to become valuable tools for monitoring and management of wild-bee populations.

Measuring the diversity of bees was provided in two rapeseed fields (Dešenice, Miletice) in western Bohemia, where various species of bees were caught at eight different locations during two periods (March - April, May June). The bees were collected by using Moerickeho cups (yellow traps) with a concentrated solution of water and detergent.