In this project I studied the importance of pollinators in the reproduction of Gastrodia ‘long column’ and the preferences of New Zealand’s native and exotic insect pollinators. This was done in order to determine the specialisation of insect pollination in urban Christchurch. This knowledge can contribute to crop pollination, conservation efforts, and an understanding of the main drivers of the common floral traits (small, white flowers) in New Zealand. The strength of the relationship between the common traits of the New Zealand flora and the preferences of the native invertebrate fauna is not known. Traditionally it had been thought that New Zealand’s insects lack strong preferences, however recent work has not supported this. Changing landscapes worldwide have led to declines in pollinator numbers. Additionally, in New Zealand Apis mellifera numbers have declined as a result of the arrival of the varroa mite and it is important that alternative pollinating species for wild and agricultural pollination are identified. To address these needs I examined the abundance and preferences of insect pollinators in modified landscapes in and around Christchurch testing: the reliance upon pollinators by the undescribed native orchid Gastrodia ‘long column’, the preferences of New Zealand’s native and introduced insect pollinators in an extensive observational study, the results of which I further tested using a subsequent manipulative experiment of petal colour (according to human vision) at the Christchurch Botanical Gardens Sampling at the Christchurch Botanical Gardens, University of Canterbury, and Port Hills across a four month observation period (January – April 2012), revealed that native pollinators preferred white native flowers and exotic pollinators preferred not-white exotic flowers when data were grouped according to insect provenance. A more detailed visit-level analysis found that two native bees, Hylaeus and Leioproctus, showed a significant preference for native flowers; the exotic bees, Apis mellifera and Bombus species, preferred purple over white flowers; and Melangyna novaezelandiae (a native hoverfly) preferred white over yellow. However, a series of experimental arrays to present controlled choices between pairs of flower types at the Christchurch Botanical Gardens (14 December 2012 – 22 January 2013) did not find significant preferences by native or exotic insect pollinators between white and yellow flowers. This may have been a result of the plant species chosen, as a correlation between pollinator preference and plant species has been shown elsewhere. Visitation to experimental arrays was both low and highly skewed, with over half of all visits made by Lasioglossum bees and 615 of the 669 visits made by native species. This may have meant that lack of significant results were representing the choices of native insects generally and Lasioglossum bees specifically. A bagging experiment from 30 January 2012 – 16 February 2012 demonstrated the dependence of the undescribed native orchid species, Gastrodia ‘long column’, upon pollinating animals. Fruit set of most plants worldwide depends upon pollination (by wind or animal-transfer of pollen). In this case study final fruit sets were significantly reduced on bagged inflorescences, while open flowers had surprisingly high natural fruit sets (>75%). A week of observations (29 January 2012 – 6 February 2012) revealed that Gastrodia ‘long column’ was predominantly visited by Lasioglossum bees, and remarkably bee numbers were high enough in a residential property in the middle of Christchurch city for high fruit set on unmanipulated plants. The results of my thesis indicate that pollinators are important in the reproductive system of Gastrodia ‘long column’, suggesting that other previously overlooked plant species may also rely upon insect pollinators. Furthermore, the relative importance of native pollinating insects is high for native plants even when examined in an urban setting. New Zealand pollinators have preferences for certain floral traits which show trends when grouped broadly, but vary when considered at the insect species level. This contrasts with traditional views of unspecialised insect pollinators which lack preferences in regards to the plants they visit. Further work which serves to increase current understanding of the underlying mechanisms of pollination specialisation in New Zealand may wish to focus on single insect species. By identifying particular preferences of pollinators and the underlying ‘native’ traits selected for, alternative options to crop pollination may be found, targeted management strategies implemented, and the strength of the relationship between pollinator preferences and the traits of the plants they visit determined.

A utilização de sequências do DNA mitocondrial (mtDNA) como marcadores moleculares na investigação da diversidade genética e evolução é muito difundida, auxiliando na realização de inferências em inúmeros trabalhos. Apesar de sua inegável importância, a utilização dessas sequências como marcadores moleculares suscita algumas questões. A heteroplasmia, por exemplo, é reconhecida como um desafio na utilização de sequências do mtDNA. Este estado ocorre quando um organismo apresenta diferentes haplótipos mitocondriais. Em um trabalho anterior, foram encontrados indícios que sugeriam a presença de heteroplasmia na espécie de abelha Bombus morio. O trabalho atual investigou de forma mais detalhada a presença de heteroplasmia nessa espécie, assim como fatores que podem influenciar na identificação desse estado. Os resultados obtidos confirmaram a existência de heteroplasmia nessa espécie, e identificaram que alguns haplótipos heteroplásmicos foram compartilhados entre indivíduos de localidades distintas. Esses haplótipos heteroplásmicos compartilhados sugerem a existência de heteroplasmia estável em B. morio, o que pode influenciar inferências evolutivas, e em especial, os estudos populacionais. Também foi detectada a presença de NUMTs, pseudogenes nucleares resultantes da transferência de sequências do mtDNA para o genoma nuclear. Esses NUMTs apresentaram grande divergência de sequência em relação aos haplótipos mitocondriais, o que poderia afetar análises filogenéticas e populacionais, além da identificação de espécies por meio do DNA barcoding. Ainda, erros de amplificação podem ser falsamente interpretados como variação intraindividual do DNA mitocondrial (mtDNA), superestimando o número de haplótipos, principalmente quando polimerases de baixa fidelidade são utilizadas. Por fim, os resultados observados neste trabalho sugerem que a utilização de sequências do mtDNA deve ser utilizada de forma cautelosa, e indícios de heteroplasmia, como a presença de picos duplos, não devem ser ignorados. Quando essas evidências são observadas investigações mais detalhadas devem ser aplicadas, a fim de aferir qual a sua origem, e, no caso da heteroplasmia ser confirmada, quais as possíveis consequências produzidas pela presença desse estado / The mitochondrial DNA sequences (mtDNA) have been widely applied as molecular markers in the investigation of genetic diversity and evolution. Despite its undeniable importance, the use of these sequences as molecular markers may present some drawbacks. Heteroplasmy, for example, is recognized as a challenge. This state occurs when an individual has different mitochondrial haplotypes. In a previous work, evidences suggesting the presence of heteroplasmy in the bumblebee Bombus morio were verified. The present work investigated in more detail the presence of heteroplasmy in this species, as well as factors that may influence the identification of this state. The results confirmed the existence of heteroplasmy in this species, and identified that some heteroplasmic haplotypes were shared between individuals from different locations. These shared heteroplasmic haplotypes suggest the existence of stable heteroplasmy in B. morio, which may interfere in evolutionary inferences, especially in population studies. NUMTs, nuclear pseudogenes resulting from the transfer of mtDNA sequences to the nuclear genome, were also detected. These NUMTs showed great sequence divergence from mitochondrial haplotypes, which could affect phylogenetic and population analyzes, as well as species identification through DNA barcoding. In addition, it was observed that amplification errors might be misinterpreted as mtDNA intraindividual variation and overestimates the number of intraindividual haplotypes, especially when low fidelity polymerases are used. Finally, the results observed in this study suggest that the use of mtDNA sequences should be used carefully, and evidences of heteroplasmy, such as the presence of double peaks, should not be ignored. Additional investigations should be applied in case of heteroplasmy evidences to ascertain your source and the consequences of the presence of this state

Temperature variation affects most biological parameters from the molecular level to community structure and dynamics. Current studies on thermal biology assess how populations vary in response to environmental temperature, which can help determine how populations differentially respond to climate change. To date, temperature fluctuation effects on endothermic poikilotherms such as the common eastern bumblebee (Bombus impatiens) are unknown even though bumblebees are the most important natural pollinators in North America. A cold-acclimation experiment with B. impatiens colonies revealed individuals acclimated to 5°C or 10°C at night did not differ in resting metabolic rate, flight metabolic rate, wingbeat frequency, or morphological measurements, compared to the control group. Moreover, an infrared camera showed that all colonies maintained maximum nest temperature consistently above 36.8°C. A latitudinal sampling of flight metabolic rate and morphological measurements of B. impatiens from four locations spanning Ontario (N 45°; W 75°) to North Carolina (N 34°; W 77°) indicated no latitudinal trend in the measured variables. This study shows that bumblebees are well equipped to face a wide range of environmental temperatures, both in the short term and long term, and can use a combination of behavioural and physiological mechanisms to regulate body and nest temperatures. These results are reassuring on the direct effects of climate change on bumblebee ecology, but further studies on the indirect effect of temperature variation on North American bumblebees are required to predict future ecosystem dynamics.

Temperature variation affects most biological parameters from the molecular level to community structure and dynamics. Current studies on thermal biology assess how populations vary in response to environmental temperature, which can help determine how populations differentially respond to climate change. To date, temperature fluctuation effects on endothermic poikilotherms such as the common eastern bumblebee (Bombus impatiens) are unknown even though bumblebees are the most important natural pollinators in North America. A cold-acclimation experiment with B. impatiens colonies revealed individuals acclimated to 5°C or 10°C at night did not differ in resting metabolic rate, flight metabolic rate, wingbeat frequency, or morphological measurements, compared to the control group. Moreover, an infrared camera showed that all colonies maintained maximum nest temperature consistently above 36.8°C. A latitudinal sampling of flight metabolic rate and morphological measurements of B. impatiens from four locations spanning Ontario (N 45°; W 75°) to North Carolina (N 34°; W 77°) indicated no latitudinal trend in the measured variables. This study shows that bumblebees are well equipped to face a wide range of environmental temperatures, both in the short term and long term, and can use a combination of behavioural and physiological mechanisms to regulate body and nest temperatures. These results are reassuring on the direct effects of climate change on bumblebee ecology, but further studies on the indirect effect of temperature variation on North American bumblebees are required to predict future ecosystem dynamics.

A utilização de sequências do DNA mitocondrial (mtDNA) como marcadores moleculares na investigação da diversidade genética e evolução é muito difundida, auxiliando na realização de inferências em inúmeros trabalhos. Apesar de sua inegável importância, a utilização dessas sequências como marcadores moleculares suscita algumas questões. A heteroplasmia, por exemplo, é reconhecida como um desafio na utilização de sequências do mtDNA. Este estado ocorre quando um organismo apresenta diferentes haplótipos mitocondriais. Em um trabalho anterior, foram encontrados indícios que sugeriam a presença de heteroplasmia na espécie de abelha Bombus morio. O trabalho atual investigou de forma mais detalhada a presença de heteroplasmia nessa espécie, assim como fatores que podem influenciar na identificação desse estado. Os resultados obtidos confirmaram a existência de heteroplasmia nessa espécie, e identificaram que alguns haplótipos heteroplásmicos foram compartilhados entre indivíduos de localidades distintas. Esses haplótipos heteroplásmicos compartilhados sugerem a existência de heteroplasmia estável em B. morio, o que pode influenciar inferências evolutivas, e em especial, os estudos populacionais. Também foi detectada a presença de NUMTs, pseudogenes nucleares resultantes da transferência de sequências do mtDNA para o genoma nuclear. Esses NUMTs apresentaram grande divergência de sequência em relação aos haplótipos mitocondriais, o que poderia afetar análises filogenéticas e populacionais, além da identificação de espécies por meio do DNA barcoding. Ainda, erros de amplificação podem ser falsamente interpretados como variação intraindividual do DNA mitocondrial (mtDNA), superestimando o número de haplótipos, principalmente quando polimerases de baixa fidelidade são utilizadas. Por fim, os resultados observados neste trabalho sugerem que a utilização de sequências do mtDNA deve ser utilizada de forma cautelosa, e indícios de heteroplasmia, como a presença de picos duplos, não devem ser ignorados. Quando essas evidências são observadas investigações mais detalhadas devem ser aplicadas, a fim de aferir qual a sua origem, e, no caso da heteroplasmia ser confirmada, quais as possíveis consequências produzidas pela presença desse estado / The mitochondrial DNA sequences (mtDNA) have been widely applied as molecular markers in the investigation of genetic diversity and evolution. Despite its undeniable importance, the use of these sequences as molecular markers may present some drawbacks. Heteroplasmy, for example, is recognized as a challenge. This state occurs when an individual has different mitochondrial haplotypes. In a previous work, evidences suggesting the presence of heteroplasmy in the bumblebee Bombus morio were verified. The present work investigated in more detail the presence of heteroplasmy in this species, as well as factors that may influence the identification of this state. The results confirmed the existence of heteroplasmy in this species, and identified that some heteroplasmic haplotypes were shared between individuals from different locations. These shared heteroplasmic haplotypes suggest the existence of stable heteroplasmy in B. morio, which may interfere in evolutionary inferences, especially in population studies. NUMTs, nuclear pseudogenes resulting from the transfer of mtDNA sequences to the nuclear genome, were also detected. These NUMTs showed great sequence divergence from mitochondrial haplotypes, which could affect phylogenetic and population analyzes, as well as species identification through DNA barcoding. In addition, it was observed that amplification errors might be misinterpreted as mtDNA intraindividual variation and overestimates the number of intraindividual haplotypes, especially when low fidelity polymerases are used. Finally, the results observed in this study suggest that the use of mtDNA sequences should be used carefully, and evidences of heteroplasmy, such as the presence of double peaks, should not be ignored. Additional investigations should be applied in case of heteroplasmy evidences to ascertain your source and the consequences of the presence of this state

Temperature variation affects most biological parameters from the molecular level to community structure and dynamics. Current studies on thermal biology assess how populations vary in response to environmental temperature, which can help determine how populations differentially respond to climate change. To date, temperature fluctuation effects on endothermic poikilotherms such as the common eastern bumblebee (Bombus impatiens) are unknown even though bumblebees are the most important natural pollinators in North America. A cold-acclimation experiment with B. impatiens colonies revealed individuals acclimated to 5°C or 10°C at night did not differ in resting metabolic rate, flight metabolic rate, wingbeat frequency, or morphological measurements, compared to the control group. Moreover, an infrared camera showed that all colonies maintained maximum nest temperature consistently above 36.8°C. A latitudinal sampling of flight metabolic rate and morphological measurements of B. impatiens from four locations spanning Ontario (N 45°; W 75°) to North Carolina (N 34°; W 77°) indicated no latitudinal trend in the measured variables. This study shows that bumblebees are well equipped to face a wide range of environmental temperatures, both in the short term and long term, and can use a combination of behavioural and physiological mechanisms to regulate body and nest temperatures. These results are reassuring on the direct effects of climate change on bumblebee ecology, but further studies on the indirect effect of temperature variation on North American bumblebees are required to predict future ecosystem dynamics.

Bumble bees provide valuable pollination services to many agricultural crops and wild flower species. Consequently, evidence that wild populations are in decline has caused widespread concern. Among multiple causal factors, some have singled out neonicotinoid pesticides as potentially a major contributor to these declines. Bumble bees are exposed to neonicotinoids, such as imidacloprid and thiamethoxam, whilst foraging for nectar and pollen from treated crops. For neonicotinoids to cause population decline, the typical residues that bumble bees encounter in the field (defined here as between 1–12 μg kg-1) should be capable of reducing colony success by detrimentally impacting demographically relevant endpoints such as reproduction and worker performance. Whether field-realistic neonicotinoids are capable of causing such effects is yet to be fully established. The overall aim of this thesis was to investigate the effects of field-realistic neonicotinoids on endpoints of demographic importance and improve understanding of the effect mechanisms of neonicotinoids in bumble bees. Laboratory experiments were conducted with Bombus terrestris L. exposed to dietary neonicotinoids up to 98 μg kg-1. Results showed that food consumption and production of brood (eggs and larvae) in queenless B. terrestris microcolonies were significantly reduced by the two highest concentrations of imidacloprid and thiamethoxam tested (39, 98 μg kg-1), but only imidacloprid produced a negative effect when concentrations were in the typical field-realistic range. Imidacloprid’s affect on microcolonies was mirrored in queenright colonies where field-realistic concentrations substantively reduced both feeding and brood production. It was postulated that the detrimental effects of imidacloprid on brood production emerge principally from nutrient limitation imposed by the failure of individuals to feed. Removing imidacloprid from the bees’ diet resulted in the recovery of feeding and brood production in queenright colonies, even when previously exposed to high doses (98 μg kg-1). Investigation into the effect mechanisms of imidacloprid in B. terrestris revealed that cytochrome P450 enzymes are not important for metabolism of the neonicotinoid in adult workers. A transcriptomic analysis indicated B. terrestris exhibit a general stress response to imidacloprid, characterised by the alteration in expression of genes involved in, for example, metabolism and storage of energy. The thesis findings raise further concern about the threat of imidacloprid to wild bumble bees. However, they also suggest that some demographically important endpoints are resilient to imidacloprid as a realistic pulsed exposure, and that bumble bees may be less sensitive to field-realistic concentrations of thiamethoxam. Further research, which is required to fully establish the demographic consequences for bumble bees of exposure to neonicotinoids, can be developed based on the foundation of work presented here.

Tesis entregada a la Universidad de Chile en cumplimiento parcial de los requisitos para optar al grado de Magíster en Ciencias Biológicas. / La medida en que los caracteres florales evolucionan bajo selección mediada por polinizadores ha sido ampliamente estudiada en la literatura. Sin embargo, la relación entre selección y efectividad de los polinizadores es un tema poco estudiado. El presente trabajo busca identificar y cuantificar las fuerzas selectivas ejercidas por abejas y colibríes sobre Erythranthe lutea, una planta herbácea altoandina de Chile central. Una de las principales preguntas en este estudio es evaluar si la selección mediada por polinizador sigue el principio del polinizador más eficiente de Stebbins. Para hacerlo, examinamos la selección putativa sobre tres rasgos florales de E. lutea durante dos años consecutivos a través de la función sexual femenina. Para examinar posibles efectos aditivos y no aditivos de los polinizadores, implementamos cuatro tratamientos en un diseño factorial de 2 x 2 usando Picaflores (P) y Abejas (A) como factores principales. De esta forma, probamos la selección mediada por polinizadores en los siguientes tratamientos: P+A+, P+A-, P-A+, P-A-. Nuestros resultados indican que el polinizador más eficiente fue el abejorro exótico Bombus terrestris, una de las especies responsables de la mayor fuerza selectiva impuesta por abejas. En particular, las abejas promueven la reproducción de flores con corolas más grandes en el período 2016. Sin embargo, durante el período 2017 se detectaron efectos no aditivos, lo que indica que el efecto de las abejas estaba supeditado a la presencia de picaflores y viceversa, lo que impide la evaluación de los efectos independientes. Si bien E. lutea parece estar evolucionando según el principio de Stebbins, es necesario demostrar que el fitness ganado después de la especialización de los rasgos florales excede al costo en fitness asociado a la pérdida de otros polinizadores menos eficientes. La observación de que B. terrestris fue el polinizador más eficiente en el ensamble de polinizadores sugiere que los polinizadores introducidos pueden impulsar nuevos y rápidos procesos de adaptación floral. En consecuencia, comprender el impacto evolutivo de los polinizadores exóticos puede ayudar al diseño de políticas de conservación frente al intercambio biótico y a la homogeneización a gran escala. / The extent to which floral characters evolve under pollinator-mediated selection has been extensively studied in the literature. However, the relationship between selection and pollinator effectiveness is a less understood subject. The present study aims to identify and quantify the selective force imposed by bees and hummingbirds upon Erythranthe lutea, a high Andean herbaceous plant from central Chile. A major question in this study is to evaluate whether pollinatormediated selection follows the Stebbins` principle of the most efficient pollinator. In doing so, we examined putative selection upon three flower traits of E. lutea during two consecutive years through the female sex function. To examine potential additive or nonadditive pollinator effects, we set four treatments in a 2 x 2 factorial design using hummingbirds (H) and bees (B) as main factors. In this way, we tested for pollinator mediated selection in the following treatments: H+B+, H+B-, H-B+, H-B-. Our results indicate that the most efficient pollinator was the exotic bumblebee Bombus terrestris, one of the responsible species of the higher selective force imposed by bees. In particular, bees promoted the reproduction of flowers with large-sized corollas in the 2016 period. However, during the 2017 nonadditive effects were detected, indicating that the effect of bees were contingent to the presence of bees and viceversa, which prevented the assessment of independent effects. While E. lutea seems to be evolving under Stebbins’ principle, it is necessary to demonstrate that the gained fitness after floral traits specialization exceeds the fitness cost associated to the loss of other less efficient pollinators. The observation that B. terrestris was the most efficient pollinator in the pollinator assemblage suggests that introduced pollinators can drive new and rapid floral adaptation processes. In consequence, understanding the evolutionary impact of exotic pollinators can help to the design of conservation policies in the face of biotic exchange and broad scale homogenization. / Proyecto FONDECYT 1150112.

Bees and wasps perform learning flights when departing their nest for the first few times or a newly discovered food source. Several studies have described the occurances and structure of these flights in several species, but few have examined how the insects systematically vary the characteristics of their learning flights in various conditions in order to aid the acquisition of visual information. This is best done in a species where individuals and nests can be easily manipulated and tested repeatedly. The aim of this thesis was therefore to investigate learning flights in bumblebees, where we have a good understanding of the structure and variability of flights from previous work and can design controlled experiments. I explored the similarities and differences of learning flights of workers and male bumblebees, observing their departures from the nest or an artificial flower. A second objective was to examine how differences in the learning flights affect the bumblebees’ ability to return the learnt location. The experiments were conducted inside a large greenhouse, under natural light regimes, with two large tables placed far apart, one for simulating the ground from which bees emerged when departing their nest, and the other representing a feeding site with an artificial flower. Female bumblebees performed shorter learning flights when leaving a flower than when leaving their nest, although both locations displayed similar visual scenes. At both locations, the duration and trajectory length of learning flights decreased over successive visits, but the decrease was faster at the flower location than at the nest. Bumblebees fixated both their nest and the flower during their learning flights as well as the landmarks available around the two locations, which suggests that they learned the position of the goal relative to these landmarks. When the nest and the flower were hidden and only three cylinders were shown as landmarks in tests, bees searched as accurately for the nest as for the flower. However, they were more persistent when searching for the nest than for the flower, which was not predicted from the variation of learning flights at the nest and flower locations. Another situation in which bumblebees varied the characteristics of their learning flights, but without an impact on their performance when recalling the learnt information, was after visiting flowers filled with low and high sucrose rewards. The bees performed longer learning flights after drinking at a highly rewarded flower. When departing a poorly rewarded flower, bumblebees did not fixate the flower during their learning flights. Nevertheless, the bees were able to return to both the poorly rewarded flower and the highly rewarded flower equally fast. Given the above findings, it is not evident how different durations or trajectory lengths of bumblebee learning flights might be linked to variations in learning of goal locations. Finally, I show that bumblebees of either sex decide to perform learning flights at locations that are of importance to them. Whilst the female workers always performed learning flights when departing their natal nest, the males did not and simply flew away in a straight line. However, when leaving a flower, the males did perform learning flights with characteristics similar to those of the females’ learning flights. They were also able to return to the flower, showing similar approach trajectories as workers. The thesis discusses these findings in the light of ideas and hypotheses that are linked to differential investment in learning which were observed in the various conditions here. It is also discussed why bumblebees used fixations in different ways when learning about the visual environment surrounding goals that are important to them. Whilst many results are parsimonious with the requirements for learning and active flight control to aid the acquisition of visual information, motivation also seems to play a role in varying the occurances and features of learning flights, such as seen in the bees’ greater persistence to search for their nest than for a flower.

Genomic imprinting, the parent-of-origin specific silencing of alleles, plays an important role in phenotypic plasticity and consequently evolution. The leading explanation for genomic imprinting is Haig's conflict theory, which suggests that alleles from each parent have evolved under different selectional pressures, resulting in the differential expression of patrigenes and matrigenes. Previous studies have mainly used mammals and flowering plants to test Haig’s theory. However, there is a lack of independent evidence to support the theory. My PhD thesis attempts to conduct an independent test of Haig’s conflict theory using buff tailed bumblebee Bombus terrestris. A methylation system to facilitate genomic imprinting has not been found in this species. Therefore the first aim of the study was to establish the presence of a functional methylation system in B. terrestris before testing Haig's conflict theory using worker reproduction in queen-less colonies. The initial finding is that a methylation system exists in B. terrestris. The next study, investigating the presence of methylated genes, revealed differential methylation patterns in caste and life stages. Finally, genes involved with worker reproduction in a range of social insects were identified, but distinguishing the matrigene and the patrigene for each gene was unsuccessful. Therefore the final study investigating the presence of imprinted genes in B. terrestris and whether they conform to the expression patterns hypothesised by Haig’s conflict theory could not be analysed. Although this study did not provide conclusive evidence to support Haig’s conflict theory, the presence of methylation in genes involved with worker reproduction in reproducing and non-reproducing B. terrestris workers suggests that further analysis is needed. With adequate evidence, proving Haig’s conflict theory will not only expand our knowledge of invertebrate methylation, but also our understanding of conflict within social insect societies and our knowledge of how genomic imprinting affects phenotypic plasticity.