Abstract
Flower colour throughout the world is generally linked to various functional purposes, pollinator attraction in particular. Current knowledge regarding the functional significance of flower colour is, however, mainly based on studies from boreal, temperate, and tropical regions. This knowledge is not readily applicable to polar regions. Similar to lower latitudes, flower colour is often observed in polar regions, as well as flower colour polymorphisms. Pollinator attraction are considered to be the main purpose of flower colour, and the lack of pollinators in polar regions, where many plant species produce coloured flowers, causes the significance of colour to be unclear. Further, flower colour pigments are often quite energy expensive to produce, and in polar regions, nutrients, light and resources are limited. Hence, the question; what is the role of flower colour in reproductive success in polar regions? This thesis addresses this question of ‘What is the role of flower colour in polar regions?’ through several studies related to various aspects of floral reproduction in polar regions. This is a broad topic, and little background knowledge exists. To deal with this, a broad approach was also used, covering different geographical scales, from bipolar (Chapter Three), arctic (Chapter Two) and regional (Svalbard; Chapter Six) scales, to the population scale (Chapter Four and Five).
Information from literature on hundreds of species (Chapter Two) was compiled, field experiments and thermal imaging were completed on specific plant populations (Chapter Three - Five), and, at the genetic level, amplified fragment length polymorphism (AFLPs) were conducted within a species (Papaver dahlianum; Chapter Six). Except for one study (Chapter Four), which included results from Campbell Island (Subantarctic), most information is gathered from the Arctic, in particular from the high-arctic archipelago, Svalbard. The thesis is thus somewhat biased towards the Arctic. There were several initial hypotheses around why we observe flower colour polymorphism in polar regions. One possibility was that the few pollinators present may be more specific or more efficient than formerly believed, and sufficient to drive and sustain flower colour polymorphisms in polar regions (addressed in Chapters Two, Three, Five and Six). Alternatively, these colour morphs could be remnants from ancestor populations living under a different pollinator regime, and without any current function (historical patterns are partly addressed in Chapter Two and Six). However, as the polar climate is thermodynamically unfavourable and energy budgets are tight, flower colour pigments are often too costly to produce without being linked to a functional purpose. Hence, another possibility was that flower colour is linked to functions aside from signalling to attract pollinators (addressed in Chapter Four and Five).
The diversity and distribution of blue-purple flowers in the Arctic were correlated with the diversity and distribution of specialist pollinators (bumblebees) with this colour preference. The highest number of both blue-purple flowers and bumblebee diversity were registered in non-glaciated arctic regions, where populations may have persisted since the arctic biome originated (Chapter Two). These results indicate that although arctic bumblebees are few and rarely active, their presence and activity is sufficient to impact colonization efficiency in the Arctic, and in certain areas may drive further selection and diversification of flower colours. The most common flower colours throughout the Arctic were shown to be white, blue-purple and yellow (Chapter Two). White and yellow are reported to attract more general, less efficient pollinators such as Diptera. In Svalbard, no known specialist pollinators exist. To investigate possible colour preference among the putative pollinators in Svalbard, insect traps with different colours were distributed at four different sites. Eight different families of Diptera were caught. Mycetophilidae were the most common, and showed a preference for the colour red, whereas the Muscidae and Syrphidae were more common in yellow and white coloured traps. Diptera species are already known to act as pollinators of white and yellow flowers, and as such are likely performing some pollination in Svalbard. However Mycetophilidae are rarely noted for pollination, and the extent to which these Diptera are involved in outcrossing pollination of flowering species on Svalbard remains unknown. The effects of flower colour and floral heating were tested in Adventdalen, Svalbard, using Papaver dahlianum which has white and yellow flower colour morphs (Chapter Four). Floral heating effects were also briefly investigated for Pleurophyllum spp, Stilbocarpa polaris, Bulbinella rossii and Anisotome latifolia on subantarctic Campbell Island. Flower temperature was measured and compared with environmental conditions, including ambient air temperature. In P. dahlianum, yellow flowers were warmer in some situations, whereas white flowers were warmer in others. Floral shape was also of some significance. Dark purple flowers on Campbell Island were found to be significantly warmer than ambient temperatures under high levels of solar radiation. Seed production was proposed to vary between colour morphs of P. dahlianum in Svalbard due to differences in insect visitation and heating based on colour (Chapter Five). Yellow flowers were expected to produce more seeds due to being more attractive and warmer than white flowers. Seed production was measured using different pollination treatments (forced selfing, hand pollination, forced outcrossing) at four different sites within the two P. dahlianum colour morphs. P. dahlianum was capable of setting seed autogamously. The largest seeds were collected from Adventdalen, but there were no significant differences of seed size between flower colours. There were no significant differences between white and yellow flowered plants in terms of relative reproductive effort. Thus, colour is assumed to have little effect on reproductive output when few seasons are considered. Assortative mating, where mating could be structured around flower colour was investigated in Chapter Six. Genetic samples were collected from nineteen populations of P. dahlianum around Svalbard. Genetic fingerprinting via AFLPS was used to assess whether colour was associated with structuring of the population (Chapter Six). There was no genetic pattern based around colour, at regional, local or the population scale. There did appear to be several widespread genetic lineages present in Svalbard, which may represent genetic groups with reduced reproductive compatibility or cryptic species.
The existence of flower colour in polar regions was concluded to be partly due to the recent colonisation processes of the areas, whereas the role held by flower colour in polar regions was variable, and strongly dependent on the season of investigation. Due to seasonal conditions differing greatly between years, flower colour polymorphisms in Svalbard are likely maintained by different conditions favouring different morphs in different seasons.