The Evolution and Development of Nectar Spurs

Nectar spurs are tubular structures usually borne from perianth tissue, which contain, or appear to contain, a floral reward such as nectar or oil. Nectar spurs have convergently evolved in many disparate angiosperm lineages, often in association with long-proboscid pollinators. The spur provides a potential mechanism through which reproductive isolation (a key precursor to speciation) can arise in sympatry – it is hypothesised that a change in spur length could evoke a shift in a plant’s primary pollinator, partially isolating the mutant plant from its conspecifics. Perhaps as a result of this mechanism, spurred taxa tend to be more species rich than related non-spurred taxa, suggesting the nectar spur to be an evolutionary ‘key innovation’ – a trait that promotes a high rate of speciation. Despite the evolutionary and ecological significance of nectar spurs, and despite them being common to many plant families, surprisingly little is known about how nectar spurs develop, and the genetics underpinning that development remain unexplored in most systems. In the absence of a spurred model organism, this thesis aimed to advance our understanding of nectar spur development in the toadflax genus *Linaria* (Plantaginaceae) – a sister taxon to the well-studied but spurless snapdragon genus *Antirrhinum*. Candidate genes hypothesised to play a role in spur initiation were selected based on a literature review and the results of a recent transcriptomic analysis comparing *A. majus* to *L. vulgaris*. Candidate gene function was explored through RT-qPCR-based analysis of gene expression in the developing flowers of the long-spurred *L. becerrae* and its short-spurred relative *L. clementei*; through heterologous expression of *Linaria* candidates in tobacco; through virus-induced gene silencing (VIGS) of candidates in *L. vulgaris*; and through the attempted overexpression of one candidate in *L. vulgaris*. Protocols for VIGS and stable transformation were trialled and developed for use in *L. vulgaris*, marking the first successful attempts to transform this species. In addition to exploring genes hypothesised to play a role in spur *initiation*, candidate genes controlling spur *length* were sought. The long-spurred *L. becerrae* and short-spurred *L. clementei* were used to generate a segregating F2 population, which formed the basis of a bulked segregant analysis (BSA). Spur length was found to be a genetically complex trait, potentially controlled by many quantitative trait loci (QTLs). Clusters of single nucleotide polymorphisms (SNPs) in chromosomes 3 and 4 appear to co-segregate with spur length, and the potential functions of genes associated with these SNPs were explored. The genomic dataset used in the BSA lays the groundwork for a more comprehensive QTL analysis in this system, taking the specific genotype and spur length of each individually sequenced F2 plant into account. Beyond the Plantaginaceae, spur development was investigated at the cellular level in *Diascia barberae* (Scrophulariaceae), a close relative of *Linaria* which evolved its spurs independently. A seemingly novel mechanism of spur development was revealed, whereby cell divisions in the developing *Diascia* spur cease much later than in previously characterised systems. Overall, this work provides new insights into spur development in *Linaria* and *Diascia*; demonstrates that *L. vulgaris* is somewhat amenable to transient and stable transformation; and presents a genomic dataset that is expected to shed further light on the genes controlling spur length in *Linaria*.