Unable to escape their predators, plants have developed a wide range of complex defence strategies throughout their millions of years of evolution. These can be broadly categorized into strategies that are always there (constitute) and present when provoked (induced), as well as strategies that impact predators direct or indirectly. One significant class of direct defence strategies which has been studied are plant specialized secondary metabolites. Included in this group are cyanogenic glucosides(CNglcs), the topic of this thesis, which deter predators by facilitating the release of poisonous cyanide during an herbivore attack. CNglcs have been found throughout the plant kingdom, from ancient ferns to gymnosperms and angiosperms. Some of the world’s most economically important cultivated crop plants are cyanogenic, including representatives from the legume family and fruit trees. While being cyanogenic is an advantage for crop plants due to increased resistance towards herbivores, it also encompasses the risk of cyanide poisoning of humans and livestock. Therefore, studying cyanogenic glucosides is key to understanding and developing the next generation of crop plants.To comprehend and regulate C Nglcs in plants, it is necessary to understand the enzymatic pathways surrounding them. CNglcs are synthesised in the biosynthesis pathway, which initiates with an amino acids precursor and has oximes and cyanohydrins (α-hydroxynitriles) as key intermediates. For some cyanogenic species, the genes responsible for synthesising CNglcs are co-localized in a functional gene cluster. These gene clusters are not only copies of homologous genes, which result from gene duplication, but consist of non-homologous genes encoding different types of enzymes of the same biosynthetic pathway. This makes CNglcs a very dynamic class of defence compounds to study. After biosynthesis, the CNglcs are stable within the plant cell. However, during cell disruption from an herbivore attack, the CNglcs get into contact with specific β-glucosidases, which initiates the bioactivation pathway. This results in the release of toxic hydrogen cyanide (HCN). To prevent autotoxicity, the plants have additionally developed a detoxification pathway, where the poisonous compounds are transformed into ammonia and amino acids, which can then be reused in primary metabolism. These three constitute the main pathways in the core function of CNglcs -plant defence- . However, recent studies have indicated that CNglcs are not only involved in plant defence, but also in functions such as seed germination, modulators of oxidative stress, flower development and bud breaking as well as carbon and nitrogen transporters. For the latter, two pathways have recently been identified in some plant species which turn over cyanogenic glucosides and return carbon and ammonia back into the primary metabolism. These have been named the “alternative turnover pathway” and the “recycling pathway”. The purpose of this PhD thesis was to deepen the common knowledge of the pathways surrounding CNglcs in two plant species: Almond (Prunus dulcis) and Lima bean (Phaseolus lunatus), as well as identifying low or non-cyanogenic varieties of common vetch (Vicia sativa) for commercial use. Almond (Prunus dulcis) is an economically important fruit tree, which contains the CNglcs prunasin and amygdalin. It was studied in two genotypes, sweet and bitter, which differs in the presence of amygdalin in the kernel. Almond has a mostly characterized biosynthesis, bioactivation and detoxification pathway, including the recent discovery of five helix-loop-helix transcription factors involved in regulating the biosynthesis of CNglcs in the two genotypes. The enzymes involved in the putative alternative turnover pathway have been suggested, but not proven beforehand, but the enzymes involved in a putative recycling pathway have not been described in almond. In this work,these pathways were explored even further by studying the transcriptome and metabolite levels in both a bitter and a sweet almond seed during development into a seedling. By identifying differentially expressed genes between the sweet and bitter genotype, and comparing them to metabolite levels, six putative genes were proposed in the pathways. Specifically, these were one putative amidase and two putative decarboxylases in the alternative turnover pathway and two putative glutathione transferases in the putative recycling pathway. In addition, a new putative amygdalin hydrolase was proposed to be active in seeds. The second and third species of plants studied in this thesis belong to the legume family. The first of these is Lima bean (Phaseolus lunatus), which is an important crop plant in many tropical and subtropical regions around the world. Lima bean has been shown to possess the CNglcs linamarin and lotaustralin, but the genes involved in its biosynthesis pathway have not been elucidated yet. In this thesis, the genes responsible for forming linamarin and lotaustralin were identified via cloning experiments using sequence homology and a gene cluster organisation in a close relative of lima bean. Notable in this discovery was that the second step of the biosynthesis pathway was performed by a different enzyme subfamily than previously described in other cyanogenic plants. Additionally, it was concluded that the occurrence of linamarin and lotaustralin as functional chemical defence compounds appears restricted to species belonging to the closely related Polystachios and Lunatus groups of the Phaseolus genus.