The surface of plant leaves, termed phyllosphere, is a ubiquitous microbial habitat that harbours diverse communities of microorganisms. Although a growing body of experimental evidence demonstrates that these microorganisms can have an influential role in host physiology, the ecological processes that drive the assembly of natural phyllosphere communities remain poorly understood. Leptospermum scoparium (mānuka) is an indigenous New Zealand tea tree widely known for the non-peroxide antibacterial properties of its honey. However, the host physiological traits associated with these properties exhibit variation that remain unexplained despite decades of research. Considering a preliminary study that identified spatially persistent host association in the mānuka phyllosphere microbiome with patterns congruent with those of a microbial community under strong host selection, the primary objective of this research was to generate a holistic understanding of the ecological processes underpinning community assembly in the natural mānuka phyllosphere. Since the host specificity of the mānuka phyllosphere microbiome was unquantified, this PhD thesis research began with a multi-species, spatially hierarchical survey of a native forest to understand the relative influence of host species identity versus distance on the phyllosphere microbiome of mānuka and ecologically similar, adjacent native plant species. The results revealed that the relative influence of host species identity on the phyllosphere microbiome was quantitatively stronger in mānuka compared to other plant species, and mānuka species-specificity was not associated with leaf morphological traits. Using a pair of morphologically indistinguishable and naturally co-occurring plant species (mānuka and Kunzea ericoides [kānuka]), I then explored the relative influence of host species identity and leaf morphology on inter-host dispersal. Specifically, I addressed a longstanding yet under-examined hypothesis that the relative strength of the phyllosphere microbiome as a source of dispersing microorganisms (i.e., source-strength) is contingent on leaf morphology. The results revealed considerable spatial heterogeneity among morphologically indistinguishable leaves at small spatial scales and suggested that mānuka may act as a stronger source of phyllosphere microorganisms than kānuka. These findings suggest that source-strength is determined by the quantitative difference in the relative strength of host selection among plant species. To contextualise this apparent spatial stability and host species-specificity of the mānuka phyllosphere microbiome, temporal variation was investigated by replicate sampling in three different seasons. My results revealed that the relative influence of individual host trees was larger than the season, and specific phyllosphere taxa persisted across time. Additionally, my results revealed an increased core microbiome during summer flowering, suggesting an association between host selection strength and host phenology. Lastly, I explored the relationship between the mānuka phyllosphere microbiome and mānuka honey quality. I sampled trees and honey from three adjacent mānuka populations known to exhibit visually discrete phenological traits (i.e., flowering time). My results revealed correlations between phyllosphere community composition and chemical properties of mānuka honey, including the primary constituent of mānuka’s non-peroxide antibacterial properties (i.e., methylglyoxal). Through the incorporation of spatial and temporal sampling designs, as well as a multi-disciplinary case study, this thesis provides a holistic understanding of the relative influence of host selection (abiotic and biotic), dispersal (short- and long-distance), and climate, on the assembly of the mānuka phyllosphere microbiome. These results also provide new perspectives on prevailing controversies (e.g., host selection vs. dispersal), address unverified hypotheses (e.g., source-strength), and illustrate a path forward that will allow the emergence of a coherent and generalisable understanding of phyllosphere microbial ecology.