Positive associations between short-term exposure to ambient particulate matter (PM) pollution and cardio-respiratory morbidity and mortality have been established in epidemiologic studies. However, scientific uncertainties remain regarding which PM constituents and sources are most harmful to the exposed population. While tentative evidence of positive links between certain PM constituents and sources with specific health outcomes exists, significant heterogeneity in study findings remains. The chemical composition and emission source of air pollution vary not only temporally, but also geographically. Thus, substantial research on these characteristics under different atmospheres (e.g., Asian atmosphere) is warranted to enhance our understanding of PM-related health effects. / In Hong Kong, air pollution levels often exceed the World Health Organization’s air quality guidelines, posing serious public health threat. Although two decades of active research have associated PM mass (weights) concentration in Hong Kong with elevated risk of daily mortality and emergency hospital admissions, individual chemical constituents and sources responsible for the adverse health effects associated with PM mass have rarely been examined. This thesis attempted to reduce the current scientific uncertainty by making use of the speciation data for PM with aerodynamic diameter ≤ 10 μm (PM₁₀) to examine the associations between chemical constituents and daily cardio-respiratory emergency hospital admissions in Hong Kong between 2001 and 2008. It also capitalized on Positive Matrix Factorization (PMF) source apportionment model to create an alternative measure of PM concentrations that quantified the relative contribution of PM₁₀ sources. This enabled the assessment of health risks associated with exposure to particle mixture from specific sources. / Time-series analyses conducted in this work showed evidence of positive links of emergency hospitalizations with multiple PM₁₀ constituents for various exposure lags examined. In multipollutant models adjusting for gaseous co-pollutants, three groups of constituents were significantly associated with increased risk of cause-specific hospitalizations. They included: combustion-related constituents (i.e., elemental carbon, organic matter, potassium ion, manganese, nitrate ion, arsenic, lead), sea salt-related constituents (i.e., sodium ion, chloride ion, magnesium), and constituents related to soil/road dust (i.e., aluminum, iron). Significant associations were most evident between November to April. PMF model identified eight PM₁₀ sources: 1) vehicle exhaust, 2) soil/road dust, 3) regional combustion, 4) residual oil, 5) fresh sea salt, 6) aged sea salt, 7) secondary nitrate, and 8) secondary sulfate. Health effect analysis demonstrated that all PM₁₀ sources, except fresh sea salt, were positively associated with emergency hospitalizations. Combustion-related sources were positively associated with risk of cardiovascular hospitalizations at shorter cumulative lags than with respiratory hospitalization. Sources that primarily generate coarser particles (i.e., soil/road dust and aged sea salt) were linked to respiratory hospitalizations at shorter cumulative lags than combustion sources that emit fine particles, which were associated with respiratory hospitalization at delayed cumulative lags. / This work may help prioritize future toxicological research on the biologic mechanisms linking PM pollution to cardio-respiratory health. It also stresses the importance of regulation and reduction of traffic and other combustion-related emissions, reconfiguration of urban environment to reduce personal exposure to traffic emissions, as well as establishment of a coordinated and robust regional-scale air quality management plan. Constituent- and source-based air quality standards and policy strategies should be considered, supplementary to standards for total PM mass (e.g., PM₁₀ and PM₂.₅), to effectively protect the population from air pollution mixture. For example, policy measures aiming at controlling anthropogenic sources of coarse particles (e.g., soil/road dust, precursor gases for atmospheric conversion of aged sea salt) should be advisable. / 以往流行病學研究已經建立大氣懸浮顆粒（PM）污染短期暴露和心血管﹑呼吸系統疾病的發病率以及死亡率之間的關聯，但仍然不能確定究竟那種PM化學成分和排放源對暴露人群最有害。初步證據已經顯示了個別PM化學成分和排放源與某些疾病的相關聯，但是，現今研究結果之間依然存在顯著差異。此外，顆粒污染物的化學成分和排放源會隨地理空間不同而變化，因此有必要在不同地域的大氣環境下（如亞洲）進行研究，從而提高我們對PM影響健康的認識。 / 香港的空氣質量常超過世界衛生組織規定的可接受空氣質量標準，空氣污染已經成為一個嚴重的公共衛生問題。香港二十年的空氣污染研究已顯示PM質量濃度會增加死亡率和緊急入院的風險，然而仍然缺乏PM化學成分和排放源對健康影響的研究。本港環境保護署擁有十多年連續測量的可吸入懸浮粒子（PM₁₀）成分數據，因此本論文利用這數據與公共衛生數據來填補這個學術空白。具體的，本論文探索香港2001年至2008年間的PM化學成分與期間每日心血管和呼吸系統緊急入院的關聯。本論文還利用正交矩陣因子分解法（PMF），即大氣顆粒物的源解析技術，進行顆粒源分離，已建立替代單元以分別代表PM₁₀不同排放源的相對貢獻建立一個替代措施量化PM₁₀的成分中不同來源相對的貢獻，最終評估暴露於特定來源的PM₁₀混合成分的健康風險。 / 本文中的時間序列分析研究結果表明緊急入院率與多個PM化學成分在不同的暴露滯後時間上均存在正關聯。本論文的多種污染物時間序列的模型對氣體的共污染性進行了控制，研究發現以下三种顆粒物對某些具體疾病的緊急入院風險有顯著增加︰燃燒有關的顆粒（即︰碳元素，有機物，鉀離子，錳，硝酸根離子，砷，鉛），與海鹽有關的顆粒（即︰鈉離子，氯離子，鎂），以及與土壤/道路粉塵有關的顆粒（即︰鋁，鐵），這種相關性在寒冷的季節更為明顯（即每年的11月至4月）。PMF鑑定了八個PM₁₀的排放源或排放組合，包括︰1）汽車排氣，2）土壤/道路粉塵，3）區域燃燒產物，4）渣油，5）鮮海鹽，6）陳海鹽，7）二次硝酸鹽，和8）二次硫酸鹽。分析發現，除鮮海鹽外，其他PM₁₀的排放源與緊急入院呈正關係。研究顯示與燃燒有關的排放源能增加即時因心血管疾病緊急入院的風險，而主要含有粗顆粒的排放源能增加即時呼吸系統疾病緊急入院的風險，且在作用時間上比燃燒排放源（主要含有細顆粒）的危害更快。 / 本論文所提供的研究成果有助於重新定位未來有關空氣污染與心肺健康的生物學機制毒理學研究方向。研究結果強調監管和降低交通以及其他有關燃燒排放的重要性，利用合理城市環境結構及配置來減少居民暴露交通空氣污染的風險，以及建立區域性的空氣質量管理計劃。此外，本研究還建議除了PM質量濃度（如PM₁₀和PM₂.₅）的空氣質量標準以外，針對PM化學成分和排放源的空氣質量標準也應加以考慮，以有效地保護居民免受空氣污染的危害，例如，應考慮有關粗顆粒（如土壤/道路麈和前驅氣體）的人為來源的政策措施。 / Pun, Chit. / Thesis Ph.D. Chinese University of Hong Kong 2014. / Includes bibliographical references (leaves 108-133). / Abstracts also in Chinese. / Title from PDF title page (viewed on 18, October, 2016). / Pun, Chit. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only. / Detailed summary in vernacular field only.