It is common knowledge that of all the regulated automotive emissions, particulate emissions are most difficult to quantify as they comprise a complex mixture of particles of varying size and composition, each of which may be influenced by many external factors including engine technology, fuel composition, air-to-fuel ratio, lubricant oil, after-treatment and the act of measurement itself. The aim of the present work is to provide further guidance into better understanding the production mechanisms of such emissions in spark-ignition engines fueled with compressed natural gas. In particular, extensive experimental investigations were designed with the aim to isolate the contribution of the fuel from that of lubricant oil to particle emissions. This because the common thought is that particulate emerging from the engine derives mainly from fuel, otherwise the contribute of lubricant oil cannot be neglected or underestimated, especially when the fuel itself produces low levels of soot emissions, such as in the case of premixed natural gas. The fuel-derived contribution was studied by analyzing the influence that the air-to-fuel ratio and the natural gas composition have on soot emitted from a single-cylinder spark-ignition engine. To achieve this purpose, precise real-time methane/propane mixtures were realized by means of an innovative experimental set-up. The results were compared with pure methane and propane, as well as with natural gas. The mechanisms of oil-derived soot formation were investigated by realizing experiments where lubricant oil was injected either into the intake manifold or directly into the combustion chamber of an optically-accessible version of the aforementioned engine, requiring no lubrication, in order to mimic the different ways by which lubricant may reach the combustion chamber. The influences on soot emissions were assessed in terms of particulate number, mass and size distributions. Gaseous emissions and engine performance were also analyzed in order to globally monitor the combustion process.