One major drawback of spark-ignition direct-injection (SIDI) engines is increased particulate matter (PM) emissions at high load, due to increased wall wetting and a reduction in available mixture preparation time when compared to port-fuel injection (PFI). It is therefore necessary to understand the mechanics behind injection strategies which are capable of reducing these emissions while also maintaining the performance and efficiency of the engine. Splitting the fuel delivery into two or more injections is a proven way of working towards this goal, however, many different injection permutations are possible and as such there is no clear consensus on what constitutes an ideal strategy for any given objective. In this study, the effect of the timing of the first and second injections for an evenly split dual injection strategy are investigated in an optical SIDI engine running at 1200 RPM with an unthrottled intake. Performance parameters derived from in-cylinder pressure data are analysed alongside high-speed natural flame luminescence images in order to obtain relationships between engine output and the physical properties associated with flame propagation. It was found that the best performing injection strategy consisted of one injection relatively early in the intake stroke and one injection towards the middle of the compression stroke, likely due to the combination of both ample mixture formation time and a high level of spray induced turbulence. Interestingly the trends found in the results are significantly non-monotonic, as there are a series of peaks and troughs in indicated mean effective pressure (IMEP) and combustion phasing (CA50) as the injection timings are varied. The relationships between these indicated performance parameters and flame propagation were also analysed, where it was discovered that flame speed is very closely correlated with CA50, regardless of the value of IMEP. Additionally, an elliptical approximation was calculated for each flame in an attempt to establish a link between injection timing and flame shape. It was found that retarded second injections lead to a less circular flame, potentially as a result of decreased charge homogeneity causing local regions of varied laminar flame speed throughout the combustion chamber. Furthermore, analysis of eccentricity and CA50 graphs suggest that there is a link between retarded combustion phasing and non-circular flame fronts.