The early flame kernel development period has a strong influence on the ultimate performance and emission characteristics of spark ignition engines. The fibre-optic instrumented spark plug, FOSP, is a tool used to characterise the early flame kernel development period without the need to modify production engines. Simultaneous in-cylinder pressure, fibre-optic spark plug and secondary ignition system voltage measurements have been made in the GM 2.8 L and the high swirl 3.1 L production engines modified to run on a single cylinder. The secondary ignition system voltages indicate that restrikes are occurring and that spark anemometry is a promising tool to extract information about the flow near the spark plug at the time of ignition. Further development of the technique is, however, required. In regard to FOSP data processing, it was found that i) the flame growth rate should be measured relative to the ignition point instead of the flame kernel centroid to account for the presence of restrikes, and ii) the FOSP measures a mass-weighted convection velocity. Normalisation of the measured convection velocity by the mean piston speed and unburned gas density at the time of ignition accounts for a large part of the convection velocity variations with engine speed and load; the so-normalised velocity was found to decay during the compression stroke. The initial flame growth rate was found to be strongly dependent on the convection velocity, and to increase during the compression stroke. Cyclic variations in the flame growth rate were found to correlate well with LPP and the 0-1% and 0-50% mass burn durations.