On 1992 July 10 the European Space Agency's spaceprobe Giotto passed the nucleus of the comparatively inactive comet 26P/Grigg-Skjellerup at a relative velocity of 14 km S-I. This Giotto Extended Mission (GEM) followed a highly successful encounter in 1986 with Comet 1PlHalley. We present results returned from the Optical Probe Experiment (OPE) and in particular consider data gathered by the channels sensitive to the scattering of solar light by cometary dust grains, in emission-free continuum bands. Owing to the demise of the Halley Multicolour Camera (HMC) during the Halley encounter, and the low number of impacts registered by the Dust Impact Detection System (DIDSY), OPE data offer the best indication of the actual encounter geometry. We find that it is likely that Giotto was on the sunward side of the shadow tenmnator plane at closest approach, with our modelling results suggesting that Giotto passed -100 km from the nucleus (although distances of up to 300 km cannot be ruled out). We investigate possible causes of the striking 'spike' features, or 'events', in the OPE data. While scattering of sunlight from ejecta particles as a result of dust impacts on the spacecraft body cannot be ruled out, considerations of the hypervelocity impact mechanisms and impact geometry show that this explanation is not without problems, and more investigation is needed before it can be conclusively accepted. As an alternative solution, we find that the complex data profiles can be fitted by jet activity in the innennost coma (which was not resolvable by ground-based observers). One particular event occurring at least 1000 km from the nucleus can be fitted if the OPE line of sight passes close to a nucleus fragment of radius 10-100 m which is situated around 50 km from the spacecraft and which is producing a small dust coma.