A 260 m length of the British end of the service tunnel for the Channel tunnel was driven over a period of two and a half months early in 1975. The operation of the tunnelling machine was observed as it advanced through the Lower Chalk and is described in terms of its principal operations, namely cutting, steering, spoil removal and lining. In addition to the normal operation of tunnelling, exploratory probe holes were drilled ahead of the tunnel face. The aim of this study was to assess the suitability of the tunnelling system for the specific task of tunnelling in chalk and to establish which of its operations most constrained the rate of tunnelling. The quality of the chalk was measured at intervals using the technique of face-indexing and samples from the face and spoil were analysed. These tests indicated the near-homogeneous nature of the Chalk Marl throughout the drive. The tunnelling machine is shown to be capable of excavating and lining at a rate of 2.4 m/h on average. Probing ahead, even with optimum dreilling rates, represents 16 per cent of the total construction time so effectively reduces the tunnelling rate to 2.0 m/h on average. The study identified the critical operations and indicated that quite simple modifications could well increase overall tunnelling rates to about 3 m/h; and even further improvement appears possible.