Following our successful survey of the Northern Hubble Deep Field at 6.7 and 15 mu, we propose to survey the Southern Hubble Deep with CAM at the same wavelengths. Data reduction on the Northern HDF is nearing completion and we hope to release the data received to date in November. We believe that we have detected a number of HDF galaxies both at 6.7 and 15 mu We have been carrying out a detailed study of the Hubble Deep Field (HDF) since its release and have published a paper on the nature of the faint galaxies detected in the HDF (Mobasher et al 1996). The key point for ISO is the very high proportion of very blue, star-forming galaxies in the sample. 80% of the galaxies have colours consistent with being in the process of undergoing vigorous star formation and it is a reasonable expectation that a high proportion of them will have strong far infrared excesses. We have used our standard starburst spectral energy distributions (Pearson and Rowan-Robinson 1996, MN in press) to predict the 6.7 and 15 mu fluxes from the HDF galaxies. We predict that the most sensitive CAM detectors (LW2 and LW3) are capable of detecting tens of HDF galaxies with S/N=5 in integrations of 2 hours per detector for each of the 3 WF fields. This is close to the conf- usion limit for CAM, so that longer integrations are unlikely to be worthwhile. The software we are already developing for analysis of the ELAIS data has been adapted for analysis of the HDF fields and we have a strong team to carry out this data analysis at ICSTM, supported by the other teams of the consortium. Our expertise in the analysis of ISO survey data, coupled with the detailed work we have been carrying out on the HDF, ensures that high-quality data and science analysis products would be delivered to the community if this proposal were successful.