This research aimed to examine whether prelingually deaf, orally educated children are capable of phonological coding as suggested by Dodd and Hermelin (1977). Their experiment was successfully replicated, but analysis of stimulus material and subjects' errors indicated use of visual coding. This was explored further in Experiment 2, which studied the preferred coding strategies of deaf and hearing children. Although both groups most favoured semantic associates, the deaf preferred visual matches as second choice whereas the hearing selected homophones. Both groups also differed in their third and fourth choices. In the subsequent experiment, the Stroop effect was used to assess coding strategy. However, subjects did not exhibit the classic interference effect and no definite conclusions could be drawn. Experiment 4 investigated the Dodd et al. (1983) hypothesis that the recency effect observed with auditory and lipread stimuli reflects phonological coding. However, it was demonstrated that recency is also obtained with visual stimuli requiring temporal integration. This contradicts the phonological coding interpretation. Experiment 5 examined the effect of accent on deaf Scottish and English subjects' ability to remember homophones. Subjects were presented with Scottish and English homophones and words homophonic in both accents. Scottish subjects made significantly more errors matching Scottish than English homophones; English subjects' performance was comparable on both, demonstrating visual coding. In Experiment 6, rhyming visually dissimilar word pairs were presented. Deaf children made similar numbers of errors matching these as they did with letter position control words. Subjects also performed equally on word pairs which were only visually alike and pairs which rhymed and were visually similar. Thus phonological similarity added little to matching performance beyond that provided by visual similarity. It is concluded that no evidence was found to support phonological coding in the deaf subjects used. Implications for teaching methods are discussed.