Abstract Objectives: Delirium is associated with extensive cognitive deficits, yet recent studies have proposed that there may be cognitive abilities which are preserved throughout its course. The present study aimed to determine whether crystallised cognitive abilities would remain stable as fluid cognition declined in the presence of delirium. Methods: Patients with delirium (n=15), and controls with no delirium (n=19), were tested on two sessions. The change in score between sessions was calculated and compared between-groups. Crystallised cognition was measured using a revised version of Castles and Coltheart’s test of reading and a newly devised colour preference task. Fluid abilities were measured using the digit span forward and backward. Diagnosis of delirium was done via the CAM and severity was tracked using the MDAS. Results: Mann Whitney U tests indicated that change scores for crystallised abilities on the reading test were not significantly larger in the delirium group in line with the hypothesis but were on colour preference. Contradictory to the hypothesis fluid abilities did not change significantly more in the presence of delirium. Post-test power calculations indicated that there was 39 -61% power to detect medium to large effect sizes indicating in insufficient power. Conclusions: The results indicated that colour preference was not stable in delirium, yet previous studies have found it to be preserved in dementia. Although further evidence is required this may be a useful domain for differential diagnosis. Methodological issues and limited power make it difficult to infer any concrete conclusions from the present study.