reasons. Stabs were mostly performed on only one finger, and in particular one hand, to avoid disturbing the paralysed arm. The patient was subjected to frequent stabs when blood glucose concentrations remained satisfactory. No device was used to limit penetration of the skin, which might have been too deep. He also had clinical evidence of small vessel disease affecting his arms. Although pulp infection and gangrene are unusual complications offinger pricks to obtain capillary blood, they are still possible and may be prevented. Just as Young et al emphasised the importance ofinspecting insulin injection sites regularly,5 we would suggest that similar caution be paid to sites used to obtain capillary blood. A proper technique must be adhered to, and in particular it would seem wise to use several fingers in rotation and a device to limit skin penetration. Repeated stabs should be kept to a minimum. The site chosen should be clean and warm, but alcohol swabs should not be used, as repeated use tends to toughen the skin. If impaired blood supply to the finger is suspected other sites such as earlobes should be considered seriously.