The serum concentrations of cortisol, 17α-hydroxypregnenolone, 17α-hydroxyprogesterone, 21-deoxycortisol and 11-deoxycortisol were measured in 19 healthy dogs, 15 dogs with pituitary-dependent hypercortisolism (pdh) and eight dogs with other diseases before and one hour after an injection of synthetic adrenocorticotrophic hormone (acth). At both times the dogs with pdh had significantly higher concentrations of cortisol, 17α-hydroxypregnenolone, 17α-hydroxyprogesterone and 21-deoxycortisol than the healthy dogs. Basal 11-deoxycortisol concentrations were also significantly higher in dogs with pdh compared with healthy dogs. When compared with the dogs with other diseases, the dogs with pdh had significantly higher basal and post-acth cortisol and basal 21-deoxycortisol, and significantly lower post-acth 11-deoxycortisol concentrations. The dogs with other diseases had significantly higher post-acth cortisol, 17α-hydroxyprogesterone and 11-deoxycortisol concentrations than the healthy dogs. In general, the post-acth concentrations of 17α-hydroxypregnenolone, 17α-hydroxyprogesterone, 11-deoxycortisol and 21-deoxycortisol were more variable than the post-acth concentrations of cortisol, resulting in large overlaps of the concentrations of these hormones between the three groups. A two-graph receiver operating characteristic (roc) analysis was used to maximise the sensitivity and specificity of each hormone for diagnosing hypercortisolism; it showed that the post-acth concentration of cortisol had the highest sensitivity and specificity. The overlaps between the healthy dogs, the dogs with pdh and the dogs with other diseases suggested that the individual precursor hormones would not be useful as a screening test for hypercortisolism.