The Carroll rating scales (CRS) was developed as a self rating instrument for depression, closely matching the information content and specific items of the Hamilton rating scales (HRS). The CRS was found to have acceptable face validity and reliability. The concurrent validity of the CRS was acceptable, based on comparisons with the HRS and the Beck Depression Inventory (BDI). The internal consistency of the CRS was very similar to that of the HRS. The CRS contained information about HRS scores beyond what could be predicted from BDI scores, but the BDI did not predict HRS scores beyond what could be predicted from CRS scores. The CRS and BDI scores were strongly correlated and both had access to a subjective dimension of depression that could not be predicted from HRS scores. The complementary uses of self ratings and observer ratings are evident from these results. The CRS may be a useful alternative to the BDI as a self rating scale, with the additional advantage of closer correspondence to the HRS.