In the field of art history, there is a great potential for the use of art image databases. However, there is little literature that deals specifically with the needs and desires of art historians when it comes to art image databases. This thesis fills the gap in the literature with a user study of the use of art image databases by art historians. It includes a survey of what databases are used, why they are used, and to what extent they are used by art history professors. Results indicate that many issues in the usability and design of currently available art image databases exist, and the implications of these issues are discussed. These issues include image quality, more diversity of images, and the need for accurate and consistent metadata. The need to use multiple art databases is explored, as well as the use of the art image database as a finding aid versus the art image database as a research tool.