A good model of a disorder is one that: (a) mimics, although in a simpler form than the full-blown clinical case, the fundamentals of the behavioral characteristics, in this case of people with Attention-Deficit/Hyperactivity Disorder (AD/HD;face validity); (b) conforms with a theoretical rationale for the disorder (construct validity); and (c) is able to predict aspects of behavior, genetics and neurobiology previously uncharted in the clinics (predictive validity). This article discusses the Spontaneously Hypertensive Rat (SHR) and some other putative animal models of AD/HD. It is argued that although other strains and species may be hyperactive and/or show attention deficits following genetic, environmental or pharmacological interventions, the SHR is presently the only strain shown to have the major behavioral symptoms of AD/HD. This dues not mean that investigating other models cannot give valuable information. (C) 2000 Elsevier Science Ltd. All rights reserved.