A previous review of alternative treatments (Tx) of ADHD--those other than psychoactive medication and behavioral/psychosocial Tx--was supplemented with an additional literature search focused on adults with ADHD. Twenty-four alternative Tx were identified, ranging in scientific documentation from discrediting controlled studies through mere hypotheses to positive controlled double-blind clinical trials. Many of them are applicable only to a specific subgroup. Although oligoantigenic (few-foods) diets have convincing double-blind evidence of efficacy for a properly selected subgroup of children, they do not appear promising for adults. Enzyme-potentiated desensitization, relaxation/EMG biofeedback, and deleading also have controlled evidence of efficacy. Iron supplementation, magnesium supplementation, Chinese herbals, EEG biofeedback, massage, meditation, mirror feedback, channel-specific perceptual training, and vestibular stimulation all have promising prospective pilot data, many of these tests reasonably controlled. Single-vitamin megadosage has some intriguing pilot trial data. Zinc supplementation is hypothetically supported by systematic case-control data, but no systematic clinical trial. Laser acupuncture has promising unpublished pilot data and may be more applicable to adults than children. Essential fatty acid supplementation has promising systematic case-control data, but clinical trials are equivocal. RDA vitamin supplementation, non-Chinese herbals, homeopathic remedies, and antifungal therapy have no systematic data in ADHD. Megadose multivitamin combinations are probably ineffective for most patients and are possibly dangerous. Simple sugar restriction seems ineffective. Amino acid supplementation is mildly effective in the short term, but not beyond 2-3 months. Thyroid treatment is effective in the presence of documented thyroid abnormality. Some alternative Tx of ADHD are effective or probably effective, but mainly for certain patients. In some cases, they are the Tx of choice, and initial evaluation should consider the relevant etiologies. A few have failed to prove effective in controlled trials. Most need research to determine whether they are effective and/or to define the applicable subgroup. Some of them, although not safer than standard Tx, may be preferable for an etiologic subgroup.