A series of myelographies performed with a smaller than customary dose (3.75 g) of metrizamide was compared with a series using the larger customary dose. While little decrease in the incidence of headache and vomiting resulted from the decreased dose, there was a heartening drop in the incidence of psychoneurologic side effects. Little difference in diagnostic quality between the two series resulted when the contrast agent was injected close to the site of main interest, but total spinal canal myelography performed with the low dose is often inadequate.