The noninvasive discrimination of myelin disease from axonal loss and other pathologic confounds remains an unsolved problem in multiple sclerosis but may be possible through magnetic resonance quantitation of the intramyelinic water compartment. Technical challenges have limited the study of this approach in the spinal cord, a common site of involvement in multiple sclerosis. This technical note reports the test-retest reproducibility of a short T2-based estimate of myelin content in human spinal cord in vivo.