Spin-lattice (T1) relaxation times are conventionally estimated using inversion recovery or saturation recovery sequences. Such "recovery" methods are limited in magnetic resonance imaging by the long times required to collect multiple points along the T1 relaxation curve. This problem can be overcome by the use of "one-shot" methods, which collect all points along the relaxation curve in a single excitation. We have compared the relative efficiency of these methods, on the basis of the signal-to-noise ratio obtained in the calculated T1 image from an exam time of fixed duration. We have found that a one-shot method using stimulated echoes has a poor efficiency. However, a method based on a technique first proposed by Look and Locker has an efficiency that is almost equal to the inversion recovery method and therefore possesses highly favorable properties for T1 imaging.