Recently, Bothner-By and co-workers proposed a new method for measuring homonuclear NOE effects under spin-locked conditions, which they named the CAMELSPIN experiment (I). We prefer to refer to this experiment as rotatingframe Overhauser enhancement spectroscopy (ROESY) since this name suggests a similarity with the closely related 2D NOESY experiment (2, 3). The ROESY method is particularly suitable for molecules that have a motional correlation time, TV, near the condition UT, = 1, where w is the angular Larmor frequency. In this case the laboratory-frame NOE effect is near zero, whereas the rotating-frame NOE (under spin-locked conditions) is always positive and monotonically increases for increasing value of T,, and can be significantly large under those conditions. The pulse scheme of the ROESY experiment is sketched in Fig. 1. At the end of the evolution period, a strong rf field is switched on. for a duration 7,. During this mixing time, T,, projections of the magnetization vectors present at the end of the evolution period, t,, onto the vector of the effective rf field remain spin locked along this effective t-f field vector. Spin exchange among spin-locked magnetization components of different nuclei can then occur, analogously to spin exchange in the regular NOE experiments, where magnetizations are aligned along the -tz axis. The spectral density functions describing spin exchange in the ROESY experiment differ from those in the NOESY experiment and cause the effect to be positive for all values of 7,. We will analyze some experimental considerations that are relevant to this new technique. In particular, we will consider the effect of limited rf field strength of the spin-locking field as well as artifacts that can occur in this type of experiment. First, as is the case for the NOESY experiment, it is very advantageous to record the ROESY spectrum in the 2D absorption mode. .4s will be shown below, there is an additional advantage in this type of experiment to the use of absorption-mode spectra: distinction between positive and negative peaks facilitates the identification of certain types of artifacts. A 2D hypercomplex Fourier transformation, as outlined by Miiller and Ernst (4) and States et al. (5) is conveniently used in this type of experiment. A simple description of this procedure is presented elsewhere (6). A minimum of four experiments with phase cycling according to Table 1 is required