We have been interested recently in exploring relaxation under multiple-pulse NMR as a means of characterizing polymer morphology. For such work it is important to know the effectiveness of homonuclear decoupling for different experimental conditions. One measure of how well the dipolar interactions are averaged is given by the linewidth in the chemical-shift spectrum recorded using multiple-pulse sequences, since the largest error Hamiltonians analyzed using the Magnus expansion are frequently those related to the dipolar interaction (1-3). Pressure-crystallized linear polyethylene (LPE) is a useful sample in this context because it represents one of the most strongly coupled proton spin systems available. An additional advantage is that crystallization of LPE under high pressure (4.86 GPa in this case) results in a material which is roughly 95% crystalline as indicated by density measurements (4). This largely eliminates complications arising from the superposition of crystalline and amorphous lineshapes. Details concerning the preparation of the sample have been published (4). In the spectrum of pressure-crystallized LPE obtained from a free induction decay, the full width at half maximum of the resonance is 80 kHz, or 400 ppm for a ‘H Larmor frequency of 200 MHz. The multiple-pulse spectrum recorded with BR-24 (2) on a Bruker CXP-200 spectrometer is shown in Fig. 1. The methylene ‘H shielding tensor displays axial symmetry, in agreement with reported methylene tensors for several small organic compounds (5). As is also the case with these smaller compounds, uII arises from the most shielded orientation. The observed anisotropy of the shielding tensor is 6.9 f 0.2 ppm. Gerstein (6) has previously reported the multiple-pulse spectrum of a high-density LPE obtained through MREV-8 (7, 8) with a cycle time of 2 1 ps. The reported value of the anisotropy is 4.72 ppm, but the axially symmetric character of the tensor is not as well defined as that in Fig. 1. We ascribe most of the improvement to our use of a spherical sample, which minimizes broadening from bulk magnetic susceptibility effects (9, IO). In addition, BR-24 is generally recognized to be more effective than MREV8 in homonuclear decoupling (3) since the design of BR-24 minimizes contributions from higher order error Hamiltonians. The data in Table 1 illustrate the superiority of BR-24 in analyzing the pressure-crystallized LPE. The cycle-time dependence of linewidth using MREV-8 is considerably stronger than that of BR-24, with the shortest cycle times offering the most effective elimination of dipolar couplings. An