Two models have been proposed to explain the two melting endotherms commonly observed in polyethers: (1) the crystals with the lower melting temperature may transform into a higher melting form during heating; (2) the two melting peaks come from separate populations of crystals. Our TEM studies of thin films of PEEK have shown two crystal populations, with widely differing lamellar thicknesses, which can occur in quite separate areas of the samples. These results strongly support the second model.