It wm found that all pieces of poly(methy1 methacrylate) have structure and aniaotropy, no matter how they were made. There ia no continuum, only differences of opinion about what the discontinuities represent. Judging from these and other results and pending the reaulta of further experiments, the author ia of the opinion that he hae uncovered both macromolecular and micellar domains in poly(methy1 methacrylate).Regardless of wishes or expectations, a consolidated polymer is not always a homogeneous continuum. Stading in 1949,' if not before, it has been shown electron microscopically that a polymer may manifest various discrete domains according to kind, treatment, and properties. Coldfractured, compression-molded polyacrylonitrile, for example, had been depicted in replica as domains roughly globular in shape, varying in average diameter from about 200 to sdme 840 A. The domains were related to those shown in the dried emulsion from which the consolidated specimen was molded. In a self-supporting film cast from a solution of such a polymer, however, much smaller units dominated the picture.2.a The small units were calculated on the basis of the bulk density to be of the order of 100,OOO in average molecular weight. This value was commensurate with that obtained by light scattering.Since variations in electron microscopical technique were involved, it is repeated here that the fracture surfaces were replicated with gelatin which in turn was replicated with silica, but there was no shadowing with a heavy metal. On the other hand, the specimens of dried emulsion and cast f i l m were shadowed with a heavy metal, but no replicas were involved.Evidently, particulations of the preparatory materials did not enter the picture in either case. Evidently, too, the domains shown in the fracture surface and in the dried emulsion were families of macromolecules. They were termed micelles by Eirich and Mark4 in 1950. The boundaries of both discrete macromolecules and micelles may be observed in various portions of an acrylic fiber,5 that is, at least two of the levels6 of macromolecular organization must be considered. Multimacromolecular units can now be explained in the light of Gillespie's recent publi~ation.~