Light microscope observations provide further evidence of transcellular strands in sieve elements. Specimens were prepared by a new method for stripping sieve tubes from phloem tissue after fixation in aldehyde solutions. Straight-sided membranous structures appear in sieve elements and are interpreted as the boundary membranes of transcellular strands. Occasionally it is possible to follow these structures through a sieve plate. The proposal that transcellular strands are artefacts due to diffraction lines is discussed and, in relation to present results, is proved incorrect. In the interference microscope, colour contrast is demonstrated between a surrounding matrix and the strands at a sieve plate; since this colour is a function of density and thickness in the specimen, structures which display a particular colour must be real. Spherical granules, some of which are more refractive than others, and fibrils, which may occur in parallel groups to form fibrillar strands, are sometimes seen within straight-sided boundaries, but more often these constituents occur in the lumina of sieve elements. Fibrils may spread out to fill the whole of the sieve element and fan-like arrays of fibrils are seen issuing from strands in sieve elements and phloem exudate. Membranous and fibrillar strands are orientated parallel to the long axis in fixed, unstained sieve elements. Subsequent staining reveals an additional constituent in the same sieve elements which is like a typical slime plug and is distinguishable from strands by a lack of structure and heavier staining. Autoradiographs of phloem exudate show an association between the distribution of mobile carbon-14 and parts of transcellular strands. Mobile carbon-14 within strands appears to be orientated in parallel lines but there is no evidence of a particulate distribution of the isotope. This result suggests the mobile carbon-14 followed the distribution of the parallel fibrils and is not associated with plastids or granules. A pattern of displaced contents in sieve elements, which is repeated in superposed cells, indicates the presence of resistance to pressure change along intact sieve tubes. Such resistance would prevent movement of solution in response to a pressure gradient and is consequently incompatible with the mass-flow theory. This result can be explained if the movement of vacuolar fluid through sieve pores is blocked by cytoplasm which introduces a resistance to pressure change between cells along a sieve tube.