A previous paper (K. M. Entwistle and N. J. Terrill, J. Mater. Sci 35 (2000) 1675) reported measurements of the micro-fibril angle of the cellulose fibres in the S2 layers in softwood. The method involved irradiating a specimen with X-rays directed at 45 degrees to the radial and to the transverse direction. The determination of the micro-fibril angle from the distribution of intensity round the (002) diffraction circle assumed an idealised structure in which all the cell walls were oriented precisely in either the radial or the transverse direction. The real cell structure diverges significantly from this ideal. The error involved in assuming the ideal structure has been assessed by making a quantitative image analysis of the length and orientations of the cell walls in a section of the specimen comparable in cross-section to that irradiated by the X-rays. The position of the S2 intensity peaks round the (002) diffraction circle has been calculated and compared with those predicted from the idealised structure. It is demonstrated that the error involved in extracting a value for the micro-fibril angle from the measured intensities using the assumed ideal cell structure of the previous paper (cited above) is not more than one degree in the range of micro-fibril angles 20 degrees -30 degrees. So the approximate analysis is adequate for many purposes.