The distribution of fillers, pigments and other materials on the surface of modern papers is of major importance in regards to their ultimate performance. Determining the distributions of these materials however is not possible with standard bulk chemical techniques. One technique which is well suited to this task is digital mapping by WDS (Wavelength Dispersive Spectroscopy) x-ray spectroscopy. By scanning an electron beam at discrete intervals across a sample and recording the intensity of the resultant x-rays, a detailed digital map of element distribution in two dimensions can be constructed. This technique allows rapid simultaneous determination of multiple element distributions with a minimal processing required in most cases. Distributions may be determined with both EDS and WDS, but the greater resolving power and lower detection limits of WDS makes it the essential for determining the distribution of starch¹, one of the major constituents on paper surfaces. This procedure was used to determine the distribution of starch , silicon and calcium in several papers. The papers were prepared by carbon coating and were attached to glass slides using carbon tape. Inorganic constituents such as SiO₂ and CaCO₃ generally produce sufficient x-rays (Ca K1 and Si K1) to be determined directly, with the finished map ready for image processing. Starch (composed mainly of carbon, hydrogen and oxygen), however does not produce sufficient X-rays to allow direct measurement. To allow visualization of starch it must be first complexed with iodine, which produces enough x-rays to allow mapping². A complication in starch determination is the presence of a calcium peak (Ca K1,3 ) near the iodine peak (I L1) which necessitates the removal of the calcium produced background. The simplest method of removing this background is to detune the spectrometer to a position above and below the iodine peak and linearly interpolate the background at the position of the Iodine peak. This procedure triples the amount of time required for map acquisition, but in some cases it may be possible to collect background signals at the same time as peak signals^3,4. Once the element distribution images are acquired, they may be analyzed with image processing software. Examples of the determination of overall coverage are given.