On-stream X-ray fluorescence analyzers have been available for about 20 years. They have been used with great success and they have made an effective process control possible in many concentrators. They have, however, one severe drawback. Normally the sample that ought to be measured is in slurry form and it would be difficult and expensive to prepare it into more conventional form used in ordinary laboratory practice. The main constituent of slurry is water and normally some kind of plastic film is needed to keep slurry apart from the measuring system at least when the slurry system is malfunctioning. Already these two things are sufficient to set a limit to the lightest element that can be measured with reasonable accuracy and precision. This means that minerals that contain only such low atomic number elements as e.g. Mg, Al, Si and P remain out of reach. Difficulties arise also in cases where the interesting element is not only in the solid material but also in the solution part of the slurry. In great many cases the interesting element occurs in several minerals at the same time which causes certain problems with the accuracy. Fortunately this problem does not set such a basic limitation as a too low Z does unless only one of these minerals is of interest, which may make the whole method worthless.