With the growing use of solar energy for heating purposes, an increasing number of solar radiation measurements have to be made. As the measured data is used as a basis for dimensioning solar energy installations, it is of considerable economic importance that the measuring instruments should give reliable data. This report describes an investigation which has had the dual aims of comparing the performance of a number of different makes of pyranometer and of determining a suitable level of delivery inspection and the degree of necessary regular calibration of the instruments. The following makes and types of instrument have been examined: Eppley PSP, Kipp & Zonen CM5, Schenk Star 8101, Lintronic Dome 615, Lambda Li-Cor 200S and Hollis MR5. Linearity, tilt angle sensitivity, temperature dependence, cosine response and azimuth variations of the pyranometers were investigated, together with the effect of variations in solar spectral power distribution on the instruments. The instruments have also been calibrated outdoors and subjected to environmental tests. Delivery inspection should cover all the above characteristics at least for untested designs. Each instrument's levelling arrangements (spirit level) should be checked and, if necessary, adjusted, when checking the azimuth variations. Temperature dependence should always be measured, and correction should be applied if necessary when making the measurements. The instruments should be calibrated outdoors under conditions which are similar to those under which they will be working.