The general. magnetic field of the Sun has been examined, using the McClean spectrograph of the Cambridge Observatories, crossed with a quartz Lummer plate of excellent quality. which gives a resolving power of about 600,000. This equipment was combined with a polarimetric device, similar to that previously suggested by Wood, consisting of a thick quartz plate and polaroid. Using two carefully selected Fraunhofer lines, one can in this way measure with great accuracy possible small wave-length displacements caused by weak Zeeman effects, the measurements being strictly differential and not affected by possible local Doppler effects. This arrangement was used in the Summer of 1949 to obtain spectrograms from two points on the Sun's central meridian at + 45° latitude. There were about 60 plates with altogether about 1000 interference points to be measured. The m.s.e. for a single determination of the magnetic field strength is about ± 6 gauss, and for each set of two corresponding plates about ± 3 gauss. For each of the two points on the Sun's meridian 15 pairs of plates have been measured. The internal m.s.e. of all these measurements is ± 0 .8 gauss for each hemisphere. The final discussion shows that at the time of the observations (1949 August–September) any possible polar field strength was too small to be measured within a m.s.e. of ± 0.9 gauss. The scatter of the observations, shown in Fig. 3 ( a ) and Fig. 3 ( b ), is within the random error of measurement. A field of 1–2 gauss should have made itself evident beyond all doubt on these diagrams.