Although the influence of magnetic fields on the structure and evolution of stars has, to a great extent, been demonstrated theoretically, observational evidence for this in non-degenerated stars is still rather scarce and is mainly restricted to bright objects (V < 10). Stellar magnetic fields are commonly measured on the basis of circular spectropolarimetry at high/middle resolution across the profile of metal lines. The level of sensitivity of telescopes and spectrographs at present makes this still an almost impossible method for faint stars. In principle, stellar magnetic fields can also be measured on the basis of low-resolution spectropolarimetry, and very important results have been obtained at the 8-m European Southern Observatory telescopes with the Focal Reducer and Low-Dispersion Spectrograph (FORS1). The trade-off between signal-to-noise ratio (S/N) and spectral resolution in measuring stellar magnetic fields justifies an attempt, presented here, to perform these measurements at the 4.5-m William Herschel Telescope. One of the stars with the weakest known magnetic field, HD 3360, and the magnetic chemically peculiar stars, HD 10783, HD 74521 and HD 201601, have been observed with the Intermediate Dispersion Spectrograph and Imaging System (ISIS) in the 3785–4480 Å range. The measured stellar magnetic fields, from Stokes I and V spectra with S/N > 600, show an internal error of ≤50 G when selecting the whole interval and ≤200 G within a Balmer line. Ripples in the Stokes V spectra of HD 3360 result in an instrumental positive magnetic field certainly no larger than 80 G.