Profiles of He 1 4471 predicted by various authors using model atmospheres and recent theory of the Stark broadening of He I are compared with published profiles in B stars and with some newly acquired profiles. The far wings match well in all cases, but the match for the forbidden component [He i] 4469 is poor, the predicted forbidden line being more prominent than is observed. The predicted cores for He 1 4471 are too narrow and too shallow at spectral types earlier than about B5. It seems that the predicted cores are too shallow because the hypothesis of LTE is not valid. The theoretical problem posed by the He I lines is reviewed and it is shown that in the theoretical treatment the use of simplifying assumptions may give a false estimate of the departures from LTE. Some results from an exploration of the He I problem are used to indicate the importance of evaluating correctly the balance between radiative and collision control in the formation of He I lines in B stars. Until these theoretical difficulties are cleared up, use of the equivalent widths of the He I lines to estimate the helium abundance in B type atmospheres may be a dubious procedure.