HeI emission in the Orion Nebula and implications for primordial helium abundance

We apply a recently developed theoretical model of helium emission to observations of both the Orion Nebula and a sample of extragalactic H II regions. In the Orion analysis, we eliminate some weak and blended lines and compare theory and observation for our reduced line list. With our best theoretical model we find an average difference between theoretical and observed intensities < I-pred /I-obs - 1 > = 65%. We argue that both the red and blue ends of the spectrum may have been inadequately corrected for reddening. For the 22 highest quality lines, with 3499 angstrom <= k <= 6678 angstrom, our best model predicts observations to an average of 3.8%. We also perform an analysis of the reported observational errors and conclude that they have been underestimated. In the extragalactic analysis, we demonstrate the likelihood of a large systematic error in the reported data and discuss possible causes. This systematic error is at least as large as the errors associated with nearly all attempts to calculate the primordial helium abundance from such observations. Our Orion analysis suggests that the problem does not lie in the theoretical models. We demonstrate a correlation between equivalent width and apparent helium abundance of lines from extragalactic sources that is most likely due to underlying stellar absorption. Finally, we present fits to collisionless case B He I emissivities as well as the relative contributions due to collisional excitations out of the metastable 2s(3)S term.