4.4. Newer results

In the last 5 years, I have been actively working on the helium problem, making use of the excellent spectroscopy of H II galaxies by Terlevich & Melnick (Campbell, Terlevich & Melnick 1986; Terlevich et al. 1991), new observational data with emphasis on obtaining high signal:noise in 6678 (which is the easiest line to interpret but relatively weak) and what seem to be the best data in the literature, rediscussed in a uniform manner (Simonson 1990; Pagel 1991). Furthermore, we have investigated the correlation with nitrogen as well as oxygen since the correlation with N/H seems to be somewhat better (Pagel, Terlevich & Melnick 1986). Fig. 3 shows one of our best spectra, secured with the Anglo-Australian Telescope in 1988, which gives accurate electron density and S⁺ / S⁺⁺ as well as 6678 / H and then in combination with blue-yellow spectrophotometry by Terlevich and his associates gives a very secure measurement of the He/H ratio (Pagel & Simonson 1989). So far we have spectra of this sort for only three objects and more are needed.

Figure 3. Red spectrum of the H II galaxy UM 461 (Terlevich et al. 1991) taken with the Anglo-Australian. Telescope in April 1988 by Pagel, Simonson & Terlevich, with identifications of emission lines. Narrower spikes are cosmic-ray events in the CCD detector. The spectral resolving power is about 2000.

Fig. 4 shows our regression relations of helium with oxygen and nitrogen in low-abundance extragalactic H II regions with maximum-likelihood linear regression lines and error limits equivalent to ±1. The regressions are

(13)

and

(14)

Figure 4. Regressions of helium mass fraction against oxygen and nitrogen abundance, respectively, in irregular and blue compact (or H II) galaxies with oxygen up to 1/4 solar. Maximum-likelihood regression lines are shown with alternatives equivalent to ±1 errors. Sizes of plotted symbols indicate their weights; responding error bars typical of hig-weighted data in the same part of the diagram are shown in just two cases. For a tabulation of the data and their sources, see Pagel (1991).

The regression against oxygen has a remarkably steep slope (corresponding to dY / dZ = 6.5±2) and suggestions of either a flattening off towards higher abundances or significant scatter which is absent (or at least not noticeable compared to errors) in the regression with nitrogen (the latter, however, certainly does not continue linearly beyond the range of the diagram; e.g. Orion has Y 0.27, 10⁷ N/H 700!). The reason for the better correlation with N/H could be local pollution by winds from Wolf-Rayet stars in the embedded cluster which produce additional He and N overlying the basic correlation with oxygen noted by Peimbert and his colleagues, whose conclusions we basically confirm. The pollution hypothesis (Pagel, Terlevich & Melnick 1986; Pagel 1987a, b) is supported by a detailed survey of NGC 5253 by Walsh & Roy (1989), but there are still some ambiguities (Simonson 1990). If the effect is not due to pollution, it could perhaps arise from differential galactic enrichment in oxygen, from short-lived massive stars, on the one hand, and in nitrogen and helium, by planetary nebulae from longer-lived intermediate-mass stars, on the other (cf. Edmunds & Pagel 1978; Steigman, Gallagher & Schramm 1990).