SIMBAD references

We have performed narrowband imaging of NGC 4631 in the [O III] 5007 and Hα lines, and long-slit spectroscopy using the Low-Resolution Imaging Spectrograph at the W. M. Keck Observatory. Hα and [O III] λ5007 are detected far from the disk, with the highest [O III]/Hα ratios coincident with the soft X-ray emission. Spectroscopy reveals that the ionization is higher than that of the Milky Way diffuse ionized gas (DIG), with an average ratio [O III]/Hβ∼1, rising to 6 in one region. We use the Balmer decrement to probe extinction in the halo. Extinction appears as high as 5 kpc above the disk. Distinct smaller scale variations of [O III], Hβ, and [S II] indicate that we are viewing multiple zones with distinct ionization conditions along the line of sight, in some cases separated by dust. This suggests that the gas has multiple phases and that the ionization may be much higher in some zones. A model using photoionized clouds reproduces the average trends in the line ratios. The inferred densities and pressures are high for halo gas but consistent with observations of other phases that may be colocated with this gas. A combination of high- and low-ionization gas is best for explaining the majority of the fluctuations. The high-ionization phase could be in shocks, but the velocity resolution is not adequate to prove or reject this possibility. Gas at 6x10⁵ K discovered by ROSAT by Wang and colleagues in 1995 could be a source of radiatively cooling gas. We present a new model of optical line emission from cooling clouds and compare it to the data. We show that it can produce part or all of the observed [O III] and fluctuations, may generate part or all of the EDIG gas itself, and may contribute to the photoionization of the upper halo. Our most important conclusions are that the halo is complex and multicomponent and that average line-of-sight line ratios alone cannot be used to determine the ionization conditions.