|Annu. Rev. Astron. Astrophys. 1984. 22:
Copyright © 1984 by . All rights reserved
2.1 The Definition of HI Content
``H I content'' is a loosely defined term used to gauge the abundance of the interstellar gas within a galaxy. Strictly speaking, it should be a synonym for the H I mass MH, a parameter that does not reflect the total amount of gas in interstellar space (a large fraction of which is in the form of molecular hydrogen or of a diffuse, ionized component). H I may actually account for only about one half of a galaxy's interstellar gas (and that fraction may be strongly type dependent), but its flux and distribution, which can be determined with relative ease, are frequently used to diagnose traumas in the recent past. Anomalies in these parameters are usually assumed to be symptomatic of comparable anomalies in the total gas content; it should be kept in perspective, however, that atomic, ionized, and molecular components are distributed differently in a galaxy's gravitational potential well, and they are vulnerable in different degrees to the action of external forces.
In the literature, the measure of H I content has been associated with a variety of observational and derived quantities, most frequently the H I mass MH, the hydrogen mass to luminosity ratio MH / L, or the H I surface density MH / D2, where D is the linear diameter. The simple quantity MH appears at first glance to be the most attractive, but as a distance-dependent quantity, it is severely affected by uncertainties in the Hubble constant. Moreover, more massive galaxies also have, on the average, greater H I masses, although the fractional H I masses may not vary. Use of the distance-independent ratios MH / L or MH / D2 has been more popular, but not without controversy.
Historically, the first comparison between field and cluster galaxies was made by Davies & Lewis (28) for galaxies in the Virgo cluster. Their results were immediately questioned (17) as being caused by a luminosity dependence on the ratio MH / L, in the sense that more luminous galaxies, which characterize the cluster sample but not the field, have intrinsically lower MH / L values simply because of their higher luminosities. This residual dependence between MH / L and L has been much debated since the publication of Davies & Lewis' paper (16, 55, 58, 98). As we discuss later, the improved statistics in the Virgo data have removed any doubt of the cluster's relative deficiency, but the Malmquist effect (whereby more distant samples are populated by more luminous galaxies) is a serious problem comparison of galaxies in differing ranges of absolute magnitude.
Further concerns arise in applying the measure of MH / L to the fainter galaxies (m > + 14.0) because of uncertainties in the measured magnitudes. Photoelectric magnitudes for the large number of galaxies being surveyed are unavailable. For the fainter galaxies, the only source of magnitudes is the Catalog of Galaxies and Clusters of Galaxies (CGCG; 119). Conversion to standard photometric systems, such as that of Holmberg (66), for objects for which direct measurements on those standard systems are not available has been a common practice in the literature (12, 30, 84). However, while these conversions may produce reliable results for the larger, brighter galaxies for which there is substantial overlap between the observational sample and the standard sample, the extrapolations necessary for application to smaller, fainter galaxies are very insecure. When properly corrected for systematic biases, especially the anomalies in the first volume, CGCG magnitudes are actually quite reliable, involving random errors on the order of 0.25 mag or less (45).
Another serious problem with the use of MH / L is the necessity for good morphological classification, as this ratio is highly type dependent. As a practical matter, classification is difficult for distant small angular diameter galaxies, especially where large-scale, high dynamic range plate material is not available. Specifically, classification using the Palomar Sky Survey material is very uncertain for galaxies more distant than cz > 6000 or with diameters smaller than one arcmin. A prime example of preliminary misclassification is provided by NGC 5902, once thought to be a gas-rich lenticular but shown later to be an Sb galaxy with a low surface brightness disk (39).
One can correct for the residual dependence of MH / L on luminosity by replacing luminosity as the second variable by the surface magnitude defined as SM = m + 5 log a, where m is the apparent magnitude corrected for galactic extinction and internal absorption, and a is the optical major diameter (47). Because of the way L is folded into SM, the relation between MH / L and SM nearly reduces to one between MH and D alone.
Several authors have noticed the near-independence of the hybrid surface density MH / D2 of spiral disks on morphological type and the constancy of this parameter when averaged over the disk (22, 63, 98). Apparently, the diameter of a spiral disk is a much more important diagnostic parameter for the H I mass than are the luminosity and morphological type. Both MH and D are disk properties, while L refers to the combined contributions of both the disk and the relatively gas-free spheroidal bulge. The increasing importance of the bulge component in earlier morphological types reduces the values of MH / L. A question yet to be answered is whether the morphological dependence still remains if a disk luminosity is used in the computation of the H I mass-to-luminosity ratio.
It is interesting to note the case of the low surface brightness spirals studied by Romanishin et al. (91), a category of relatively isolated objects. When their H I content is judged by the average MH / L ratio applicable for all late spirals, they appear overabundant by more than a factor of 2 (91, 105). If gauged in terms of their sizes, however, they have perfectly normal H I masses when compared with other isolated galaxies.
It should be underscored that MH / D2 is a hybrid quantity, as D is an Optical diameter. For isolated galaxies, Hewitt et al. (63) have shown that the H I extent is well correlated with the optical diameter: with near-independent on morphological type, they find 70% of the H I mass to be confined within 1.2 times the blue diameter as taken from the Uppsala General Catalog of Galaxies (81a). However, MH / D2 is not perfectly invariant. A small residual dependence on D can be detected, in the sense that larger disks have slightly lower MH / D2. The correction of such residuals introduces a distance-scale-dependent bias, which is fortunately very small. The parameter MH / D2 may provide the most sensitive estimate the normalcy of a galaxy's H I content.