6.2. The End of a Conspiracy?

Early studies of DM in spirals showed that rotation curves were remarkably flat and featureless. Well outside the optical radius, the halo must provide virtually all of the rotation velocity, whereas in the inner regions maximum disk solutions indicated that the disk was providing most of the rotation velocity. Even with less massive disks, an apreciable fraction of the inner velocity is atributable to the disk, so the flatness of rotation curves was puzzling. It appeared that the disk and halo were somehow combining to produce a rotation velocity that stayed constant with radius.

Attempts to decompose the disk and halo (and nuclear bulge if appropriate) confirmed the view that the disk and halo were making comparable contributions to the rotation velocity inside the optical radius. Bahcall and Casertano (1985) summarized results for 8 spiral galaxies and found that the ratio of disk mass to halo mass within R₂₅ was close to unity for all of the galaxies studied. Carignan and Freeman (1985) obtained a similar result for 3 galaxies. In view of the flatness of HI rotation curves, this result was not surprising. In order for the disk and halo to both contribute significantly to the overall rotation velocity, while at the same time producing a flat rotation curve, the disk and halo masses within the optical radius are necessarily comparable. This apparent fine-tuning between the disk and halo was termed the ``disk-halo conspiracy'' (e.g. van Albada and Sancisi 1986; Sancisi and van Albada 1987). Such a conspiracy, if genuine, would presumably be telling us something important about galaxy formation and the nature of the DM (e.g. Burstein and Sarazin 1983; Bahcall and Casertano 1985).

Some of the results described in Section 5 above indicate cracks in this conspiracy. For instance, in some dwarf irregulars it seems the halo mass is appreciably greater than that of the disk inside the optical radius. Other evidence has come to light which also weakens this conspiracy. Persic and Salucci (1988) devised a method of using the slope of optical rotation curves to estimate the DM fraction within R₂₅. I return to this and subsequent work below, but note for now that this study suggested an increase in the DM fraction with decreasing spiral galaxy luminosity. Thus, in the inner regions of bright spirals, the disk is the dynamically dominant component, whereas the halo contributes significantly to the rotation curve only at larger radii. In lower luminosity galaxies, the halo is more likely to dominate at all radii.

This suggests that it is principally in the brightest spirals that one expects some kind of feature in the rotation curve where the transition from disk-dominated to halo-dominated dynamics is reached. This motivated Salucci and Frenk (1989) to investigate the magnitude of such an effect. On the basis of Persic and Salucci's (1988) result, they predicted a drop in the rotation curve of a few tens of km s^-1 immediately outside the disk of the brightest spirals, whereas in faint, halo-dominated galaxies they predicted the rotation curve remain flat or possibly rise beyond the optical radius. Salucci and Frenk (1989) found some examples of this effect in published HI rotation curves. Sancisi and van Albada (1987) and Baev, Makov and Fridman (1988) also remarked on this phenomenon.

At about the same time, Casertano and van Gorkom (1991) were obtaining HI rotation curves of a number of galaxies. Their approach was to use relatively short integration times to get crude rotation curves that could be explored further if interesting behavior was uncovered. This allowed a large number of such curves to be obtained. Combining their results with published curves, Casertano and van Gorkom (1991) also identified the drop in the velocities of bright spirals outside the optical radius. Moreover, they found a correlation between the maximum rotational velocity and the slope of the rotation curve beyond the optical radius. (Since spiral galaxy luminosity and velocity are linked through the Tully-Fisher relation, this is equivalent to a correlation between rotation curve shape and luminosity.) As predicted by Salucci and Frenk (1989), the rotation velocities of the brightest galaxies fell beyond he optical radius, whereas fainter galaxies showed flat or rising curves. This is illustrated in Figure 2. It is worth noting that a similar trend was suggested on the basis of optical rotation curves by Rubin et al. (1985). A summary of this earlier work as well as more recent results from optical rotation curves is given by Persic and Salucci (1991b).

Figure 2. The families of rotation curves described by Casertano and van Gorkom (1991). Dwarf spirals typically have rising curves, whereas the rotation curves of bright spirals are often falling and exhibit the ``disk bump.'' Figure courtesy of S. Casertano.

Whether these results herald the end of the disk-halo conspiracy depend to a large extent on how one defines the conspiracy. It is clear that not all rotation curves are flat and that their behavior is easily understood in terms of a systematic variation of DM fraction with galaxy luminosity. This variation has a couple of possible of physical explanations (Dekel and Silk 1986; Ashman 1990a), neither of which are conspiritorial. The fact that the mass of DM and visible material within the optical radius is comparable at all can also be understood in terms of simple models of galaxy formation (Fall and Efstathiou 1980). Some interplay between dissipating gas and the dark halo may be required to explain the featurelessness of spiral galaxy rotation curves (Blumenthal et al. 1986; Athanassoula 1988; Ryden 1991), but in my opinion this does not constitute a conspiracy.