No. 7, December 2003 THE ING NEWSLETTER

SCIENCE

First Evidence for an Extended Dark Halo in the Dwarf Spheroidal

Jan T. Kleyna, Mark I. Wilkinson, Gerard Gilmore, N. Wyn Evans (IoA)

ver the past several years, we map out the true masses and shapes shape and anisotropy (Wilkinson et al., have been engaged in a project of dSph matter distributions. 2002), it was possible to lift the O to obtain velocities at large degeneracy between Draco’s mass and radii in the Draco and Ursa Minor Draco orbital anisotropy. In these models, the dwarf spheroidal (dSph) galaxies. overall velocity normalisation was fitted Draco and UMi are low-luminosity by the projected central dispersion, the With the commissioning of the (L≈2 ×10 5L ) galaxies about 70 kpc halo shape parameter α could vary O. AUTOFIB2/WYFFOS instrument on from from mass follows light (α =1) to from the . Stellar velocity the WHT, it became possible to obtain constant density (α = –2), and the measurements in the centres of Draco simultaneous spectra of about a and UMi suggest a central mass-to- logarithmic anisotropy parameter ν hundred stars over a one degree field, ν light ratio M/L≈10 2 M . /L . (Aaronson, could be radially anisotropic ( >0) or O O overcoming the problem of Galactic tangentially anisotropic (ν < 0). The 1983; Armandroff, Olszewski, & Pryor, contamination near the outer limits of likelihood contours of Figure 1 shows 1995; Hargreaves et al., 1996). If this the dSphs’ stellar distribution. In four the result of our modelling: Draco is excess mass takes the form of dark nights in June 2000, we were able to matter, then Draco, UMi, and other fit best with an isotropic orbital measure the velocities of 159 Draco distribution in a halo that becomes dSphs with large M/L should be member stars, extending nearly to the approximately isothermal (α ≈0) at excellent laboratories in which to King tidal radius (Kleyna et al., 2001). large radii. Both a mass-follows-light study structure formation and dark From these data, it is apparent that distribution and a completely flat halo matter haloes: low mass galaxies like Draco’s velocity dispersion remains flat density are ruled out at the ~2.5σ the dSphs are probably the basic or even increases with radius, strongly level. The best fit mass, M~8×107M , components from which all larger suggesting the presence of an extended O. is similar to the Jeans estimate. structures form, and an understanding dark halo. An isotropic Jeans equation of the low-mass end of the galaxy mass estimate of the mass contained spectrum provides an important within Draco’s light distribution gives Ursa Minor constraint for evaluating Cold Dark 8 M~10 MO. , with a mean mass-to-light Matter (CDM) and other theoretical ≈ ratio M/L 500 M O. /L O.. Ursa Minor (UMi) resembles Draco models of structure formation. in size, luminosity, and velocity By performing a maximum likelihood dispersion. Unlike Draco, it is elongated In the past, stellar velocity fit of Draco to a family of dynamical and appears to have a second peak measurements have been concentrated models parameterised by the halo along the major axis. Often, this peak in the cores of dSphs, and dynamical modelling has largely been limited to Figure 1. Likelihood contours fitting the central velocity dispersion of the fit of our Draco data to to an isotropic mass-follows-light King the two-parameter α,ν models profile. However, the assumption that of Wilkinson et al. (2002). The mass follows light is known to be contours are at enclosed two- incorrect for virtually all other galaxies, dimensional χ2 probabilities and the assumption of isotropy masks of 0.68, 0.90, 0.95, and 0.997. the crucial degeneracy between The most likely value is anisotropy and mass. Thus, a prime indicated by a plus sign. The objective of this work is to obtain stellar top and right panels of each plot represent the probability velocities at large projected radii within distributions of α and ν, Draco and other northern dSphs. respectively; the median of Combined with modelling methods that each distribution is relax the assumption that mass follows represented by a square, and light and permit varying halo shapes, the triangles show the 1σ, this new data set should allow us to 2σ and (for ν) 3σ limits.

4 THE ING NEWSLETTER No. 7, December 2003

is attributed to tidal disruption, though Figure 2. Result of search for a plausible mechanism for this has kinematic sub-populations in UMi. not been proposed. In May 2002, we Coutours are linearly spaced undertook a 4-night AF2/WYFFOS stellar isopleths; the second peak run to obtain large-radius stellar of UMi’s stellar population is velocities in UMi, with the aim of visible above and to the left of the fitting for the halo shape and orbital centre. Gray stars are UMi red anisotropy using our α , ν models. giant branch member stars with measured velocities. The filled However, 2.5 nights were clouded out, circles represent points where a and poor seeing limited the quality of model with a kinematically cold the data of the remaining 1.5 nights. sub-population is at least 1000 times more likely than a model Though our data was insufficient for composed of a single 8.8 km s–1 detailed modelling, we were Gaussian. The size of each dot is nevertheless able to obtain a number proportional to the logarithm of of velocities in the vicinity of UMi’s the relative likelihood. second density peak. After combining our data set with previously published an anisotropic velocity distribution and References: UMi velocities, we noted that the a dark halo that is isothermal in the velocity histogram of the clump limit of large radii. In Ursa Minor, we Aaronson, M., 1983, ApJ, 266, L11. appeared narrower than the dispersion show that the second peak in the Armandroff, T. E., Olszewski, E. W., of UMi as a whole (Kleyna et al., 2003). stellar density has a cold kinematical Pryor, C., 1995, AJ, 110, 2131. Accordingly, we modelled UMi’s signature. This signature strongly Hargreaves, J. C., Gilmore, G., Irwin, M. velocity distribution as the sum of two suggests that the feature is a persistent J., Carter, D., 1996, MNRAS, 282, 305. Gaussians: a Gaussian subpopulation clump sloshing back and forth within Kleyna, J. T., Wilkinson, M. I., Evans, N. with adjustable normalisation, width a core, and is inconsistent W., Gilmore G., 2001, ApJ, 563, L115. and mean, and an 8.8 km s–1 Gaussian with the cusped halos that are predicted Kleyna, J. T., Wilkinson, M. I., Gilmore representing the bulk of UMi’s stars. by Cold Dark Matter theory. G., Evans, N. W., 2003, ApJ, 588, L21. We then scanned the face of UMi to ¤ Wilkinson M. I., Kleyna, J. T., Evans, N. determine where there was a signature Jan Kleyna ([email protected]) W., Gilmore G., 2002, MNRAS, 330, 778. of a kinematical subpopulation. As suggested by the histograms, only the region near the second clump contained The SAURON Deep Field: Investigating statistically significant ( p=99.45%) evidence of a second kinematical the Diffuse Lyman-α Halo of “Blob1” population (Figure 2). in SSA 22 We note that a dynamically coherent population can survive inside a cored halo, because sinusoidal orbits in the R. G. Bower1, S. L. Morris1, R. Bacon2, R. Wilman1, M. Sullivan1, (nearly) harmonic potential of a core do not diverge over time. However, S. Chapman3, R. L. Davies4, P. T. de Zeeuw 5 kinematical substructure would be quickly smeared out if UMi’s halo had 1: Physics Department, University of Durham. 2: CRAL-Observatoire, Lyon. 3: California a density cusp, as predicted by CDM. Institute of Technology. 4: Dept. of Astrophysics, University of Oxford. 5: Sterrewacht Leiden. Detailed dynamical simulations demonstrate that a cold clump could ecent studies of star-forming spectrograph, we study the formation survive for a Hubble time in a objects in the early , of the most massive galaxies in the 5×107 M UMi-like dSph if the halo measuring their clustering O. R Universe. The primary target is the properties and determining their has a core larger than ~500 pc. If the bright Ly-α emission line halo in the halo has a cusp, however, all evidence luminosity functions, have shown that conspicuous SSA 22 super-cluster at of substructure is erased within several these galaxies are key to understanding hundred million years. the star formation and metal z = 3.07– 3.11 (Steidel et al., 2000). enrichment history of the universe The highly-obscured very luminous and the role of galactic “super-winds” submillimeter galaxy found by Summary in regulating the conversion of baryons SCUBA near the centre of this halo into stars. Using large-radius velocity data probably is an example of a forming obtained using AF2/WYFFOS, we In this article, we describe how, using massive elliptical galaxy (Chapman show that the Draco dSph possesses the SAURON integral field et al., 2001).

5