1: \documentclass{aastex}
2: \usepackage{emulateapj5}
3:
4: \shorttitle{Stellar kinematics in Leo\,II}
5: \shortauthors{Koch et al.}
6: \slugcomment{Accepted for publication in the AJ, April 25, 2007}
7: \def\kms{\,km\,s$^{-1}$}
8: \def\p1{Paper\,I}
9:
10: \begin{document}
11:
12: \title{Stellar kinematics in the remote Leo\,II dwarf spheroidal galaxy -- \\
13: Another brick in the wall\altaffilmark{1}}
14:
15:
16: \author{Andreas Koch\altaffilmark{2,3},
17: Jan T.~Kleyna\altaffilmark{4},
18: Mark I.~Wilkinson\altaffilmark{5,6},
19: Eva K.~Grebel\altaffilmark{2,7}, \\
20: Gerard F.~Gilmore\altaffilmark{5},
21: N.~Wyn Evans\altaffilmark{5},
22: Rosemary F.~G.~Wyse\altaffilmark{8},
23: Daniel R.~Harbeck\altaffilmark{9}}
24: \email{akoch@astro.ucla.edu}
25:
26: %
27: \altaffiltext{1}{Based on observations collected at the European Southern
28: Observatory at Paranal, Chile; proposal 171.B-0520(A), and on observations
29: made through the Isaac Newton Groups' Wide Field Camera Survey Programme
30: with the Isaac Newton Telescope operated on the island of La Palma by the
31: Isaac Newton Group in the Spanish Observatorio del Roque de los Muchachos
32: of the Instituto de Astrofisica de Canarias.}
33: \altaffiltext{2}{Department of Physics and Astronomy, Astronomical Institute of the University of Basel,
34: Binningen, Switzerland}
35: \altaffiltext{3}{UCLA, Department of Physics and Astronomy, Los Angeles, CA, USA}
36: \altaffiltext{4}{Institute for Astronomy, University of Hawaii, Honolulu, HI, USA}
37: \altaffiltext{5}{Institute of Astronomy, Cambridge University, Cambridge, UK}
38: \altaffiltext{6}{Department of Physics and Astronomy, University of Leicester, Leicester, UK}
39: \altaffiltext{7}{Astronomisches Rechen-Institut, Zentrum f\"ur Astronomie Heidelberg,
40: University of Heidelberg, Heidelberg, Germany}
41: \altaffiltext{8}{The John Hopkins University, Baltimore, MD, USA}
42: \altaffiltext{9}{Department of Astronomy, University of Wisconsin, Madison, WI, USA}
43:
44:
45: \begin{abstract}
46: We present the projected velocity dispersion profile for the remote (d=233\,kpc)
47: Galactic dwarf spheroidal (dSph) galaxy Leo\,II, based on 171 discrete
48: stellar radial velocities that were obtained from medium-resolution
49: spectroscopy using the FLAMES/GIRAFFE spectrograph at the European
50: Southern Observatory, Chile. The dispersion profile of those stars with good
51: membership probabilities is essentially
52: flat with an amplitude of 6.6$\pm$0.7\kms\ over the full radial extent
53: of our data, which probe to the stellar boundary of this galaxy. We
54: find no evidence of any significant apparent rotation or velocity
55: asymmetry which suggests that tidal effects cannot be invoked to
56: explain Leo\,II's properties. From basic mass modeling, employing Jeans'
57: equation, we derive a mass out to the limiting radius of ($2.7\pm
58: 0.5)\times10^7\,M_{\odot}$ and a global mass to light ratio of 27--45 in
59: solar units, depending on the adopted total luminosity. A cored halo
60: profile and a mild amount of tangential velocity anisotropy is
61: found to account well for Leo\,II's observed kinematics, although we
62: cannot exclude the possibility of a cusped halo with radially varying
63: velocity anisotropy.
64: All in all, this galaxy exhibits dark
65: matter properties which appear to be concordant with the other dSph
66: satellites of the Milky Way, namely a halo mass profile which is
67: consistent with a central core and a total mass which is similar to
68: the common mass scale seen in other dSphs.
69: \end{abstract}
70:
71: \keywords{Galaxies: kinematics --- Galaxies: dwarf ---
72: Galaxies: stellar content --- Galaxies: structure ---
73: Galaxies: individual (\objectname{Leo\,II}) --- Local Group}
74:
75: \section{Introduction}
76:
77: In the two decades since the seminal work of Aaronson (1983) on the
78: velocity dispersion of the Draco dwarf spheroidal galaxy (dSph), it
79: has become observationally well-established that all of these lowest
80: luminosity galaxies exhibit large overall velocity dispersions (of the
81: order of 10\kms). Further, their radial velocity dispersion profiles
82: are generally flat to very large radii (Kleyna et al. 2002, 2004; Koch
83: et al. 2007a; Mu\~noz et al. 2006; Walker et al. 2006; though see
84: Wilkinson et al. 2004 for the case of Draco and Ursa Minor). Coupled
85: with their relatively large characteristic radii (of the order of
86: hundreds of parsecs; see Fig.~1 in Gilmore et al. 2007) and low,
87: star-cluster-like luminosities (Mateo 1998), these observations have
88: led to the conclusion that these systems may be dominated by large
89: amounts of dark matter on all spatial scales (see Mateo 1998; Gilmore
90: et al. 2007 for recent reviews). Their estimated mass-to-light (M/L)
91: ratios are often large, with values as high as several hundreds being
92: reported in the literature\footnote{Throughout this paper, we will
93: state all values of M/L in terms of $V$-band luminosity and solar
94: units.}. The recent addition of eight new dSph candidates to the
95: census of Milky Way satellites (Belokurov et al. 2006,2007; Irwin et
96: al. 2007; Zucker et al. 2006a,b) has highlighted the bi-modal nature
97: of the distribution of sizes for low-luminosity stellar systems. All
98: known star clusters exhibit half-light radii smaller than 30pc while
99: no dSph has a stellar core radius smaller than about 120pc (Belokurov et
100: al. 2007; Gilmore et al. 2007). Given the apparent absence of dark
101: matter in star clusters, this bi-modality adds circumstantial weight
102: to the suggestion that dSphs are the smallest stellar systems to
103: contain dynamically significant quantities of dark matter, making them
104: particularly valuable for studies of its physical properties.
105: %
106:
107: On the other hand, the above M/L estimates presuppose that the dSphs
108: are in dynamical equilibrium, an assumption which is directly
109: supported in, for example, the case of the Draco dSph by lack of any
110: significant depth extent along our line of sight (e.g., Klessen et
111: al. 2003). It has been argued, though, that the flat dispersion
112: profiles and high apparent M/L ratios could be the results of tidal
113: sculpting of these galaxies (e.g., Read et al. 2006a, 2006b, Westfall
114: et al. 2006). While there are a few examples of close dSphs whose
115: shape or kinematics appear to be affected by tides (e.g.,
116: Ursa Major\,II [Zucker et al. 2006];
117: Ursa Minor [Palma et al. 2003]), the outer dSph satellites might be
118: expected to be relatively unscathed by tidal perturbations. Galactic
119: tides, if acting, would predominantly unsettle the outermost regions
120: near the stellar boundary of these systems. Moreover, the Draco dSph,
121: though being one of the close Galactic satellites, does not show any
122: indication of structural perturbations (Odenkirchen et al. 2001;
123: Segall et al. 2007). Coupled with the negligible ratio of ordered
124: stellar rotational velocity (as a manifestation of tides; see Section 3
125: below) to the random motion in dSphs, the
126: observations suggest that these galaxies are in fact the most dark
127: matter dominated systems found to date. It is nevertheless valuable to
128: compare the observed properties of dSphs with alternative models, and
129: scenarios in which the dSphs are the remnants of tidally disrupted,
130: dark-matter-free stellar systems have been investigated by a number of
131: authors (e.g. Kroupa 1997; Kuhn \& Miller 1989; Klessen \& Zhao 2002).
132:
133: Initially, information about the dSphs' dark matter content was mainly
134: gleaned from estimates of their central velocity dispersion.
135: Fortunately, in recent years wide field surveys have provided the
136: opportunity to analyze stellar data covering the entire spatial extent
137: of the dSphs and to investigate stellar kinematics even beyond
138: their formal tidal limits as derived from King profile fits to their
139: stellar density distributions (usually some tens of arc minutes; e.g.,
140: Odenkirchen et al. 2001; Kleyna et al. 2003, 2004; Tolstoy et al. 2004; Chapman et al. 2005;
141: Mateo 2005; Mu\~noz et al. 2006; Walker et al. 2006; Wilkinson et
142: al. 2006 and references therein; Koch et al. 2007a). Such studies are
143: essential to reliably distinguish between the dark matter dominated
144: and tidal disturbance scenarios.
145:
146: The mass-density profiles of low-mass galaxies provide valuable
147: information about the spatial distribution of dark matter on small
148: scales. It has been suggested that halo density profiles, in
149: particular the differentiation between cored and cusped halos, can
150: provide an insight into the actual nature of cold dark matter (CDM;
151: Read \& Gilmore 2005; Colafrancesco et al. 2007; Gilmore et
152: al. 2007). Numerical CDM simulations with high spatial resolution
153: predict dark matter halos that exhibit a central density cusp,
154: $\rho(r)\propto r^{-\alpha}$ with $\alpha$=1--1.5 (e.g., Navarro,
155: Frenk \& White 1995; hereafter NFW; {Diemand et al. 2005}).
156: By
157: contrast, observations of the kinematics of massive galaxies are fully
158: consistent with flat density cores (e.g., Gentile et al. 2004;
159: Spekkens et al. 2005), which also appear to account for most of the
160: available data for dSphs (\L okas 2002; Kleyna et al. 2003; Read \&
161: Gilmore 2005; Goerdt et al. 2006; S\'anchez-Salcedo et al. 2006;
162: Strigari et al. 2006; Wilkinson et al. 2006; Koch et al. 2007a, {Wu 2007}). The
163: presence of central cores in the dark matter haloes of dSphs has
164: important implications for the properties of dark matter as it may
165: imply a maximum phase space density for dark matter which can be used
166: to discriminate between competing dark matter candidates (CDM, Warm
167: Dark Matter, self-interacting dark matter, etc.). One should bear in
168: mind, however, that in the case of dSphs such inferences are subject
169: to low-number statistics and uncertainties in the models and do not
170: take account of the dynamical evolution of the dwarf galaxies'
171: progenitor systems (Grebel et al. 2003).
172:
173: In the light of all the above, the remote dSph Leo\,II provides an
174: interesting test bench to investigate the dark matter content of Local
175: Group dSphs. At its present-day Galactocentric distance of 233\,kpc
176: (Bellazzini et al. 2005) it is unlikely to have experienced severe
177: Galactic tides unless it is on a very eccentric orbit that brings it
178: close to the Milky Way. Previous kinematic data for this satellite
179: were assembled in the study of Vogt et al. (1995; hereafter V95) for
180: 31 stars within the core radius at $2.9\arcmin$ (Irwin \&
181: Hatzidimitriou 1995). From these data, V95 found a M/L of 7 and
182: concluded that Leo\,II must be embedded within a massive dark matter halo,
183: although its dark matter properties are not extreme compared to those
184: of other dSphs. On the other hand, Coleman et al. (2007), using photometry
185: from the Sloan Digital Sky Survey,
186: estimate a M/L of the order of 100, but predicate this
187: value on the assumption that Leo\,II is purely dark matter dominated.
188:
189: In this Paper, we shall investigate the dynamics of this particular
190: galaxy by analyzing data for a large number of targets that reach out
191: to the nominal outer optical radius at $\sim9\arcmin$ (Irwin \& Hatzidimitriou
192: 1995; Coleman et al. 2007) thus enabling us to perform a full
193: kinematical analysis of Leo\,II and to derive detailed information on its
194: density distribution, its global M/L ratio and thus to answer the
195: question of how well-behaved or extreme this galaxy is in terms of its
196: dark matter content.
197:
198: This paper is organized as follows: our data, their reduction and the
199: derivation of individual velocities are described in \textsection
200: 2. We investigate the extent of Leo\,II's apparent rotation in
201: \textsection 3 and determine the galaxy's radial velocity dispersion
202: profile in \textsection 4. These results allow us to derive mass and
203: density profiles in \textsection 5, while \textsection 6 is dedicated
204: to the question of whether Leo\,II's density profile shows a cusp or a
205: core, as well as the importance of velocity anisotropy. Finally, we
206: summarize our findings in \textsection 7.
207:
208: \section{Observations and reduction}
209: %
210:
211: In the framework of the ESO Large Programme 171.B-0520(A) (principal
212: investigator: G.~F. Gilmore), which aims at elucidating the kinematic
213: and chemical characteristics of Galactic dSphs, five fields in Leo\,II
214: were observed using the FLAMES multi-object facility at ESO's Very
215: Large Telescope (Pasquini et al. 2002). Within the same programme, we
216: also analysed the Carina dSph, which is described in detail in Koch et
217: al. (2006). In a first paper concerning Leo\,II that made use of these
218: data we presented this galaxy's spectroscopic metallicity and age
219: distributions, which were derived using the near-infrared calcium
220: triplet (CaT) calibration method (Koch et al. 2007b, hereafter \p1).
221: %
222: For details on our observing strategy, the data reduction and
223: calibration techniques, we refer the reader to \p1. In this section we
224: briefly summarize the main steps taken.
225: %
226: \subsection{Target selection and acquisition}
227: %
228: The wide field of view of the FLAMES instrument, with a diameter of
229: 25$\arcmin$, enables one to cover the whole projected area of the
230: Leo\,II dSph with one single pointing, since its nominal tidal radius
231: lies at 8$\farcm$7 (Irwin \& Hatzidimitriou 1995). However, in order
232: to achieve good sampling out to large radii and to yield more
233: information about the stellar content near and beyond the tidal
234: radius, we observed five different (overlapping) fields in several
235: instrument configurations, typically offset by 5$\arcmin$ with respect
236: to each other (see Tables 1 and 2 in \p1).
237:
238: Our targets were drawn from photometry that was obtained by the
239: Cambridge Astronomical Survey Unit\footnote{see
240: http://www.ast.cam.ac.uk/$\sim$wfcsur} (CASU; Irwin \& Lewis 2001) at
241: the 2.5\,m Isaac Newton Telescope (INT) on La Palma, Spain. From this
242: data set we selected red giant candidates with magnitudes ranging from
243: the tip of the red giant branch (RGB) at V$\sim$18.5\,mag down to
244: 2.5\,mag below the RGB tip, reaching as deep as 21\,mag in apparent
245: $V$-band magnitude.
246: %
247: The total selected number of red giant candidates we could thus
248: observe was 200. Their positions on the sky are depicted in Fig.~1.
249:
250: The observations themselves were carried out using the identical
251: strategy and instrumental setups as described in \p1, i.e., we used
252: the FLAMES instrument in combination with the GIRAFFE multifibre
253: spectrograph in ``low-resolution''-mode centered around the
254: near-infrared CaT at 8550\AA. {This set up provides a full wavelength
255: coverage of 8206--9400\AA\ and a resolving power of $R\sim 6500$
256: per resolution element.}
257: Although we aimed at exposing each
258: configuration for 6\,hrs to reach a minimum nominal signal-to-noise
259: (S/N) ratio of 20\,{pixel$^{-1}$} at our spectral resolution of $\sim 6500$, the
260: major part of the nights was hampered by bad sky
261: conditions. Consequently, the median S/N achieved after processing the
262: spectra is 12\,{pixel$^{-1}$, as defined from the variance in continuum bandpasses
263: surrounding CaT region, from which we will determine our velocities in Sect.~2.3.}.
264:
265: \subsection{Data reduction}
266:
267: Details of the reduction process are given in \p1. In summary, we used
268: version 1.09 of the FLAMES data reduction system, girbldrs, and the
269: associated pipeline version 1.05 (Blecha et al.\ 2000). After
270: standard bias correction and flatfielding, the spectra were extracted
271: by summing the pixels along a slit of width 1\,pixel. The final
272: rebinning to the linear wavelength regime was done using Th-Ar
273: calibration spectra taken during the daytime.
274:
275: Sky subtraction was facilitated by the allocation of about 20 fibers
276: per configuration to blank sky.
277: Because the pipeline did not perform reliable sky
278: subtraction at the time the reduction was performed, custom reduction
279: software was written to subtract sky background. First, the sky
280: fibers were median-combined to produce an average sky for each
281: field. Next, the sky was subtracted from each spectrum by modeling the
282: observed spectrum as a second order polynomial plus the average sky,
283: and minimizing the integral over wavelength of the absolute residual.
284: The regions of the spectrum containing the lines of the CaT
285: were excluded from the residual calculation because they are not
286: expected to conform to our simple model of the spectrum. We estimate
287: the
288: accuracy of the final sky subtraction to be on the order of $\sim3\%$,
289: defined as the 1$\sigma$-dispersion of each
290: sky-subtracted spectrum divided by the median of the sky
291: spectrum.
292:
293: Our data set was then completed by the co-addition of the
294: dispersion-corrected and sky-subtracted science frames, weighted by
295: the exposures' individual S/N, and subsequent rectification of the
296: continuum.
297: %
298: \subsection{Radial velocities and Membership estimates}
299: %
300: In order to separate Leo\,II's RGB stars from Galactic foreground
301: stars (Wyse et al. 2006), we determined the individual radial
302: velocities of each target star by means of cross-correlation of the
303: three calcium triplet lines against synthetic Gaussian template
304: spectra using IRAF's\footnote{IRAF is distributed by the National
305: Optical Astronomy Observatories, which are operated by the Association
306: of Universities for Research in Astronomy, Inc., under cooperative
307: agreement with the National Science Foundation.} FXCOR task (see
308: e.g., Kleyna et al. 2004). The templates were synthesized adopting
309: representative equivalent widths of the CaT in red giants. The typical
310: median velocity error achieved in this way is 2.4\,km\,s$^{-1}$.
311: Taking into account the possibility that the errors returned by FXCOR,
312: which are based on the Tonry-Davis R-value (Tonry \& Davis 1979), may
313: in fact be under- or overestimated (e.g., Mateo et al. 1998; Kleyna et
314: al. 2002), we {pursued two different tests of our achieved accuracy.
315: First, we determined
316: a multiplicative constant to the formal velocity
317: uncertainties by deriving } radial velocities from
318: the first two of the Ca lines at $\lambda\lambda$\,8498, 8452\,\AA\
319: and separately from the third line at 8662\,\AA. The constant was
320: then obtained by requiring a reduced $\chi^2$ of unity between these
321: measurements. This procedure resulted in a re-scaling of the FXCOR
322: velocity errors by a factor of 0.6.
323: {A second estimate of our velocity errors will be
324: given in Sect.~2.4 by comparing our individual measurements
325: to the published values from V95.}
326: A plot of velocity errors versus
327: apparent magnitude and {S/N} is shown in Fig.~2.
328:
329: Given Leo\,II's systemic velocity of $\sim76$\kms\ (V95), our velocity
330: data set may contain a number of Galactic foreground stars along the
331: galaxy's line of sight. However, owing to our selection of target
332: stars whose colors and luminosities are consistent with their being
333: members of Leo\,II's red giant branch and due to the much lower Galactic
334: field star density, Leo\,II's velocity peak clearly stands out against
335: the Galactic contribution (see Fig.~3; and Fig.~4 in \p1). From the
336: Besan\c con synthetic Galactic model (Robin et al. 2003) we would
337: expect no more than 4 Galactic foreground stars within the our
338: color-magnitude selection box and with velocities between 40 and
339: 120\kms, thus coinciding with
340: Leo\,II's intrinsic population (see also \p1).
341:
342: An initial fit of a Gaussian velocity peak to the Leo\,II data yields
343: a mean radial velocity and velocity dispersion of 78.7\kms\ and
344: 7.6\kms. Our sample contains 23 apparent radial velocity non-members,
345: which deviate by more than 5\,$\sigma$ from this initial fit and thus
346: were rejected from further analyses. The mean heliocentric velocity
347: and line-of-sight velocity dispersion were finally determined via an
348: iterative error-weighted maximum-likelihood fit assuming a Gaussian
349: velocity distribution (e.g., Kleyna et al. 2002; Koch et al. 2007a).
350: From this we find Leo\,II's mean radial velocity to be
351: (79.1$\,\pm\,$0.6)\kms\ with a global velocity dispersion of
352: (6.6$\,\pm\,$0.7)\kms. These values are in good agreement with the
353: measurements of V95, who determined a mean velocity of
354: (76.0\,$\pm$\,1.3)\kms\ and a central velocity dispersion of
355: (6.7$\,\pm\,$1.1)\,km\,s$^{-1}$ based on 31 high-resolution spectra of
356: stars within the core radius at 2$\farcm$9 (Irwin \& Hatzidimitriou
357: 1995). If one adopts a conservative $2\sigma$-cut in the velocity
358: distribution in order to define membership of Leo\,II, 158 red giants are
359: included in our sample. Relaxing the criterion to a $\pm 3\sigma$-cut
360: yields 171 radial velocity member candidates. Given the low
361: likelihood of interlopers in our sample, we will apply the
362: $3\sigma$-cut throughout the following analyses.
363:
364: One noteworthy feature in the velocity histogram (Fig.~3, right panel)
365: is the occurrence of an excess of stars at $\sim$85\kms, whose
366: velocities are approximately 0.9\,$\sigma$ higher than the sample
367: mean. This peak deviates by 1.8\,$\sigma$ (taken as $\sqrt{N}$) from
368: the best-fit curve. We note, however,
369: that there is no apparent spatial correlation of the targets in this
370: velocity range, nor is there any particular separation in color-magnitude
371: space. In order to further investigate the possibility that
372: we might be seeing an indication of a separate cold feature in our
373: data, we ran an extensive number of Monte Carlo tests, where we
374: generated normally distributed velocity samples centered on the
375: observed mean velocity of Leo\,II assuming the previously derived
376: dispersion, and which were additionally varied by the measurement
377: errors. It turns out that $\sim$30\% of the random data sets exhibit
378: peaks in any one bin that deviate by as much as the observed 1.8\,$\sigma$ so that
379: the observed excess of stars is not significant above the
380: 1\,$\sigma$-level.
381: Therefore we conclude that the potential second peak in Fig.~3 is most
382: likely not a real velocity feature, but may rather be due to
383: statistical fluctuations.
384: %
385: \subsection{Comparison with other data}
386: %
387: Out of the 31 stars observed by V95, 28 coincide with targets in our
388: sample. Since their data are based on high-resolution spectroscopy
389: obtained with the HIRES spectrograph at the Keck telescope, it is
390: worthwhile to employ the stars in common between the two data sets to
391: assess the quality and consistency of our own measurements. Fig.~4
392: shows a comparison of the radial velocities of these 28 stars as
393: measured in this work to those from V95. An error-weighted linear
394: least squares fit to these data yields the relation $v_{\rm
395: \,V95}\,=\,(0.98\,\pm\,0.10)\,v_{\rm
396: \,This\ work}\,+\,(0.5\,\pm\,7.6)$ with a r.m.s. scatter of 4.5\kms.
397: {Although the latter value yields an estimate of the expected
398: uncertainty in the velocities, it is clear that the errors of both studies and
399: the individual S/N ratios have to be considered before one employs a simple re-scaling
400: of our values to match the quoted r.m.s.\footnote{We note that the HIRES data of V95,
401: despite their considerably higher spectral resolution than our data, are also of low S/N,
402: which will contribute a non-negligible fraction to the observed scatter in Fig.~4. }. Hence, we
403: determined scale factors $A$ in three bins of S/N (see right panel Fig.~4) such as
404: to yield an agreement between our mean velocity errors and the scatter in terms of
405: $\chi^2 = \sum \frac{(v_{\rm This\,\,work} - v_{\rm V95})^2}{A^2\,\sigma_{\rm This\,\,work}^2 + \sigma_{\rm V95}^2} $,
406: where the sum is over the stars in common. As a result, we rescaled our final errors by
407: 1.46 (S/N$\le$20), 1.31 ($20\le$S/N$\le 40$) and 1.18 (S/N$>$40). This leads to the uncertainties
408: finally presented in Table~1.}
409:
410:
411: Recently, Bosler et al. (2007) have performed a low-resolution study
412: of the metallicity distribution of Leo\,II. However, their work did not
413: aim at a kinematical analysis of this galaxy and yielded a velocity
414: resolution of 54\kms\ per resolution element (at the region of the
415: CaT) and a median measurement uncertainty of 8.5\kms. Hence, and
416: since these authors did not publish individual error estimates of
417: their measurements, we did not include the 37 stars that our
418: observations have in common with their set of 74 red giants in the
419: comparison in Fig.~4. Nonetheless, an analogous fit to that described
420: above above yields $v_{\rm \,B07}\,=\,(0.93\,\pm\,0.22)\,v_{\rm
421: \,This\ work}\,+\,(12.2\,\pm\,17.5)$ and a r.m.s. scatter of 10.0\kms.
422:
423: In the light of the respective uncertainties associated with each of
424: the published data sets, there is no significant systematic deviation
425: present in our measurements relative to those from V95 and the results
426: agree to within the errors. On the other hand, there is an indication
427: of a systematic zero point offset of the velocities of Bosler et
428: al. (2007). However, this offset is within the uncertainties and can
429: be attributed to the lower resolution of the latter data set.
430: %
431: \section{Tests for apparent rotation}
432: %
433: It is often suggested that the apparent rotation of a system when
434: viewed in projection can be considered as characteristic of tidal
435: perturbation. N-body simulations, (e.g., Oh et al. 1995) indicate that
436: the action of the tidal field of the Galaxy on a dSph leads to
437: velocity gradients in the outer regions of the satellite. When viewed
438: in projection, these dynamical effects will lead to a systematic change
439: in the mean velocity along the dSphs's major axis, thus mimicking
440: rotation (Piatek \& Pryor 1995; Oh et al. 1995; Johnston et al. 1995;
441: Mateo et al. 1998; Read et al. 2006a). Hence, the most efficient way
442: of gathering evidence whether a particular dSph is being affected by
443: tides sufficiently strongly that its kinematics are being modified is
444: to look for any sign of apparent rotation in its outer regions. To
445: date, all but one of the dSphs of the Local Group show no significant
446: velocity gradients. The sole exception is Carina, which exhibits some
447: evidence of a difference in the mean velocity at either end of the
448: major axis, for stars beyond the nominal tidal radius (Mu\~noz et
449: al. 2006).
450:
451: With its present-day Galactocentric distance of (233$\pm$15)\,kpc
452: (Bellazzini et al. 2005), Leo\,II is not expected to have been recently
453: affected by tides, and will only have been perturbed in the past if
454: its orbit brought it into much closer proximity to the Galaxy. This
455: is consistent with its observed regular structure, as pointed out by
456: Coleman et al. (2007). However, since V95 argue that
457: Leo\,II may not have an unusually high dark matter content compared to the
458: inner Galactic satellites, it cannot be entirely ruled out that it may
459: have been tidally influenced during its evolution despite the presence
460: of a (possibly) considerable amount of dark matter.
461:
462: In order to investigate the question of tides, we implemented two
463: tests for a significant indication of rotation in the Leo\,II velocities.
464: First of all, in Fig.~5 we plot the run of radial velocity as a
465: function of major axis distance. In addition, mean radial velocities
466: have been determined in radial bins whose widths were chosen so as to
467: maintain a constant number of stars per bin (lower panel of Fig.~5).
468: If tides had in fact significantly perturbed the outer regions of Leo\,II,
469: one would expect to see a radial gradient in the velocities. In
470: particular, potential tidal distortions of the galaxy's outskirts,
471: e.g., reflected in tidal tails, could lead to distinct east-west
472: asymmetries in the velocity distributions and one would expect
473: excesses of high and low velocity stars (w.r.t. the systemic mean)
474: near the outer boundary on opposite sides of the galaxy.
475: %
476: It is worth noticing that although 13 of our 200 targeted stars lie
477: outside of the King-tidal radius of 8.7$\arcmin$, only two of these
478: lie within 5$\sigma$ of the systemic velocity.
479: %
480: Recent evidence suggests that Leo\,II may in fact be slightly more
481: extended, with a nominal tidal radius of 9.2$\arcmin$ (Coleman et
482: al. 2007), but as these values rely strongly on the concept
483: of fitting a particular functional form to surface photometry, such
484: differences remain insignificant and will not be pursued any further
485: in the present work (see Koch et al. 2007a for a detailed discussion).
486: %
487: In order to ascertain that we have not missed any potential gradient
488: or high- or low-velocity feature in our data due to our selection
489: criteria, we repeated the averaging in each radial bin after applying
490: several cuts, i.e., at 3\,$\sigma$, 5\,$\sigma$, and 10\,$\sigma$ of
491: the sample mean. For each cut, the data were rebinned
492: to assure a constant number of stars per bin.
493: However, as Fig.~5 implies, there is no significant
494: radial gradient discernible in the kinematics of Leo\,II for any of the
495: velocity cuts. An error weighted least squares fit to our velocity
496: data yields a slope of (1.2$\pm$0.9)\kms\,kpc$^{-1}$. Further, those stars with
497: higher and lower velocities than the systemic mean of Leo\,II are equally distributed
498: across the galaxy and there is no apparent indication of any localized
499: kinematic excess.
500:
501: To investigate the question of the extent to which any apparent
502: rotational signal in terms of velocity gradients can be excluded, we
503: proceeded a step further and computed the mean radial velocity
504: difference of stars on either side of bisecting lines passing through
505: each individual target. In the presence of a distinct apparent
506: galactic rotation, such a plot should exhibit a clear sinusoidal
507: signal (e.g., Walker et al. 2006; Koch et al. 2007a). These velocity
508: differences are shown in Fig.~6 versus the position angle of the
509: respective bisectors.
510: %
511: In fact, there is a peak with an amplitude of $\sim$2\kms\ seen in
512: our data, which is of the same order of magnitude as the radial
513: gradient discussed above. Moreover, the maximum of this distribution
514: occurs at a position angle of 16.5$\degr\pm2.4\degr$, where the error
515: was obtained through bootstrap resampling. In the light of the large
516: uncertainty of Irwin \& Hatzidimtriou's (1995) quoted systemic
517: position angle, i.e., (12$\degr\pm10\degr$), we cannot reject a
518: coincidence between the apparent rotational signal in our velocity
519: data and Leo\,II's minor axis.
520: %
521: To assess the actual reality of such a kinematic gradient, we
522: performed 10$^4$ Monte Carlo runs, in which we generated random
523: samples of velocities at the fixed sky positions
524: of our targets. The velocities for these tests were drawn from a
525: normal distribution, assuming Leo\,II's mean radial velocity and velocity
526: dispersion, and additionally allowing for a variation by the observed measurement
527: errors.
528: %
529: By means of these simulations we find that 87\% of the random samples
530: exhibit maximum velocity differences larger than our observed
531: value {(Fig.~6, bottom panel)}. Hence, the apparent rotational signal we detect is only
532: significant at a 13\% (0.17\,$\sigma$) confidence level.
533:
534: As Walker et al. (2006) discuss, apparent rotation can also originate
535: from relative transverse motion of a dSph due to the bulk proper
536: motion of the system. As there are no extant observations of Leo\,II's
537: proper motion, owing to its large distance, we cannot approach the
538: question of whether the relative motion between the Sun and Leo\,II would
539: be sufficient to account for the full amplitude of our observed
540: rotation.
541: %
542: All in all, there is no kinematical evidence of any significant
543: velocity gradients in Leo\,II, either due to rotational support or
544: produced by Galactic tides. It therefore appears that tides most
545: likely have not affected Leo\,II to any significant degree.
546:
547: \section{Radial variation of the dispersion}
548: %
549: In order to derive the radial velocity dispersion profile of Leo\,II we
550: employed a maximum likelihood technique, which is described in more
551: detail in Kleyna et al. (2004) and Koch et al. (2007a). In essence,
552: we radially binned our data such that the same number of stars per bin
553: was maintained. The binsize was defined so as to ensure a
554: sufficiently large number of stars per bin, and we chose to include no
555: fewer than 12 data points in each bin. A Gaussian velocity
556: distribution centered on the single systemic mean velocity is then
557: assumed for each radial bin. This distribution is convolved with the
558: observational errors and also with an additional component, which
559: accounts for the Galactic foreground contamination. The latter is
560: represented by a uniform interloper velocity distribution contributing
561: a fraction $f_{int}$ to each individual bin (see eq. 2 in Koch et
562: al. 2007a).
563: %
564: As the velocities of our non-member stars also appear to be well fit
565: by a power-law distribution with an exponent of $-0.6$, we also
566: considered such an interloper distribution in our determinations of
567: the dispersion profile. Nevertheless, it turned out that the
568: resulting velocity dispersion profiles are indistinguishable in
569: practice and $f_{int}$ is typically compatible with zero in all
570: bins. Finally, the error bounds on the dispersion were calculated by
571: numerically integrating the total probability of the data set and
572: finding the corresponding 68\% confidence intervals. Fig.~7 displays
573: the radial dispersion profile thus obtained.
574:
575: Given the results of Sect.~3, we neglected the effect of a
576: rotationally-induced inflation of the dispersion in our calculations.
577: Another critical issue in the proper dynamical study of a stellar
578: system is the influence of any significant binary population. The
579: presence of any such component in kinematical data leads to an
580: additional, artificial inflation of the observed line of sight
581: velocity dispersion. Furthermore, the high-velocity extension of a
582: binary distribution increases the deviation from a Gaussian so that
583: the assumption of a zero binary fraction yields larger error
584: estimates, which may invalidate the assessment of any radial
585: information derived from the dispersion profiles.
586:
587: However, it has been shown via Monte Carlo simulations in the past
588: that the influence of binary systems in dSphs is in fact negligible,
589: as the purely binary induced dispersion tends to be small compared to
590: the overall large velocity dispersions found in these galaxies
591: (Hargreaves et al. 1996; Olszewski et al. 1996). Furthermore, repeat
592: observations of red giant velocities in the Draco dSph (Kleyna et
593: al. 2002) and in Fornax (Walker et al. 2006) suggest that the impact
594: of binaries on the measured velocity dispersion is negligible. These
595: kinematical studies did not support an overall binary content larger
596: than 40\% in those dSphs. Also, Kleyna et al. (2002) note that the fraction of
597: dynamically significant binaries in the Draco sample amounted to less
598: than 5\%. Likewise, Koch et al. (2007a) showed for the case of the
599: Leo\,I dSph that the assumption of a larger binary fraction does not significantly
600: alter the resulting velocity dispersion profile. Finally, there seems
601: also to be no evidence that binaries would have significantly affected
602: Leo\,II's internal kinematics, as revealed by the Monte Carlo simulations
603: of V95. If we interpret observed velocity differences as an
604: indication of binarity, there is only one object with a difference in
605: velocity between our measurement and that of V95 that is in excess of
606: 2$\sigma$ (taken as the measurement uncertainty). If one was to
607: interpret this in terms of binarity (not accounting for any selection
608: effects in the data), this would also suggest a low binary fraction in
609: Leo\,II. {The lack of repeat measurements of the stars in our data set inhibits
610: an assessment of potential binaries.}
611: %
612: Hence, we will proceed by neglecting the presence of any binary
613: population in our calculations.
614:
615: The resultant velocity dispersion profile (Fig.~7) is essentially flat
616: out to the last data point and at most subject to fluctuations within
617: the measurement errors. The last point, shown at a median distance of
618: 0.5\,kpc in Fig.~7, includes our outermost observed targets at
619: locations of 1.5\,kpc from the galaxy's center. Overall, the profile
620: is well described by a constant value of 6.6\,\kms, which is in
621: agreement with Leo\,II's global value derived in Sect.~2.3.
622: Hence it appears that Leo\,II does in fact show the same behavior as the
623: majority of the dSphs analysed to date, in terms of an essentially
624: flat velocity dispersion profile (Wilkinson et al. 2006 and references
625: therein). This suggests that, similar to the other dSphs, this galaxy
626: is also a dark matter dominated system out to the largest scales. We
627: will turn to the dynamical implications of this finding by deriving
628: Leo\,II's mass profile and mass to light ratio in Section~5.
629: %
630: \subsection{Correlation with metallicities}
631: For 52 of the targets around Leo\,II's velocity peak, metallicities were
632: derived in \p1. From these it was inferred that Leo\,II exhibits a wide
633: range of metallicities, covering at least 1.3\,dex. The derived
634: age-metallicity relation indicates that Leo\,II experienced constant star
635: formation over an extended period. Overall, this galaxy appears to
636: have formed stars from 15\,Gyr until 2\,Gyr ago, although its dominant
637: stellar population is of intermediate age, around 9\,Gyr (Mighell \&
638: Rich 1996).
639: %
640: When split into a metal-poor and a metal-rich population by dividing
641: at the median metallicity, it was shown in \p1 that these stellar
642: components do not differ significantly in their spatial distributions,
643: i.e., there is no apparent metallicity gradient present in
644: Leo\,II. Moreover, the stellar ages do not appear to show spatial trends.
645: It is then worth noticing that the star formation histories derived from
646: deep, but central, pencil beam surveys such as the HST studies of
647: Mighell \& Rich (1996) and Dolphin (2002) are also consistent with our findings
648: in \p1 from a wide-field survey. This again argues in favor of the lack of any
649: considerable population gradient between the center and Leo\,II's outskirts.
650: %
651: In dwarf galaxies with extended star formation histories, the younger
652: and/or more metal-rich populations generally tend to be more centrally
653: concentrated (e.g., Harbeck et al.\ 2003).
654: %
655: Moreover, different
656: stellar populations in dSphs may also have significantly different
657: kinematics, as in the case of Sculptor (Tolstoy et al. 2004) or CVn
658: (Ibata et al. 2006).
659:
660: In our data, there is no apparent trend of velocity with metallicity
661: discernible (top panel of Fig.~8). In the following, we also
662: separated our velocity data according to their previously assigned
663: metallicities. The respective radial velocity histogram is shown in
664: Fig.~8 (left panel). It turns out that the metal poor stars have
665: slightly higher velocities on average, but the difference between them
666: and the metal richer member candidates is marginal. Using the same
667: formalism as outlined in Section~2.3, we find the mean velocity of the
668: metal poor stars to be (79.1$\,\pm\,$1.6)\kms\ versus
669: (76.9$\,\pm\,$1.0)\kms\ for the metal rich component. The overall
670: velocity dispersions for the two sample are indistinguishable and are
671: consistent with that obtained using the entire velocity data set. In
672: addition, the data were divided into three radial bins (see Fig.~8,
673: right panel). In this case, the metal poor (and presumably older; see
674: \p1) component has a higher velocity dispersion in each bin.
675: However, given the low number of stars per bin (of order ten at our
676: chosen binning) and the derived measurement uncertainties, any
677: difference in the dispersion between both metallicity populations is
678: in fact insignificant. Thus we conclude that Leo\,II's stellar
679: populations are kinematically indistinguishable and well-mixed in
680: phase-space, which is fully consistent with the lack of any obvious
681: metallicity or age gradient (Bellazzini et al. 2005; \p1).
682:
683:
684: \section{Mass, light and mass-to-light estimates}
685: In order to define a mass estimator that can be readily implemented
686: we integrated the
687: Jeans equation (Binney \& Tremaine 1998, eqs. 4-54 ff.; Koch et
688: al. 2007a) under the assumption of spherical symmetry and an isotropic
689: velocity distribution. Since our data cover the entire face of the
690: galaxy, this procedure has the advantage of exploiting the full
691: kinematic information that has been gathered in the previous sections
692: as opposed to mass estimates that rely on the single value of a
693: central velocity dispersion (e.g., V95; Coleman et al. 2007 for the
694: case of Leo\,II). Moreover, these methods of ``core-fitting'' make the
695: assumption that the mass distribution follows that of the light
696: (Richstone \& Tremaine 1986; V95), which has been shown to be invalid in
697: the case of the Draco dSph (Kleyna et al. 2002). Nevertheless, a
698: limitation of the simple formalism employed in this section is the
699: neglect of the degeneracy between velocity dispersion and anisotropy
700: (Wilkinson et al. 2002), which persists unless large data sets and/or
701: higher moments of the velocity distribution are considered (e.g.,
702: Binney \& Mamon 1982; \L okas \& Mamon 2003).
703: We will account for this effect in detail in Sect.~6.
704:
705: The surface brightness profile required for the mass determinations
706: was adopted from a fit to the data of Irwin \& Hatzidimitriou (1995). It is
707: well fit by a Plummer profile with a characteristic radius of
708: $2.76\arcmin\pm 0.09\arcmin$ (see Fig.~9), where the error was
709: determined via bootstrap resampling (e.g., Mackey \& Gilmore 2003).
710: This value is in fact very similar to Leo\,II's core radius from a King
711: (1966) profile fit (Irwin \& Hatzidimitriou 1995; Coleman et al. 2007)
712: and use of the latter value yields analogous results for the mass
713: computations.
714: %
715: As implied by the results from the previous sections the dispersion
716: profile was assumed to be radially constant.
717:
718: The resultant mass profile of Leo\,II is shown in Fig.~9. From this we
719: estimate its total mass out to the tidal radius at 8.7$\arcmin$
720: (0.6\,kpc) to be $(2.7\,\pm\,0.5)\times10^7$\,M$_{\odot}$. The
721: quoted errors were obtained from a Monte Carlo simulation of the mass
722: calculation and take into account the 1$\sigma$ confidence interval
723: of the velocity dispersion as well as the
724: uncertainty in the parameters of the light profile.
725:
726: Also shown in Fig.~9 is the density profile that we obtained from the
727: above dynamical considerations. It is characterised by a central
728: density of
729: $\rho_0=(3.4\,\pm\,0.4)\times10^8\,M_{\odot}$\,kpc$^{-3}$, which is
730: in good agreement with the value of
731: $(4.0\,\pm\,1.1)\times10^8\,M_{\odot}$\,kpc$^{-3}$ derived by V95
732: from their analysis of Leo\,II's very center.\footnote{Coleman et
733: al. (2007) yield a lower value of
734: $\rho_0=(1.3\,\pm\,0.4)\times10^8\,M_{\odot}$\,kpc$^{-3}$ under the
735: assumption of a Gaussian mass and light distribution.}
736: %
737: Leo\,II's density distribution resembles a cored profile in the innermost regions,
738: although we note that our innermost target star lies at a distance of
739: 0.5$\arcmin$ (0.03\,kpc) from the center so that any assessment of the
740: ``central'' density has to rely on an extrapolation.
741: The profile reaches a power-law slope of $-1$ at approximately
742: 0.13\,kpc and converges towards a slope of $-2$ for distances larger
743: than 0.5\,kpc.
744:
745: With these mass estimates at hand, we derive the satellite's M/L ratio
746: by adopting its luminosity from the literature. Thence the (M/L) (in
747: solar units) varies from 26.6$^{+5.5}_{-4.5}$ for
748: L$_V$\,=\,(9.9$\pm$0.2)$\times$10$^5$\,L$_{\odot}$ (V95),
749: 35.6$^{+8.8}_{-7.7}$
750: (L$_V$\,=\,(7.4$\pm$2.0)$\times$10$^5$\,L$_{\odot}$; Coleman et
751: al. 2007) to 45.4$^{+11.7}_{-10.4}$ for
752: L$_V$\,=\,(5.8$\pm$1.8)$\times$10$^5$\,L$_{\odot}$ (Mateo 1998)\footnote{For the subsequent
753: discussions of Leo\,II's M/L, e.g., Gilmore et al. (2007), we adopt the latter value using Mateo's (1998)
754: luminosity estimate, since this value is, in turn, based on the measurements of Irwin \& Hatzidimitriou
755: (1995), whose surface brightness profile we employed.}. In all cases,
756: a M/L of this order of magnitude strongly suggests that Leo\,II is
757: governed by a considerable amount of dark matter on all spatial
758: scales.
759: {Mateo et al. (1998) estimate that the {\em stellar} (M/L)$_V$ for a galaxy like Leo\,II with
760: prominent old and intermediate age populations (see also \p1) is of the order of 1--1.4.
761: If, on the other hand, Leo\,II contained only an old, single-age globular cluster-like stellar population,
762: then Mateo et al. (1998) indicate that the M/L ratio of Leo\,II
763: would be higher than the value of the galaxy's actual population by a factor of $\sim$1.3.
764: Clearly, these results and our observations show that the dark matter content of Leo\,II is
765: well above that expected in a purely old stellar population.}
766: From their photometric structural analysis of Leo\,II, Coleman et
767: al. (2007) derive its M/L via two assumptions: for the case of mass
768: follows light they end up with a (M/L)$_V$ of 7, whereas for the other
769: extreme of a pure dark matter dominance, this value may be as high as
770: $\sim$125. The fact that our full kinematical analysis
771: points towards the higher end of these directions underscores the dark
772: matter dominated state of Leo\,II, with at most mild tidal perturbation.
773: {Since the spatial resolution of our data and the resulting method
774: is an improvement over the traditional core-fitting methods, which assume a constant
775: M/L, we show in Fig.~9 (bottom panel) also the theoretical M/L profile over the full range
776: covered by our observations and using
777: the previously derived parameters. Already at Leo\,II's center the M/L is an order of magnitude higher than
778: the purely stellar estimates.}
779:
780: Given the above results, Leo\,II fits well remarkably well on the (M/L)
781: versus $M_V$ relation for dSphs (see Fig.~5 in Gilmore et al. 2007),
782: which is an oft-used argument in favor of all dSphs being embedded in
783: dark halos of the same mass. Although the exact value of such a common
784: mass scale is subject to various assumptions such as the M/L of the
785: pure stellar component, this underlying halo tends to be of the order
786: of 3$\times$10$^7$\,M$_{\odot}$ (Mateo et al. 1993; Wilkinson et
787: al. 2006; Gilmore et al. 2007).
788:
789: \section{Velocity Anisotropy -- cusp versus core}
790:
791: All the mass computations in Sect.~5 relied on the assumption of an
792: isotropic velocity tensor, i.e., the anisotropy parameter
793: $\beta=1-{\left<v^2_{\theta}\right>}/{\left<v^2_r\right>}$ was assumed
794: to be zero. Nonetheless, there is compelling evidence that the shapes
795: and kinematics of most dSphs necessitate non-negligible amounts of
796: velocity anisotropy on all spatial scales (e.g., \L okas 2002; Kleyna
797: et al. 2002; Wilkinson et al. 2004; Klimentowski et al. 2007). As the
798: neglect of a non-isotropic velocity distribution can considerably
799: affect estimates of the underlying dark halo mass, we explore in the
800: following the possible parameter space of Leo\,II's velocity anisotropy.
801: In doing this we follow the same formalism as outlined in Koch et
802: al. (2007a) (see also \L okas 2002). That is, under the assumption of
803: a (constant) anisotropy parameter and for different plausible mass
804: profiles, we numerically determined a theoretical velocity dispersion
805: profile; the model parameters were then optimized in a least-squares
806: sense to yield the best representation of the observations from
807: Sect.~4.
808:
809: \subsection{Density cusps}
810: Motivated by the cosmological simulations of Navarro, Frenk \& White
811: (1995; NFW), the density profiles of dark matter halos are often
812: described by a convenient function with only one free parameter, the
813: ``characteristic density''. For simulated haloes, this
814: parameterisation has proven applicable to a wide range of halo masses
815: from the smallest scales, such as the dSphs' dark matter associations,
816: up to the large scale structures of galaxy clusters. We first adopted
817: this cuspy halo mass model,
818: %
819: \begin{equation}
820: M(r)\,\propto\,M_v\,\left\{\ln (1+cr/r_v)\,-\,\frac{cr/r_v}{1+cr/r_v}\right\},
821: \end{equation}
822: %
823: in our computations of an empirical dispersion profile.
824: In our representation, the only free parameter is the virial halo mass, $M_v $,
825: and all other parameters, namely virial radius $r_v$ and concentration $c$, are
826: empirically scaled with this mass according to the relations obtained
827: by Jing \& Soto (2000) (see \L okas \& Mamon 2001; eqs. 9--12 in Koch et
828: al. 2007a for details of the parameterisations).
829:
830: The best-fit results for varying degrees of anisotropy are overlaid on
831: the observed dispersion profiles in Fig.~10. For all choices of
832: $\beta$, the virial mass of the NFW-halo is of the order
833: 2.3$\times$10$^8$M$_{\odot}$, corresponding to a formal virial radius
834: of about 12\,kpc (which is much
835: larger than the stellar limiting radius, allowing for the factor of $\sim$10 in mass out to the
836: virial radius). As for the anisotropy parameter, the reduced
837: $\chi^2$ for a $\beta$ of ($-1$, $-$0.5, 0, 0.5) amounts to (0.29,
838: 0.21, 0.22, 0.47). The best coincidence of prediction and observation
839: is obtained when a slight amount of tangential anisotropy
840: ($\beta\sim-0.25$) is included.
841:
842: \subsection{Cored density profile}
843: The majority of presently available observational data of
844: low-luminosity galaxies indicates that their kinematics and structures
845: are consistent with cored density profiles (e.g., \L okas 2002; Kleyna
846: et al. 2003; Salucci et al. 2003; S\'anchez-Salcedo et al. 2006).
847: %
848: Hence, we also computed dispersion profiles in the same manner as
849: above for a halo profile with variable inner logarithmic slope
850: (Hernquist 1990; Read \& Gilmore 2005; eq. 13 in Koch et al. 2007a):
851: \begin{equation}
852: \rho(r)\,\propto\,C\,\left(\frac{r}{r_s}\right)^{-\gamma}\,\left[1+\left(\frac{r}{r_s}\right)^{\alpha}\,
853: \right]^{(\gamma-3)/\alpha}
854: \end{equation}
855: The parameters were then determined such as to match the observed dispersion
856: profile for different values of the anisotropy $\beta$.
857:
858: As it turns out, the ``best-fit'' for all choices of anisotropy was obtained for a slope parameter
859: of $\gamma\simeq0$, which corresponds to a pure core, as opposed to
860: the cuspy halo class ($\gamma=1$). Moreover, a scale radius $r_s$ of
861: 0.3\,kpc and a large smoothness parameter $\alpha$ (which determines
862: the sharpness of the transition toward the density profile at large
863: radii) of $\sim$3.7 were found to yield the best representation of our
864: data (see lower panels of Fig.~10). For this cored model, the values
865: of the reduced $\chi^2$ are (0.21, 0.20, 0.21, 0.31) for constant
866: anisotropies of ($-1$, $-$0.5, 0, 0.5) and, as for the cuspy halos,
867: our data ideally require some tangential anisotropy of the same order
868: of magnitude, that is $\beta\sim-0.25$.
869:
870: It is worth noticing that our data do not allow us to unequivocally
871: constrain the shape of the velocity anisotropy tensor at large radii
872: for either choice of halo profile. However, we can conclude from
873: Fig.~10 that strongly radial anisotropy is not suitable to account for
874: the dispersion profile towards the central regions.
875:
876: \subsection{Cusp or Core ?}
877: For both profiles investigated above, the mass at the galaxy's
878: limiting radius amounts to $\sim2.0\times10^7$M$_{\odot}$ and is
879: thus slightly smaller than that derived from a purely isotropic
880: velocity distribution in Sect.~5\footnote{The assumption
881: of velocity isotropy would generally lead to
882: an overestimate of the mass at large radii, if Leo\,II were indeed tangentially
883: anisotropic, whereas if it were radially anisotropic in the outer parts,
884: we would underestimate the mass assuming isotropy.}.
885:
886: Judging by the values of the $\chi^2$ statistics, a cored density
887: profile is favored to account for the Leo\,II observed kinematics; however,
888: the difference between this and the cuspy model is small. In
889: particular, there is an indication of a leveling off of the
890: predictions for the dispersion towards larger radii for the cored model, a trend that
891: merits further investigation from a large sample of stars at and
892: beyond the limiting radius of this galaxy.
893:
894: In addition, we re-ran the computations with a velocity anisotropy
895: which varies radially according to the prescription of Osipkov (1979)
896: and Merritt (1985), i.e., $\beta(r)=r^2/(r^2+r_a^2)$, which yields radial
897: anisotropy for the case of $r>r_a$.
898: The resulting
899: curves are shown in the right panels of Fig.~10. While the NFW halo
900: requires an anisotropy radius $r_a$ of $\sim$0.7\,kpc and thus larger
901: than Leo\,II's tidal radius, the cored profile is best associated with a
902: $r_a$ of $\sim$0.2\,kpc. The respective $\chi^2$ statistics are then
903: 0.22 and 0.19, thus again marginally favoring a cored halo\footnote{We note
904: that a radially varying tangential anisotropy $\beta(r)=r^2/(r^2-r_a^2)$ (Merritt 1985;
905: his model II) yields essentially the same results.}
906:
907: As a comparison with extant data of the density profiles of Local
908: Group dSphs indicates, the majority of these systems are consistent
909: with cored inner halo profiles, which is similar to our findings for
910: Leo\,II (Wilkinson et al. 2006; Gilmore et al. 2007 -- see their
911: Fig.~4). While the outer density profile of Leo\,II is similar to those of
912: the other LG dSphs analysed to date,
913: Leo\,II apparently exhibits a mildly steeper central density profile than
914: the other dwarfs, due to the apparently smaller size of its central
915: core.
916:
917: From their photometric analysis Coleman et al. (2007) argue that Leo\,II's
918: M/L ratio has an upper limit of $125^{+56}_{-51}$ under the assumption
919: of a constant density core out to their tidal radius. Although our value derived in the previous
920: Section is lower by a factor of approximately 3, we note that the core
921: size we obtain in the density profile is correspondingly smaller.
922:
923: \section{Summary}
924: %
925: We have obtained a large data set of radial velocity measurements for
926: about 171 red giant member candidates in the remote Leo\,II dSph, thereby
927: increasing existing published samples more than fivefold (V95). These
928: data extend in radial distance out to the galaxy's limiting radius,
929: which is an important prerequisite for kinematic studies, since Leo\,II,
930: as the second most remote dSph satellite of the Milky Way, is valuable
931: for an assessment of the potential role of Galactic tides. With its
932: moderate present-day systemic velocity, Leo\,II can be considered as bound
933: to the Galaxy and yet there is no indication of any tidal
934: perturbation. This lack of perturbations is expected if this dSph has
935: maintained its present large distance over long time-scales. Our
936: study precludes any apparent rotation of Leo\,II's stellar component;
937: moreover we do not detect a velocity gradient, nor any asymmetry in
938: the sense that high and low velocity outliers preferentially reside in
939: particular directions along the major axis, which might reflect the
940: action of tides. Although other dSphs may well exhibit ``extratidal''
941: material (e.g., Carina; Mu\~noz et al. 2006) or be in an advanced
942: state of dissolution (e.g., Sagittarius; Ibata et al. 1994), these
943: effects are clearly not seen in Leo\,II.
944: %
945: Thus we must conclude that this galaxy is a purely pressure-supported
946: system which has not been significantly affected by tides over the
947: course of its evolution.
948:
949: The radial velocity dispersion profile and resulting density profiles
950: show a resemblance to the respective profiles of the majority of the
951: other Local Group dSphs, and display the well-established features of
952: these systems, namely a dispersion which is flat out to the reach of
953: the present data and a density profile that is consistent with a cored
954: halo mass distribution. We note, however, that our present data do not
955: unambiguously distinguish between the possibilities of a cusped or
956: cored density profile in terms of the $\chi^2$-statistics of the
957: best-fit to our observed dispersion profile. Both cored and cusped
958: halo models favor the presence of a mild amount of tangential
959: velocity anisotropy in the central regions of the galaxy.
960:
961: All in all, we find that Leo\,II is a typical dSph in terms of the values
962: of its total mass of $\sim$3$\times$10$^7$\,M$_{\odot}$ and of its
963: characteristic central density of the order of
964: $3.4\times10^8\,M_{\odot}$\,kpc$^{-3}$ (Gilmore et al. 2007).
965: Depending on the adopted total luminosity of Leo\,II, we estimate its M/L
966: ratio to lie in the range 25 to 50. This is, in conjunction with the
967: flatness of the dispersion profile, striking evidence that Leo\,II is a
968: dark matter dominated system. Moreover, the value we find is
969: consistent with the idea that all dSphs are embedded in a dark matter
970: halo with one single characteristic mass scale. Taking into account
971: that all nearby dSph galaxies contain very old stellar populations
972: (Grebel \& Gallagher 2004), it thus appears reasonable to consider
973: these objects as the smallest surviving condensations of dark matter
974: with associated stellar populations in the universe (Read et
975: al. 2006b).
976:
977:
978: \acknowledgments
979:
980: A.K. and E.K.G. are grateful for support by the Swiss National Science
981: Foundation through grants 200020-105260 and 200020-113697.
982: M.I.W. acknowledges the Particle Physics and Astronomy Research
983: Council and the Royal Society for financial support.
984:
985: \begin{thebibliography}{}
986: %
987: \bibitem[aaronson83]{a1} Aaronson, M. 1983, ApJ, 266, L11
988: %
989: \bibitem[Bellazzini et al.(2005)]{Bellazzini05} Bellazzini, M.,
990: Gennari, N., \& Ferraro, F.~R.\ 2005, \mnras, 360, 185
991: %
992: \bibitem[Belokurov et al.(2006)]{belokurov06} Belokurov, V., et al., 2006, ApJ, 647, L111
993: %
994: \bibitem[Belokurov et al.(2007)]{belokurov07} Belokurov, V., et
995: al., 2007, \apj, 654, 897
996: %
997: \bibitem[Binney \& Mamon 1982]{bm82} Binney, J., Mamon, G.~A. 1982, MNRAS, 200, 361
998: %
999: \bibitem[Binney1998]{b5} Binney, J., \& Tremaine S. 1998, ``Galactic
1000: Dynamics'', Princeton University Press
1001: %
1002: \bibitem{blecha00} Blecha, A., Cayatte, V., North, P., Royer, F., \& Simond, G. 2000, Data-reduction software for GIRAFFE, the VLT medium-resolution
1003: multi-object fiber-fed spectrograph, in Optical and IR Telescope Instrumentation and Detectors, ed. I.Masanori, \& A.F.Moorwood, SPIE, 4008, 467
1004: %
1005: \bibitem[Bosler2007]{b7} Bosler, T.~L., Smecker-Hane, T.~A., \& Stetson, P.~B. 2007, MNRAS,
1006: submitted (astro-ph/0608197)
1007: %
1008: \bibitem[chapman]{c4} Chapman, S.~C., Ibata, R., Lewis, G.~F., Ferguson, A.~M.~N., Irwin, M.,
1009: McConnachie, A., \& Tanvir, N. 2005, ApJ, 632, L87
1010: %
1011: \bibitem[colafrancesco 2007]{colafr07} Colafrancesco, S., Profumo, S., \& Ullio, P. 2007,
1012: Phys. Rev. D, 75, 023513
1013: %
1014: \bibitem[Coleman et al. 2007]{coleman07} Coleman, M.~C., Jordi, K., Koch, A., Grebel, E.~K. \&
1015: Rix, H.-W., 2007, AJ, submitted
1016: %
1017: \bibitem[Diemand et al.(2005)]{2005MNRAS.364..665D} Diemand, J., Zemp, M.,
1018: Moore, B., Stadel, J., \& Carollo, C.~M.\ 2005, \mnras, 364, 665
1019: %
1020: \bibitem[dolphin 2002]{dolphin2002} Dolphin, A.~E. 2002, MNRAS, 332, 91
1021: %
1022: \bibitem[gentile et al. 2004]{gentile04} Gentile, G., Salucci, P., Klein, U., Vergani, D., \& Kalberla, P.
1023: 2004, MNRAS, 351, 903
1024: %
1025: \bibitem[Gilmore et al. 2007]{gilmore07} Gilmore, G., Wilkinson, M.~I., Kleyna, J.~T., Wyse, R.~F.~G.,
1026: Koch, A., Evans, N.~W., \& Grebel, E.~K. 2007, ApJ, in press (astro-ph/0703308)
1027: %
1028: \bibitem[Goerdt06]{g6} Goerdt, T., Moore, B., Read, J.~I., Stadel, J., \& Zemp, M. 2006, MNRAS, 368,
1029: 1073
1030: %
1031: \bibitem[grebel03]{g5}Grebel, E.~K., Gallagher, J.~S., \& Harbeck, D.~R. 2003, AJ, 125, 1926
1032: %
1033: \bibitem[grebel 2004]{grebel04} Grebel, E.~K., \& Gallagher, J.~S., ApJ, 610, L89
1034: %
1035: \bibitem[hargreaves1996]{h2} Hargreaves, J.~C., Gilmore, G., \& Annan, J.~D. 1996, MNRAS, 279, 108
1036: %
1037: \bibitem[hernquist]{h5} Hernquist, L. 1990, ApJ, 356, 359
1038: %
1039: \bibitem[ibata et al. 1994]{ibata1994} Ibata, R.~A., Gilmore,
1040: G., \& Irwin, M.~J.\ 1994, Nature, 370, 194
1041: %
1042: \bibitem[Ibata et al. 2006]{ibata06} Ibata, R.~Chapman, S., Irwin, M., Lewis, G., \&
1043: Martin, N. 2006, MNRAS, 373, L70
1044: %
1045: \bibitem[Irwin1995]{i2} Irwin, M. \& Hatzidimitriou, D. 1995, MNRAS, 277, 1254
1046: %
1047: \bibitem{irwin01} Irwin, M., \& Lewis, J. 2001, New Ast. Rev., 45, 105
1048: %
1049: \bibitem{irwin07} Irwin, M., et al., 2007, \apj, 656, L13
1050: %
1051: \bibitem[jing00]{j1} Jing, Y.~P., \& Suto, Y. 2000, ApJ, 529, L69
1052: %
1053: \bibitem[johnston95]{j2} Johnston, K.~V., Spergel, D.~N., \& Hernquist, L. 1995, ApJ, 451, 598
1054: %
1055: \bibitem[king66]{k1} King, I.~R. 1966, AJ, 71, 64
1056: %
1057: \bibitem[klessen02]{k3} Klessen, R.~S., \& Zhao, H. 2002, ApJ, 566, 838
1058: %
1059: \bibitem[klessen03]{k4} Klessen, R.~S., Grebel, E.~K., \& Harbeck, D.~R. 2003, ApJ, 589, 798
1060: %
1061: \bibitem[kleyna2002]{kleyna2002} Kleyna, J.~T., Wilkinson, M.~I., Evans, N.~W., Gilmore, G.F., \&
1062: Frayn, C. 2002, MNRAS, 330, 792
1063: %
1064: \bibitem[kleyna2003]{k7} Kleyna, J.~T., Wilkinson, M.~I., Gilmore, G., \& Evans, N.~W. 2003,
1065: ApJ, 588, L21
1066: %
1067: \bibitem[Kleyna2004]{k8} Kleyna, J.~T., Wilkinson, M.~I., Evans, N.~W. \& Gilmore, G.~F.
1068: 2004, MNRAS, 354, L66
1069: %
1070: \bibitem[Klimentowski et al. 2007]{klimentowski07} Klimentowski, J., \L okas, E.~L., Kazantzidis, S.,
1071: Prada, F., Mayer, L., \& Mamon, G.~A. 2007, MNRAS, subm. (astro-ph/0611296)
1072: %
1073: \bibitem{koch06} Koch, A., Grebel, E.~K., Wyse, R.~F.~G., Kleyna, J.~T., Wilkinson, M.~I., Harbeck, D.~R., Gilmore, G.~F., \& Evans, N.~W.
1074: 2006, AJ, 131, 895
1075: %
1076: \bibitem[Koch et al.(2007a)]{Koch07a} Koch, A., Wilkinson, M.I., Kleyna, J.T., Gilmore, G.F., Grebel, E.K., Mackey, D., \&
1077: Evans, N.W., \& Wyse, R.F.G. 2007a, ApJ, 657, 241
1078: %
1079: \bibitem{koch07b} Koch, A., Grebel, E.~K., Kleyna, J.~T., Wilkinson, M.~I., Harbeck, D.~R., Gilmore, D.~R., Wyse, R.~F.~G., \& Evans, N.~W.
1080: 2007b, AJ, 133, 270; (\p1)
1081: %
1082: \bibitem[Kroupa(1997)]{Kr97} Kroupa, P., 1997, New Astronomy, 2, 139
1083: %
1084: \bibitem[Kuhn \& Miller 1989]{kuhn89} Kuhn, J.~R., Miller, R.~H., 1989, ApJ, 341, L41
1085: %
1086: \bibitem[lokasmamon01]{l3} \L okas, E.~L., \& Mamon, G.~A. 2001, MNRAS, 321, 155
1087: %
1088: \bibitem[lokas02]{l5} \L okas, E.~L. 2002, MNRAS, 333, 697
1089: %
1090: \bibitem[lokas03]{l6} \L okas, E.~L., \& Mamon, G.~A. 2003, MNRAS, 343, 401
1091: %
1092: \bibitem[Mateo et al. 1993]{Mateo93} Mateo, M., Olszewski, E.~W., Pryor, C., Welch, D.~L., \&
1093: Fisher, P. 1993, AJ, 105, 510
1094: %
1095: \bibitem[Mackey \& Gilmore 2003]{Mackey03} Mackey, D.~A., \& Gilmore, G.~F. 2003, MNRAS, 338, 85
1096: %
1097: \bibitem[Mateo2005]{m7} Mateo, M. 2005, in IAU Colloquium 198, Near-field cosmology with dwarf elliptical galaxies, ed.
1098: H.~Jerjen \& B.~Binggeli (Cambridge: CUP), 52
1099: %
1100: \bibitem[mateo1998]{m5} Mateo, M. 1998, AR\&A, 36, 435
1101: %
1102: \bibitem[Mateo1998]{m6} Mateo, M., Olszewski, E.~W., Vogt, S.~S., \& Keane,
1103: M.~J. 1998, AJ, 116, 2315
1104: %
1105: \bibitem{m9} Merrit, D. 1985, AJ, 90, 1027
1106: %
1107: \bibitem[Mighell \& Rich(1996)]{Mighell96} Mighell, K.~J., \&
1108: Rich, R.~M.\ 1996, \aj, 111, 777
1109: %
1110: \bibitem[munoz06]{munoz06} Mu\~noz, R.~R., et al. 2006, ApJ, 649, 201
1111: %
1112: \bibitem[Navarro95]{n2} Navarro, J.~F., Frenk, C.~S. \& White, S.~D.~M. 1995, MNRAS, 275, 720
1113: (NFW)
1114: %
1115: \bibitem[odenkirchen01]{o1} Odenkirchen, M., et al. 2001, AJ, 122, 2538
1116: %
1117: \bibitem[oh95]{o2} Oh, K.~S., Lin, D.~N.~C., \& Aarseth, S.~J. 1995, ApJ, 442, 142
1118: %
1119: \bibitem[Olszewski1996]{o3} Olszewski, E.~W., Pryor, C., Armandroff, T.~E. 1996, AJ, 111, 750
1120: %
1121: \bibitem{o3.5} Osipkov, L.~P. 1979, Pis'ma Astron. Zh., 5, 77
1122: %
1123: \bibitem[Palma 2003]{palma03} Palma, C., Majewski, S.~R., Siegel, M.~H., Patterson, R.~J.,
1124: Ostheimer, J.~C. \& Link, R. 2003, AJ, 125, 1352
1125: %
1126: \bibitem[Pasquini et al.(2002)]{Pasquini02} Pasquini, L., et al.\
1127: 2002, The ESO Messenger, 110, 1
1128: %
1129: \bibitem[piatek95]{P3} Piatek, S., \& Pryor, C. 1995, AJ, 109, 1071
1130: %
1131: \bibitem[read2005]{r0} Read, J.~I., \& Gilmore, G. 2005, MNRAS, 356, 107
1132: %
1133: \bibitem[read200a6]{r1} Read, J.~I., Wilkinson, M.~I., Evans, N.~W., Gilmore, G., Kleyna, J.~T. 2006a,
1134: MNRAS, 367, 387
1135: %
1136: \bibitem[read2006b]{read2} Read, J.~I., Wilkinson, M.~I., Evans, N.~W., Gilmore, G., Kleyna, J.~T. 2006b,
1137: MNRAS, 366, 429
1138: %
1139: \bibitem[Richstone1986]{r2} Richstone, D.~O., \& Tremaine, S. 1986, AJ, 92, 72
1140: %
1141: \bibitem{robin03} Robin, A.~C., Reyl\'e, S., Derri\`ere, S., \& Picaud, S. 2003, \aap,
1142: 409, 523
1143: %
1144: \bibitem[salucci et al. 2003]{salucci03} Salucci, P., Walter, F., \& Borriello, A. 2003, A\&A, 409, 53
1145: %
1146: \bibitem[sanchez-salcedo]{sanch} S\'anchez-Salcedo, F.~J., Reye-Iturbide, J., \& Hernandez, X.
1147: 2006, MNRAS, 370, 1829
1148: %
1149: \bibitem[segall 2007]{segall} Segall, M., Ibata, R., Irwin, M., Martin, N., \& Chapman, S. 2007,
1150: MNRAS, 375, 831
1151: %
1152: \bibitem[spekkens05]{s2} Spekkens, K., Giovanelli, R., \& Haynes, M.~P. 2005, AJ, 129, 2119
1153: %
1154: \bibitem[Strigari et al.(2006)]{Strigari06} Strigari, L.~E.,
1155: Bullock, J.~S., Kaplinghat, M., Kravtsov, A.~V., Gnedin, O.~Y.,
1156: Abazajian, K., \& Klypin, A.~A.\ 2006, \apj, 652, 306
1157: %
1158: \bibitem[Tolstoy et al. 2004]{tolstoy04} Tolstoy, E., et al. 2004, ApJ, 617, L119
1159: %
1160: \bibitem[tonry]{t0} Tonry, J.~L., \& Davis, M. 1979, \aj, 84, 1511
1161: %
1162: \bibitem[Vogt1995]{v1}Vogt, S.~S., Mateo, M., Olszewski, E.~W., \& Keane, M.~J.
1163: 1995, AJ, 109, 151 (V95)
1164: %
1165: \bibitem[Walker2006]{w1}Walker, M.~G., Mateo., M., Olszewski, E.~W.,
1166: Bernstein, R.~A., Wang, X., \& Woodroofe, M. 2006, AJ, 131, 2114
1167: %
1168: \bibitem[westfall06]{w5} Westfall, K.~B., Majewski, S.~R., Ostheimer, J.~C., Frinchaboy, P.~M.,
1169: Kunkel, W.~E., Patterson, R.~J., \& Link, R. 2006, AJ, 131, 375
1170: %
1171: \bibitem[Wilkinson et al.(2002)]{wilkinson02} Wilkinson, M.~I., Kleyna, J., Evans, N.~W., Gilmore, G., 2002, MNRAS, 330, 778
1172: %
1173: \bibitem[Wilkinson2004]{w8}Wilkinson, M.~I, Kleyna, J.~T., Evans, N.~W., Gilmore, G., Irwin, M.~J., \&
1174: Grebel, E.~K. 2004, ApJ, 611, L21
1175: %
1176: \bibitem[Wilkinson et al.(2006)]{Wilkinson06a} Wilkinson, M.~I.,
1177: Kleyna, J.~T., Evans, N.~W., Gilmore, G.~F., Read, J.~I., Koch, A.,
1178: Grebel, E.~K., \& Irwin, M.~J.\ 2006, in EAS Publications Series 20,
1179: Mass Profiles and Shapes of
1180: Cosmological Structures, ed.\ G.A. Mamon,
1181: F.\ Combes, C.\ Deffayet, \& B.\ Fort (Paris: EDP Sciences), 105
1182: %
1183: \bibitem[Wu07]{wu07} Wu, X. 2007, AJ, subm. (astro-ph/0702233)
1184: %
1185: \bibitem[Wyse06]{wyse06} Wyse, R.F.G., Gilmore, G.F., Norris, J.E., Wilkinson, M.I., Kleyna, J.T.,
1186: Koch, A., Evans, N.W., \&
1187: Grebel, E.K. 2006, ApJ, 639, L13
1188: %
1189: \bibitem[Zucker et al. 2006a]{zucker06a} Zucker, D.~B., et al. 2006, ApJ, 643, L103
1190: %
1191: \bibitem[Zucker et al. 2006b]{zucker06b} Zucker, D.~B., et al. 2006, ApJ, 650, L41
1192: %
1193: \end{thebibliography}
1194:
1195: \clearpage
1196: %
1197: %
1198: %
1199: \begin{figure}
1200: \epsscale{0.8}
1201: \plotone{f1.eps}
1202: \caption{Distribution of our targets on the sky, centered on Leo\,II.
1203: Apparent radial velocity non-members are designated
1204: with open circles. The solid and dashed lines indicate Leo\,II's core and nominal King tidal radius
1205: at 2$\farcm$9 and 8$\farcm$7 (Irwin \& Hatzidimitriou 1995).}
1206: \end{figure}
1207: %
1208: %
1209: %
1210: \begin{figure}
1211: \epsscale{1.0}
1212: \plotone{f2.eps}
1213: \caption{Velocity errors versus V-band magnitude (left panel) and {Signal to Noise ratio per pixel
1214: (right panel). The dashed lines in the right panel indicate our formal mean velocity errors in each S/N
1215: bin, whereas the solid lines show the r.m.s. discrepancy between our measurements and those of V95
1216: (open squares),
1217: which we use to re-scale our uncertainties in Sect.~2.4.}}
1218: \end{figure}
1219: %
1220: %
1221: %
1222: \begin{figure}
1223: \epsscale{1.0}
1224: \plotone{f3.eps}
1225: \caption{Histograms of the radial velocities of stars in our survey of Leo\,~II.
1226: The left panel displays the full sample, whereas the right panel only
1227: shows stars around Leo\,II's systemic velocity. The best-fit Gaussian to
1228: this distribution is shown as a solid line.}
1229: \end{figure}
1230: %
1231: %
1232: %
1233: \begin{figure}
1234: \epsscale{0.6}
1235: \plotone{f4.eps}
1236: \caption{Comparison of our velocity measurements to those
1237: estimated by Vogt et al. (1995) for the 28 stars in common between the
1238: two data sets. {The scale factors to the error bars have not been applied in this
1239: diagram to illustrate our internal formal measurement errors}. The dashed line is unity.}
1240: \end{figure}
1241: %
1242: %
1243: %
1244: \begin{figure}
1245: \epsscale{0.8}
1246: \plotone{f5.eps}
1247: \caption{Top panels: Radial velocity versus major axis radial distance.
1248: The right panel shows those stars around Leo\,II's mean systemic velocity, while
1249: the full data are included in the left panel. Members within (outside) a 3\,$\sigma$
1250: cut are shown as filled (open) symbols.
1251: The bottom panel
1252: displays mean radial velocities obtained in radial bins with a constant number of stars, where we
1253: adopted different cuts in velocity to consider the presence of high- or low-velocity member stars:
1254: 3\,$\sigma$ (open circles), 5\,$\sigma$ (solid squares), and 10\,$\sigma$ (filled diamonds).
1255: No significant radial gradient is discernible in our data. {Monte Carlo tests show
1256: that the hint of a change
1257: in the mean velocity between the outer bins at each end of the major axis
1258: is not statistically significant (see text for details).}
1259: Note the different scales in each of the subplots.}
1260: \end{figure}
1261: %
1262: %
1263: %
1264: \begin{figure}
1265: \epsscale{0.85}
1266: \plotone{f6a.eps}\plotone{f6b.eps}
1267: \caption{Test for apparent rotation {in our observations (top panel) and a sample of
1268: Monte Carlo simulations (bottom panel).}
1269: The points with error bars show the mean velocity difference
1270: of stars on either side of bisecting lines at the respective position
1271: angles. The best-fit sinusoid is displayed as a solid line. The
1272: discernible rotation amplitude at a position angle of $\sim$23$\degr$
1273: is statistically insignificant -- 87\% of the random samples as shown in the bottom panel
1274: produce a comparable amplitude or larger. Indicated
1275: as a dashed line in the top panel is the minor axis position angle of 12$\degr$ from
1276: Irwin \& Hatzidimitriou (1995).}
1277: \end{figure}
1278: %
1279: %
1280: %
1281: \begin{figure}
1282: \plotone{f7.eps}
1283: \caption{Leo\,II's radial velocity dispersion profile obtained for a
1284: 3\,$\sigma$ membership criterion. In each panel, each bin contains the
1285: indicated constant number of stars.}
1286: \end{figure}
1287: %
1288: %
1289: %
1290: \begin{figure}
1291: \epsscale{1.0}
1292: \plotone{f8a.eps}
1293: \plottwo{f8b.eps}{f8c.eps}
1294: \caption{Upper panel: Velocity distribution for a subsample of our data for which there is
1295: extant metallicity information (\p1). The velocity histograms and
1296: dispersion profiles in the lower panels have been separated into a
1297: metal-poor and a metal-rich component, which exhibit only marginally
1298: different kinematics.}
1299: \end{figure}
1300: %
1301: %
1302: %
1303: \begin{figure}
1304: \epsscale{0.8}
1305: \plotone{f9.eps}\plotone{f9b.eps}
1306: \caption{Mass (middle left panel) and density (middle right) profile
1307: of Leo\,II, obtained from Jeans equation under the assumption of an isotropic velocity distribution
1308: and a constant velocity dispersion. Error bounds originate from the uncertainties in
1309: the light and velocity dispersion profile.
1310: The surface brightness profile entering the calculations
1311: is shown in the top panel (after Irwin \& Hatzidimitriou 1995). The solid line is a best-fit
1312: Plummer profile. At Leo\,II's distance, 1$\arcmin$ corresponds to about 0.07\,kpc.
1313: {Based on mass and light profiles, we derive the M/L distribution shown in the lower panel}.}
1314: \end{figure}
1315: %
1316: %
1317: %
1318: \begin{figure}
1319: \epsscale{0.8}
1320: \plotone{f10.eps}
1321: \caption{Dispersion profiles derived
1322: under the assumption of varying degrees of velocity anisotropy
1323: $\beta$. The upper and lower panels adopt two different density
1324: models, namely a NFW cusp (upper panels) and a cored halo-profile
1325: (lower panels), respectively. While the left panels maintain constant
1326: values of $\beta$, the right panels were determined for a velocity
1327: anisotropy which varies radially according to the Osipkov-Merritt
1328: (O.M.) prescription with an anisotropy radius as stated. }
1329: \end{figure}
1330: %
1331: %
1332: %
1333: \clearpage
1334: %
1335: %
1336: %
1337: \begin{table}
1338: \begin{center}
1339: \caption{Measured properties of radial velocity members in Leo\,II}
1340: \begin{footnotesize}
1341: \begin{tabular}{lccrrrcc}
1342: \hline
1343: \hline
1344: Star & $\alpha$ (J2000) & $\delta$ (J2000) & r\,[$\arcmin$] & v$_{\rm r}$ &
1345: $\sigma_{\rm v_r}$ & [Fe/H] & $\sigma_[Fe/H]$ \\
1346: \hline
1347: T\_18 & 11 13 03.5 & 22 11 35.0 & 7.06 & 48.49 & 3.18 & $-1.34$ & 0.18\\
1348: T\_19 & 11 13 23.0 & 22 09 23.0 & 1.64 & 78.18 & 5.04 & $-$1.62 & 0.52\\
1349: V\_19 & 11 13 35.6 & 22 09 06.0 & 1.68 & 76.33 & 8.59 & \nodata & \nodata \\
1350: T\_20 & 11 13 32.1 & 22 09 12.0 & 0.77 & 90.10 & 16.29 & \nodata & \nodata \\
1351: T\_21 & 11 13 22.7 & 22 08 09.0 & 2.11 & 83.99 & 3.53 & $-$1.75 & 0.25 \\
1352: \hline
1353: \end{tabular}
1354: \end{footnotesize}
1355: \end{center}
1356: {\footnotesize Note. --- This Table is published in its entirety in the electronic edition of the {\it
1357: Astronomical Journal}.
1358: A portion is shown here for guidance regarding its form and content. Metallicities are those from \p1.}
1359: \end{table}
1360: %
1361:
1362: \end{document}
1363: