1: %\documentclass[12pt,preprint]{aastex}
2: \documentclass{emulateapj}
3:
4: %% manuscript produces a one-column, double-spaced document:
5:
6: %\documentclass[manuscript]{aastex}
7:
8: %% preprint2 produces a double-column, single-spaced document:
9:
10: %\documentclass[preprint2]{aastex}
11:
12: %% \documentclass[preprint2,longabstract]{aastex}
13:
14: %% If you are submitting to a journal that translates manuscripts into
15: %% SGML, you need to follow certain guidelines when preparing your
16: %% macros. See the AASTeX v5.x Author Guide for information.
17:
18: \newcommand{\vdag}{(v)^\dagger}
19: \newcommand{\myemail}{skywalker@galaxy.far.far.away}
20:
21:
22: \begin{document}
23:
24: %% LaTeX will automatically break titles if they run longer than
25: %% one line. However, you may use \\ to force a line break if you
26: %% desire.
27:
28: \title{Four New Stellar Debris Streams in the Galactic Halo}
29:
30: \author{C. J. Grillmair}
31: \affil{Spitzer Science Center, 1200 E. California Blvd., Pasadena, CA 91125}
32: \email{carl@ipac.caltech.edu}
33:
34:
35: \begin{abstract}
36:
37: We report on the detection of four new stellar debris streams and a
38: new dwarf galaxy candidate in the Sloan Digital Sky Survey. Three of
39: the streams, ranging between 3 and 15 kpc in distance and spanning
40: between 37\arcdeg~ and 84\arcdeg~ on the sky, are very narrow and are
41: most probably tidal streams originating in extant or disrupted
42: globular clusters. The fourth stream is much broader, roughly 45 kpc
43: distant, at least 53\arcdeg~ in length, and is most likely the tidal
44: debris from a dwarf galaxy. As the streams each span multiple
45: constellations, we extend tradition and designate them the Acheron,
46: Cocytos, Lethe, and Styx streams. At the same distance and apparently
47: embedded in the Styx stream is a $\sim 1$ kpc-wide concentration of
48: stars with an apparently similar color-magnitude distribution which we
49: designate Bootes III. Given its very low surface density, its location
50: within the stream, and its apparently disturbed morphology, we argue
51: that Bootes III may be the progenitor of Styx and in possibly the
52: final throes of tidal dissolution. While the current data do not
53: permit strong constraints, preliminary orbit estimates for the streams
54: do not point to any likely progenitors among the known globular
55: clusters and dwarf galaxies.
56:
57:
58:
59: \end{abstract}
60:
61: %% Keywords should appear after the \end{abstract} command. The uncommented
62: %% example has been keyed in ApJ style. See the instructions to authors
63: %% for the journal to which you are submitting your paper to determine
64: %% what keyword punctuation is appropriate.
65:
66: \keywords{globular clusters: general --- Galaxy: Structure --- Galaxy: Halo}
67:
68: \section{Introduction}
69:
70: Despite the once common belief that the stellar debris streams
71: produced by tidal stripping of dwarf galaxies and globular clusters
72: would be quickly dispersed by molecular cloud scattering, orbital
73: precession, and phase mixing, recent observations of our Galaxy and
74: others have shown that such streams are both common and evidently
75: long-lived. The Sloan Digital Sky Survey (SDSS) has proven to be a
76: particularly remarkable resource for finding such streams, and for
77: studying Galactic structure at a level of detail which we
78: cannot hope to match in any other galaxy. In addition to the large
79: scale features attributed to past galaxy accretion events
80: \citep{yann03,maje2003,roch04,
81: grill2006d, belokurov2006b, grill2006e,
82: belokurov2007}, SDSS data has been used to detect the remarkably
83: strong tidal tails of Palomar 5 \citep{oden2001,rock2002, oden2003,
84: grill2006b} and NGC 5466 \citep{belokurov2006a, grill2006a}, as
85: well as the presumed globular cluster stream GD-1 \citep{grill2006c}.
86: Though spectroscopic follow-up and detailed numerical calculations
87: have yet to be carried out for most of these streams, they will no
88: doubt become important for constraining the three dimensional shape of
89: the Galactic potential. Globular cluster streams will be particularly
90: important since they are dynamically very cold \citep{comb99} and
91: therefore useful for constraining not only the global Galactic
92: potential but also its lumpiness \citep{mura99}.
93:
94: In this paper we continue our search of the SDSS database for more
95: extended structures in the Galactic halo. We describe our
96: analysis procedure in Section \ref{analysis}. We discuss four
97: new stellar streams and a possible dwarf galaxy progenitor in Section
98: \ref{discussion} and put initial constraints on their orbits in
99: Section \ref{orbit}. We make concluding remarks Section
100: \ref{conclusions}.
101:
102:
103: \section{Data Analysis} \label{analysis}
104:
105: Data comprising $g,r,$ and $i$ photometry for $7 \times 10^7$ stars in
106: the region $108\arcdeg~ < \alpha < 270\arcdeg~$ and $-4\arcdeg~ < \delta
107: < 65\arcdeg~$ were extracted from the SDSS DR5 database using the SDSS
108: CasJobs query system. The data were analyzed using the matched filter
109: technique employed by \citet{grill2006a}, \citet{grill2006b}, \citet{
110: grill2006c}, \citet{grill2006d}, and \citet{grill2006e}, which itself is a
111: variation on the optimal filtering
112: technique described by \citet{rock2002}. This technique is made
113: necessary by the fact that, over the magnitude range and over the
114: region of sky we are considering, the surface densities of foreground
115: stars are some three orders of magnitude greater than the surface
116: densities of known stellar debris streams. Applied in the
117: color-magnitude (CM) domain, the matched filter is a means by which we can
118: optimally differentiate between streams and foreground populations.
119:
120: Our filtering technique departs somewhat from that of
121: \citet{rock2002}, who were primarily interested in searching for
122: debris from a known and relatively well characterized progenitor. By
123: contrast, our present goal is to survey the sky for discrete but
124: hitherto unknown stellar populations. Since we are interested in
125: detecting streams throughout the Galactic halo, we also need to
126: account for the effects of survey completeness as we search larger and
127: larger volumes. Consequently, rather than using the observed
128: color-magnitude distribution (CMD) for stars of interest (e.g.
129: \citet{rock2002}), we generate template distributions which are based
130: on the observed color-magnitude sequences of several globular clusters
131: situated within the SDSS survey area. Specifically, we measure normal
132: points lying along the $g-r$ and $g-i$ color-magnitude sequence of
133: each cluster and then interpolate to compute the expected color at any $g$
134: magnitude. Using mean photometric errors as a function of magnitude
135: (measured in relatively sparse regions of the survey area where source
136: crowding is not an issue) we broaden the globular cluster sequences by
137: convolving with appropriate Gaussians at each magnitude level. We also
138: apply a fixed broadening of $\sigma = 0.02$ mag at all magnitudes to
139: account for the intrinsic spread in the colors of giant branch stars.
140:
141: Since we have no {\it a priori} knowledge concerning the luminosity
142: function of stars in streams, and since we need to decouple observed
143: luminosity functions from the survey completeness, we adopt a general
144: form for the luminosity function based on the very deep luminosity
145: function of $\omega$ Cen \citep{demarchi1999}, converted to the Sloan
146: system using the empirical transformations of \citet{jordi2006}. We
147: find that the exact form of the luminosity function is not
148: particularly important. Experiments with the somewhat steeper
149: luminosity functions one might expect for tidally stripped stars
150: \citep{koch2004} yield no perceptible improvements over the range of
151: absolute magnitudes considered here.
152:
153: Comparing the observed luminosity function in the outskirts of M 13
154: with the much deeper $\Omega$ Cen luminosity function, we derive an
155: approximate completeness function which is unity at $g = 22$, 0.5 at
156: $g = 23.3$, and 0 at $g = 24.4$. While the actual completeness will
157: vary across the survey area, practical concerns require that we use a
158: single completeness function for the entire field. Since we impose a
159: cutoff at $g = 22.5$, small modifications to the form of
160: our completeness function have only very minor effects on the results.
161: Our expectation is that mismatches between our template
162: color-magnitude sequences and the largely unknown distributions of
163: stars in stellar streams will have a much larger effect on our
164: sensitivity to discrete populations.
165:
166: Once the color-magnitude distribution of the stars of interest has
167: been constructed, an optimal filter requires that this distribution be
168: divided by the corresponding distribution of field stars (e.g.
169: \citet{rock2002}). We sample the field star distribution over various
170: portions of DR5, binning the stars in $g$ and $g-i$ (or $g-r$), and
171: then slightly smoothing over the bins with a Gaussian of kernel
172: $\sigma = 0.02$ mag. Figure 1 shows a template filter based on the CMD
173: of M 13 at its nominal distance of 7.7 kpc \citep{harris96}. To avoid
174: numerical issues in relatively unpopulated regions of the
175: color-magnitude diagram, we set the filter to zero for stars more than
176: 6$\sigma$ from the color-magnitude sequence. Examination of Figure 1
177: shows that the most highly weighted stars are those at the main
178: sequence turn-off. Stars fainter and redward of the turn-off are much
179: less favored, though their integrated contribution remains
180: significant. By virtue of both their relatively small numbers, and of
181: colors that are indistinguishable from the bulk of the foreground
182: population, the subgiant and lower giant branch stars are given
183: comparatively little weight.
184:
185:
186: \begin{figure}
187: \epsscale{1.0}
188: \plotone{f1.eps}
189: \caption{Panel (a): An example of a matched filter based on the
190: SDSS color-magnitude distribution of stars in M 13, but with a luminosity
191: function based on $\Omega$ Cen. The stretch is logarithmic. The peak response of
192: the filter occurs on the blue side of the main sequence turn-off, where
193: there are relatively few foreground stars. The subgiant and lower giant branch are somewhat deweighted due to the very large number
194: of intervening foreground stars in this color range. The width of the
195: filter is based primarily on the mean SDSS photometric errors as a function of
196: magnitude. Stars lying more than $6\sigma$ from the sequence are given a weight of zero. Panel (b): The color-magnitude distribution of stars lying within 10\arcdeg~ of the eastern half of the Cocytos stream, with the filter in panel (a) overlaid. \label{fig1}}
197: \end{figure}
198:
199:
200:
201:
202: We used all stars with $15 < g < 22.5$, and we dereddened the SDSS
203: photometry as a function of position on the sky using the DIRBE/IRAS
204: dust maps of \citet{schleg98}. In an initial survey, a single Hess
205: diagram for field stars was generated using roughly half the Sloan
206: survey area in regions where no streams are currently known to exist.
207: We applied the filters to the entire survey area, and the resulting
208: filtered star counts were summed by location on the sky to produce two
209: dimensional, filtered surface density maps. Once the streams were
210: detected, we optimized the filters for individual streams. Since the
211: CM distribution of the field star population varies over the survey
212: area, we expect that a filter incorporating the distribution of only
213: nearby field stars will enhance the signal-to-noise ratio of the
214: streams. For each stream we sampled the field star population within
215: 10\arcdeg~ of the stream, and with one exception, extending along only
216: the eastern and western halves of each stream. This has the effect of
217: increasing the measured signal-to-noise ratio by a few percent beyond
218: what one can achieve using a single, survey-wide field star
219: distribution. Panel (b) of Figure 1 shows an example of one such field
220: star distribution. Further improvements may be possible by more finely
221: subdividing the field star populations, or modeling the foreground
222: population as a function of sky position, but that is beyond the scope
223: of this paper.
224:
225: In Figure 2 we show the filtered star count distributions after
226: shifting the optimal filters by -1.4, +0.6, +1.2 (M 13-based filter) and +3.2
227: mags (M 15-based filter). The field area is shown in the Sloan Survey
228: coordinate system to improve visibility and reduce the distortions that
229: a projection in the equatorial or Galactic coordinate systems would
230: entail. The images have been binned to a pixel size of
231: $0.1\arcdeg~$ and smoothed using a Gaussian kernel with $\sigma =
232: 0.2\arcdeg$.
233:
234: \begin{figure*}
235: \epsscale{1.0}
236: \plotone{f2.eps}
237: \caption{Matched-filtered surface density maps of stars in the eastern 2/3rds of
238: the DR5 Sloan Digital Sky Survey field. The stretch is logarithmic,
239: and darker areas
240: indicate higher surface densities. For panel (a) a 5th-order
241: polynomial fit has been subtracted from the surface densities for
242: presentation purposes. For the other three panels a 7th-order
243: polynomial surface fit has
244: been has been subtracted. All fields have been
245: smoothed with a Gaussian kernel of width 0.2\arcdeg~. The white areas
246: designate areas of missing data. Panels (a) through (d) result from
247: shifting the M13 or M 15-matched filters by -1.4, +0.8, +1.1, and
248: +3.2 magnitudes, respectively, and the corresponding distances
249: (assuming $d_{M13} = 7.7$ kpc and $d_{M15} = 10.3$ kpc) are
250: indicated. \label{fig2}}
251: \end{figure*}
252:
253:
254:
255:
256: The filtered surface density maps are the sum of maps generated using
257: $g - r$ and $g - i$ filters as these colors best measure the turn-off
258: and main sequence stars of interest. To improve the visibility of the
259: streams in Figure 2, each image has been background subtracted by
260: first masking out globular clusters and dwarf galaxies and then
261: fitting a 5th or 7th order polynomial surface. These surface fits are
262: shown in Figure 3. For the nearest of the streams, a 5th order
263: polynomial fit is found to be sufficient to remove the rise in the
264: number of disk stars at low Galactic latitudes. For the remaining
265: streams a 7th order polynomial was used to subdue the increasing
266: contribution from the Sagittarius stream. As is
267: evident in Figure 3, there are no high-frequency features in the
268: surface fits that could could be held to account for the streams
269: visible in Figure 2. We emphasize that these background subtractions
270: are purely for the purposes of presentation and we make no further use
271: of them in our subsequent analysis.
272:
273:
274: \begin{figure}
275: \epsscale{1.0}
276: \plotone{f3.eps}
277: \caption{Polynomial surface fits used to remove the background in
278: Figure 2. Panel (a) is the result of using a 5-th order polynomial
279: fit, while all
280: other panels employ a 7th-order polynomial to reduce the effect of the
281: Sagittarius stream. None of the streams in Figure 2 can be associated
282: with the much lower frequency undulations visible here. \label{fig3}}
283: \end{figure}
284:
285:
286:
287:
288: We compared Figure 2 with the reddening map of \citet{schleg98}
289: to ensure that apparent stellar over-densities are not due to
290: localized changes in extinction. The reddening map covering the field
291: of interest is shown in Figure 4. There is no apparent correlation between
292: the new streams and the applied reddening corrections. The maximum
293: values of $E(B-V)$ in the regions subtended by the new streams are
294: $\sim 0.13$, with typical values of $< 0.05$. Rerunning the matched
295: filter analysis without reddening corrections has no significant
296: effect on the location or the apparent strengths of the new features.
297:
298:
299: \begin{figure}
300: \epsscale{1.0}
301: \plotone{f4.eps}
302: \caption{$E(B-V)$ over the region of sky shown in Figure 2, as determined from the
303: the DIRBE/IRAS dust maps of \citet{schleg98}. The stretch is linear,
304: with darker regions corresponding to higher reddening. The reddening varies from $< 0.05$ mag to 0.13 mag at the
305: locations of the four streams. \label{fig4}}
306: \end{figure}
307:
308:
309:
310: We have also compared Figure 2 with similar maps made using the
311: SDSS DR5 galaxy database to ascertain the extent to which confusion
312: between stars and galaxies at faint magnitudes could contribute
313: to the features we see. In none of the four cases presented here is
314: there any indication of similar extended features in the distribution of
315: galaxies.
316:
317:
318: \section{Discussion} \label{discussion}
319:
320: Visible in Figure 2 are several well known tidal features, as well as
321: four new stellar debris streams. The new streams are much
322: less pronounced than the Sagittarius, Pal 5, or GD-1 streams, and have
323: average surface densities of between 5 and 50 stars deg$^{-2}$ The
324: streams were initially detected and are most easily distinguished by
325: viewing a rapid sequence of filtered images in which the filter is
326: successively shifted to fainter and fainter magnitudes. The streams
327: become apparent to the eye as linear features which often
328: move from one side of the survey area to the other as one moves
329: outward in distance.
330:
331: All of the new streams span multiple constellations and none can be
332: securely identified with a known progenitor at this time. For
333: convenience, we extend traditional nomenclature and name the
334: new streams after four mythical rivers in Hades: the Acheron (river of
335: sorrow), Cocytos (lamentation), Lethe (forgetfulness), and Styx (hate).
336:
337:
338: \subsection{Acheron \label{acheron}}
339:
340: Visible in panel (a) of Figure 2 is a fairly narrow stream of stars
341: extending some $37\arcdeg~$ across the southeastern corner of the
342: DR5 survey area. The stream extends from the southern edge of Serpens
343: Caput [($\lambda, \eta$) = (45\arcdeg, -36\arcdeg), (R.A., dec.) =
344: (230\arcdeg, -2\arcdeg)] to the center of Hercules [($\lambda, \eta$) =
345: (63\arcdeg, 22\arcdeg), (R.A., dec.) = (259\arcdeg, 21\arcdeg)], and is
346: truncated at both the southern and eastern ends by the limits of the
347: available data.
348:
349: The stream is much less pronounced than (for example) GD-1
350: \citep{grill2006c} and in places the signal-to-noise ratio is almost
351: vanishingly small. To better quantify the significance of the
352: detection, we apply the following test: (i) We trace along the length
353: of the putative stream, connecting the high points in the surface
354: density distribution with segments that match the general curvature of
355: the stream (though in this case there is almost none). (ii) We create
356: a mask image of this trace, setting pixels along the trace and
357: laterally out to 0.25\arcdeg~ in each direction to unity. All other
358: mask pixels are set to zero. This width is chosen to roughly match the
359: apparent width of the stream. (iii) We break the mask into seven
360: stream segments, each approximately 5\arcdeg~ in length. (iv) For each
361: stream segment, we successively shift the mask in the $\lambda$
362: direction across a representative portion of the sky, 0.1\arcdeg~ at a
363: time, and multiply the mask by the filtered image in Figure 2. (v) For
364: each segment, we fit a one dimensional, 3rd order polynomial to the
365: mean filtered star counts as a function of lateral distance from the
366: stream and subtract it. We exclude from this fit the region within
367: 0.5\arcdeg~ of the stream. (vi) For each lateral offset, we compute
368: the median of the background-subtracted responses over all segments.
369: This last calculation serves as a continuity constraint and prevents
370: strong biasing due to (for example) a single populous star cluster in
371: in any one segment.
372:
373: More succinctly, if we define $f(i,j)$ as the filtered star counts in
374: a pixel with indices $i$ and $j$, and $m(i,j)$ as the stream-tracing
375: mask, then the signal $s$ for stream segment $k$, shifted laterally by
376: offset $d$, is:
377:
378: \begin{equation}
379: s(d,k) = \frac{\sum_{i,j}[f(i,j) \times m(i+d,j)]}{ n(d,k)},
380: \end{equation}
381:
382: where the summation is over all valid pixels (i.e. excluding
383: portions of the Survey footprint with missing data). The total stream
384: signal at offset $d$ is then:
385:
386: \begin{equation}
387: T(d) = {\rm median} [s(d,k)]_{k=1,...l},
388: \end{equation}
389:
390: where $l$ is the number of stream segments.
391:
392: We plot the run of $T$ versus lateral offset for Acheron in Figure 5.
393: The signal due to the stream is clearly visible as a broad peak
394: approximately centered on a lateral offset of 0\arcdeg. We note that
395: perfect centering is not expected; aside from measurement
396: uncertainties in regions with very little signal, the lateral profiles
397: of real tidal streams need not be symmetric, depending on such factors
398: as our line of sight to the stream and what portion of the stream's
399: orbit is being observed. If we regard $T$ beyond 1\arcdeg~ from the
400: stream as being due to random clumping of stars and therefore
401: a reasonable measure of the noise, we find $\sigma_{T} \approx 5$,
402: where the standard deviation is measured after binning over the
403: 0.5\arcdeg~ width of the mask. Integrating over the region $-1\arcdeg
404: < d < 1\arcdeg$ we find that we have detected the stream at the $\sim
405: 10 \sigma$ level. For comparison, a similar test applied to GD-1
406: yields a signal-to-noise ratio of $\sim 13$.
407:
408:
409: \begin{figure}
410: \epsscale{1.0}
411: \plotone{f5.eps}
412: \caption{Stream signal $T$ (described in the text) as a function of $\lambda$ offset from the Acheron stream. The dashed line shows the results of a similar test using an artificial stream with a Gaussian cross section and
413: FWHM = 0.9\arcdeg. \label{fig5}}
414: \end{figure}
415:
416:
417:
418: Also shown in Figure 5 is an identical test applied to an artificial
419: tidal stream at the same location as Acheron and with a Gaussian cross
420: section. The Gaussian which yields the best match to the observed
421: stream profile has a full width at half maximum (FWHM) of 0.9\arcdeg.
422: While the physical cross section of the stream need not be Gaussian,
423: the convolved artificial stream profile matches the actual stream
424: profile reasonably well, and we use the best fitting FWHM as a
425: convenient measure of the stream's breadth. At a distance of 3.6 kpc
426: (see below), this corresponds to a physical width of 60 pc. This is
427: similar to the widths measured for the tidal tails of the globular
428: clusters Pal 5 and NGC 5466
429: \citep{grill2006b,belokurov2006a,grill2006a} and the presumed cluster
430: remnant GD-1 \citep{grill2006c}. On the other hand, the width is much
431: narrower than the tidal arms of the Sagittarius dwarf
432: \citep{maje2003,mart2004} or the presumed dwarf galaxy streams
433: discussed by \citet{grill2006d}, \citet{belokurov2006b},
434: \citet{belokurov2007}, and \citet{grill2006e}. This is consistent with
435: the hypothesis that the stars making up the stream have very low random
436: velocities, and that they were weakly stripped from a relatively small
437: potential. Combining this with an orbit which passes low over the
438: Galactic bulge (see below) suggests that the parent body is (or was) a
439: globular cluster.
440:
441: Following \citet{grill2006c} we shift the main sequence of the M
442: 13-based filter brightward and faintward to estimate the stream's
443: distance. To avoid potential problems related to a difference in age
444: between M 13 and the stream stars, we use only the portion of the
445: filter with $19.5 < g < 22.5$, where the bright cutoff is 0.8 mags
446: below M 13's main sequence turn-off. This significantly reduces the
447: contrast between the stream and the background (since turn-off stars
448: contribute a substantial portion of the signal) but still provides
449: sufficient signal to enable a reasonably precise measurement of peak
450: contrast. We find that the strength of the southern end of the stream
451: peaks at a magnitude shift of -1.53 mags, the central portion has the
452: highest contrast at -1.58 mags, and the northernmost portion of the
453: stream peaks at -1.73 mags. Adopting a distance to M 13 of 7.7 kpc
454: \citep{harris96}, this puts the southern end of the stream at a
455: heliocentric distance of 3.8 kpc, while the northern end is at 3.5
456: kpc. While the match between the color-magnitude distributions of
457: stars in M 13 and in the stream is uncertain, the {\it relative}
458: line-of-sight distances along the stream should be fairly robust;
459: we estimate our random measurement uncertainties to be $\approx 10\%$.
460:
461: {\it Absolute} distance estimates using this method depend not only on
462: the uncertainty in the distance to M 13 but also on the CMD of
463: foreground stars and the degree to which the metallicity (and hence
464: color) of M 13's main sequence matches that of the stream. We have
465: only a very coarse estimate of the latter, namely the maximum contrast
466: obtained for the stream when the star counts are processed with
467: matched filters made from different globular clusters. If as a rough
468: estimate of this uncertainty we take half the $g$ magnitude offset
469: (0.46 mag) at a fixed color between the (dereddened) main sequence
470: loci of M13 and M 15 (with [Fe/H] of -1.54 and -2.25, respectively),
471: and combine this with a 5\% uncertainty in the distance to M 13
472: \citep{grundahl1998}, we arrive at a probable lower bound on the
473: systematic distance uncertainty of 11\%.
474:
475:
476: Integrating the locally background subtracted, filtered star counts
477: over a width of $\approx 1\arcdeg~$ we find the total number of stars
478: in the discernible stream to be $1300 \pm 200$. For stars with $g <
479: 22.5$ the average surface density is $50 \pm 5$ stars deg$^{-2}$, with
480: occasional peaks of over 100 stars deg$^{-2}$. While these surface
481: densities are similar to those found by \citet{grill2006c} for GD-1,
482: Acheron appears considerably less pronounced. This is simply a
483: consequence of the much larger number of contaminating foreground
484: stars near the Galactic plane.
485:
486:
487: \subsection{Cocytos \label{cocytos}}
488:
489:
490: Visible in panel (b) of Figure 2 is a faint, narrow stream extending
491: from Virgo [($\lambda, \eta$) = (1\arcdeg, -35\arcdeg), (R.A., dec.) =
492: (186\arcdeg, -3\arcdeg)] in the south to Hercules [($\lambda, \eta$) =
493: (65\arcdeg, 20\arcdeg), (R.A., dec.) = (259\arcdeg, 20\arcdeg)] in the east.
494: The 80\arcdeg~~ length of the stream is again limited by the extent of
495: the SDSS survey area.
496:
497: In Figure 6 we show the run of $T$ with lateral distance from the stream. In
498: this case we have shifted our stream mask at a 45\arcdeg~ angle across
499: Figure 2 and divided the stream mask into 12, $\sim 5\arcdeg$-long
500: segments. We find that $\sigma_{T} \approx 5.0$ at $|d| > 1\arcdeg~$ and,
501: integrating from $-1\arcdeg < d < 1\arcdeg$, determine that we have
502: detected the stream at the $\sim 8\sigma$ level. Generating an artificial
503: stream and applying the same test, we find that we obtain the best
504: match to the observed profile using a Gaussian with FWHM =
505: 0.7\arcdeg~. At a distance of 11 kpc (see below) this corresponds to a physical
506: width of 140 pc. This is again similar to known globular cluster
507: streams and argues that the progenitor of Cocytos is (or was) a
508: globular cluster.
509:
510: \begin{figure}
511: \epsscale{1.0}
512: \plotone{f6.eps}
513: \caption{As in Figure 5, but for the Cocytos stream. The artificial
514: stream that best matches the observed $T$ profile has a Gaussian
515: profile with FWHM = 0.7\arcdeg. \label{fig6}}
516: \end{figure}
517:
518:
519:
520: The strength of both the southern and eastern ends of the stream peak
521: at an M 13 main sequence magnitude shift of 0.9 mags. The estimated
522: distance to the stream is therefore $11 \pm 2$ kpc. Integrating the
523: background subtracted, weighted star counts over a width of $\approx
524: 1\arcdeg~$ we find the total number of stars in the discernible stream
525: to be $500 \pm 100$. For stars with $g < 22.5$ the average surface
526: density is between 5 and 8 stars deg$^{-2}$.
527:
528:
529: \subsection{Lethe \label{lethe}}
530:
531: Panel (c) of Figure 2 shows a faint stream extending from
532: Hercules [($\lambda, \eta$) = (64\arcdeg, 19\arcdeg), (R.A., dec.) =
533: (258\arcdeg, 20\arcdeg)] to Leo [($\lambda,\eta$) = (-13\arcdeg,
534: -14\arcdeg), (R.A., dec.) = (171\arcdeg, 18\arcdeg)]. The stream extends
535: from the eastern limit of the survey and appears to fade substantially
536: in Leo, where it crosses the Sagittarius stream. There may be a
537: continuation of the stream south of the Sagittarius stream but we
538: have been unable to identify it with any confidence.
539:
540: The apparent strength of Lethe relative to the distribution of
541: foreground stars peaks at an M 13-relative offset of 1.3 mags at the
542: eastern end, and about 1.0 mag at the western end. This puts the stream at
543: a distance of between 12.2 and 13.4 kpc. Figure 7 shows the run of
544: $T(d)$ for the stream, where again we have translated the stream mask at a
545: 45\arcdeg~ angle across the filtered imaged in Figure 2 and subdivided
546: the mask into 12 stream segments. In this case we find $\sigma_{T} =
547: 1.8$ at $|d| > 1\arcdeg~$, and integrating within this region we find
548: that Lethe is detected at
549: the $7\sigma$ level. The best matching Gaussian stream shown in Figure 7
550: has a FWHM of 0.4\arcdeg, which at 13 kpc corresponds to a physical
551: width of 95 pc. Once again we conclude that Lethe is the debris
552: stream of a globular cluster.
553:
554: \begin{figure}
555: \epsscale{1.0}
556: \plotone{f7.eps}
557: \caption{As in Figure 5, but for the Lethe stream. The artificial
558: stream that best matches the observed $T$ profile has a Gaussian
559: profile with FWHM = 0.4\arcdeg. \label{fig7}}
560: \end{figure}
561:
562:
563:
564:
565: Integrating the background subtracted, weighted star counts over a
566: width of $\approx 1\arcdeg~$ we find the total number of stars in the
567: discernible stream to be $1100 \pm 300$. For stars with $g < 22.5$
568: the average surface density is 12 stars deg$^{-2}$, with the more
569: pronounced regions of the stream having $\sim 30$ stars deg$^{-2}$.
570:
571:
572: \subsection{Styx} \label{styx}
573:
574: Visible in panel (d) of Figure 2 is a broad stream extending west from
575: Hercules [($\lambda, \eta$) = (63\arcdeg, 21\arcdeg), (R.A., dec.) =
576: (259\arcdeg, +21\arcdeg)] to where it is overwhelmed by the much more
577: populous Sagittarius stream in Coma Berenices [($\lambda, \eta$)
578: = (8\arcdeg, -12\arcdeg), (R.A., dec.) = (194\arcdeg, +20\arcdeg)].
579: The contrast for Styx is improved if, instead of using a matched
580: filter based on M 13, we use the SDSS CMD of M 15. This
581: suggests that the CMD of stars in Styx is bluer than that of M
582: 13. At the distance of Styx the stellar main sequence is almost
583: entirely beyond the SDSS limiting magnitude. Consequently the main
584: sequence fitting technique for estimating distance is not
585: usable and we are forced to rely on turn-off and subgiant stars. Using
586: the M 15 filter, the apparent contrast of the stream peaks at
587: magnitude offsets of +2.8 mag at the western end, +3.2 mag in the
588: central portion, and +3.4 at the eastern end. Adopting a distance to M
589: 15 of 10.3 kpc \citep{harris96}, this translates to estimated
590: distances of 38, 45, and 50 kpc, respectively. Owing to both the lack
591: of a direct main sequence comparison and to the rather distended
592: nature of the stream (which makes foreground estimation problematic),
593: we regard these distances as very approximate until such time as
594: deeper photometry becomes available.
595:
596: Figure 8 shows the run of $T$ with lateral offset. For Styx we have
597: used a mask width of 1\arcdeg, divided the mask into 12, $\sim
598: 4\arcdeg$-long segments, and shifted the mask solely in the $\eta$
599: direction. In this case we find $\sigma_{T} = 0.18$ at $|d| >
600: 3\arcdeg~$, and integrating from -3\arcdeg~ to +3\arcdeg~ we find that
601: Styx is detected at the $\sim 13\sigma$ level. The stream profile is
602: noticeably asymmetric, with a fairly sharp northern edge but an excess
603: of stars extending some 4\arcdeg~ to the south. This is qualitatively
604: similar to the morphologies of tidal tails in N-body simulations
605: (e.g. \citet{choi2007}). The artificial stream which best matches the
606: primary component of Styx has a FWHM of 3.3\arcdeg. At a distance of
607: 45 kpc this corresponds to 2.6 kpc. This is much broader than the
608: presumed globular cluster streams above, but is similar to the the
609: dwarf galaxy streams discussed by \citet{grill2006d},
610: \citet{belokurov2007}, and \citet{grill2006e}, though narrower by half
611: compared to the Sagittarius stream
612: \citep{maje2003,mart2004,belokurov2006b}. We conclude that this stream
613: is most likely debris from a dwarf galaxy.
614:
615: \begin{figure}
616: \epsscale{1.0}
617: \plotone{f8.eps}
618: \caption{As in Figure 5, but for the Styx stream. The artificial
619: stream that best matches the observed $T$ profile over the region $
620: -3\arcdeg < d < 3\arcdeg$ has a Gaussian
621: profile with FWHM = 3.3\arcdeg. The stream is
622: clearly asymmetric, with an extended spray of stars to the south, and
623: a fairly sharp cut off on the northern side.\label{fig8}}
624: \end{figure}
625:
626:
627:
628: \subsection{Bootes III: A New Dwarf Galaxy?} \label{bootes3}
629:
630: With a contrast maximum at very nearly the same distance ($\sim 46$
631: kpc) as the Styx stream is a relatively compact feature at [($\lambda,
632: \eta$) = (21.6\arcdeg, -3.5\arcdeg), (R.A., dec) =(209.3\arcdeg,
633: 26.8\arcdeg)]. The object has a filtered star count surface density
634: many times higher than any visible portion of Styx and is clearly
635: distinct within the stream. Panel (a) of Figure 9 shows the filtered
636: star count distribution in the immediate vicinity of this object.
637: Galaxy cluster ACO 1824 lies within 3\arcmin~ of the location of this
638: object\citep{abell1989}; could the apparent over-density be due to
639: SDSS misclassification of stars at faint magnitudes? In panel (b) of
640: Figure 9 we show the distribution of objects classified as galaxies in
641: DR5, where we have used a filter identical to that used in panel (a).
642: There is little correspondence between the two distributions, and
643: certainly no concentration of objects at the same position. We
644: conclude that this over-density is due to an equidistant collection of
645: stars orbiting in the outskirts of our own Galaxy and we henceforth
646: designate it Bootes III.
647:
648: \begin{figure}
649: \epsscale{1.0}
650: \plotone{f9.eps}
651: \caption{(a) Matched-filtered surface density map in the immediate vicinity of Bootes
652: III, smoothed with a Gaussian kernel of width 0.1\arcdeg, at an M
653: 15-relative magnitude shift of 3.2 mags. Darker areas indicate higher
654: surface densities and the stretch is linear. Though at very
655: different distances, NGC 5466, NGC 5272, and the Canes
656: Venatici dwarf \citep{zucker2006} are also visible by virtue of their
657: high surface densities and the overlap of parts of their CMDs with
658: the M 15 filter. (b) The distribution of objects classified as
659: galaxies in DR5, where we have used a filter identical to that used in
660: panel (a). \label{fig9}}
661: \end{figure}
662:
663:
664:
665: Figure 10 shows a contour plot of the filtered star counts in and
666: around Bootes III; the object appears somewhat double-lobed, possibly
667: disturbed, and extends $\approx 1.5\arcdeg~$ from east to west. At 46
668: kpc this corresponds to a spatial extent of about $\sim 1$ kpc. If the
669: stellar populations in the eastern and western lobes of Bootes III are
670: identical, then the difference in the filter shifts required to
671: maximize the apparent contrast indicates that the eastern lobe of the
672: galaxy is some 3 kpc closer to us than the western lobe. If the two
673: lobes are indeed part of the same structure then Bootes III must be
674: highly elongated along the line of sight. Figure 11 shows a surface
675: density profile of Bootes III, where we have counted all stars with $g
676: < 22.7$ and $-1 < g-i < 1$. The galaxy is evidently quite extended,
677: with a power-law surface density profile that goes as $\sigma \propto
678: r^{-1.0 \pm 0.2}$. Integrating the background-subtracted counts out
679: to 1\arcdeg~ we find a total of 302 stars with $g \le 22.7$ which we
680: can reasonably attribute to Bootes III.
681:
682: \begin{figure}
683: \epsscale{1.0}
684: \plotone{f10.eps}
685: \caption{Contour plot of a region centered on Bootes III. The underlying image has been smoothed with a Gaussian kernel of 6\arcmin. Contours are
686: spaced linearly, and the strong source to the northeast
687: of Bootes III is NGC 5466.}
688: \end{figure}
689:
690:
691: \begin{figure}
692: \epsscale{1.0}
693: \plotone{f11.eps}
694: \caption{The (unfiltered) surface density profile of all stars with $g < 22.7$ and $-1 < g -i < 1$. The center of Bootes III is taken to be at (R.A., dec) = (209.281\arcdeg, 26.775\arcdeg). The surface densities have been background subtracted using the measured counts in an annulus extending from 1.5\arcdeg~ to 2\arcdeg.}
695: \end{figure}
696:
697:
698:
699: As is the case for Styx, using the SDSS color-magnitude distribution
700: of stars in M 15 as the basis for the matched filter yields a slightly
701: higher contrast between the galaxy and the foreground population. The
702: inference is that the age and/or metallicity of Bootes III and the
703: stream are more similar to that of M 15 than M 13. At 46 kpc the
704: galaxy is revealed almost entirely by subgiant and turn-off stars;
705: removing the red giant branch from the filter has little effect on the
706: apparent contrast. Figure 12 shows the color-magnitude distribution of
707: all stars within 1\arcdeg~ of the center of Bootes III. Here we have
708: subtracted the distribution of stars in several regions around Bootes
709: III over an area of $\sim 60$ square degrees. Comparing with the CM
710: loci of M 13 and M 15 (shifted vertically to a distance of 46 kpc),
711: there is a clear overdensity of stars along the expected positions of
712: the main sequence turn-off, the subgiant branch, and the lower giant
713: branch. Consistent with the filter responses above, the CM locus of M
714: 13 is evidently $\sim 0.1$ mag too red to match the apparent
715: distribution. The M 15 locus clearly provides a better match to
716: the data.
717:
718: \begin{figure}
719: \epsscale{1.0}
720: \plotone{f12.eps}
721: \caption{The background-subtracted, color-magnitude distribution of
722: all stars lying within 1\arcdeg~ of the center of Bootes III. Darker
723: areas indicate enhancements over the background distribution. Stars
724: were binned in color and magnitude and smoothed with a Gaussian kernel of
725: $\sigma = 0.1$. The distribution was background-subtracted using
726: similarly binned and smoothed color-magnitude distributions in
727: regions spaced around Bootes III and the two nearby globular clusters
728: and covering a total of $\sim 60$ square degrees. Also shown are the
729: $g, g-i$ and $g, g-r$ color-magnitude loci for M 13 (to the red) and M
730: 15 (to the blue), as derived from SDSS photometry.}
731: \end{figure}
732:
733:
734:
735:
736: In Figure 13 we show the CMD of all stars within 0.8\arcdeg~ of the
737: center of Bootes III. While the turn-off, subgiant, and red giant
738: branches of Bootes III are lost among the unsubtracted foreground
739: stars in this figure, there is a clear concentration of stars at the
740: expected location of the blue horizontal branch (BHB). Fitting the BHB
741: sequence tabulated for the SDSS system by \citet{sirko2004}, we find
742: $(M-m)_0 = 18.35 \pm 0.01$. Corresponding to a distance of 46.7 kpc,
743: this is in excellent agreement with our maximum contrast distance
744: estimate above. In Figure 14 we show the distribution of stars
745: selected to have colors and magnitudes consistent with Bootes III's
746: BHB. There is an apparent enhancement of such stars across the face of
747: Bootes III, though just as for the turn-off stars sampled using the
748: matched filter, the distribution of BHB stars is not very centrally
749: concentrated. The BHB star distribution is considerably more extended
750: in the east-west direction (much like the filtered star counts in
751: Figures 9 and 10), subtending nearly 2\arcdeg on the sky.
752:
753: \begin{figure}
754: \epsscale{1.0}
755: \plotone{f13.eps}
756: \caption{The color-magnitude distribution of stars within 0.8\arcdeg~
757: of the center of Bootes III. A blue horizontal branch is clearly
758: visible at $(g, g - i) \approx 19, -0.25$). The solid line shows the
759: predicted SDSS blue horizontal branch sequence of \citet{sirko2004},
760: shifted vertically by 18.35 mags.}
761: \end{figure}
762:
763:
764: \begin{figure}
765: \epsscale{1.0}
766: \plotone{f14.eps}
767: \caption{The sky distribution of candidate Bootes III blue horizontal branch stars. Based on the CMD in Figure 13, the stars are selected to have $18.7 < g < 19.3$ and $-0.55 < g - i < -0.07$.}
768: \end{figure}
769:
770:
771:
772: All the evidence is consistent with Bootes III being a dwarf galaxy (or
773: a remnant thereof). Combined with its apparent location at the same
774: distance as Styx and very nearly in the middle of it, we infer that
775: Bootes III is both physically associated with the stream and quite
776: possibly its progenitor. Its broad spatial extent, its low surface
777: density and power-law profile, its possibly disturbed morphology, and
778: its location within Styx, all suggest that Bootes III may be in or
779: nearing the final throes of tidal dissolution.
780:
781:
782: \section{Constraints on Orbits} \label{orbit}
783:
784: Though a lack of velocity information prevents us from tightly
785: constraining the orbits of the streams, the apparent orientations of
786: the streams combined with our distance estimates can yield some
787: initial constraints. We use the Galactic model of \citet{allen91},
788: which assumes a spherical halo potential. We employ a least squares
789: method to fit both the stream orientations on the sky and our
790: estimated distances. The tangential velocities at each point are
791: primarily constrained by the projected paths of the streams while our
792: relative distance estimates help to limit the range of possible radial
793: velocities at any point.
794:
795: We fit to a number of normal points lying along the estimated
796: centerlines of each stream. We adopt a solar Galactocentric distance
797: of 8.5 kpc, and distance uncertainties of 1.0, 2.5, 2.5, and 10 kpc for
798: Acheron, Cocytos, Lethe, and Styx, respectively. We choose fiducial
799: points in each stream for radial velocity predictions based in part on
800: the apparent strength of the stream at those locations. When selecting
801: targets for follow-up spectroscopy, these regions will presumably have
802: the largest concentrations of stream stars and thus the highest
803: targeting priority. The predicted heliocentric radial velocities and
804: proper motions at the fiducial points are listed in Table 1. The
805: uncertainties correspond to the 90\% confidence interval for each
806: parameter and do not take account of inaccuracies in the model
807: potential. Predictions are provided for both prograde
808: and retrograde orbits. Three-view projections of the best fit orbits
809: are shown in Figure 15.
810:
811: \begin{figure}
812: \epsscale{1.0}
813: \plotone{f15.eps}
814: \caption{Orbit projections for the four new streams in X, Y, Z
815: Galactic coordinates. The red curve corresponds to Acheron, yellow
816: to Cocytos, green to Lethe, and Styx is represented by the blue
817: curve. The heavy lines show the portions of the orbits visible in
818: Figure 1. The thin solid curves show the orbits
819: integrated forward in time while the dotted portions of the curves
820: show the same orbit integrated backwards. The Sun's position at (X,Y,Z) =
821: (8.5,0,0) kpc is indicated.}
822: \end{figure}
823:
824:
825:
826:
827: Do the computed orbits suggest possible associations between the new
828: streams and known residents of the halo? Integrating orbits for a
829: sufficiently long time, one is almost certain to pass near the current
830: positions of known globular clusters. However, given our limited
831: knowledge of the Galactic potential and the rather rudimentary orbit
832: constraints determined above, such an exercise would clearly be
833: pointless. However, if we limit ourselves to integrating orbits no
834: more than once around the Galaxy, we can match the computed positions,
835: distances, and radial velocities against the 150 globular clusters
836: compiled by \citet{harris96}. Setting rejection limits of 5\arcdeg~
837: in sky position, 5 kpc or a 30\% difference in distance, and 50 km
838: s$^{-1}$ in radial velocity, we can say that there are no known
839: globular clusters that match our orbits simultaneously in position,
840: distance, and radial velocity. Similarly, other than Bootes
841: III there are no known dwarf galaxies that lie along the integrated
842: orbit of Styx. While the uncertainties in our orbit estimates do not
843: allow us to make strong statements at this point, our results are
844: consistent with the hypothesis that the newly detected globular
845: cluster streams are remnants of clusters that have long since
846: been whittled away to nothing.
847:
848:
849:
850:
851:
852: \section{Conclusions} \label{conclusions}
853:
854: Applying optimal contrast filtering techniques to SDSS data, we have
855: detected four new stellar streams in the Galactic halo, along with a
856: probable dwarf galaxy in possibly the final throes of tidal
857: disruption. Three of the streams are spatially very narrow and are
858: likely to be the remnants of extant or disrupted globular clusters.
859: The fourth stream is much broader, similar in width to the Orphan and
860: Anticenter streams, and presumably constitutes the tidal debris of a
861: dwarf galaxy. Bootes III, the new dwarf galaxy candidate, appears to
862: lie near the middle of this stream and may well be its
863: progenitor. Based on an apparently good match to the color-magnitude
864: distribution of stars in M 13 and M 15, we conclude that the stars
865: making up the streams are old and metal poor, with the Styx stream and
866: Bootes III being the oldest and/or most metal poor.
867:
868: A determination of the nature and the properties of our dwarf galaxy
869: candidate with deep imaging are the subject of a forthcoming paper
870: \citep{grill2008}. If indeed Bootes III is in the final throes of
871: tidal dissolution, it will be a particularly interesting target for
872: detailed kinematic studies. Refinement of the stream orbits will require radial
873: velocity measurements of individual stars, though given the very low
874: stellar surface densities in these streams, this will necessarily be
875: an ongoing task. In this respect, these streams may be
876: particularly well suited for upcoming spectroscopic survey instruments such
877: as LAMOST.
878:
879:
880: \acknowledgments
881:
882: The author is grateful to an anonymous referee for numerous
883: recommendations that greatly improved both the presentation and the
884: quality of the results. Funding for the SDSS and SDSS-II has been
885: provided by the Alfred P. Sloan Foundation, the Participating
886: Institutions, the National Science Foundation, the U.S. Department of
887: Energy, the National Aeronautics and Space Administration, the
888: Japanese Monbukagakusho, the Max Planck Society, and the Higher
889: Education Funding Council for England. The SDSS Web Site is
890: http://www.sdss.org/.
891:
892: The SDSS is managed by the Astrophysical Research Consortium for the
893: Participating Institutions. The Participating Institutions are the
894: American Museum of Natural History, Astrophysical Institute Potsdam,
895: University of Basel, University of Cambridge, Case Western Reserve
896: University, University of Chicago, Drexel University, Fermilab, the
897: Institute for Advanced Study, the Japan Participation Group, Johns
898: Hopkins University, the Joint Institute for Nuclear Astrophysics, the
899: Kavli Institute for Particle Astrophysics and Cosmology, the Korean
900: Scientist Group, the Chinese Academy of Sciences (LAMOST), Los Alamos
901: National Laboratory, the Max-Planck-Institute for Astronomy (MPIA),
902: the Max-Planck-Institute for Astrophysics (MPA), New Mexico State
903: University, Ohio State University, University of Pittsburgh,
904: University of Portsmouth, Princeton University, the United States
905: Naval Observatory, and the University of Washington.
906:
907:
908: {\it Facilities:} \facility{Sloan}.
909:
910:
911: \begin{thebibliography}{}
912:
913: \bibitem[Abell, Corwin, \& Olowin (1989)]{abell1989} Abell, G. O.,
914: Corwin, H. G. Jr., \& Olowin, R. P. 1989, \apjs, 70, 1
915:
916: \bibitem[Allen \& Santillan (1991)]{allen91} Allen, C., \& Santillan,
917: A. 1991, {\it Rev. Mex. Astron. Astrofis.}, 22, 255
918:
919: \bibitem[Belokurov et al. (2006a)]{belokurov2006a} Belokurov, V., Evans, N. W.,
920: Irwin, M. J., Hewett, P. C., \& Wilkinson, M. I. 2006a, \apjl, 637, 29
921:
922: \bibitem[Belokurov et al. (2006b)]{belokurov2006b} Belokurov, V., et
923: al. 2006b, \apjl, 642, 137
924:
925: \bibitem[Belokurov et al. (2007)]{belokurov2007} Belokurov, V., et
926: al. 2007, \apj, 658, 337
927:
928: \bibitem[Choi, Weinberg, \& Katz (2007)]{choi2007} Choi, J.-H.,
929: Weinberg, M.~D., \& Katz, N. 2007, \mnras, 381, 987
930:
931: \bibitem[Combes, Leon, \& Meylan (1999)]{comb99} Combes, F., Leon, S.,
932: \& Meylan, G. 1999, \aap, 352, 149
933:
934: \bibitem[de Marchi (1999)]{demarchi1999} de Marchi, G. 1999, \aj, 117,
935: 303
936:
937: \bibitem[Gnedin \& Ostriker (1997)]{gnedin1997} Gnedin, O. Y., \&
938: Ostriker, J. P. 1997, \apj, 474, 223
939:
940: \bibitem[Grillmair (2006a)]{grill2006d} Grillmair, C. J. 2006a, \apjl, 645, 37
941:
942: \bibitem[Grillmair (2006b)]{grill2006e} Grillmair, C. J. 2006b, \apjl, 651, 29
943:
944: \bibitem[Grillmair \& Johnson (2006)]{grill2006a} Grillmair, C. J., \&
945: Johnson, R. 2006, \apjl, 639, 17
946:
947: \bibitem[Grillmair \& Dionatos (2006a)]{grill2006b} Grillmair, C. J., \&
948: Dionatos, O. 2006a, \apjl, 641, 37
949:
950: \bibitem[Grillmair \& Dionatos (2006b)]{grill2006c} Grillmair, C. J., \&
951: Dionatos, O. 2006b, \apjl, 643, 17
952:
953: \bibitem[Grillmair, Hamam, \& Laher (2008, in preparation)]{grill2008} Grillmair, C. J., Hamam,
954: N., \& Laher, R. 2008, in preparation.
955:
956: \bibitem[Grundahl et al. (1998)]{grundahl1998} Grundahl, F.,
957: Vandenberg, D. A., \& Anderson, M. I. 1998, \apjl, 500, 179
958:
959: \bibitem[Harris (1996)]{harris96} Harris, W. E. 1996, \aj, 112, 1487
960:
961: \bibitem[Jordi, Grebel, \& Ammon (2006)] {jordi2006} Jordi, K.,
962: Grebel, E. K., \& Ammon, K. 2006, \aa, 460, 339
963:
964: \bibitem[Koch et al. (2004)]{koch2004} Koch, A., Grebel, E. K.,
965: Odenkirchen, M., Martinez-Delgado, D., \& Caldwell, J. A. R. 2004,
966: \aj, 128, 2274
967:
968: \bibitem[Martinez-Delgado et al. (2004)]{mart2004} Martinez-Delgado,
969: D., Gomez-Flechoso, M., Aparicio, A., \& Carrera, R. 2004, \apj, 601,
970: 242
971:
972: \bibitem[Majewski et al. (2003)]{maje2003} Majewski, S. R., Skrutskie,
973: M. F., Weinberg, M. D., \& Ostheimer, J. C. 2003, \apj, 599, 1082
974:
975: \bibitem[Murali \& Dubinski (1999)]{mura99} Murali, C., \& Dubinski,
976: J. 1999, \aj, 118, 911
977:
978: \bibitem[Odenkirchen et al. (2001)]{oden2001} Odenkirchen, M., et
979: al. 2001, \apjl, 548, 1650
980:
981: \bibitem[Odenkirchen et al. (2003)]{oden2003} Odenkirchen, M. et
982: al. 2003, \aj, 126, 2385
983:
984: \bibitem[Rocha-Pinto et al. (2004)]{roch04} Rocha-Pinto, H. J.,
985: Majewski, S. R., Skrutskie, M. F., Crane, J. D., Patterson,
986: R. J. 2004, \apj, 615, 732
987:
988: \bibitem[Rockosi et al. (2002)]{rock2002} Rockosi, C. M. et al. 2002,
989: \aj, 124, 349
990:
991: \bibitem[Schlegel, Finkbeiner, \& Davis (1998)]{schleg98} Schlegel,
992: D. J., Finkbeiner, D. P., \& Davis, M. 1998, \apj, 500, 525
993:
994: \bibitem[Sirko et al. (2004)]{sirko2004} Sirko, E., Goodman, J.,
995: Knapp, G. R., Brinkman, J., Ivezic, Z., Knerr, E. J., Schlegel, D.,
996: Schneider, D. P., \& York, D. G. 2004, \aj, 127, 899
997:
998: \bibitem[Yanny et al. (2003)]{yann03} Yanny, B. et al. 2003, \apj,
999: 588, 824
1000:
1001: \bibitem[Zucker et al. (2006)]{zucker2006} Zucker, D. B. et al. 2006,
1002: \apjl, 643, L103
1003:
1004: \end{thebibliography}
1005:
1006:
1007: \begin{deluxetable}{lrrcccccccccc}
1008: \tablecaption{Predicted Motions and Orbit Parameters}
1009: \tablecolumns{11}
1010: \tablewidth{0pc}
1011: \tablehead{
1012: \multicolumn{1}{c} {Stream} &
1013: \multicolumn{2}{c} {Fiducial Point} &
1014: \multicolumn{3}{c} {Prograde Orbit} &
1015: \multicolumn{3}{c} {Retrograde Orbit} \\
1016: \multicolumn{1}{c} {}&
1017: \multicolumn{1}{c} {R.A.} &
1018: \multicolumn{1}{c} {dec} &
1019: \multicolumn{1}{c} {$v_r$} &
1020: \multicolumn{1}{c} {$\mu_\alpha$ cos($\delta$)} &
1021: \multicolumn{1}{c} {$\mu_\delta$} &
1022: \multicolumn{1}{c} {$v_r$} &
1023: \multicolumn{1}{c} {$\mu_\alpha$ cos($\delta$)} &
1024: \multicolumn{1}{c} {$\mu_\delta$} &
1025: \multicolumn{1}{c} {$R_{peri}$} &
1026: \multicolumn{1}{c} {$R_{apo}$} \\
1027: \multicolumn{1}{c} {} &
1028: \multicolumn{2}{c} {J2000} &
1029: \multicolumn{1}{c} {km s$^{-1}$} &
1030: \multicolumn{1}{c} {mas yr$^{-1}$} &
1031: \multicolumn{1}{c} {mas yr$^{-1}$} &
1032: \multicolumn{1}{c} {km s$^{-1}$} &
1033: \multicolumn{1}{c} {mas yr$^{-1}$} &
1034: \multicolumn{1}{c} {mas yr$^{-1}$} &
1035: \multicolumn{1}{c} {kpc} &
1036: \multicolumn{1}{c} {kpc}}
1037: \startdata
1038: Acheron & 15 50 24 & +9 48 39 & $-240^{+99}_{-78}$ & $-3.5 \pm 0.5$ &
1039: $-4.2 \pm 0.3$ & $136^{+104}_{-74}$ & $-12.5 \pm 0.5$ & $-13.3 \pm 0.3$ & $3.5 \pm 0.8 $& $9.2 \pm 3.3 $\\
1040: Cocytos & 16 29 21 & +26 40 8 & $ -142 \pm 10$ & $0.51 \pm 0.07$ &
1041: $-3.80 \pm 0.05$ & $-92 \pm 10$ & $-5.8 \pm 0.1$ & $-1.0 \pm 0.05$ & $ 4.9 \pm 0.2 $ & $12.5 \pm 0.2 $\\
1042: Lethe & 16 15 45 & +29 56 45 & $ -134 \pm 13$ & $+0.67 \pm 0.19$ & $ -3.46 \pm 0.04$ & $-105 \pm 13$ & $ -5.1 \pm 0.2$ & $ -0.40 \pm 0.03$ & $7.7 \pm 0.4$ & $17.3 \pm 0.5 $\\
1043: Styx & 13 56 24 & +26 48 00 & $ -42^{+44}_{-191}$ & $ -0.23^{+0.09}_{-0.17}$ & $ -0.6 \pm 0.03$ & $-22^{+182}_{-44}$ & $-1.05 \pm 0.09$ & $ -0.95 \pm 0.03$ &$13.7 \pm 3.9$ & $45^{+38}_{-1}$\\
1044: \enddata
1045: \end{deluxetable}
1046:
1047:
1048: \clearpage
1049:
1050: %% Use the figure environment and \plotone or \plottwo to include
1051: %% figures and captions in your electronic submission.
1052: %% To embed the sample graphics in
1053: %% the file, uncomment the \plotone, \plottwo, and
1054: %% \includegraphics commands
1055: %%
1056: %% If you need a layout that cannot be achieved with \plotone or
1057: %% \plottwo, you can invoke the graphicx package directly with the
1058: %% \includegraphics command or use \plotfiddle. For more information,
1059: %% please see the tutorial on "Using Electronic Art with AASTeX" in the
1060: %% documentation section at the AASTeX Web site,
1061: %% http://www.journals.uchicago.edu/AAS/AASTeX.
1062: %%
1063: %% The examples below also include sample markup for submission of
1064: %% supplemental electronic materials. As always, be sure to check
1065: %% the instructions to authors for the journal you are submitting to
1066: %% for specific submissions guidelines as they vary from
1067: %% journal to journal.
1068:
1069: %% This example uses \plotone to include an EPS file scaled to
1070: %% 80% of its natural size with \epsscale. Its caption
1071: %% has been written to indicate that additional figure parts will be
1072: %% available in the electronic journal.
1073:
1074:
1075:
1076:
1077: \end{document}
1078:
1079: