1: \documentclass[12pt,preprint]{aastex}
2:
3: %% manuscript produces a one-column, double-spaced document:
4:
5: %% \documentclass[manuscript]{aastex}
6:
7: %% preprint2 produces a double-column, single-spaced document:
8:
9: %% \documentclass[preprint2]{aastex}
10:
11: %% Sometimes a paper's abstract is too long to fit on the
12: %% title page in preprint2 mode. When that is the case,
13: %% use the longabstract style option.
14:
15: %% \documentclass[preprint2,longabstract]{aastex}
16:
17:
18: \newcommand{\msol}{\mathrm{M_{\odot}}}
19: \newcommand{\rsol}{\mathrm{R_{\odot}}}
20: \newcommand{\thco}{^{13}\mathrm{CO}}
21: \newcommand{\twco}{^{12}\mathrm{CO}}
22: \newcommand{\hii}{\mathrm{H\,\scriptstyle II}}
23: \newcommand{\kms}{{\mathrm{km \, s^{-1}}}}
24: \newcommand{\de}{{^{\circ}}}
25: \newcommand{\ls}{\mathrel{\raise0.35ex\hbox{$\scriptstyle <$}\kern-0.6em
26: \lower0.40ex\hbox{{$\scriptstyle \sim$}}}}
27: \newcommand{\dv}{|{\Delta v}|}
28:
29: \slugcomment{For submission to Ap J L }
30:
31: \shorttitle{GMCs in Spiral Arms}
32: \shortauthors{Stark and Lee}
33:
34: \begin{document}
35:
36: \title{Giant Molecular Clouds are More Concentrated
37: to Spiral Arms than Smaller Clouds
38: }
39:
40:
41: \author{Antony A. Stark}
42: \affil{Smithsonian Astrophysical Observatory, Cambridge MA 02138}
43: \email{aas@cfa.harvard.edu}
44:
45: \and
46:
47: \author{Youngung Lee}
48: \affil{Korea Astronomy Observatory, Taeduk Radio Astronomy
49: Observatory, Daejeon, Korea}
50: \email{yulee@trao.re.kr}
51:
52: \begin{abstract}
53: From our catalog of Milky Way molecular clouds,
54: created using a temperature thresholding algorithm
55: on the Bell Laboratories $\thco$ Survey, we
56: have extracted two subsets:
57: (1)
58: Giant Molecular Clouds (GMCs),
59: clouds that are definitely larger
60: than $10^5 \, \msol$, even if they are at their ``near distance'',
61: and
62: (2) clouds that are definitely smaller than
63: $10^5 \, \msol$, even if they are at their ``far distance''.
64: The positions and velocities of these clouds are compared to
65: the loci of spiral arms in $\ell,\, v$ space.
66: The velocity separation of each cloud from the
67: nearest spiral arm is introduced as a ``concentration statistic''.
68: Almost all of the GMCs are found near spiral arms.
69: The density of smaller clouds is enhanced near spiral arms,
70: but some clouds ($\sim\,10\%$) are unassociated with any
71: spiral arm.
72: The median velocity separation between
73: a GMC and the nearest spiral arm is $3.4\pm0.6\, \kms$,
74: whereas the median separation between smaller clouds
75: and the nearest spiral arm is $5.5\pm0.2\, \kms$.
76:
77: \end{abstract}
78:
79: \keywords{Galaxy: structure---ISM: clouds---ISM: molecules}
80:
81:
82: \section{Introduction}
83:
84: The Galaxy as a whole affects star formation
85: through the mechanism of spiral structure.
86: Stars in galactic disks
87: tend to develop spiral-shaped density waves that arise
88: either
89: through spontaneous instability
90: or through
91: gravitational perturbations from
92: nearby galaxies or a central bar.
93: The spiral density wave modulates the gravitational
94: potential in the disk, and the interstellar medium reacts
95: non-linearly to this varying potential,
96: gathering and concentrating the gas.
97: Giant Molecular Clouds (GMCs) form, leading to a local increase in
98: the star formation rate and the creation of giant $\hii$ regions.
99: The cloud formation process is not well understood, and
100: is the subject of
101: ongoing investigation \citep{
102: elmegreen00,pringle01,hartmann01,zhang02,elmegreen02,ostriker04}.
103:
104: In this {\em Letter}, we measure the
105: degree to which GMCs and smaller molecular clouds
106: are concentrated in the spiral arms of
107: the Milky Way, to provide a quantitative comparison with theoretical
108: models of molecular cloud formation.
109: In \S 2, we take the Bell Laboratories $\thco$ Survey data
110: \citep{lee01} and use a
111: brightness temperature thresholding algorithm to create a
112: catalog of molecular clouds.
113: From this catalog, we extract a set of clouds that
114: are definitely GMCs, and a set of clouds that are definitely
115: smaller than GMCs.
116: In \S 3, we use surveys of neutral and ionized hydrogen to
117: define the spiral arms.
118: In \S 4, we compare the locations of clouds
119: in $(\ell,\,v)$-space with the spiral arm loci.
120: We show that the GMCs and the
121: smaller clouds have significantly different distributions.
122: The GMCs are more concentrated to the spiral arms than
123: the smaller clouds.
124:
125:
126: \section{Cloud Identification and Mass Estimation}
127:
128: A catalog of clouds was generated from the
129: Bell Laboratories $\thco$ survey data
130: \citep{lee01} using a
131: thresholding technique
132: described in \citet{stark05c}.
133: This survey contains $2.3\,\times\,10^7$ pixels of $\thco$ data,
134: each $0.05\de \,\times\,0.05\de\,\times\,0.68\,\kms$ in size.
135: Survey pixels having $T_R^{*} > T_{\mathrm{th}}$
136: are identified
137: and grouped together in $(\ell, \, b, \, v)$
138: space to make a ``cloud", where all
139: the pixels constituting the cloud are above the threshold
140: and also adjacent to at least one other pixel that
141: is also above the threshold.
142: A ``cloud" is then a connected volume of pixels,
143: all of which are above the threshold.
144: Applying the thresholding method with
145: $T_{\mathrm{th}} = 1 \mathrm K $ on the
146: Bell Laboratories $\thco$ Survey
147: yields a catalog of 1,400 clouds.
148: As further described in \citet{stark05c}, we assign several
149: possible distances to each cloud based on our knowledge
150: of the velocity field of the Galaxy.
151: Essentially, the distances correspond to the
152: ``near'' and ``far'' points of
153: the distance ambiguity (cf. Mihalas 1968\nocite{mihalas68}),
154: plus some
155: additional uncertainty due to
156: random motions.
157: Each cloud has a range of possible distances.
158:
159: We want to distinguish GMCs from smaller clouds, but the
160: distances are uncertain, and so are the luminosities and estimated masses.
161: We can, however, derive a subset of the catalog that
162: contains only GMCs, and another subset that that contains
163: only clouds that are definitely smaller than a GMC.
164: For the purposes of this {\em Letter}, we define a GMC
165: as a member of our catalog with
166: $L(\thco) \,>\, 5 \times 10^3 \, \mathrm{K \, km \, s^{-1} \, pc^2}$.
167: The mass corresponding to this luminosity is
168: $1 \times 10^5\, \mathrm{\msol}$, based on the
169: mass-luminosity relation
170: $M_c = [20 \, \mathrm{\msol / K \, km \, s^{-1} \, pc^2}] L(\thco)$
171: derived
172: in \citet{stark05c}
173: for clouds identified by the method we use here.
174: For each cloud in the catalog, we calculate the
175: range of luminosities corresponding to
176: the range of possible distances.
177: If all these possible luminosities exceed the luminosity
178: threshold
179: $L(\thco) \,>\, 5 \times 10^3 \, \mathrm{K \, km \, s^{-1} \, pc^2}$,
180: the cloud is definitely a
181: GMC, regardless of the distance uncertainties.
182: Applying this criterion to the catalog yields 56 GMCs.
183: If all the possible luminosities fall below the threshold,
184: the cloud is definitely not a GMC, and
185: the cloud is included in the ``definite small cloud'' set.
186: This second criterion yields 1257 small clouds.
187: Our catalog also contains 87 clouds that do not
188: fulfill either criterion. These are moderately-large
189: clouds whose distance is not well determined, and
190: we exclude them from further consideration.
191:
192: \section{Spiral Arm Loci}
193:
194: Spiral arms in the Milky Way have been identified
195: in surveys of 21 cm atomic hydrogen and radio
196: recombination lines. Here we adopt the analyses of
197: \citet{reifenstein70},
198: \citet{burton70},
199: \citet{shane72},
200: \citet{lindblad73},
201: and
202: \citet{simonson76}, to determine the locations
203: of the spiral arms in
204: $(\ell, \, v)$ space:
205: we take locations defining the spiral arms from
206: these references, and interpolate using cubic spline functions.
207: The loci of the arms are well-determined for
208: $ 20\de \, < \, \ell \, < \, 140\de$, but not
209: for the Galactic Center Region.
210: We therefore exclude from further analysis all clouds with
211: $ \ell < 20\de$.
212: This reduces the samples to 39 GMCs and 932 smaller clouds.
213: The locations of the clouds, and the locations of the
214: spiral arms, are plotted in figure 1.
215: The spiral arms are identified by letters, as in \citet{cohen80}.
216:
217: Drawing the spiral arms onto the $(\ell, \, v)$
218: diagram of the Milky Way has been controversial. The criteria adopted
219: by the above authors for the placement of the arms
220: are subjective and ill-defined. Arms ``A'' and
221: ``B'', the local arm and the Lindblad Ring,
222: are not large-scale features of the Galaxy, but local
223: spurs that loom large because they are close to the Sun.
224: Spiral arms ``A'' through ``F'' are adopted here because
225: they are traditional and because there is no alternative; the analysis
226: below will, however, be seen to support this initial choice.
227:
228: Almost all the GMCs (red circles) lie close to
229: spiral arms. The most notable exceptions are the GMCs in
230: the bridge between the Sagittarius (C) and Scutum (D)
231: arms near $\ell = 38\de$, $v = 90\,\kms$. The local
232: spiral arm and Lindblad Ring (A and B) contain no GMCs within
233: the area of sky covered by the Bell Labs survey.
234: As noted by \citet{cohen80},
235: the Perseus arm (F) is
236: particularly distinct and well-separated from surrounding
237: material.
238:
239: \section{Concentration of Clouds to the Arms}
240:
241: There is at least one spiral arm at some velocity
242: for each value of $\ell$ in figure 1.
243: We can therefore define a ``concentration statistic'' for each
244: cloud: the absolute value of the separation in velocity
245: between that cloud and the nearest spiral arm, $\dv$.
246: The set of $\dv$ for each of our cloud samples is the
247: distribution of separations of those clouds
248: from the spiral arm loci. These distributions are plotted
249: in figure 2 for the GMC and small cloud samples.
250: The small cloud distribution has a
251: long tail, containing about 10\% of all clouds,
252: that extends to large values of $\dv$.
253: Most of these
254: are near $\ell \,\approx \,55\de, \,v\,\approx\,30\,\kms$,
255: where we have a long line of sight that falls between
256: spiral arms and includes the tangent velocity, where the
257: phase space density is high.
258: These are clear examples of interarm clouds.
259: All but five of the GMCs have $\dv \, < \, 14 \, \kms$.
260: Of those five, four are clouds
261: in the Sagittarius-Scutum bridge \citep{cohen80}.
262:
263: The Kolmogorov-Smirnov statistic \citep{press92} is the largest
264: vertical separation
265: between the two distributions, as shown in figure 2.
266: The hypothesis that the two sets of $\dv$ are drawn
267: from the same parent distribution is
268: rejected at the 94\% level by the Kolmogorov-Smirnov
269: test --- it is highly likely that the GMCs
270: are distributed differently with respect to the spiral arms
271: than are the small clouds.
272: The two distributions have significantly different medians.
273: The median $\dv$ for the
274: GMCs is $3.4\pm0.6\, \kms$,
275: whereas the median $\dv$ for the small clouds
276: is $5.5\pm0.2\, \kms$. The errors in the medians are estimated
277: by the bootstrap method \citep{efron83}.
278:
279: We will compare these results with
280: the concentration statistics of
281: objects that have the same radial distribution and
282: velocity dispersion as
283: molecular clouds but
284: are azimuthally symmetric, and therefore have no
285: concentration to spiral arms.
286: The radial distribution can be parameterized
287: by:
288: \begin{equation}
289: \rho \, \propto \, R^\alpha \,
290: \mathrm{exp}\left(-{{R}\over{R_\mathrm{s}}}\right) \, ,
291: \end{equation}
292: where $\rho$ is the surface density of objects, $R$ is radius from
293: the galactic center, and $\alpha$ and $R_\mathrm{s}$ are
294: parameters.
295: The galactic rotation curve is taken to be flat, at a
296: velocity $\Theta_0$.
297: The velocity dispersion is Gaussian, with a
298: one-dimensional root-mean-square of $\sigma_v$ \citep[e.g.,][]{stark89b}.
299: A set of Monte Carlo objects were generated by
300: the following steps:
301: (1) Choose a set of parameters, $\alpha$, $R_\mathrm{s}$, $\Theta_0$
302: and $\sigma_v$.
303: (2) Use a random number generator
304: to create a set of objects in two dimensions that have
305: a radial distribution given by equation 1.
306: (3) Calculate the $\ell$ and $v$ of each
307: object as seen from the Sun at $\rsol = 8\,\mathrm{kpc}$.
308: (4) Generate a random Gaussian deviate for each
309: object from a distribution
310: with $\sigma \, = \, \sigma_v$ and zero mean, and add it to each $v$.
311: (5) Compare the resulting set of random $(\ell, v)$
312: values to the catalog of actual molecular clouds, using
313: a two-dimensional Kolmogorov-Smirnov test \citep{press92}.
314: (6) Go back to step 1, varying parameters to maximize the
315: similarity between the random set of objects and
316: the observed catalog.
317:
318: This procedure results in the parameters
319: $\Theta_0 \, = \, 215 \, \kms$,
320: $R_\mathrm{s} \, = \, 1.7 \, \mathrm{kpc}$, $\alpha = 2.3$,
321: and $\sigma_v \,=\, 7.7\,\kms$.
322: This value of $\Theta_0$ should not be
323: taken to be a measure of the Sun's rotational velocity. If
324: the rotation curve were parameterized as
325: $\Theta(R) \,\approx\, \Theta_0 \,+\, (R/\rsol)\Theta_1 \,+\,\ldots\,$,
326: then
327: $\Theta_1$ cannot be constrained by fitting to objects
328: in the inner Galaxy: galactic kinematics looks the same
329: with a solid-body term added to the rotation
330: curve.
331: These data do not constrain $\alpha$ very well either, since
332: our cutoff at $\ell\, = \, 20\de $ eliminates much of the molecular
333: hole at $R\,\approx\,3.5\,\mathrm{kpc}$. The shape of the
334: inner edge of this hole is parameterized by $\alpha$, and
335: almost all values of $\alpha$ between 2 and 3 fit well.
336: The quick drop-off of the molecular ring to larger radii,
337: $R_\mathrm{s} \, = \, 1.7 \, \mathrm{kpc}$, is typical of
338: fits to first quadrant molecular line surveys
339: \citep{burton78,solomon79}.
340: The value of
341: $\sigma_v \,=\, 7.7\,\kms$ is consistent with the value
342: $\sigma_v \,=\, 7.8^{+0.6}_{-0.5}\,\kms$ determined for
343: moderately-large
344: clouds near the Sun by \citet{stark89b}.
345:
346: The distribution of $\dv$ values for the Monte Carlo objects
347: is plotted in green in figure 2, and their median value of $|\Delta v|$
348: is $7.81\pm0.05\,\kms$. This value is robust, in the sense that essentially
349: any azimuthally-symmetric distribution of random objects
350: in the galactic plane that we have tried will produce
351: a median value of $|\Delta v|$ that is greater than $7\,\kms$.
352: The small molecular clouds are more concentrated to spiral arms than
353: random objects, and the GMCs are even more concentrated than the
354: small clouds.
355:
356: \section{Discussion}
357:
358: It is unfortunate that we cannot observe the Milky Way from the
359: outside, and see the spatial relationship between
360: the molecular clouds and the spiral arms directly.
361: Interferometric observations of CO in other galaxies
362: having well-defined spiral structure
363: shows that the CO emission is
364: concentrated to the spiral arms \citep{murgia05}.
365: In the Milky Way, most of the CO emission is from GMCs.
366: This is illustrated in figure 3. Here
367: we have taken the catalog subset discussed in \citet{stark05c},
368: where the clouds were selected to have well-determined distances,
369: and further restricted that subset to clouds with $\ell\,>\,20\de$.
370: Figure 3 shows the distribution of virial masses for these clouds.
371: The vast majority of the CO-emitting material, over 85\%, is in the GMCs.
372: Since the CO luminosity per molecule is approximately constant
373: \citep{liszt84}, the majority of the CO luminosity will
374: also arise in GMCs.
375: It seems likely that this is true of all spiral galaxies, and
376: that the concentration of CO to spiral arms is
377: principally a concentration of GMCs to the spiral arms.
378:
379: Given the median velocity separations derived in \S4, we
380: can estimate the typical size of the spatial deviations
381: from the spiral arms.
382: The distributions of $\Delta v$ values are approximately
383: Gaussian for small values, ignoring the long tails.
384: For purposes of this estimate, we assume that
385: the spatial deviations from the arms are also Gaussian.
386: Let $\sigma_a^2$ be the Gaussian velocity variance; its
387: square root is proportional
388: to the observed median, $m$:
389: \begin{equation}
390: 2\int_0^m\,{{1}\over{\sigma_a\,\sqrt{2\pi}}}\,\mathrm{e}^{-{{v^2}\over
391: {2\,\sigma_a^2}}}\,\mathrm{d}v \,=\,
392: \mathrm{erf}\left({{m}\over{\sqrt{2}\,\sigma_a}}\right)
393: \,=\,{{1}\over{2}} \,,
394: \end{equation}
395: or $\sigma_a = m/(b\sqrt2) = 1.48 \, m$, where
396: $\mathrm{erf}(b)\,\equiv\,{{1}\over{2}}$. So
397: $\sigma_a\,\approx\,8.2\,\kms$ for the small clouds and
398: $\sigma_a\,\approx\,5.0\,\kms$ for the GMCs.
399: We know that the one-dimensional root-mean-square velocity
400: dispersions of clouds are $\sigma_v \, \approx \, 7.8 \,
401: \kms$ for moderate-sized clouds ($M\,\ls\,10^{5}\,\msol$,
402: Stark \& Brand 1989) \nocite{stark89b}
403: and
404: $\sigma_v \, \approx \, 4 \, \kms$ for GMCs ($M\,>\,10^{5.5}\,\msol$,
405: Stark \& Lee 2005). \nocite{stark05c}
406: This dispersion will, approximately, add in quadrature with
407: the amount of velocity separation caused by the spatial
408: deviations and the rotation
409: curve, $\sigma_r$:
410: \begin{equation}
411: \sigma_a^2 \,\approx\, \sigma_v^2 + \sigma_r^2 \, .
412: \end{equation}
413: A typical spatial deviation of the clouds from the spiral arm is
414: then
415: \begin{equation}
416: \Delta r \,\sim\, {{\sigma_r}
417: \over{|{{\mathrm{d}v}\over{\mathrm{d}r}}|}} \, \approx \,
418: {{\sqrt{\sigma_a^2 - \sigma_v^2}}
419: \over{|{{\mathrm{d}v}\over{\mathrm{d}r}}|}} \,\sim \,50 \, \mathrm{pc} ,
420: \end{equation}
421: where
422: ${|{{\mathrm{d}v}\over{\mathrm{d}r}}|}\,\sim\,
423: 50\,\kms\,\mathrm{kpc}^{-1}$
424: is a typical change in radial velocity with distance due to galactic
425: rotation. The estimated separation,
426: $\Delta r \,\sim\, 50\, \mathrm{pc}$, which applies to both the small
427: clouds and
428: the GMCs, is small compared to the scale of a spiral arm.
429: We do not want to make too much of this rough estimate, but
430: we can say that
431: aside from the long tail in the
432: distribution of $\Delta v$ values,
433: the data are consistent with
434: the hypothesis that the spatial concentration of most
435: molecular clouds to the spiral arms is high.
436:
437: The complete picture of the interaction of molecular clouds
438: with spiral arms is a dynamical process in six-dimensional
439: phase space. The Bell Laboratories $\thco$ Survey provides
440: statistical information about two spatial and one velocity
441: dimension. We have seen in \citet{stark05c} that, on average,
442: the GMCs are tightly concentrated to the galactic plane,
443: with a FWHM scaleheight about 30 pc. We now see that
444: these same GMCs are concentrated to the spiral arms
445: as well. Most clouds are associated
446: with spiral arms, aside from the
447: outliers that are examples of interarm clouds.
448: Among clouds near the arms, the GMCs are significantly more
449: concentrated to the arms in velocity than are the small clouds.
450:
451:
452:
453: \acknowledgments
454:
455: We thank the members of
456: the Bell Laboratories Radio Physics Research Group during the
457: ten-year period of the Bell Labs $\thco$ Survey: R. W. Wilson,
458: J. Bally, D. Mumma, W. Bent, W. Langer, G. R. Knapp,
459: and M. Pound.
460: We thank N. Tothill for comments.
461: This work was supported by Basic Research Program
462: R01-2003-000-10513-0 of KOSEF, Republic of Korea,
463: and
464: by the William Rollins Endowment Fund
465: of the Smithsonian Institution.
466:
467:
468: \begin{thebibliography}{22}
469: \expandafter\ifx\csname natexlab\endcsname\relax\def\natexlab#1{#1}\fi
470:
471: \bibitem[{{Burton} \& {Gordon}(1978)}]{burton78}
472: {Burton}, W.~B. \& {Gordon}, M.~A. 1978, \aap, 63, 7
473:
474: \bibitem[{{Burton} \& {Shane}(1970)}]{burton70}
475: {Burton}, W.~B. \& {Shane}, W.~W. 1970, in IAU Symp. 38: The Spiral Structure
476: of our Galaxy, 397
477:
478: \bibitem[{{Cohen} {et~al.}(1980){Cohen}, {Cong}, {Dame}, \&
479: {Thaddeus}}]{cohen80}
480: {Cohen}, R.~S., {Cong}, H., {Dame}, T.~M., \& {Thaddeus}, P. 1980, \apjl, 239,
481: L53
482:
483: \bibitem[{{Efron} \& {Gong}(1983)}]{efron83}
484: {Efron}, B. \& {Gong}, G. 1983, The American Statistician, 37, 36
485:
486: \bibitem[{{Elmegreen}(2000)}]{elmegreen00}
487: {Elmegreen}, B.~G. 2000, \apj, 530, 277
488:
489: \bibitem[{{Elmegreen}(2002)}]{elmegreen02}
490: ---. 2002, \apj, 577, 206
491:
492: \bibitem[{{Hartmann} {et~al.}(2001){Hartmann}, {Ballesteros-Paredes}, \&
493: {Bergin}}]{hartmann01}
494: {Hartmann}, L., {Ballesteros-Paredes}, J., \& {Bergin}, E.~A. 2001, \apj, 562,
495: 852
496:
497: \bibitem[{Lee {et~al.}(2001)Lee, Stark, Kim, \& Moon}]{lee01}
498: Lee, Y., Stark, A.~A., Kim, H.-G., \& Moon, D.-S. 2001, \apjs, 136, 137
499:
500: \bibitem[{{Lindblad} {et~al.}(1973){Lindblad}, {Grape}, {Sandqvist}, \&
501: {Schober}}]{lindblad73}
502: {Lindblad}, P.~O., {Grape}, K., {Sandqvist}, A., \& {Schober}, J. 1973, \aap,
503: 24, 309
504:
505: \bibitem[{{Liszt}(1984)}]{liszt84}
506: {Liszt}, H.~S. 1984, Comments on Astrophysics, 10, 137
507:
508: \bibitem[{{Mihalas}(1968)}]{mihalas68}
509: {Mihalas}, D. 1968, {Galactic Astronomy} (W. H. Freeman and Company)
510:
511: \bibitem[{{Murgia} {et~al.}(2005){Murgia}, {Helfer}, {Ekers}, {Blitz},
512: {Moscadelli}, {Wong}, \& {Paladino}}]{murgia05}
513: {Murgia}, M., {Helfer}, T.~T., {Ekers}, R., {Blitz}, L., {Moscadelli}, L.,
514: {Wong}, T., \& {Paladino}, R. 2005, \aap, 437, 389
515:
516: \bibitem[{{Ostriker} \& {Kim}(2004)}]{ostriker04}
517: {Ostriker}, E.~C. \& {Kim}, W.-T. 2004, in ASP Conf. Ser. 317: Milky Way
518: Surveys: The Structure and Evolution of our Galaxy, 248
519:
520: \bibitem[{{Press} {et~al.}(1992){Press}, Teukolsky, Vetterling, \&
521: Flannery}]{press92}
522: {Press}, W.~H., Teukolsky, S.~A., Vetterling, W.~T., \& Flannery, B.~P. 1992,
523: {Numerical Recipies in C} (Cambridge University Press)
524:
525: \bibitem[{{Pringle} {et~al.}(2001){Pringle}, {Allen}, \& {Lubow}}]{pringle01}
526: {Pringle}, J.~E., {Allen}, R.~J., \& {Lubow}, S.~H. 2001, \mnras, 327, 663
527:
528: \bibitem[{{Reifenstein} {et~al.}(1970){Reifenstein}, {Wilson}, {Burke},
529: {Mezger}, \& {Altenhoff}}]{reifenstein70}
530: {Reifenstein}, E.~C., {Wilson}, T.~L., {Burke}, B.~F., {Mezger}, P.~G., \&
531: {Altenhoff}, W.~J. 1970, \aap, 4, 357
532:
533: \bibitem[{{Shane}(1972)}]{shane72}
534: {Shane}, W.~W. 1972, \aap, 16, 118
535:
536: \bibitem[{{Simonson}(1976)}]{simonson76}
537: {Simonson}, S.~C. 1976, \aap, 46, 261
538:
539: \bibitem[{{Solomon} {et~al.}(1979){Solomon}, {Sanders}, \& {Scoville,
540: N.~Z.}}]{solomon79}
541: {Solomon}, P.~M., {Sanders}, D.~B., \& {Scoville, N.~Z.} 1979, \apjl, 232
542:
543: \bibitem[{Stark \& Brand(1989)}]{stark89b}
544: Stark, A.~A. \& Brand, J. 1989, \apj, 339, 763
545:
546: \bibitem[{{Stark} \& {Lee}(2005)}]{stark05c}
547: {Stark}, A.~A. \& {Lee}, Y. 2005, \apjl, 619, L159
548:
549: \bibitem[{{Zhang} {et~al.}(2002){Zhang}, {Song}, {Yang}, {Chen}, {Zhang}, {Li},
550: \& {Cui}}]{zhang02}
551: {Zhang}, T., {Song}, G., {Yang}, Z., {Chen}, L., {Zhang}, B., {Li}, N., \&
552: {Cui}, J. 2002, Chinese Astronomy and Astrophysics, 26, 276
553:
554: \end{thebibliography}
555:
556:
557:
558:
559: \clearpage
560:
561:
562:
563: \begin{figure}
564: \plotone{f1.eps}
565: \caption{Distribution in $\ell$ and $v$ of two sets
566: of clouds selected from
567: the Bell Laboratories $\thco$ Survey.
568: The area of each symbol is proportional to the
569: velocity width of the corresponding cloud.
570: The red clouds have a mass $>\,10^5\,\msol$,
571: even if they are at their near distance; the
572: blue clouds have a mass
573: $<\,10^5\,\msol$, even if they are at their far distance.
574: The green curves indicate the loci of spiral arms:
575: A and B are the local arm, C is the Sagittarius arm,
576: D is the Scutum arm, E is the 3 kpc arm, and F is the Perseus
577: arm.
578: \label{fig1}}
579: \end{figure}
580:
581: \clearpage
582:
583: \begin{figure}
584: \plotone{f2.eps}
585: \caption{Cumulative distribution of separation velocities from the
586: nearest spiral arm.
587: For each cloud in figure 1, the velocity separation, $|\Delta v|$,
588: from the
589: nearest spiral arm is determined. The two distributions of these
590: velocities for 39 GMCs and for 932 smaller clouds are plotted here.
591: The ordinate is the
592: fraction of clouds of each size having a separation
593: velocity less than the value on the abscissa.
594: The value of the Kolmogorov-Smirnov statistic is shown.
595: The green line is the distribution of separation velocities
596: for a set of azimuthally-symmetric random objects that have the
597: same radial distribution as the molecular clouds.
598: \label{fig2}}
599: \end{figure}
600:
601: \clearpage
602:
603: \begin{figure}
604: \plotone{f3.eps}
605: \caption{
606: Distribution of estimated virial masses of a cloud sample from the
607: Bell Laboratories $\thco$ Survey. This subset of
608: clouds was selected to have small distance uncertainty
609: in \citet{stark05c}. The red curves show the distribution
610: if all the clouds are at their far distance, and the blue
611: curves show the distribution if all the clouds are at their
612: near distance.
613: The lower panel shows a histogram of
614: the cloud mass distribution in logarithmic bins. The
615: coarse sampling grid of the survey may cause it to
616: miss some clouds smaller than $\sim \, 10^3 \, \msol$.
617: The upper panel shows the integrated mass distribution.
618: Over 85\% of the total mass is in GMCs ($M \, > \, 10^5 \, \msol$).
619: \label{fig3}}
620: \end{figure}
621:
622:
623: \end{document}
624:
625: %%
626: %% End of file `spiralarm.tex'.
627: