1: \documentclass[12pt,preprint]{aastex}
2: \pdfoutput=1
3:
4: \newcommand{\eqref}[1]{(\ref{#1})}
5:
6: \shorttitle{Gravitational Instability and Star Formation}
7: \shortauthors{Yang et al.}
8:
9: \begin{document}
10:
11:
12:
13: \title{Large-Scale Gravitational Instability and Star Formation\\
14: in the Large Magellanic Cloud}
15:
16:
17:
18: \author{Chao-Chin Yang, Robert A. Gruendl, and You-Hua Chu}
19: \affil{Department of Astronomy, University of Illinois, Urbana, IL~61801}
20: \email{cyang8@astro.uiuc.edu, gruendl@astro.uiuc.edu, chu@astro.uiuc.edu}
21:
22: \and
23:
24: \author{Mordecai-Mark Mac Low}
25: \affil{Department of Astrophysics, American Museum of Natural History,
26: New York, NY~10024-5192}
27: \email{mordecai@amnh.org}
28:
29: \and
30:
31: \author{Yasuo Fukui}
32: \affil{Department of Physics and Astrophysics, Nagoya University,
33: Chikusa-ku, Nagoya~464-8602, Japan}
34: \email{fukui@a.phys.nagoya-u.ac.jp}
35:
36:
37:
38: \begin{abstract}
39: Large-scale star formation in disk galaxies is hypothesized to be driven by
40: global gravitational instability.
41: The observed gas surface density is commonly used to compute the strength of
42: gravitational instability, but according to this criterion star formation often
43: appears to occur in gravitationally stable regions.
44: One possible reason is that the stellar contribution to the instability has
45: been neglected.
46: We have examined the gravitational instability of the Large Magellanic Cloud
47: (LMC) considering the gas alone, and considering the combination of collisional
48: gas and collisionless stars.
49: We compare the gravitationally unstable regions with the on-going star
50: formation revealed by \emph{Spitzer} observations of young stellar objects.
51: Although only 62\% of the massive young stellar object candidates are in
52: regions where the gas alone is unstable, some 85\% lie in regions unstable due
53: to the combination of gas and stars.
54: The combined stability analysis better describes where star formation occurs.
55: In agreement with other observations and numerical models, a small fraction of
56: the star formation occurs in regions with gravitational stability parameter
57: $Q > 1$.
58: We further measure the dependence of the star formation timescale on the
59: strength of gravitational instability, and quantitatively compare it to the
60: exponential dependence expected from numerical simulations.
61: \end{abstract}
62:
63:
64:
65: \keywords{Galaxies: kinematics and dynamics
66: --- galaxies: ISM
67: --- galaxies: stellar content
68: --- Magellanic Clouds
69: --- stars: formation}
70:
71:
72:
73: \section{INTRODUCTION}
74:
75: Observations of nearby galaxies suggest that there exists a gas surface density
76: threshold for star formation in a galactic disk \citep{rK89,MK01}.
77: This threshold is hypothesized to be formed by the threshold for
78: gravitational stability of the disk, although alternative explanations
79: have been offered \citep[e.g.,][]{jS04}.
80: Gravitationally unstable regions are commonly identified by the Toomre
81: criterion \citep{aT64,GL65}, which considers the stability of a
82: single-component disk against axisymmetric perturbations.
83: However, this criterion, when applied to observed gas surface
84: densities, does not always
85: correctly predict the locations of star formation \citep[e.g.,][]{MK01,WB02}.
86: Stellar contribution to the gravitational instability could be important;
87: for example, the giant molecular clouds in the Large Magellanic Cloud (LMC)
88: have been shown to correlate with stellar surface density, indicating the
89: increase of star formation towards regions of higher stellar gravity
90: \citep{yF07}.
91:
92: To crudely include the stellar contribution, \citet{JS84} derived the stability
93: criterion for a two-fluid disk.
94: This stability criterion differs from the Toomre criterion by only a linear
95: correction factor, when the gas-to-stellar ratios of surface density and
96: velocity dispersion are constant across a galaxy, or when the gas and stellar
97: components are only weakly coupled \citep{WS94}.
98: To correctly include the stellar contribution, \citet{cG92} and
99: \citet{rR01} analyzed the
100: stability of a composite disk consisting of a collisional gas component and a
101: population of collisionless stellar components.
102: The resulting instability criterion has been used to explain the dust
103: morphology of edge-on disk galaxies \citep*{DY04}.
104:
105: We study the relationship between gravitational stability and star
106: formation in the LMC.
107: The LMC is chosen because high linear resolution can be obtained
108: \citep[1\arcsec\ corresponds to 0.25~pc at 50~kpc,][]{mF99}, its internal and
109: foreground extinctions are small, and its moderate inclination allows
110: spatial and kinematical mapping of the disk with minimum confusion along the
111: line of sight.
112: Both stars and the interstellar medium (ISM) of the LMC have been extensively
113: surveyed.
114: Recent \emph{Spitzer Space Telescope} observations have been used to identify
115: massive young stellar objects (YSOs), which mark the sites of current star
116: formation (R.~A. Gruendl et al., in preparation; B.~A. Whitney et al.,
117: in preparation).
118: Massive YSOs provide better probes than \ion{H}{2} regions to study the
119: relationship between the ISM and star formation because their stellar energy
120: feedback has not significantly altered the distributions and physical
121: conditions of the ambient ISM.
122:
123: In this paper, we investigate the relationship between global star formation
124: and gravitational instability in the LMC using two different stability
125: analyses: assuming that the gaseous disk is decoupled from the stellar disk and
126: considering only the gravity of the gas (\S~\ref{S:Q_g}), or assuming a
127: collisionless stellar disk and a collisional gas disk and using the
128: \citet{rR01} stability criterion (\S~\ref{S:Qsg}).
129: The results are discussed in \S~\ref{S:disc}.
130:
131:
132:
133: \section{GRAVITATIONAL INSTABILITY} \label{S:GI}
134:
135: \subsection{Gas Alone} \label{S:Q_g}
136: A thin, differentially rotating gaseous disk is unstable against
137: axisymmetric perturbations when
138: \begin{equation} \label{E:Q_g}
139: Q_g \equiv \frac{\kappa c_g}{\pi G\Sigma_g} < 1,
140: \end{equation}
141: where $Q_g$ is generally referred to as the Toomre $Q$ parameter for a
142: gaseous disk, $\kappa$ is the epicyclic frequency, $c_g$ is the isothermal
143: sound speed of the gas, $G$ is the gravitational constant, and
144: $\Sigma_g$ is the unperturbed surface density of the gas \citep{GL65}.
145: %\citep[see][]{BT87}.
146:
147: To apply this analysis to a disk galaxy, approximations need to be made.
148: First, rather than consisting of a single isothermal component, the real ISM
149: consists of gas with temperatures ranging from 10~K to $10^7$~K in multiple
150: phases including cold atomic and molecular, warm atomic and ionized, and hot
151: ionized gas.
152: However, the gas mass is dominated by cold atomic and molecular material in
153: turbulent motion, except possibly in star-forming regions where much of the ISM
154: is ionized.
155: Therefore, we consider only the neutral atomic and molecular gas mass.
156: The atomic medium of galaxies generally has nearly constant velocity dispersion
157: \citep*{VS82,SV84,DH90,PR07}, and has a vertical thickness that is well modeled
158: by assuming hydrostatic equilibrium with the gravitational potential at the
159: velocity dispersion \citep[e.g.,][]{sM95}.
160: We assume the gas disk of the LMC has an effective sound speed representative
161: of the velocity dispersion $c_g = 5$~km~s$^{-1}$ \citep*{DB06,PR07}.
162: Second, $\Sigma_g$ in equation~\eqref{E:Q_g} is the \emph{unperturbed} gas
163: surface density, while the observed gas surface density is \emph{perturbed}.
164: The surface density of a gravitationally unstable region grows exponentially as
165: it collapses; thus, using the observed surface density in
166: equation~\eqref{E:Q_g} will correctly diagnose gravitationally unstable
167: regions, but $Q_g$ may be underestimated if collapse has set in.
168: Therefore, we adopt the observed gas surface density for $\Sigma_g$ in
169: equation~\eqref{E:Q_g} to identify gravitationally unstable regions.
170:
171: The neutral atomic and molecular components of the ISM in the LMC have been
172: well surveyed.
173: For the neutral atomic component, we use the \ion{H}{1} column density map
174: derived from the combined data set of the Australia Telescope Compact Array and
175: the Parkes multibeam receiver \citep{KS03}.
176: This \ion{H}{1} map has a resolution of $\sim$1\arcmin, or 15~pc.
177: For the molecular component, we use the NANTEN CO survey of the LMC conducted
178: by \citet{FM99,FM01}.
179: These maps have a resolution of $\sim$2\farcm6, or 40~pc.
180: We scale maps of integrated column density by the CO-to-H$_2$ conversion
181: factor, $X = 5.4\times10^{20}$~H$_2$ atoms cm$^{-2}$ (K km s$^{-1}$)$^{-1}$
182: \citep{BF07} to obtain the H$_2$ mass distribution.
183: Both the \ion{H}{1} and H$_2$ maps are clipped at the 1$\sigma$ level
184: to prevent negative densities and minimize the effects of noise.
185: Furthermore, we account for He and heavier elements by assuming that
186: the mass fraction of H is $X \sim 0.7$.
187: From the \ion{H}{1} map, we find a total mass for the neutral atomic component
188: of $5.1\times10^8~M_\sun$, consistent with the value found by \citet{KS98}.
189: For the molecular component, we find a total mass of $6.6\times10^7~M_\sun$,
190: which is on the high side of the 4--7$\times10^7~M_\sun$ given by
191: \citet{nM99,nM01}, as we used a higher CO-to-H$_2$ conversion factor and
192: included He and heavy elements.
193: We add these two maps to obtain a map of total gas surface density with a
194: resolution of 40~pc pixel$^{-1}$.
195: The resulting map is shown in Figure~\ref{Fi:maps}a.
196:
197: The remaining parameter needed in equation~\eqref{E:Q_g} is the
198: epicyclic frequency $\kappa$, defined by
199: \begin{equation}
200: \kappa^2 \equiv \frac{2V^2}{R^2}\left(1 + \frac{R}{V}\frac{dV}{dR}\right),
201: \end{equation}
202: where $R$ is the galactocentric distance and $V = V(R)$ is the circular
203: velocity as a function of $R$.
204: To calculate $\kappa$ at a given $R$ in the LMC, we use the rotation curve
205: derived by \citet{KS98}.
206: The rotation curve is a best-fit to the \ion{H}{1} and the carbon star
207: \citep{KD97} measurements inside and outside of about 3.2~kpc, respectively.
208: We find the derivatives of the circular velocity and thus the epicyclic
209: frequency by central finite differences with an increment of 250~pc, then we
210: use a cubic spline to interpolate the discrete results to a continuous domain.
211: Finally, the center and the orientation of the disk of the LMC need to be
212: specified.
213: To be consistent with the \ion{H}{1} kinematics, we adopt the same center and
214: orientation determined by \citet{KS98}.
215: The kinematic center of the LMC disk is at $\alpha = 05^{\rm h}$17\fm6 and
216: $\delta = -69$\arcdeg02\arcmin\ (J2000) with an inclination of $i = 33\degr$
217: and a line of nodes at a P.A. of $-12\degr$.
218: Using this geometry, the total gas surface density map is deprojected to
219: determine the galactocentric distance of each pixel, allowing us to construct a
220: map of $Q_g$ calculated from equation~\eqref{E:Q_g}.
221: This map is shown in Figure~\ref{Fi:maps}b.
222:
223: The contours in Figure~\ref{Fi:maps}b delineate the critical boundary
224: $Q_g = 1$ that surrounds regions gravitationally unstable due to gas alone.
225: Also marked in the figure are candidate massive YSOs from R.~A. Gruendl et al.
226: (in preparation), indicating regions of on-going star formation.
227: In the top panel of Figure~\ref{Fi:hist}, we show the number distribution of
228: YSO candidates and the corresponding cumulative fraction with respect to $Q_g$.
229: While $\sim$62\% (153 of 245) of the YSO candidates do lie in regions with
230: $Q_g < 1$, the rest are distributed in regions with values as high as $Q_g \sim
231: 3$.
232:
233: Some of the candidates with $Q_g < 1$ might be argued not to result from
234: large-scale gravitational instabilities because the unstable regions in which
235: they reside are too small.
236: If we exclude regions $\la 100$~pc in diameter, only 47\% (116) of the YSO
237: candidates would be associated with gravitationally unstable regions by this
238: criterion.
239: A frequently used empirical correction for the contributions from the stellar
240: component \citep{JS84,WS94} and the effect of the disk scale height
241: \citep{JS84} is to raise the critical value to $Q_g \sim 1.4$
242: \citep{rK89,MK01}.
243: Even if we apply this linear correction factor to the Toomre criterion,
244: $\sim$25~\% of the YSO candidates still appear to reside in stable regions.
245: Considering the stability of only the gas disk or applying a simple linear
246: correction to the Toomre criterion may not be sufficient to account for star
247: formation activity.
248:
249: \placefigure{Fi:hist}
250:
251:
252:
253: \subsection{Gas and Stars Together} \label{S:Qsg}
254:
255: We now also consider the contribution of stars to gravitational instability.
256: To include them, we follow Rafikov's (2001) treatment of a disk galaxy
257: consisting of a collisional gas disk and a collisionless stellar disk.
258: The instability condition becomes
259: \begin{equation} \label{E:Qsg}
260: \frac{1}{Q_{sg}}
261: \equiv \frac{2}{Q_s}\frac{1}{q}\left[1 - e^{-q^2}I_0\left(q^2\right)\right] +
262: \frac{2}{Q_g}R\frac{q}{1 + q^2 R^2}
263: > 1.
264: \end{equation}
265: In equation~\eqref{E:Qsg}, $Q_g$ is the stability parameter for the gas derived
266: by \citet{GL65}, as defined in equation~\eqref{E:Q_g}, while
267: \begin{equation}
268: Q_s \equiv \frac{\kappa\sigma_s}{\pi G\Sigma_s}
269: \end{equation}
270: is the stability parameter for the stars derived by \citet{aT64}, where
271: $\sigma_s$ is the stellar radial velocity dispersion, $\Sigma_s$ is the stellar
272: surface density;
273: $R \equiv c_g / \sigma_s$ and $q \equiv k\sigma_s / \kappa$ are two
274: dimensionless parameters, with $k$ being the wavenumber of the axisymmetric
275: perturbations;
276: and $I_0$ is the Bessel function of order zero.
277:
278: Similarly to our procedure in \S~\ref{S:Q_g}, we use local observed
279: values to evaluate $Q_{sg}$.
280: To estimate the stellar surface density $\Sigma_s$, we use the number density
281: of red giant branch (RGB) and asymptotic giant branch (AGB) stars, as they are
282: part of the old stellar population and therefore trace the overall mass
283: distribution of the stellar disk.
284: RGB and AGB stars are luminous and distinctive in the $[J - K_s]$ vs.\ $K_s$
285: color-magnitude diagram and much less confused by faint background galaxies or
286: Galactic stars.
287: To select these stars, we follow a procedure similar to that outlined by
288: \citet{rV01} but use only the Two Micron All Sky Survey Point Source Catalogue
289: \citep{mS06} and the criteria: $[J - K_s] < (22 - [K_s]) / 10.5$ and
290: $[K_s] < 14.5$.
291: The source counts are then binned with a resolution of 40~pc pixel$^{-1}$ and
292: Gaussian smoothed with a dispersion of 100~pc.
293: The smoothing scale is chosen to minimize pixel-to-pixel variations due to the
294: small number density of the tracer RGB and AGB stars, while retaining good
295: spatial resolution for the analysis; this scale is also small compared to the
296: large-scale gravitational instability in the analysis.
297: In the resulting map, a faint Galactic background with a gradient is present.
298: Therefore, we fit and subtract this background with exponential and linear
299: dependencies on the Galactic latitude and longitude, respectively.
300: Finally, by adopting the total stellar mass of $2 \times 10^9~M_\sun$
301: estimated by \cite{KS98}, we normalize this map to obtain the absolute stellar
302: surface density map shown in Figure~\ref{Fi:maps}c
303: \citep[cf. Fig.~2 of][]{rV01}.
304: We adopt a constant stellar radial velocity dispersion of 15~km~s$^{-1}$, after
305: considering the velocity dispersions of carbon stars
306: \citep{KD97,AN00,GG00,VA02}, red supergiants \citep{PM89,OM07}, and young
307: globular clusters \citep{FI83}.
308:
309: With these values assessed, the parameter $Q_{sg}$ defined in
310: equation~\eqref{E:Qsg} now depends only on the perturbation wavenumber $k$, or
311: equivalently, the perturbation wavelength $\lambda = 2\pi / k$ at
312: each pixel (see Fig.~\ref{Fi:qsgwl}).
313: Since $Q_{sg}$ decreases with the degree of gravitational instability and our
314: objective is to locate gravitationally unstable regions, we define its value at
315: each pixel as its global minimum with respect to $\lambda$.
316: The derived $Q_{sg}$ map of the LMC is shown in Figure~\ref{Fi:maps}d, with the
317: identified YSOs overlaid.
318:
319: Each point is unstable to perturbations over a finite range of wavelengths, as
320: shown by Figure~\ref{Fi:qsgwl}.
321: If the size of a gravitationally unstable region is smaller than the minimum
322: unstable wavelength $\lambda_\mathrm{min}$, no perturbation inside the region
323: can have long enough wavelength to drive an instability.
324: Therefore, we also find $\lambda_\mathrm{min}$ in each pixel, and compare this
325: value with the size of each region containing YSO candidates by visual
326: inspection.
327: The contours of $Q_{sg} = 1$ are plotted over grayscale presentations of
328: $\lambda_\mathrm{min}$ in Figure~\ref{Fi:lmin} so that the sizes of the
329: $Q_{sg} < 1$ regions can be compared with the minimum unstable wavelengths
330: directly.
331:
332: \placefigure{Fi:qsgwl}
333:
334: Figure~\ref{Fi:maps}d shows that, because of the stellar contribution, a much
335: larger fraction of the galaxy is unstable than would have been deduced just
336: from the gas contribution shown in Figure~\ref{Fi:maps}b.
337: The vast majority of the massive YSO candidates are in fact located within
338: gravitationally unstable regions, in sharp contrast to the results in
339: \S~\ref{S:Q_g} where only the gas is considered.
340: Particularly noticeable is the vicinity of the LMC bar region, where on-going
341: star formation is observed.
342: This bar region, having low gas surface densities and high epicyclic
343: frequencies, is stable when only the gas disk is considered, but becomes
344: unstable when the stellar component is added.
345: The bottom panel of Figure~\ref{Fi:hist} shows the number distribution of YSO
346: candidates and the corresponding cumulative fraction with respect to $Q_{sg}$.
347: About 86\% (212 of 245) of the YSO candidates are located in regions where
348: $Q_{sg} < 1$.
349: Figure~\ref{Fi:lmin} shows 12 of the 212 YSOs in probably stable regions having
350: $Q_{sg} < 1$ but region size $< \lambda_\mathrm{min}$.
351: Removing these 12 still leaves 82\% of the YSOs whose formation is attributed
352: to gravitational instability of the disk of stars and gas.
353: One caution is that finite thickness of the disk may lower the critical
354: $Q_{sg}$ value slightly, depending on the scale height of the disk relative to
355: the unstable wavelengths \citep{JS84}.
356: Given an estimated scale height of $\sim$180~pc \citep{KD99} and
357: kiloparsec-scale unstable wavelengths in the LMC, this effect may be small.
358: The results demonstrate that the stellar contribution to gravitational
359: instability can be significant.
360: Gravitational instability of the full disk appears responsible for most
361: large-scale star formation activity.
362:
363: \placefigure{Fi:lmin}
364:
365:
366:
367: \section{DISCUSSION AND CONCLUSIONS} \label{S:disc}
368:
369: \citet{jS04} has argued that a constant gas surface density threshold is a good
370: indicator in predicting the edge of the star forming disk.
371: He suggested that, for the LMC disk, this threshold is
372: $\sim$4~$M_\sun$ pc$^{-2}$.
373: Figure~\ref{Fi:maps}a shows the contours of 4~$M_\sun$ pc$^{-2}$ in total
374: neutral gas surface density.
375: The contours encompass the majority of the LMC disk except the central cavities
376: of the supergiant shells LMC-4 and LMC-8 \citep{jM80}.
377: Clearly, all the massive YSO candidates are located where the gas surface
378: density is greater than 4~$M_\sun$ pc$^{-2}$, but this constant surface density
379: threshold does not account for the distribution of the star forming sites,
380: especially in the inner parts of the LMC disk.
381: Therefore, Schaye's (2004) constant surface density threshold is a necessary
382: but not sufficient condition for star formation.
383: Detailed analysis of the large-scale gravitational instability, as shown in
384: \S~\ref{S:GI}, is still needed.
385:
386: Another measure of the response of star formation to disk instability is the
387: relation between $Q_{sg}$ and the effective timescale for star formation
388: $\tau_{sf}$.
389: This can be approximated if we assume the average lifetime of massive YSOs is
390: constant.
391: We normalize the number of YSOs, $N_*$, within a region having a given range of
392: $Q_{sg}$ by the area, as represented by the total number of pixels
393: $N_\mathrm{pix}$ within that region.
394: Figure~\ref{Fi:rate} presents $N_* / N_\mathrm{pix}$, which is $\propto
395: \tau_{sf}^{-1}$, for a number of $Q_{sg}$ bins.
396: The errors in the ratio are estimated assuming Poisson statistics on the number
397: of YSOs and number of pixels, while the horizontal error bars show the size of
398: the bins.
399: A clear log-linear relationship is found, implying that the response of star
400: formation depends exponentially on the value of $Q_{sg}$.
401: \citet*{LM05} also found an exponential dependence
402: $\tau_{sf} \propto \exp(\alpha Q_{sg})$, with $\alpha = 4.2 \pm 0.3$ from
403: global numerical simulations of galactic disks.
404: A linear fit to the data in Figure~\ref{Fi:rate} gives a slope that can be
405: interpreted as a local value of $\alpha = 2.7 \pm 0.2$.
406: This behavior appears to be a general property of nonlinear gravitational
407: instability.
408:
409:
410: \placefigure{Fi:rate}
411:
412:
413: We have made the simple assumption that the structure of the LMC is a
414: thin, flat disk with negligible stellar halo. In reality, tidal features due
415: to galaxy-galaxy interactions do exist in the outer parts of the disk
416: \citep[e.g.,][]{rV01}. Furthermore, at some locations, there
417: exist multiple \ion{H}{1} velocity components, which we have simply
418: summed to obtain total gas surface density.
419: Other uncertainties stem from our assumed constant gas effective sound
420: speed (or velocity dispersion) and
421: stellar radial velocity dispersion.
422: These affect the relative contribution of each component to the gravitational
423: instability: the higher the local velocity dispersion, the less
424: contribution a component has.
425: The sound speed of the gas we adopt, 5~km s$^{-1}$, is a typical average for
426: the neutral gas.
427: The radial velocity dispersion of stars is also uncertain.
428: As gravitational instability is much more sensitive to the gas component due to
429: its low sound speed, it is less affected by the uncertainty in the stellar
430: velocity dispersion.
431: Yet another effect to consider is the scale height of the disk, which tends to
432: stabilize the disk \citep{aT64,JS84}.
433: Given all these uncertainties, nevertheless, the exponential dependence of star
434: formation rate on $Q_{sg}$ that we find appears to be robust.
435:
436: From our analysis of the gravitational instability of the LMC, we conclude that
437: the contribution of the stellar disk to gravitational instability
438: cannot be ignored \citep{JS84,cG92,rR01}. Taking it into account in
439: the case of the LMC, we find that $\sim$85\% of the massive YSO
440: candidates lie in gravitationally unstable regions, implying that star
441: formation occurs predominantly in these regions. It
442: appears that gravitational instability of the disk drives most
443: large-scale star formation, as proposed by \citet{bE02}, \citet{aK03}, and
444: \citet{LM05}.
445:
446:
447: \acknowledgments
448: This research was supported by NASA grants JPL1264494 and JPL1290956,
449: and NSF grant AST03-07854.
450: We would like to thank Dr.~Charles Gammie for his feedback and discussions in
451: this work.
452: We also thank the anonymous referee for the prompt review and useful
453: suggestions.
454:
455:
456: \begin{thebibliography}{}
457:
458: \bibitem[Alves \& Nelson(2000)]{AN00}
459: Alves, D.~R., \& Nelson, C.~A.\ 2000, \apj, 542, 789
460:
461: \bibitem[Blitz et al.(2007)]{BF07}
462: Blitz, L., Fukui, Y., Kawamura, A., Leroy, A., Mizuno, N., \&
463: Rosolowsky, E.\ 2007, in Protostars and Planets V,
464: ed. B. Reipurth, D. Jewitt, \& K. Keil (Tucson: Univ. of Arizona Press), 81
465: (astro-ph/0602600)
466:
467: \bibitem[Dalcanton et al.(2004)Dalcanton, Yoachim, \& Bernstein]{DY04}
468: Dalcanton, J.~J., Yoachim, P., \& Bernstein, R.~A.\ 2004, \apj, 608, 189
469:
470: \bibitem[Dib et al.(2006)Dib, Bell, \& Burkert]{DB06}
471: Dib, S., Bell, E., \& Burkert, A.\ 2006, \apj, 638, 797
472:
473: \bibitem[Dickey et al.(1990)Dickey, Hanson, \& Helou]{DH90}
474: Dickey, J.~M., Hanson, M.~M., \& Helou, G.\ 1990, \apj, 352, 522
475:
476: \bibitem[Elmegreen(2002)]{bE02}
477: Elmegreen, B.~G.\ 2002, \apj, 577, 206
478:
479: \bibitem[Feast(1999)]{mF99}
480: Feast, M.\ 1999, in IAU Symp.~190, New Views of the Magellanic Clouds,
481: ed. Y.-H. Chu, N.~B. Suntzeff, J.~E. Hesser \& D.~A. Bohlender
482: (San Francisco: ASP), 542
483:
484: \bibitem[Freeman et al.(1983)]{FI83}
485: Freeman, K.~C., Illingworth, G., \& Oemler, A., Jr.\ 1983, \apj, 272, 488
486:
487: \bibitem[Fukui(2007)]{yF07}
488: Fukui, Y.\ 2007, in IAU Symp.~237, Triggered Star Formation in a Turbulent ISM,
489: ed. B.~G. Elmegreen \& J. Palous (Cambridge: Cambridge Univ. Press), 31
490:
491: \bibitem[Fukui et al.(2001)]{FM01}
492: Fukui, Y., Mizuno, N., Yamaguchi, R., Mizuno, A., \& Onishi, T.\ 2001,
493: \pasj, 53, L41
494:
495: \bibitem[Fukui et al.(1999)]{FM99}
496: Fukui, Y., et al.\ 1999, \pasj, 51, 745
497:
498: \bibitem[Gammie(1992)]{cG92}
499: Gammie, C.~F.\ 1992, Ph.D.~Thesis, Princeton University
500:
501: \bibitem[Goldreich \& Lynden-Bell(1965)]{GL65}
502: Goldreich, P., \& Lynden-Bell, D.\ 1965, \mnras, 130, 97
503:
504: \bibitem[Graff et al.(2000)]{GG00}
505: Graff, D.~S., Gould, A.~P., Suntzeff, N.~B., Schommer, R.~A., \&
506: Hardy, E.\ 2000, \apj, 540, 211
507:
508: \bibitem[Jog \& Solomon(1984)]{JS84}
509: Jog, C.~J., \& Solomon, P.~M.\ 1984, \apj, 276, 114
510:
511: \bibitem[Kennicutt(1989)]{rK89}
512: Kennicutt, R.~C., Jr.\ 1989, \apj, 344, 685
513:
514: \bibitem[Kim et al.(1998)]{KS98}
515: Kim, S., Staveley-Smith, L., Dopita, M.~A., Freeman, K.~C., Sault, R.~J.,
516: Kesteven, M.~J., \& McConnell, D.\ 1998, \apj, 503, 674
517:
518: \bibitem[Kim et al.(1999)]{KD99}
519: Kim, S., Dopita, M.~A., Staveley-Smith, L., \& Bessell, M.~S.\ 1999,
520: \aj, 118, 2797
521:
522: \bibitem[Kim et al.(2003)]{KS03}
523: Kim, S., Staveley-Smith, L., Dopita, M.~A., Sault, R.~J., Freeman, K.~C.,
524: Lee, Y., \& Chu, Y.-H.\ 2003, \apjs, 148, 473
525:
526: \bibitem[Kravtsov(2003)]{aK03}
527: Kravtsov, A.~V.\ 2003, \apjl, 590, L1
528:
529: \bibitem[Kunkel et al.(1997)]{KD97}
530: Kunkel, W.~E., Demers, S., Irwin, M.~J., \& Albert, L.\ 1997, \apjl, 488, L129
531:
532: \bibitem[Li et al.(2005)Li, Mac Low, \& Klessen]{LM05}
533: Li, Y., Mac Low, M.-M., \& Klessen, R.~S.\ 2005, \apj, 626, 823
534:
535: \bibitem[Malhotra(1995)]{sM95}
536: Malhotra, S.\ 1995, \apj, 448, 138
537:
538: \bibitem[Martin \& Kennicutt(2001)]{MK01}
539: Martin, C.~L., \& Kennicutt, R.~C., Jr.\ 2001, \apj, 555, 301
540:
541: \bibitem[Meaburn(1980)]{jM80}
542: Meaburn, J.\ 1980, \mnras, 192, 365
543:
544: \bibitem[Mizuno et al.(1999)]{nM99}
545: Mizuno, N., et al.\ 1999, in Star Formation 1999, ed. T.~Nakamoto
546: (Nobeyama:Nobeyama Radio Observatory), 56
547:
548: \bibitem[Mizuno et al.(2001)]{nM01}
549: Mizuno, N., et al.\ 2001, \pasj, 53, 971
550:
551: \bibitem[Olsen \& Massey(2007)]{OM07}
552: Olsen, K.~A.~G., \& Massey, P.\ 2007, \apjl, 656, L61
553:
554: \bibitem[Petric \& Rupen(2007)]{PR07}
555: Petric, A. O., \& Rupen, M. P.\ 2007, AJ, in press (astro-ph/0704.0279)
556:
557: \bibitem[Prevot et al.(1989)]{PM89}
558: Prevot, L., Martin, N., \& Rousseau, J.\ 1989, \aap, 225, 303
559:
560: \bibitem[Rafikov(2001)]{rR01}
561: Rafikov, R.~R.\ 2001, \mnras, 323, 445
562:
563: \bibitem[Schaye(2004)]{jS04}
564: Schaye, J.\ 2004, \apj, 609, 667
565:
566: \bibitem[Shostak \& van der Kruit(1984)]{SV84}
567: Shostak, G.~S., \& van der Kruit, P.~C.\ 1984, \aap, 132, 20
568:
569: \bibitem[Skrutskie et al.(2006)]{mS06}
570: Skrutskie, M.~F., et al.\ 2006, \aj, 131, 1163
571:
572: \bibitem[Toomre(1964)]{aT64}
573: Toomre, A.\ 1964, \apj, 139, 1217
574:
575: \bibitem[van der Kruit \& Shostak(1982)]{VS82}
576: van der Kruit, P.~C., \& Shostak, G.~S.\ 1982, \aap, 105, 351
577:
578: \bibitem[van der Marel(2001)]{rV01}
579: van der Marel, R.~P.\ 2001, \aj, 122, 1827
580:
581: \bibitem[van der Marel et al.(2002)]{VA02}
582: van der Marel, R.~P., Alves, D.~R., Hardy, E., \& Suntzeff, N.~B.\ 2002, \aj,
583: 124, 2639
584:
585: \bibitem[Wang \& Silk(1994)]{WS94}
586: Wang, B., \& Silk, J.\ 1994, \apj, 427, 759
587:
588: \bibitem[Wong \& Blitz(2002)]{WB02}
589: Wong, T., \& Blitz, L.\ 2002, \apj, 569, 157
590:
591: \end{thebibliography}
592:
593: \clearpage
594:
595: \begin{figure}
596: \plotone{f1}
597: \caption{(a) Total gas surface density distribution in the Large Magellanic
598: Cloud. Only neutral species are included in this map. The grayscale runs
599: linearly from 0 to 100~$M_\sun$ pc$^{-2}$. The contour lines show a
600: surface density of 4~$M_\sun$ pc$^{-2}$.
601: (b) Comparison between massive star-forming sites traced by
602: young stellar object candidates (\emph{red dots}) and regions
603: where the gas alone is gravitationally unstable.
604: The Toomre parameter for the gas $Q_g$ is shown in grayscale, running
605: logarithmically from 5.0 to 0.2.
606: The \emph{solid} lines delineate the critical value of $Q_g = 1$ inside
607: which the regions are gravitationally unstable.
608: (c) Total stellar surface density distribution.
609: The grayscale runs linearly from 0 to 200~$M_\sun$ pc$^{-2}$.
610: (d) As in (b), but for regions where the gas and stars together
611: are gravitationally unstable, rather than just the gas alone.}
612: \label{Fi:maps}
613: \end{figure}
614:
615: \clearpage
616:
617: \begin{figure}
618: \plotone{f2}
619: \caption{Number distribution of YSO candidates with respect to the Toomre
620: stability parameter of the pixel in which they are located.
621: The \emph{top} and \emph{bottom} panels show respectively the Toomre
622: parameters for the gas alone $Q_g$, and the stars and gas together
623: $Q_{sg}$.
624: The \emph{solid} lines show the cumulative fractions of YSOs with
625: decreasing $Q$.
626: The vertical \emph{dotted} lines denote the critical value $Q = 1$.}
627: \label{Fi:hist}
628: \end{figure}
629:
630: \clearpage
631:
632: \begin{figure}
633: \plotone{f3}
634: \caption{Schematic plot of the $Q_{sg}$ parameter as a function of
635: wavelength $\lambda$ of radial perturbations.
636: The disk of stars and gas is stable against perturbation of a given
637: $\lambda$ when $Q_{sg} > 1$; otherwise, it is unstable.
638: We measure the value of $Q_{sg}$ at its minimum in each pixel, occurring
639: at $\lambda = \lambda_\mathrm{crit}$ (point A).
640: We also find the minimum wavelength $\lambda_\mathrm{min}$ such that the
641: disk is marginally unstable (point B).}
642: \label{Fi:qsgwl}
643: \end{figure}
644:
645: \clearpage
646:
647: \begin{figure}
648: \plotone{f4}
649: \caption{Map of the minimum unstable wavelength $\lambda_\mathrm{min}$,
650: shown in grayscale.
651: The solid contours delineate the critical boundary of $Q_{sg} = 1$, while
652: the red dots mark the massive YSO candidates.
653: $\lambda_\mathrm{min}$ is defined by $\lambda_\mathrm{crit}$
654: (see Fig.~\ref{Fi:qsgwl}) when $Q_{sg} > 1$.}
655: \label{Fi:lmin}
656: \end{figure}
657:
658: \clearpage
659:
660: \begin{figure}
661: \plotone{f5}
662: \caption{Ratio of the number of YSO candidates $N_*$ to the number of
663: pixels $N_\mathrm{pix}$ in each $Q_{sg}$ bin, denoted by squares.
664: The vertical error bars are estimated by Poisson statistics,
665: while the horizontal error bars show the bin sizes.
666: The \emph{solid} line is the best-fit, of slope $-2.7 \pm 0.2$.}
667: \label{Fi:rate}
668: \end{figure}
669:
670: \end{document}
671:
672: