1: %\documentclass[preprint2]{aastex}
2: %\documentclass[12pt,preprint]{aastex}
3: \documentclass{emulateapj}
4:
5: %\usepackage{amssymb}
6: \usepackage{amsmath}
7: \usepackage{graphicx}
8:
9: \newcommand{\lO} {\lambda_{\Omega}}
10: \newcommand{\led} {\lambda_{e}}
11: \newcommand{\cAi} {c_{Ai}}
12: \newcommand{\cAn} {c_{An}}
13: \newcommand{\RM} {R_{M}}
14: \newcommand{\mbfB}{\mathbf{B}}
15: \newcommand{\mbfu}{\mathbf{u}}
16: \newcommand{\mbfui}{\mathbf{u}_i}
17: \newcommand{\mbfun}{\mathbf{u}_n}
18: \newcommand{\mbfz}{\mathbf{z}}
19: \newcommand{\mbfn}{\mathbf{n}}
20: \newcommand{\mbfk}{\mathbf{k}}
21: \newcommand{\mbfx}{\mathbf{x}}
22: \newcommand{\mbfv}{\mathbf{v}}
23: \newcommand{\mbfnabla}{\mathbf{\nabla}}
24: \newcommand{\f} {\frac}
25: \newcommand{\ddx}{\frac{\partial}{\partial x}}
26: \newcommand{\ddy}{\frac{\partial}{\partial y}}
27: \newcommand{\ddt}{\frac{\partial}{\partial t}}
28:
29: \begin{document}
30:
31: \title{Formation of Structure in Molecular Clouds: A Case Study}
32:
33: \author{Fabian Heitsch\altaffilmark{1,2}}
34: \author{Andreas Burkert\altaffilmark{1}}
35: \author{Lee W. Hartmann\altaffilmark{3}}
36: \author{Adrianne D. Slyz\altaffilmark{4}}
37: \author{Julien E.G. Devriendt\altaffilmark{5}}
38: \altaffiltext{1}{University Observatory Munich, Scheinerstr. 1, 81679 Munich, Germany}
39: \altaffiltext{2}{U Wisconsin-Madison, 475 N Charter St, Madison,
40: WI 53706, U.S.A.}
41: \altaffiltext{3}{Harvard-Smithsonian Center for Astrophysics, 60 Garden St, MS42,
42: Cambridge, MA 02138, U.S.A.}
43: \altaffiltext{4}{CRAL, Ecole Normale Superieure, 69364, Lyon, Cedex 07, France}
44: \altaffiltext{5}{CRAL, Observatoire de Lyon 9, Avenue Charles Andre,
45: 69561 St-Genis Laval Cedex, France}
46:
47: \lefthead{Heitsch et al.}
48: \righthead{Formation of Structured Molecular Clouds}
49:
50: \begin{abstract}
51: Molecular clouds (MCs) are highly structured and ``turbulent''. Colliding gas streams
52: of atomic hydrogen have been suggested as a possible source of MCs, imprinting
53: the filamentary structure as a consequence of dynamical and thermal instabilities.
54: We present a 2D numerical analysis of MC formation via
55: converging HI flows. Even with modest flow speeds and completely uniform inflows,
56: non-linear density perturbations as possible precursors of MCs arise.
57: Thus, we suggest that MCs are inevitably formed with
58: substantial structure, e.g., strong density and velocity fluctuations,
59: which provide the initial conditions for subsequent gravitational collapse
60: and star formation in a variety of galactic and extragalactic environments.
61:
62: \end{abstract}
63: \keywords{turbulence --- methods:numerical
64: --- ISM:clouds --- ISM:kinematics and dynamics--- stars:formation}
65:
66: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
67: %
68: %\section{Motivation}
69: %
70: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
71: \section{Motivation}\label{s:motivation}
72:
73: Molecular clouds (MCs) in our Galaxy are complex and highly-structured,
74: with broad, non-thermal linewidths suggesting substantial turbulent motion
75: \citep{FAP1990,WBM2000}.
76: Thus, MCs very likely are not static entities
77: and might not necessarily be in an equilibrium state, but
78: their properties could well be determined by their formation process.
79: The importance of initial conditions for cloud structure is emphasized
80: by observational and theoretical evidence for short cloud "lifetimes"
81: %meaning that the molecular phase lasts approximately for a turbulent timescale
82: (\citealp{BHV1999}; \citealp{ELM2000}; \citealp{HBB2001};
83: \citealp{HAR2002}).
84:
85: Flows are ubiquitous in the interstellar medium (ISM) due to the energy
86: input by stars - photoionization, winds, and supernovae. In principle, they
87: can pile up atomic gas to form MCs.
88: Shock waves propagating into the warm ISM
89: will fragment in the presence of thermal instability and linear perturbations
90: \citep{KOI2000,KOI2002,KIB2004,HEP1999,HEP2000} and allow H$_2$-formation within a few
91: Myrs \citep{BHR2004} in a plane-parallel geometry.
92: We envisage the colliding flows less as e.g. interacting shells, but as
93: (more or less) coherent gas streams from turbulent motions on scales
94: of the order of the Galactic gaseous disk height \citep{BHV1999,HBB2001}.
95: Parametrizing the inflows as a
96: ram pressure allowed \citet{HWG2003,HWC2004} to study the
97: fragmentation and collapse of an externally pressurized slab.
98:
99: In this paper we present a study of the generation of filaments
100: and turbulence in atomic clouds -- which may be precursors
101: of MCs -- as a consequence of their formation process,
102: extending the model of large-scale colliding HI-flows outlined
103: by \citet{BHV1999} and \citet{BUR2004}. We discuss the dominant dynamical
104: and thermal instabilities leading to turbulent flows and fragmentation
105: of an initially completely uniform flow.
106: Resulting non-thermal
107: linewidths in the cold gas phases (the progenitors of MCs)
108: reproduce observed values, emphasizing the ease with
109: which turbulent and filamentary structures can be formed in the ISM.
110:
111: This study is a "proof of concept", discussing the dominant instabilities
112: leading to non-linear density perturbations and quantifying the timescales that
113: are necessary to reach conditions under which molecules -- and eventually stars --
114: could form. Since turbulence is thought to be one of the main agents
115: controlling star formation (\citealp{LAR1981}; \citealp{MAK2004}; \citealp{ELS2004}),
116: and consequent density fluctuations surely are crucial to gravitational fragmentation,
117: star formation theory can benefit from a better understanding of the
118: structure initially present in MCs.
119: Detailed aspects and parameter studies will be presented in a forthcoming
120: paper \citep{HSD2005}.
121:
122: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
123: %
124: %\section{Physical Mechanisms}
125: %
126: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
127: \section{Physical Mechanisms}
128: We restricted the models to hydrodynamics including thermal
129: instabilities, leaving out the effects of gravity and magnetic fields.
130: Gravity would eventually lead to
131: further fragmentation, and magnetic fields would be expected to have a stabilizing effect.
132: For this regime, then, we identify three relevant instabilities:
133:
134: (1) The Non-linear Thin Shell Instability (NTSI, \citealp{VIS1994})
135: arises in a shock-bounded slab, when ripples in a two-dimensional slab
136: focus incoming shocked material and produce density fluctuations.
137: The growth rate is $\sim c_sk(k\Delta)^{1/2}$, where $c_s$ is the sound speed,
138: $k$ is the wave number along the slab, and $\Delta$ is the amplitude of the spatial
139: perturbation of the slab.
140: Numerical studies focused on the generation of substructure via Kelvin-Helmholtz-modes
141: \citep{BLM1996}, on the role of gravity \citep{HSW1986} and on the effect of
142: the cooling strength \citep{HUE2003}.
143: \citet{WAF1998,WAF2000} discussed the interaction of stellar winds, and
144: \citet{KLW1998} investigated cloud collisions. The latter authors included
145: magnetic fields, albeit only partially.
146:
147: (2) The flows deflected at the slab will cause Kelvin-Helmholtz Instabilities
148: (KHI), which have been studied at great length analytically and
149: numerically. If the velocity profile across the shear
150: layer is given by a step function, and if the densities are constant across the layer,
151: the growth rate is $k\Delta U$, where $\Delta U$ is the velocity difference.
152: If aligned with the flow, magnetic tension forces can stabilize against the KHI.
153: %(e.g. \citealp{WZH2004}).
154:
155: (3) The Thermal Instability (TI, \citealp{FIE1965}) rests on the
156: balancing of heating and cooling processes in the ISM.
157: The TI develops an isobaric condensation mode
158: and an acoustic mode, which -- under ISM-conditions -- is mostly damped.
159: The condensation mode's growth rate is independent of the wave length, however,
160: since it is an isobaric mode, smaller perturbations will grow
161: first \citep{BUL2000}. A lower growth scale is set by heat conduction, whose
162: scale needs to be resolved \citep{KIA2004}.
163: The signature of the TI are fragmentation and clumping as long as the
164: sound crossing time is smaller than the cooling time scale.
165: The TI can drive turbulence in an otherwise quiescent
166: medium \citep{AUH2005} -- even continuously, if an episodic heating source
167: is available \citep{KNA2002,KNB2002}.
168:
169: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
170: %
171: %\section{Numerical Method}
172: %
173: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
174: \section{Numerical Method}
175: All three instabilities grow fastest or at least first on the smallest
176: scales. This poses a dire challenge for the numerical method.
177: %, since the
178: %spatial discretization determines the smallest scales, and thus the onset
179: %of the instabilities could well be determined by numerical effects.
180: We chose a method based on the 2nd order BGK formalism
181: (\citealp{PRX1993}; \citealp{SLP1999}; \citealp{HZS2004}; \citealp{SDB2005}),
182: allowing control of viscosity and heat conduction. The statistical properties
183: of the models are invariant under changes of viscosity, heat conduction and
184: grid resolution,
185: although the flow patterns change in detail --- as to be expected in a
186: turbulent environment \citep{HSD2005}. The linear resolution varies between
187: $512$ and $2048$ cells. The heating and cooling rates are restricted to optically thin
188: atomic lines following \citet{WHM1995}, so that we are able to study the precursors of MCs
189: up to the point when they could form H$_2$. Dust extinction becomes important above
190: column densities of $N(\mbox{HI})\approx 1.2\times 10^{21}$cm$^{-2}$, which are
191: reached only in the dense cold regions of the models. Thus, we can use the unattenuated
192: UV radiation field for grain heating \citep{WHM1995} for most of the simulation domain,
193: expecting substantial uncertainties in cooling rates only for the densest regions.
194: The ionization degree is derived from a balance between ionization by cosmic rays and
195: recombination, assuming that Ly $\alpha$ photons are directly reabsorbed.
196:
197:
198: Two opposing, uniform, identical flows in the $x$-$y$ computational plane initially
199: collide head-on at a sinusoidal interface with wave number $k_y=1$ and amplitude $\Delta$.
200: The incoming flows are in thermal equilibrium.
201: The system is thermally unstable for densities $1\lesssim n\lesssim 10$cm$^{-3}$.
202: The cooling curve covers a density range of
203: $10^{-2} \leq n \leq 10^3$ cm$^{-3}$ and a temperature range of
204: $30\leq T \leq 1.8\times10^4$ K. The box side length is $44$pc. Thus,
205: Coriolis forces from Galactic rotation are negligible.
206: For a interface with $\Delta=0$, a cold
207: high-density slab devoid of inner structure would form.
208: The onset of the dynamical instabilities thus can solely be
209: controlled by varying the amplitude of the interface perturbation. This
210: allows us to test turbulence generations under minimally favourable
211: conditions.
212:
213: Model series B has $n_0=1.0$cm$^{-3}$ and $T_0=2.5\times 10^3$K,
214: series C has $n_0=0.5$cm$^{-3}$ and $T_0=5.3\times 10^3$K. The amplitude of the
215: interface perturbation corresponds to $2.5$\% of the box length, i.e. $1.1$pc in the
216: horizontal direction.
217: The second digit of the model name gives the Mach number of the inflow.
218: The inflow velocity varies between $6<v_{in}<17$ km s$^{-1}$.
219:
220: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
221: %
222: %\section{Results}
223: %
224: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
225: \section{Results}
226: We chose to present a few models containing the salient features of
227: cloud formation via colliding flows. We ask how hard it is to generate
228: non-linear turbulent density perturbations in an otherwise uniform flow.
229:
230: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
231: %\subsection{Dominant Instabilities}
232: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
233: \subsection{Dominant Instabilities}
234: The structures generated in colliding flows depend strongly on the initial parameters
235: (Fig.~\ref{f:4instab})\footnote{All plots without time-dependence are taken at the
236: endpoint of the corresponding model, i.e. at $15$Myrs for model B3, and
237: at $19$Myrs for all other models.}, as a result of the dominating instability.
238: This is not surprising, since all three instabilities at work have different signatures.
239: For high-density, low-velocity inflows (model B1, upper left panel), the
240: TI dominates and leads to fast cooling, manifested
241: in a coherent slab of cold gas. This situation comes closest to the 1D
242: plane-parallel slab.
243:
244: \begin{figure}[h]
245: \includegraphics[width=0.49\columnwidth]{./f1a_c.eps}
246: \hfill
247: \includegraphics[width=0.49\columnwidth]{./f1b_c.eps}
248: \vfill
249: \includegraphics[width=0.49\columnwidth]{./f1c_c.eps}
250: \hfill
251: \includegraphics[width=0.49\columnwidth]{./f1d_c.eps}
252: \caption{\label{f:4instab}Temperature maps ($\log T$ in K) for models representative
253: of the dominating instability (see text). From upper left to lower right: B1
254: (TI), C1 (KHI), C2, (KHI+NTSI), B3 (NTSI).}
255: \end{figure}
256:
257: Reducing the density (model C1, upper right)
258: leads to slightly less efficient cooling, thus giving the dynamical instabilities
259: time to work, in this case dominated by the KHI. The eddies are visible at
260: the flanks of the initial slab.
261: Increasing the inflow speed while keeping all other parameters fixed
262: increases the vertical $x$-momentum transport,
263: so that for model C2 (lower left), the NTSI will arise. Its typical
264: signature are long strands of denser (and colder) gas predominantly
265: along the flow direction (Hueckstaedt 2003). The KHI modes that are still
266: discernible here have vanished in model B3. Here, the NTSI dominates the
267: dynamics almost completely.
268: Oblique colliding flows (not shown) lead to a nearly instantaneous break-up
269: of the initial slab, because they excite KHI modes at the scale of the initial,
270: non-linear perturbation. The NTSI dominates more and more with higher
271: Mach numbers (up to Mach $4$, not shown). This morphological discussion will
272: be quantified in \citet{HSD2005}.
273:
274: Despite the symmetric initial conditions, all models develop large-scale
275: asymmetries, resulting from an amplification of slight differences
276: at machine accuracy between the upper and lower half of the domain. The main
277: culprit is the strong cooling. Without cooling (i.e. for an adiabatic or isothermal
278: equation of state), or for 1D problems with cooling, the code preserves perfect symmetry.
279:
280: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
281: %\subsection{Mass Distribution}
282: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
283: \subsection{Mass Distribution}
284:
285: Figure~\ref{f:massspec} shows the mass spectrum of cores.
286: Cores are defined as coherent regions with densities $n>100$cm$^{-3}$.
287: Each of the histograms comprises approximately $50$ cores. The dashed
288: line denotes a spectral index of $-1.7$, as observed for observed
289: molecular cores \citep{HBS1998}. The resolution limit for both histograms lies at
290: $1.8\mbox{M}_\odot$. Note that (a) these "cores" correspond to cold HI regions, not
291: molecular cores, and that (b) the masses are per length in our 2D models.
292: A direct comparison between these 2D spectra and observed 3D mass spectra
293: requires assumptions concerning what structures the
294: filaments would correspond to in three dimensions. Assuming that the probability density
295: functions are the same for 2D and 3D, the mass spectra
296: are expected to be flatter in 3D \citep{CHS2001}.
297:
298: \begin{figure}[h]
299: \includegraphics[width=0.49\columnwidth]{./f2a.eps}
300: \hfill
301: \includegraphics[width=0.49\columnwidth]{./f2b.eps}
302: \caption{\label{f:massspec}
303: {\em Left:} Mass-histogram of objects with densities
304: $n>100$cm$^{-3}$, i.e. likely precursors of molecular fragments, for
305: models C2 and B3. The dashed line denotes a power
306: law with exponent $-1.7$. The resolution limit is $M_r=1.8$M$_\odot$.
307: {\em Right:} Histogram of the local Jeans number $L/\lambda_J$ for
308: the cold regions, for models C2 and B3.
309: Clearly, none of the dense objects would be gravitationally bound.}
310: \end{figure}
311:
312: Would these cold cores be gravitationally unstable? Although the models do not include
313: self-gravity, we can estimate the thermal Jeans length
314: $\lambda_J\equiv (\pi/(G\rho))^{1/2}c_s$
315: and its turbulent counterpart $\lambda_{t}\equiv (\pi/(G\rho))^{1/2}\langle v^2\rangle^{1/2}$.
316: The ratio of the cores size over the core's Jeans length $n_J\equiv L/\lambda_J<1$, i.e.,
317: none of the cores would be gravitationally unstable (Fig.\ref{f:massspec}, right).
318: Since $\lambda_J \propto T^{1/2}/(P/T)^{1/2} \propto T$ for an isobaric contraction,
319: the histogram of Figure~\ref{f:massspec} (right) would
320: shift by a factor of $3$ to larger Jeans masses if we cooled the gas down to $T\approx 10$K,
321: thus still yielding Jeans-stable objects. The core size is determined by the geometric mean
322: of the longest and shortest radius.
323:
324: However, if we determine the "global" Jeans number of the cold gas --- given by the thickness
325: of the slab over the global Jeans length derived from the mean density and temperature in
326: the cold gas --- after $10$Myrs for model B3 and $15$Myrs for model C2,
327: the length scale of the cold gas $L>\lambda_J$, i.e. the "slab" would become gravitationally
328: unstable without turbulence (Fig.~\ref{f:jeansmass}).
329: Note that these quantities are global
330: measures in the sense that they do not refer to isolated cold regions. The turbulent
331: Jeans length $\lambda_{t}>L$ for all times and models. Global (i.e. large-scale)
332: gravitational effects could still have a crucial effect on the system \citep{BUH2004}, especially
333: once the inflow stops.
334:
335: \begin{figure}
336: \centerline{\includegraphics[width=0.99\columnwidth]{./f3.eps}}
337: \caption{\label{f:jeansmass}Average length scale (thick
338: lines), turbulent Jeans length (medium lines) and thermal Jeans length (thin lines) for
339: models C2 and B3 against time.}
340: \end{figure}
341:
342: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
343: %\subsection{Linewidths}
344: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
345: \subsection{Linewidths and Kinetic Energy Modes}
346:
347: A primary observable of interstellar clouds is the line-of-sight
348: velocity dispersion $\sigma_v$. MCs consistently show
349: non-thermal linewidths of a few km s$^{-1}$ \citep{WBM2000},
350: that -- together with temperatures of $T\approx 10$K -- are generally
351: interpreted as supersonic turbulence in those clouds.
352: The linewidths in our models are consistent with the observed
353: values (Fig.~\ref{f:linewidths}).
354: The "observed" linewidth is derived from the density-weighted histogram of the
355: line-of-sight velocity dispersion in the cold gas at $T<100$K (filled symbols).
356: Comparing this to the {\em internal} linewidth of coherent cold regions (open symbols),
357: we note a marked offset between the two values. The sound speed
358: of the cold gas ranges around $0.8$km s$^{-1}$. Thus, the {\em internal} velocity
359: dispersions do not reach Mach numbers ${\cal M} > 1$. Hence, the "supersonic"
360: linewidths are a consequence of cold regions moving with respect to each other,
361: but not a result of internal supersonic turbulence in the cold gas which eventually
362: would be hosting star formation. This is consistent with the argument by
363: \citet{HAR2002}, that because of the ages and small spatial dispersions of
364: young stars in Taurus, their velocity dispersions relative to their natal gas
365: are very likely subsonic. Note that the turbulent linewidths $\sigma_v$
366: amount only to a fraction of the inflow velocity.
367:
368: \begin{figure}[h]
369: \centerline{\includegraphics[width=0.99\columnwidth]{./f4.eps}}
370: \caption{\label{f:linewidths}Mean velocity
371: dispersion in cold gas at $T<100$K (filled symbols) and in the cold coherent
372: regions (open symbols) against the inflow velocity.}
373: \end{figure}
374:
375: Figure~\ref{f:ecmpsol} shows the compressible, the solenoidal and the total
376: specific kinetic energy for the whole domain (left) and for the cold gas ($T<100$K,
377: right). Because of the highly compressible initial conditions, the total specific
378: kinetic energy is dominated by compressible modes. However, within the bounding
379: shocks, the highly compressible
380: inflows are converted efficiently into solenoidal motions.
381: Note that because of the 2D geometry,
382: the ratio of solenoidal over compressible kinetic energy is only a lower limit.
383: An extension of Figure~\ref{f:ecmpsol} to radiative losses and internal energy
384: allows us to estimate the overall efficiency of turbulence generation in
385: MCs \citep{HSD2005}.
386:
387: \begin{figure}[h]
388: \centerline{\includegraphics[width=0.99\columnwidth]{./f5.eps}}
389: \caption{\label{f:ecmpsol}Specific kinetic energy fraction against time for models
390: C2 and B3, split up into compressional, solenoidal
391: and total specific kinetic energy. {\em Left:} For the whole domain.
392: {\em Right:} for the cold gas ($T<100$K). The specific kinetic energy is
393: normalized to the value at $t=0$.}
394: \end{figure}
395:
396:
397: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
398: %
399: %\section{Summary}
400: %
401: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
402: \section{Summary}
403:
404: Even for completely uniform inflows,
405: we have shown that the combination of dynamical and thermal instabilities
406: efficiently generates non-linear density perturbations that
407: seed structure of eventual MCs. There is a direct correlation between the
408: morphology of the resulting clouds and the dominating instability.
409:
410: While our "cloud" would be gravitationally unstable, the isolated cold
411: regions would still be stable against gravity. Fragmentation and turbulent
412: mixing driven by the incoming warm gas would prevent the global collapse of the
413: cloud -- under the caveat that a finite extent of the cloud in the vertical
414: direction might lead to edge effects resulting in collapse \citep{BUH2004}.
415: Linewidths reached in the cold gas are consistent with observed values of
416: a few km s$^{-1}$ (Fig~\ref{f:linewidths}) and reach only a fraction of the inflow
417: speed. The internal linewidths, however, are generally subsonic. Thus,
418: the label "supersonic turbulence" refers to the velocities with respect to the cold
419: gas, but does {\em not} necessarily give a hydrodynamically accurate description of
420: the cold gas.
421:
422: Although we adopted a specific cooling curve and thus set the physical regime for our
423: models, we expect the mechanism to work on a variety of scales. The surface density
424: of the cold gas should give us a rough estimate of the amount of stars forming later
425: on. Even though the cold gas mass depends strongly on the turbulent evolution of
426: the slab, it correlates strongly with the inflow momentum. In this sense, colliding
427: flows not only could explain the rather quiescent star formation events as in Taurus,
428: but would be a suitable model for generating star bursts in galaxy mergers.
429:
430: \acknowledgements
431: We enjoyed the discussions with J.~Gallagher, R.~Klessen, L.~Sparke and
432: E.~Zweibel. We thank the referee for a speedy and very constructive report.
433: Computations were performed on ariadne built
434: by S.~Jansen at the Department of Astronomy, UW-Madison, and at the
435: NCSA (AST040026). This work was supported by the NSF (AST-0328821).
436:
437: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
438: %
439: %\references
440: %
441: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
442:
443: \begin{thebibliography}{}
444:
445: \bibitem[Audit \& Hennebelle(2005)]
446: {AUH2005}
447: Audit, E., Hennebell, P.\ 2005, A\&A, 433, 1
448: \bibitem[Ballesteros-Paredes et al.(1999)]
449: {BHV1999}
450: Ballesteros-Paredes, J., Hartmann, L., V\'{a}zquez-Semadeni, E.\ 1999, ApJ, 527, 285
451: \bibitem[Bergin et al.(2004)]
452: {BHR2004}
453: Bergin, E.~A., Hartmann, L.~W., Raymond, J.~C., Ballesteros-Paredes, J.\ 2004,
454: ApJ, 612, 921
455: %\bibitem[Blitz \& Williams(1997)]
456: % {BLW1997}
457: % Blitz, L., \& Williams, J.~P.\ 1997, \apjl, 488, L145
458: \bibitem[Blondin \& Marks(1996)]
459: {BLM1996}
460: Blondin, J.~M., Marks, B.~S.\ 1996, New Ast., 1, 235
461: %\bibitem[Boldyrev(2002)]
462: % {BOL2002}
463: % Boldyrev, S.\ 2002, ApJ, 569, 841
464: %\bibitem[Boldyrev et al.(2002)]
465: % {BNP2002}
466: % Boldyrev, S., Nordlund, \AA., Padoan, P.\ 2002, ApJ, 573, 678
467: \bibitem[Burkert \& Hartmann(2004)]
468: {BUH2004}
469: Burkert, A., Hartmann, L.\ 2004, ApJ, 616, 288
470: \bibitem[Burkert \& Lin(2000)]
471: {BUL2000}
472: Burkert, A., Lin, D.~N.~C.\ 2000, ApJ, 537, 270
473: \bibitem[Burkert(2004)]
474: {BUR2004}
475: Burkert, A.\ 2004, ASP Conf.~Ser.~322:
476: The Formation and Evolution of Massive Young Star Clusters, 322, 489
477: %\bibitem[Cazaux et al.(2005)]
478: % {CCT2005}
479: % Cazaux, S., Caselli, P., Tielens, A.~G.~G.~M., Le Bourlot, J., Walmsley, M.\
480: % 2005, J. Phys., 6, 155
481: \bibitem[Chappell \& Scalo(2001)]
482: {CHS2001}
483: Chappell, D., Scalo, S.\ 2001, \mnras, 325, 1
484: %\bibitem[Cho \& Lazarian(2003)]
485: % {CHL2003}
486: % Cho, J., Lazarian, A.\ 2003, MNRAS, 345, 325
487: \bibitem[Elmegreen(2000)]
488: {ELM2000}
489: Elmegreen, B.~G.\ 2000, \apj, 530, 277
490: \bibitem[Elmegreen \& Scalo(2004)]
491: {ELS2004}
492: Elmegreen, B.~G., Scalo, J.\ 2004, ARAA, 42, 211
493: \bibitem[Falgarone \& Philips(1990)]
494: {FAP1990}
495: Falgarone, E., Philips, T.~G.\ 1990, ApJ, 359, 344
496: \bibitem[Field(1965)]
497: {FIE1965}
498: Field, G.~B.\ 1965, ApJ, 142, 531
499: \bibitem[Hartmann et al.(2001)]
500: {HBB2001}
501: Hartmann, L., Ballesteros-Paredes, J., Bergin, E.~A.\ 2001, ApJ, 562, 852
502: \bibitem[Hartmann(2002)]
503: {HAR2002}
504: Hartmann, L.\ 2002, ApJ, 578, 914
505: \bibitem[Heithausen et al.(1998)]
506: {HBS1998}
507: Heithausen, A., Bensch, F., Stutzki, J., Falgarone, E., Panis, J.~F.\ 1998, A\&A, 331, L65
508: \bibitem[Heitsch et al.(2004)]
509: {HZS2004}
510: Heitsch, F., Zweibel, E.~G., Slyz, A.~D., \& Devriendt, J.~E.~G.\ 2004, \apj, 603, 165
511: \bibitem[Heitsch et al.(2005)]
512: {HSD2005}
513: Heitsch, F., Slyz, A.~D., Devriendt, J.~E.~G., Hartmann, L., Burkert, A.\ 2005, in prep.
514: \bibitem[Hennebelle \& P\'{e}rault(1999)]
515: {HEP1999}
516: Hennebelle, P., P\'{e}rault, M.\ 1999, A\&A, 351, 309
517: \bibitem[Hennebelle \& P\'{e}rault(2000)]
518: {HEP2000}
519: Hennebelle, P., P\'{e}rault, M.\ 2000, A\&A, 359, 1124
520: \bibitem[Hennebelle et al.(2003)]
521: {HWG2003}
522: Hennebelle, P., Whitworth, A.~P., Gladwin, P.~P., \& Andr{\' e}, P.\ 2003,
523: \mnras, 340, 870
524: \bibitem[Hennebelle et al.(2004)]
525: {HWC2004}
526: Hennebelle, P., Whitworth, A.~P., Cha, S.-H., \& Goodwin, S.~P.\ 2004,
527: \mnras, 348, 687
528: \bibitem[Hueckstaedt(2003)]
529: {HUE2003}
530: Hueckstaedt, R.~M.\ 2003, New Ast., 8, 295
531: \bibitem[Hunter et al.(1986)]
532: {HSW1986}
533: Hunter Jr., J.~H., Sandford II, M.~T., Whitaker, R., Klein, R.I.\
534: 1986, \apj, 305, 309
535: \bibitem[Klein \& Woods(1998)]
536: {KLW1998}
537: Klein, R.~I., Woods, D.~T.\ 1998, ApJ, 497, 777
538: \bibitem[Koyama \& Inutsuka(2000)]
539: {KOI2000}
540: Koyama, H., \& Inutsuka, S.\ 2002, \apj, 532, 980
541: \bibitem[Koyama \& Inutsuka(2002)]
542: {KOI2002}
543: Koyama, H., \& Inutsuka, S.\ 2002, \apjl, 564, L97
544: \bibitem[Koyama \& Inutsuka(2004a)]
545: {KIA2004}
546: Koyama, H., \& Inutsuka, S.\ 2004, \apjl, 602, L25
547: \bibitem[Koyama \& Inutsuka(2004b)]
548: {KIB2004}
549: Koyama, H., \& Inutsuka, S.\ 2004, RMxAC, 22, 26
550: \bibitem[Kritsuk \& Norman(2002a)]
551: {KNA2002}
552: Kritsuk, A.~G., \& Norman, M.~L.\ 2002, \apjl, 569, L127
553: \bibitem[Kritsuk \& Norman(2002b)]
554: {KNB2002}
555: Kritsuk, A.~G., \& Norman, M.~L.\ 2002, \apjl, 580, L51
556: \bibitem[Larson(1981)]
557: {LAR1981}
558: Larson, R.~B.\ 1981, MNRAS, 194, 809
559: \bibitem[Mac Low \& Klessen(2004)]
560: {MAK2004}
561: Mac Low, M.-M., \& Klessen, R.~S.\ 2004, Reviews of Modern Physics, 76, 125
562: \bibitem[Prendergast \& Xu(1993)]
563: {PRX1993}
564: Prendergast, K. H., Xu, K.,\ 1993, J. Comp. Phys., 109, 53
565: \bibitem[Slyz \& Prendergast(1999)]
566: {SLP1999}
567: Slyz, A.~D., Prendergast, K.~H.\ 1999, A\&AS, 139, 199
568: \bibitem[Slyz et al.(2005)]
569: {SDB2005}
570: Slyz, A.~D., Devriendt, J.~E.~G., Bryan, G., \& Silk, J.\ 2005, \mnras, 356, 737
571: \bibitem[Vishniac(1994)]
572: {VIS1994}
573: Vishniac, E.~T.\ 1994, ApJ, 428, 186
574: \bibitem[Walder \& Folini(1998)]
575: {WAF1998}
576: Walder, R., Folini, D.\ 1998, ApSS, 260, 215
577: \bibitem[Walder \& Folini(2000)]
578: {WAF2000}
579: Walder, R., Folini, D.\ 2000, ApSS, 274, 343
580: %\bibitem[Watson et al.(2004)]
581: % {WZH2004}
582: % Watson, C., Zweibel, E.~G., Heitsch, F., \& Churchwell, E.\ 2004, \apj, 608, 274
583: \bibitem[Williams et al.(2000)]
584: {WBM2000}
585: Williams, J.~P., Blitz, L., \& McKee, C.~F.\ 2000, Protostars and Planets IV, 97
586: \bibitem[Wolfire et al.(1995)]
587: {WHM1995}
588: Wolfire, M.~G., Hollenbach, D., McKee, C.~F., Tielens, A.~G.~G.~M.,
589: Bakes, E.~L.~O.\ 1995,
590: ApJ, 443, 152
591: \end{thebibliography}
592:
593: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
594: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
595:
596: \end{document}
597:
598: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
599: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
600:
601:
602: