astro-ph0011122/ms.tex
1: %\documentstyle[aaspp4,aabib]{article}
2: %\documentstyle[aas2pp4,twoside,aabib]{article}
3: %\documentstyle[aasms4,twoside,aabib]{article}
4: %\documentstyle[aasms4,twoside]{article}
5: \documentstyle[aas2pp4,twoside]{article}
6: \setlength{\oddsidemargin}{0.2cm}
7: \setlength{\evensidemargin}{-1.0cm}
8: \input epsf
9: %
10: \def\s2n{S^{\prime}/N}
11: \def\vecr{{\bf r}}
12: \def\vecdr{{\bf \Delta r}}
13: \def\dr{{\Delta r}}
14: \def\aq{{\langle Q \rangle}}
15: \def\jco{{J=1-0 $^{13}$CO}}
16: \def\co{$^{13}$CO}
17: %
18: 
19: 
20: \begin{document}
21: \title{The Turbulent Shock Origin of Proto--Stellar Cores}
22: 
23: \author{Paolo Padoan\footnote{ppadoan@cfa.harvard.edu},
24: \affil{Harvard-Smithsonian Center for Astrophysics, Cambridge, MA 02138}
25: Mika Juvela\footnote{juvela@astro.helsinki.fi},
26: \affil{Helsinki University Observatory,T\"ahtitorninm\"aki, P.O.Box 14,
27:        SF-00014 University of Helsinki, Finland}
28: Alyssa A. Goodman\footnote{agoodman@cfa.harvard.edu}
29: \affil{Harvard-Smithsonian Center for Astrophysics, Cambridge, MA 02138}
30: and \AA ke Nordlund\footnote{aake@astro.ku.dk}}
31: \affil{Astronomical Observatory and Theoretical Astrophysics Center, \\
32:        Juliane Maries Vej 30, DK-2100 Copenhagen, Denmark}
33: 
34: 
35: \begin{abstract}
36: 
37: 
38: The fragmentation of molecular clouds (MC) into proto--stellar cores 
39: is a central aspect of the process of star formation. Because of the 
40: turbulent nature of super--sonic motions in MCs, it has been suggested 
41: that dense structures such as filaments and clumps are formed by 
42: shocks in a turbulent flow. In this work we present strong
43: evidence in favor of the turbulent origin of the fragmentation
44: of MCs.
45: 
46: The most generic result of turbulent fragmentation is that dense
47: post shock gas traces a gas component with a smaller velocity dispersion
48: than lower density gas, since shocks correspond to regions of converging
49: flows, where the kinetic energy of the turbulent motion is dissipated.
50: 
51: Using synthetic maps of spectra of molecular transitions, computed
52: from the results of numerical simulations of super--sonic turbulence,
53: we show that the dependence of velocity dispersion on gas density
54: generates an observable relation between the rms velocity centroid 
55: and the integrated intensity (column density), $\sigma(V_0)-I$, which is 
56: indeed found in the observational data. The comparison 
57: between the theoretical model (maps of synthetic $^{13}$CO spectra), 
58: with $^{13}$CO maps from the Perseus, Rosette and Taurus MC complexes, 
59: shows excellent agreement in the $\sigma(V_0)-I$ relation.
60: 
61: The $\sigma(V_0)-I$ relation of different observational
62: maps with the same total rms velocity are remarkably
63: similar, which is a strong indication of their origin from a very
64: general property of the fluid equations, such as the turbulent
65: fragmentation process.
66: 
67: 
68: \end{abstract}
69: 
70: \keywords{
71: turbulence -- ISM: kinematics and dynamics -- individual (Perseus, Rosette, Taurus);
72: radio astronomy: interstellar: lines
73: }
74: 
75: 
76: \section{Introduction}
77: 
78: 
79: The importance of turbulence in the process of star formation was recognized long 
80: ago (Von Weizs{\"a}cker 1951), and was discussed in a seminal paper
81: \nocite{von51} 
82: by Larson (1981). Several successive works have tried to use the 
83: observational data to relate different properties of MCs with the
84: physics of laboratory and numerical turbulence, such as power
85: spectra of kinetic energy \footnote{See Leung, Kutner \& Mead 1982; Myers 1983; 
86: Quiroga 1983;  Sanders, Scoville \& Solomon 1985;   Goldsmith \& Arquilla 1985; 
87: Dame et al.\ 1986; Falgarone \& P\'{e}rault 1987.}
88: \nocite{Leung+82,Myers83,Quiroga83}
89: \nocite{Sanders+85} 
90: \nocite{Goldsmith+Arquilla85,Dame+86,Falgarone+Perault87}, 
91: probability distributions of velocity and velocity differences \footnote{ 
92: See Scalo 1984; Kleiner \& Dickman 1985, 1987; Hobson 1992; Miesch \& 
93: Bally 1994; Miesch \& Scalo 1995; Lis et al. 1996; Miesch, Scalo \& 
94: Bally 1999.}, intermittency (Falgarone \& Phillips 1990; Falgarone
95: \& Puget 1995; Falgarone, Pineau Des Forets, \& Roueff 1995),
96: \nocite{Falgarone+Phillips90} \nocite{Falgarone+Puget95} \nocite{Falgarone+95}
97: and self--similarity \footnote{See Beech 1987; Bazell \& D\'{e}sert 1988, Scalo 1990; 
98: Dickman, Horvath \& Margulis 1990; Falgarone, Phillips \& Walker 1991; 
99: Zimmermann, Stutzki \& Winnewisser 1992; Henriksen 1991; 
100: Hetem \& Lepine 1993; Vogelaar \& Wakker 1994; Elmegreen \& Falgarone 1996.}. 
101: \nocite{Beech87}  \nocite{Bazell+Desert88}  \nocite{Scalo90}
102: \nocite{Dickman+90}  \nocite{Falgarone+91} \nocite{Zimmermann+92} 
103: \nocite{Henriksen91} \nocite{Hetem+Lepine93} \nocite{Vogelaar+Wakker94} 
104: \nocite{Elmegreen+Falgarone96}
105: 
106: During the last decade, numerical simulations of transonic turbulence
107: (Passot \& Pouquet 1987; Passot, Pouquet \& Woodward 1988; L\'{e}orat, 
108: Passot \& Pouquet 1990; Lee, Lele \& Moin 1991 \nocite{Lee+91}; Porter, 
109: Pouquet \& Woodward 1992 \nocite{Porter+94}; Kimura \& Tosa 1993; 
110: Porter, Woodward \& Pouquet 1994; Vazquez--Semadeni 1994; Passot, 
111: Vazquez--Semadeni \& Pouquet 1995) 
112: and highly super--sonic turbulence (Passot, V\'{a}zquez--Semadeni, Pouquet
113: 1995; V\'{a}zquez--Semadeni, Passot \& 
114: Pouquet 1996; Padoan \& Nordlund 1997, 1999; Stone, Ostriker \& 
115: Gammie 1998; MacLow et al. 1998; Ostriker, Gammie \& Stone 1999; Padoan, Zweibel \& 
116: Nordlund 2000; Klessen, Heitsch \& MacLow 2000), on relatively high--resolution 
117: numerical grids, have become available, and very detailed comparisons 
118: between observational data with turbulence models of MCs have been
119: performed (Padoan, Jones \& Nordlund 1997; Padoan et al. 1998; Padoan \& 
120: Nordlund 1999; Padoan et al. 1999; Rosolowsky et al. 1999; Padoan, 
121: Rosolowsky \& Goodman 2000). 
122: 
123: Recent numerical studies of super--sonic
124: magneto--hydrodynamic (MHD) turbulence have brought new
125: understanding of the physics of turbulence. The most important
126: results are:
127: 
128: \nocite{MacLow_Puebla98} \nocite{Stone+98} \nocite{Kimura+Tosa93} 
129: \nocite{Lee+91} \nocite{Vazquez-Semadeni+96} \nocite{Klessen+99}
130: \nocite{Padoan+97ext} \nocite{Padoan+98per} \nocite{Padoan+Nordlund98MHD}
131: \nocite{Padoan+98cat} \nocite{Padoan+Nordlund97MHD} \nocite{Ostriker+99}
132: \nocite{Passot+95} \nocite{MacLow+98} \nocite{MacLow99}
133: \nocite{Rosolowsky+99} \nocite{Padoan+2000SCF}
134: 
135: 
136: \begin{itemize}
137: \item Super--sonic turbulence decays in approximately one dynamical 
138: time, independent of the magnetic field strength (Padoan \& Nordlund 1997, 1999;
139: MacLow et al. 1998; Stone, Ostriker \& Gammie 1998; MacLow 1999).
140: \item The probability distribution of gas density in isothermal turbulence is well 
141: approximated by a Log--Normal distribution, whose standard deviation 
142: is a function of the rms Mach number of the flow (Vazquez--Semadeni 1994;
143: Padoan 1995; Padoan, Jones \& Nordlund 1997; Scalo et al. 1998; Passot
144: \& Vazquez--Semadeni 1998; Nordlund \& Padoan 1999; Ostriker, Gammie \& Stone 1999).
145: \item Super--sonic isothermal turbulence generates a complex system 
146: of shocks which fragment the gas very efficiently into high density
147: sheets, filaments, and cores (this is the general result of any numerical 
148: simulation of super--sonic turbulence). 
149: \item Super--Alfv\'{e}nic turbulence provides a good description of the
150: dynamics of MCs, and an explanation for the origin of dense cores with
151: magnetic field strength consistent with Zeeman splitting observations. 
152: (Padoan \& Nordlund 1997, 1999).\footnote{This particular result is 
153: supported almost exclusively by our work. A significant fraction of the 
154: astrophysical community still favors the traditional idea that a rather 
155: strong magnetic field supports MCs against their gravitational collapse. 
156: An example of numerical work that favors the traditional
157: idea of magnetic support is Ostriker, Gammie \& Stone (1999).}  
158: \end{itemize}  
159: 
160: \nocite{Scalo+98} \nocite{Vazquez-Semadeni94}
161: 
162: We call ``turbulent fragmentation'' the process of generation of high 
163: density structures by turbulent shocks. Since random super--sonic 
164: motions are ubiquitous in MCs, turbulent fragmentation cannot
165: be avoided: it is a direct consequence of the observational evidence. 
166: In some analytical studies it is tacitly 
167: assumed that turbulent fragmentation can be neglected if the kinetic 
168: energy of random motions is in rough equipartition with the magnetic 
169: energy. This assumption is wrong, because motions along magnetic field 
170: lines are unavoidable, and so turbulent fragmentation occurs via 
171: super--sonic compressions along the magnetic field lines, as discussed
172: by Gammie \& Ostriker (1996) and Padoan \& Nordlund (1997, 1999).
173: 
174: \nocite{Gammie+Ostriker96}
175: 
176: Numerical simulations of turbulence have been used to discuss the turbulent
177: origin of MC structures by Passot \& Pouquet (1987). Vazquez--Semadeni, 
178: Passot \& Pouquet (1996) and Ballesteros--Paredes, Hartmann \& Vazquez--Semadeni 
179: (1999) have used two dimensional simulations to argue that MCs are formed by 
180: turbulence. Padoan \& Nordlund (1997, 1999) have shown that super--sonic
181: and super--Alfv\'{e}nic turbulence can explain the origin of magnetized cores
182: in MCs, including the observed field strength--density ($B-n$) 
183: relation (Myers \& Goodman 1988; Fiebig \& G\"{u}sten 1989; Crutcher 1999).
184: The idea that proto--stellar cores and stars are formed in turbulent shocks 
185: has been previously discussed by Elmegreen (1993). In that work, an analysis
186: of the gravitational instability of the cores can be found. 
187: Here, we present new strong observational evidence in favor of the 
188: turbulent shock origin of proto--stellar cores. Such evidence
189: is based on the fact that dense post shock gas traces a gas component 
190: with a smaller velocity dispersion than lower density gas, since it maps
191: regions of converging flows, where the kinetic energy of the turbulent 
192: motion is dissipated.
193: 
194: \nocite{Passot+Pouquet87} \nocite{Ballesteros+99} \nocite{Myers+Goodman88} 
195: \nocite{Fiebig+Gusten89} \nocite{Crutcher99} \nocite{Elmegreen93}
196: 
197: In \S 2 and 3, numerical simulations and observational data, used in this
198: work, are briefly described. In \S 4 we compute the rms flow velocity
199: as a function of the gas density, in simulations of super--sonic turbulence,
200: and show that it decreases for increasing values of the gas density.
201: In \S 5 we show that such general property generates an observable 
202: relation between the rms velocity centroid and the integrated intensity 
203: (roughly proportional to the surface density), for the \jco\ transition,
204: and in \S 6 the same relation is found in the observational data. Results are 
205: discussed in \S 7, and conclusions are summarized in \S 8.
206: 
207: 
208: \section{Numerical Models}
209: 
210: 
211: The numerical models used in this work are based on the results
212: of numerical simulations of super--Alfv\'{e}nic and highly super--sonic
213: MHD turbulence, run on a 128$^3$ computational mesh, with periodic
214: boundary conditions. As in our previous works, the initial density 
215: and magnetic fields are uniform; the initial velocity is random, 
216: generated in Fourier space with power only on the large scale. 
217: We also apply an external random force, to drive the turbulence at a 
218: roughly constant rms Mach number of the flow. This force is generated 
219: in Fourier space, with power only on small wave numbers ($1<k<2$), 
220: as the initial velocity. The isothermal equation of state is used. 
221: Descriptions of the numerical code used to solve the MHD equations
222: can be found in Galsgaard \& Nordlund (1996); Nordlund, Stein 
223: \& Galsgaard (1996); Nordlund \& Galsgaard (1997); Padoan \& Nordlund (1999). 
224:  
225: \nocite{Nordlund+96}
226: \nocite{Nordlund+Galsgaard97}
227: \nocite{Galsgaard+Nordlund96}
228: \nocite{Padoan+2000AD}
229: 
230: In this work we neglect the effect of self--gravity, although that can be 
231: described with a different version of our code (Padoan et al. (2000)). 
232: Here we compare the relative velocity of regions of MC complexes as a 
233: function of their gas density or column density. Such regions are distributed 
234: across the full extension of the MC complexes, that is several pc. In our numerical 
235: models, driven continuously by an external force on the large scale, self--gravity 
236: is responsible for the collapse of gravitationally bound cores formed by the turbulent 
237: flow, but does not affect significantly the large scale flow. Since we assume 
238: that large scale motions in MC complexes are due to turbulence, and not to a 
239: gravitational collapse, we can neglect self--gravity. Similarly, we have neglected
240: the effect of ambipolar drift, since it is not relevant for motions on the scale
241: of several pc, although that is computed in a different version of our code,
242: using the strong coupling approximation (see Padoan, Zweibel \& Nordlund 2000). 
243:  
244: In order to scale the models to physical units, we use the following 
245: empirical Larson type relations, as in our previous works: \nocite{Larson81}
246: 
247: \begin{equation}
248: {\cal{M}}=2.0\left(\frac{L}{1pc}\right)^{0.5},
249: \end{equation} 
250: where $\cal{M}$ is the rms sonic Mach number of the flow, and
251: a temperature $T=10$~K is assumed, and
252: \begin{equation}
253: \langle n \rangle=2.0\times10^3\left(\frac{L}{1pc}\right)^{-1},
254: \end{equation} 
255: where the gas density $n$ is expressed in cm$^{-3}$.
256: The rms sonic Mach number is an input parameter of the
257: numerical simulations, and can be used to re-scale them
258: to physical units. When comparing theoretical models with
259: observations, $\cal{M}$ is in fact the only parameter that
260: we need to match (or its two dimensional equivalent on the maps,
261: $\sigma_v$ --see \S 3). In the absence of self--gravity and magnetic field,
262: statistical properties of turbulent flows (very large Reynolds 
263: number) with the same value of $\cal{M}$ should be universal, and
264: independent of the average density (for isothermal flows without self--gravity). 
265: If the magnetic field is
266: present, the rms Alfv\'{e}nic Mach number of the flow is also
267: an input parameter of the numerical simulations (it determines
268: the magnetic field strength). The physical unit of velocity in 
269: the code is the isothermal speed of sound, $C_s$, and the 
270: physical unit of the magnetic field is 
271: $C_s(4\pi\langle\rho\rangle)^{\frac{1}{2}}$ (cgs).
272: 
273: In this work we use two numerical models with rms velocity
274: 3.4 and 1.7 km/s, which corresponds to ${\cal M}\approx 13.0$ 
275: and 6.5 respectively. Using the Larson type relations (1) and (2)
276: we get $L\approx 42$ and 10.5 pc and $\langle n \rangle\approx 48$ 
277: and 190 cm$^{-3}$. The one dimensional rms velocity for the two models
278: is $\sigma_v=2.0$ and 1.0 km/s, which are also recovered from the 
279: analysis of the synthetic spectral maps computed with these models.
280: The values of $\sigma_v$ have been chosen for the appropriate
281: comparison with the observational data presented in \S~3.\footnote{
282: The Taurus MC complex has $\sigma_v\approx 1.0$~km/s, and $L\approx 40$~pc,
283: while the Larson type relation (1) would give $L\approx 12.6$~pc.
284: For the purpose of this work we are interested in comparing the 
285: observations with numerical models with similar rms Mach number, 
286: and we do not try to match the physical extension of each MC complex.}
287: The magnetic field strength is $B\approx 5$~$\mu$G in both models.
288: 
289: Maps of synthetic spectra of molecular transitions are computed, using 
290: a non--LTE Monte Carlo radiative transfer code (Juvela 1997, 1998),
291: from the three dimensional density and velocity fields generated in
292: the numerical MHD experiments.
293: The method of computing synthetic spectra was presented in 
294: Padoan et al. (1998). For the purpose of this work we use only 
295: one molecular transition, namely \jco\ . Uniform temperature, $T=10$~K 
296: is assumed for these radiative transfer calculations, in agreement
297: with the isothermal equation of state used in the MHD calculations.
298: We are presently computing thermal equilibrium models of MCs that
299: we will use in a future work to study the effect of realistic 
300: temperature variations in molecular spectra. 
301: 
302: Since the spectral noise
303: resulting from uncertainties in our radiative transfer calculations
304: is always much smaller than the typical observational noise, the
305: comparison of synthetic spectra with observational data can be done
306: only after adding noise to the synthetic spectra, and the effect
307: of noise on the statistical properties of the spectra need to be quantified
308: (see \S 4).
309: 
310: 
311: \section{Observational Data}
312: 
313: 
314: We choose to use the \jco\ transition because it samples the range
315: of values of column density we are here interested in, and also because
316: several large maps of molecular clouds are available in
317: this transition.
318: We compare maps of \jco\ synthetic spectra with some of the largest
319: observational \jco\ spectral maps in the literature from the following 
320: Galactic regions: the Perseus MC complex (Billawala, Bally \& Sutherland 
321: 1997), the Taurus MC complex (Mizuno et al. 1995), the Rosette MC complex
322: (Blitz \& Stark 1986; Heyer et al. 2000). These MC complexes
323: have an extension of approximately 30--50~pc, and radial velocity 
324: dispersions in the range 1--2.4~km/s. We define the total rms velocity of a
325: map as the rms velocity weighted with the total spectrum of the map, $T(v)$:
326: \begin{equation}
327: \sigma_v=\sqrt{\sum_v\,(v-\bar{v})^2\,T(v)\,dv\over{\sum_v T(v)\,dv}},
328: \end{equation}
329: where
330: \begin{equation}
331: \bar{v}={\sum_v v\,T(v)\,dv\over{\sum_v T(v)\,dv}}.
332: \end{equation}
333: The Blitz \& Stark map of the Rosette MC complex has $\sigma_v=2.4$~km/s,
334: but, limited to the region that matches the more recent Heyer et al.
335: map, the value is $\sigma_v=2.0$~km/s (for both maps). The full map
336: of the Perseus MC complex also yields a value of $\sigma_v=2.0$~km/s.
337: Taurus is instead much less turbulent, despite its large spatial extent,
338: with $\sigma_v=1.0$~km/s. 
339: 
340: The angular resolution is inversely proportional to the diameter of the
341: antenna: 4 m for the Taurus map, 7 m for the Perseus and the Blitz \& Stark
342: Rosette maps, and 14 m for the Heyer et al. map of Rosette.
343: Assuming a distance of 140 pc for Taurus, 300 pc for Perseus, and 
344: 1600 pc for Rosette, the spatial resolution of the maps is 0.1 pc
345: for Taurus, 0.15 pc for Perseus, 0.84 pc for the Blitz \& Stark map
346: of Rosette, and 0.42 pc for the Heyer et al. map of Rosette.   
347: The spectral resolution is 0.1 km/s for Taurus, 0.273 km/s for Perseus,
348: 0.68 km/s for the Blitz \& Stark map of Rosette, and 0.06 km/s for
349: the Heyer et al. map of the same cloud. 
350: 
351: Rms noise $N$ and average spectrum quality $\aq$ (see Padoan, Rosolowsky
352: \& Goodman 2000) also vary from 
353: map to map. The spectrum quality, $Q$, is related to the 
354: signal--to--noise, $S/N$. It is defined as the ratio of the rms 
355: signal (over the whole spectrum or inside a velocity window) 
356: to the rms noise, $N$:
357: \begin{equation}
358: Q={\sqrt{\sum_v\,T(v)^2 dv}\over{N}}
359: \end{equation}
360: The usual definition of $S/N$ is based on Gaussian fits of the spectra,
361: which we prefer to avoid because the \jco\ transition typically yields
362: spectra with significant non--Gaussian shape and multiple components. 
363: The spectrum quality is a sort of signal--to--noise weighted over the
364: whole spectrum. The relation between $Q$ and $S/N$ is discussed in
365: Padoan, Goodman \& Rosolowsky (2000). We call average spectrum
366: quality $\aq$ the value of $Q$ averaged over the whole map. Values of
367: $\aq$, $N$, resolutions and $\sigma_v$ are listed in Table~1 for
368: all the observational maps. In the following, when different maps 
369: are compared with each other, we make noise and velocity resolution
370: equal in the different maps, by adding noise and reducing the velocity 
371: resolution where necessary.
372: 
373: 
374: 
375: \section{Velocity Dispersion Versus Gas Density in Super--Sonic Turbulence}
376: 
377: 
378: Turbulent fragmentation generates a complex system of dense post shock sheets,
379: filaments and cores, reminiscent of structures observed in molecular cloud
380: (MC) maps. An example of a two dimensional projection of the three dimensional
381: density field from a simulation of super--sonic turbulence is shown in Figure~1. 
382: In previous works (Padoan, Jones \& Nordlund 1997; Padoan \& Nordlund 1997, 
383: 1999), we have shown that, besides this morphological similarity, the 
384: density field of numerical super--sonic turbulence has important statistical 
385: properties in agreement with the density field of observed MCs.
386: 
387: Figure~2 (left panel) is a two dimensional section (no projection) of the 
388: same three-dimensional density field used in Figure~1. A complex system of 
389: filaments is apparent, with a number of high density cores inside the
390: filaments. In three dimensional super--sonic turbulence, filaments
391: are formed by two dimensional compressions (at the intersection of sheets)
392: and the densest cores are formed by three dimensional compressions. 
393: Most of the ``filaments'' in two dimensional sections like Figure~2, are 
394: instead two dimensional cuts through sheets, and most of the cores are local 
395: density maxima, due to fluctuations in the shock velocity (they
396: usually corresponds to strongly curved segments of filaments).
397: Such density maxima are often unstable to gravitational collapse,
398: and are the origin of proto--stellar cores. 
399: 
400: Since the dense gas originates in shocks, that is in regions of converging flows,
401: it should move with significantly lower velocity than the
402: lower density turbulent flow. This is illustrated in the right panel
403: of Figure~2, which shows the magnitude of the flow velocity on the same
404: plane as the left panel. Dark blue is low velocity, and dark red high
405: velocity. It is clear that high density filaments trace regions of low
406: velocity, at the intersections of high velocity ``blobs''. 
407: This general property of super--sonic turbulence is quantified by
408: the dependence of the rms flow velocity, $\sigma(v)$, on the gas density:
409: \begin{equation}
410: \sigma(v)=\langle (v-\bar{v})^2 |\rho \rangle,
411: \end{equation}
412: We repeat for different values of $\rho$, and than plot the result
413: versus the gas density. Expression (6) means that the
414: rms flow velocity is obtained as an average over the whole computational
415: box, using only positions where the gas density has a value contained in
416: an interval centered around $\rho$. The average is repeated for different
417: intervals of values of $\rho$, to span the whole range of densities.
418: 
419: The plot is shown in Figure~3, at time
420: $t/t_{dyn}=0.07$ (square symbols) and $t/t_{dyn}=1.0$ (asterisks),
421: where $t/t_{dyn}$ is the time in units of the dynamical time (defined
422: as the size of the computational box divided by the rms flow velocity), 
423: and the density is uniform in the initial conditions, at $t/t_{dyn}=0.0$.
424: At $t/t_{dyn}=1.0$, regions of $\rho\approx \langle \rho\rangle$ have 
425: $\sigma(v)$ comparable to the total rms velocity, $\sigma_v=3.0$~km/s, 
426: while regions with 10 times higher density have much smaller rms velocity, 
427: $\sigma(v)\approx 2.3$~km/sec, and at  $\rho=100\,\langle \rho\rangle$
428: $\sigma(v)\approx 0.9$~km/s. The rms velocity conditioned to the gas density
429: decreases sharply with increasing gas density already at a very early time,
430: $t/t_{dyn}=0.07$, when the density field is still very smooth.
431: 
432: 
433: 
434: \section{Rms Velocity Centroid versus Integrated Intensity}
435: 
436: Observational spectral maps of MCs, do not provide a direct estimate
437: of the three dimensional velocity field and gas density. Only radial
438: velocity and velocity--integrated intensity (which is roughly proportional to the
439: surface density), averaged along each line of sight, are directly available from the data.
440: However, the dependence of the rms velocity centroid on the integrated 
441: intensity should resemble the $\sigma(v)-\rho$ relation, since lines of
442: sight of high intensity are usually dominated by one or more dense cores. 
443: The velocity centroid, $V_0$, is the average velocity along an
444: individual line of sight ${\bf x}$ on the map:
445: \begin{equation}
446: V_0({\bf x})={\sum_v v\, T(v,{\bf x})dv \over {\sum_v T(v,{\bf x})dv}},
447: \end{equation}
448: where $T(v,{\bf x})$ is the signal (antenna temperature), at the velocity 
449: $v$ and position ${\bf x}$, and $dv$ is the width of the velocity channel.
450: The integrated intensity, $I({\bf x})$, is:
451: \begin{equation}
452: I({\bf x})=\sum_v T(v,{\bf x})dv.
453: \end{equation}
454: 
455: We compute the rms velocity centroid, $\sigma(V_0)$, conditioned
456: on $I$:
457: \begin{equation}
458: \sigma(V_0)=\langle (V_0-\bar{V_0})^2 |I \rangle_{\bf x},
459: \end{equation}
460: (spatial average) and plot it against $I$. Expression (9) means that
461: the rms velocity centroid is obtained as an average over the whole map,
462: selecting all lines--of--sight on the map with values of $I$ 
463: inside a given interval, [$I,I+dI$]. This rms velocity centroid
464: is not equivalent to a local line width, because it is obtained
465: as an average over the whole map. The plot is shown in 
466: Figure~4, for two maps of \jco\ synthetic spectra with different 
467: total line widths $\sigma_v$ (defined in \S 3). Figure~4 shows
468: that $\sigma(V_0)$ decreases significantly for increasing values of $I$.
469: The general property of super--sonic turbulence, namely high density 
470: gas moves relatively slowly, is therefore apparent also in the 
471: observable relation $\sigma(V_0)-I$.
472: The relation $\sigma(V_0)-I$ is affected by noise and by the width of
473: the velocity channels, which must be taken into account when comparing
474: different maps. Noise has been added to the synthetic 
475: spectra used for computing the plot in Figure~4, to yield a value of the 
476: spectrum quality comparable to a typical value found in the observational
477: data used in this work, $\aq=3.5$ (see \S 3 for the definition
478: of $\aq$). 
479: 
480: In Figure~5 we show the effect of noise (left panel) and 
481: spectral resolution (right panel). The effect of increasing the noise
482: (decreasing the value of $\aq$) is that of making the $\sigma(V_0)-I$
483: relation steeper, because the uncertainty in the determination of
484: the velocity centroids due to noise contributes to the dispersion of 
485: velocity centroid values, and the effect increases at decreasing values of
486: integrated intensity $I$. Lower spectral resolution further increases
487: the same effect. However, we have verified that if the noise is low
488: enough, $\aq>20$, both noise and velocity resolution have no effect
489: on the $\sigma(V_0)-I$ relation, and the squared symbols in the left 
490: panel of Figure~5 correspond therefore to the intrinsic relations.
491: We can conclude that this observable relation truly originates from 
492: the three dimensional $\sigma(v)-\rho$ relation, with only a partial 
493: contribution from noise. 
494: 
495: We have also verified that spatial resolution does not affect 
496: our results. The spatial resolution can be decreased significantly,
497: by rebinning the map to a smaller number of spectra, without any 
498: appreciable variation in the 
499: $\sigma(V_0)-I$ relation. As the spatial resolution is decreased,
500: however, the statistical sample (number of spectra) decreases, and 
501: statistical fluctuations (deviations around the high resolution 
502: $\sigma(V_0)-I$ relation) become progressively more important.
503: 
504: 
505: 
506: \section{Observational $\sigma(V_0)-I$ Relation}
507: 
508: We have shown in the previous section that the $\sigma(V_0)-I$ 
509: relation is sensitive to the value of the rms noise and to the
510: spectral resolution. In the following plots, where we compare 
511: different spectral maps from observations and models, we have therefore
512: added noise to the spectra and increased the velocity channel width $dv$
513: to match the map with the largest noise and $dv$. We have not modified 
514: the spatial resolution in any map, since that has no systematic 
515: effect on the $\sigma(V_0)-I$ relation, as commented above. 
516: 
517: The left panel of Figure~6 shows the $\sigma(V_0)-I$ relation 
518: for the  maps of MC complexes introduced in \S 3. For the 
519: Rosette MC complex we have used the portion of the full 
520: Blitz \& Stark map that matches the Heyer et al. map.
521: The two models used for the comparison have $\sigma_v=2.0$~km/s,
522: similar to Rosette and Perseus, and $\sigma_v=1.0$~km/s, similar 
523: to Taurus. It is remarkable that the Rosette and the Perseus MC 
524: complexes have indistinguishable $\sigma(V_0)-I$ relations,
525: which are also coincident with the theoretical prediction 
526: (square symbols), for the same value of $\sigma_v$. 
527: The result for Taurus is also in good agreement with the 
528: $\sigma_v=1.0$~km/s model. Horizontal shifts could be expected 
529: in the plot, since different MC complexes can have different 
530: surface density. However, MCs and MC complexes are known to 
531: approximately follow the Larson relation between density
532: and size (equation (2)), which implies roughly constant surface 
533: density, and small horizontal shifts in the plot. 
534: 
535: To compute the plots in the left panel of Figure~6, all maps 
536: have been treated to make their rms noise and velocity resolution 
537: equal. This is achieved by rebinning spectral profiles into a smaller 
538: number of velocity channels, and by adding noise, when necessary. In order
539: to check that the $\sigma(V_0)-I$ relations from different maps
540: treated in this way are really comparable, we have computed the
541: $\sigma(V_0)-I$ relation for the Rosette MC complex using both
542: the Heyer et al. map, and the portion of the Blitz \& Stark map
543: that matches the region covered by the Heyer et al. map. The result
544: is plotted in the right panel of Figure~6. The velocity resolution
545: of the Heyer et al. map has been decreased from $dv=0.06$~km/s 
546: to $dv=0.68$~km/s, and noise has been added, to match exactly 
547: the velocity resolution and rms noise in the Blitz \& Stark map. 
548: As can be seen in the right panel of Figure~6, after
549: this drastic treatment of the higher resolution map, the 
550: $\sigma(V_0)-I$ relations for the two maps are practically
551: indistinguishable from each other, which support the validity
552: of the comparison of different maps (left panel of Figure~6.
553: In the right panel, the case of the full Blitz \& Stark map is 
554: also plotted (square symbols). The full map has a higher total
555: rms velocity, $\sigma_v=1.4$~km/s, than the portion that matches 
556: the Heyer et al. map, and its $\sigma(V_0)-I$ relation is 
557: therefore steeper, as expected. 
558: 
559: 
560: 
561: \section{Discussion and Conclusions}
562: 
563: 
564: 
565: The origin of proto--stellar cores is a fundamental problem in our 
566: understanding of the process of star formation. Models that have
567: been proposed to describe proto-stellar cores are based on i) static
568: or quasi--static equilibrium (e.g. Curry \& Mckee 2000; Jason \& 
569: Pudritz 2000), ii) thermal instability (e.g. Yoshii \& Sabano 1980; Gilden 
570: 1984; Graziani \& Black 1987), iii) gravitational instability through
571: ambipolar diffusion (e.g. Basu \& Mouschovias 1994; Nakamura, Hanawa, \& 
572: Nakano 1995; Indebetouw \& Zweibel 2000; Ciolek \& Basu 2000), 
573: iv) non--linear Alfv\'{e}n waves (e.g. Carlberg \& Pudritz 1990; 
574: Elmegreen 1990, 1997, 1999), v) clump collisions (e.g. Gilden 1984; 
575: Kimura \& Tosa 1996), vi) super--sonic turbulence (e.g Elmegreen 1993; 
576: Klessen, Burkert and Bate 1998; Klessen, Heitsch, \& Mac Low 2000;
577: Padoan et al. 2000).
578: 
579: Many detailed comparisons between observational data and models,
580: which support the idea of the turbulent origin of the structure and 
581: kinematics of molecular clouds, have been presented in our previous 
582: papers (Padoan, Jones \& Nordlund 1997; Padoan et al. 1998; Padoan 
583: et al. 1999; Padoan \& Nordlund 1999; Padoan, Goodman \& Rosolowsky
584: 2000). Models of numerical turbulence can be 
585: compared with the observations by computing i) synthetic stellar 
586: extinction measurements, ii) synthetic spectral maps of molecular 
587: transitions, iii) synthetic Zeeman splitting measurements, iv) 
588: synthetic polarization maps.
589: 
590: Maps of synthetic spectra contain a lot of information about
591: the kinematic, the structure and the thermal properties of 
592: molecular clouds, and can be analyzed with different statistical
593: tools. In this work we have presented a new statistical method
594: to analyze spectral--line maps, which is very useful because it
595: probes directly a very general property of super--sonic turbulence,
596: that is the fact that dense gas traces a gas component with a 
597: smaller velocity dispersion than lower density gas. This property 
598: arises because the gas density is enhanced in regions where the 
599: large scale turbulent flow converges (compressions) and the kinetic 
600: energy of the turbulent flow is dissipated by shocks.
601: If local compressions are instead due to local instabilities (e.g. 
602: gravitational instability, or gravitational instability mediated
603: by ambipolar diffusion), and the large scale motions are only the
604: consequence of local instabilities (as it should be in a 
605: self--consistent picture), gas density increases with the flow 
606: velocity dispersion (see for example the model by Indebetouw \& 
607: Zweibel 2000), contrary to the observational evidence presented 
608: in this work.
609: 
610: \nocite{Indebetouw+Zweibel2000}
611: 
612: The main conclusion of this work is that every model
613: for the origin of molecular cloud structure and proto--stellar 
614: cores should be tested against the $\sigma(V_0)-I$ relation.
615: Turbulent fragmentation provides a realistic scenario for the
616: origin of proto--stellar cores, which satisfies this new 
617: observational constraint.  
618: 
619: 
620: \nocite{Passot+88}
621: 
622: 
623: 
624: \acknowledgements
625: 
626: We are grateful to Edith Falgarone and Phil Myers for discussions 
627: that stimulated this work, and to the referee for a number
628: of useful comments. This work was supported by NSF grant AST-9721455.
629: \AA ke Nordlund acknowledges partial support by the Danish National 
630: Research Foundation through its establishment of the Theoretical 
631: Astrophysics Center.
632: 
633: \clearpage
634: 
635: \begin{thebibliography}{}
636: 
637: \bibitem[\protect\astroncite{{Ballesteros-Paredes}
638:   et~al.}{1999}]{Ballesteros+99}
639: {Ballesteros-Paredes}, J., {Hartmann}, L., {V{\'a}zquez-Semadeni}, E. 1999,
640:   ApJ, 527, 285
641: 
642: \bibitem[\protect\astroncite{Bazell \& D\'{e}sert}{1988}]{Bazell+Desert88}
643: Bazell, D., D\'{e}sert, F.~X. 1988, ApJ, 333, 353
644: 
645: \bibitem[\protect\astroncite{{Beech}}{1987}]{Beech87}
646: {Beech}, M. 1987, Ap\&SS, 133, 193
647: 
648: \bibitem[\protect\astroncite{{Crutcher}}{1999}]{Crutcher99}
649: {Crutcher}, R.~M. 1999, ApJ, 520, 706
650: 
651: \bibitem[\protect\astroncite{Dame et~al.}{1986}]{Dame+86}
652: Dame, T.~M., Elmegreen, B.~G., Cohen, R.~S., Thaddeus, P. 1986, ApJ, 305, 892
653: 
654: \bibitem[\protect\astroncite{Dickman et~al.}{1990}]{Dickman+90}
655: Dickman, R.~L., M., M., {Horvath}, M.~A. 1990, ApJ, 365, 586
656: 
657: \bibitem[\protect\astroncite{Elmegreen}{1993}]{Elmegreen93}
658: Elmegreen, B.~G. 1993, ApJ, 419, 29
659: 
660: \bibitem[\protect\astroncite{Elmegreen \&
661:   Falgarone}{1996}]{Elmegreen+Falgarone96}
662: Elmegreen, B.~G., Falgarone, E. 1996, ApJ, 471, 816
663: 
664: \bibitem[\protect\astroncite{Falgarone \&
665:   P\'{e}rault}{1987}]{Falgarone+Perault87}
666: Falgarone, E., P\'{e}rault, M. 1987,
667: \newblock in G. Morfil, M. Scholer (eds.), Physical Processes in Interstellar
668:   Clouds, Reidel, Dordrecht, ~59
669: 
670: \bibitem[\protect\astroncite{Falgarone \&
671:   Phillips}{1990}]{Falgarone+Phillips90}
672: Falgarone, E., Phillips, T.~G. 1990, ApJ, 359, 344
673: 
674: \bibitem[\protect\astroncite{Falgarone et~al.}{1991}]{Falgarone+91}
675: Falgarone, E., Phillips, T.~G., Walker, C. 1991, ApJ, 378, 186
676: 
677: \bibitem[\protect\astroncite{Falgarone et~al.}{1995}]{Falgarone+95}
678: Falgarone, E., Pineau~des Forets, G., Roueff, E. 1995, A\&A, 300, 870
679: 
680: \bibitem[\protect\astroncite{Falgarone \& Puget}{1995}]{Falgarone+Puget95}
681: Falgarone, E., Puget, J. 1995, A\&A, 293, 840
682: 
683: \bibitem[\protect\astroncite{Fiebig \& G\"{u}sten}{1989}]{Fiebig+Gusten89}
684: Fiebig, D., G\"{u}sten, R. 1989, A\&A, 214, 333
685: 
686: \bibitem[\protect\astroncite{Galsgaard \&
687:   Nordlund}{1996}]{Galsgaard+Nordlund96}
688: Galsgaard, K., Nordlund, A. 1996, J. Geophys. Res., 101(A6), 13445
689: 
690: \bibitem[\protect\astroncite{Gammie \& Ostriker}{1996}]{Gammie+Ostriker96}
691: Gammie, C.~F., Ostriker, E.~C. 1996, ApJ, 466, 814
692: 
693: \bibitem[\protect\astroncite{Goldsmith \&
694:   Arquilla}{1985}]{Goldsmith+Arquilla85}
695: Goldsmith, P.~F., Arquilla, R. 1985, ApJ, 297, 436
696: 
697: \bibitem[\protect\astroncite{{Henriksen}}{1991}]{Henriksen91}
698: {Henriksen}, R.~N. 1991, ApJ, 377, 500
699: 
700: \bibitem[\protect\astroncite{{Hetem} \& {Lepine}}{1993}]{Hetem+Lepine93}
701: {Hetem}, A., J., {Lepine}, J. R.~D. 1993, A\&A, 270, 451
702: 
703: \bibitem[\protect\astroncite{{Indebetouw} \&
704:   {Zweibel}}{2000}]{Indebetouw+Zweibel2000}
705: {Indebetouw}, R.~., {Zweibel}, E.~G. 2000, ApJ, 532, 361
706: 
707: \bibitem[\protect\astroncite{Kimura \& Tosa}{1993}]{Kimura+Tosa93}
708: Kimura, T., Tosa, M. 1993, ApJ, 406, 512
709: 
710: \bibitem[\protect\astroncite{Klessen et~al.}{1999}]{Klessen+99}
711: Klessen, R.~S., Heitsch, F., MacLow, M.~M. 1999,  astro--ph/9911068
712: 
713: \bibitem[\protect\astroncite{Larson}{1981}]{Larson81}
714: Larson, R.~B. 1981, MNRAS, 194, 809
715: 
716: \bibitem[\protect\astroncite{Lee et~al.}{1991}]{Lee+91}
717: Lee, S., Lele, S.~K., Moin, P. 1991, Phys. Fluids A, 3, 657
718: 
719: \bibitem[\protect\astroncite{Leung et~al.}{1982}]{Leung+82}
720: Leung, C.~M., Kutner, M.~L., Mead, K.~N. 1982, ApJ, 262, 583
721: 
722: \bibitem[\protect\astroncite{{Mac Low}}{1999}]{MacLow99}
723: {Mac Low}, M. 1999, ApJ, 524, 169
724: 
725: \bibitem[\protect\astroncite{{Mac Low} et~al.}{1998}]{MacLow+98}
726: {Mac Low}, M., {Smith}, M.~D., {Klessen}, R.~S., {Burkert}, A. 1998, ApJS, 261,
727:   195
728: 
729: \bibitem[\protect\astroncite{Mac~Low}{1998}]{MacLow_Puebla98}
730: Mac~Low, M.-M. 1998,
731: \newblock in J. Franco, A. Carrami{\~n}ana (eds.), Interstellar Turbulence,
732:   Cambridge University Press
733: 
734: \bibitem[\protect\astroncite{Myers}{1983}]{Myers83}
735: Myers, P.~C. 1983, ApJ, 270, 105
736: 
737: \bibitem[\protect\astroncite{Myers \& Goodman}{1988}]{Myers+Goodman88}
738: Myers, P.~C., Goodman, A.~A. 1988, ApJ, 326, L27
739: 
740: \bibitem[\protect\astroncite{Nordlund \&
741:   Galsgaard}{1997}]{Nordlund+Galsgaard97}
742: Nordlund, A., Galsgaard, K. 1997, A 3D MHD Code for Parallel Computers,
743:   technical report, Astronomical Observatory, Copenhagen University
744: 
745: \bibitem[\protect\astroncite{Nordlund et~al.}{1996}]{Nordlund+96}
746: Nordlund, A., Stein, R.~F., Galsgaard, K. 1996, in Lecture Notes in Computer
747:   Science 1041, ed. J. Wazniewsky, (Heidelberg: Springer), 450
748: 
749: \bibitem[\protect\astroncite{{Ostriker} et~al.}{1999}]{Ostriker+99}
750: {Ostriker}, E.~C., {Gammie}, C.~F., {Stone}, J.~M. 1999, ApJ, 513, 259
751: 
752: \bibitem[\protect\astroncite{Padoan et~al.}{1999}]{Padoan+98per}
753: Padoan, P., Bally, J., Billawala, Y., Juvela, M., Nordlund, {\AA}. 1999, ApJ,
754:   525, 318
755: 
756: \bibitem[\protect\astroncite{Padoan et~al.}{1997}]{Padoan+97ext}
757: Padoan, P., Jones, B., Nordlund, {\AA}. 1997, ApJ, 474, 730
758: 
759: \bibitem[\protect\astroncite{Padoan et~al.}{1998}]{Padoan+98cat}
760: Padoan, P., Juvela, M., Bally, J., Nordlund, {\AA}. 1998, ApJ, 504, 300
761: 
762: \bibitem[\protect\astroncite{Padoan \& Nordlund}{1997}]{Padoan+Nordlund97MHD}
763: Padoan, P., Nordlund, {\AA}. 1997,  astro--ph/9706176
764: 
765: \bibitem[\protect\astroncite{Padoan \& Nordlund}{1999}]{Padoan+Nordlund98MHD}
766: Padoan, P., Nordlund, {\AA}. 1999, ApJ, 526, 279
767: 
768: \bibitem[\protect\astroncite{Padoan et~al.}{2001}]{Padoan+2000FRAG}
769: Padoan, P., Nordlund, \AA ., R\"{o}gnvaldsson, \"{O}., Goodman, A.~A. 2000, ApJ (submitted)
770: 
771: \bibitem[\protect\astroncite{Padoan et~al.}{2000a}]{Padoan+2000SCF}
772: Padoan, P., Rosolowsky, E, W., Goodman, A.~A. 2000, ApJ (in press)
773: 
774: \bibitem[\protect\astroncite{Padoan et~al.}{2000b}]{Padoan+2000AD}
775: Padoan, P., Zweibel, E., Nordlund, {\AA}. 2000, ApJ, 540, 332
776: 
777: \bibitem[\protect\astroncite{Passot \& Pouquet}{1987}]{Passot+Pouquet87}
778: Passot, T., Pouquet, A. 1987, J. Fluid Mech., 181, 441
779: 
780: \bibitem[\protect\astroncite{Passot et~al.}{1988}]{Passot+88}
781: Passot, T., Pouquet, A., Woodward, P. 1988, 197, 228
782: 
783: \bibitem[\protect\astroncite{Passot et~al.}{1995}]{Passot+95}
784: Passot, T., V\'{a}zquez-Semadeni, E., A., P. 1995, ApJ, 455, 536
785: 
786: \bibitem[\protect\astroncite{Porter et~al.}{1994}]{Porter+94}
787: Porter, D.~H., Pouquet, A., Woodward, P.~R. 1994, Phys. Fluids, 6, 2133
788: 
789: \bibitem[\protect\astroncite{Quiroga}{1983}]{Quiroga83}
790: Quiroga, R.~J. 1983, Ap. Sp. Sci., 93, 37
791: 
792: \bibitem[\protect\astroncite{Rosolowsky et~al.}{1999}]{Rosolowsky+99}
793: Rosolowsky, E, W., Goodman, A.~A., Wilner, D.~J., Williams, J.~P. 1999, ApJ,
794:   524, 887
795: 
796: \bibitem[\protect\astroncite{Sanders et~al.}{1985}]{Sanders+85}
797: Sanders, D.~B., Scoville, N.~Z., Solomon, P.~M. 1985, ApJ, 289, 372
798: 
799: \bibitem[\protect\astroncite{Scalo}{1990}]{Scalo90}
800: Scalo, J.~M. 1990,
801: \newblock in R. Capuzzo-Dolcetta, C. Chiosi, A.~D. Fazio (eds.), Physical
802:   Processes in Fragmentation and Star Formation, (Kluwer : Dordrecht,  p.151
803: 
804: \bibitem[\protect\astroncite{Scalo et~al.}{1998}]{Scalo+98}
805: Scalo, J.~M., V\'{a}zquez-Semadeni, E., Chappell, D., Passot, T. 1998, ApJ,
806:   504, 835
807: 
808: \bibitem[\protect\astroncite{{Stone} et~al.}{1998}]{Stone+98}
809: {Stone}, J.~M., {Ostriker}, E.~C., {Gammie}, C.~F. 1998, ApJ, 508, L99
810: 
811: \bibitem[\protect\astroncite{V\'{a}zquez-Semadeni}{1994}]{Vazquez-Semadeni94}
812: V\'{a}zquez-Semadeni, E. 1994, ApJ, 423, 681
813: 
814: \bibitem[\protect\astroncite{V\'{a}zquez-Semadeni
815:   et~al.}{1996}]{Vazquez-Semadeni+96}
816: V\'{a}zquez-Semadeni, E., Passot, T., Pouquet, A. 1996, ApJ, 473, 881
817: 
818: \bibitem[\protect\astroncite{{Vogelaar} \& {Wakker}}{1994}]{Vogelaar+Wakker94}
819: {Vogelaar}, M. G.~R., {Wakker}, B.~P. 1994, A\&A, 291, 557
820: 
821: \bibitem[\protect\astroncite{{von Weizs{\"a}cker}}{1951}]{von51}
822: {von Weizs{\"a}cker}, C.~F. 1951, ApJ, 114, 165
823: 
824: \bibitem[\protect\astroncite{{Zimmermann} et~al.}{1992}]{Zimmermann+92}
825: {Zimmermann}, T., {Stutzki}, J., {Winnewisser}, G. 1992,
826: \newblock in Evolution of Interstellar Matter and Dynamics of Galaxies,  254
827: 
828: \end{thebibliography}
829: 
830: 
831: \clearpage
832: 
833: 
834: 
835: 
836: \onecolumn
837: 
838: {\bf TABLE AND FIGURE CAPTIONS:} \\
839: 
840: {\bf Table~1:} Cloud name; maximum spatial extension; 
841: total rms velocity; telescope beam size; velocity
842: channel width; rms noise; average spectrum quality 
843: (signal--to--noise); bibliographic reference. \\  
844:                
845: {\bf Figure \ref{fig1}:} Two dimensional projection of the three 
846: dimensional density field from a simulation of isothermal
847: super--sonic turbulence with rms Mach number $\sim 10$. \\ 
848: 
849: {\bf Figure \ref{fig2}:} Left panel: Two dimensional section
850: (no projection) of the same density field used for Figure~1.
851: Most filaments are sections of sheets, and most dense cores
852: are density maxima inside curved segments of filaments, formed
853: by fluctuations in the shock velocity. Right panel: Modulus of
854: the three dimensional flow velocity on the same two dimensional plane as in the 
855: left panel. Dark blue is low velocity, and dark red high velocity.
856: The dense filaments on the left panel are commonly found in regions 
857: of small flow velocity, at the intersections of patches of high
858: velocity. \\
859: 
860: {\bf Figure \ref{fig3}:} Rms flow velocity, conditioned to
861: gas density, versus the gas density, computed from a simulation
862: of super--sonic turbulence with rms Mach number $\sim 10$.
863: Squared symbols are for an early time, just 7\% of the 
864: dynamical time after an initial condition with uniform density;
865: asterisks are for a time equal to a dynamical time. Density values
866: are binned over 20 logarithmic intervals between the average
867: and the highest density. Within each interval, the density
868: grows by a factor of 1.04 at the early time, and 1.27 at one 
869: dynamical time. \\
870: 
871: {\bf Figure \ref{fig4}:} Rms velocity centroid, conditioned
872: to integrated intensity, versus the integrated intensity
873: (see text for details). Two maps of synthetic spectra from
874: MHD turbulence simulations are used, with different values
875: of the total rms radial velocity (intensity weighted),
876: $\sigma_v=1.0$~km/s (diamonds), and $\sigma_v=2.0$~km/s 
877: (triangles). The model with larger rms velocity has also 
878: a slightly larger maximum intensity (surface density), 
879: although both models are rescaled into physical units
880: using the Larson relations (see \S 2). Integrated intensity 
881: values are binned over 12 linear intervals between the lowest
882: and the highest values in the map. Within each interval, 
883: the density increment is $\approx 1$~K~km/s (for the 
884: $\sigma_v=1.0$~km/s model) and $\approx 2$~K~km/s (for the 
885: $\sigma_v=2.0$~km/s model). \\
886: 
887: 
888: {\bf Figure \ref{fig5}:} Effect of noise and spectral
889: resolution. Left panel: Conditioned rms velocity
890: centroid versus intensity. Triangle and diamond
891: symbols are the same as in Figure~4, while squared symbols
892: have higher spectrum quality $\aq$ (signal--to--noise). 
893: Different values of $\aq$ are obtained by adding different
894: levels of noise to the synthetic spectra. $\aq=3.5$ is typical
895: of the observational maps used in this work. The plot is 
896: not sensitive to the decreasing noise, for $\aq >20$.
897: Right panel: Same as left panel, but the squared symbols
898: are for lower spectral resolution (the velocity channels
899: of the synthetic spectra is increased from $dv=0.2$~km/s
900: to $dv=0.6$~km/s). Clearly, both the level of noise and
901: the spectral resolution must be taken into account when 
902: comparing different maps. Integrated intensity 
903: values are binned over 12 linear intervals between the lowest
904: and the highest values in the map, as in Figure~4. \\
905: 
906: {\bf Figure \ref{fig6}:} Observational $\sigma(V_0)-I$
907: relation. Left panel: Conditioned rms velocity centroid 
908: versus intensity for different MC complexes: Rosette, 
909: Perseus, and Taurus. Square symbols are for maps of 
910: synthetic spectra with total rms velocity $\sigma_v=1.0$~km/s
911: and 2.0~km/s. Right panel: Comparison of the two maps
912: of the Rosette MC complex (see text for details). Integrated 
913: intensity values are binned over 10 linear intervals between 
914: the lowest and the highest values in the map. \\
915: 
916: 
917: 
918: \clearpage
919: \begin{table}
920: \begin{tabular}{lccccccl}
921: \hline
922: \hline
923: MC       & $L$ [pc] & $\sigma_v [km/s]$ & $dx$ [pc]  & $dv$ [km/s] & $N$ [K] & $\aq$ & reference  \\ 
924: \hline
925: Taurus   &   38      &   1.0             &  0.10       &   0.10      &  0.24   &  2.3  & Mizuno et al. (1995)   \\ 
926: Perseus  &   27      &   2.0             &  0.15       &   0.27      &  0.24   &  2.8  & Billawala et al. (1997) \\ 
927: Rosette  &   52      &   2.4             &  0.84       &   0.68      &  0.20   &  2.1  & Blitz \& Stark (1986) \\ 
928: Rosette  &   36      &   2.2             &  0.84       &   0.68      &  0.19   &  4.0  & Blitz \& Stark (matching region) \\ 
929: Rosette  &   36      &   2.0             &  0.42       &   0.06      &  0.12   &  3.8  & Heyer et al. (2000) \\
930: \hline
931: \end{tabular}
932: \caption{}
933: \end{table}
934: 
935: 
936: 
937: 
938: \clearpage
939: \begin{figure}
940: \centerline{\epsfxsize=13cm \epsfbox{fig1.eps}}
941: \caption[]{}
942: \label{fig1}
943: \end{figure}
944: 
945: \clearpage
946: \begin{figure}
947: \centerline{\epsfxsize=18cm \epsfbox{fig2.eps}}
948: \caption[]{}
949: \label{fig2}
950: \end{figure}
951: 
952: \clearpage
953: \begin{figure}
954: \centerline{\epsfxsize=13cm \epsfbox{fig3.eps}}
955: \caption[]{}
956: \label{fig3}
957: \end{figure}
958: 
959: \clearpage
960: \begin{figure}
961: \centerline{\epsfxsize=13cm \epsfbox{fig4.eps}}
962: \caption[]{}
963: \label{fig4}
964: \end{figure}
965: 
966: \clearpage
967: \begin{figure}
968: {\epsfxsize=8.cm \epsfbox{fig5a.eps}}
969: {\epsfxsize=8.cm \epsfbox{fig5b.eps}}
970: \caption[]{}
971: \label{fig5}
972: \end{figure}
973: 
974: 
975: \clearpage
976: \begin{figure}
977: {\epsfxsize=8cm \epsfbox{fig6a.eps}}
978: {\epsfxsize=8cm \epsfbox{fig6b.eps}}
979: \caption[]{}
980: \label{fig6}
981: \end{figure}
982: 
983: 
984: 
985: \end{document}
986: 
987: