1: \documentclass[11pt,twoside]{article}
2:
3: \usepackage{asp2004}
4: \usepackage{epsf}
5: \usepackage{epsfig}
6: \usepackage{lscape}
7:
8: \markboth{M. Haverkorn}{Turbulence in the Galactic warm ISM}
9:
10: \pagestyle{myheadings}
11: \setcounter{equation}{0}
12: \setcounter{figure}{0}
13: \setcounter{footnote}{0}
14: \setcounter{section}{0}
15: \setcounter{table}{0}
16:
17: \begin{document}
18: \title{Characteristics of the Galactic magneto-ionized ISM from
19: Faraday rotation}
20:
21: \author{Marijke Haverkorn}
22: \affil{Jansky fellow, National Radio Astronomy Observatory}
23: \affil{Astronomy Department University of California at Berkeley, 601
24: Campbell Hall, Berkeley CA 94720, USA}
25:
26: \begin{abstract}
27: Faraday rotation measurements of polarized extragalactic sources probe
28: the Galactic magnetized, ionized interstellar medium. Rotation
29: measures of these sources behind the inner Galactic plane are used to
30: explore characteristics of the structure in the spiral arms and in
31: interarm regions. Structure in the spiral arms has a characteristic
32: outer scale of a few parsecs only, whereas interarm regions typically
33: show structure up to scales of hundreds of parsecs. The data indicate
34: that in the spiral arms, the random component of the magnetic field
35: dominates over the regular field, but in the interarm regions the
36: random and regular field components may be comparable, and a few times
37: weaker than the random magnetic field in the spiral arms.
38: \end{abstract}
39:
40: \section{Introduction}
41:
42: The discussion whether tiny scale structure in atomic, molecular or
43: ionized form is part of a power law power spectrum or is made up of
44: overdense discrete structures is ongoing (see e.g.\ Heiles, Mason,
45: Deshpande, this volume). Therefore, in addition to the study of the
46: tiny structures themselves, studies to determine the power spectra of
47: the neutral and ionized medium are relevant to have a framework in
48: which the existence of tiny scale structure can be tested.
49:
50: The velocity and density power spectra of neutral gas are extensively
51: probed as part of molecular cloud evolution and star formation studies
52: \citep{es04}. For the warm ionized component of the interstellar
53: medium (ISM), electron density studies mostly show Kolmogorov spectra
54: \citep{ars95}. The ionized gas dynamics and structure are expected to
55: be heavily influenced by magnetic fields threading the medium, the
56: characteristics of which are still very uncertain.
57:
58: In this paper, we discuss fluctuations in the warm magneto-ionized
59: medium probed by way of Faraday rotation, to estimate typical scales of
60: structure and magnetic field strengths in spiral arms and in interarm
61: regions.
62:
63: \section{Data analysis of polarized extragalactic point sources}
64:
65: Polarized radiation from extragalactic point sources is altered by
66: Faraday rotation when propagating through the Milky Way plane, which
67: makes these point sources a good probe of the structure in the
68: Galaxy's magnetic field and electron density.\footnote{Faraday
69: rotation describes the rotation of the angle of linear polarization
70: $\phi$ due to birefringence for left and right handed circular
71: polarization in a magnetized, ionized medium. Faraday rotation is
72: wavelength dependent: $\phi \propto\mbox{RM}\,\lambda^2$, where rotation
73: measure RM is RM~$=0.81\int n_e[\mbox{cm}^{-3}] \, B_{\parallel}[\mu
74: \mbox{G}] \, dl[\mbox{pc}]$, with $n_e$ thermal electron density,
75: $B_{\parallel}$ magnetic field strength parallel to the line of sight
76: and $dl$ path length.}
77:
78: However, as these sources are irregularly spaced on the sky,
79: performing a Fourier transform to obtain the typical scales of
80: structure introduces artifacts. Instead, the second order structure
81: function (SF) of rotation measure RM can be used, which is defined as
82: $D_{\mbox{RM}}(dr) = \langle
83: (\mbox{RM}(r)-\mbox{RM}(r+dr))^2\rangle_r$, where $dr$ is the
84: separation of two sources on the sky, and $\langle\rangle_r$ is an
85: average over every position $r$ which contains a source with a
86: neighboring source in a bin around $dr$ away. For a power law power
87: spectrum in RM, the SF will be a power law with a certain smallest
88: dissipation scale $l_d$ which is much smaller than the scales probed
89: in this paper, and an outer scale $l_0$ which is the maximum scale
90: found in the turbulence, believed to be the dominant scale of energy
91: input.
92:
93: The data we use are from the Southern Galactic Plane Survey (SGPS,
94: McClure-Griffiths et al.\ 2005, Haverkorn et al.\ 2006a), a neutral
95: hydrogen and full-po\-la\-ri\-za\-tion 1.4~GHz continuum survey of the
96: Galactic plane, which spans an area of $253\deg < l < 357\deg$ and
97: $|b| < 1.5\deg$ and contains 148 polarized sources with an unambiguous
98: RM measurement (Brown et al.\ 2006). The data are obtained with the
99: Australia Telescope Compact Array (ATCA) and the Parkes 64m
100: single-dish telescope, and are publicly available (ATCA data only for
101: polarized continuum
102: data)\footnote{http://www.atnf.csiro.au/research/cont/sgps/queryForm.html}.
103:
104: Lines of sight through discrete structures like H~{\sc ii} regions and
105: supernova remnants are biased due to the large electron density and
106: possibly magnetic field, which increases $|$RM$|$ in this direction
107: \citep{mwk03}. Therefore, we have used the total intensity 1.4~GHz
108: radio data from the ATCA combined with Parkes single-dish data to
109: determine which extragalactic sources have a sight line passing through
110: a supernova remnant or H~{\sc ii} region. The data from these sources
111: (about 15\% of the total sample) were then discarded.
112:
113: SFs are sensitive to large-scale gradients in electron density across
114: the field of view. In addition, the geometrical component of the
115: change in magnetic field can to first order be approximated with a
116: linear contribution. A plane in RM is subtracted from the area in
117: which a SF is computed to correct for these effects.
118:
119: Because the SGPS data probe the inner Galaxy, which includes a number
120: of spiral arms, they are well-suited to study differences in the
121: structure in the ISM in spiral arms and in interarm regions. The SGPS
122: sources are used to construct SFs for different lines of sight
123: preferentially through spiral arms or mostly through interarm regions,
124: estimated from the spiral arm positions in
125: \citet{cl02}. Figure~\ref{f:sf} shows the SFs in lines of sight
126: primarily going through interarm regions (upper panels) and in lines
127: of sight dominated by spiral arms (lower panels).
128:
129: %***********************
130: \begin{figure}[t]
131: \centerline{\psfig{figure=figure1.ps,width=.85\textwidth}}
132: \caption{Structure functions of RM for Galactic interarm regions (top)
133: and spiral arms (bottom). The dotted lines are linear fits to
134: the rising parts of the SFs (top) and horizontal lines
135: (bottom).}
136: \label{f:sf}
137: \end{figure}
138: %***********************
139:
140: \section{Typical scales of structure}
141: \label{s:out}
142:
143: \subsection{The turnover scale of structure functions}
144:
145: The difference between the structure in RM in spiral arms and the
146: structure in interarm regions is obvious: the spiral arm SFs are flat,
147: while in interarm regions the SFs rise to a certain turnover in the
148: SF. The location of the turnover is interpreted as the largest angular
149: scale of structure in the interarm regions. With the argument that the
150: largest angular scales in RM are probably coming from nearby, this
151: outer scale corresponds to spatial scales of about 100-200~pc. For the
152: spiral arms we can only estimate an upper limit for the outer scale of
153: structure, i.e.\ the smallest scale we probe. In this way, we estimate
154: the outer scale of structure in the spiral arms to be smaller than
155: about 10~pc \citep{hgb06}.
156:
157:
158: \begin{table}
159: \begin{center}
160: \begin{tabular}{cccccccccc}
161: region & range in l & $r_{out}$ & $p$ & $\sigma$ & $r_{out}^d$ & $B_0$ & $C_B^2$ & $B_{ran}$ \\
162: & [$^{\circ}]$ & [pc] & [\%] & [rad m$^{-2}$] & [pc] & [$\mu$G] & & [$\mu$G]\\
163: \hline
164: Inter1 & 255 - 281 & 100 & 5.5 & 200 & 1 & 2.8 & 10 & 2.1\\
165: Carina & 281 - 292 & $<$ 17 & 1.5 & 250 & 7 & 3.5 & 18 & 2.8\\
166: Inter2 & 292 - 308 & 170 & 3.7 & 250 & 3 & 3.8 & 3 & 1.2\\
167: Crux & 308 - 317 & $<$ 40 & 3.3 & 160 & 6 & 3.2 & 100 & 6.7\\
168: Inter3 & 317 - 327 & 220 & 2.5 & 225 & 5 & 3.7 & 100 & 6.7\\
169: \hline
170: \end{tabular}
171: \caption{ISM parameters for three interarm regions and the Carina and
172: Crux spiral arms. The parameter $r_{out}$ is the outer scale
173: determined from the turnover of the structure functions; $p$ is
174: observed polarization degree, and $\sigma$ is the standard deviation
175: in RM. The outer scale as determined from depolarization is given by
176: $r_{out}^d$, $B_0$ is the parallel component of the regular magnetic
177: field, $C_B^2$ is the amplitude of the magnetic field spectrum given
178: in 10$^{-13}$ m$^{-2/3}$ $\mu$G$^2$, and $B_{ran}$ the resulting
179: random magnetic field strength.\label{t:out}}
180: \end{center}
181: \end{table}
182:
183: \subsection{Depolarization of point sources}
184:
185: %***********************
186: \begin{figure}[t]
187: \centerline{\psfig{figure=figure2.ps,width=\textwidth}}
188: \caption{Top: SGPS field of view, where the circles are RMs of
189: extragalactic sources. RM minimum and maximum is
190: $\pm$1000~rad~m$^{-2}$. Bottom: same field and same sources,
191: but the circles denote degree of polarization. Minimum degree
192: of polarization is 0.4\%, the maximum 13.7\%.}
193: \label{f:depol}
194: \end{figure}
195: %***********************
196: %***********************
197: \begin{figure}[t]
198: \centerline{\psfig{figure=figure3.ps,width=.55\textwidth}}
199: \caption{Polarization degree $p$ against $|$RM$|$ for each source.}
200: \label{f:rm_p}
201: \end{figure}
202: %***********************
203:
204: An independent estimate of the outer scales of the structure can be
205: made from depolarization of extragalactic point sources. This
206: depolarization is caused by variability in polarization angle on
207: angular scales smaller than the size of the (unresolved) source. The
208: variance in polarization angle within a telescope beam will decrease
209: the polarization degree of the source.
210:
211: Variations in polarization angle causing partial depolarization are
212: expected to arise within any polarized extragalactic source
213: itself. Indeed, no source in our sample exhibits the intrinsic maximum
214: degree of polarization of around 70\%, but instead observed
215: polarization degrees are typically under 10\%. However, a Galactic
216: component to this depolarization has been detected as
217: well. Figure~\ref{f:depol} shows RMs in the upper panel and
218: polarization degree $p$ in the bottom panel, and a clear
219: anticorrelation between $|$RM$|$ and $p$ is visible especially at the
220: lower longitudes. This trend is also evident in Figure~\ref{f:rm_p},
221: which shows the degree of polarization of every source as a function
222: of its RM. As the scale of the structure in RM and $p$ is several
223: degrees, this cannot be intrinsic to the sources but instead must be
224: caused by the Galactic ISM.
225:
226: The Galactic component of this depolarization can be estimated
227: assuming a power-law power spectrum of RM fluctuations, in the
228: approximation that the outer scale of structure $r_{out}$ is much
229: larger than the source size $r_{src}$ which is the case here. Adapted
230: from Tribble (1991), the depolarization by a power spectrum of RM
231: fluctuations is given by the degree of polarization $p$ as:
232: \begin{equation}
233: \langle |\frac{p(\lambda)}{p_0}|^2\rangle \approx 1 - 4 \sigma^2
234: \lambda^4 2^{m/2} \left(\frac{r_{src}}{r_{out}}\right)^m
235: \Gamma(1+\frac{m}{2})
236: \label{e:tribble}
237: \end{equation}
238: where $p_0$ is the intrinsic polarization degree of the extragalactic
239: source radiation when it exits the source and $\sigma$, $m$ and
240: $r_{out}$ are defined via the structure function
241: \[
242: D_{\mbox{RM}}(r) = \left\{
243: \begin{array}{ll}
244: 2 \sigma^2 (r/r_{out})^m & \mbox{for}~r < r_{out} \\
245: 2 \sigma^2 & \mbox{for}~r > r_{out}
246: \end{array}
247: \right.
248: \]
249:
250: The average degree of polarization in the studied regions is given in
251: Table~\ref{t:out}. The spiral arms seem to be more depolarized than
252: the interarm regions, with some possible confusion closest to the
253: Galactic center due to superposition of arms and interarm regions
254: along the line of sight.
255:
256: The amount of intrinsic depolarization resulting in polarization
257: degree $p_0$ can be estimated from the extragalactic sources observed
258: in and around the LMC (Gaensler et al.\ 2005) to be 10.4\%, which we
259: assume is the average polarization degree of point sources for which
260: all depolarization is intrinsic. With these assumptions, the
261: depolarization beyond 10.4\% is due to the variations in Galactic RM
262: across the face of the source, which is on average 6~arcsec (Gaensler
263: et al.\ 2005). This percentage is higher than the actual average
264: degree of polarization due to a selection of strong, highly-polarized
265: sources over weak, weakly polarized ones. However, as we are
266: interested in the relative depolarization only, this selection effect
267: does not influence our conclusions.
268:
269: In the spiral arms it is straightforward to use Eq.~(\ref{e:tribble})
270: to obtain the outer scale $r^d_{out}$ needed to obtain the observed
271: depolarization. We assume Kolmogorov turbulence ($m = 5/3$, however,
272: see Section~\ref{s:slope}), and determine the value of the RM standard
273: deviation $\sigma$ from the SF saturation level. The distance is
274: chosen to be the average distance to the region probed, which has a
275: large error due to the large spatial extent of the gas.
276:
277: For the interarm regions we observe a shallow spectrum. Assuming that
278: this spectrum turns over to a steeper Kolmogorov spectrum towards
279: small scales (see Section~\ref{s:amp}), the Kolmogorov slope on small
280: scales will dominate the depolarization of the point sources. Then,
281: using linear fits to the rising parts of the slopes, we can calculate
282: at which $r_{out}$ and $\sigma_{RM}$ the depolarization given by
283: Eq.~(\ref{e:tribble}) equals the observed depolarization in
284: Table~\ref{t:out}. This $r_{out}$ is the outer scale of the Kolmogorov
285: turbulence, i.e.\ the scale at which the Kolmogorov slope turns over
286: into a shallower slope. This scale is given in Table~\ref{t:out} as
287: the $r_{out}^d$ in the interarm regions, and is consistent with the
288: outer scale of Kolmogorov turbulence found in the spiral arms of a few
289: parsecs.
290:
291: \section{Amplitude of magnetic field fluctuations}\label{s:amp}
292:
293: While the turnover in the SF of RM corresponds to the outer scale of
294: structure, the {\it amplitude} of the SF gives information about the
295: magnitude of the magnetic field fluctuations in the medium.
296:
297: Minter \& Spangler (1996, MS96) developed a formalism with which to
298: describe the SF of RM assuming power spectra in magnetic field and in
299: electron density fluctuations which are zero-mean, isotropic and
300: Gaussian. Assuming Kolmogorov turbulence, MS96 find that the RM
301: structure function can be described as:
302: \begin{eqnarray}
303: D_{RM}\!\!\!&=&\!\!\! \left[\!\right. 251.226\left[ \left({\small
304: \frac{n_0}{0.1~\mbox{cm}^{-3}}}\right)^2\! \left({\small
305: \frac{C_B^2}{10^{-13}\mbox{ m}^{-2/3}\mu\mbox{G}^2}}\right)
306: \!+\! \left({\small \frac{B_0}{\mu\mbox{G}}}\right)^2\!
307: \left({\small\frac{C_n^2}{10^{-3} \mbox{ m}^{-20/3}}}\right)\right]
308: \nonumber\\
309: &+&\!\!\!\!23.043\!\left({\small\frac{C_n^2}{10^{-3} \mbox{ m}^{-20/3}}}\right)\!
310: \left({\small \frac{C_B^2}{10^{-13}\mbox{ m}^{-2/3}\mu\mbox{G}^2}}\right)\!
311: \left({\small \frac{l_0}{\mbox{pc}}}\right)^{2/3} \!\!\left.\right]
312: \!*\!\left({\small \frac{L}{\mbox{kpc}}}\right)^{8/3}\!
313: \left({\small \frac{\delta\theta}{\mbox{deg}}}\right)^{5/3}
314: \nonumber
315: \end{eqnarray}
316: where $n_0$ is the mean electron density, $B_{0}$ is the mean magnetic
317: field strength along the line of sight, $l_0$ the outer scale of
318: structure, $L$ the length of the line of sight and the magnetic field
319: and density fluctuations are described by power laws with the same
320: outer scale and spectral index such that
321: \[
322: % \left<\delta n(\mathbf{r_0})\delta n(\mathbf{r_0+r}) \right> &=&
323: % \int d^3q \frac{C_n^2 \mbox{e}^{-i\mathbf{q}\cdot\mathbf{r}}}{(q_0^2+q^2)^{\alpha/2}}\\
324: \left<\delta B_i(\mathbf{r_0})\delta B_i(\mathbf{r_0+r}) \right> =
325: \int d^3q \frac{C_B^2 \mbox{e}^{-i\mathbf{q}\cdot\mathbf{r}}}{(q_0^2+q^2)^{\alpha/2}}
326: \]
327: and a similar expression for $\left<\delta n(\mathbf{r_0})\delta
328: n(\mathbf{r_0+r}) \right>$. Assuming that on smaller scales the
329: magnetic field power spectrum follows the observed Kolmogorov spectrum
330: of electron density, the observed SFs will turn over to steeper slopes
331: towards smaller scales. The constraint that the Kolmogorov SF on the
332: small scales and the shallower SF on larger scales must have the same
333: amplitude at turnover scale $l_0$ yields
334:
335: \[
336: D_{RM}(\delta\theta) = \left\{
337: \begin{array}{ll}
338: A \delta\theta^{5/3} & \mbox{for}~\delta\theta \le l_0/L \\
339: A \delta\theta^{m} (l_0/L) & \mbox{for}~\delta\theta \ge l_0/L
340: \end{array}
341: \right.
342: \]
343:
344: where $m$ is the spectral index of the shallower SF. Following
345: this formalism, we can derive the amplitude of the magnetic field
346: fluctuations in the spiral arms and interarm regions, which is shown
347: as the dotted lines in Figure~\ref{f:sf}, while the input and output
348: parameters for the computation are are given in Table~\ref{t:out}. The
349: mean electron density was determined from the \citet{cl02} electron
350: density model, $B_{\parallel}$ was calculated assuming a constant
351: circular field with a strength $B_{reg}=4~\mu$G \citep{bbm96}, and the
352: outer scale of Kolmogorov turbulence $r_{out}$ is taken as 5~pc in
353: both arms and interarms, consistent with the rough estimates in
354: Table~\ref{t:out}. The path length is chosen as the distance to the
355: point for which 90\% of the electron density along the line of sight
356: is contained in the path length.
357:
358: This procedure allows us to calculate the random magnetic field
359: coefficient $C_B^2$ and the corresponding random magnetic field
360: strengths of about 7~$\mu$G in the spiral arms and 2~$\mu$G in the
361: interarm regions, see Table~\ref{t:out}. Caution needs to be taken
362: that the uncertainties in the input parameters are large so that the
363: magnetic field strength values are also fairly uncertain. However, the
364: analysis indicates that the random magnetic field component is
365: consistent with a constant in all interarm regions much lower than the
366: value in the spiral arms.
367:
368: Random magnetic field strength in the spiral arms exceeding that in
369: the interarms has been observed in some external galaxies (e.g.\
370: NGC~4631, Beck \& Hoernes 1996; IC342, Krause, Hummel \& Beck 1989),
371: although the situation is not clear in the Milky Way. It is expected
372: for spiral galaxies with weak dynamos (Shukurov 1998).
373:
374:
375: \section{Speculations on the nature of the structure}
376: \label{s:slope}
377:
378: If the above assumption is correct and the computed SFs will turn over
379: to a (steeper) Kolmogorov spectrum at smaller scales, several
380: mechanisms can be responsible for creating the shallow slopes.
381:
382: Superposition of two spiral arms with similar spatial outer scales at
383: different distances will yield a shallow transition SF at scales just
384: smaller than the saturation scale. However, lines of sight with more
385: spiral arm superpositions (higher longitude) should give shallower
386: spectra, contrary to what is observed. Furthermore, other observations
387: in the outer Galactic plane (Sun \& Han 2004) and at higher Galactic
388: latitudes (Haverkorn et al.\ 2003) which find shallow slopes without
389: possible spiral arm superpositions argue against this explanation.
390:
391: Discrete structures with internal turbulence within a turbulent
392: medium, such as H~{\sc ii} regions, can also explain the
393: observations. In this case, on small scales (i.e.\ scales smaller than
394: the size of the region) turbulence in the H~{\sc ii} regions would
395: dominate the SF, whereas on larger scales these clouds would just add
396: a constant 'noise' term, which makes the total slope shallower.
397:
398: Shallow SF slopes can also be caused by a transition from 2D to 3D
399: turbulence as suggested by MS96. They invoke physical sheets of gas in
400: which turbulence cannot operate perpendicular to the sheets, which is
401: a logical choice for the region of the sky that they probe, which has
402: a large H~{\sc ii} region near by. For data over a significant part of
403: the plane, this interpretation is not likely.
404:
405: A plausible option is multiple scales of energy input in the interarm
406: regions: for supernova-driven turbulence, the outer scale is believed
407: to be about 100~pc (as observed). However, if energy sources such as
408: stellar winds or outflows, interstellar shocks or H~{\sc ii} regions
409: input a significant amount of energy into the interstellar turbulence
410: on smaller scales (typically parsecs, Mac Low 2004), this may flatten
411: the SF on scales of order 1~pc to scales of order 100~pc, as observed.
412:
413: Alternatively, the power spectrum of magnetic field fluctuations may
414: not follow the Kolmogorov scaling at all. If this is the case, the
415: difference between spiral arms and interarm regions may be due to the
416: absence and presence of a strong regular magnetic field in the arms
417: and interarms, respectively, causing a different spectral index
418: (e.g. Schekochihin et al.\ 2004), or due to a transition between
419: subsonic and supersonic turbulence.
420:
421: \section{Summary and conclusions}
422: \label{s:sum}
423:
424: Faraday rotation measurements of polarized extragalactic sources
425: behind the inner Galactic plane in the fourth quadrant are used to
426: study the characteristics of the magnetized, ionized interstellar
427: medium in the plane, in particular in the spiral arms and in interarm
428: regions. Structure functions show that the typical outer scale of
429: structure in the spiral arms is a few parsecs, whereas in the interarm
430: regions fluctuations up to hundreds of parsecs in size are
431: observed. Partial depolarization of the extragalactic sources by the
432: ISM is used to derive a turbulent outer scale of a few parsecs,
433: assuming Kolmogorov-like turbulence. From the saturation amplitudes of
434: the structure functions the strength of the random magnetic field
435: component in the spiral arms and interarm regions is
436: estimated. Assuming an equal regular magnetic field strength in both
437: arms and interarms, it is found that the random field in the arms is
438: about 3 times stronger than that in the interarm regions. In this
439: case, the random magnetic field dominates in the spiral arms, whereas
440: the regular and random components are similar in the interarm regions.
441:
442:
443: \acknowledgements
444:
445: The ATCA is part of the Australia Telescope, which is funded by the
446: Commonwealth of Australia for operation as a National Facility managed
447: by CSIRO. The author would like to thank Bryan Gaensler, Jo-Anne
448: Brown, Alexander Schekochihin, Stanislav Boldyrev, Steve Spangler, and
449: Joel Weisberg for stimulating discussions. M.H. acknowledges support
450: from the National Radio Astronomy Observatory (NRAO), which is
451: operated by Associated Universities Inc., under cooperative agreement
452: with the National Science Foundation.
453:
454: \begin{thebibliography}{}
455: \bibitem[Armstrong et al.(1995)]{ars95}
456: Armstrong, J. W., Rickett, B. J., \& Spangler, S. R. 1995, ApJ,
457: 443, 209
458: \bibitem[Beck et al.(1996)]{bbm96}
459: Beck, R., Brandenburg, A., Moss, D., et al. 1996, ARA\&A, 34, 155
460: \bibitem[Beck \& Hoernes(1996)]{bh96}
461: Beck, R., \& Hoernes, P. 1996, Nat, 379, 47
462: \bibitem[Brown et al.(2006)]{bhg06}
463: Brown, J. C., Haverkorn, M., Gaensler, B. M., Taylor, A.~R., et al.
464: 2006, ApJL submitted
465: \bibitem[Cordes \& Lazio(2002)]{cl02}
466: Cordes, J.~M., \& Lazio, T.~J.~W. 2002, preprint (astro-ph/0207156)
467: \bibitem[Elmegreen \& Scalo(2004)]{es04}
468: Elmegreen, B.~G., \& Scalo, J. 2004, ARA\&A, 42, 211
469: \bibitem[Gaensler et al.(2005)]{ghs05}
470: Gaensler, B.M., Haverkorn, M., Staveley-Smith, L., et al. 2005,
471: Science, 307, 1610
472: \bibitem[Haverkorn et al.(2006a)]{hgm06}
473: Haverkorn, M., Gaensler, B.~M., McClure-Griffiths, N.~M., et
474: al. 2006a, ApJ, in press
475: \bibitem[Haverkorn et al.(2006b)]{hgb06}
476: Haverkorn, M., Gaensler, B.~M., Brown, J. C., Bizunok, N.,
477: et al. 2006b, ApJL, 637, 33
478: \bibitem[Haverkorn et al.(2003)]{hkb03}
479: Haverkorn, M., Katgert, P., de Bruyn, A. G. 2003, A\&A, 403, 1045
480: %SFs
481: \bibitem[Krause et al.(1989)]{khb89}
482: Krause, M., Hummel, E., \& Beck, R. 1989, A\&A, 217, 4
483: \bibitem[Mac Low(2004)]{m04}
484: Mac Low, M.-M. 2004, Ap\&SS, 289, 323
485: \bibitem[McClure-Griffith et al.(2005)]{mdg05}
486: McClure-Griffiths, N.~M., Dickey, J.~M., Gaensler, B.~M., et
487: al. 2005, ApJS, 158, 178
488: \bibitem[Minter \& Spangler(1996)]{ms96}
489: Minter, A. H., \& Spangler, S. R. 1996, ApJ, 458, 194
490: \bibitem[Mitra et al.(2003)]{mwk03}
491: Mitra, D., Wielebinski, R., Kramer, M., \& Jessner, A. 2003, A\&A,
492: 398, 993
493: \bibitem[Schekochihin et al.(2004)]{sct04}
494: Schekochihin, A.~A., Cowley, S.~C., Taylor, S.~F., et al. 2004, ApJ,
495: 612, 276
496: \bibitem[Shukurov(1998)]{s98}
497: Shukurov, A. 1998, MNRAS, 299, 21
498: \bibitem[Sun \& Han(2004)]{sh04}
499: Sun, X.~H, \& Han, J.~L. 2004, in The Magnetized Interstellar
500: Medium, ed.\ B.~Uyan\i ker, W.~Reich, R.~Wielebinski
501: (Katlenburg-Lindau: Copernicus GmbH), 25
502: \bibitem[Tribble(1991)]{t91}
503: Tribble, P.~C. 1991, MNRAS, 250, 726
504:
505: \end{thebibliography}
506:
507: \end{document}
508: