Astron. Astrophys. 354, 853–873 (2000) ASTRONOMY AND High ...aa.springer.de/papers/0354003/2300853.pdf · Astron. Astrophys. 354, 853–873 (2000) ASTRONOMY AND ASTROPHYSICS High–resolution

Astron. Astrophys. 354, 853–873 (2000) ASTRONOMYAND

ASTROPHYSICS

High–resolution imaging of compact high–velocity clouds

R. Braun1 and W.B. Burton2

1 Netherlands Foundation for Research in Astronomy, P.O. Box 2, 7990 AA Dwingeloo, The Netherlands2 Sterrewacht Leiden, P.O. Box 9513, 2300 RA Leiden, The Netherlands

Received 4 October 1999 / Accepted 22 December 1999

Abstract. Six examples of the compact, isolated Hi high–velocity clouds (CHVCs) identified by Braun & Burton (1999),but only marginally resolved in single–dish data, have been im-aged with the Westerbork Synthesis Radio Telescope. The 65confirmed objects in this class define a dynamically cold sys-tem, with a global minimum for the velocity dispersion of only70 km s−1, found in the Local Group Standard of Rest. Thepopulation is in-falling at 100 km s−1 toward the Local Groupbarycenter. These objects have a characteristic morphology, inwhich one or more compact cores is embedded in a diffuse halo.The compact cores typically account for 40% of the Hi line fluxwhile covering some 15% of the source area. The narrow linewidth of all core components allows unambiguous identifica-tion of these with the cool condensed phase of Hi , the CNM,with kinetic temperature near 100 K, while the halos appear torepresent a shielding column of warm diffuse Hi , the WNM,with temperature near 8000 K. We detect a core with one ofthe narrowest Hi emission lines ever observed, with intrinsicFWHM of no more than 2 km s−1 and 75 K brightness. Froma comparison of column and volume densities for this featurewe derive a distance in the range 0.5 to 1 Mpc. We determinea metallicity for this same object of 0.04 to 0.07 solar. Compa-rably high distances are implied by demanding the stability ofobjects with multiple cores, which show relative velocities aslarge as 70 km s−1 on 30 arcmin scales. Many of the compactcores show systematic velocity gradients along the major axisof their elliptical extent which are well-fit by circular rotation ina flattened disk system. Two out of three of the derived rotationcurves are well-fit by Navarro, Frenk & White (1997) cold darkmatter profiles. These kinematic signatures imply a high dark-to-visible mass ratio of 10–50, forD = 0.7 Mpc, which scalesas1/D. The implied dark matter halos dominate the mass vol-ume density within the central 2 kpc (10 arcmin) of each source,providing a sufficent hydrostatic pressure to allow CNM con-densation. The CHVC properties are similar in many respectsto those of the Local Group dwarf irregular galaxies, exceptingthe presence of a high surface brightness stellar population.

Key words: ISM: clouds – Galaxy: kinematics and dynamics –galaxies: abundances – galaxies: ISM – galaxies: Local Group

Send offprint requests to: R. Braun ([email protected])

1. Introduction

The possibility of an extragalactic deployment of high–velocityclouds has been considered in various contexts by (among oth-ers) Oort (1966, 1970, 1981), Verschuur (1975), Eichler (1976),Einasto et al. (1976), Giovanelli (1981), Bajaja et al. (1987),Burton (1997), Wakker & van Woerden (1997), Blitz et al.(1999), and Braun & Burton (1999). Blitz et al. (1999) inter-pret several general HVC properties in terms of the hierarchicalstructure formation and evolution of galaxies. In this context,the extended HVC complexes would be the nearby objects cur-rently undergoing accretion onto the Galaxy, while the morecompact, isolated ones would be their distant counterparts inthe Local Group environment.

It is striking that the population of anomalous velocityH i splits naturally into two rather distinct components whenobserved with the spatial sampling and sensitivity of modernsurveys, like the Leiden/Dwingeloo Survey (LDS: Hartmann &Burton 1997) in the North and the Parkes Multibeam Survey(see Putman & Gibson 1999) in the South. With this new per-spective, it has become clear that the large majority of the 561objects previously cataloged as distinct HVCs by Wakker & vanWoerden (1991) are the local, low–contrast maxima of extendeddiffuse complexes with angular sizes of tens of degrees. In ad-dition to this diffuse, yet lumpy, component there appear to beseveral hundred intrinsically compact objects with angular sizesof about 1 degree. Objects of intermediate angular size appearto be rather rare.

Braun & Burton (1999) identified and confirmed 65 exam-ples of compact, isolated high–velocity clouds (the CHVCs)which plausibly represent a homogeneous subsample of thehigh–velocity clouds, in a single physical state, and arguablybefore their physical properties have been strongly influencedby the radiation field of the Milky Way or of M31, or by agravitational encounter with one of these major systems. Braunand Burton showed, in particular, that the velocity dispersionof the CHVC sample is minimized in a reference frame consis-tent within the observational errors to the Local Group Standardof Rest. This minimization provides a quantitative demonstra-tion of Local–Group deployment. Within this frame, the CHVCensemble is dynamically quite cold, with a dispersion of only70 km s−1, although strongly infalling into the Local Groupbarycenter at a velocity of about 100 km s−1.

854 R. Braun & W.B. Burton: High–resolution imaging of compact high–velocity clouds

Most HVCs have been identified simply by theiranomalous–velocity Hi emission in total–power surveys at anangular resolution of 0.5 or coarser. Information on charac-teristic intrinsic linear scales, on resolved spectral propertieswhich might reveal, for example, opacity information, and onsuch kinematic properties as intrinsic widths or possible rota-tion, have been largely unavailable. Of the sample of 65 com-pact, isolated HVCs catalogued by Braun & Burton (1999),only two had been subject to interferometric imaging. Wakker& Schwarz (1991) used the Westerbork array to show that bothCHVC 114−10−430 and CHVC 111−06−466 are character-ized by a core/halo morphology, with only about 40% of thesingle–dish flux recovered on angular scales of tens of arcmin,and, furthermore, that the linewidths of the single–dish spectraof these two sources were resolved into components of some 5km s−1 width or less. Both of the imaged systems display sys-tematic velocity gradients along the major axis of an ellipticalH i distribution, which Wakker and Schwarz judged to be sug-gestive of rotation in self–gravitating systems at Local Groupdistances.

If the CHVC objects are in fact a population of unevolvedsub-dwarf galaxies scattered throughout the Local Group, thenthey might reveal some morphological characteristics whichwould not be consistent with the expectations of other suggestedscenarios, in particular for objects ejected by a galactic fountain(e.g. Shapiro & Field 1976; Bregman 1980) or located withinthe Galactic halo (Moore et al. 1999).

The Westerbork imaging discussed in this paper reveals acharacteristic core/halo morphology for the CHVCs, very nar-row linewidths in the cores, and, in many cases, a signatureof rotation. These properties are strongly suggestive of self-gravitating systems at Local Group distances, and specificallyresemble the gaseous components of some dwarf galaxies. Onthe contrary, if the CHVC objects were produced relatively lo-cally by an energetic mechanism responsible for a galactic foun-tain, then their Hi properties might have been expected to in-clude large linewidths, motions not ordered by rotation, and acharacteristic morphology other than that of a core/halo.

High–resolution imaging also makes it possible to providespecific targets for optical observations. Deep optical imag-ing would help clarify the distinction between the CHVCs and(sub–) dwarf galaxies, and any indication of a stellar populationwould allow a direct distance determination. If the CHVCs are atLocal Group distances, then the diffuse Hα emission surround-ing them is expected to be weaker than that associated with high–velocity gas in complexes lying within the halo of the Milky Wayor extending from the Magellanic Clouds (Bland-Hawthorn &Maloney 1999). Furthermore, high–resolution imaging is nec-essary in order to interpret existing observations aimed at detect-ing an HVC in absorption against a star in the halo of the MilkyWay or against an extragalactic background source of contin-uum radiation. In either case, the absorption experiment targetsan extremely small area. It is not clear if the failure to detectabsorption in at least some of the published accounts should beattributed to the HVC in question not in fact covering the contin-uum source, rather than to the metallicity characterisics of the

HVC (for the cases of a negative result toward an extragalacticbackground sources) or to the HVC being at a larger distance(for the cases of a negative result toward halo stars). The re-sults shown below indicate that only a small portion of eachobject is characterized by the high column density Hi emissionfrom the CHVC cores, and thus that selection of suitable probesfor absorption–line measurements must take the actual columndensity distribution into account.

If the compact HVCs discussed here are distributed through-out the Local Group as primordial objects, either surviving asremnants of the formation of the Local Group galaxies, or stillraining in to these systems in a continuing evolution, then suchobjects would also be expected near other groups of galaxies.Several searches for such objects beyond the Local Group arecurrently in progress. The results presented here on the smallangular sizes of the CHVCs generally and the even smaller sur-face covering factor of the high column density cores must beconsidered when designing an optimum search strategy or whencalculating the expected detection statistics for a given experi-ment directed at other galaxies or groups of galaxies. A conclu-sive detection experiment of this type has not yet been carriedout.

Our discussion is organized as follows. We begin by de-scribing the method of sample selection in Sect. 2, proceed witha description of the newly acquired observations in Sect. 3 andcontinue with a presentation and discussion of our results inSects. 4 and 5. We close by briefly summarizing our results inSect. 6.

2. Sample selection

Because high–resolution Hi imaging had previously only beenobtained for two objects of the CHVC class, we chose a sam-ple from the catalog of Braun & Burton (1999) which spanneda wide range of the source properties, paying particular atten-tion to source position both on the sky and in radial velocity, aswell as to the physical attributes of linewidth and total Hi fluxdensity. Although only six CHVC sources were imaged in ourprogram, the sources are distributed widely in galactic coordi-nates, span radial velocities of−275 < vLSR < +165 km s−1,vary in linewidth from 6 to 95 km s−1, and in line flux from25 to 300 Jy km s−1. The source properties are summarized inTable 1.

3. Data

Observations of the six CHVC fields were obtained with theWSRT array between 12/11/98 and 17/12/98. One twelve–hourintegration was obtained for each field in the standard WSRTarray configuration having a shortest baseline of 36 meters. On–source integration was bracketed by half–hour observations ofthe calibration sources 3C286 and 3C48. At the time of theobservations, 11 of the 14 telescopes of the array were equippedwith the upgraded receivers having a system temperature ofabout 27 K in the 1150 to 1850 MHz band. (As of 11/3/99the entire array is equipped with a receiver package covering

R. Braun & W.B. Burton: High–resolution imaging of compact high–velocity clouds 855

Table 1.Compact, isolated high–velocity clouds imaged with the WSRT.

Name RA(2000) Dec(2000) LDS flux LDS FWHM LDS structureCHVC lll ± bb ± vvv (h m) ( ′) (Jy km s−1) (km s−1) (a× b @ PA)

CHVC 069+04−223 19 50.1 33 41 86 34.0 1.0 × 0.9 @ 330

CHVC 115+13−275 22 56.9 74 33 105 95.4 1.1 × 0.6 @ 114

CHVC 125+41−207 12 24.0 75 36 235 5.9 1.3 × 0.7 @ 292

CHVC 191+60+093 10 36.9 34 10 66: 29.6 1.0 × 0.9 @ 170

CHVC 204+30+075 08 27.5 20 09 305: 33.9 0.8 × 0.6 @ 215

CHVC 230+61+165 10 55.2 15 28 26: 29.3 1.4 × 0.7 @ 182

the range from 250 MHz to 8 GHz with eight feed systemsmounted in a prime–focus turret.) The interim WSRT correlatorsystem available at the time was used to provide 256 uniformlyweighted spectral channels in two linear polarizations across2.5 MHz centered on thevLSR velocity of each source. Theeffective velocity resolution was 1.2 times the channel spacingof 2.06 km s−1.

Standard gain and bandpass calibrations were carried outafter editing the data and deleting all baselines effected by shad-owing. A component model for the continuum emission fromeach field was derived from an image made from the averageof the line–free spectral channels. This model was subtracteddirectly from the visibility data. The block of spectral channelscontaining line emission was imaged with a visibility–basedclean deconvolution proceeding down to a flux level of twicethe rms noise level at a variety of spatial resolutions. Uniformweighting of the visibility data was employed, together with aseries of Gaussian tapers decreasing to 30% amplitude at pro-jected baselines of 1.25, 2.5, 5, and 10 kλ, resulting in spatialresolutions of about 120, 60, 30, and 20 arcsec, respectively. Ina few cases, some residual continuum emission was still presentin the data cubes; in those cases, several spectral channels fromboth edges of the cube were averaged together and subtractedfrom the entire cube.

The typical rms noise levels in the deconvolved cubes were2.0, 1.7, 1.4, and 1.3 mJy per beam per spectral channel atthe four spatial resolutions noted above. The correspondingbrightness sensitivities were, respectively, 0.085, 0.29, 0.95,and 2.0 K in a single channel of 2.06 km s−1 width. (Fluxper beam and brightness temperature are related as usual byS = 2kBTBΩB/λ2, orSmJy/Beam = 0.65ΩasTB/λ2

cm for Ωas,the beam area in arcsec2.) Expressed as an optically thin Hi col-umn density, the sensitivities correspond to 0.32, 1.1, 3.6, and7.4×1018 cm−2, respectively, for emission filling the beam andextending over a single velocity channel. While some Hi emis-sion profiles originating in the cool atomic phase are indeedas narrow as the single–channel velocity width (as we willsee below), emission profiles originating in the warm neutralphase will have substantially broader linewidths. For example,the thermal linewidth of a component with kinetic temperature,Tk = 8000 K, will be 21 km s−1 FWHM. At this velocityresolution, our column density sensitivity is degraded to 1.0,3.5, 11, and 24×1018 cm−2. This somewhat counter–intuitivecorrespondence of column density sensitivity with the expected

linewidth must be borne in mind. Narrow lines are significantlyeasier to detect than broad ones.

Moment images (zero, first, and second) were generatedfrom each cube after employing a blanking criterion for inclu-sion of each pixel in the weighted sum. This involved demandinga brightness in excess of about 2σ after smoothing the cube bya factor of three both spatially and in velocity. Images of inte-grated emission were corrected for the primary–beam responseof the instrument, which is well–approximated (at 1420 MHz)by a circular Gaussian with 2110 arcsec FWHM.

While structures extending over as much as 10 arcmin ina single spectral channel were adequately recovered in the re-sulting images, there were also indications of the presence ofmore diffuse features which could not be adequately imaged.These were apparent due to the artifacts they induced; specifi-cally, the so–called short–spacing bowl surrounding regions ofbright extended emission. Even with careful windowing duringthe deconvolution it was not possible to completely eliminatesuch artifacts. Consequently, a significant fraction of the totalflux detected in the single–dish observations could not be re-covered. The integrated Hi flux detected in the reconstructedimages after primary beam correction varied from less than 1%to as much as 55% of that detected in the LDS. The percentagesof total fluxes contributed by the dense cores are indicated foreach observed field in Table 2.

A crude attempt was made to correct the interferometric im-ages of integrated Hi for the non–detection of diffuse emissionfeatures. An elliptical Gaussian with dimensions and orienta-tion as measured in the LDS and an integral flux sufficient torecover the LDS total was added to the interferometric images.These parameters are summarized for the six fields in Table 2.The resulting composite images are presented in turn below. Thelowest column–density contours in these composite images fol-low the elliptical outlines of the LDS source model, while thehigher levels are dominated by the compact structures detectedinterferometrically. Although not appropriate for detailed anal-ysis on spatial scales between about 10 and 30 arcmin, thesecomposite images are consistent with all of the data currentlyin hand and are indicative of the column densities likely to bepresent in diffuse structures as well as their overall spatial ex-tent. In at least one case (discussed below) we were able to verifywith independent data that the resulting composite provided areasonable representation of the source structure and extent.


Table 2.Properties of CHVCs imaged with the WSRT.

Name WSRT flux Core flux Core area DistanceCHVC lll ± bb ± vvv (Jy km s−1) % % (kpc)

CHVC 069+04−223 21.9 26 3.4 >(40–50)CHVC 115+13−275 20.7 20 10. >(40-100)CHVC 125+41−207 130. 55 25 500–1000CHVC 191+60+093 0.22 0.3 0.2 >7:CHVC 204+30+075 116. 38 48 >(40–130)CHVC 230+61+165 0.65 2.5 0.3 >4:

4. Results

In this section, we briefly summarize the structural and kine-matic data obtained for each of the six observed fields using theWSRT.

4.1. CHVC 069+04−223

The moment images and a few representative spectra for thisfield are shown at 1 arcmin resolution in Fig. 1. The integratedH i distribution is dominated by a bright elongated concentra-tion (clump A) of some 15 arcmin extent. Several fainter knotslie about 10 arcmin further to the South, and are connected toeach other by a low brightness bridge of emission (more clearlyseen in Fig. 2). Spectra toward some of the brightest emissionregions all have (i) peak brightnesses of about 2 K, (ii) relativelybroad line profiles of about 15 km s−1 FWHM, and (iii) a cen-troid velocity near−235 km s−1. At several localNHI maximawithin clump A the profiles become doubly peaked. These willbe referred to again below. The velocity field displays at leastone systematic trend: a moderate velocity gradient, orientedalong PA 30 (East of North) along the major axis of clump Afrom about−230 to−240 km s−1 and extending over some10 arcmin.

The distribution of profile linewidth shows that clump A hasa moderately high linewidth throughout, but also has severaldistinct regions with apparent dispersions as high as 8 km s−1.These correspond to regions of line splitting rather than to sim-ple line broadening, while the overall velocity centroid is un-changed. Two such regions are observed, centered near(α, δ) =(19h49m50s, 3352′) and(α, δ) = (19h49m55s, 3354′). Thetwo peaks in the line profile are separated by about 10 km s−1

in both cases, while the angular extent is only some 90 arcsec.The integrated Hi distribution in CHVC 069+04−223, af-

ter inclusion of a simple representation of the diffuse emisisonderived from the LDS data, is shown in Fig. 2. The brightestknots have column densities just in excess of 1020 cm−2, whilethe diffuse underlying envelope peaks at about5 × 1018 cm−2.

4.2. CHVC 115+13−275

The basic data for this field are shown at 2 arcmin resolution inFig. 3. This object is resolved in the WSRT data into a collec-tion of approximately 10 sub–structures. Each clump is betweenabout 1 and 10 arcmin in size, while the ensemble of clumps

is distributed over a region of about 30 arcmin diameter. Spec-tra towards several of the clumps show that each feature has apeak brightness of about 1 K and an intrinsic velocity width ofbetween 10 and 15 km s−1 FWHM. On the other hand, each fea-ture has a distinct centroid velocity so that the collection spansthe interval of−300 to−220 km s−1. The global distributionof the line–of–sight velocities shows no obvious pattern. High,low, and intermediate velocity clumps are distributed across thefield. Some of the most extreme velocities do occur in closespatial proximity, however, near(α, δ) = (23h55m, 7438′);but comparable velocities are also seen elsewhere.

More careful examination of the individual clumps revealsthat many of them have significant velocity gradients orientedpreferentially along their long axes. This is true for the clumpscentered near(α, δ) = (22h55m, 7432′), (23h01m, 7427′),(22h57.m5, 7442′), and(22h57.m5, 7415′). The typical gra-dient observed is some 10 km s−1 over an angular extent of5 arcmin.

The image of velocity dispersion shown in the lower–leftpanel of Fig. 3 serves primarily to illustrate the regions wheremultiple velocity components overlap along the line of sight.The morphology of the regions having multiple–velocity com-ponents does not seem peculiar, but instead suggests simplyline–of–sight overlap of distinct features rather than a physicalinteraction.

The composite image of integrated Hi inCHVC 115+13−275 is shown in Fig. 4. Column densitiesin the knots do not exceed about5 × 1019 cm−2, whilethe broad underlying distribution peaks between 5 and10 × 1018 cm−2

4.3. CHVC 125+41−207

The compact high–velocity cloud CHVC 125+41−207 is par-ticularly interesting. Because of the high brightness of thissource, higher angular resolution could be employed for thedata presentation. Basic data at 28 arcsec resolution are shownin Fig. 5. The source has a complex filamentary morphology ex-tending over some 45 arcmin. Several compact cores of only 1 to2 arcmin extent are seen within the Eastern segment (clump A)of the complex. The spectrum toward the brightest compact coreis quite remarkable, in having a peak brightness of 75 K as wellas a linewidth which is completely unresolved with our effec-tive velocity resolution of 2.47 km s−1. This width is one ofthe narrowest ever measured for the Hi emission line. Velocity


CHVC069+04-223 MOM1

Grey scale range= -245 to -215 km/sPeak contour = -2.4645E+05 m/s Levs = 1.000E+03 * (-240, -236, -232, -228, -224,-220, 0)

DE

CL

INA

TIO

N (

J200

0)

RIGHT ASCENSION (J2000)19 51 00 50 30 00 49 30 00

34 05

00

33 55

50

45

40

35

30

-240 -230 -220

CHVC069+04-223 MOM0

Grey scale flux range= -10.0 100.0 10^18CM2Peak contour flux = 2.6232E+02 10^18CM2Levs = 1.000E+00 * (20, 40, 60, 80, 100)

DE

CL

INA

TIO

N (

J200

0)


34 05

00

33 55

50

45

40

35

30

0 50 100

CHVC069+04-223 MOM2

Grey scale range, sigma = 1 to 9 km/s

DE

CL

INA

TIO

N (

J200

0)


34 05

00

33 55

50

45

40

35

30

2 4 6 8

RA: 19 50 18.895 DEC: 33 54 37.30

KE

LV

IN

VELO-LSR km/s-260 -250 -240 -230 -220

2.2

1.9

1.6

1.2

0.9

0.6

0.3

0.0

RA: 19 50 49.247 DEC: 33 33 57.69

KE

LV

IN

VELO-LSR km/s-260 -250 -240 -230 -220

1.9

1.6

1.4

1.1

0.8

0.5

0.3

0.0

-0.3

RA: 19 49 53.192 DEC: 33 53 37.65

KE

LV

IN

VELO-LSR km/s-260 -250 -240 -230 -220

2.0

1.7

1.4

1.1

0.8

0.6

0.3

0.0

C

B

A

B

C

A

-232

-236

-228

Fig. 1. Imaging data for CHVC 069+04−223 at 1 arcmin and 2 km s−1 resolution. Upper left panel: apparent integrated Hi (assuming negligibleopacity), with contours at 20, 40, 60, 80, and 100×1018 cm−2 and a linear grey–scale extending from−10 to 100×1018 cm−2. Upper rightpanel: brightness temperature spectra at the indicated positions. Lower left panel: intensity weighted line–of–sight velocity,vLSR, with contoursat −240, −236, −232, −228, −224, and−220 km s−1 and a linear grey–scale extending from−245 to −215 km s−1. Lower right panel:intensity weighted distribution of squared velocity, corresponding to the velocity dispersion of a Gaussian profile, with a linear grey–scaleextending from 1 to 9 km s−1.

channels adjacent to the line peak have intensities down to about20% of maximum while being spaced by only 2.06 km s−1. Aspectrum toward the more diffuse western knot (clump B) has amore modest peak brightness of some 30 K and a linewidth ofabout 6 km s−1 FWHM. The most prominent systematic trend

in the velocity field is a North–South gradient over clump Bof about 8 km s−1 over 10 arcmin. Clump A has a shallowerNorth–South velocity gradient of about half this magnitude. Thefainter filaments have velocities which do not follow this simplepattern.


CHVC069+04-223 MOM0+DWL

Grey scale flux range= 0.0 100.0 10^18CM2Peak contour flux = 1.1676E+02 10^18CM2Levs = 1.000E+00 * (1, 2, 5, 10, 20, 50, 100)

DE

CL

INA

TIO

N (

J200

0)

RIGHT ASCENSION (J2000)19 53 52 51 50 49 48 47

34 20

10

00

33 50

40

30

20

10

00

0 50 100

Fig. 2.Column density distribution of Hi inCHVC 069+04−223 at 1 arcmin resolutionreconstructed from LDS and WSRT data.Contours are drawn at 1, 2, 5, 10, 20, 50,and 100×1018 cm−2; a linear grey–scaleextends from0 to 100×1018 cm−2.

The image of velocity dispersion delineates those regionswith extremely narrow linewidth from those which are onlymoderately narrow. Both of the compact cores in the Easterncomplex as well as several other more diffuse structures arecompletely unresolved in velocity. Clump B is less extremein this regard. One small region of clump B, centered near(α, δ) = (12h23m18s, 7530′), deserves special mention. Thisminor local maximum in theNHI distribution displays clearvelocity splitting in the emission profile, amounting to some5 km s−1 over a region of about an arcmin in diameter, as shownin Fig. 6. A second region displaying a similar phenomenon iscentered near(α, δ) = (12h21m35s, 7532′). Both regions areabout 90 arcsec in diameter.

The composite distribution ofapparentH i column densityis shown in Fig. 7. The emphasis on the term “apparent” is par-ticularly apt: as we will see below, the compact cores are likelyto have opacities of about 2, so that actual column densities areabout twice the apparent ones in these directions. Even so, theapparent column density exceeds4 × 1020 cm−2 over much ofthe source, while the underlying halo reaches values of morethan5 × 1018 cm−2 near the knots.

In this particular case, we can obtain some quality assess-ment of our simple composite image ofNHI, in which we have

attempted to compensate for the limited sensitivity to diffusestructures inherent in our interferometric data. This source hasrecently been observed with the Lovell Telescope at JodrellBank by de Vries et al. (1999). Comparison of the5×1018 cm−2

contour in the Jodrell Bank image with our composite imageshows good agreement in both source size and orientation.

4.4. CHVC 204+30+075

The data for this compact HVC are shown at 1 arcmin res-olution in Fig. 8. Several moderately bright clumps are seenover a region of some 35 arcmin in size. The large angular ex-tent of this object in a given velocity channel challenges recon-struction of the brightness distribution with the single 12–hourWSRT coverage which we obtained. Some questions remain re-garding the reconstruction fidelity, particularly in regard to themost Northerly feature which extends down to the 10% levelof the primary beam, and as such, will have had any imperfec-tions magnified by almost a factor of 10 in the displayed result.The spectra toward the more compact local maxima have peakbrightnesses between 2 K and 8 K and linewidths which aregenerally some 15 km s−1 FWHM. Doubly–peaked spectra areseen in two regions; in one case where two components seem to


CHVC115+13-275 MOM1

Grey scale range = -305 to -225.0 km/sPeak contour = -3.1828E+05 m/s Levs = 1.000E+03 * (-300, -290, -280, -270, -260,-250, -240, -230, 0)

DE

CL

INA

TIO

N (

J200

0)


74 45

40

35

30

25

20

15

10

-300 -280 -260 -240CHVC115+13-275 MOM2


DE

CL

INA

TIO

N (

J200

0)


74 45

40

35

30

25

20

15

10

10 20 30

CHVC115+13-275 MOM0

Grey scale flux range= -5.00 50.00 10^18CM2Peak contour flux = 5.9183E+01 10^18CM2Levs = 1.000E+00 * (5, 10, 20, 40)

DE

CL

INA

TIO

N (

J200

0)


74 45

40

35

30

25

20

15

10

0 20 40

RA: 23 00 29.758 DEC: 74 26 12.46

KE

LV

IN

VELO-LSR km/s-320 -300 -280 -260 -240 -220

1.7

1.3

1.0

0.7

0.3

0.0

-0.3

RA: 22 55 05.291 DEC: 74 38 37.59

KE

LV

IN

VELO-LSR km/s-320 -300 -280 -260 -240 -220

1.2

1.0

0.7

0.5

0.2

0.0

RA: 22 57 31.379 DEC: 74 40 39.18

KE

LV

IN

VELO-LSR km/s-320 -300 -280 -260 -240 -220

1.2

0.9

0.7

0.5

0.2

0.0

-0.2

A

C

B

A

C

B

-270

-250-280

-290

-270

-250

-260

-260

-270

-240-250

-260

Fig. 3. Imaging data for CHVC 115+13−275 at 2 arcmin and 2 km s−1 resolution. Upper left panel: apparent integrated Hi (assumingnegligible opacity), with contours at 5, 10, 20, and 40×1018cm−2 and a linear grey–scale extending from−5 to 50×1018cm−2. Upper rightpanel: brightness temperature spectra at the indicated positions. Lower left panel: intensity weighted line–of–sight veloctiy,vLSR, with contoursat −300 to−230 km s−1 in steps of 10 km s−1 and a linear grey–scale extending from−305 to −225 km s−1. Lower right panel: intensityweighted distribution of squared velocity, corresponding to the velocity dispersion of a Gaussian profile, with a linear grey–scale extending from1 to 35 km s−1.

overlap along the line of sight and in another where localized linesplitting is observed (as noted below). The large elliptical featurein the South–central part of the field (clump A) shows a well–defined North–South velocity gradient running from about 55 to80 km s−1 over some 12 arcmin. The large elliptical feature in

the North–East (clump B) has a similar gradient along PA−35

running from 45 to 70 km s−1 over 20 arcmin. Four other majorclumps with more nearly discrete line–of–sight velocities fill outthis system. Velocity dispersions are typically substantially lessthan 10 km s−1, except near(α, δ) = (08h28m30s, 20) where



Grey scale flux range= 0.00 50.00 10^18CM2Peak contour flux = 6.9349E+01 10^18CM2Levs = 1.000E+00 * (1, 2, 5, 10, 20, 50)

DE

CL

INA

TIO

N (

J200

0)

RIGHT ASCENSION (J2000)23 05 00 22 55 50

75 10

00

74 50

40

30

20

10

00

0 20 40

Fig. 4. Column density distribution of Hi inCHVC 115+13−275 at 2 arcmin resolutionreconstructed using LDS and WSRT data.Contours are drawn at 1, 2, 5, 10, 20, and 50×1018 cm−2 and a linear grey–scale extendsfrom 0 to 50×1018 cm−2.

clump A appears to overlap with the smaller South–Easterlyclump, and near(α, δ) = (08h26m06s, 2017′30′′) where twovelocity peaks separated by about 20 km s−1 are seen over the60 arcsec extent of this local maxima inNHI.

Column densities of a few times1020 cm−2 are seen to-ward the brighter knots in the composite reconstruction shownin Fig. 9. A very substantial diffuse component is inferred forthis field, reaching column densities of a few times1019 cm−2

over about 30 arcmin near the source centroid.

4.5. CHVC 191+60+093 and CHVC 230+61+165

The two final sources in our sample are both quite faint andprimarily diffuse, albeit small in angular size. The WSRT dataat 2 arcmin resolution are shown in Fig. 10. Low–brightnessclumps of 5 to 10 arcmin extent are detected in both thesecases. Peak brightnesses of less than 0.5 K are seen, coupledwith moderately narrow linewidths of some 6 km s−1 FWHM.The Northern clump of CHVC 191+60+093 has a moderatelystrong velocity gradient along PA= 0 extending from 87km s−1 to 95 km s−1 over 5 arcmin. The velocity field ofCHVC 230+61+165 shows no clear systematic trend. Thecomposite reconstructions shown in Fig. 11 and Fig. 12 indicate

peak column densities of only about1019 cm−2, embedded ina diffuse component of only 2–5×1018 cm−2.

5. Discussion

5.1. Overall morphology and kinematics

The six compact, isolated high–velocity clouds which we haveimaged with moderately high spatial resolution span a widerange of source morphologies, but nevertheless share severalnoteworthy properties. In each case we detect a number of rel-atively compact clumps which range in size from a few arcminto perhaps 20 arcmin in extent. These are local enhancementsin the column density distribution with peak column densitiesin the range1019 to 1021 cm−2. Each clump is characterized byits own line–of–sight velocity and relatively narrow line widthsat 60 arcsec resolution in the range of 2 to 15 km s−1 FWHM.At the narrow linewidth extreme, very stringent limits on theH i kinetic temperature are implied which we discuss below.But even linewidths as large as 15 km s−1 FWHM are signifi-cantly less than the 21 km s−1 FWHM which corresponds to the8000 K equilibrium temperature of the Warm Neutral Medium(WNM). Based on these modest linewidths we can already con-


CHVC125+41-207 MOM1

Grey scale range = -212 to -198 km/sPeak contour = -2.2082E+05 m/s Levs = 1.000E+03 * (-210, -208, -206, -204, -202,-200, 0)

DE

CL

INA

TIO

N (

J200

0)

RIGHT ASCENSION (J2000)12 30 28 26 24 22 20

75 45

40

35

30

25

20

15

-210 -205 -200 CHVC125+41-207 MOM2

Grey scale range, sigma = 0.5 to 5.0 km/s

DE

CL

INA

TIO

N (

J200

0)


75 45

40

35

30

25

20

15

1 2 3 4 5

CHVC125+41-207 MOM0

Grey scale flux range= -30.0 300.0 10^18CM2Peak contour flux = 4.9352E+02 10^18CM2Levs = 1.000E+00 * (50, 150, 250, 350)

DE

CLI

NA

TIO

N (J

2000

)


75 45

40

35

30

25

20

15

0 100 200 300

RA: 12 26 48.059 DEC: 75 31 36.53

KE

LV

IN

VELO-LSR km/s-225 -220 -215 -210 -205 -200 -195

67

53

40

27

13

0

RA: 12 22 37.310 DEC: 75 31 39.55

KE

LV

IN

VELO-LSR km/s-225 -220 -215 -210 -205 -200 -195

28

22

17

11

6

0

A B

A B

-206-208

-204

-206

-202

-200-202

-204

Fig. 5. Imaging data for CHVC 125+41−207 at 28 arcsec and 2 km s−1 resolution. Upper left panel: apparent integrated Hi (assuming negligibleopacity), with contours at 50, 150, 250, and 350×1018 cm−2 and a linear grey–scale extending from−30 to 300×1018 cm−2. Upper rightpanel: brightness temperature spectra at the indicated positions. The extremely narrow linewidths — the spectrum shown on the left in this panelis one of the narrowest Hi emission lines ever observed — robustly constrain the kinetic temperatures. Lower left panel: intensity weightedline-of-sight veloctiy,vLSR, with contours at−210, −208, −206, −204, −202, and−200 km s−1 and a linear grey–scale extending from−212to −198 km s−1. Lower right panel: intensity weighted distribution of squared velocity, corresponding to the velocity dispersion of a Gaussianprofile, with a linear grey–scale extending from 0.5 to 5 km s−1.

clude that the Hi in the CHVC clumps must be in the form ofthe Cool Neutral Medium (CNM) with typical equilibrium tem-peratures in the range of 50 to 200 K (e.g. Wolfire et al. 1995a,1995b). Each clump has a smoothly varying line–of–sight ve-locity across its extent, often in the form of a systematic gradientalong the major axis of an elliptical distribution. We will addressthe best–resolved cases of velocity gradients below.

Most of the imaged CHVC systems are composed of mul-tiple clumps. Since each clump has a distinct velocity, the total

CHVC linewidth is determined primarily by the velocity spreadbetween the clumps, and to a lesser degree by systematic veloc-ity gradients, rather than by the intrinsic linewidth of a singleentity. There are a few cases where there appears to be par-tial line–of–sight overlap of two distinct clumps, leading to adoubly–peaked emission profile within a single 60 arcsec beam.In the most extreme case studied so far, CHVC 115+13−275,the inter–clump velocity spread amounts to some 70 km s−1.This case will also be addressed in more detail below.


DE

CL

INA

TIO

N (

J200

0)

RIGHT ASCENSION (J2000)12 23 40 35 30 25 20 15 10 05

75 31 00

30 45

30

15

00

29 45

30

15

00

28 45

CHVC125+41-207

KE

LV

IN

-220

-200

20100

Fig. 6. Brightness temperature spectrafor a small region (2×2 arcmin) ofCHVC 125+41−207 at 28 arcsec and2 km s−1 resolution. Note the localized re-gion of line splitting extending over about1 arcmin.

In addition to the clumps, we see evidence, in all cases stud-ied, for a substantial diffuse halo component of Hi emission.The clumps account for as little as 1% to as much as 55% ofthe total Hi emission which was detected in the single–dishobservations, although a typical flux fraction is about 40%, aswas also found previously by Wakker and Schwarz (1991) forthe two CHVCs which they imaged. The halos have typical spa-tial FWHM of about 1 degree and Hi column densities withinthe inner 30 arcmin between 2×1018 and 2×1019 cm−2. InTable 2 we also include an indication of the surface area of thehigh–colum–density cores relative to the diffuse halos. This wasdefined as the ratio of the area with column density exceeding5×1018cm−2 in the WSRT-only images to the source area de-termined in the LDS. These values vary over a wide range, butare typically lower by a factor of about two from the fractionalflux, with a mean of 15%.

The data currently available suggest a characteristic two–phase structure. The diffuse morphology of the halo is consistentwith a component of WNM shielding the embedded clumps ofCNM. Thermodynamic modeling of the Hi (which we addressin the next subsection) actually requires such a nested geome-try of the warm and cool Hi phases when the entire structureis immersed in an isotropic radiation field. Comparison of theH i emission spectra observed interferometrically with thoseseen in total power supports this conjecture, since the diffusecomponents generally have linewidths exceeding the 21 km s−1

FWHM thermal linewidth of an 8000 K gas. This remains a dif-ficult issue to address in full with the limited data currently inhand. High–resolution, total–power imaging with the up–graded

Arecibo telescope would allow direct assessment of the degreeof sub–structure as well as the local lineshape of the halo com-ponent. This could enable unambiguous determination of thenature of the halo gas.

5.2. Narrow linewidths and opaque cores inCHVC 125+41−207

A tight correlation of Hi emission brightness temperature withH i opacity has been established for both the Galaxy and forM31 (Braun & Walterbos 1992); a comparable correlation per-tains for the resolved high–brightness Hi structures seen inmany nearby galaxy disks (Braun 1997, 1998) which have atypical size of about 150 pc and a velocity FWHM of lessthan about 5 km s−1. The high brightness temperatures seenin CHVC 125+41−207, amounting to about 75 K, are thereforealready an indication that relatively high Hi opacities are likelyto be present. In this case a lower limit to the Hi opacity inthe bright cores can be inferred by comparing the kinetic tem-perature which follows from the linewidth with the emissionbrightness temperature. As illustrated in Fig. 5, the linewidth inthe cores of clump A is completely unresolved with the avail-able 2.47 km s−1 resolution. An upper limit to the intrinsiclinewidth of about 2 km s−1 FWHM corresponds to an upperlimit on the kinetic temperature of 85 K. For an isothermal gasthe Hi brightness temperature,TB, is related to the spin tem-perature,TS, by TB = TS(1 − e−τ ); assuming equality of thekinetic and spin temperatures allows a lower limit on the opac-ity to be derived ofτ ≥ 2. This must be regarded as a firm



Grey scale flux range= 0.0 300.0 10^18CM2Peak contour flux = 4.4419E+02 10^18CM2Levs = 1.000E+00 * (1, 2, 5, 10, 20, 50, 100, 200,400)

DE

CL

INA

TIO

N (

J200

0)

RIGHT ASCENSION (J2000)12 30 25 20 15

76 10

00

75 50

40

30

20

10

00

0 100 200 300

Mrk 205

Fig. 7. Column density distribution of Hi in CHVC 125+41−207 at 28 arcsec resolution reconstructed from the LDS and WSRT data. Contoursare drawn at 1, 2, 5, 10, 20, 50, 100, 200, and 400×1018 cm−2; a linear grey–scale extends from0 to 300×1018 cm−2.

lower limit since any line broadening by turbulence, which thelinewidth observed already constrains to be of order 1 km s−1

or less, would only serve to bring the implied kinetic tempera-ture even closer to the observed brightness temperature, yieldinga higherτ . Similarly, excitation effects which might suppress

the Hi spin temperature with respect to the kinetic temperaturewould only serve to exasperate this problem. (The interestingpossibilities for constraining ISM physics with this remarkablesource have prompted us to request new WSRT observationswith substantially higher velocity resolution.)


CHVC204+30+075 MOM1

Grey scale range = 30 to 90 km/sPeak contour = 9.8716E+04 m/s Levs = 1.000E+03 * (40, 45, 50, 55, 60, 65, 70,75, 80, 85)

DE

CL

INA

TIO

N (

J200

0)

RIGHT ASCENSION (J2000)08 29 30 00 28 30 00 27 30 00 26 30 00 25 30

20 40

30

20

10

00

19 50

40 60 80CHVC204+30+075 MOM2


DE

CL

INA

TIO

N (

J200

0)


20 40

30

20

10

00

19 50

2 4 6 8 10

CHVC204+30+075 MOM0

Grey scale flux range= -10.0 150.0 10^18CM2Peak contour flux = 3.2950E+02 10^18CM2Levs = 1.000E+00 * (20, 50, 100, 200, 300)

DE

CL

INA

TIO

N (

J200

0)


20 40

30

20

10

00

19 50

0 50 100 150RA: 08 28 02.746 DEC: 20 36 40.43

KE

LV

IN

VELO-LSR km/s30 40 50 60 70 80 90

8.4

7.2

6.0

4.8

3.6

2.4

1.2

0.0

-1.2

RA: 08 27 39.922 DEC: 20 07 20.00

KE

LV

IN

VELO-LSR km/s30 40 50 60 70 80 90

2.1

1.6

1.0

0.5

0.0

RA: 08 28 22.492 DEC: 20 01 54.92

KE

LV

IN

VELO-LSR km/s30 40 50 60 70 80 90

2.1

1.8

1.4

1.1

0.7

0.4

0.0

-0.4

RA: 08 27 17.213 DEC: 19 58 57.57

KE

LV

IN

VELO-LSR km/s30 40 50 60 70 80 90

2.0

1.3

0.7

0.0

-0.7

A B

C D

A

C

B

D

6570 75

60

55

8085

7065

6055

5040

Fig. 8. Imaging data for CHVC 204+30+075 at 1 arcmin and 2 km s−1 resolution. Upper left panel: apparent integrated Hi (assuming negligibleopacity), with contours at 20, 50, 100, 200, and 300×1018 cm−2 and a linear grey–scale extending from−10 to 150×1018 cm−2. Upper rightpanel: brightness temperature spectra at the indicated positions. Lower left panel: intensity weighted line–of–sight velocity,vLSR, with contoursat 40 to 85 km s−1 in steps of 5 km s−1 and a linear grey–scale extending from 30 to 90 km s−1. Lower right panel: intensity weighted distributionof squared velocity, corresponding to the velocity dispersion of a Gaussian profile, with a linear grey–scale extending from 1 to 10 km s−1.

Plausible lower limits to the actual Hi column density to-ward the two compact cores in clump A of CHVC 125+41−207are therefore about1021 cm−2 (where the apparent column issome4.5 × 1020 cm−2). The angular size of these features canbe determined approximately from the apparent column den-

sity distribution shown in Fig. 5. Mean spatial FWHM extentsof about 90 arcsec are observed, although their flat–topped ap-pearance suggests that opacity effects may already be partiallyresponsible for this measured size. The FWHM of the actual col-umn density distribution is likely to be somewhat smaller. These


CHVC204+30+075 MOM0+DWL

Grey scale flux range= 0.0 300.0 10^18CM2Peak contour flux = 3.4062E+02 10^18CM2Levs = 1.000E+00 * (1, 2, 5, 10, 20, 50, 100, 200)

DE

CL

INA

TIO

N (

J200

0)


20 40

30

20

10

00

19 50

40

30

0 100 200 300

Fig. 9. Column density distribution of Hi inCHVC 204+30+075 at 1 arcmin resolutionreconstructed from LDS and WSRT data.Contours are drawn at 1, 2, 5, 10, 20, 50,100, and 200×1018 cm−2; a linear grey–scale extends from0 to 300×1018 cm−2.

compact cores are seen in projection against more extended re-gions with apparent column densities of 1 to 2×1020 cm−2.

The thermodynamics of Hi in a variety of astrophysicalsettings has been studied by Wolfire et al. (1995a, 1995b). Astable cool phase of Hi is expected when both(i) a sufficientcolumn of shielding gas is present, and(ii) the thermal pressureis sufficiently high. Equilibrium kinetic temperatures of 85 Kare predicted to occur over a range of volume densities, de-pending on the dust–to–gas ratio, on the gas–phase metallicity,and on the intensity of the interstellar radiation field. These de-pendancies are illustrated in Figs. 5–7 of Wolfire et al. (1995a).Since the thermal pressure in the mid–plane of our Galaxy inthe solar neighborhood is fairly well–constrained to be aboutP/k = 2000 cm−3 K and this pressure is believed to declinesubstantially with height above the plane, an upper limit to thevolume density of some 24 cm−3 can be assigned to Hi with85 K kinetic temperature.

Assuming that the opaque cores in CHVC 125+41−207have a roughly spherical geometry, which is indeed suggestedby their projected spatial appearance and that the volume fill-ing factor of cool Hi is unity, then the column density, angularsize, and volume density can be used to estimate the distanceto the source, according toD = NH/(nHθ). This provides a

lower limit to the distance to CHVC 125+41−207 of 31 kpc. Ofcourse at a distance of 30 kpc, the ambient thermal pressure islikely to have declined by more than an order of magnitude fromthe Galaxy mid–plane value (e.g. Wolfire et al. 1995b), leadingto a corresponding linear increase in the distance estimate. At apotential distance of few 100 kpc, it becomes important to re-consider the ionization and heating conditions that would apply.

Specific calculations of equilibrium Hi conditions withinthe Local Group environment were made available to us byWolfire, Sternberg, Hollenbach, and McKee (private communi-cation) for two bracketing values of the neutral shielding columndensity, namely 1 and 10×1019 cm−2, a metallicity of 0.1 so-lar (which we discuss further below), and a dust–to–gas massratio of 0.1 times the solar neighborhood value. Equilibriumvolume densities atTk = 85 K of 3.5 and 0.65 cm−3 are foundfor shielding columns of 1 and 10×1019 cm−2, respectively, ascan be seen in Fig. 13. The corresponding distances are in therange of 210 to 1100 kpc. Comparison with Fig. 5 suggests thatshielding columns in the higher part of this range are most rel-evant for the two opaque cores under consideration, suggestinga distance in the range 0.5 to 1 Mpc.

We stress that the resulting distance estimate is not depen-dent on details of the thermodynamic calculation, but only on


CHVC230+61+165 MOM0

Grey scale range= -0.5 to 5.0 10^18 cm^Levs = 1.0 * [2, 5] 10^18 cm^2

DE

CL

INA

TIO

N (

J200

0)


15 35

30

25

20

15

10

05

0 2 4

CHVC230+61+165 MOM1

Grey scale range = 150 to 165 km/sPeak contour = 1.7324E+05 m/s Levs = 1.000E+03 * (152, 155, 158, 161)

DE

CL

INA

TIO

N (

J200

0)


15 35

30

25

20

15

10

05

155 160 165

CHVC191+60+093 MOM0

Grey scale range = 0.0 to 5.0 10^18 cm^Levs = 1.0 * [2, 5, 10] 10^18 cm^2

DE

CL

INA

TIO

N (

J200

0)


34 20

15

10

05

00

0 2 4

CHVC191+60+093 MOM1

Grey scale range = 85 to 100 km/sPeak contour = 1.0878E+05 m/s Levs = 1.000E+03 * (88, 91, 94, 97)

DE

CL

INA

TIO

N (

J200

0)


34 20

15

10

05

00

90 95 100

H191+60 RA: 10 37 27.5 DEC: 34 15 59.

Mill

iKE

LV

IN

VELO-LSR km/s75 80 85 90 95 100 105 110

480

420

360

300

240

180

120

60

0

H230+61 RA: 10 55 14.1 DEC: 15 21 59.

Mill

iKE

LV

IN

VELO-LSR km/s130 135 140 145 150 155 160 165 170

270

216

162

108

54

0

-54

BA

9491

158

155

A B

Fig. 10. Imaging data for CHVC 191+60+093and CHVC 230+61+165 at 2 arcmin and2 km s−1 resolution. Upper panels: apparent in-tegrated Hi (assuming negligible opacity), withcontours at 2, 5, and 10×1018 cm−2 and a lineargrey–scale extending from0 to 5×1018 cm−2.Middle panels: intensity weighted line–of–sightvelocities,vLSR, with contours at 88, 91, 94, and97 km s−1 and a linear grey–scale extending from85 to 100 km s−1 for CHVC 191+60+093, andwith contours at 152, 155, 158, and 161 km s−1

and a linear grey–scale extending from 150 to165 km s−1 for CHVC 230+61+165. Lower pan-els: brightness temperature spectra at the indi-cated positions.

the assumed thermal pressure in the source environment. Plac-ing CHVC 125+41−207 at a distance of, say, 3 kpc would re-quire assigning it a volume density of some 250 cm−3. Thewell–constrained temperature ofTk = 85 K, would then implya thermal pressure,P/k = 21000 cm−3 K, a value vastly ex-ceeding those observed in the local ISM. One might argue thata high pressure may be related to the high speed re–entry ofthe object into the Galaxy given its line–of–sight velocity in theLocal Standard of Rest frame of−200 km s−1. It is difficult toreconcile such a scenario with one of the narrowest Hi emis-sion lines ever detected, showing no measurable component ofturbulent line broadening. A moderately low thermal pressure,on the other hand, amounting toP/k ∼ 100 cm−3 K for this

high–latitude, non–turbulent source, seems plausible on rathergeneral grounds. A source distance in the range 0.5 to 1 Mpcfollows directly. In this case a consistent solution for the heatingand ionization of the source is also in hand, as outlined above.

Unfortunately, this method of distance determination cannotbe applied more generally in our current sub–sample. None ofthe cores in the other CHVCs we have imaged has a compara-bly high opacity. The typical peak brightness seen in other coresamounts to only about 2 K, suggesting opacities of about 2%, forTk ∼ 100 K. With such low opacity, the likelihood of encounter-ing unit filling factor of the gas along the line-of-sight becomesvanishingly small, so that the comparison of column and vol-ume densities gives only a very weak lower limit to the source



Grey scale flux range= 0.00 15.00 10^18CM2Peak contour flux = 1.6546E+01 10^18CM2Levs = 1.000E+00 * (1, 2, 5, 10)

DE

CL

INA

TIO

N (

J200

0)

RIGHT ASCENSION (J2000)10 40 38 36 34 32

35 00

34 30

00

33 30

0 5 10 15

Fig. 11. Column density distribution ofH i in CHVC 191+60+093 at 2 arcmin reso-lution reconstructed using LDS and WSRTdata. Contours are drawn at 1, 2, 5, and 10×1018 cm−2 and a linear grey–scale extendsfrom 0 to 15×1018 cm−2.

distance. Applying the same method blindly to the other coresin our current sample and assuming a volume density of 1 cm−3

throughout, returns lower limits to the distance which vary sub-stantially within an individual CHVC as shown in Table 2. Themethod is particularly suspect for the low brightness cores (0.5 Kpeak) of CHVC 191+60+093 and CHVC 230+61+165, wherelower limits of 7 and 4 kpc are found. Elsewhere, lower limitsto the distance in the range 40 to 130 kpc are implied.

5.3. Metal abundance in CHVC 125+41−207

The source CHVC 125+41−207 is one of the few compactHVCs for which it has been possible to measure Mgii absorp-tion against a background UV source. Bowen & Blades (1993)detected unsaturated Mgii absorption at a corrected velocity ofvLSR = −209 km s−1 (Bowen et al. 1995) toward the Seyfertgalaxy Mrk 205 at(α, δ) = (12h21m44.s12, 7518′38.′′25).They derive a Mgii column density of0.7±0.1×1013 cm−2 forthis component. This line of sight passes just to the South–Eastof the highNHI portion of clump B shown in Fig. 7. The columndensity we infer in this direction from our composite image is5 ± 1 × 1018 cm−2. Wakker & Van Woerden (1997) quote avalue of3.0 ± 0.7 × 1018 cm−2 in a nearby direction. Analysis

of the recent Jodrell Bank Hi data (de Vries et al. 1999) wouldprovide an improved estimate. Jenkins et al. (1986) provide ex-tensive data for the gas–phase abundance of Mgii relative toH i in the solar neighborhood. They derive relative gas–phaseabundances of log(NMgII/NHI) = −4.7and−5.0, respectively,for low– and high–volume density lines of sight. The relativeabundance we infer, log(NMgII/NHI) = −5.8, is about a factorof 0.07 down from the local ISM value appropriate for low vol-ume densities. Passing as it does through only the outer reachesof the atomic halo of CHVC 125+41−207, this seems the mostappropriate for comparison.

We have taken care to refer to comparable measure-ments of gas–phase metal abundance in comparing theCHVC 125+41−207 data with that of the solar neighbor-hood. The local ISM value is believed to be substan-tially depleted by deposition on dust grains from its totalvalue, log(NMgII/NHI) = −4.4. If the dust abundance ofCHVC 125+41−207 is less than that in the solar neighborhood,which is quite likely in most formation scenarios, then dust de-pletion would presumably be less effective and the metal abun-dance might then lie somewhere in the range 0.04 to 0.07 relativeto the solar value.



Grey scale flux range= 0.00 10.00 10^18CM2Peak contour flux = 8.9717E+00 10^18CM2Levs = 1.000E+00 * (1, 2, 5, 8)

DE

CL

INA

TIO

N (

J200

0)

RIGHT ASCENSION (J2000)10 59 58 57 56 55 54 53 52 51

16 00

15 30

00

14 30

0 5 10

Fig. 12. Column density distribution ofH i in CHVC 230+61+165 at 2 arcmin res-olution reconstructed from LDS and WSRTdata. Contours are drawn at 1, 2, 5, and 8×1018 cm−2 and a linear grey–scale extendsfrom 0 to 10×1018 cm−2.

5.4. Localized line splitting

An interesting kinematic phenomenon to emerge from thesehigh–resolution observations of the compact high–velocityclouds is the occurrence of several spatially localized regionswhich display splitting of the Hi emission profiles without arelated change in the velocity of the centroid of the profile. Anexample of the phenomenon is given by the series of spectrashown in Fig. 6. Five such regions were detected, two in eachof CHVC 069+04−223 and CHVC 125+41−207, and one inCHVC 204+30+075. Each of these cases is associated with arelatively high column density, in excess of about1020 cm−2.These regions are quite distinct from those instances of en-hanced linewidth which are the result of localized line–of–sightoverlap of extended features having different velocities. Thatphenomenon is also observed on several occasions (as notedabove) but it can typically be recognized as such by the pres-ence of two distinct velocity systems which are each spatiallyextended and have uncorrelated morphology.

The angular extent of these split–line regions is confinedto 1 or 2 arcmin in all cases. The degree of velocity splittingvaries from only about 5 km s−1 in CHVC 125+41−207, to10 km s−1 in CHVC 069+04−223, and to as high as 20 km s−1

in CHVC 204+30+075.

Small depressions in the line core of Galactic Hi emissionfeatures have often been interpreted as evidence for so–calledself–absorption by the cool opaque core of a warmer semi–opaque atomic structure. Such an interpretation does not seemappropriate for the phenomena under discussion, since the linestrength between the peaks declines to a rather low brightnesstemperature. The spatially localized and symmetric distributionof line shapes seen in Fig. 6 seems more suggestive of an orga-nized inflow or outflow with a substantial degree of sphericalsymmetry. At a distance of 700 kpc, these regions would havelinear sizes of 300 pc, making them comparable to some of thelow luminosity super–shells detected in nearby galaxies whichare powered by the supernovae and stellar winds of a stellarassociation (e.g. Mashchenko et al. 1999).

5.5. Velocity gradients

A common kinematic pattern observed within the CHVCclumps which we have imaged is a velocity gradient orientedalong the major axis of a roughly elliptical distribution. Spe-cific examples of such gradients were noted above for severalof the objects imaged. The observed magnitudes of the veloc-ity gradients vary from 0.5 to 2 km s−1 arcmin−1. The cases


Fig. 13. Equilibrium temperature curves for Hi in an intergalacticradiation field at a metallicity of 0.1 solar, a dust–to–gas mass ratio of0.1 times the solar neighborhood value and two values of the neutralshielding column density:1019 cm−2 indicated by the solid line and1020 cm−2 indicated by the dashed line. The 85 K kinetic temperatureof the opaque cores in CHVC 125+41−207 is indicated by the dottedline, as is the 8000 K typical temperature of the WNM.

which are best resolved are clump A of CHVC 069+04−223and clumps A and B of CHVC 204+30+075, each of which isabout 20 arcmin in angular size. This same pattern was alsoseen in the two CHVCs imaged by Wakker & Schwarz (1991),namely CHVC 114−10−430 and CHVC 111−06−466.

Since these well–resolved examples of systematic gradientsin the velocity fields are reminiscent in form and amplitude ofthe “spider” diagrams seen for Hi distributions in some dwarfgalaxies, we carried out standard tilted–ring fits to assess theextent to which they could also be modeled by rotation in aflattened disk system. In the usual way, we began by allowingall kinematic parameters to vary freely from ring to ring, andthen solved sequentially for the best–fitting kinematic center, forthe systemic velocity, for the position angle of the receding line–of–nodes, and, finally, for the kinematic inclination. Holding allof these best–fitting parameters fixed, we then fit only for therotation velocity of each ring.

Robust solutions for circular rotation were found in allthree cases and these are shown in Fig. 14. The fits displayslowly rising rotation velocity with radius; the rise is contin-uous to some 15 km s−1 in the case of CHVC 069+04−223A,and flattens out at 500 to 600 arcsec to values of 15 and20 km s−1 for CHVC 204+30+075A and B. An estimate ofthe contained dynamical mass is given byMdyn = Rv2/G =2.3×105Rkpcv

2km/s, while the mass of gas,Mgas = 1.4 MHI =

3.2 × 105S D2Mpc, whereS is the integrated Hi flux in units

of Jy km s−1 and there is an assumed contribution by He-lium of 40% by mass. At an assumed distance of 0.7 Mpc,these three clumps haveMdyn = 108.1, 108.1, and108.3 M,Mgas = 107.1, 106.5, and106.9 M, and dark–to–visible massratios ofΓ = 10, 36, and 29, respectively. The derived value ofΓ scales with the assumed distance as1/D, given the depen-dencies noted above.

Fig. 14. Derived rotation velocities in three elliptical cores,CHVC 069+04−223A and CHVC 204+30+075A and B. The best-fitting position angle, inclination, and systemic velocity are indicatedat the top right of each panel. These were subsequently held fixed inderiving the rotation velocity as function of radius. Solid lines are therotation curves of NFW cold dark matter halos of the indicated mass.

In addition to this crude estimate of dynamical mass, we cancompare our derived rotation curves with those expected for acold dark matter halo as parameterized by Navarro, Frenk &White (NFW, 1997). These authors find that such halos tend tohave a “universal” shape in their extensive numerical simula-tions, so that the density profile and resulting rotation velocityare determined simply from the halo mass. We have calculatedhalo properties using the (web–retrieved 19/09/99 version) codeof Navarro, Frenk & White (1997) assumingΩ0=0.3,λ0=0.7,H0= 60 km s−1 Mpc−1 and a cluster abundance normalizedfluctuation spectrum. The early plateau and subsequent rise inthe derived rotation velocity of CHVC 069+04−223A is notwell-fit by an NFW halo. For comparison, the expected rotationcurve of a107.7 M (within the virial radius of 3.4 kpc) NFWhalo is overlaid on the data. While this curve crudely matchesthe last few measured points, the shape at smaller radii deviatessubstantially. On the other hand, rather good agreement withthe data is found for both CHVC 204+30+075A and B withhalo masses of107.8, and108.2 M (within 9.3 and 12.6 kpc)at an assumed distance of 0.7 Mpc. These rotation curves areoverlaid on the data in Fig. 14. Note how well the shape of theserotation curves is reproduced by the standard NFW halo profile.

5.6. CHVC stability

The very high total linewidth seen for CHVC 115+13−275 wasfound to arise from the broad range of distinct line–of–sight ve-


locities of the ten individual clumps which make up this source.The angular separation of the clumps is some 30 arcmin, whiletheir velocity centroids are separated by as much as 70 km s−1.Since this source is very isolated in position and velocity (see theappropriate panel of Fig. 1 of Braun & Burton 1999) a chancesuperposition of ten unrelated components is unlikely.

It is interesting to consider the stability of this collection ofclumps under several scenarios. If, for instance, this collectionwere located at a distance of 5 kpc, the distribution of clumpswould have a linear diameter of 44 pc and would double insize on a timescale of only5 × 105 yr. Such a short dynamicaltimescale would imply that we were witnessing a rather specialmoment in the evolution of this source. If, on the other hand,this source were located at a distance of 0.7 Mpc and wereself–gravitating, the angular radius and half–velocity width of15 arcmin and 35 km s−1 could be used to calculate a dynamicalmass fromMdyn = Rv2/G = 2.3 × 105Rkpcv

2km/s = 108.93

M. The corresponding mass of gas at this distance,Mgas =1.4 MHI = 3.2×105S D2

Mpc is Mgas = 107.22 M. The dark–to–visible ratio at this assumed distance isΓ = 51, scaling, as inthe previous subsection, with1/D.

A comparable analysis can be applied to CHVC 043−13 −302, for which Arecibo imaging data was obtained by Gio-vanelli (1981). This object has only two primary core compo-nents which extend over about 40 arcmin, and are separated by60 km s−1 in velocity, while the total line flux is 260 Jy-km s−1.Using the half-separations and line flux yieldsMdyn = 108.92

M, Mgas = 107.61 M andΓ = 20 for D = 0.7 Mpc. Sinceonly two core components are involved, geometric effects oforientation are more likely to play a role in this calculation thanin the case of CHVC 115+13−275.

5.7. Comparison with nearby galaxies

The Hi imaging obtained here for our CHVC sub–sample re-veals striking similarities between the gas properties of theCHVCs and those of low mass galaxies. The comparison withthe Local Group dwarf irregulars, Leo A and Sag DIG, studied athigh resolution by Young & Lo (1996, 1997b) is particularly apt.Both of these objects display the same general morphology andkinematics as the CHVCs: high column density clumps char-acterized by narrow linewidths (about 8 km s−1 FWHM) andangular sizes of a few arcmin embedded in a diffuse halo of lowercolumn density gas with broader linewidths (about 20 km s−1

FWHM) extending over 10’s of arcmin. Young and Lo reachthe same conclusion presented here regarding the Hi phasecontent of the dwarf irregulars, namely that cool condensationsof CNM (with Tk ∼100 K) are found within WNM envelopes(with Tk ∼8000 K). The peak brightness temperatures detectedfrom the CNM in Leo A, 73 K, are almost identical to those wesee in the compact clumps of CHVC 125+41−207.

The velocity fields of both Leo A and Sag DIG are alsovery similar to those of the CHVCs. Modest velocity gradientsof comparable magnitude are observed (about 10 km s−1 over10 arcmin) along one position angle in addition to a substan-tial unstructured component. Unlike the CHVCs, these galaxies

contain comparable gas and stellar masses, each correspondingto about 108 M in the case of Leo A and about 107 M inSag DIG.

Young and Lo also comment on a small region of about2.5 arcmin diameter within Leo A which they suggest might bethe expansion or contraction signature of an Hi shell, compa-rable to the five cases of line–splitting within CHVCs which wedescribe above.

The dwarf elliptical companions of M31, NGC 185 andNGC 205 (Young & Lo 1997a) are rather different in theirgas properties. While both the CNM and WNM componentsof H i are seen in these two cases, all of the neutral gas is con-centrated to the central five arcmin (1 kpc) of each galaxy andthe global kinematics are less systematic. The strong interac-tion of these objects with M31 (NGC 205 in particular is on anorbit that is likely to have recently passed through the disk ofM31) and the high mass density of stars (stellar masses about100 times greater than the 105.5M in gas) makes a direct com-parison with the CHVCs less obvious.

An intermediate and less–perturbed example is given by themore isolated Local Group dwarf spheroidal LGS 3 (Young &Lo 1997b). In this case, the stellar mass is likely only a fewtimes that of the 105.8 M in gas. Here the Hi is also confinedto only the inner 5 arcmin (corresponding to 1 kpc), showslittle systematic kinematics, and the cool condensed phase isalmost entirely absent. Although peak Hi column densities ofsomewhat more than 1020 cm−2 are observed over a region ofa few arcmin extent, there is no sign of ongoing star formation.

Extending the comparison to even higher mass systems pro-vides some further insight into the conditions necessary for al-lowing the condensation of CNM clumps in a galactic environ-ment. Several low mass spirals were observed in the high reso-lution H i sample of Braun (1995, 1997), including the nearbySBm galaxy NGC 2366 and two of the nearest low surfacebrightness spirals, NGC 247 (of type Sc) and NGC 4236 (SBd).A filamentary network of CNM gas is seen in all these systems,which accounts for about 80% of the Hi line flux from onlyabout 15% of the surface area withinR25 (the radius where theaverage face-on surface brightness of light in the B band hasdeclined to 25 mag arsec−2). Just beyondR25 the flux fractionof CNM gas disappears entirely, even though the WNM con-tinues out to a radius of almost twiceR25. The result is thatthe total flux fraction due to CNM in these galaxies is about70%, substantially more than the 40% CNM fraction in theCHVCs. Comparison with the Hi thermodynamics of Wolfireet al. (1995a) suggest that the edge of the CNM disk shouldoccur where the thermal pressure has declined to a value ofPmin, below which CNM condensation is no longer possible.The numerical value ofPmin depends on the dust and gas phasemetal abundance, the radiation field intensity and the shieldingcolumn of neutral gas. Values forPmin/k of about 200 cm−3 Kseem indicated under conditions of moderate metallicity and ra-diation field-intensity expected for the low mass spirals above.Comparison with Fig. 13 suggests that slightly lower values forPmin/k of about 100 cm−3 K should be expected under the low


metallicity, low field–intensity conditions that might apply tothe CHVCs.

What pressures might be expected in the mid-plane of adisk–like self–gravitating distribution of Hi ? The concept ofhydrostatic pressure equilibrium demands that the weight of theoverlying medium is balanced by the sum of the contributionsto the gas pressure. This can be expressed as∫ ∞

z

4πGΣtot(z′)ρgas(z′)dz′ = Ptot(z)

= PB + PCR + Pturb + Pth + . . . (1)

whereρgas(z) is the gas volume density at heightz from themid-plane. In this form,Σtot refers to the mass surface densitybetweenz′ = 0 to z′ = z, not betweenz′ = −z to z′ = z. Thetotal pressure is composed of magnetic, cosmic ray, and turbu-lent components in addition to the thermal component which isof relevance to the calculation of the HI kinetic temperature. Ina high mass galactic disk, the thermal pressure might amount toonly about 1/4 of the total pressure, while in a low mass CHVCwith few if any sources of internal turbulence and energetic par-ticles, the thermal pressure might be the dominant component.If for simplicity we assume a constant gas density,ρgas, for−s/2 < z < + s/2 for some full disk thickness,s, thenΣtot(z) = (ρgas + ρstars + ρDM) × z for z less thans/2. Forthe moment we will neglect contributions to the mass surfacedensity by stars or dark matter, which if present would lead toproportionally higher pressures. Eq. 1 then becomes simply

Ptot(0) = πGρ2gass

2/2 = πG(µmHNH)2/2. (2)

Inserting numerical values into Eq. 2 yields

Ptot(0)/k = 41(NH/1020)2 cm−3 K. (3)

Thus, whenever the face–on Hi column density exceeds about1020 cm−2 in a disk–like system we can expect to see the coolcondensations of the CNM. Given the quadratic dependence oncolumn density in Eqs. 1–3, we might expect the transition frompurely WNM to a CNM/WNM mix to be quite abrupt. Thisexpectation is borne out quite dramatically at the edge of theCNM disks in spiral galaxies which indeed occurs very near aface–on Hi column density of 1–2×1020 cm−2. If a significantcontribution to the mass surface density within one gas scaleheight were due to stars or dark matter then the minimum columndensity needed for CNM condensation would decrease. For theCHVCs, our current data suggest a transition column densitynearer 1019 cm−2, rather than the 1020 cm−2 seen in nearbyspirals. This may be an indication that the mass surface densityof dark matter is already a significant contribution.

We can assess this possibility by considering the mass vol-ume density of the implied dark halos. The density distributionof the108.2 M NFW halo which gave a good fit to the derivedrotation velocity of the core in CHVC 204+30+075B (Fig. 14) isshown in Fig. 15. At a thermal pressure ofP/k = 100 cm−3 K,the 8000 K WNM halos of the CNM cores will have a volumedensity of about 0.01 cm−3. Comparison with Fig. 15 showsthat the dark halo will provide a dominant contribution to the

Fig. 15. Mass volume density, in units of hydrogen nuclei per cubiccentimeter, as a function of radius for an NFW cold dark matter haloof mass108.2 M within the 12.6 kpc virial radius.

mass density out to several kpc. Within the central region, ofperhaps 2 kpc (or 10 arcmin atD = 0.7 Mpc) in diameter,the mass density is enhanced by more than a factor of 10 overthat of the gas alone. In consequence, we can expect to achievevalues ofPmin/k of about 100 cm−3 K in the CHVCs wheneverthe Hi column densities exceed a few times 1019 cm−2.

6. Summary

We present some of the first high–resolution Hi imaging ofa sub–sample of the compact high–velocity clouds (CHVCs).Previously we had demonstrated that the spatial and kinematicproperties of the CHVCs were strongly suggestive of a LocalGroup origin (Braun & Burton 1999), although the typical dis-tance could not be well–constrained from the low resolution(0.5) data then in hand.

Our imaging has revealed that these objects have a charac-teristic morphology, consisting of one or more quiescent, low–dispersion compact cores embedded in a diffuse, warmer halo.This is consistent with what was seen previously by Wakker &Schwarz (1991) for the two CHVCs which they imaged. Thecompact cores can be unambiguously identified with the coolneutral medium (CNM) of condensed atomic hydrogen, sincetheir linewidths are in all cases significantly narrower than eventhe thermal linewidth of the warm neutral medium (WNM). Inone object we detect some of the narrowest Hi emission lines yetseen, with an intrinsic FWHM of 2 km s−1. The halo linewidths,while difficult to measure directly, appear broad enough to beconsistent with a WNM origin. Such a nested geometry of CNMand WNM is in agreement with the expectation from thermody-namics since a significant neutral shielding column is requiredfor the stable existence of CNM when exposed to an ionizingradiation field (e.g. Wolfire et al. 1995a, 1995b). The flux frac-tion in compact cores varies between about 1% and 50% in ourCHVC sub–sample, but is typically 40%, as also seen in thetwo cases studied by Wakker and Schwarz (1991). The surfacecovering factor of the moderate column density core emission


relative to the diffuse halo is usually a factor two lower, namelyabout 15%.

High–resolution imaging of the Hi column density distri-bution of CHVC 125+41−207 has permitted an improved deter-mination of the metal abundance in this object when combinedwith a previous measurement of unsaturated Mgii absorption(Bowen & Blades 1993; Bowen et al. 1995) of between 0.04and 0.07 solar.

The current program has also allowed meaningful distanceestimates for CHVCs to be made for the first time. Distancesin the range 0.5 to 1 Mpc follow from two independent ap-proaches. The first method utilizes the fact that we have resolvedcool, opaque knots of Hi within one of the six imaged systems(CHVC 125+41−207). By equating the observed angular sizeof these knots with the line–of–sight depth, we derive a dis-tance from a comparison of the column and volume densities.Estimates of the column density follow directly from our data,while specific thermodynamic calculations for volume densitiesunder the relevant physical conditions were made available tous by Wolfire, Sternberg, Hollenbach, and McKee (private com-munication). The second approach allows only a combinationof source distance and dark matter fraction to be derived, sinceit relies on a comparison of dynamical with visible mass underthe assumption that the system is gravitationally bound. For thisapproach, the CHVC 115+13−275 system proved particularlyilluminating, since a collection of some 10 compact cores isdistributed over a 30 arcmin diameter region within a commondiffuse envelope, while having discrete line–of–sight velocitiesdistributed over a 70 km s−1 range. At an assumed distance,D = 0.7 Mpc, the dark–to–visible mass ratio isΓ = 51, withΓ scaling as1/D.

A common feature observed in the CHVCs is that the com-pact cores have an elliptical shape with a clear velocity gradientaligned with the major axis. These kinematic data can be well–modeled by rotation in flattened disk systems. Rotation curvefits to the best–resolved cores within CHVC 069+04−223 andCHVC 204+30+075, yield rather high dark-to-visible mass ra-tios of Γ = 10–40 at D = 0.7 Mpc, withΓ scaling, as above,with 1/D. Moreover, the shape of the rotation curves in two ofthese three cases is very well described by that due to a colddark matter halo as parameterized by Navarro et al. (1997).

Well–resolved imaging has also revealed another intriguingaspect of the CHVCs. Five instances of localized line splittingare detected. These are areas of 60 to 90 arcsec extent in regionsof moderately high column density in which the Hi emissionprofile becomes doubly peaked in velocity, without displayinga shift in the velocity centroid. This appears to be a signature oflocalized outflow or inflow. At a distance of 0.7 Mpc they wouldcorrespond to feaures of a few hundred pc in extent, comparablein size to the stellar–wind–driven bubbles surrounding youngstellar associations.

Comparison of the CHVC properties with those of nearbylow mass galaxies shows striking similarities, particularly withthe Local Group dwarf irregulars. The required thermal pressurefor condensation of the cool CNM cores,Pmin/k ∼ 100 cm−3 K,can be achieved in hydrostatic equilibrium of a self–gravitating

disk–like distribution at the observed shielding Hi column ofWNM of about 1019 cm−2 if the significant mass density of thedark matter halos is taken into account within the inner 2 kpc.

Our study of the compact high–velocity–clouds has helpedto clarify the nature of these objects. All of the evidence points tothese being strongly dark–matter–dominated, self–gravitatingobjects of modest mass in the Local Group potential at dis-tances between perhaps 0.3 and 1 Mpc with physical dimen-sions in Hi of about 10 kpc. Typical Hi masses of a few times107 M are indicated, with associated dark masses of some109 M. The total number of such objects associated with theLocal Group, while still uncertain, may be about 200. It seemslikely that these are the objects which recent numerical simu-lations suggest should still be associated with massive galaxiesin a poor environment at the current epoch (Klypin et al. 1999;Moore et al. 1999).

The star formation history within these objects is still verymuch a mystery. High stellar densities, like those seen in dwarfspheroidal systems in the Local Group, are clearly not associ-ated with the centroids of the gas distribution. A large spatialseparation of the gas and associated stellar distribution, duefor example to tidal effects, is not expected to be a commoncircumstance since the population is dynamically so cold. Thepeak Hi column densities currently observed within the CHVCcores only rarely exceed 1020 cm−2. It is plausible that thesevalues are insufficient to allow widespread star formation at thistime. On the other hand, finite yet low metallicities have beenmeasured in at least one source, indicating that some enrichmenthas clearly taken place in the past. It is quite possible that asso-ciated old stellar populations with a low surface density will befound in deep searches which can distinguish such stars fromthe Galactic foreground.

Acknowledgements.We are grateful to M.G. Wolfire, A. Sternberg, D.Hollenbach, and C.F. McKee for providing the equilibrium tempera-ture curves for Hi in an intergalactic radiation field shown in Fig. 13as well as stimulating correspondence on the ISM in general, and toU. Schwarz and H. van Woerden for showing us, in advance of pub-lication, the Jodrell Bank data of de Vries et al. (1999) on the sourceCHVC 125+41−207. The Westerbork Synthesis Radio Telescope isoperated by the Netherlands Foundation for Research in Astronomy,under contract with the Netherlands Organization for Scientific Re-search.

References

Bajaja E., Morras R., Poppel W.G.L., 1987, Pub. Astr. Inst. Czech. Ac.Sci. 69, 237

Bland-Hawthorn J., Maloney P.R., 1999, In: Gibson B.K., Putman M.E.(eds.) Stromlo Workshop on High-Velocity Clouds. ASP Conf.166, p. 212

Blitz L., Spergel D.N., Teuben P.J., Hartmann D., Burton W.B., 1999,ApJ 514, 818

Bowen D.V., Blades J.C., 1993, ApJ 403, L55Bowen D.V., Blades J.C., Pettini M., 1995, ApJ 448, 662Braun R., 1995, A&AS 114, 409Braun R., 1997, ApJ 484, 637Braun R., 1998, In: Franco J., Carraminana A. (eds.) Interstellar Tur-

bulence. Cambridge University Press, Cambridge, p. 52


Braun R., Burton W.B., 1999, A&A 341, 437Braun R., Walterbos R.A.M., 1992, ApJ 386, 120Bregman J.N., 1980, ApJ 236, 577Burton W.B., 1997, In: Lesch H., Dettmann R.-J., Mebold U., Schlick-

eiser R. (eds.) The Physics of the Galactic Halo. Akademie Verlag,Berlin, p. 15

de Vries G., Schwarz U.J., Smoker J., Davies R.D., van Woerden H.,1999, private communication

Eichler D., 1976, ApJ 208, 694Einasto J., Haud U., Joeveer M., Kaasik A., 1976, MNRAS 177, 357Giovanelli R., 1981, AJ 86, 1468Hartmann D., Burton W.B., 1997, Atlas of Galactic Neutral Hydrogen.

Cambridge University Press, CambridgeJenkins E.B., Savage B.D., Spitzer L., 1986, ApJ 301, 35Klypin A., Kravtsov A.V., Valenzuela O., Prada F., 1999, ApJ 522, 82Mashchenko S.Y., Thilker D.A., Braun R., 1999, A&A 343, 352Moore B., Ghigna S., Governato G., et al., 1999, ApJ 524, 19

Navarro J.F., Frenk C.S., White S.D.M., 1997, ApJ 490, 493Oort J.H., 1966, Bull. Astr. Inst. Netherlands 18, 421Oort J.H., 1970, A&A 7, 381Oort J.H., 1981, A&A 94, 359Putman M.E., Gibson B.K., 1999, PASA 16, 70Shapiro P.R., Field G.B., 1976, ApJ 205, 762Verschuur G.L., 1975, ARA&A 13, 257Wakker B.P., Schwarz U., 1991, A&A 250, 484Wakker B.P., van Woerden H., 1991, A&A 250, 509Wakker B.P., van Woerden H., 1997, ARA&A 35, 217Wolfire M.G., Hollenbach D., McKee C.F., Tielens A.G.G.M., Bakes

E.L.O., 1995a, ApJ 443, 152Wolfire M.G., McKee C.F., Hollenbach D., Tielens A.G.G.M., 1995b,

ApJ 453, 673Young L.M., Lo K.Y., 1996, ApJ 462, 203Young L.M., Lo K.Y., 1997a, ApJ 476, 127Young L.M., Lo K.Y., 1997b, ApJ 490, 710

Astron. Astrophys. 354, 853–873 (2000) ASTRONOMY AND High ...aa.springer.de/papers/0354003/2300853.pdf · Astron. Astrophys. 354, 853–873 (2000) ASTRONOMY AND ASTROPHYSICS High–resolution

Documents