Star formation in globules in IC1396

Abstract

We present a large-scale study of the IC1396 region using new deep NIR and optical images, complemented by 2MASS data. For ten globules in IC1396 we determine (H-K, J-H) colour-colour diagrams and identify the young stellar population. Five of these globules contain a rich population of reddened objects, most of them probably young stellar objects. Two new HH objects (HH865 and HH864) could be identified by means of [SII] emission, one of them a parsec-scale flow. Using star counts based on 2MASS data we create an extinction map of the whole region. This map is used to identify 25 globules and to estimate their mass. The globule masses show a significant increase with the distance from the exciting O6.5V star HD206267. We explain this correlation by the enhanced radiation pressure close to this star, leading to evaporation of the nearby clouds and hence smaller globule masses. We see evidence that the radiation from HD206267 has a major impact on the star formation activity in these globules. Comment: 13 pages, 14 figures, A&A accepted, version with high resolution figures at http://www.dias.ie/~df/

Full text

Page 1

arXiv:astro-ph/0411706v1 25 Nov 2004
Astronomy & Astrophysics manuscript no. 1791
(DOI: will be inserted by hand later)
November 29, 2010
Star formation in globules in IC1396
Dirk Froebrich1⋆, Alexander Scholz2, Jochen Eisl¨ offel2and GarethC. Murphy1
1Dublin Institute for Advanced Studies, 5 Merrion Square, Dublin 2, Ireland
2Th¨ uringer Landessternwarte Tautenburg, Sternwarte 5, D-07778 Tautenburg, Germany
Received sooner ; accepted later
Abstract. We present a large-scale study of the IC1396 region using new deep NIR and optical images, complemented by
2MASS data. For ten globules in IC1396 we determine (H-K, J-H) colour-colour diagrams and identify the young stellar
population. Five of these globules contain a rich population of reddened objects, most of them probably young stellar objects.
Two new HH objects (HH865 and HH864) could be identified by means of [SII] emission, one of them a parsec-scale flow.
Using star counts based on 2MASS data we create an extinction map of the whole region. This map is used to identify 25
globules and to estimate their mass. The globule masses show a significant increase with the distance from the exciting O6.5V
star HD206267. We explain this correlation by the enhanced radiation pressure close to this star, leading to evaporation of the
nearby clouds and hence smaller globule masses. We see evidence that the radiation from HD206267 has a major impact on
the star formation activity in these globules.
Key words. stars: formation – stars: winds, outflows – ISM: Herbig-Haro objects – ISM: jets and outflows — Individual:
IC1396
1. Introduction
Star formation takes place not only in giant molecular clouds
but also in small isolated globules. Works e.g. by Sugitani et
al. (1991), Schwartz et al. (1991) and others showed that such
places are associated with youngstellar objects. The identifica-
tion of young stellar clusters or outflow activity in such glob-
ules hints an ongoing star formation process. Star formation in
globules might be induced by the propagation of an ionisation
shock front, the so-called radiation driven implosion mecha-
nism (Reipurth 1983). However, it is not fully clear what the
main properties are that influence star formation within these
globules (e.g. density, mass, or size of the globule, strength of
the ionisation shock front). Further, it would be interesting to
see if and how these properties influence the number and clus-
tering of the formingstars. The investigationof a larger, homo-
geneous, and as far as possible unbiased sample of globules is
an ideal way to obtain a deeper understanding of these issues.
IC1396 is one of the youngest and most active HII re-
gions in the CepOB2 group of loosely clustered OB stars
(Schwartz et al. 1991). The nebula is excited by the O6.5V
star HD206267 (Walborn & Panek 1984) and contains 15
small clouds and globules associated with red IRAS sources
(Schwartz et al. 1991) at a distance of about 750pc (Matthews
Send offprint requests to: df@cp.dias.ie
⋆VisitingAstronomer at theGerman-Spanish Astronomical Centre,
Calar Alto, operated by the Max-Planck-Institut f¨ ur Astronomie,
Heidelberg, jointly with the Spanish National Commission for
Astronomy.
1979). Large-scale observations of this region in the rotational
CO lines were done by Patel et al. (1995) and Weikard et al.
(1996). The numerous sharp-rimmed clouds and the relative
proximity make this region an ideal place to study star forma-
tion in a large, homogeneous sample of small globules.
Some of the globules in IC1396 were already investigated
in detail (IC1396N: Beltran et al. 2002, Nisini et al. 2001,
Codella et al. 2000; IC1396W: Froebrich & Scholz 2003;
IC1396A or the Elephant Trunk Nebula: Nakano et al. 1989,
Hessman et al. 1995, Reach et al. 2004) and/or are known
to harbour outflow sources (IRAS21388+5622, Duvert et al.
1990, Sugitani et al. 1997, De Vries et al. 2002, Ogura et al.
2002; IRAS22051+5848, Reipurth & Bally 2001). Relatively
little is known about most of the other globules, which are
therefore targets for this new study. As shown e.g. in Froebrich
& Scholz (2003), deep NIR imaging in JHK and the construc-
tionof(H-K,J-H)colour-colourdiagramscanrevealembedded
young objects or clusters in such globules. These data can be
used in conjunction with star counts (see e.g. Kiss et al. 2000)
to estimate the extinction and hence determine the mass of the
globules. Simultaneous observations of the region in narrow-
band filters centred either on the 1-0S(1) line of molecular hy-
drogenat 2.122µm orthe[SII]lines at671.6and673.1nmwill
uncover outflow activity from young stars.
This paper is structured as follows: In Sect.2 we describe
our data obtained in the optical and NIR and specify the data
reduction process. Our photometry in the NIR images with
emphasis on (H-K, J-H) colour-colour diagrams is shown in
Sect.3. The creation of NIR extinction maps to estimate the

Page 2

2 D. Froebrich et al.: Star formation in globules in IC1396
mass of the globules is described in Sect.4, followed by a dis-
cussion of the determined globule properties in Sect.5 and the
characterisation of the newly discovered outflows in Sect.6.
Finally, the results are discussed and summarised in Sect.7.
Some more detailed technical descriptions of the data analysis
procedures can be found in AppendicesA-C.
2. Observations and data reduction
2.1. Near infrared data
The near infrared (NIR) observations were made towards 10
out of the 15 IRAS sources listed in Schwartz et al. (1991).
Two of them have been studied before (IC1396N, Beltran et
al. 2002, Nisini et al. 2001, Codella et al. 2000; IC1396W
Froebrich & Scholz 2003), and another three globules could
not be observed due to bad weather conditions during our run.
The observed globules are listed on the top of Table1. One
main purpose of the NIR data was to construct colour-colour
diagrams for each globule in which we want to detect reddened
sources. This can be done by comparing the measured colours
of the targets with that of the main sequence. To get an esti-
mate for the colours of the main sequence, we observed main
sequencestars with knownspectral type in the same way as our
IC1396 targets. These standard stars were selected from the
SIMBAD database and deliver an unreddened main sequence
for our colour-colour diagrams. We preferred to define this
mainsequencebyobservationsratherthanwiththeoreticalevo-
lutionarytracks,becausethisavoidsmismatchesfrominconsis-
tent photometric systems, which can be substantial, as we have
shown in Froebrich & Scholz (2003). We selected the standard
stars so that they are generally less than 10degrees away from
the IC1396 region. Thus, they can be observed roughly at the
sameairmass asIC1396,whichexcludessystematic offsetsbe-
cause of differential extinction.
Our near infrared data were obtained on six nights from the
18th to 23rd of July in 2003 with the 2.2-m telescope on Calar
Alto, Spain. We observed with the MAGIC camera (Herbst et
al. 1993) in its wide-field mode (6′.92×6′.92 FoV). Using a
3x3 dither pattern around the central coordinates with a shift
of half a detector size we obtained 13′.5×13′.5 sized mosaics
of the IC1396 globules. This field size is comparable with the
typical diameter of small globules in this region (see Table1).
For the standard star observations, we chose a smaller shift
between the images to reduce overhead times. All fields were
observed in J, H, and K′, the IC1396 globules additionally in
a narrow-band filter centred on the 1-0S(1) line of molecular
hydrogen at 2.122µm (hereafter called the H2-filter). The K′
filter is very similar to the 2MASS Ks band and will be called
K in the following. The total per-pixel integration time in the
IC1396 mosaics was 324s in each broad-band filter. The inte-
gration times for the H2-filter were at least 2160s per pixel, in
most cases we reached 3240s per pixel. For the standard stars,
appropriateintegrationtimes werechosento avoiddetectorsat-
uration.
The weather conditions duringthe observations of the stan-
dard stars were photometric. Our ten globules in IC1396 were
also mainly observed under photometric conditions. For each
globule, we obtained at least one full set of mosaics (JHK), and
we took care to assure that at least one of these sets was ob-
servedunderphotometricconditions,sothatacalibrationofthe
wholedatasetis possible.Incaseofcirrus,someoftheglobules
wereobservedlongertoobtaina uniformlimitingmagnitudein
all fields. The narrow-band observations of the globules in the
H2-filter were done mostly under non-photometric conditions
orathigherairmass.Theseeingconditionsduringourrunwere
excellent (below 1′′), but we are limited by the large pixel scale
of 1′′.6 per pixel. All broad-band images, IC1396 globules as
well as standard stars, were obtained in the airmass range be-
tween 1.07 and 1.5. The positions of the observed globules are
shown in Fig.1 (small squares).
Our standard NIR data reduction included flat-fielding (us-
ing sky-flats), sky-subtraction, and co-addition to mosaics. To
co-center the images into a mosaic we used all stars in the
field to ensure a high astrometric accuracy.Mosaicing and sky-
subtraction was done using the xdimsum package in IRAF1
(Stanford et al. 1995). For each filter and globule we produced
mosaics of the 3x3 dither patterns. Additionally, all single mo-
saics of each filter were stacked to a deep mosaic. A plate so-
lution for each mosaic was obtained using the 2MASS sources
in the field.
2.2. Optical data
A large-scale optical survey of the IC1396 region was car-
ried out with the Schmidt camera at the 2-m telescope of the
Th¨ uringer Landessternwarte Tautenburg in two observing runs
(9.-12. Sept. 1999 and 21.-25. June 2004). We observed with
a narrow-band filter centred on the [SII] emission line as well
as with a broad-band I-filter, to be able to identify emission
regions. These optical observations cover about 12sqdeg. The
survey field is shown in Fig.1. A standard data reduction was
performed including bias subtraction and flat-field correction.
Emission features were searched for by comparing the narrow-
band [SII] images with the I-band images.
3. Photometry
3.1. Technique
For source detection, we applied the SExtractor (Bertin &
Arnouts 1996) to our deep K-band mosaic of the globule. The
relative offsets between the positions in this ’master’ mosaic
and all other mosaics of the same field were determined by
measuring the pixel coordinates of a bright star in each image.
Applying these offsets to the source catalogue of the ’master’
mosaic, we obtained catalogues for all mosaics of the field.
Subsequently, we performed aperture photometry for each ob-
ject in the catalogues using the daophot package within IRAF.
Because of the large pixel scale, the seeing is more or less con-
stant in all mosaics. Therefore, we decided to hold the aperture
for the photometry constant for all mosaics. Sky coordinates
1IRAF is distributed by the National Optical Astronomy
Observatories, which are operated by the Association of Universities
for Research in Astronomy, Inc., under cooperative agreement with
the National Science Foundation.

Page 3

D. Froebrich et al.: Star formation in globules in IC1396
. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . .
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
22.021.921.8
3
. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . . . . . . . . . .
. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .
21.721.6
R.A. (J2000)
. . . . . . . .. . . . . . . .
+
+
. . . . . . . .
+
. . . . . . . . . . . . . . . .
+
+
+
+
. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . . . . . . . . . .
21.5
. . . . . . . .. . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . . . . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. . . . . . . .
. . . . . . . .. . . . . . . .. . . . . . . .
+
. . . . . . . .. . . . . . . .
+
+
. . . . . . . .. . . . . . . .. . . . . . . .
. . . . . . . .. . . . . . . .
+
. . . . . . . . . . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . . . . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .
+
. . . . . . . .
+
+
+
+
. . . . . . . .. . . . . . . .
+
+
. . . . . . . .
+
. . . . . . . . . . . . . . . . . . . . . . . .
+
+
. . . . . . . . . . . . . . . .. . . . . . . .
+
. . . . . . . .. . . . . . . .
+
+
+
++
+
+
+
. . . . . . . .
+
. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .
+
+
++
. . . . . . . . . . . . . . . .. . . . . . . .. . . . . . . . . . . . . . . .. . . . . . . .. . . . . . . . . . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . . . .. . . . . .
+
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
11
+
+
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
10
+
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
++
+ +
+
+
8
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .9
+
+
+
+
+
+
+
+
+
+
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .7
+
+
+
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
4
5
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
+
+
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
3
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .2
+
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .6
++
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .1
++
n
+
12
n
13
n
+
+
14
+
n
+
+
+
+
16
+
+
n
17
+
n
+
19
+
n
+
21
+
++
n
22
??
+
+
+
+
??
+
23
n
24
+
n
25
+
+
+
+
+
+
+
+
+
+
+
+ ++
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
++
+
+ +
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+++
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
++
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
++
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
++
+
+
+
+
+
+
+
+
+
++
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
++
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
++
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
++
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
++
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
++
+ +
+
+
+
+
+
+
+
++
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
22.1
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
+
21.4
56
57
58
59
Dec (J2000)
Fig.1. IC1396 field investigated in this paper. The backgroundimage is the extinction map obtained by accumulated star counts
usingthe 2MASSJ-banddata(seeSect.4).Blackrepresentsregionswith highextinction,grayzeroextinction,andwhitenegative
extinction (caused by star clusters). The small squares mark the fields observed with MAGIC in JHK and in the 1-0S(1) line of
H2(IC1396W, the western most field, is marked also, since it is included in the discussion). Circles correspond to the globules
listed in the bottom half of Table1 (a few globules are slightly outside the region shown here). The numbers at the side of circles
and small squares are identical to the globule numbers used in Table1. The area marked with dashed lines was observed with
the Schmidt telescope in Tautenburg in I and [SII]. The very dark, circular region at about (21h.73:58.8◦) is not a globule but the
very bright star µCep, which caused non-detections of stars in its vicinity. A similar fake globule is generated by the star ζ Cep
at (22h.19:58.2◦). The + signs mark all IRAS sources in the field.
for the sources were determined by applying the plate solu-
tion to the pixel coordinates.As the result we obtained for each
globule a database with sky coordinates and instrumental mag-
nitudes in JHK and H2.
The instrumental magnitudes of the standard stars were
measured in a very similar way as the IC1396 object values,
i.e. with aperture photometry using daophot. Since all these
stars were observed under photometric conditions and airmass
similar to theIC1396objectfields, theircolourscanbe directly
compared with them. According to the spectral types given in
the literature, our standard stars span a spectral range from B0
to K8, hence they can be used to establish a main sequence in
the colour-colour diagram.
Tests were performed to verify the reliability of our instru-
mental photometry (e.g. repeated observations, field overlap).
Only in one of the ten fields (IRAS21539+5821) were larger
systematic photometry mismatches found, and hence the field
was excludedfromfurtheranalysis.For a moredetaileddiscus-
sion see AppendixA.
3.2. Colour-colour diagrams and reddened sources
We show colour-colour diagrams for five of the nine globules,
which contain more than 20 reddened objects, in Figs.2-6.
These figures contain a) datapoints for the IC1396 objects as
small dots, b) the main sequence datapoints as filled triangles,
c) the extinction path calculated with the extinction law given
by Mathis (1990) as dotted lines, d) reddened objects as filled
squaresandarrows.Toallowareliablecomparisonbetweenthe
objects in IC1396 and the standard stars, these diagrams are
shown in instrumental magnitudes, which avoids colour mis-
matches. A rough transformation into the 2MASS photomet-
ric system can be obtained by adding about 0.3mag to the H-
K values. The diagrams show all datapoints in the 13′.5×13′.5
field for each globule. We generally selected objects with H-
K>0.5mag as reddened sources. A more detailed discussion
ofthecolour-colourdiagramsandtransformationofinstrumen-
tal to 2MASS colours can be found in AppendixB.
As noted above, we found a large population of reddened
objects for five of the nine globules. The remaining four glob-
ules contain only a small number of reddened sources. To as-
sess thenatureofalltheseredobjects,we estimatedthenumber
of field stars in the direction of IC1396 using the Besanc ¸on
Galaxy model of Robin et al. (2003). This model is avail-
able online2and is able to generate a photometry catalogue
of objects in a certain Galactic direction with user-defined
colour constraints. Our selection criterion for red objects was
H-K>0.5mag in instrumental colours, which translates into
about H-K>0.8mag in absolute colours.
2http://www.obs-besancon.fr/www/modele/modele ang.html

Page 4

4D. Froebrich et al.: Star formation in globules in IC1396
Fig.2. Colour-colour diagram of the IRAS21324+5716 glob-
ule. Small dots are objects in and aroundthe globule. Filled tri-
angles are field stars with known spectral type (which mark the
unreddened main sequence). Filled squares, arrows up (lower
limit in J-H), and arrows right (lower limit in H-K and arbi-
trary J-H) mark reddened objects. The dashed line shows the
criterion used to select these red sources. Dotted lines indicate
the interstellar extinction path according to the extinction law
of Mathis (1990). The two nearly vertical dotted lines mark ex-
tinctionsofAV=10and20mag.Note thatthe colourscales are
in instrumental units.
Fig.3. Colour-colour diagram of the IRAS21346+5714 glob-
ule. Symbols as in Fig.2.
In a first step, we executed several simulations with stan-
dard Galactic optical extinction (0.7mag/kpc) and found that
the number of objects with H-K>0.8mag (and even with H-
K>0.5mag)is zero.This Galaxymodel,however,doesnotin-
clude ultracool dwarfs with spectral types L and T. According
to the DUSTY models of Chabrier et al. (2000), evolved ul-
tracool objects with Teff=1700K have an H-K colour above
0.8mag, and thus could be selected as red objects with our
diagrams. On the other hand, objects with Teffbelow 1300K
are too faint to be seen in our images. For this range of ef-
fective temperatures, Gizis et al. (2000) give a space density
of 2.11 · 10−3pc−3. Our limiting magnitude constrains the
Fig.4. Colour-colour diagram of the IRAS21352+5715 glob-
ule. Symbols as in Fig.2.
Fig.5. Colour-colour diagram of the IRAS21445+5712 glob-
ule. Symbols as in Fig.2.
distance at which we could detect these ultracool objects to
<50pc. With these values, the number of L-type objects in a
1sqare degree field is ≈0.03. From this estimate, we conclude
that the number of Galactic L-type stars in our red sample is
negligible.
For this first calculation, however, we made the assump-
tion of uniform Galactic extinction, which is not given in our
fields, since we observe regions with high extinction in the di-
rection of IC1396. The Besanc ¸on Galaxy model is able to in-
clude clouds with a given distance and extinction. In a second
set of simulations, we therefore assumed a cloud at a distance
ofIC1396(750pc)withvariousextinctionvaluesrangingfrom
AV=3mag to AV=10mag. The basic effect of such an addi-
tional cloud is that the main sequence in the colour-colour dia-
gram is split. While the foregroundobjects remain unreddened,
thebackgroundstarsareshiftedalongthereddeningpath.From
these simulations, we found that the background stars begin to
exceed the H-K=0.8mag limit if the extinction of the cloud is
higher than AV=5mag. Thus, we expect a significant number
of reddened background stars for globules with AV>5mag.
According to the Galaxy model, these reddened objects will be

Page 5

D. Froebrich et al.: Star formation in globules in IC13965
Fig.6. Colour-colour diagram of the IRAS22051+5848 glob-
ule.Symbolsas inFig.2.Forthis globuletheselectioncriterion
forthe reddenedobjects was shifted to H-K>0.7mag,because
all objects in this field show reddening of 0.2mag.
mostly M type giants for a cloud with AV=5mag, but with
stronger extinction stars with earlier spectral types also would
appear to have H-K>0.8mag.
Based on these results, we can make the following conclu-
sions about the nature of the reddened objects: We can exclude
that our reddened objects contain a significant fraction of fore-
ground stars. This includes also ultracool dwarfs with spec-
tral types L and T. For globules with high extinction, we ex-
pect a significant numberof highly reddened backgroundstars,
mostly late-type giants, along the reddening path. The remain-
ing reddened sources, particularly the objects whose colours
place them below the reddening path in the colour-colour di-
agram, should be young stellar objects (YSO) associated with
the globules in IC1396, which are intrinsically red because of
excess emission from circumstellar matter. From photometry
alone, it is not possible to unambiguously distinguish between
YSOs and backgroundstars.
Five of our nine globules clearly show a cumulation of red-
dened objects, whereas the number of reddened objects in the
remaining four globules is <10. It is now interesting to anal-
yse whether an excessive number of red objects is caused by a
high density of YSOs in the globule or a high cloud extinction,
leadingtoahighnumberofbackgroundstarswhichappearred-
dened. From the position of the reddened objects in the colour-
colour diagram,we are able to get a roughidea of whether they
are mostly YSOs or background stars, since background stars
will be reddenedalong the reddeningpath, whereas YSOs may
also appear below the extinction path. We define as YSO can-
didates all objects whose photometry does not rule out the pos-
sibility that they are located below the reddeningband.This in-
cludes all objects with photometryin all three bands, which are
below the reddening band, but also objects with upper limit in
J, that could be located below the extinction band, and all red-
dened objects for which only K-band magnitudesare available.
For each globule, we counted the number of YSO candidates
and the total number of reddened objects. For globules with-
out a large reddened population, it was found that only 77%
of the red sources are YSO candidates, whereas the remain-
ing objects are in the extinction path. On the other hand, for
globules with a large reddenedpopulation, the fraction of YSO
candidates is 68%, i.e. significantly lower than in the ’empty’
globules. The most straightforward interpretation of this result
is that the globuleswith manyred objects haveboth a highden-
sity of YSO and strong extinction, leading to a high number of
reddened backgroundstars. The other globules possess smaller
extinction values and fewer young stellar objects.
Assuming a distance of 750pc for IC1396 and no signifi-
cantextinction,ourreddenedtargetshaveabsoluteJ-bandmag-
nitudes between 3.1 and 7.3mag. This constrains the masses
for these objects to a range roughly between 0.05 and 0.9M⊙
(Baraffe et al. 1998). Note that there are also no brighter
sources significantly below the reddening path but with H-
K<0.5mag. Hence no unextincted Herbig AeBe stars are
present. Since the globules show J-band extinction of at least
0.9mag (see Table1), there could be a few higher mass stars in
our sample. On the other hand, it is unlikely that our faintest
objects have substellar masses, rather than just being strongly
influenced by extinction.
4. Extinction maps
4.1. Method
We determined extinction maps of the region 21h.3 to 22h.2 in
R.A. and 56◦to 60◦in DEC (J2000) using accumulated star
counts (Wolf-diagrams) in the 2MASS database. An example
Wolf-diagram is shown in Fig.C.1. Stars are counted in 3′×3′
boxes every 20′′down to the completness limit of the cata-
logue. A co-centred 1◦x1◦sized field was chosen as ”unex-
tincted” comparison. Minimum and maximum traceable ex-
tinction values are calculated and shown in Fig.C.2, depend-
ing on the resolution used. We can trace about one to 20mag
optical extinction. Details of the star count method and the de-
terminationof the detectionlimits can be foundin AppendixC.
4.2. Results
The extinction map of the whole IC1396 field shows several
regions of high extinction (see Fig.1). These are the globules
around the star HD206267, and a further group towards the
north-east. We defined a ’globule’ as an object with an ex-
tinction of at least 3σ above the noise level and a size of at
least 9 square arcminutes. With these criteria, an automated
search using the SExtractor (Bertin & Arnouts 1996) revealed
20 globules in IC1396. Five of these globules were contained
in the target list for our near-infraredsurveywith MAGIC from
Schwartz et al. (1991). The globules in this list that could not
be detected in our extinction maps possess sizes and masses
below our detection limits (see also Patel et al. 1995). Adding
the remaining five objects from this list, there are 25 dark
clouds in the field. The large region of high extinction at about
(21h.73:58.8◦) is caused by the bright star µCep, which pre-
vented the detection of stars in its vicinity. Similarly the star
ζ Cep at (22h.19:58.2◦) causes such a fake globule. The full list
of all globules with their coordinatesand knownidentifications

Page 6

6 D. Froebrich et al.: Star formation in globules in IC1396
is given in Table1. Figure7 shows as examples the extinction
maps for the globules where we detected a large number of
reddened sources.
To measure the mass we integrated the total extinction in
the globules (Atot
one sigma noise level where considered, and the outer radius
of the globule was taken as the size given by the SExtractor
software (FWHM, assuming a Gaussian core). From these
integrated extinction values we determined the mass of the
globules using the fact that the column density N(H) of hy-
drogen atoms can be expressed as 6.831021cm−2×AV/RV
where RV is typically 3.0 (see e.g. Mathis 1990). Using a
distance of 750pc for the globules, and the 20′′pixel scale
in our maps, the total mass in a globule can be determined
from the integrated optical extinction (Atot
Mglob[M⊙]=0.098Atot
tion values are averagedover the box-size of 3′×3′. If the dust
is concentratedin a smaller area, the obtainedmasses are lower
limits. The error of the estimated mass depends on the mean
extinction within the globule compared to the noise level in
our map. As a typical value we find that the mean extinction is
about 5σ of the noise level and hence the uncertainties in the
inferred masses are about 20%. In the case of IC1396N we
can compare our mass estimate with literature values. Wilking
et al. (1993) estimate 380M⊙and Serabyn et al. (1993) give
480±120M⊙ within the central 0.3pc. Given the large un-
certainties and the fact that the object is rather small and the
mass is concentrated close to the actual star forming core (see
e.g. Codella et al. 2001), our lower limit for the whole glob-
ule of 300M⊙ agrees well with these estimates. Weikard et
al. (1996) give an estimate of 12000M⊙ for the total mass
of their mapped region (about six square degrees centred on
HD206267). The mass estimated from our extinction maps in
the same field is 9000M⊙, reasonably close.
Within the errors of about 20% the total optical extinc-
tion values and globule masses, obtained from the J- and H-
band 2MASS data, are consistent. Some of the masses calcu-
lated with the K-band data differ by a larger amount. There
are two groups of objects: 1) The mass from the K-band data
is much smaller than the mass obtained from the J- and H-
band data. The reason could be that the globule contains an
embedded cluster of stars. These stars are easier to detect in K
and hence decrease the apparent extinction. In these cases the
given masses MJ,Hare lower limits for the globule mass. 2)
The mass from the K-band data is much larger than the mass
obtained from the J- and H-band data. An explanation could be
thattheconversionofextinctionintheK-bandtoopticalextinc-
tion cannot be performed following Mathis et al. (1990) that
assume an opacity index of 1.7. A lower value might be valid
in these globules. Note that this effect could counter-balance1)
in the case of an embedded cluster.
V). Only regions with an extinction above the
V) in the globule by:
V[mag]. Note that the obtained extinc-
5. Globule properties
With the results of Sect. 3 and 4, a homogeneous sample of
globules is now at our disposal (see Table 1). To evaluate pos-
sible trends in this dataset, we selected the following glob-
ule properties and tested them for correlations: projected dis-
tance of the globule from the exciting O star HD206267, pro-
jected size of the globule, number of YSO candidates within
the field of the globule, peak extinction value, total mass, mean
column density within the globule (total mass divided by the
size). For each pair of these properties, we fitted their relation-
ship linearly and calculated the correlation coefficient r. Since
the value t = r
?
Student’s t-distribution, this gives us the probability that the
two parameters show indeed a significant correlation. In or-
der to perform these tests, we set globule sizes to 30 square
arcminutes and masses to 120M⊙, if they could not be deter-
mined. These values are the upper limits from our detection
procedure for globules in the extinction maps. As peak extinc-
tion, we used the values from the J-band, since they have the
best signal-to-noise ratio.
(N−2)
1−r2 (N: number of datapoints) follows
Thebestcorrelationsareobtainedwhencomparingsizeand
mass, peak extinction and mass, and size and peak extinction.
The false alarm probabilities (FAP) for linear relationships be-
tween these three parameters are below 10−3%. These are ex-
pected correlations, since the mass is proportional to the mean
extinction and the size of the globule. The correlation of size
and peak extinction might be due to our method that smooths
the extinction over a 3′×3′field, naturally leading to lower
peak extinction values for small globules.
Additionally, we found a non-obvious correlation between
mass M and distance d from the exciting star: Globules far-
ther away fromHD206267have on averagelarger masses. The
least-squarefit givesM[M⊙] = 8.6d[pc]+84,andtheFAP for
this correlation is 0.078%. We repeated this test excluding the
four globules with d > 40pc, because it might be that these re-
gions are too far away from HD206267 to be influenced by its
radiation.Without the data points for these globules, the FAP is
still only 2.2%. Thus, there is a significant connectionbetween
globule mass and distance from the exciting star (see Fig. 8).
There are two possible explanations for this correlation: 1)
The mass loss rate of a cloud induced by photo-ionisation of
theexternallayersis inverselyproportionaltothedistancefrom
the exciting star (see e.g Codella et al. 2001). Hence, more dis-
tant globules do not lose as much mass due to evaporation. 2)
Globules closer to the exciting star contain more young objects
(see below)andhencethe mass estimatesof theseglobuleslead
totoolowvaluesduetothestarcountmethod.Connectedtothe
distance-mass correlationare correlationsofdistance and glob-
ule size as well as distanceandpeakextinctionvalue.These are
consequences of the interrelations described in the above para-
graph.
The remaining combinations of parameters show no signif-
icant correlations, i.e. with FAP below 10%. Particularly, we
see no correlation between distance and number of YSO can-
didates. This can be used to rule out the second explanation
for the mass-distance correlation. This relationship is therefore
most likely related to mass loss via photo-evaporation, as ex-
plained above. For all correlation tests that use the number of
YSO candidates, we are restricted to those globules in which
red objects were identified (see Table1). Thus, small number
statistics hamper the correlation tests for these parameters.

Page 7

D. Froebrich et al.: Star formation in globules in IC13967
Fig.7. Contour plots of the extinction obtained from the J-band 2MASS data for the globules 11 (upper left), 3 (upper right),
9 (lower left), and 4 and 5 (lower right). Contours start at 1mag optical extinction and increase by 1mag. Conversion from
extinction in the J-band to optical extinction was done according to Mathis et al. (1990). Dashed lines mark the borders of the
observedNIR fields. + signs indicate reddenedobjects in the reddeningpath, while circles mark the YSO candidates, and arrows
theprincipledirectiontowardsthe O6.5Vstar HD206267.Note thatthecontoursareslightlysmoothedandhencepeakextinction
values are somewhat lower than in Table1.
How do these findings fit in a scenario where the radiation
pressure of the O star is the primary triggering mechanism for
star formationin the globules?Ina first orderinterpretation,we
assume that the density of young objects S/A3/2(S - numer of
YSO candidates; A - projected size of the globule) is positively
correlated with the mass of the globule M and the pressure P
exertedbytheradiationfluxΦ fromthestar: S/A3/2∝ M×P.
The radiation pressure or ionising flux is inverselyproportional
to the square of the distance (P ∝ Φ ∝ d−2; Codella et al.
2001) and the mass is proportional to the distance (M ∝ d as
shown in Fig.8). Hence we should expect a correlation of the
star density with the inverse distance from the star: S/A3/2∝
M × P ∝ d−1. Indeed, such a correlation is seen tentatively
in our data. A linear fit gives S/A3/2= 0.555d−1, with a FAP
of 2.5% (see Fig. 9). If we exclude the one globule with d >
40pc, the FAP increases to 5.0%. Although this result clearly
needs to be substantiated with more datapoints, it tentatively
confirms our initial assumption that the star forming activity is
driven by the radiation pressure of the O star.
We re-examined the conclusion of Sect.3.2 where we
found that globules with a rich population of young stars
also show high extinction leading to a high number of red-
dened backgroundobjects. Althoughthere is no correlationbe-
tween extinction AJand the number of YSO candidates N,
the data show a clear tendency: All globules with N <10 have
AJ≤1.1mag. On the other hand, 50% of the globules with
N >10 exhibit AJ≥2mag. Thus, globules that harbour many
red objects tend to show high extinction values, confirmingour
result from Sect.3.2.
We further investigated the positions of the YSO candi-
dates in the globules with respect to the direction of the O6.5V
star HD206267. Figure7 shows the positions of the reddened
sources overplotted on the extinction map for the five globules
where a large number of reddened objects was found (the two
overlapping fields are combined in one figure). In three of the
fields the YSO candidates (circles) seem to be preferentially
positioned towards the direction of the exciting O6.5V star.
Objects within the reddeningpath (plus signs) seem to be more
concentrated towards the high extinction regions. This picture
agrees well with the expectations of triggered star formation
via radiation-driven implosion, although this behaviour is not
present in all investigated globules.

Page 8

8 D. Froebrich et al.: Star formation in globules in IC1396
Table 1. All detected globules in the IC1396 region. The upper part of the table lists the objects observed with MAGIC in JHK,
while in the lower part the additional globules detected in our extinction maps are given. Column4 lists the projected distance
of the globule from the exciting star HD206267. The size in square arcminutes is the value we obtained from the SExtractor
software (a ? marks the globules that were not detected in the extinction maps.). Column 6 contains the number of reddened
objects (in brackets: number of YSO candidates as defined in Sect.3.2). In columns 7 to 9 we give the peak extinction values in
the three filters obtained from our extinction maps, while columns 10 to 12 list the masses estimated for the globules.∗The given
masses are measured for both globules together.
Nr.Name(s)
(α;δ)(J2000)
[hm]
Distance
[pc]
Size
[2′]
Red starsAJ
AH
[mag]
AK
MJ
MH
[M⊙]
MK
[◦ ′]
1 IRAS21246+5743
IC1396W
IRAS21312+5736
IRAS21324+5716
LDN1093
LDN1098
IRAS21346+5714∗
IRAS21352+5715∗
LDN1099
LDN1105
IRAS21354+5823
LDN1116
IRAS21388+5622
IRAS21428+5802
LDN1130
IRAS21445+5712
IC1396E
IRAS21539+5821
IRAS22051+5848
LDN1165
LDN1164
21 2657 5824.7114 31(18)3.1 1.91.3 515 541401
2
3
21 33
21 34
57 50
57 32
11.7
8.6
? 9(7)
27(20)
0.9
2.0
0.7
1.2
0.7
1.150 224182207
4
5
21 36
21 37
57 28
57 30
5.4
3.1
?
?
51(38)
36(22)
1.1
0.8
1.0
0.7
0.7
0.6
120
120
119
119
74
74
6 21 37 58 3715.2? 4(3) 0.9 0.81.1
7
8
21 40
21 44
56 36
58 17
12.3
13.8
?
?
7(5)
10(8)
0.9
1.1
0.7
0.9
0.9
0.9
9 21 4657 2613.2? 24(16)0.90.8 0.7120141195
10
11
21 55
22 07
58 35
59 08
32.9
55.9
115
226
– 2.0
2.7
1.3
2.1
0.9
1.5
374
788
351
714
458
84751(36)
12
13
14
15
16
17
18
19
20
21
22
23
24
25
21 25
21 25
21 28
21 33
21 33
21 38
21 40
21 40
21 41
21 47
21 49
22 01
22 08
22 08
57 53
58 37
57 31
59 30
58 09
56 07
59 34
58 20
59 36
57 46
56 43
58 54
58 23
58 31
25.5
29.3
19.4
29.2
14.0
18.6
28.2
11.6
28.6
14.9
21.3
44.7
53.7
54.1
77
50
50
39
46
64
45
84
87
147
70
270
74
36
1.8
1.4
1.5
1.1
1.4
1.4
1.6
1.5
1.6
1.4
1.2
2.3
1.7
1.3
1.4
0.9
1.1
1.1
1.1
1.0
1.2
1.2
1.2
1.0
0.9
1.7
1.3
1.1
1.5
1.1
1.0
1.1
1.2
0.8
1.1
1.0
1.2
0.9
0.8
1.2
0.9
0.9
279
160
273
86
189
298
265
285
390
385
207
927
402
201
285
158
273
146
182
299
303
294
353
403
199
875
382
205
187
109
233
160
206
372
305
271
452
498
311
933
376
247
LDN1086
LDN1102
LDN1088
LDN1131
IC1396N
LDN1131
LDN1129
LDN1153
LBN 102.84+02.07
LBN 102.84+02.07
6. New HH objects and H2outflows
In our observed field there are several known Herbig-Haro ob-
jects andoutflows.Thesearemainlythe well investigatedflows
in the IC1396N globule (HH589-595 found by Ogura et al.
2002 and HH777-780discoveredby Reipurth et al. 2003). The
H2outflow from IC1396W which was shown in Froebrich &
Scholz (2003) has no published optical emission counterpart.
Connected to IRAS21388+5622 is the HH588 object (Ogura
et al. 2002). In the IRAS22051+5848 field a giant outflow
HH354 was found by Reipurth & Bally (2001).
In our [SII] images we were able to re-discover all of the
known HH objects around IC1396N. The whole field is full
of extended emission line objects (visible as excess emission
in the [SII] images compared to the I-band fluxes). These fil-
aments are excited by the strong UV emission of the O6.5V
star HD206267. Many of these filaments show a similar ap-
pearance to some of the HH objects near IC1396N. Hence, it
is very difficult to decide from the shape of the emission alone
Table 2. Coordinates of the brightest knots in the newly de-
tected Herbig-Haro objects.∗Counterparts of the optical knots
detected in H2in Froebrich & Scholz (2003).
ObjectH∗
2
α(J2000)
[hms]
δ(J2000)
[◦ ′ ′′]
HH588NE321 41 00.0
21 41 01.0
21 44 28.5
21 44 29.3
21 45 10.5
21 26 01.4
21 26 02.0
21 26 07.9
21 26 21.3
21 26 18.6
56 37 19
56 37 25
57 32 01
57 32 24
57 29 51
57 56 09
57 56 09
57 56 03
57 57 40
57 57 12
HH865A
HH865B
HH864A2-j
2-j
5
2-b
2-d
HH864B
HH864C
if we see an outflow or just UV excited filaments. This might

Page 9

D. Froebrich et al.: Star formation in globules in IC13969
Fig.8. Globule masses vs. distance from the exciting O star.
A linear least-square fit is shown as dashed line. This corre-
lation has a FAP of 0.078%. If we exclude the four globules
with d>40pc, the correlation is still significant with a FAP of
2.2%.
Fig.9. Diagram of star density vs. inverse distance from the
excitingO star. A linearleast squarefit is shown as dashedline.
Thecorrelationhasa FAP of2.5%. Ifweexcludeglobuleswith
d>40pc a FAP of 5.0% is still reached.
be the reason why HH777 was not mentioned by Ogura et al.
(2002) although it is clearly visible in their image.
We also detect the HH354 and HH588 flows from
IRAS22051+5848and IRAS21388+5622, respectively. There
is an additional emission feature 1′.4 east of HH588NE2, con-
sisting of two fuzzy blobs (for positions see Table2). This ob-
ject was slightly outside the field shown in Ogura et al. (2002).
The knot-like appearance of this object suggests an HH object
(HH588NE3) connected to the HH588 flow, rather than being
a UV-excitedemissionfeature(seeFig.10). Noneofthe optical
emission features of HH588 is detected in our H2images.
We detected two new groups of HH objects in our op-
tical field. We found the H2 flow from IC1396W also in
[SII] and call it HH864 (see Fig.11). The bright H2knot 2-j
(Froebrich & Scholz 2003) has a [SII] counterpart (HH864A)
which shows a bow shape with two maxima in the emission.
Fig.10. [SII] image of the HH588 region. The newly detected
emission knots HH588NE3 are seen in the upper left corner.
Fig.11. Our [SII] image of the outflow from IC1396W
(HH864) shows clearly some of the brightest emission knots,
previously detected in H2by Froebrich & Scholz (2003).
The two knots (2-b and 2-d) in the north-eastern lobe are vis-
ible in the optical emission line also (HH864C). Knot 5 (50′′
east of 2-l) has an optical counterpart as well (HH864B).
It is still not clear if this emission knot is connected to the
IRAS21246+5743 source or another (unknown) source in the
IC1396W globule.
We find in our optical field a giant flow emerging from
the IRAS21445+5712 source, which is also called IC1396E.
There are two bright bow shocks heading to the north-west of
the source, HH865A and HH865B (see Fig.12). The distance
ofabout0.2◦oftheterminatingbowHH865A fromthesource
makes this flow 2.6pc in length in one lobe (assuming a dis-
tance of 725pc). No counterflow is found in our images. The
globule was also observed in H2. The closer of the two bow
shocks is unfortunately just outside our NIR field and hence
could not be detected. There is no H2emission in our image
that might be connected to this flow. Only a few very faint fea-
tures are found south and east of the IRAS source. These are
UV excited emission coinciding with cloud borders.
In IRAS21312+5736 we see some faint H2 filaments
that might be UV excited. The field of IRAS21324+5716
is full of bright H2 emission features coinciding with the
optical emission. This applies also for the neighbouring
fields IRAS21346+5714 and IRAS21352+5715. No H2fea-

Page 10

10 D. Froebrich et al.: Star formation in globules in IC1396
Fig.12. The new flow HH865 in the [SII] filter, heading away
from the possible driving source IRAS21445+5712.
tures are detected near IRAS21354+5823, IRAS21428+5802,
IRAS21539+5821and IRAS22051+5848.
7. Summary and discussion
NIR observations of IC1396 in conjunction with extinction
maps obtained from 2MASS data reveal star forming activ-
ity and a large number of globules in this region. Twenty five
globules were identified using our extinction maps and the list
of Schwartz et al. (1991). Four of them were previously un-
catalogued in the SIMBAD database. In all but four cases the
masses (or at least lower limits) of the globules could be de-
termined. Also the size could be measured properly for all but
seven objects.
For nine globules observed deeper than 2MASS in this
work, the content of heavily reddened objects was derived by
meansofcolour-colourdiagrams.Fiveoftheseglobulesexhibit
a rich population of red objects. At least half of these objects
are good candidates for young stellar objects, the remaining
half is probably contaminated by reddened background stars.
The five globules with many red objects include the targets
with the highest extinction values, suggesting a correlation of
the strength of the star formation activity with the mass of the
globule.
Star formation in small globules is often thought to be
strongly influenced by the radiation pressure of a nearby bright
star. It was therefore investigated how the globule properties in
IC1396dependonthedistancefromtheOstarHD206267.The
masses of the globules show a clear positive correlation with
the distance from this star. We conclude that evaporation due
to photo-ionisationaffects the mass distribution of the globules
around HD206267. Our data are consistent with a scenario in
which the radiation pressure from the O star regulates the star
forming activity (expressed as density of young sources) in the
globules,inthe sensethatthe radiationpressurecompressesthe
gas and thus leads to enhanced star formation.
Our optical data lead to the discovery of several new HH
objects. These are the counterpart of the previously known H2
emission of the flow from IC1396W and a new parsec scale
flow from IC1396E. Further a new emission knot belongingto
the known HH588 object was discovered.
Acknowledgements. We thank J.Woitas for providing observing time
during his TLS run for this project. D.Froebrich and G.C.Murphy re-
ceived funding by the Cosmo Grid project, funded by the Program for
Research in Third Level Institutions administered by the Irish Higher
Education Authority under the National Development Plan and with
assistance from the European Regional Development Fund. A.Scholz
work was partially funded by Deutsche Forschungsgemeinschaft
(DFG) grants Ei409/11-1 and 11-2 to J.Eisl¨ offel. This publica-
tion makes use of data products from the Two Micron All Sky
Survey, which is a joint project of the University of Massachusetts
and the Infrared Processing and Analysis Center/California Institute
of Technology, funded by the National Aeronautics and Space
Administration and the National Science Foundation. This research
has made use of the SIMBAD database, operated at CDS, Strasbourg,
France.
Appendix A: Reliability of instrumental
photometry
Several tests were executed to evaluate the reliability of the
instrumental photometry. Since we observed each globule at
least twice (and six of them three times), we can compare the
instrumental magnitudes from different mosaics of the same
region to look for systematic mismatches. With one excep-
tion (see below), we found good agreement within ±0.05mag.
The photometry from mosaics which were taken under non-
photometricconditions (see Sect. 2.1) shows the highest devia-
tions, as expected. These deviations, however, affect all broad-
band mosaics equally, thus they are not visible in the colours.
For one field (IRAS21539+5821)we see large systematic pho-
tometry mismatches between all mosaics, which are also vis-
ible in the colours. The most probable explanation is variable
weather conditions, e.g. cirrus clouds which we did not recog-
nise during the observations. We excluded this field from all
further analysis.
As mentioned in Sect. 2.1, all broad-band images were ob-
tained at low airmass; the maximumairmass difference is 0.45.
Since the NIR extinction coefficient on Calar Alto is usually
below0.1mag/airmass(Hopp&Fern´ andez2002),thedifferen-
tial extinction offsets between the instrumental magnitudes of
different fields are safely below 0.05mag. More important, the
extinction coefficient in the NIR is nearly wavelength indepen-
dent. Therefore, differential extinction does not significantly
affect the colours of our targets. Thus, we did not perform an
extinction correction. Another test of our photometry was en-
abledbyfieldoverlapofthemosaicsaroundIRAS21346+5714
and IRAS21352+5715. Again we found no significant offsets
between the colours of the targets which were detected in both
fields. For these reasons, we conclude that the instrumental
colours of the targets observed under photometric conditions
are reliable.
If available, the deep, stacked mosaics of the globules were
calibrated by measuring the magnitude offsets to a mosaic ob-
tained under photometric conditions and applying these offsets
to the instrumental magnitudes of the deep mosaics. Thus, the
photometry of the deep mosaics is now directly comparable

Page 11

D. Froebrich et al.: Star formation in globules in IC1396 11
with all other data obtained under photometric conditions, in
particular with the standard star main sequence.
Appendix B: Colour-colour diagrams and
absolute calibration
For the colour-colour diagrams, we used the photometry from
the deep mosaics if more than one mosaic is available, other-
wise the photometry from the single mosaic. We only consider
objects with errors below 0.2mag. There are, however, only
very few objects (typically below 1%) with larger errors.
The bulk of the IC1396 objects in the diagrams coincides
with the late-type end of the main sequence. There is one ex-
ception – Fig.6 – which will be discussed below. This is ex-
pected because most stars are late-type. Moreover, giant stars
also concentrate towards the late-type end of the main se-
quence. Since our main sequence ends at spectral type K8, it
is not surprising that there are many field objects above the up-
per end of this sequence.These could be M-type dwarf or giant
stars. The bulk of the field objects is (againwith oneexception)
always in the same position in each diagram, which makes us
confident that the colours from the different fields can be com-
pared with each other.
There is one field (IRAS22051+5848, see Fig. 6), where
the bulk of the datapoints is clearly offset by about 0.2mag
in (H-K) and 0.2mag in (J-H), i.e. roughly in the direction of
the reddening vector. This field shows large-scale structures
of nebulosity and large voids without any objects, indicative
of strong extinction. Thus, in this field background objects
are probably significantly reddened because their light must
pass through extended regions of dust. The 0.2mag offset in
both colours then corresponds to a mean visual extinction of
AV≈ 2mag.
We are now interested in detecting objects whose position
in the colour-colour diagram indicates significant intrinsic red-
dening, i.e whose position is clearly shifted in the direction of
the extinction path or even below these lines. The following
criteria were used to select such reddened objects:
a) If an object is detected in all three filters, its H-
K colour should exceed 0.5mag: H-K>0.5mag. (For the
IRAS22051+5848 field, we require H-K>0.7mag, because
all objects in this field show a 0.2mag shift in H-K, see above.)
Objects that satisfy this condition are shown as filled squares
in the diagrams.
b) If the object is only detected in H and K,
we demand again that H-K>0.5mag (or >0.7mag for
IRAS22051+5848). These objects are marked with an ’arrow
up’, and their J-H colour is a lower limit estimated from the
H-band photometry and our sensitivity limit in the J-band.
c) If the object is only detected in K, we can only deter-
mine a lower limit for the H-K colour by subtracting the K-
band photometry from the sensitivity limit in the H-band. We
require that this lower limit is >0.5mag (or >0.7mag for
IRAS22051+5848). These objects are shown with an ’arrow
to the right’, and their J-H colours are chosen arbitrarily.
We selected all objects that satisfy one of these condi-
tions and examined these targets in the images. We rejected
all objects that are clearly not star-shaped (i.e. galaxies or H2
emission knots), as well as spurious detections (i.e. cosmics or
spikes of bright stars). The remaining objects are good candi-
dates for objects with intrinsic reddening in the globules. This
selection might be incomplete,since some youngobjects could
have H-K<0.5, and thus would fall out of our selection crite-
ria. As noted above, the colour-colour diagrams contain only
objects with errors below 0.2mag. With larger error bars a re-
liable separation between reddened and unreddened objects is
not possible. In most cases, however, the objects with errors
>0.2mag do not appear strongly reddened.
After this selection process, the globules clearly fall in
two groups: Four of them show only very few reddened ob-
jects, i.e. their number is ≤10. The remaining five glob-
ules harbour more than 20 reddened objects, in two cases
(IRAS21346+5714, IRAS22051+5848) the number of red-
denedobjects is largerthan 50.For these five globuleswe show
the colour-colour diagrams in Figs. 2-6.
As noted above, the colours in the diagrams are instrumen-
tal. We determined,however,the offsetsto thephotometricsys-
temofthe2MASScatalogue,whichis availableonline3.Forall
detected objects, we searched for counterparts in the 2MASS
database. In each field, several hundred common objects were
identified. For these objects, we determined the average offset
∆ between 2MASS photometry and our instrumental magni-
tudes in J, H, and K. These offsets (called ∆J, ∆H, ∆Kin the
following)includea zero-pointas well as extinctioncorrection.
The offsets in a certain band are not constant for all glob-
ules; they differ by as much as 0.3mag, depending on the
weather conditions during the observations (see above). The
difference between the offsets, in particular the values of ∆J−
∆H and ∆H − ∆K, are comparable for all globules, con-
firming again that the relative colours are reliable. We obtain
∆J−∆H= 0.03±0.06and ∆H−∆K= 0.29±0.08.Thus,
our colour-colour diagrams can be roughly transformed into
the 2MASS photometric system by adding 0.3mag to the H-
K values. We note, however, that this transformation does not
include colour terms, which means that the offsets are proba-
bly not constant for all colours, in particular not for the main
sequence and the reddened targets (see Froebrich & Scholz
2003). This was the main reason why we kept the diagrams
in the instrumental colours, as described above.
The absolute calibration also allows us to estimate the sen-
sitivity limit of our survey, i.e. the faintest objects for which
a 5σ detection is possible. We reach about 17.0mag in J,
16.5mag inH, and 16.0magin K. Forcomparison,the2MASS
catalogue for this region contains objects down to 16.9mag in
J, 16.1mag in H, and 15.6mag in K. Our images are thus sig-
nificantly deeper than the 2MASS catalogue for this region,
particularly in the H- and K-band.
Appendix C: Extinction map determination
A determination of absolute magnitudes of the stars within the
globules, as well as an estimation of the dust mass in the glob-
ules, requires a measurement of the extinction. As discussed
above, the assumption of a uniform Galactic extinction is not
3see http://www.ipac.caltech.edu/2mass

Page 12

12 D. Froebrich et al.: Star formation in globules in IC1396
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
....................................... . . . . . . . .
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. . . . . . . . . . . . .
. .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
−0.5
0.0
0.5
1.0
1.5
2.0
logN
10 1214 16
apparent brightness [mag]
Fig.C.1. Example of a Wolf diagram for the three filters (J -
solid; H - dashed; K - dotted) at α=21h52m00sδ =57◦00′00′′
(J2000). The accumulated number of stars N is averaged over
a 1◦x1◦field and normalised to an area of nine square arcmin-
utes. The vertical lines correspond to the completeness limit in
JHK (line typeas fortheWolf diagrams)at whichwe measured
the extinction.The mean slope <X> varies with the filter (0.36
for J and 0.34 for H and K) and also with the position on the
sky. As can be seen the slope is also not constant, due to ex-
tinction from the ISM (the slope in J ranges from 0.38 (at 11 to
12mag) to 0.30 (at 15 to 16mag)).
valid in our field. There are local extinctionenhancementscon-
nected to the globules and even within these small clouds the
dust and so the extinction is not uniformly distributed. Hence
we need to determine the local extinction enhancements for
our globules, compared to the neighbouring star field. These
extinction enhancements will be used to estimate the mass of
the globules by adopting a distance of 750pc. Measuring the
extinction from colour-colour diagrams is relatively difficult
due to the photometric errors and the large scatter of the main
sequence (see e.g. our Figs.2-6). Hence this method only al-
lows a rough estimate of at best ±3mag of the optical ex-
tinction averaged over the whole globule and is in particular
not able to determine the extinction/mass distribution within
the globules. A more accurate method is to create accumula-
tive star counts (Wolf diagrams) as described e.g. in Kiss et
al. (2000). Since we expect high extinction values within the
globules (AV≥5mag), creating these diagrams from NIR ob-
servations obviously is the best choise. The 2MASS database
provides an ideal basis for this purpose.
Using star counts to estimate the extinctionis based on sev-
eral assumptions: 1) The stars are equally distributed and all
apparent voids or less densely populated regions are caused
by extinction. 2) All stars possess the same absolute bright-
ness. 3) The completeness limit of the input catalogue does not
depend on the position in the sky. Since in general these as-
sumptions are not valid, we performedour star counts in a way
that ensures as small as possible errors due to the assumptions
made. 1) As reference or control field we choose a 1◦x1◦field
around the position where the stars are counted. This running
average ensures that large scale variations in the average star
numbers due to Galactic position or large scale structures are
Fig.C.2. Minimum and maximum values of optical extinction
that couldbe traced with ourextinctionmaps obtainedfromthe
2MASS data in JHK, using different box sizes. AVvalues are
determinedfromextinctioninJHKusingtheconversionfactors
given in Mathis et al. (1990).
corrected. 2) The box size in which we count the stars for each
position was varied between 1′×1′and 5′×5′. In boxes larger
than 3′×3′of the order of 100stars are enclosed on average.
This ensures that we enclose a ’typical’ sample of stars in our
box, that represents the ’average’ star very well. 3) The com-
pleteness limit of the 2MASS catalogue varies over the field.
Star counts where hence only performed down to a magnitude
towhichthecatalogueis completeoverthewholefield(J=16.6,
H=15.6, K=15.2).
Inorderto investigatenot onlythe extinctiondistributionin
our globules but also the large scale dust distribution in the en-
tire CepOB2 region, we created an extiction map of the whole
field, ranging from 21h.3 to 22h.2 in R.A. and from 56◦to 60◦
in DEC (J2000). In Fig.1 we show an example of the extinc-
tion map obtained from the J-band 2MASS data. Since we had
a large amount of data to process we decided to parallelise the
problem. We wrote an MPI-parallel C++ code in order to get
maximum performance from our 32-node hyperthreaded clus-
ter of the so-called ”Beowulf” type. It took 13 hours to process
all the data on this cluster - without the parallelisation of the
problem it would have taken of the order of 60 times as long
(about 33 days).
For the star counts we selected all 2MASS objects in our
field with a signal-to-noise ratio of larger than five (quality flag
A, B, or C). With this selection criteria we extracted 841908,
783474, and 635478 objects in JHK, respectively. This con-
verts to an average density of stars of about 1star per 20′′x20′′
field. Hence, we performed the star counts every 20′′in or-
der to obtain the final extinction map. This method ensures
a maximum gain of information from the 2MASS catalogue.
FigureC.1 shows an example of a Wolf diagram in each of
the three filters at α=21h52m00sδ =57◦00′00′′(J2000). In
order to find the best compromise between box-size and ex-
tinction regime that can be traced, the box-sizes were varied in
steps of 0′.25. Depending on the box-size the average number
of stars in the boxes varies and hence we can trace different

Page 13

D. Froebrich et al.: Star formation in globules in IC1396 13
extinction regimes. In Fig.C.2 we plot the minimum and max-
imum value of traceable extinction (converted to AV) with our
method in IC1396, depending on the chosen box-size and the
three 2MASS filters. The lower limit (Amin
the three sigma noise level of the determined extinction in the
wholeCepOB2field. This limit canbe transformedintoa ratio
Fσ/Fbackby
λ
) is calculated from
Amin
λ
= −1
X∗ log
?
1 −
Fσ
Fback
?
.
(C.1)
Fσis the noise of the number of stars in the star count map. If
this number of stars is missing in the presence of Fbackstars
in the comparison field, an apparent extinction of Amin
tected. The factor X is the slope in the Wolf-diagram. Note that
this factor, as well as the number of background stars varies
with the position and filter. The maximum tracable extinction
Amax
λ
can now be determined by assuming that only Fσstars
are present.
λ
is de-
Amax
λ
= −1
X∗ log
?
Fσ
Fback
?
(C.2)
This is a conservative assumption, because Fσrepresents ac-
tually the noise of the number of background and foreground
stars, but only the latter determines the maximum of tracable
extinction. Note: If we assume that the scatter of foreground
stars accounts only for half of Fσ, the maximum tracable ex-
tinction values increase by about 4, 6, and 10mag optical ex-
tinction for JHK, respectively (assuming X=0.3). We selected
a box-size of 3′×3′to perform all our measurements. The
lower limit (Amin
λ
) is also comparable to the error of our de-
termined extinction values. As can be seen in Fig.C.2 this is
much lower than the estimated accuracy of ±3mag when the
extinction is inferred from the colour-colour diagrams.
References
Baraffe, I., Chabrier, G. Allard, F. & Hauschildt, P.H. 1998, A&A,
337, 403
Beltr´ an, M.T., Girart, J.M., Estalella, R., Ho, P.T.P. & Palau, A. 2002,
ApJ, 573, 246
Bertin, E. & Arnouts, S. 1996, A&AS, 117, 393
Chabrier, G., Baraffe, I., Allard, F. & Hauschildt, P. 2000, ApJ, 542,
464
Codella, C., Bachiller, R., Nisini, B., Saraceno, P. & Testi, L. 2001,
A&A, 376, 271
De Vries, C.H., Narayanan, G. & Snell, R.L. 2002, ApJ, 577, 798
Duvert, G., Cernicharo, J., Bachiller, R. & G´ omez-Gonz´ alez, J. 1990,
A&A, 233, 190
Froebrich, D. & Scholz, A. 2003, A&A, 407, 207
Gizis, J.E., Monet, D.G., Reid, I.N., Kirkpatrick, J.D., Liebert, J. &
Williams, R.J. 2000, AJ, 120, 1085
Herbst, T.M., Beckwith, S.V., Birk, C., Hippler, S., McCaughrean,
M.J., Mannucci, F. & Wolf, J. 1993, SPIE, 1946, 605
Hessman, F.V., Beckwith, S.V.W., Bender, R., Eisl¨ offel, J., G¨ otz, W.
& Guenther, E. 1995, A&A, 299, 464
Hopp, U. & Fern´ andez, M. 2002, Calar Alto Newsletter Nr. 4
Kiss, C,. T´ oth, L.V., Mo´ or, A., Sato, F., Nikolic, S. & Wouterloot,
J.G.A. 2000, A&A, 363, 755
Mathis, J.S. 1990, ARA&A, 28, 37
Matthews, H.I. 1979, A&A, 75, 345
Nakano, M., Tomita, Y., Ohtani, H., Ogura, K. & Sofue, Y. 1989,
PASJ, 41, 1073
Nisini, B., Massi, F., Vitali, F., Giannini, T., et al. 2001, A&A, 376,
553
Ogura, K., Sugitani, K. & Pickles, A. 2002, AJ, 123, 2597
Patel, N.A., Goldsmith, P.F., Snell, R.L., Hezel, T. & Xie, T. 1995,
ApJ, 447, 721
Reach, W.T., Rho, J., Young, E., et al. 2004, ApJS, 154, 385
Reipurth, B. 1983, A&A, 117, 183
Reipurth, B., Armond, T., Raga, A. & Bally, J. 2003, ApJ, 593, 47
Reipurth, B. & Bally, J. 2001, ARA&A, 39, 403
Robin, A. C., Reyl´ e, C., Derriere, S. & Picaud, S. 2003, A&A, 409,
523
Schwartz, R.D., Wilking, B.A. & Giulbudagian, A.L. 1991, ApJ, 370,
263
Serabyn, E., Guesten, R. & Mundy, L. 1993, ApJ, 404, 247
Stanford, S.A., Eisenhardt, P.R.M. & Dickinson, M. 1995, ApJ, 450,
512
Sugitani, K., Fukui, Y. & Ogura, K. 1991, ApJS, 77, 59
Sugitani, K., Morita, K.-I., Nakano, M., Tamura, M. & Ogura, K.
1997, ApJ, 486, 141
Weikard, H., Wouterloot, J.G.A., Castets, A., Winnewisser, G. &
Sugitani, K. 1996, A&A, 309, 581
Wilking, B., Mundy, L., McMullin, J., Hezel, T. & Keene, J. 1993, AJ,
106, 205
Walborn, N.R. & Panek, R.J. 1984, ApJ, 286, 718