2MASS Extended Sources in the Zone of Avoidance

A new high-resolution near-infrared mapping effort, the Two Micron All Sky Survey (2MASS), is now underway and will provide a complete census of galaxies as faint as 13.5 mag (3 mJy) at 2.2 µm for most of the sky, and ~12.1 mag (10 mJy) for regions veiled by the Milky Way. This census has already discovered nearby galaxies previously hidden behind our Galaxy, and will allow delineation of large-scale structures in the distribution of galaxies across the whole sky. Here we report the detection and discovery of new extended sources from this survey for fields incorporating the Galactic plane at longitudes between 40 and 70°. Follow up H I 21-cm and optical spectroscopy observations provide positive confirmation for 14 of the new 2MASS galaxies. We perform an internal completeness and reliability analysis for the sample, consisting of 7000 sources in ~1000 deg² of area, including galaxies and Galactic nebulae from the W51 GMC. The area-normalized detection rate is ~1-2 galaxies per deg² brighter than 12.1 mag (10 mJy), roughly constant with Galactic latitude throughout the "Zone of Avoidance", of which 85-95% are newly discovered sources. In conjunction with the deep HI surveys, 2MASS will greatly increase the current census of galaxies hidden behind the Milky Way. Moreover, owing to its sensitivity to elliptical and other gas-poor galaxies, 2MASS will provide a key complementary data set to that of the gas-rich-sensitive H I surveys of the Milky Way Galaxy, potentially uncovering nearby galaxies critical to the local gravity and mass density fields.

Key words: galaxies: general --- ISM: HII regions---ISM: W51---astronomical databases: surveys: infrared radiation

Abstract

3.1 Source Classification
3.2 A Montage of Galaxies and Nebulae in the ZoA
3.3 Constructing an Extended Source Catalog
3.3.1 Extinction & Color vs. Galactic Latitude
3.3.2 Completeness & Reliability
3.3.3 Source Counts
3.3.4 Previously Catalogued Extended Sources

References

Table 1 -- Parameters of 2MASS extended Sources in the ZoA
Table 2 -- Extragalactic & Galactic Extended Source Counts: b > 0°
Table 3 -- Extragalactic & Galactic Extended Source Counts: b < 0°

List of Figures
 

Over the last two decades astronomers have learned that the local Universe exhibits a complex topology, characterized by groups and clusters of galaxies aligned along sheets and filamentary structures. The structures are often portrayed as "walls" of galaxies or "bubbles and voids," with sizes ranging from 30 to 150 million light years across (deLapparent, Geller & Huchra 1986; Giovanelli & Haynes 1988; Geller & Huchra 1989; da Costa et al. 1994; Di Nella et al. 1997). Moreover, nearby galaxies, including the Milky Way and Andromeda (M31), exhibit a peculiar flowing motion with respect to the Cosmic Background Radiation, with the signature being a large-angular scale dipole feature in the Cosmic Microwave Background (e.g.. Smoot et al. 1977). In the simplest of interpretations, the dipole is dominated (after correcting for the motion of the Sun through the Galaxy) by a streaming motion of galaxies toward what is frequently called the "Great Attractor "(Lynden-Bell et al. 1988; Kolatt et al. 1995), located somewhere behind the Galactic center region, probably associated with the massive rich cluster Abell 3627 at ~5000 km/s, at l ~325°, b ~-7° (Kraan-Kortweg et al. 1996).

The velocity field is complicated by infall motion of the Local Group of galaxies toward the nearby Virgo cluster (Aaronson et al. 1982) as well as a peculiar motion towards the Galactic anti-center region, the so-called "Local Velocity Anomaly" (Faber & Burstein 1988; Lu & Freudling 1995) and the Puppis constellation, also located deep in the plane of the Milky Way, at l ~ 240°, |b| < 10° (Lahav et al. 1993; Kraan-Kortweg & Huchtmeier 1992). If galaxies are the tracers of mass in the local Universe, then in order to decode these complex phenomena it is essential that we derive a complete census of galaxies local to the Milky Way.

This objective has been frustrated by the presence of the Galaxy itselfùthe Milky Way's gas, dust and stars obscure to one degree or another about one-half of the entire sky, including ~20% that is significantly veiled from view. Traditional visual spectroscopic and imaging surveys have usually avoided the plane of the Galaxy, hence the "Zone of Avoidance" (hereafter, ZoA) refers to the region of the sky from 10° to 30° from the Galactic plane at |b| = 0° (Note: the zone also depends on Galactic longitude, with the zone area increasing toward the Galactic center at l ~ 0°.) As such, not only is a significant fraction of the local Universe gravity field undefined, but the observed fraction is split into two disconnected parts, which may lead to biases in interpreting the structure. Once it was realized that galaxies in the ZoA play an important role in deciphering the Supergalactic plane and the local Universe in general, new methods were utilized to see through the veil of the Milky Way.

Although there are ongoing efforts to penetrate the ZoA down to |b| ~10° using deep optical imaging that targets low-obscuration "windows" (e.g., Kraan-Korteweg & Woudt 1999; Kraan-Korteweg et al. 1996), the best solution is to observe the ZoA at longer wavelengths that are less affected by dust attenuation. The Infrared Astronomical Satellite (IRAS) provided one of the first uniform and complete, to ~1 to 2 Jy at 60µm, windows into the ZoA, revealing many nearby infrared-bright galaxies near or within the Milky Way (Rowan-Robinson et al. 1990; Saunders et al. 1991; Kaiser et al. 1991; Fisher et al. 1995; Lu et al. 1990; Oliver et al. 1996; Canavezes et al. 1998; Branchini et al. 1999). However, far-infrared surveys, such as the IRAS 1.2 Jy PSC and the PSCz, are ultimately limited by their poor spatial resolution and bias towards dusty luminous galaxies (such as late-type spirals and starbursts).

One promising approach to finding galaxies in the ZoA is to look for redshifted 21-cm neutral hydrogen line emission (H I). Many studies to detect galaxies in the ZoA in H I have been carried out (Henning 1992; Pantoja et al. 1997; Kraan-Korteweg & Huchtmeier 1992; Lu & Freudling 1995), with some extensive surveys of the ZoA underway, including the Dwingeloo Obscured Galaxy Survey (e.g., Kraan-Korteweg et al. 1994; Henning et al. 1998) and the Parkes multibeam survey (HIPASS; Staveley-Smith et al. 1997; Henning et al. 1999; Juraszek 1999). The H I surveys have been successful at finding new galaxies in the ZoA, including a handful of very large nearby galaxies, like Dwingeloo 1, located just beyond the Local Group (Kraan-Korteweg et al. 1994). These surveys, however, are only sensitive to gas-rich nearby spirals (e.g., Dwingeloo 1 and Maffei 2), while their sensitivity drops as the inverse square of the distance of the sources. For example, the noise limits of the all-sky Parkes (shallow) survey (Kilborn et al. 1999) only permit the detection of a typical spiral with 10⁹ solar masses of H I to a redshift of about 2000 km/s. Deeper surveys in the Galactic plane are now in progress that will achieve better sensitivity, extending the search volume out to 10,000 km/s for galaxies with very high H I masses (Kraan-Korteweg et al. 1998; Staveley-Smith et al. 1998; Kraan-Korteweg & Woudt 1999).

Near-infrared surveys have the advantage of being sensitive to most types of galaxies, including gas-poor spheroidals. In addition, being imaging surveys they detect typical spiral galaxies out to ~20,000 km/s and the largest galaxies about five times farther. Tracing elliptical galaxies is important for solving the large-scale gravity and velocity fields in the local Universe because they trace the densest regions. The IRAS and H I surveys, by contrast, are insensitive to these largest cluster galaxies. Two large near-infrared surveys are currently underway, the Deep NIR Southern Sky Survey (DENIS; Epstein et al. 1997; Kraan-Korteweg et al. 1998; Staveley-Smith et al. 1998; Kraan-Korteweg & Woudt 1999, and the Two Micron All Sky Survey (hereafter, 2MASS; Skrutskie 1997). Both projects nicely complement the IRAS and H I surveys, since they are sensitive to both spiral and elliptical-type galaxies as faint as K=12^th mag (10 mJy) in the ZoA.

2MASS is a ground-based, imaging survey that utilizes the near-infrared atmospheric band windows at J (1.25µm), H (1.65µm) and K_s (2.17µm). One of the key scientific objectives of 2MASS is to study large-scale structures in the local Universe. Near-infrared surveys are able to address these issues owing to their unique properties: (1) e.g., 2MASS is a highly uniform all sky survey, so a complete census is possible; (2) the mass distribution of galaxies is dominated by the older stellar population, which in turn emit most of their light in the near-infrared; (3) interstellar extinction within galaxies is considerably less than that observed at optical wavelengths; similarly extinction from gas and dust in the Milky Way is greatly reduced in the near-infrared, providing a window into and through the ZoA; (4) the spatial resolution of 2MASS/DENIS allows detection and identification of the more numerous small galaxy members of clusters, and (5) for highly inclined spirals, distances may be inferred using the infrared Tully-Fisher relation.

The combined sensitivity and spatial resolution of 2MASS will lead to detection of over 1 million galaxies out to redshifts >30,000 km/s for most of the sky (Jarrett et al. 2000). In the ZoA 2MASS will discover many new galaxies, including those belonging or in close proximity to the Local Group. For example, a large (~6´ diameter) nearby (775 km/s) galaxy was recently discovered with 2MASS (Hurt et al. 2000). This paper shows that for the Galactic plane, deep in the ZoA, the combined effects of source confusion and extinction ultimately limit detection to bright, 12 - 13^th mag (5 to 10 mJy at K band), nearby galaxies, z ~10,000 km/s, with "normal" early-type morphology, including spirals and ellipticals. The data set and targeted fields are introduced in §2. In §3 we present a detailed study of a large piece of sky bounding the ZoA between 40 and 70° Galactic longitude, including extended-source counts, completeness and reliability and color distribution. We illustrate the kinds of extended sources that 2MASS detects in the ZoA, including follow-up H I observations to verify the extragalactic nature of a number of sources. In §4 we discuss the impact that 2MASS will have upon understanding large-scale structure across the plane of the Milky Way.

The 2MASS Second Incremental Data Release includes large portions of the Milky Way, including the region at 40° < l < 70° and |b| < 40°, that is the targeted area of study presented in this paper. For this longitude range, the ZoA is roughly defined to be the area within 20° of the plane, or comprising about 50% of the field area. The field location with respect to the backdrop of the Milky Way is illustrated in Fig. 1. The area is bisected by the Sagittarius tangent arm extending from the Galactic center, and incorporates portions of the Cygus, Lyra, Vulpecula, Sagitta, Hercules, Aquila, Delphinus and Pegasus constellations. The star-formation region, W51 molecular cloud complex, lies at the center of the field, l ~49°, b ~ -0.5°. The field was chosen for study owing to its large dynamic range in stellar number density, from a few hundred to tens of thousands of stars per unit area; moreover, it is infused with a diverse population of Galactic sources associated with star formation. It is confusion from foreground stars that ultimately limits the sensitivity and reliability of the extended sources, both galactic and extragalactic, that can be extracted by 2MASS. Hence, the field of study (Fig 1) provides a broad test of the ability of 2MASS to detect extended sources subject to vastly different environments and confusion levels. To date, the 2MASS survey has scanned over 1000 deg² of the defined area, corresponding to about 1/4 of the total possible area, consisting of ~1880 individual scans (each scan is 6° × 0.15°) determined to be of high quality as measured with photometric and sensitivity proxies.

The region of study ranges in stellar number density (hereafter shortened to "density") from a mere 1000 stars deg^-2 brighter than 14.0 mag at K_s (1.8 mJy) to an enormous 40,000 stars deg^-2 near the Galactic plane. Stars typically outnumber galaxies by a factor of 100 to 10,000 in the ZoA; consequently, it is a challenge to find galaxies in these crowded fields. The 2MASS project employs automated star-galaxy discriminator techniques to reduce the ratio of stars to galaxies to a more manageable 2:1 or 1:1 ratio (detailed in Jarrett et al. 2000). It is then a straightforward process to cull out the remaining non-extended sources (such as triple stars) using visual examination of images, or more sophisticated automated methods, such as decision trees or neural nets. We also compared our source lists with astronomical databases to identify previously catalogued objects, such as planetary nebulae.

2MASS supplies a variety of integrated flux measurements for a given extended source. In this study we use the integrated flux based on the elliptical aperture best fit to the K_s-band isophote corresponding to 20.0 mag arcsec^-2 in surface brightness (roughly equivalent to 1-sigma of the RMS background noise in the K_s-band image). The same "fiducial" aperture (position, size and orientation) is applied to each band, allowing cross-comparison between bands.

A 3-color JHK_s mosaic of the subsection field concentrating on the most extreme ZoA, |b| < 10°, is shown in Fig. 2. The stellar foreground confusion is noticeably most severe near b = 0°. Nonetheless, 2MASS is able to find extended sources peering through the Milky Way: four typical galaxy detections are highlighted in Figure 2. The presence of gas and dust is clearly seen near the Galactic plane, |b| < 5°, with an H I gas column density gradient vectored west toward the Galactic center.

The major limitation to galaxy detection by 2MASS in the ZoA is "confusion" noise from stars. The confusion noise becomes appreciable within an area ±5° of the Galactic plane (Fig. 2). The end result is a loss of sensitivity as the surface density of stars increases exponentially. Correspondingly, a profusion of multiple star groupings, including double and triple stars, mimic galaxies and are falsely identified by the automated extended source detection and extraction routines. Nevertheless, 2MASS can identify galaxies as faint as 13^th mag at K_s (~4.4 mJy) for |b| > 5 - 10°, and as faint as 12.1 mag at K_s (~10 mJy) down to |b| ~ 2 to 3°. A different sort of problem arises from the fact that galaxies and Galactic extended sources are sometimes difficult to discriminate from each other based on their morphology and colors. H II regions are found predominantly at galactic latitudes <3°. For this study, the number of extended-sources brighter than K_s = 13^th mag (~4.4 mJy), including galaxies, Galactic nebulae and multiple groupings of stars, is ~7,000 in total.

Finally, to firmly establish the extra-galactic nature of 2MASS extended sources deep in the ZoA, we are carrying out systematic follow up observations of sources brighter than ~12.5 at K_s using radio telescopes at Arecibo and Nancay, and optical telescopes at the F.L. Whipple (FLWO) and Palomar observatories. Here we report preliminary (0.38 - 0.75 µm) spectroscopic results for 8 sources using the FLWO Tillinghast reflector, where measured redshift uncertainties range between 20 and 40 km/s, and for an additional 8 sources (mostly invisible at optical wavelengths due to dust obscuration) 21-cm spectroscopy using the new Gregorian reflector system on the 305-m Arecibo telescope, achieving 21 km/s resolution after smoothing with 0.9 mJy rms.

Our objective is to assess the quality of the 2MASS extended-source data, to forecast the quality of the catalogs that are to be ultimately produced, and to identify galaxies located behind the Milky Way. To this end, we have collected all of the sources in the defined area (º2) that are deemed "extended" (i.e., resolved) with respect to the point spread function (PSF) by the 2MASS project, as of the Second Incremental Data Release (e.g., http://www.ipac.caltech.edu/2mass/releases/second/index.html). We have looked at all sources brighter than 13^th mag at K_s, which is slightly fainter than the 2MASS confusion noise limit in the ZoA.

To determine the nature of the object, in particular, extragalactic versus stellar, we have visually examined each source using 2MASS data. We compare the NIR images with independently acquired optical-wavelength image data. The Digitized Sky Survey (DSS), for example, is well matched in surface brightness and resolution to that of 2MASS for normal-colored galaxies. Some sources can be cross-identified with astronomical databases (e.g., NED, ZCAT and SIMBAD). Finally, for some cases in which the reddening is severe (e.g., at |b| < 5-10°, the DSS is largely ineffective), we obtained additional radio, optical or deep infrared data. Previously identified/catalogued sources in the ZoA tend to be Galactic nebulae, such as H II regions, which have very red colors, J-K_s > 1.5, redder than most 2MASS extragalactic sources. It is not possible to clearly identify every source detected in the survey. In particular, "fuzzy" objects at very low galactic latitude are likely to be Galactic nebulae (e.g., H II regions), representing the primary "contaminant" to any extragalactic catalog that extends below |b| < 3°.

For simplicity and owing to the difficulty with discriminating galaxies from Galactic nebulae, we visually categorize extended source candidates as either (1) extended (including galaxies and Galactic nebulae), (2) false (double & triple stars, artifacts), and (3) unknown. Identification of multiple stars is generally straightforward owing to their relatively "blue" J-K_s colors with respect to extragalactic objects, which suffer a greater level of reddening due to foreground Galactic extinction. Artifacts arise from two primary sources: bright stars (e.g., Altair and Vega are in close proximity to the field of study) and transient events (e.g., meteor streaks); see Jarrett et al. (2000) and the 2MASS Explanatory Supplement (Cutri et al. 2000, online at http://www.ipac.caltech.edu/2mass/releases/docs.html) for more details. Most of the artifacts can be eliminated by applying optimized algorithms that recognize their signatures, but there will always be artifacts that are difficult to eliminate with automated routines. Although time-consuming, visual inspection remains the most reliable method for artifact elimination.

With this classification scheme we are able to assess the absolute reliability and internal completeness of a 2MASS extended source "catalog", as well as quantify the color gradient of galaxies with Galactic latitude. One should bear in mind that the final 2MASS extended source catalog will contain mostly (>98%) galaxies at |b| > 20°, a mixture of (~90%) galaxies and non-extended sources (~10%) at 5° < |b|< 20°, and finally, both extragalactic (~40%) and Galactic extended sources (~40%) and artifacts (~10-20%) at |b| < 5° , for a large range in longitude. There are inevitably small patches of the Galactic plane where the reliability is likely to be very poor; for example, in the Galactic center region, 350° < l < 10°, the confusion noise overwhelms any attempt at distinguishing extended sources from foreground stars, decreasing the reliability to <5%.

A colorful "zoo" of galaxies are identified in our ZoA field, exhibiting a wide mix in morphology and surface brightness. Deep in the Galactic plane we encounter extremely red and bright fuzzy objects, many of which are associated with the W51 molecular cloud complex. Examples of typical extended sources found in this ZoA region are shown in Fig. 3, with their basic infrared parameters given in Table 1. This sample represents only 1% of the total number of galaxies and nebulae brighter than 12^th mag (10 mJy) at 2.2 µm identified in the field between 40° and 70° longitude. At the "gray"-edge of the ZoA (e.g., upper panel, Fig. 3), about one-quarter 2MASS extended sources are previously catalogued galaxies (see §3.4 for details). As the source density and visual extinction increase, see bottom panels, Fig. 3, only a few catalogued objects remainù mostly faint IRAS point sourceùprecisely where 2MASS will greatly extend our current knowledge of the ZoA.

The 2MASS extended sources have intrinsically red colors, J-K_s > 1(§3.3.1; see also Jarrett et al. 2000) which make them stand out compared to the foreground stellar population. In the ZoA, interstellar reddening amplifies this color contrast, except in the unusual case in which stellar light contamination from the foreground population will slightly offset the measured galaxy colors toward the blue: for example, the galaxy cores of objects #10 and #23 in Fig. 3 are contaminated by foreground "blue" stars. The 2MASS processing attempts to automatically remove as much foreground contaminant light as possible by identifying point sources and subtracting their light profiles (Jarrett et al. 2000). Most of the galaxies in Table 1 have been followed up with deeper imaging and spectroscopic data, detailed below.

The HI-observation galaxies were selected based on their morphology: spiral galaxy profiles and surface brightness that would suggest gas-rich content. The resultant heliocentric redshifts and line widths for the sources are given in Table 1. One galaxy candidate did not have an H I detection, which means either the source is (1) a gas-poor early-type spiral or elliptical galaxy, (2) a Galactic extended source containing little neutral atomic hydrogen, or (3) outside the range of redshifts 0 < cz < 10,000 km/s surveyed. Several large, >90° in diameter, galaxies are newly discovered, and we have obtained H I line emission profiles (see Table 1). One is a very bright early-type spiral, K ~ 9^th mag: #13 from Fig 3; note that at least 15% of the flux is cut-off due to the western image edge. Likewise, several other new H I-confirmed galaxies are located deep in the ZoA. The largest of which is ~70° in size based on the 20 K_s mag arcsec^-2 isophote, found at a Galactic latitude +6.9° (#26 in Fig 3; Table 1). We highlight this galaxy in Fig. 4.

The galaxy (#26, Fig. 3) is located at a Galactic latitude of 2.6°. In Fig. 4 we compare the 2MASS infrared with the Digitized Sky Survey optical image to demonstrate the effect of extinction on the surface brightness of the background galaxy. It has a K_s brightness of 10.30 mag (54 mJy), and an unusually red J-K_s color of 1.51 mag. The galaxy appears to be an ~Sb or later-type spiral, from which we would expect an intrinsic color of between 0.9 and 1.1. The inferred reddening is therefore approximately 0.4 to 0.6 mag in J-K_s, corresponding to a total visual extinction of 3 to 4 mag -- the visual light is attenuated by a factor of 15 to 40, while the 2.2µm light is attenuated by only a factor ~2. We may compare this rough empirical estimate with what is deduced from coarse-resolution H I, IRAS and COBE surveys of the field, all of which trace the interstellar dust extinction to one degree or another. The H I column density, obtained from the Dwingloo H I survey; Hartmann & Burton (1997), along the line of sight to the galaxy is 0.686 × 10²² cm^-2, translating to a visual extinction of ~3.6 mag (assuming A_v = N_HI × 5.3 × 10^-22 cm²; see Bohlin, Savage & Drake 1978). At considerably better spatial resolution, IRAS and COBE/DIRBE far-infrared maps suggest that the visual extinction is slightly higher, at ~4.9 mags. The far-infrared derived extinction implies that, based on its infrared morphology, the galaxy is even bluer and later type than crudely estimated. Correcting for extinction, the K_s-band isophotal integrated flux of the galaxy becomes 9.6 - 9.9 mag (75 - 100 mJy), suggesting that it is located nearby if we assume an L* luminosity.

The galaxy has a strong H I detection, top panel Fig. 4 (and Table 1 for the line flux), indicating that it is indeed a nearby gas-rich spiral at 2722 km/s. In addition, we have obtained higher spatial resolution K_s-band (2.2 µm) observations with the Palomar 200" telescope to more accurately gauge the morphology and extent of the galaxy. These data, presented in Fig 5, achieve a sensitivity limit of 21.6 mag/arcsec² (1.6 µJy/ arcsec²) in surface brightness, or about 10 times fainter than the 2MASS limit. The deep K-band images corroborate the 2MASS finding of a late-type spiral galaxy with a highly inclined orientation. The light profile exponentially extends down to ~21 mag/arcsec² (2.8 µJy/ arcsec²), translating to a total size of ~70´´ as measured at 2.2 µm.

Finally, to illustrate that 2MASS will adeptly detect Galactic nebulae, in addition to extragalactic fuzz, we show the 2MASS 3-color wide-field image of the W51 giant molecular cloud complex (Fig. 6). Both in size and mass, W51 is one of the most significant GMCs in the Milky Way, and not surprisingly, is undergoing a vigorous starburst (Carpenter & Saunders 1998; Koo 1997 & 1999). The 2MASS automated extended source detector splits the W51 complex into dozens of smaller (<1´) sources, including dense star clusters, H II regions, compact radio sources (demarked in Fig 6), diffuse nebulae and multiple "mini-cluster" groupings of stars. For the region studied in this paper, the W51 complex dominates the extended source number counts for |b| < 2° (see §3.3.3 below).

To demonstrate the usefulness of 2MASS in probing through the ZoA, in this paper we construct an extended source "catalog" for sources brighter than 13^th mag (4 mJy) located between 40° and 70° Galactic longitude, and ±40° latitude. The catalog can then be evaluated for completeness and reliability, and finally basic global measurements may be carried out.

The 2MASS extended source database is designed to err on the side of completeness, particularly for fields subject to high source confusion. Consequently, in order to derive a reliable catalog of extended sources in the ZoA, it is necessary to apply thresholds on key star-galaxy discrimination parameters, and to further cull out artifacts by masking large areas around bright stars. The most powerful parameter that separates galaxies from stellar sources is the "gscore" (see Jarrett et al. 2000, 2000b). This parameter is derived from a machine-assisted oblique decision tree network that utilizes a nine-dimensional parameter space, including the radial "width", intensity-weighted moments, area, symmetry, brightness, and J-H and H-K_s colors of the object. As noted earlier, color is an effective separator between foreground stars and background-reddened galaxies. The "gscore" ranges from 1==extended to 2==pointlike. A "gscore" threshold of 1.4 turns out to be a good compromise between improved reliability and stable completeness.

After applying this "gscore" threshold, and masking regions around bright stars (K_s < 5^th mag) appropriate to the confusion noise level, the overall reliability of the catalog is typically better than 90% for |b| > 10°. Double stars account for most of the non-extended "contamination" to the catalog, while particularly stubborn artifacts account for 1 to 3% of the total. For ~1000 deg² of area, we catalog ~7000 extended sources with K_s < 13 (4 mJy), and ~2000 with K_s < 12.1 (10 mJy). A sky plot of the sources is given in Fig. 7, where the left panel (a) corresponds to K_s < 12.1 mag and the right panel (b) to K_s < 13.0 mag. The gray regions denote areas that either have yet to be surveyed by 2MASS or have been observed but whose data was deemed poor quality (and are scheduled for future re-observation). The diamond symbols represent cataloged galaxy clusters. Note the dense clustering of sources at northern latitudes, from 20 to 40°; in §3.3.3 and §4 we will address this apparent clustering. Near the galactic plane, the distinct cluster of points near l ~ 50° and b ~ -1° corresponds to W51 (Fig 6). Note the apparent decreasing completeness starting at |b| < 10° for K_s < 13 (Fig. 7b). For K_s < 12.1 (Fig 7a), however, the catalog is much more uniform throughout, suggesting that 2MASS is complete for these flux levels (cf. §3.3.3).

In order to access asymmetries in the number counts above and below the Galactic plane, we split the source catalog into "northern" and "southern" hemisphere components. Table II (north) and Table III (south) details the key global properties of the catalog, including the stellar number "density" (stars per deg² brighter than 14^th mag at K_s), area (deg²), mean J-K_s color, mean "H I" gas column density (3 × 10²¹ atoms per cm²; derived from the maps of Hartmann & Burton 1997), total number of extended source candidates ("Nx"), completeness ("C"), reliability ("R") and area normalized source "counts" (cumulative number per deg²). A cautionary note about the area normalization: arriving at the correct integrated area for a given latitude interval requires great care, including corrections for coverage gaps, scan-to-scan overlap, masked circular areas due to moderately bright stars (5^th < K_s < 9^th), and masked areas (both circular and rectangular) due to bright stars (K_s < 5th mag). A more detailed prescription for computing the areal coverage for 2MASS extended source data is given in Jarrett et al. (2000b).

 3.3.1 Extinction & Color vs Galactic Latitude

The mean J-K_s color for the northern and southern catalog is shown in Fig 8 (upper panels), where the filled circles correspond to the north (b > 0°) and the open circles to the south (b < 0°). Here we have computed the mean color within the specified latitude range (see Tables 1 & 2) and for 40° < l < 70°. For comparison, the corresponding mean hydrogen gas column density is also shown (solid line denotes the north, dotted line the south). The atomic gas serves as a proxy measure to the interstellar dust extinction (Bohlin, Savage & Drake 1978; see also §3.2), assuming that the 21-cm emission line is optically thin.

First converting the gas column density to equivalent visual extinction (multiplying by 5.3 × 10^-22 cm²; Bohlin, Savage & Drake 1978; Burstein & Heiles 1982; Bernard et al. 1992) and then to an equivalent (J-K_s) extinction using E(J-K_s) = 0.54 and R = 3.1, the predicted extinction nearly matches, to better than ~5%, the measured J-K_s of extended sources in the ZoA for |b| > 5° . The lower panels in Fig 8 show the J-K_s color corrected for extinction based on the HI column density, revealing a uniform color distribution down to |b| ~ 5°. This result is not surprising given that nearly all of the extended sources are galaxies located behind the total column of neutral hydrogen gas and the typical J-K_s color for unreddened normal galaxies is about ~1.05 (dashed line in Fig 8; see also Jarrett et al. 2000 & 2000b). The extinction is roughly uniform above and below the Galactic plane as the northern vs. southern sources have about the same color vs. b distribution down to |b| ~ 3°.

Below 3° the extinction is undoubtedly very patchy (e.g., Fig. 2 & 6) and contamination from intrinsically red extended sources (e.g., H II regions) is severe, rendering the HI-corrected colors as unreliableùat low Galactic latitude nebulae associated with the Milky Way dominate the integrated colors of extended sources. An alternative method at estimating the extinction within 3° of the Galactic plane is to use the mean colors of the non-extended field stars. See, for example, Fig 2 where the reddening gradient is dominated by colors of stars. Although there is no shortage of stars from which to estimate ensemble averages, the main disadvantage of this method is that the mean color will represent a mix of sources along the lines of sight, and therefore subject to a wide range in dust obscuration. The solution is to instead use the colors of extragalactic objects, which are located behind the total column of dust of the Milky Way. Even then, this method requires a large sample of "normal" galaxies to average down systematics arising from the intrinsic color and internal extinction of the galaxy sample.

Producing a catalog is an exercise in balancing completeness and reliability. Flux and star-galaxy parameter thresholds are chosen to maximize the extended source completeness while also minimizing contamination from non-extended or false sources. In addition to artifacts from bright stars, the primary contaminants are double stars and triple+ stars for the Galactic plane fields. Hence, the reliability is a strong function of the stellar number density or simply background "confusion". Our strategy is to err on the side of completeness, with an internal completeness better than 90%, giving a larger set of extended source candidates that can then be visually examined.

Completeness, in an absolute sense, refers to the ratio of sources that are detected to the total number of sources that should be detected to some flux limit and surface brightness. In practice it is very difficult to establish the latter, since it requires an independent data set that is more sensitive and is, in itself, reliable. For "normal" early-type galaxies the 2MASS extended source catalog has been estimated to be at least 95% complete in an absolute sense for most of the sky where the stellar source density is moderate to low (|b| > 30°; Jarrett et al. 2000). Note that 2MASS is not sensitive to late-type spiral galaxies, including those of Sc and Sd types, particularly if they are oriented face-on, thereby decreasing their overall surface brightness (see Jarrett 2000). For the Galactic plane, stellar confusion is the most significant and deleterious effect that limits both detection, characterization and discrimination, rendering a final completeness that is considerably less than 95%. The relative loss in sources (or decreasing detection rate) as a function of galactic latitude (or stellar number density) can be estimated by measuring the "internal" completeness.

The internal completeness is defined to be the ratio or percentage of detected and verified extended sources that are classified as extended versus the total number of detected extended sources. We emphasize that for the ZoA we can only estimate the fractional loss of sources that we would normally detect at high galactic latitude. Moreover, any "internal" measure must necessarily underestimate the true loss in sources with galactic latitude. Some galaxies, for example, are simply not detected by 2MASS in the ZoA due to a variety of reasons, including source confusion in which stars directly obscure the background source, bright star masking, and surface brightness. Late-type spirals are much fainter than early-type spirals or ellipticals in the near-infrared. Since we cannot reliably estimate the number of background sources that are not detected by 2MASS, any measure of "completeness" can only approximate the true completenessù more specifically, it represents an upper limit. In addition, a small percentage of sources, ~1 to 3% of the total, cannot be reliably identified as stellar, Galactic or extragalactic. Since they only contribute at most a few % uncertainty in the completeness and reliability for |b| > 3°, we ignore their contribution to the overall completeness and reliability. For the most confused regions where the fraction of "unknown" sources rises to 10 to 30%, their contribution is significant (see below).

Reliability is defined to be the ratio or percentage of detected and real verified extended sources versus the sum of detected and real extended plus verified false sources that pass our "gscore" thresholding. Note that the reliability depends on the presence of real extended sourcesùthe more extended source there are in some region (e.g., a galaxy cluster) the higher the reliability will beùthe reliability is a coupling between the density of real extended sources and the level of contamination as given by the stellar source density. For the Galactic plane, the confusion noise drives down the sensitivity, decreasing the total number of extended sources that are detected, while at the same time the number of sources that dominate contamination is on the rise. Hence, the reliability is a strong function of the stellar number density. As with the completeness calculation, sources classified as "unknown" are ignored here, representing a statistically insignificant number of sources for |b| > 3°.

Figure 9 shows the achieved internal "completeness" and reliability for the combined "north" and "south" catalogues. We have not applied any corrections for extinction. The internal completeness holds steady between 90 and 95% for most of the ZoA. The reliability also holds steady between 90 - 95% all the way down to |b| ~ 5 to 7° (corresponding to 10⁴ stars deg^-2 brighter than 14^th mag at K_s). But, as the confusion noise dramatically rises for |b| < 5°, the reliability decreases to ~40%. Note also that the 40% reliability is inflated due to the presence of W51, a large source of fuzzy objects less than 1° from the Galactic plane.

Since the actual total number of sources is small, it is feasible to visually inspect each extended sources candidate and eliminate most of the false sources (e.g., multiple stars) yielding an corrected reliability greater than 90%. There will still remain a large number of sources whose nature cannot be discerned from visual inspection or otherwise using 2MASS parametric information: ~28% of the extended sources at |b| < 3° are classified as "unknown". These sources can significantly alter the true reliability. But, as a final note we should point out that the very definition of what is "extended", at least with respect to the PSF, breaks down as the source density blows up near the Galactic plane. In the Galactic center, just about everything is "extended" as nebulosity and walls of stars merge in projection on the sky.

Figure 10 shows the cumulative number of extended sources per deg² for flux thresholds K_s = 12.1 and 13.0 (4 and 10 mJy, respectively). We have not applied any corrections for dust extinction. The northern sample (b > 0°) is denoted with filled circles and the southern sample (b < 0°) with open circles. Three trends are clear: (1) an "excess" of sources between +30° and +40° latitude, (2) declining counts between 5-12° latitude, and (3) an upturn in the declining source counts for bright sources near the Galactic plane, |b| < 3°. The average trend for |b| > 5° is ~8-10 galaxies per deg² brighter than 13^th mag at K_s and ~1-2 galaxies per deg² brighter than 12^th mag at K_s. Finally, there is statistically significant difference between the source counts in the northern (b > 0°) and southern (b < 0°) fields. The asymmetry in counts is most likely due to real large-scale structure from background galaxy clusters (see below).

The excess at high galactic latitude comes from the presence of a number of prominent galaxy clusters (see Fig 7). A large-scale structure, extending from 2000 - 5000 km/s, has been previously identified by Marzke et al. (1996). More generally, the northern galaxy counts (b > 0°) tend to be greater than that of the south (b < 0°), highlighting the apparent "wall" of galaxies to the north of the Galactic plane.

The 30 to 50% decline in source counts between 5° and 12°, corresponding to an increase in stellar number density of 3200 to 10,000 stars per deg² brighter than 14^th mag at K_s, tracks the increasing background confusion noise and subsequent loss in completeness for extragalactic objects. The extended source completeness, specifically the internal component, was first shown in Figure 9 to be ~85% to 90% for |b| < 10°. Since the number counts measure both the internal and absolute completeness, assuming a smooth spatial distribution of galaxies, we infer that 2MASS is not detecting from 15 to 35% of the total possible number of extended sources (i.e., galaxies) brighter than 12^th mag at K_s deep in the ZoA, |b| ~ 5° and l ~ 50°.

The upturn in the source counts for bright sources at low Galactic latitude, |b| < 3°, is due to foreground Galactic nebulae, including H II regions, planetary and diffuse nebulae, which are predominantly confined to the galactic plane near |b| = 0°. The W51 GMC is the most prominent cluster of Galactic nebulae and dense clusters of young stars in the study area (cf. Fig 6). At these source densities, >20,000 stars per deg² brighter than 14^th mag at K_s, the confusion noise is so large that detection of background galaxies is limited to only the nearest or intrinsically brightest galaxies (e.g., Fig. 5). Conversely, star formation regions, and most GMCs, tend to be confined to the Galactic plane with scale heights < 0.1 kpc. Both effects result in an excess of extended source counts deep in the ZoA, at least for l ~ 50°, with patchy variation throughout the disk of the Milky Way.

The NASA Extragalactic Database (NED) provides a means for comparison between 2MASS galaxy detections and previously catalogued extragalactic sources. The primary objective is to estimate the number of new galaxies that 2MASS will add to NED, and in particular, to the traditionally incomplete ZoA. We also utilize SIMBAD to search for Galactic extended objects (e.g., H II regions), which are a significant component to the total number of extended sources at the highest source densities. Figure 11 summarizes the 2MASS vs. NED/SIMBAD cross-comparison. Here we have split the extended sources into (1) galaxies, (2) radio & infrared sources (which may also be galaxies, but are more likely Galactic in nature), and (3) planetary nebulae. The left upper/lower panels correspond to the southern fields (b < 0°), and the right upper/lower panels to the northern fields (b > 0°). For b > 20°, upwards of 30-60% of the 2MASS sources are previously catalogued galaxies -- most of which belong to the cluster of galaxies discussed in §3.3.3. The percentage drops dramatically as we enter the ZoA, where only 5 to 25% are previously catalogued galaxies for 5° < |b| < 20°. For galaxies, the percentage drops to zero near the Galactic plane. The number of radio/infrared and planetary nebulae increase to 5 - 10% of the total at these extreme stellar source densities. It is likely that the true percentage of Galactic extended sources at |b| < 3° is at least 50%, reflecting the incompleteness of previously catalogued sources in the Galactic plane.

The plane of the Milky Way presents a number of obstacles toward detection of background galaxies. At optical wavelengths, the dust obscuration attenuates the already diffuse light coming from galaxies (relative to stars). At longer wavelengths, such as the far-infrared (e.g., IRAS), the dust is transparent but limitations arising from spatial resolution and sensitivity result in missed sources. Radio surveys provide a clear window toward detection of nearby gas-rich (i.e., late-type spiral) galaxies, but currently lack the sensitivity to detect more distant spirals (cz > 5,000 km/s) or gas-poor early-type galaxies. The near-infrared, in contrast, offers a sort of compromise between the optical and radio bands. Dust attenuation is an order of magnitude less at 2.2 &micro;m compared to optical bands, while the spatial and sensitivity limits of surveys such as 2MASS and DENIS allow detection of both spiral and elliptical galaxies background to the ZoA out to redshifts ~10-20,000 km/s (cf. Kraan-Korteweg et al. 1998; Mamon et al. 1999; Schroeder, Kraan-Korteweg & Mamon, 1999). As demonstrated in Fig 11, 2MASS will increase the number of new galaxy detections in the ZoA by a factor of ~3. Moreover, for galaxies brighter than 12.1 mag (10 mJy) and sizes >14´´, 2MASS will provide a complete (~80-90%) sample at a rate of ~2 galaxies per deg² penetrating the ZoA down to |b| ~ 5°. Lifting the veil of the Milky Way, a more complete map of large-scale structure will be possible with follow-up redshift observations of 2MASS galaxies.

Given the detection rate in this region of the Milky Way, we may estimate the total number of galaxies that 2MASS will find in the entire ZoA, excluding ±20° around the Galactic center. Integrating the rates for |b| < 20° gives between 12,600 and 25,000 galaxies brighter than 12.1 mag (10 mJy) at K_s, and deeper in the plane, |b| < 10°, between 6,400 and 12,800 galaxies brighter than 12.1 mag (10 mJy) at K_s. From 75 to 95% of these galaxies will be newly catalogued by 2MASS, with a reasonable fraction having H I counterparts coming from the deep 21-cm surveys. We may expect more new galaxies since the region studied in this work, 40° < l < 70°, has a much higher stellar density, translating to higher confusion noise and lower sensitivity, than in the Galactic anti-center region. Moreover, the ZoA region studied here is likely to be underdense in galaxies in comparison to the anti-center/Supercluster fields and the region encompassing the Great Attractor (e.g. Kolatt et al. 1995). We expect many of these sources to be in common with ongoing H I surveys, although it is worth noting that none of the Arecibo H I detections in the present work is sufficiently strong that they would have been detected in the shallow H I surveys of the past few years. We can expect at least half to be detected only by 2MASS.

The 2MASS census of galaxies in the ZoA will provide the necessary information toward identification of galaxy clusters veiled by the Milky Way. These clusters in turn will help establish the large-scale structures extending into the ZoA and probe the mass density of the Hydra-Centaurus-Pavo-Indus-Telescopium wall of galaxies. By relating the observed infall velocity to the distribution of matter, we can estimate the mean mass density of the Universe.

The redshift distribution for previously catalogued galaxies and from this work is given in Table 1 and illustrated in Fig 12, ranging between 2000 and 50,000 km/s. It should be noted that many of the galaxies in Table 1 have redshifts between 5000 to 8000 km/s, which is where we would expect an extension of the large-scale "Great Wall" (Geller & Huchra 1989) to be located if the western extent (see Marzke et al 1996) were beginning to close off as another "bubble" feature. Although the distribution is not from a complete sample of galaxies, the results do show peaks for galaxies at 5500 and at 8500 km/s. In spatial projection (Fig. 7) we see large-angular scale structures extending from b ~ 35° down to 20° (but may extend as far as 10°), with a clear clustering of galaxies near l ~ 60° and b ~ 19°. Unfortunately the current gaps in the 2MASS coverage hides the true extent of this cluster, as well as any large-scale structures in the southern fields. The more complete 2MASS map, which will be available in the near future, and a more complete redshift census of galaxies in these fields, is required to substantiate this hypothesis. Finally, redshift-independent distance measures (e.g., H-band Tully-Fisher) are needed to establish cluster space densities that are non-biased.

Not only will 2MASS and DENIS help reveal large-scale structures hidden behind the Milky Way, but it is still possible that a large previously undetected Local Group galaxy will be uncovered by these near-infrared survey or the H I surveys in upcoming years as these surveys reach their peak in productivity. A galaxy comparable in size to that of M31 will have enough mass to play a significant role in the gravity field of the Local Group, conceivably the source of complex and puzzling peculiar velocities observed in the Local Group and other nearby groups. It is more likely that smaller Local Group galaxies will be discovered, a prospect that would also be important to understanding the Local Group of galaxies. Already 2MASS has discovered large angular-sized galaxies near the Local Group (Hurt et al. 2000; Howard 2000). What we can say with confidence is that the zone of avoidance is now a zone of discovery.

Some ~1000 deg² surveyed by 2MASS between 40° and 70° galactic longitude and ±40° about the Galactic plane include over 7000 verified extended sources brighter than 13.0 at K_s (4.4 mJy) and more than 1600 sources brighter than 12.1 at K_s (10 mJy). For the ZoA, |b| < 20°, most of the sources are newly discovered, with a small fraction, ~5-25%, previously catalogued in NED-catalogued. For a number of new 2MASS galaxies we acquired confirmation observations, including deep K-band imaging and spectroscopy at I-band and radio wavelengths.

On average, 2MASS discovers galaxies at the rate of ~8-10 per deg² brighter than 13^th mag (4 mJy) at K_s and ~2 galaxies per deg² brighter than 12^th mag (10 mJy) at K_s. Near the Galactic plane, |b| < 3°, the extended source counts are dominated by Galactic H II regions and large low surface brightness nebulae. The W51 GMC (l ~ 49°, b ~ -0.5°) is clearly detected and split into dozens of 2MASS extended sources. Most important, 2MASS discovers new galaxies near the Galactic plane, at least down to stellar source densities as high as 10,000 stars per deg² brighter than 14^th mag at K_s. At the final completion of the 2MASS survey, we project that over 20,000 "normal" galaxies will be detected in the ZoA, |b| < 20°, at least half of which will be detectable in H I in deep 21-cm radio surveys.

Aaronson, M., Huchra, J., Mould, J., Schechter, P.L., & Tully, R.B. 1982, ApJ, 258, 64.

Bernard, J., Boulanger, F., Desert, F., & Puget, J. 1992, A&A, 263, 258.

Bohlin, R., Savage, B., & Drake, J. 1978, ApJ, 224, 132.

Branchini, E., Teodoro, L., Frenk, C.S., et al., 1999, MNRAS, 308, 1.

Burstein, D. & Heiles, C. 1982, AJ, 87, 1165.

Carpenter, J. & Sanders, D. 1998, ApJ, 116, 1856.

Canavezes, A., Springel, V., Oliver, S.J., et al., 1998, MNRAS, 297, 777.

Cutri, R.M., et. al., 2000, The 2MASS Explanatory Supplement.

Da Costa, L.N. et al. 1994, ApJ, 424, L1.

deLapparent, V., Geller, M.J. & Huchra, J.P. 1986, ApJ, L1.

de Vaucouleurs, G., de Vaucouleurs, A., Corwin, H.G., Buta, R,. Paturel, G., & Fouque, P., 1991, Third Reference Catalogue of Bright Galaxies.

Di Nella, H., Couch, W., Parker, Q. & Paturel, G. 1997, MNRAS 287, 472.

Epchtein, N. et al. 1997, The Messenger, 87, 27.

Faber, S. M., & Burstein, D. 1988, In The Vatican study week on large scale motions in the Universe, ed. G. V. Coyne, V. C. Rubin (Princeton: Princeton Univ. Press), 115.

Fisher, K. B., Huchra, J. P., Strauss, M. A., Davis, M., Yahil, A. & Schlegel, D. 1995, ApJ.S. 100, 69.

Geller, M. & Huchra, J. 1989, Science 246, 897.

Giovanelli, R. & Haynes, M. 1988, IAU-Symp, 130, 113.

Hartmann, D. & Burton, W.B. 1997, Leiden/Dwingeloo survey of HI in the galaxy. Cambridge Univ. Press.

Henning, P. 1992, ApJ, 78, 365.

Henning, P., et al. 1998, AJ, 115, 584.

Henning, P., Staveley-Smith, L, Kraan-Korteweg, R, & Sadler, E. 1999, PASA, 16, 1.

Howard, E., 2000, AJ, submitted.

Hurt, R., Jarrett, T.H., Kirkpatrick, D., Cutri, R.M., Skrutskie, M., Schneider, S., Skrutskie, M., & van Driel, W. 2000, AJ, submitted.

Jarrett, T.H. 2000, PASP, in press (Aug issue)

Jarrett, T.H., Chester, T., Cutri, R., Schneider, S., Skrutskie, M. & Huchra, J. 2000, AJ, 119, 2498.

Jarrett, T.H., Chester, T., Cutri, R., Schneider, S., Skrutskie, M. & Huchra, J. 2000b, AJ, in preparation.

Juraszek, S. 1999, PASA, 16, 1.

Kaiser, N., et al. 1991, MNRAS, 252, 1.

Kilborn, V., Webster, R.L., & Staveley-Smith, L. 1999, PASA, 16, 8.

Kolatt, T., Dekel, a., & Lahav, O. 1995, MNRAS, 275, 797.

Koo, B.C. 1997, ApJS, 108, 489.

Koo, B.C. 1999, ApJ, 518, 760.

Kraan-Korteweg, R., et al. 1994, Nature 372, 77.

Kraan-Korteweg, R., Woudt, P., Cayette, V. et al. 1996, Nature, 379, 519.

Kraan-Korteweg, R. & Huchtmeier, W. 1992, A&A, 266, 150.

Kraan-Korteweg, R. & Woudt, P. 1999, PASA, 16, 1.

Kraan-Korteweg, R., Schroeder, A., Mamon, G, & Ruphy, S. 1998, in "The Impact of Near-IR Surveys on Galactic and Extragalactic Astronomy", ed. N. Epchtein, Kluwar (Netherlands), 209.

Lahav, O., Yamada, t., Scharf, C. & Kraan-Korteweg, R. 1993, MNRAS, 262, 711.

Lu, N. & Freudling, W. 1995, ApJ, 449, 527.

Lu, N., Dow, M., Houck, J. & Salpeter, E. 1990, ApJ, 357, 388.

Lundmark, Knut, 1940, Lund Observatory.

Lynden-Bell, D., et al. 1988, ApJ, 326, 19.

Mamon, G, Borsenberger, J., Tricottet, M & Banchet, V. 1998, in "The Impact of Near-IR Surveys on Galactic and Extragalactic Astronomy", ed. N. Epchtein, Kluwar (Netharlands), 177.

Marzke, R., Huchra, J. & Geller, M. 1996, AJ, 112, 1803.

Nakanishi, K., et al. 1997, ApJ, 112, 245.

Oliver, S., et al. 1996, MNRAS, 280, 673.

Pantoja, C., Altschuler, D., Giovanardi, C. & Giovanelli, R. 1997, AJ, 113, 905.

Rowan-Robinson, M., et al. 1990, MNRAS, 247, 1.

Saunders, W., et al. 1991, Nature 349, 32.

Schneider, S.E., Thuan, T., Mangum, J. & Miller, J. 1992, ApJS, 81, 5.

Schroeder, A., Kraan-Korteweg, R. & Mamon, G. 1999, PASA, 16, 1.

Skrutskie, M., et al. 1997, in "The Impact of Large Scale Near-IR Sky Surveys", ed. Garzon et al., Kluwar (Netharlands), 210, 187.

Smoot, G., Gorenstein, M., & Muller, R. 1977, Phys. Rev. L. 39, 898.

Staveley-Smith, L, et al. 1998, AJ, 116, 2727.

Staveley-Smith, L, 1997, PASA, 14, 111.

Acknowledgements. We thank Glenn Veeder for help in obtaining the Palomar near-infrared observations and Perry Berlind and Mike Calkins for optical spectroscopy. We thank Raymond Tam with help in galaxy classification. Ron Beck, Diane Engler and Helene Huynh were instrumental in gathering 2MASS image data. We thank Geraldine Jarrett for her sharp text editing. Finally, we thank Daniel Egret for providing fast access to the SIMBAD database. The Digitized Sky Surveys were produced at the Space Telescope Science Institute. This research has made use of the NASA/IPAC Extragalactic Database (NED) which is operated by the Jet Propulsion Laboratory, California Institute of Technology, under contract with the National Aeronautics and Space Administration (NASA). This publication makes use of data products of the 2MASS, which is a joint project of the Univ. of Massachusetts and the Infrared Processing and Analysis Center, funded by the NASA and NSF. This work was supported in part by the Jet Propulsion Laboratory, California Institute of Technology, under a contract with NASA.

Table notes: index is identifying number in Fig. 3; ra & dec are the equatorial coordinates (J2000); b is the galactic latitude coordinate; 2MASS K_s mag and J-K_s color; cz is the heliocentric redshift in km s^-1; W50 is the 50% H I velocity widths in km s^-1; atomic hydrogen mass, M_HI, given in units of 10⁹ M_o, assuming H₀ = 75 km/s/Mpc. For source #10, spectroscopic measurements indicate that a foreground star is located directly in front of the galaxy core. Reference code: (1) Marzke et al. 1996; (2) RC3.9; (3) Lu et al. 1990; (4) Schneider et al. 1992; (5) Fisher et al. 1995; (6) Nakanishi et al. 1997; (7) mostly like Galactic reflection nebula.

TABLE 2 Extragalactic & Galactic Extended Source Counts: b > 0°

Table notes: "density" is the number of stars per deg² brighter than 14^th mag at K_s; "H I" is the atomic hydrogen column density (3 × 10²¹ atoms cm^-2; derived from the maps of Hartmann & Burton 1997); "Nx" the total number of extended source candidates; "C" is the internal completeness and "R" the reliability; "counts" is the area normalized extended source counts (cumulative number per deg²).

TABLE 3 Extragalactic & Galactic Extended Source Counts: b < 0°

Table notes: see Table 2 notes.

Figure 1--The visible Milky Way Galaxy. Home to some 400 thousand-million stars, including the Sun, interstellar gas and dust, nearly all of which are confined to the "disk" or plane of the Galaxy. The Galactic plane, depicted here in a scientifically-accurate drawing from the Lund Observatory (Lundmark 1940) as seen at visible wavelengths, is characterized by patchy and thick dust lanes crossing a very high surface density "disk" of stars. This region fills nearly a quarter of the observable sky, in effect a giant curtain or veil to the Universe beyond the Milky Way. 2MASS will survey the entire sky, revealing both deeply embedded star formation regions and galaxies located beyond the Milky Way. The region of the plane relevant to this work is indicated with a green box, located between 40° and 70° in longitude from the Galactic center, in the Sagittarius Arm Tangent.