THE CARNEGIE-IRVINE GALAXY SURVEY. I. OVERVIEW AND ATLAS OF OPTICAL IMAGES

To fully capture the wealth and range of the morphological properties of the local galaxy population, to maximize signal-to-noise ratio (S/N) and spatial resolution, and to ensure statistical completeness, we target a large, well-defined sample of the brightest objects optimally placed for Carnegie's facilities at Las Campanas Observatory. We impose no selection according to galaxy morphology, size, or environment. The sample, consisting of 605 objects, is formally defined by B_T ⩽ 12.9 mag and δ < 0°. The completeness level of bright galaxies to this magnitude limit is essentially 100% (Paturel et al. 2003), yielding a sample roughly comparable in size to that of the Palomar survey of northern galaxies (Ho et al. 1995, 1997a), which is desirable for future comparisons between the two hemispheres. In practice, we select objects from the Third Reference Catalogue of Bright Galaxies (RC3; de Vaucouleurs et al. 1991), with the aid of the online database HyperLeda⁸ (Paturel et al. 2003).

CGS provides a fair, statistically unbiased representation of the local galaxy population over the absolute magnitude range −16 ≲ $M_{B_T}$ ≲ −23. This is illustrated in Figure 1, which shows the B-band luminosity function calculated using the 1/V_max method of Schmidt (1968), compared with the r-band luminosity function of z ≈ 0.1 galaxies derived from the SDSS by Blanton et al. (2003). In making this comparison, we assume a typical galaxy color for Sbc galaxies (B − r = 0.67 mag; Fukugita et al. 1995), which is approximately the median morphological type of CGS, and a Hubble constant of H₀ = 73 km s⁻¹ Mpc⁻¹. The two functions agree quite well, both in shape and in normalization. The relatively large fluctuations on the faint end of the luminosity function reflect the density inhomogeneities in the local volume.

Zoom In Zoom Out Reset image size

Table 1 and Figure 2 summarize some of the basic parameters of the sample. Given the bright magnitude limit of the survey (Figure 2(a)), it is not surprising that most of the galaxies are quite nearby (median D_L = 24.9 Mpc; Figure 2(b)), luminous (median $M_{B_T} = -20.2$ mag, corrected for Galactic extinction; Figure 2(c)), and angularly large (median B-band isophotal diameter D₂₅ = 33; Figure 2(d)).

Zoom In Zoom Out Reset image size

The sample spans the full range of Hubble types in the nearby universe (Figure 3), comprising 17% ellipticals, 18% S0 and S0/a, 64% spirals, and 1% irregulars. A handful of well-known interacting systems (e.g., the "Antennae," Centaurus A, NGC 2207) are included, but the vast majority of the sample have relatively undisturbed morphologies, although the RC3 classifications designate ∼16% of the sample as "peculiar" in some form. As with the Palomar sample (Ho et al. 1997b), roughly 2/3 of the disk (S0 and spiral) galaxies are barred according to the classifications in the RC3, with the breakdown into strongly barred (SB) and weakly barred (SAB) being 37% and 29%, respectively. Approximately 10% have an outer ring or pseudoring (Buta & Crocker 1991).

Zoom In Zoom Out Reset image size

In addition to the parent sample of 605 galaxies, Table 1 also lists 11 sources that were observed early in the survey prior to it being fully defined, but that now formally do not meet the magnitude limit of the survey. We include these extra objects for completeness, but they will be omitted from future statistical analyses.

The initial processing follows standard steps for CCD data reduction within the IRAF⁹ environment, using tasks within ccdproc, which include trimming, overscan correction, bias subtraction, and flat fielding. Dark current subtraction is unnecessary. We took dark frames with integration times comparable to those of the longest science exposures, but the dark current always turns out to be negligible. For each night of each observing run, we generate a master bias frame by averaging a large number (typically ∼20) of individual bias frames. Similarly, we create a master flat-field frame with high S/N by combining a series (typically ∼6–10 per filter) of domeflats and twiflats. Through experimentation, we found that the twiflats produce more uniform illumination than the domeflats for the B and V images, whereas the domeflats are more effective for the R and I images. A small fraction of the flats contained dust specks that introduced "doughnut" features into the flattened frames; roughly ∼4% of the science images were affected by this. The flattened I-band images contain subtle, residual fringe patterns with amplitudes (peak to trough) at the level of ∼1%–2%. We remove these by subtracting an optimally scaled fringe frame with a mean zero background, which was constructed by median combining a total of 36 sky images collected over many observing runs. The fringe pattern proved to be remarkably stable throughout the course of the survey. Apart from a couple of partially dead columns near the center of the chip, the Tek#5 CCD is relatively clean cosmetically. We generate a mask for these and other less conspicuous defective regions, and use it for local interpolation to correct the bad pixels. Finally, we use an IDL version of van Dokkum's (2001) L.A.Cosmic routine to correct the pixels affected by cosmic ray hits and satellite trails; any remaining imperfections were further edited manually.

Next, we shift and align the flattened, cleaned images in each filter to a common reference frame defined by the I-band image, which generally has the sharpest seeing and is least affected by possible dust obscuration. Multiple images taken with the same filter were combined. We carefully examine the images for possible saturation near the core of the galaxy and replace any saturated pixels with their appropriately scaled counterparts from the short-exposure images. We determine the proper scale factor by comparing the background-subtracted, integrated counts of bright, isolated field stars that are unsaturated in both the long and short exposures. The affected regions are generally small, spanning a diameter of only ∼10 pixels. As this is not significantly larger than the point-spread function (PSF), this procedure has a negligible impact on any of our subsequent scientific analysis. For convenience, we flip the images so that north points up and east to the left. Lastly, we use the Astrometry.net (Lang et al. 2010) software¹⁰ to solve for the World Coordinate System (WCS) coordinates of each image and store them in the FITS header.

A little more than half of the CGS galaxies were observed under photometric conditions, as determined from observations of Landolt (1992) standard stars. We measure the photometry of the stars using a circular aperture with a fixed radius of 7'' to mimic as closely as possible Landolt's measurements. We determine the local sky of each star from a 5 pixel wide annulus outside that aperture, using only sky pixels (i.e., avoiding any nearby neighboring stars). The median photometric errors (Table 2, Column 9) for the photometric nights are 0.08, 0.04, 0.03, and 0.04 mag for the B, V, R, and I band, respectively. The images calibrated in this manner have the header keyword ZPT_LAN.

For the non-photometric observations, we adopt an indirect approach that allows us to obtain a less accurate, but still useful, photometric calibration. Our strategy is to bootstrap the instrumental magnitudes of the brighter stars within each CCD field to their corresponding photographic magnitudes as published in the second-generation Hubble Space Telescope Guide Star Catalog (GSC2.3; Lasker et al. 2008). We derive the transformation between the GSC photographic bandpasses and our standard BVRI magnitudes using the subset of field stars that were observed by us under photometric conditions. The GSC contains astrometry, photometry, and object classification for nearly a million objects. The photometry for the southern hemisphere is given primarily in three photographic bandpasses, B_J, R_F, and I_N. The typical photometric error for the stellar objects in the GSC, depending on the magnitude and passband, is 0.13–0.22 mag (Lasker et al. 2008).

Starting with the list of WCS coordinates extracted for the field stars in each CCD image, we use the Vizier server¹¹ to query the GSC for objects classified therein as stellar (class = 0). We eliminate saturated objects and stars fainter than B ≈ 22 mag. To properly compare our instrumental magnitudes with the GSC magnitudes, it is critical that we follow as closely as possible the methodology that the GSC used to derive their photometry. We cannot use conventional methods of stellar aperture photometry (e.g., using standard tasks in IRAF). As outlined in Lasker et al. (2008), the procedures used in the GSC for object detection, flux measurement, sky determination, and source deblending in the case of crowded fields (Beard & MacGillivray 1990) follow very closely those implemented in the SExtractor package (Bertin & Arnouts 1996). We therefore use SExtractor to measure the stars in our CGS images. We manually performed aperture photometry of a number of isolated stars to confirm that the SExtractor-based measurements are reliable, to better than 0.02 mag.

We intercalibrate the CGS and GSC photometric scales using data from a night with exceptional photometric stability, during which the open cluster M67 was observed. Our photometry for M67 agrees very well with that published by Montgomery et al. (1993), to better than 0.01 mag, for all filters. Using a set of 500–600 field stars selected from the science images observed throughout the course of that night, we derive a set of transformation equations between the GSC photometric bandpasses (B_J, R_F, and I_N) and our standard bandpasses (B, V, R, and I). The fitting was done with the IDL function ladfit, which uses a robust least absolute deviation method and is not sensitive to outliers. The transformation equations and residual scatter (σ) (in magnitudes) for each filter are as follows:

$$\begin{eqnarray} B &=& B_J + 0.2807 - 0.1365 (B_J - R_F)\quad \sigma = 0.2088 \end{eqnarray} \tag{ 1 }$$

$$\begin{eqnarray} B &=& B_J + 0.1499 + 3.239 \times 10^{-5} (B_J - I_N)\nonumber\qquad\quad\\ \sigma &=& 0.2490\qquad\qquad\quad\qquad\quad \end{eqnarray} \tag{ 2 }$$

$$\begin{eqnarray} &&B = 0.1489 + 1.0001 B_J\quad \sigma = 0.2597\qquad\qquad\ \end{eqnarray} \tag{ 3 }$$

$$\begin{eqnarray} &&B = R_F + 0.9700 + 0.8888 (R_F - I_N)\quad \sigma = 0.2982 \end{eqnarray} \tag{ 4 }$$

$$\begin{eqnarray} &&B = 0.4508 + 1.0468 R_F\quad \sigma = 0.3832\qquad\qquad\ \end{eqnarray} \tag{ 5 }$$

$$\begin{eqnarray} &&B = 0.8471 + 1.0488 I_N \quad \sigma = 0.6058 \qquad\qquad\ \end{eqnarray} \tag{ 6 }$$

$$\begin{eqnarray} &&V = B_J + 0.1512 - 0.6625 (B_J - R_F)\quad \sigma = 0.1140 \end{eqnarray} \tag{ 7 }$$

$$\begin{eqnarray} &&V = R_F + 0.4658 + 0.2279 (R_F - I_N)\quad \sigma = 0.1721 \end{eqnarray} \tag{ 8 }$$

$$\begin{eqnarray} &&V = B_J + 0.0060 - 0.3982 (B_J - I_N)\quad \sigma = 0.1760 \end{eqnarray} \tag{ 9 }$$

$$\begin{eqnarray} &&V = 0.8077 + 0.9825 R_F\quad \sigma = 0.1828\qquad\qquad\quad\quad \end{eqnarray} \tag{ 10 }$$

$$\begin{eqnarray} &&V = 0.7665 + 0.9232 B_J\quad \sigma = 0.2775\qquad\qquad\quad\quad \end{eqnarray} \tag{ 11 }$$

$$\begin{eqnarray} &&V = 0.0875 + 1.0499 I_N\quad \sigma = 0.3780\qquad\qquad\quad\quad\ \end{eqnarray} \tag{ 12 }$$

$$\begin{eqnarray} R &=& R_F + 0.1380 - 0.2667 (R_F - I_N)\quad\sigma = 0.0976\quad\ \ \end{eqnarray} \tag{ 13 }$$

$$\begin{eqnarray} R &=& 1.2166 + 0.9286 R_F\quad \sigma = 0.1121\qquad\qquad\quad\quad \end{eqnarray} \tag{ 14 }$$

$$\begin{eqnarray} R &=& I_N - 0.0954 + 0.3353 (B_J - I_N)\quad \sigma = 0.1257\quad\ \ \ \end{eqnarray} \tag{ 15 }$$

$$\begin{eqnarray} R &=& R_F - 0.0208 + 0.0551 (B_J - R_F)\quad\sigma = 0.1264\quad\ \ \end{eqnarray} \tag{ 16 }$$

$$\begin{eqnarray} &&\ \ R = -0.2960 + 1.0441 I_N\quad \sigma = 0.2301\qquad\qquad\quad\quad \end{eqnarray} \tag{ 17 }$$

$$\begin{eqnarray} R &=& 1.5288 + 0.8546 B_J\quad \sigma = 0.3509\qquad\qquad\quad\quad \end{eqnarray} \tag{ 18 }$$

$$\begin{eqnarray} I &=& I_N - 0.1119 + 0.0931 (R_F - I_N)\quad\sigma = 0.0924\quad\ \ \end{eqnarray} \tag{ 19 }$$

$$\begin{eqnarray} I &=& I_N - 0.0709 + 0.0152 (B_J - I_N)\quad\sigma = 0.1073\quad\ \ \end{eqnarray} \tag{ 20 }$$

$$\begin{eqnarray} \ \ \ I &=& {-}0.1713 + 1.0079 I_N\quad \sigma = 0.1164\qquad\qquad\quad\quad \end{eqnarray} \tag{ 21 }$$

$$\begin{eqnarray} I &=& R_F - 0.2323 - 0.2276 (B_J - R_F)\quad\sigma = 0.2000\quad \end{eqnarray} \tag{ 22 }$$

$$\begin{eqnarray} I &=& 1.6690 + 0.8728 R_F\quad \sigma = 0.2244\qquad\qquad\quad\quad \end{eqnarray} \tag{ 23 }$$

$$\begin{eqnarray} I &=& 2.5445 + 0.7739 B_J\quad \sigma = 0.4543 \qquad\qquad\quad\quad \end{eqnarray} \tag{ 24 }$$

For any given CCD frame, the calibration equation we choose depends on which GSC bandpasses are available for that particular field; all else being equal, we select the transformation equation that has the smallest scatter. The images calibrated in this manner have the keyword ZPT_GSC in the header. The median photometric uncertainty (Table 2, Column 9) of the GSC-based magnitudes, which is dominated by the error in the above fitting equations, is 0.21, 0.11, 0.098, and 0.092 mag for the B, V, R, and I band, respectively.

Many aspects of the survey require that we exclude signal contributed by unrelated objects, such as foreground stars and background galaxies. This is achieved by creating a mask that properly identifies the affected pixels. We begin by running SExtractor to identify all the objects in the image and classify them into stellar and nonstellar sources based on their compactness relative to the seeing. Through experimentation, we find that a "class parameter" larger than 0.3 effectively isolates the stars from the resolved objects. During this initial step, we deliberately set the contrast parameter to a moderate value (we set the threshold to be 3 σ above the background) so that galaxies with substantial internal structure (e.g., spirals and irregulars) do not get broken up into multiple pieces. This, in turn, implies that SExtractor will miss fainter stars that might be superposed on the main body of the central galaxy. We carefully examine each image visually and manually add additional objects to the segmentation image if necessary. Next, two segmentation images are created, one to isolate the unsaturated stellar objects and the other the nonstellar objects, which include background galaxies and saturated stars (including their bleed trails, if present). Because we set the contrast parameter to a conservative level, the segmentation images identify only the bright core regions of the objects to be masked and often miss their fainter outer halos, which can be substantial, especially for bright stars in the I band. To account for these regions, we need to "grow" the segmentation image of each object. From trial and error, we find that a growth of 8 pixels in radius is optimal for the stellar objects. For the nonstellar objects, we distinguish between two regimes. A growth radius of 5 pixels is sufficient for bright bleed trails. For the fainter, more extended halos, we find that the growth radius can be approximated by $R={\rm min}[15, 1.5 \ \sqrt{N/\pi }]$, where N is the number of pixels contained in the original segmentation image. Lastly, we stack together the masks of each individual filter to create a master mask for each galaxy, which is then used in all subsequent analysis, to ensure that all filters reference the same set of pixels.

We use the IRAF task psf within the daophot package and the bright, unsaturated stars identified by SExtractor in the previous step to build an empirical PSF for each image. The PSF image is a key ingredient in much of our analysis, including generation of the structure maps (Section 5) and bulge-to-disk decomposition (S. Huang et al. 2012, in preparation).

The bulk of the CGS observations have fairly good image quality. The seeing, as estimated from the full width at half-maximum (FWHM) of the radial profiles of bright, unsaturated stars, ranges between ∼05, the limit we can measure given the pixel scale of our detector, to ∼2'', with a median value of 117, 111, 101, and 096 in the B, V, R, and I bands, respectively (Figure 4). The main factors that limit the surface brightness sensitivity of the images stem from residual errors in flat fielding and uncertainties in sky determination. We have discovered that large variations in the ambient temperature and humidity throughout the night can occasionally produce water condensation on the surfaces of some of the filters. This effect imprints large-scale flat-fielding errors at the level of ∼1%. Background gradients of a roughly similar magnitude are sometimes induced by scattered light from very bright stars, either within or just outside the science frame. These anomalies can be mitigated by fitting a two-dimensional polynomial function to the background pixels (Noordermeer & van der Hulst 2007; Barazza et al. 2008; Erwin et al. 2008). A second-order polynomial function usually suffices, but in some cases the order of the polynomial may need to be as high as 5. The fitting regions are source-free pixels far from the main galaxy, selected from the image after convolving it with a Gaussian with FWHM ≈ 4''–5'' to accentuate low-spatial frequency features. This correction typically improves the flatness of the images by about a factor of two, from ∼1% to ∼0.6%. Once the background is flat, determining its value is relatively straightforward because most of the survey galaxies have sizes (median D₂₅ = 33; Figure 2(d)) that fit comfortably within the field of view of the detector (89 × 89). The larger galaxies (D₂₅ ≳ 5'–6'; ∼15% of the sample), however, pose a challenge, and we are forced to implement an indirect, less reliable strategy for sky subtraction (Paper II). The surface brightness depth of the images (Table 2, Column 10), which we define to be the isophotal intensity that is 1 σ above the sky rms, has a median value of μ ≈ 27.5, 26.9, 26.4, and 25.3 mag arcsec⁻² in the B, V, R, and I bands, respectively.

Zoom In Zoom Out Reset image size

Paper II presents the brightness profiles and isophotal parameters of the survey. A number of basic, but nonetheless useful, global parameters for the galaxies can be derived from those data. We summarize them here. The advantage of our measurements is that they are derived in a uniform, self-consistent manner.

Table 3 lists several size measurements derived from the B-band images. We choose this bandpass because it closely matches most published measurements in the literature. The quantities R₂₀, R₅₀, and R₈₀ are the radii enclosing, respectively, 20%, 50%, and 80% of the total flux, which is calculated from integrating the surface brightness profile generated using the IRAF task ellipse (see Paper II). We define the concentration parameter as C = R₈₀/R₂₀. The isophotal diameters, measured at a surface brightness level of μ_B = 25.0 and 26.5 mag arcsec⁻², are given as D₂₅ and D_26.5, while D^c₂₅ and D^c_26.5 are the corresponding values corrected for Galactic extinction (A_B in Table 1, taken from Schlegel et al. 1998) and inclination effect, following the prescription of Bottinelli et al. (1995). For galaxies with morphological types T > −3.5, we derive the inclination angle of the disk, i, using Hubble's (1926) formula

$$\begin{equation} {\rm cos}^2 i = {{(1-e)^2 - q_0^2}\over {1-q_0^2}}, \end{equation} \tag{ 26 }$$

which makes use of the apparent ellipticity (e) of the outer regions of the disk. The intrinsic flattening of the disk depends mildly on morphological type (Paturel et al. 1997), but for simplicity we adopt a constant value of q₀ = 0.2 (Noordermeer & van der Hulst 2007). The position angle (P.A.) of the photometric major axis, east of north, is also given. Unlike the other parameters in this table, both e and P.A. are measured from the I-band image rather than from the B-band image in order to minimize possible distortions from dust or patchy regions of star formation. These quantities are determined from the outer regions of the galaxy—far enough to be insensitive to the bulge and bar but not so much so to be affected by lopsidedness in the outer disk or imperfect flat fielding—where they usually converge to a constant value. The listed uncertainties represent the standard deviation about the mean. In cases where e and P.A. do not converge, we estimate them visually from the best-fitting isophotes at large radii. Lastly, we tabulate Σ, the sum of the rms fluctuations in the structure map of the B-band image, which gives a useful relative measure of the degree of high-spatial frequency fluctuations present in the image.

Table 4 presents BVRI integrated magnitudes derived from the star-cleaned images (Section 5). Two measurements are made: m₂₅ pertains to the total apparent magnitude within the μ = 25 mag arcsec⁻² isophote, whereas m_tot derives from the total flux within the largest isophote fitted to the image. The main uncertainty of the magnitudes comes from the uncertainty of the photometry zero point, with little contribution from sky error or Poisson noise. For convenience, we provide the absolute magnitudes in the B band, corrected for Galactic extinction as given in Table 1 and adopting the luminosity distances therein.

Table 4. Magnitudes Derived from CGS

Name	B25	Btot	V25	Vtot	R25	Rtot	I25	Itot	$M_{B_{25}}$	$M_{B_{\rm tot}}$
	(mag)	(mag)	(mag)	(mag)	(mag)	(mag)	(mag)	(mag)	(mag)	(mag)
(1)	(2)	(3)	(4)	(5)	(6)	(7)	(8)	(9)	(10)	(11)
ESO 009-G010	13.08 ± 0.03	12.99 ± 0.03	12.29 ± 0.02	12.23 ± 0.02	11.64 ± 0.02	11.61 ± 0.02	10.84 ± 0.03	10.82 ± 0.03	−20.18 ± 0.03	−20.27 ± 0.03
ESO 027-G001	12.67 ± 0.21	12.53 ± 0.21	12.02 ± 0.11	11.92 ± 0.11	11.48 ± 0.10	11.41 ± 0.10	10.77 ± 0.09	10.74 ± 0.09	−19.50 ± 0.21	−19.64 ± 0.21
ESO 027-G008	13.29 ± 0.21	13.22 ± 0.21	12.44 ± 0.11	12.41 ± 0.11	11.73 ± 0.10	11.72 ± 0.10	11.02 ± 0.09	11.01 ± 0.09	−19.61 ± 0.21	−19.68 ± 0.21
ESO 056-G115	...	...	...	...	...	...	...	...	...	...
ESO 060-G019	13.10 ± 0.21	12.87 ± 0.21	12.58 ± 0.11	12.43 ± 0.11	12.03 ± 0.10	11.94 ± 0.10	11.27 ± 0.09	11.21 ± 0.09	−18.67 ± 0.21	−18.90 ± 0.21
ESO 091-G003	13.13 ± 0.03	13.01 ± 0.03	12.19 ± 0.02	12.15 ± 0.02	11.48 ± 0.01	11.45 ± 0.01	10.52 ± 0.03	10.80 ± 0.03	−19.43 ± 0.03	−19.55 ± 0.03
ESO 097-G013	...	...	...	...	...	...	...	...	...	...
ESO 121-G006	13.47 ± 0.09	13.22 ± 0.09	12.73 ± 0.04	12.61 ± 0.04	12.05 ± 0.04	11.98 ± 0.04	11.16 ± 0.05	11.12 ± 0.05	−18.16 ± 0.09	−18.41 ± 0.09
ESO 121-G026	12.65 ± 0.09	12.51 ± 0.09	12.01 ± 0.04	11.94 ± 0.04	11.42 ± 0.04	11.39 ± 0.04	10.74 ± 0.05	10.73 ± 0.05	−20.59 ± 0.09	−20.73 ± 0.09
ESO 136-G012	14.15 ± 0.30	13.81 ± 0.30	13.46 ± 0.17	13.33 ± 0.17	12.89 ± 0.10	12.79 ± 0.10	11.95 ± 0.09	11.83 ± 0.09	−19.92 ± 0.30	−20.26 ± 0.30

Notes. Column 1: galaxy name. Columns 2, 4, 6, and 8: total apparent magnitudes within the 25 mag arcsec⁻² isophote. Columns 3, 5, 7, and 9: total apparent magnitudes within the last measured point. Columns 10 and 11: absolute magnitudes in the B band, corrected for Galactic extinction as listed in Table 1.

Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.

Download table as:  DataTypeset image

Table 5 lists μ_e, the surface brightness measured at the effective or half-light radius R₅₀, for each of the four filters.

Table 6 assembles multiwavelength photometric data from several large galaxy catalogs. The listed values represent total integrated fluxes for the entire galaxy. Instead of collecting data from any available literature source, we restrict our selection to only a few well-documented, homogeneous compilations. For the ultraviolet, we select FUV (1350–1750 Å) and NUV (1750–2800 Å) measurements principally from the Galaxy Evolution Explorer (GALEX) atlases of Buat et al. (2007), Dale et al. (2007), and Gil de Paz et al. (2007). At present, only ∼16% of our sample has published GALEX data, although the situation will soon improve substantially with the recent completion of the All-Sky Imaging Survey.¹² To extend the optical coverage blueward of the B band, we searched HyperLeda for U-band (3500 Å) photometry, which is available for 63% of the sample. The most complete coverage is in the near-infrared (J, H, and K_s bands) and far-infrared (12, 25, 60, and 100 μm); ∼95% of the galaxies are included in the 2MASS (Skrutskie et al. 2006) and in the source catalogs of the Infrared Astronomical Satellite (IRAS). Because our galaxies are angularly large, care is taken to choose photometry derived from studies that properly treat extended sources (Jarrett et al. 2003; Sanders et al. 2003). In addition to the flux densities in the individual IRAS bands, we also give the parameter FIR, defined as 1.26×10⁻¹⁴ (2.58S₆₀ + S₁₀₀) W m⁻², which approximates well the total flux between 42.5 and 122.5 μm (Helou et al. 1988; Rice et al. 1988). To date, ∼70% (406/605) of our sample have been observed at 3.6 and 4.5 μm using the Infrared Array Camera on Spitzer. Roughly 60% of our sample overlaps with the Spitzer Survey of Stellar Structure in Galaxies (S⁴G; Sheth et al. 2010). We intend to incorporate these mid-infrared data into CGS in the future.

Three categories of data are included in Table 7. We assemble from HyperLeda information on the internal kinematics of the galaxies, namely the apparent maximum rotation velocity of the gas (V_max), the maximum rotation velocity corrected for inclination (V_rot), and the central stellar velocity dispersion (σ_⋆).

Next, we provide three environmental indicators. Using the facilities in the NASA/IPAC Extragalactic Database (NED),¹³ we search for candidate neighbors within a radius of 750 kpc. (NED restricts environmental searches to a maximum radius of 600', and, for a handful of objects, this corresponds to less than 750 kpc.) We calculate the projected angular separation, Δθ, in units of the diameter D₂₅ (Table 3) to the nearest neighboring galaxy having an apparent magnitude brighter than B_T + 1.5 mag and a systemic velocity within v_h ± 500 km s⁻¹. These criteria select sizable objects, with luminosity ratios of at least 4:1, which are likely to be physically associated companions. When no neighbor that satisfies the above criteria is found, Δθ is given as a lower limit. Following Bournaud et al. (2005), we also calculate the tidal parameter

$$\begin{equation} t_p \equiv \log \left\lbrace \sum _{i} \frac{M_i}{M_0} \left(\frac{R_0}{D_i}\right)^{3} \right\rbrace, \end{equation} \tag{ 27 }$$

where M₀ and R₀ are the mass and size of the galaxy in question, M_i and D_i are the mass and projected separation of neighbor i, and the summation is performed over all neighbors with systemic velocities within v_h ± 500 km s⁻¹ in a region with a radius of 750 kpc. For simplicity, we assume that all galaxies have the same mass-to-light ratio and adopt the Johnson B band as the reference filter. To the extent possible, we convert photometry listed in other bandpasses to the B band, assuming colors typical of an Sbc galaxy (Fukugita et al. 1995; Peletier & de Grijs 1998). An uncertainty of up to 0.3 mag, which encompasses the extreme range of plausible K corrections for different Hubble types, introduces an error of 0.12 dex in t_p. We set R₀ = 0.5D₂₅. Lastly, we indicate whether the galaxy belongs to the field or to a known galaxy group or cluster. The main cluster in the southern sky pertinent to CGS is Fornax, and we draw our membership identifications from Ferguson (1989) and Ferguson & Sandage (1990). Most of the group assignments come from the 2MASS-based group catalog of Crook et al. (2007).

The last two columns of Table 7 summarize the neutral hydrogen content. The H i (21 cm) flux, in magnitude units, is defined such that m^c₂₁ = −2.5log f + 17.40, where the integrated line flux f is in units of Jy km s⁻¹. We adopted the correction for self-absorption as given in HyperLeda. In the optically thin limit, the H i mass is given by

$$\begin{equation} M_{{\rm H\, \mathsc{i}}} = 2.36\times 10^5 \ D_{L}^2\ f \, M_\odot, \end{equation} \tag{ 28 }$$

where D_L is the luminosity distance expressed in megaparsecs (Roberts 1962). We list the H i mass normalized to the total B-band luminosity, using magnitudes from Table 1, corrected for Galactic extinction.

Figures 7.1–7.616 present the image atlas for the 605 galaxies in CGS; we also include the 11 extra galaxies that are not part of the formal sample. One full-page figure is devoted to each galaxy, ordered sequentially following the numerical indices listed in Table 1. The six panels of each figure show the composite color image, the star-cleaned composite color image, the BVRI stacked image, the structure map of the star-cleaned B-band image, and the B − R and R − I color index maps. Darker regions on the index maps correspond to redder colors. Each image is scaled to a dimension of 1.5 D₂₅, with north up and east to the left. We use an arcsinh stretch for the color composite, star-cleaned, and stacked images, while the structure and color index maps are shown on a linear stretch. We only display three sample pages for illustration (Figures 7(a)–7(c)); the full set of figures is available in the electronic version of the paper, as well as on the project Web site http://cgs.obs.carnegiescience.edu.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

View figure set (616 panels)

Tarball of figure set images

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size