II. Observations and Data Reduction

a) Equipment

All observations presented in this paper were made with an InSb detector system for which further details are given in Appendix B. The filters used are listed in Table 1, where the effective wavelengths and bandwidths shown were determined from filter scans provided by the manufacturer. The [2.36 µm] - [2.19 µm] color is used to measure the strength of the first overtone absorption bands of CO (bandhead at 2.29 µm). Similarly, the [2.01 µm] - [2.19 µm] color measures the strength of the H₂O absorption feature centered at 1.9 µm.

Table 1. Filters

Name	0(µ)*	(µ)
J	1.24	0.28
H	1.65	0.30
K	2.22	0.41
H2O	2.01	0.07
Continuum	2.19	0.14
CO	2.36	0.08
* Referred to 77° K.

In the course of this work three silicon field lenses optimized for telescope focal ratios of 7.5, 9.6, and 13.5 were employed. An important feature of the InSb dewar was the existence of an external focus mechanism which allowed focusing of the very steep field optics (f / ~ 1 in the most extreme case) with the dewar cold directly at the telescope. A typical example of the relatively flat beam profiles obtained is shown in Figure 1. (Experience proved that a focus change of only ~ 0.01 inch could produce significant degradation of the profiles.)

Figure 1. Typical beam profile scans of a star across the aperture are shown for the filters used in this study.

The beam size used was defined by an aperture wheel in the focal plane. The field lens, aperture wheel, filters, and detector were all cooled to pumped nitrogen temperatures (T ~ 60 K). The photometer employed had an offset guider and star-sky chopper consisting of a rotating sector wheel mirror near the focal plane. The telescopes, field lenses, chopper throws, and apertures used, along with the dates of the observations, are listed in Table 2. Chopping was always in the north-south direction. The physical aperture sizes employed were 4.0, 3.0, 2.0, 1.5, and 1.0 mm. However, due to spherical aberration of the plano-convex field lenses, the effective aperture sizes were typically 3-6% smaller. Only the effective aperture sizes, determined from the half power points of stellar scans (cf. Figure 1), are listed in Table 2.

b) Galaxy Selection

The galaxies observed in this study were selected so as to fulfill one or more of the following criteria: 1) They were bright and could thus be observed in a reasonable amount of integration time - the sensitivity of the InSb detector being such that an object of 9th magnitude at K observed with a 60 inch telescope required typically one hour of integration time to achieve a statistical accuracy of ~ 0.02 mag with the CO filter. 2) They would span a range of morphological type extending from early spiral through Magellanic irregular. 3) They would include objects for which broad band UBV photometry, and/or narrow band optical line indices, and/or detailed synthesis models were available from the literature.

The galaxies selected are listed in column 1 of Table 3. The morphological types, on the modified Hubble system of de Vaucouleurs, are also listed in column l, and were obtained from the Second Reference Catalogue of Bright Galaxies (de Vaucouleurs, de Vaucouleurs, and Corwin 1976, hereafter RC2), with but two exceptions: The type for NGC 7316 (not classified in the RC2) is adopted from Huchra (1977a), and the type for NGC 5253 is taken from the original Reference Catalogue of Bright Galaxies (de Vaucouleurs and de Vaucouleurs 1964, hereafter RC1), where the galaxy is classified as Im, rather than from the RC2, where it is classified as I0. The former classification is considerably more consistent with this galaxy's broad-band colors.

NGC 224 - 53.3" measurement is mean of 53.4" and 53.2" observations. For 47.5" measurement, reference beam correction = 0.37 mag.

NGC 520 - Aperture centered on bright central lump of this morphologically peculiar galaxy, 168" east and 120" north of bright field star to southwest.

NGC 598 - Reference beam correction = 0.18 mag.

NGC1068 - Seyfert galaxy.

NGC1566 - Seyfert galaxy.

NGC2841 - For H₂O index, large aperture measurement is at 105".

NGC2903 - For H₂O index, large aperture measurement is at 105". For 36.9" (47.5") measurement, reference beam correction = 0.24. (0.33) mag.

NGC2976 - Small aperture measurement centered 19" east and 51" north, and large aperture measurement centered 48" east and 25" north, of faint star on southwest edge of galaxy. For 41.1" measurement, reference beam correction = 0.16 mag.

NGC3034 - Centered on infrared peak.

NGC3556 - Aperture centered 18" east and 3" south of superimposed star, which is included in the field. A 13.7" aperture measurement centered on the star gives K = 10.57 mag, J - H = 0.40 mag, and H - K = 0.09 mag.

NGC4258 - For 41.1" measurement, reference beam correction = 0.17 mag.

NGC4449 - Aperture centered on stellar-like nucleus, which is displaced somewhat from geometric center of the overall structure.

NGC4569 - Star projected on nucleus is included in measurement.

NGC4631 - Aperture centered on infrared peak, which is 33" east and 30" south of faint star on northern edge of galaxy. See Aaronson (1977) for further discussion.

NGC5128 - Aperture centered on infrared peak.

NGC5253 - For C60 measurements, photometer was rotated to remove bright star from reference beam.

NGC5457 - Reference beam correction = 0.21 mag.

NGC5746 - For 66.4" measurement, reference beam correction = 0.21 mag.

NGC6814 - Seyfert galaxy.

NGC6946 - For 105" measurement, reference beam correction = 0.17 mag, due to bright star in beam.

NGC7316 - Markarian 307.

NGC7465 - Markarian 313.

NGC7620 - Markarian 321.

The sample in Table 3 includes galaxies in the Local Group, the Virgo cluster, and the general field; galaxies with absolute magnitude M_V primarily between -19 and -23; and galaxies spanning a wide range of inclination angle, from face-on to edge-on. Aside from the criteria discussed above, the sample does not appear to be overly biased in any obvious fashion.

c) Observational Procedure

All galaxy measurements were made only on nights of high photometric quality while guiding on bright stars in nearby offset fields. The focal plane aperture was always centered on the position of peak infrared brightness: the galaxy was first centered visually, and this centering was then checked by maximizing either the 1.6 µm or 2.2 µm signal. A further check on the position observed for galaxies with either very diffuse centers or thick absorption lanes was made by measuring offsets to nearby field stars, and comparing these offsets with published photographs of the galaxy (Sandage 1961; RC2 and references therein). For all but 6 of the 91 galaxies listed in Table 2, there was clear correspondence between the infrared peak and either the most optically prominent part of the galaxy (i.e., the nucleus), or in the case of a number of edge-on spirals with thick dust lanes, the physical centroid of the spherical component. The exceptions included two well known galaxies with strong imbedded IR sources, NGC 3034 and 5128; three galaxies, NGC 520, 2976, and 3556, having no apparent well-defined intensity peaks - in either the optical or infrared; and NGC 4631, an edge-on spiral where the position of peak infrared intensity lies not on the most optically prominent part of the galaxy, but in the nearby area of strongest dust absorption. The location of this peak is interpreted as the true position of the nucleus (see Aaronson 1977).

d) Photometric Errors and Instrumental Corrections

The adopted photometric errors were a combination of the statistical accuracy of a particular integration, and the nightly photometric quality, as measured by the dispersion of standard star residuals. Typical errors were 0.01 - 0.03 mag. Possible non-linear response of the detector was checked for by comparing measurements of Lyr relative to much fainter standards on both the Kitt Peak 2.1-m and 0.9-m, but none was found to the 0.01 mag level. However, several sources of systematic error were identified and corrected for:

First, because galaxies are extended objects it is necessary to correct for flux in the reference beam. For a linear magnitude growth curve and a spherically symmetric galaxy, the size of this correction at a given value of projected aperture size is determined only by the ratio of beam throw to beam diameter. The corrections adopted are illustrated in Figure 2, parameterized by this ratio. The curves shown were calculated using a three-part growth curve: For log A / D(0) < - 1.3, where A is the aperture size and D(0) is the corrected face-on diameter from the RC1, ⁵ the V growth curve for NGC 224 obtained from published optical data (Section IIe) was used. For -1.3 log A / D(0) 0.1, the K growth curve for elliptical galaxies determined in Paper I was employed. For log A / D(0) > 0.1, the growth curve given by Sandage (1975) for giant ellipticals was adopted. As indicated by Figure 2 and Table 2, the typical reference beam correction ranged from 0.03 - 0.07 mag. The curve labeled 2' in Figure 2 was obtained by rotating the final growth curve about the point at log A / D(0) = 0.0 so that the magnitude change at log A / D(0) = - 1.0 was decreased by 0.20 mag. It illustrates that the chopping throws were sufficiently large so that even for significant systematic differences between galaxies in either growth curve or color-aperture relations, the adopted beam corrections introduce an error of only 0.02 mag.

Figure 2. The adopted reference beam corrections are shown as a function of log A / D(0), the ratio of aperture size to corrected face-on diameter from de Vaucouleurs and de Vaucouleurs (1964). Numbers labelling the curves refer to the ratio of beam throw to beam diameter. The curve labelled 2' is explained in the text.

While the growth curves for spiral and elliptical galaxies are similar over the relevant range of aperture sizes and beam throws (RC2), we do not expect the corrections in Figure 2 to be appropriate for an inclined spiral with position angle (P.A.) oriented east-west, if we are chopping north-south. As a check on this, and also on the real adequacy of Figure 2 for spirals, a number of measurements of flux in the reference beam itself were taken. (This was not done generally because of the time-consuming nature of the observations.) For 45° P.A. 135°, the difference between the calculated and observed correction was found to be 0.01 ± 0.015 (24 measurements), while for P.A. < 45° or P.A. > 135°, the result was 0.035 ± 0.015 (23 measurements). We have thus adopted the following simple prescription: For spirals with P.A. < 45° or P.A. > 135°, or for all spirals with ill-defined P.A.'s because of small inclination angle, the calculated correction (Figure 2) was used directly; but for spirals with 45° P.A. 135° , the calculated correction minus 0.025 mag was used. Position angles were adopted either from Danver (1942), or measured directly from Palomar Observatory Sky Survey or ESO Quick Blue Survey photographs.

For a small number of measurements the above procedure was clearly inadequate, due either to the presence of a bright field star in the reference beam, or to anomalous light distribution of the galaxy itself. In these cases (noted in Table 3), the measured reference beam flux at the telescope was the correction used.

Corrections were also applied to account for the rounded nature of, and systematic differences between, the beam profiles of the various filters (Figure 1). Owing to the wavelength-dependent transmission characteristics of the field lens, scans of the H and especially the J filter tended to be more peaked in the center than K filter scans. These effects were accounted for by convolving digitized beam scans with the ellipsoidal luminosity law of de Vaucouleurs (RC1). The corrections ranged from 0.0 - 0.02 mag for the K magnitude and H-K color, and from 0.02 - 0.05 mag for the J-H color. No such correction was required for the CO index because of the similarity between the CO and continuum filter scans. However, problems in mounting the H₂O filter resulted in a somewhat altered appearance of the H₂O beam scans (Figure 1). This effect, corrected for as with the broad-band colors, ranged from 0.0 - 0.03 mag. We believe that the uncertainty in these adopted corrections, due to reasonable variation in either the digitized scan or the galaxy surface brightness profile, is ~ 0.01 mag.

Finally, when multiaperture measurements were made, field lens aberration again resulted in aperture dependent corrections. For example, changing from a 4 mm to a 2 mm aperture, the typical correction was 0.02 - 0.03 mag for the K magnitude, and ~ 0.01 mag for the narrow band indices. A 0.03 mag correction was also applied to all J-H colors measured with a 4 mm aperture using the f/7.5 field lens, as in this case only, the effective aperture at J was found to be ~ 0.1 mm smaller than at H or K. We stress that the estimated uncertainties in the sum of all instrumental corrections are only about 0.01 mag for the narrow band indices, and about 0.02 mag for the broad-band magnitudes and colors.

The K magnitude, J-H and H-K colors, and CO and H₂O indices, corrected for all instrumental effects, are listed in columns 6-10, respectively, in Table 2. Preceeding entries in the table are: column 2 - galactic latitude and redshift taken from the RC2; column 3 - the telescope used, coded as in Table 2; column 4 - the measured projected aperture in arc sec; and column 5 - the projected aperture in units of D₀. Unless noted otherwise in Table 3, the adopted nominal errors are 0.03, 0.03, 0.04, 0.02, and 0.02 mag at K, J - H, H - K; CO, and H₂O, respectively. About half of the measurements were repeated on at least two nights, and the dispersion in these repeated values, after application of instrumental corrections, are consistent with the nominal errors.

⁵ In this paper D(0) is the corrected face-on diameter taken from the RC1, and D₀ is the same quantity but taken instead from the RC2. Note that log A / D₀ = log A / D(0) + x, where 0.0 < x < 0.3, depending on morphological type (RC2). Back.

e) The UVK Colors

To form V-K colors a literature search was made of all published UBV data; some unpublished data was also kindly made available by A. Sandage and by J. Huchra. For each galaxy, the V magnitude and U - V colors were plotted as a function of aperture size, and after fitting a smooth curve to the points, values were read off at the apertures corresponding to the infrared measurements. In cases where the optical and infrared aperture ranges did not overlap, no V - K color was formed if extrapolation of data greater than 0.2 in log A / D₀, was required, and in only four instances was there extrapolation greater than 0.1 in log A / D₀. The results are presented in Table 4, where UVK colors for 76 galaxies are listed, along with the sources from which the optical data were gathered. Note that for 12 galaxies with only sparse optical data, K magnitudes were interpolated or extrapolated to the optical apertures For these objects there is no correspondence between the values of log A / D₀ in column 2 of Table 4 and column 5 of Table 3.

Since many times considerable scatter was present in the optical measurements, some amount of subjective judgement was necessarily used in forming the UVK colors. The primary sources of data were those available from the work of de Vaucouleurs, Huchra, Sandage and Tifft (cf. Table 4). These data points were given twice the weight as those from the remaining secondary sources, and for all but four galaxies in Table 4 there was at least one measurement available from a primary source. All data from Tifft was transformed to the UBV system using the relations given in Tifft (1973); the agreement between this transformed data and that from the other primary sources was generally acceptable (i.e., within 0.10 mag). As in Paper I, we adopt a nominal error of ± 0.10 mag in the V - K color, unless the error in the K magnitude or the scatter in the V data is exceptionally large, in which case a colon is given after the V - K color and a nominal error of 0.20 is assumed.

f) Reddening and Redshift Corrections

To correct for the effects of interstellar reddening, we have used the absorption free polar-cap model of Sandage (1973) and Van de Hulst curve #15 (Johnson 1968). The adopted corrections are summarized on the left half of Table 5. The values for the color excess ratios E_H2O / A_V and E_CO / A_V were obtained by careful interpolation of the reddening curve at the effective wavelengths listed in Table 1. Note that reddening causes an increase in the measured H₂O index, but a decrease in the measured CO index.

Table 5. Adopted Corrections for Reddening and Redshift ^*

Reddening	Redshift.
AV = 0.10(csc[b] - 1), |b| 50°	KK = - 3.25 z
AV = 0.,                        |b| > 50°
EU-V / AV = 0.56	KU-V = 1.2 z,                    T 0†
	= (1.2 + 0.7 T) z,     0 T 3
EV-K / AV = 0.91	= 3.3 z,                    T 3
EJ-H / AV = 0.10	KV-K = 4.9 z,                     T 0
	= (4.9 - 0.45 T) z,     0 T 3
EH-K / AV = 0.06	= (3.55 - 0.3[T - 3]), T 3
EH2O / AV = 0.014	KJ-H = 0.5z
ECO / AV = - 0.010	KH-K = 3.6 z
	KH2O = 0.
	KCO = - 4.8z
*m0 = m - Am - Km, C0 = C - Ec - Kc
†T is a morphological type parameter following the notation in the RC2 ; T = - 5 for type E, T = 0 for type S0/a, T = 5 for type Sc, etc.

The adopted redshift corrections are listed on the right side of Table 5. The redshift corrections for the optical colors are taken from Pence (1976), and given as a function of morphological type parameter T following the notation in the RC2, where T = - 5 is for type E, T = 0 for type S0/a, T = 5 for type Sc, etc. The redshift corrections for the infrared colors are discussed in and adopted from Papers I and II. Although these corrections were based on measurements of elliptical galaxies, we have used them for spirals because, as we show below, the infrared colors depend very weakly on morphological type.

Only nine galaxies in Table 3 have galactic latitude less than 20°, the galaxy with the smallest value being NGC 6946, with b^II = 12° . No galaxies in Table 3 for which CO and H₂O indices were observed, and only four galaxies for which JHK colors were measured, have redshift z > 0.01, the largest value being that for NGC 7520, with z = 0.032. Thus, uncertainties in either the redshift or reddening corrections will not affect the conclusions of this paper.

g) Comparison with Previous Results

The majority of observations in Papers I and II were made with a completely independent set of filters and detector (the CIT system). The J - H and H - K transformations between that system and the present one (the HCO system) are discussed in Appendix A. No transformation is necessary between the systems for the K magnitudes, or the CO and H₂O indices. For comparison purposes a number of galaxies were measured in common with the two systems using the same or similar apertures. The results, after application of all instrumental effects and transformations, are summarized in Table 6. We conclude from this Table that no statistically significant calibration differences exist between the data of Papers I and II and the present paper.

Table 6. Comparison of Galaxy Observations on the HCO and CIT Systems

Number of Observations	Feature	Mean *
3	K	-0.013 ± 0.014
6	J-H	-0.007 ± 0.009
6	H-K	-0.010 ± 0.010
9	CO	0.002 ± 0.004
3	H2O	0.005 ± 0.005
* (color)HCO - (color)CTT, after transformation of the CIT data to the HCO system and application of all instrumental corrections.