The NIR is sensitive to light coming from the older stellar population and from red supergiants associated with massive star formation. Stellar light reveals the overall system morphology to be generally disturbed. The nuclei of UGC 12914/5 dominate the total near-IR emission, with prominent "red" dust lanes evident in the brighter nucleus, UGC 12915 (northernmost nucleus; Fig 3) and an H II knot ~30´´ southwest of the nucleus, spatially coincident with H-alpha line emission. A tidal-warped spiral arm of UGC 12915 is seen to the northwest of the nucleus following the radio continuum. The disk of UGC 12914 is characterized by a bright nucleus, dust lanes extending 20-30´´ from the nucleus and bright "warp" features anchoring the north-south ends of the disk. A spiral arm extending above the northern "warp" feature appears to be tidally stretched by the interaction with UGC 12915. Moreover, a large-scale ring or branch of a spiral arm encircles UGC 12914, with the brightest section passing through the southwestern bridge region (see Fig 3b). In spite of the ring presence, most of the connecting bridge region is devoid of near-infrared light. A summary of the global photometry for the system is given in Table 1 and the corresponding colors (log of the integrated flux density ratios between J & K and H & K) in Table 2.

The final coadd (average of four data sets) MIR images for the ISOCAM bands, LW1(4.5 µm ), LW2 (6.75 µm), LW8(11.4 µm), and LW3 (15.0 µm), are shown in Fig 5. Radio continuum contours are included with the LW3 grey-scale image. The total effective field of view is ~3´, with 2´´ pixels and an effective resolution of ~9´´. The nucleus of UGC 12915 is the brightest object in the field, approximately 50-60% brighter than the combined nucleus and adjacent knots comprising the disk of UGC 12914; see Tables 1 & 2 for global photometry and colors, respectively. The average 5-17 µm (LW2+LW8+LW3) flux density of the system, UGC 12914/5 + Bridge, is ~0.42 Jy (see Table 1), which is about 7% of the total IRAS 60 µm flux density (6.3 Jy) and 3% of the 100 µm flux density (13.4 Jy).

The UGC 12914/5 nuclei supply the bulk of the MIR emission; nonetheless, the "bridge" is easily seen in the LW2, LW8 and LW3 images. The bridge structure is more apparent when combining all three MIR bands into one color image, with LW2 = blue, LW8 = green and LW3 = red (Fig 6). The source of the MIR emission coming from the bridge region is associated with "Aromatic Features," the so-called Unidentified Infrared Bands, or UIB (Leger & Puget 1984; Desert et al. 1990; Schutte et al. 1998; Helou et al. 1999), commonly thought to be due to polycyclic aromatic hydrocarbon (PAH) molecules or tiny carbonaceous grains. The MIR emission peaks at 6.2 µm & 11.4 µm are spatially coincident with the 1.49 GHz radio bridge (see overlay contours, Fig 6b) and the H I line emission (contour overlay, Fig 2), corresponding to position ~00h01m39s, +23d29m30s (J2000). The integrated MIR flux of the bridge is between 10 and 40 times fainter than the nucleus of UGC 12915 (Table 1). There is no discernable stellar population in the bridge based on the LW1 image, which should trace stellar photosphere emission from late-type stars and photo-ionization from hard UV photons via the hydrogen recombination line Br-alpha at 4.1 µm. Warm dust emitting from the bridge region and abutting large-scale ring can be clearly seen in the LW2 and LW3 images (Fig 5 & 6). The flux ratio between LW1 and LW3 for the bridge is a factor of three smaller than that of the disk of UGC 12914 (see Table 2), also consistent with a paucity of 4 µm emission in the bridge. As seen in the NIR, the ring or warped spiral arm passes through (in projection) the connecting bridge. This morphology is more clearly seen in (Fig 7) which overlays contours of NIR K_s-band on gray-scales of MIR LW2 & LW3 images. The NIR light is presumably originating from an older population of stars, which would be consistent with the ring structure projecting against or belonging to a spiral arm of UGC 12914.

Cross-section profiles are used to compare relative variations in the NIR to MIR bands and in the infrared to radio continuum ratio on small spatial scales. Three cuts are made, one passing through the disk of UGC 12914, one the disk of UGC 12915, and one orthogonal to the disks, passing through the connecting bridge and H II region southwest of UGC 12915 (see Fig 7). To increase the SNR we integrate a small circular area (6´´ in radius, equivalent to ~1.5 kpc) along the profile, with 2´´ steps between adjacent areas. We plot integrated flux density (in mJy) and effective "colors," log of the flux density ratio between bands. Flux ratios are only computed for points in which the integrated flux density has an estimated SNR > 5 in both bands being measured. Colors include: log[J/K], log[H/K], log[LW2/LW3], [LW8/LW3] and log [radio/LW3].

The radial profile and color plots are shown in Fig 8. The nucleus dominates the NIR surface brightness profile (with H-band the brightest), while the MIR is characterized by three nearly equally bright peaks separated ~20´´ (~6 kpc) from each other. It is evident that the NIR morphology is very different from that of the MIR and radio. The NIR is tracing photospheric emission from the old (pre-collision) stellar population while the MIR is tracing small interstellar dust grains.

The NIR colors indicate a general reddening (in J vs. K_s) trend moving from the southern end of the disk to the northwestern end, with a prominent drop in the J and H flux density relative to K_s at the point where the northwestern "warp" feature is observed. The reddening is probably selective extinction from dust in the ISM, but may also be light coming from luminous evolved bright stars. The MIR colors are mostly flat across the disk (to within the uncertainty in the ratio), while the radio-to-LW3 (15 µm) ratio looks very different compared to the infrared colors. The radio flux density is a factor of 5 to 10 times fainter than the MIR. There is a prominent rise in the radio continuum relative to LW3 along the northern side of the nucleus, with the color ratio peaking somewhere between the nucleus and the bright northern MIR knot (~3 to 5 kpc from the nucleus). This radio "peak" may be tracing non-thermal emission from cosmic rays diffusing outward from massive star formation regions (cf. Helou & Bicay 1993; Marsh & Helou 1998). The asymmetry between the northern and southern disk regions of UGC 12914 suggest that the ISM has been unequally affected by the recent passage of the interloping galaxy UGC 12915, which Condon et al. (1993) deduce to have passed through UGC 12914 just north of the nucleus.

The disk/nucleus of UGC 12915 is remarkably different from that of its companion, UGC 12914. First, both the NIR and MIR reveal a singular broad and relatively smooth nucleus, Fig 9. The NIR colors show a reddening trend from the south toward the center of the nuclear region, but with the reddest colors seen toward the northern side of the disk, ~12´´ (4 kpc) from the nucleus. In contrast, there is a prominent "blue" rise in the J flux density near the northwest section of the tidal "hook", probably associated with massive star formation (also revealed in H-alpha line emission). In the mid-infrared the flux density profile is more asymmetric compared to that of the near-infrared (Fig 9, panels 1 vs. 3). Moreover, the MIR emission peak appears to be shifted 1 to 2´´ southeast with respect to the NIR emission peak. The peak position difference may be due to the absolute pointing uncertainty in the ISO data.

The MIR colors exhibit even more dramatic behavior. The shortest bands, LW2 and LW8, significantly strengthen (relative to LW3) moving from the south, passing through the nucleus, and to the north, with colors reaching maximum 20´´ from the nucleus, again corresponding to the northwest tidal arm region, and roughly corresponding to the NIR color gradient. This behavior is most likely the result of strengthening in the PAH bands at 6.2 and 7.7 µm (convolved by LW2) and 11.4 µm (convolved by LW8), a compelling clue that a vigorous starburst is underway in the northern quadrant of the UGC 12915 disk. The radio continuum is consistent with this interpretation, showing a steep rise in the flux density relative to LW3 passing from the southeast to the northwest, pointing to a higher dose of non-thermal radio emission (associated with massive star formation) in these regions.

The radial profile and color plots are shown in Fig 10. To summarize previous findings: the connecting bridge between the interacting galaxies includes a bright HII region ~15´´ southwest of UGC 12915, prominent H I line emission, non-thermal synchrotron radiation and, to the south in projection, a large-scale ring of UGC 12914 (~30´´ northeast of UGC 12914). The HII region is not prominent in the NIR flux density profiles (compared to the radio, for example; Fig 10, panels 1 and 3) but does appear in the color profiles (panel 2) with a sharp "blue" rise in J/K and H/K toward the northern side of the HII complex. The J/K blue trend increases radially toward the ring, while H/K is mostly flat for the ring itself. Note however that the connecting bridge between the HII region and the ring is mostly devoid of any photospheric light at 1-2 µm (see also Table 1, Fig 2) so the NIR flux ratios are dominated by noise fluctuations. In the mid-infrared and radio the HII region complex is much more extensive with a broad shoulder extending ~25´´ from the nucleus deep into the connecting bridge (Fig 10). The MIR emission reveals a hot dust component within the dividing medium. The [LW8/LW3] color is mostly flat through the bridge, excepting a sharp "red" depression south of the HII region toward the central bridge region (~30´´ from UGC 12915). The bridge colors then show a "blueward" trend moving southward into the ring area and toward the nucleus of UGC 12914. The ratio [LW2/LW3] also appears to be depressed between the HII complex and the ring (but as before, background noise may be overwhelming any real signal in the NIR). The bridge remains most prominent in the radio maps. The radio continuum peaks in a region ~25´´ in length, putatively the "central" bridge, while the MIR broadly encompasses the northeastern HII complex and the southwestern ring of UGC 12914. The net result is that the radio-to-LW3 color is strongly peaked in the central bridge, somewhat anti-correlated with MIR colors. The origin of the radio and mid-infrared emission in the bridge ISM is probably from two unrelated sources, as we will argue in º4.3.

The detailed composition of the MIR emission spectrum, shortward of 12 µm, coming from the nuclear regions is revealed with PHT-S spectroscopy (Fig 11). The Aromatic Features in emission are clearly seen at 6.2, 7.7 and 11.3 µm for the disk/nucleus of UGC 12915 (Fig 11a), dominating the total 6 to 12 µm band emission (corresponding to LW2, LW8 and partially, LW3). For comparison, a composite spectrum (Lu 1999, private communication; Helou et al. 1999) comprised of many quiescent (normal) spiral galaxies is overlaid in black (solid continuous line) to illustrate the aromatic emission features and underlying continuum shortward of 6 µm. The presence of these line emission features is strongly correlated with heating by stars (via star formation and also from older stellar populations) in normal galaxies and in infrared-luminous starburst galaxies (Boulanger 1998; Boselli et al. 1998).

The ISOPHOT integrated flux density between 5 and 8.4 µm is 210 mJy, dominated by the 6.2 and 7.7 µm emission lines. In comparison, the ISOCAM LW2 flux density is 183 mJy, as computed in a similarly sized area centered on the nucleus of UGC 12915, a difference of ~15%, indistinguishable within the measurement uncertainty. At the short end, the 2.5 to 4.5 µm spectrum is dominated by noise, while no Br-alpha recombination (at 4.1 µm) or PAH emission at 3.3 µm is discernable.

UGC 12914 is considerably fainter than its companion and consequently its spectrum (Fig 11b) is, for the most part, subjugated by background noise. Only the 6.2 and 7.7 µm lines rise from the noise and underlying continuum. The 7.7 µm line, with an integrated flux density of ~50 mJy +- 30%, is inexplicably blue-shifted with respect to the 6.2 µm line and the composite spectrum. Given the low SNR and the inconsistent line positions, the spectrum for UGC 12914 is evidently unreliable.

Infrared imaging reveals that the VV 254 system can be decomposed into three principal morphological components (with size ~10 kpc): two multi-peaked "disks" or "main disks," large-scale ring or tidal arm of UGC 12914, and connecting bridge. The energetics and general ISM environment are vastly different, with the disks subject to a vigorous starburst (UGC 12915) or normal star formation phase (UGC 12914), the ring/arms modified by an expanding shock wave from the collision, and the connecting bridge constrained by strong magnetic fields trapping electrons extracted in the encounter and channeling energetic cosmic rays from massive star formation regions. Given the relatively brief evolution from the initial encounter to the present state of the system, these regions are likely to evolve dramatically in the next ~10⁷ -10⁸ years.

For the VV 254 system, the surface brightness averaged on scales of ~3 kpc varies widely among emission bands, from optical (0.6 µm and H-alpha at 0.66 µm), NIR (1-2 µm), MIR (3 - 17 µm) and radio (20 cm). Moreover the two galaxies appear quite different as viewed in the MIR and radio. The short wavelength emission is significantly attenuated by dust absorption, rendering for example, H-alpha to UV radiation comparisons suspect in the core and disk of UGC 12915 and to a lesser extent in the disk of UGC 12914. In the MIR and radio, the nuclear region of UGC 12915 exhibits a strong peak at the position of the nucleus and a broad shoulder extending to the north into the tidal "hook" structure, also roughly corresponding to H-alpha emission peaks ( Bushouse & Werner, 1990). Numerical simulations of galaxy-galaxy interactions show that star formation is enhanced in the post-collision disks due to molecular cloud collisions (cf. Olson & Kwan 1990), which is consistent with the observed H-alpha/MIR/radio spectrum of UGC 12915 and the total FIR flux of the system. Furthermore, the correlated radio-MIR morphology in UGC 12915 (see Fig 9) is typical of star forming disks, where massive star formation produces both the UV radiation field that transiently heats dust grains to emit in the MIR and the energetic cosmic rays (via supernovae) that generate synchrotron radiation as they spiral along the magnetic field lines (cf. Helou et al. 1985, and Condon 1992 review). In the southern galaxy, UGC 12914, MIR emission delineates three nearly equal bright peaks, including the nucleus, spanning the optical/NIR disk (~18 kpc) of UGC 12914, but with a total MIR and radio flux nearly a factor of two less than its northern companion. The radio is enhanced relative to the MIR along the northeast side of the disk (third peak), corresponding to the junction between the disk and tidal arm and ring structure of UGC 12914. The MIR emission peaks roughly correspond to enhanced H-alpha emission, one bright HII complex to the southeast and a string of HII regions extending to the northwest along the disk and warp feature (cf. Bushouse & Werner 1990). The MIR peaks appear to represent "hot spot" regions undergoing massive star formation probably induced or enhanced from the interaction. Collision-triggered starbursts are a common feature of interacting galaxies, and are a prime candidate for the engines that power ultra-luminous infrared galaxies (cf. Genzel, Lutz & Tacconi, 1998; Genzel et al. 1998). In the NIR, the peaks appear more linear and resemble an extended (several kpc) bar structure not unlike the barred spiral NGC 7552, and the NIR morphology of NGC 7469 (Mazzarella et al. 1994).

The overall level of star formation in UGC 12914 appears to be half of that in UGC 12915 based on the radio and MIR fluxes. The pre-collision molecular gas mass differences and encounter geometry can account for the post-interaction star formation activity. UGC 12915 is observed to have a higher H₂ mass fraction as deduced from molecular CO data (Min Yun, private communication), and is more compact than its companion (as measured from the stellar light), making UGC 12915 comparatively more conducive to star formation.

The PHOT-S 3-12 µm spectrum reveals the nature of the MIR radiation. The emission from the nuclear regions of UGC 12915 is dominated by Aromatic Features, a signature of star formation, but generally not associated with AGN phenomena (Genzel et al. 1998; Puget & Leger, 1989). Based on the estimated time-scale of the interaction ~few × 10⁷ yrs, deduced from the spectral-steepening of the radio spectrum index of the "post-interaction" zone compared to that of the disks, and the relative mass and velocity differential between the spirals (Condon et al. 1993), the UGC 12915 starburst is early in its life-cycle.

The formation of expanding rings (Lynds & Toomre 1976; Theys & Spiegel 1977; Appleton & James 1990; see also review from Barnes, 1992) in the VV 254 encounter is likely given the conditions that it satisfies, including a near head-on collision (impact parameter r_min < r_disk) between nuclei of comparable mass and counter-rotating disks. Condon et al (1993) deduce from their radio velocity maps that UGC 12915 penetrated the disk of UGC 12914 just north of its nucleus, inducing a ring structure in the larger UGC 12914 disk. A slightly off-center encounter will result in an "open" ring, offset from the nucleus. Both optical and NIR emission (Fig 1 - 3) reveal an elliptical ring-like feature enveloping UGC 12914. The large-scale ring extends some 1.8´ (30 kpc) across, with the brightest arcs of the ring projected across the connecting bridge field. There is some hint of massive star formation in the ring from the H-alpha line emission (Bushouse & Werner 1990). The MIR (5 to 17 µm) emission clearly outlines the rings in the bridge area and to the northeast of UGC 12914 coincident with a string of HII regions defining the northern section of the ring (Fig 6). The MIR ring emission is not spatially correlated with the radio continuum. The H I line map (Fig 2) does not clearly trace the ring, even though the H I gas extends well beyond the radio continuum and the faint optical light contours. Still, it may be no coincidence that the extreme west-northwestern edge of the H I column density map corresponds to the NW ring arc.

In summary, the ringlike structure appears to consist of stars, presumably from an old disk population of the pre-collision disk, and a few isolated pockets of massive star formation, with little or no evidence of enhanced atomic gas column density or a systemic starburst associated with the expanding shock wave. It may be that massive stars will form in the future or that the molecular gas density is insufficient to fuel star formation (cf. inner ring of the Cartwheel galaxy, Higdon 1996). It is also possible that the ring is in fact a tidal arm of UGC 12914. N-body simulations of the VV 254 system are needed to help decode the nature of the ring and spiral-arm structure.

The connecting bridge region of VV 254 is remarkable in that it contains a significant atomic gas column density (~10²¹ cm^-2), accounting for nearly one-quarter of the total H I mass of the system (Fig 2). Only a few nearby systems show significant gas fractions outside of their post-collision disks, including NGC 4814/20 (van de Hulst & Hummel 1985) and NGC 7714/5 (Smith et al. 1997). Moreover, the bridge is most conspicuous in the radio, dominated by synchrotron radiation from cosmic ray electrons trapped in the post-collision magnetic field stretching between the disks. The optical and NIR clearly delineate the H II region complex southwest of UGC 12915 (see also the H-alpha images in Bushouse and Werner 1990) and the ring projected against the southern half of the bridge, but do not show otherwise any evidence of a stellar population in the bridge (cf. Fig 1 - 3). Nonetheless, consistent with a "splash" bridge formation scenario (Struck 1977), the MIR emission indicates significant dust content in the bridge, including the gap between the H II region and the ring (Fig 6). There are two possible heating mechanisms for the dust detected in the broad band ISOCAM data: UV radiation from the galaxy disks, and collisional heating from the relativistic electrons in the bridge.

Given the steep radio spectrum of the bridge, the cosmic ray heating mechanism is likely to be negligible for the following reason. Cosmic ray electrons both ionize the atomic hydrogen and free-free scatter off ions, but the power lost per electron by ionization is independent of electron energy, so the fractional losses are greatest for low-energy electrons (Pacholczyk 1970; Condon 1992). The signature for low-energy electron losses is a flat radio spectrum (ionization losses reduce the spectral index), while free-free losses do not affect the spectrum. Conversely, synchrotron and inverse-Compton losses are both proportional to energy squared, leading to a steep radio spectrum. Hence, the observed steep radio spectrum suggests that the dominant energy-loss mechanism is probably synchrotron radiation. The radio flux, however, is much smaller than the larger-bandwidth mid-infrared flux, so the synchrotron power is less then the power needed to heat the dust (which re-radiates in the MIR). Since the cosmic-ray heating loss power is less than the synchrotron power, the cosmic rays cannot provide the observed MIR power. We conclude that it is more likely that the ambient UV radiation field is the heating source of the bridge-ISM dust.

To explore the radiative heating hypothesis, we need to address two key issues: (1) is there enough dust to account for the MIR emission and (2) is the local radiation field strong enough to power the dust emission? Thanks in large part to IRAS, the gas-to-dust correlation is well established for FIR emission originating from Milky Way "cirrus" clouds and from the local solar neighborhood (Boulanger & Perault 1988; Boulanger et al. 1996). Moreover, the FIR to atomic gas column density relation persists in nearby galaxies, particularly for the diffuse (optically thin) components. For the more complicated components, dense molecular clouds and H II region, the yield is more uncertain (cf. Walterbos & Greenwalt 1996). Most importantly, the gas-to-dust correlation appears to extend to shorter wavelengths, <12 µm, including regions in which Aromatic Feature emission dominates the MIR spectrum (cf. Desert et al 1990; Jenniskens & Desert 1993). This correlation can be used to discern the nature of the MIR emission in VV 254. The bridge is spatially far enough from the bright nuclear disks and giant H II region, ~6 to 7.5 kpc, that the radiation environment of the bridge medium is analogous to that of "cirrus" clouds above the disk of the Milky Way. If the intensity of the radiation field in the bridge is similar to that of the Galactic "cirrus" (both in total intensity and in the UV photon spectrum), then the MIR should similarly scale with the atomic gas column density. The bridge interstellar radiation field is probably composed of mostly non-ionizing photons originating from the disks and, to a lesser extent, OB star producing UV photons diffusing outward from H II regions. The detailed nature of the interstellar radiation field in spiral galaxies is still a much debated and open topic (cf. Devereux & Young 1990; Xu & Helou 1995; Walterbos & Greenwalt 1996; Devereux et al. 1997), but nevertheless is unlikely to alter our analysis. We do not assume what photons are actually heating the diffuse bridge dust, but instead assume that the heating source (ambient radiation field) is similar to what is observed in the solar neighborhood and upper Galactic "cirrus" clouds.

Consider the gas content of the bridge. The atomic hydrogen column density (see Fig 2 overlay), measured by Condon et al (1993), is ~3 × 10²¹ cm^-2 (18´´ FWHM beam) centered on the bridge. For comparison, the bridge MIR integrated flux densities (20´´ diameter aperture) are given in Table 1. Using the flux convention and formalism of Boulanger & Perault 1988, we convert the flux densities to intensity: 4I, normalize to the gas column density and compute the dust emissivity, or the MIR intensity per hydrogen atom (W/H atom). The emissivity reflects the expected MIR emission given an observed hydrogen column density, assuming that the 21-cm radio emission (from which the column density is derived) is optically thin and dust emission associated with molecular gas in the bridge is relatively minor (but see below). If the bridge radiation field intensity is similar to that of the local Solar neighborhood and in high Galactic latitude "cirrus" clouds, then the gas column density-normalized emissivity should be similar to that seen in the Milky Way. The VV 254 dust-to-gas MIR emission spectrum results are given in Table 3. It is instructive to compare our MIR measurements with IRAS 12-µm solar neighborhood values (as given in table 2 of Boulanger & Perault 1988) and Galactic "cirrus" ISOCAM LW2 and LW3 values (from Bill Reach, private communication). For LW8 (11.4 µm) and LW3 (15 µm), largely overlapping with the IRAS 12 µm band, the combined dust emission per H atom agrees to within a factor of two with that seen in Galactic "cirrus" clouds, while the narrow-band LW2 and LW3 values agree very well, 20-30%, with the reference values observed in the Milky Way. Given the location of the bridge with respect to the disks and nearest H II complexes, the system geometry and the comparable stellar populations, it is reasonable to assume that the VV 254 bridge UV radiation field is comparable to the Solar neighborhood value (to within a factor of ~2), and consequently the observed MIR (LW3) emission is consistent with the expected column of dust based on the observed H I column density.

At wavelengths shortward of the IRAS bands (and LW3), we consider the detailed emission-line work of Desert et al (1990). Using IRAS and ground-based NIR and sub-mm observations of the solar neighborhood and Galactic "cirrus," Desert et al (1990) modeled the expected dust emission spectrum for a range of heating intensities of the interstellar radiation field local to the Sun. For the NIR and MIR bands, they used a two-component dust model: very small grains (continuum) and PAHs (line emission), transiently heated from single photon absorption coming from the radiation field. Employing the flux unit convention of Desert et al (1990), Table 3 lists the I /N_H specific emission values for the VV 254 bridge, which can be directly compared to the emission spectrum of the Desert et al (1990) model from their table 3 and figure 4 (after converting to F). The bridge emissivity and specific emission, spanning across the 6 to 15 µm spectrum, agrees to within 10-50% with predicted continuum emission coming from very small grains and from PAH emission line features-a very good correspondence considering the relative strengths of the emission lines and probable difference in the interstellar radiation field between VV 254 and Galactic "cirrus" fields. Finally, at the shortest wavelengths, LW1 (4-5 µm), which has only a continuum spectrum, the VV 254 upper limit emissivity value is a factor of ~3 higher than the model-predicted value. We conclude that the intensity of dust emission coming from the VV 254 bridge region is consistent with the observed atomic gas column density.

A final word of caution: since molecular gas is observed in the bridge (traced via CO line emission), the atomic gas column density must therefore underestimate the total gas column density in the bridge, and some fraction of the MIR emission must be associated with the H₂ gas (cf. Desert et al 1990; Walterbos & Greenwalt 1996; Boulanger et al. 1996). The MIR to N_H ratio, therefore, represents an upper limit (if in fact small) to the true emissivity.

Fig 1-- VV 254 as seen in the optical and radio bands. The gray-scales correspond to the Digital Sky Survey (DSS) image overlaid with radio continuum contours. The 1.49 GHz total-intensity contours are from Condon et al. (1993). The contours range from 0.1 to 7 mJy/beam (5´´ FWHM), with successive contours scaled by sqrt(2).

Fig 2-- VV 254 as seen in the near-infrared. Clockwise: panels are gray-scale/contour overlays of J band (1.3 µm) , H band (1.6 µm) , and K_s band (2.2 µm). Lower left panel shows the K_s image overlaid with contours of atomic hydrogen column density, N_H (Condon et al. 1993). Density contours range from 3.4 to 27 × 10²¹ cm^-2.

Fig 3-- Near-infrared 3-band composite image of the VV 254 system. The J, H and K_s bands are RGB color-combined as follows: J (1.3 µm) == blue, H (1.6 µm) == green, and K_s (2.2 µm) == red . (a) Contrast to show high surface brightness features; (b) contrast to show low surface brightness emission. The contour overlay corresponds to 1.49 GHz radio continuum emission. Contours range 0.1 to 7 mJy/beam (5´´ FWHM), with successive contours multiplied by sqrt(2).

Fig 4-- VV 254 as seen in the far-infrared. Clockwise: panels are gray-scale/contour overlays of IRAS 12, 25, 60 and 100 mm coadd images, resolution enhanced with IPAC deconvolution algorithms. The integrated flux densities for the combined UGC 12914/5 system are 0.43, 0.88, 6.3 and 13.4 Jy at 12, 25, 60 and100 mm, respectively.

Fig 5-- VV 254 as seen in the mid-infrared. Clockwise: panels are gray-scale/contour overlays of ISOCAM LW1 (4.1 - 5.0 µm), LW2 (5.5 - 8.5 µm), LW8 (10.9 - 11.9 µm), and LW3 (12.2 - 17.0 µm). The lower left panel (LW 3 image) includes contours of the 1.49 GHz radio continuum, ranging from 0.1 to 7 mJy/beam (5´´ FWHM).

Fig 6-- Mid-infrared 3-band composite image of the VV 254 system. The LW2, LW8 and LW3 bands are RGB color-combined as follows: LW2 (6.8 µm) == blue, LW8 (11.4 µm) == green, and LW3 (15 µm) == red. (a) Contrast to show high surface brightness features; (b) contrast to show low surface brightness emission. The contour overlay corresponds to the 1.49 GHz radio continuum emission. Contours range 0.1 to 7 mJy/beam (5´´ FWHM), with successive contours multiplied by sqrt(2).

Fig 7-- Near-infrared and mid-infrared emission from VV 254. Panels are log-stretched gray-scale images of LW2 (5.5 - 8.5 µm) and LW3 (12.2 - 17.0 µm) overlaid with K band (2.2 µm) contours, ranging from 0.01 to 0.35 mJy/arcsec. Cross-section slice cuts (see Fig 8) through the disks of UGC 12914/5 and the connecting bridge region are denoted with thick solid lines.

Fig 8-- Cross-section flux density and color profile of the UGC 12914 disk. At each point along the cross-section (see Fig 7) the flux density is integrated in a 6´´ radius circular aperture. (a) Near-infrared flux density; (b) near-infrared colors, [H/K] and [J/K]; (c) mid-infrared/radio flux density; (d) MIR/radio colors, [LW2/LW3], [LW8/LW3] and [radio/LW3].

Fig 9-- Cross-section flux density and color profile of the UGC 12915 disk. At each point along the cross-section (see Fig 7) the flux density is integrated in a 6´´ radius circular aperture. (a) Near-infrared flux density; (b) near-infrared colors, [H/K] and [J/K]; (c) mid-infrared/radio flux density; (d) MIR/radio colors, [LW2/LW3], [LW8/LW3] and [radio/LW3].

Fig 10-- Cross-section flux density and color profile of the UGC 12914/5 disks plus intergalactic bridge. At each point along the cross-section (see Fig 7) the flux density is integrated in a 6´´ radius circular aperture. (a) Near-infrared flux density; (b) near-infrared colors, [H/K] and [J/K]; (c) mid-infrared/radio flux density; (d) MIR/radio colors, [LW2/LW3], [LW8/LW3] and [radio/LW3].

Fig 11-- Mid-infrared spectra of (top panel) UGC 12915 and (bottom panel) UGC 12914. The units are flux density (Jy) vs. wavelength. The PHT-S aperture size is 24 × 24´´ with 0.1 µm per pixel element. The relative uncertainties in the pixel values are denoted by the 1-sigma error bars. The absolute uncertainty in the flux density is estimated to be 30%. A composite MIR spectrum (from Lu 1999, private communication; generated from a linear combination of spectra from 28 normal galaxies) is denoted with dashed line (redshift-corrected to the VV 254 frame of reference). The major Aromatic Feature emission lines are demarked with vertical dashes. The -widths of the ISOCAM broad band filters are also indicated.