An optical B-band image was obtained at the 1-m Jacobus Kapteyn Telescope (JKT) of Roque de los Muchachos observatory (La Palma, Spain) in 1997 November with a 24 $\mu$ m $1024\times1024$ pixels Tek#4 CCD (see Table 2). An additional Johnson-V image was taken at the 1.52-m spanish telescope at EOCA (Calar Alto, Almería, Spain) in December 1993 with a Tek $1024\times1024$ CCD with pixel size of 19 $\mu$ m. This observation was broken up into five exposures with a total integration time of 2400s. Finally, an R-band image was obtained during service time in November 1998 with the ING wide-field camera (ING-WFC) equipped with four EEV42 $\rm 2k \times 4k$ pixels CCD detectors on the 2.5-m Isaac Newton Telescope (La Palma, Spain).

Narrow-band images in the light of [O III] $\lambda$ 5007 Å ( $\lambda_{0}=5012~$ Å, FWHM=50 Å) and H $\alpha$ ( $\lambda_{0} = 6568$ Å, FWHM=95 Å) were obtained. In order to subtract the continuum, V and R-band images were used, respectively. The [O III] $\lambda$ 5007 Å image was secured for us on December 1993 during service time with a $1280\times 1180$ pixels EEV5 CCD attached to the 2.5-m Isaac Newton Telescope. The H $\alpha$ image was taken during the same service time observations that the R-band image in November 1998 using the ING-WFC camera at the 2.5-m Isaac Newton Telescope (La Palma, Spain).

High S/N ratio dome flats and exposures of the sky taken in twilight were obtained in any case. The standard procedure of bias removal, dark-current subtraction, and flat fielding using dome and sky flat-field images was performed using the ESO image processing system MIDAS for the V image and IRAF for the B and R-band images.

Atmospheric conditions were photometric during the observing runs. Band V images were flux calibrated observing repeatedly during the nights a set of standard stars, taken from the lists given by Kent (1985) and Landolt (1973). Finally, the R-band image was flux calibrated using the radial surface brightness profile published by Papaderos et al. (1996a).

The [O III] $\lambda$ 5007 Å and H $\alpha$ images flux calibration was performed as follows (see also GZG). The continuum emission was subtracted, using the V and R images, respectively. Then, we trimmed those regions covered by the slits of the b(blue) and r(red) spectra (see Table 2). The band r spectra were convolved with the transmission curves of the corresponding narrow-band filters. Then, we compared image counts and fluxes measured in the convolved spectra for several adjacent regions in order to determine the precise location of the slits. Finally, once the regions covered by the slits were precisely defined, we corrected these calibration relations for the sensitivity of the filter at the corresponding wavelength and, in the case of the H $\alpha$ flux, for the contribution of the [N II] $\lambda$ 6548 Å and [N II] $\lambda$ 6583 Å lines. The discrepancies obtained among different spectra were in all cases lower than 10 per cent.

**Table 2:** Journal of observations
	Spectroscopic Observations $^{\dagger}$
	Exp.	Slit	Range	Disp.
Telescope	time(s)		(nm)	(Å/pix)
CAHA 2.2 m	3600	#1,2,4,6b	330-580	2.6
CAHA 2.2 m	1800	#5b	330-580	2.6
CAHA 2.2 m	3600	#1,2,4,5,6r	435-704	2.6
CAHA 2.2 m	3600	#3	390-650	2.6
INT 2.5 m	1800	#7,8R	637-677	0.39
INT 2.5 m	900	#9R	637-677	0.39
	Image Observations
	Exp.	Filter	Scale	PSF
Telescope	time(s)		( $^{\prime \prime }$ /pixel)	( $^{\prime \prime }$ )
JKT 1.0 m	600	B	0 $.\!\!^{\prime\prime}$ 33	1 $.\!\!^{\prime\prime}$ 0
CAHA 1.5 m	2400	V	0 $.\!\!^{\prime\prime}$ 33	1 $.\!\!^{\prime\prime}$ 6
INT 2.5 m	900	R	0 $.\!\!^{\prime\prime}$ 333	1 $.\!\!^{\prime\prime}$ 2
INT 2.5 m	900	[O III] $\lambda$ 5007	0 $.\!\!^{\prime\prime}$ 57	2 $.\!\!^{\prime\prime}$ 5
INT 2.5 m	7200	H $\alpha$	0 $.\!\!^{\prime\prime}$ 333	1 $.\!\!^{\prime\prime}$ 2
KPNO 2.3 m	900	J	0 $.\!\!^{\prime\prime}$ 66	1 $.\!\!^{\prime\prime}$ 8
KPNO 2.3 m	360	H	0 $.\!\!^{\prime\prime}$ 66	1 $.\!\!^{\prime\prime}$ 6
KPNO 2.3 m	540	K	0 $.\!\!^{\prime\prime}$ 66	1 $.\!\!^{\prime\prime}$ 7

$^{\dagger}$ See Fig. 1 of GZG for slit orientations and positions.

3.2 Near-infrared imaging

Near-infrared images (nIR hereafter) of Mrk 86 in J( $\lambda_{0}=1.25~\mu{\rm m}$ , $FWHM=0.30~\mu{\rm m}$ ), H ( $\lambda _{0}=1.65~\mu{\rm m}$ , $FWHM=0.28~\mu{\rm m}$ ) and Ks ( $\lambda _{0}=2.15~\mu{\rm m}$ , $FWHM=0.33~\mu{\rm m}$ ) were obtained on January 1998 with the Steward Observatory near-infrared camera equipped with a $256\times 256$ NICMOS3 detector attached to the 2.3-m Bok Telescope at Kitt Peak National Observatory (Arizona, U.S.A.). The observational procedure closely follows that of Aragón-Salamanca et al. (1993) and Gil de Paz et al. (2000b). The total integration time of each subimage was broken up into background-limited sub-exposures to avoid saturation of the detector. Images of adjacent blank areas of the sky were alternated with the target frames for accurate flat-fielding. Comparable amounts of time were spent imaging the source and the sky to ensure adequate monitoring of the sky changes. Individual subexposures of the object were offset several arcseconds to improve the final result.

The reduction process include: (1) bias and dark subtraction of object and sky frames; (2) flat-fielding using normalized sky frames; (3) sky subtraction using sky frames taken before and after each exposure; (4) bad pixel removal; (5) registering of the subimages using fractional pixel shifts and (6) median combining of all individual frames. The reduction was carried out using own IRAF procedures.

The nIR images were calibrated observing standard stars from the list from Elias et al. (1982) during the nights at the same airmasses than the object. We have assumed a color independent correction between the Ks and K' bands ( $K' \equiv KM$ ; $\lambda_{0}=2.12~\mu{\rm m}$ , $FWHM=0.34~\mu{\rm m}$ ; Wainscoat & Cowie 1992). In order to check the validity of this assumption, we convolved the Ks and K' filters response functions with Planck spectral distributions at different temperatures in the range 3000-20000K, obtaining the fluxes F_Ks and F_K'. The larger difference in $2.5\times\log(F_{Ks}/F_{K'})$ within this temperature range was 0.025^m, small enough to assume this correction to be independent of the spectral energy distribution.

Thus, once the fluxes were transformed to K'-band fluxes, we converted them to the standard K-band ( $\lambda _{0}=2.19~\mu{\rm m}$ , $FWHM=0.41~\mu{\rm m}$ ) applying the standard correction given by Wainscoat & Cowie (1992), $K'-K=0.22\times(H-K)$ .

In Fig. 1 we show the Mrk 86 neighborhood in the optical-near-infrared bands studied, including the H $\alpha$ and [O III] $\lambda$ 5007 Å images. The astrometric calibration was performed using the program PLATEASTROM (García-Dabó & Gallego 1999; see also http://www.ucm.es/info/Astrof/opera/opera.html). The bright field star placed at the relative position (60 $^{\prime \prime }$ E, 30 $^{\prime \prime }$ N), saturated in the V and R-band images, was artificially removed.

3.3 Spectroscopy

Up to 14 optical long-slit spectra at 10 different slit positions were obtained. The 11 medium-low resolution spectra (b and r spectra; see Table 2) were obtained with the Boller & Chivens spectrograph at the Cassegrain focus of the 2.2-m telescope at the German-Spanish Calar Alto observatory (Almería, Spain) in January 1993. Althought special care was taken to place the slits into position, using offsets from the bright field star at north of the galaxy, there were some lack of precision. The actual positions of the slits (see Fig. 1 of GZG) were determined a posteriori comparing broad-band image spatial cuts with the spatial profiles of the spectra previously convolved with the corresponding filter transmission curves. The detector employed was a $1024\times1024$ Tek#6 CCD with a pixel size of 24 $\mu$ m. The 600grmm^-1grating chosen provided a spectral resolution of 6 Å in the light of H $\alpha$ and a reciprocal dispersion of 2.6 Å/pixel, with a slit width of 2 $\hbox{$.\!\!^{\prime\prime}$ }$ 65. The final spatial scale was 1 $\hbox{$.\!\!^{\prime\prime}$ }$ 43/pixel. The spectral coverage was around 2500 Å and the grating angle was selected to cover the blue region ( $\sim3300-5800$ Å) and red domain ( $\sim4350-7045$ Å) in two different exposures which overlapped on the H $\beta$ region. The seeing was variable during the observing run with FWHM between $1\hbox{$.\!\!^{\prime\prime}$ }5-2\hbox{$.\!\!^{\prime\prime}$ }5$ . The low airmasses at which these spectra were obtained ( $\leq1.2$ ) guarantee that no significant loss of blue light due to atmospheric refraction has ocurred. In addition, three high resolution spectra (7R, 8R, 9R) were obtained with the IDS spectrograph at the Isaac Newton Telescope (INT) of the Roque de los Muchachos Observatory (La Palma, Spain) in January 1998 with a $1024\times1024$ Tek#3 24 $\mu$ m CCD (see Table 2). The 1200R grating (1200grmm^-1) was used with an slit width of 1 $^{\prime \prime }$ , providing a spectral resolution of 0.9 Å in the light of H $\alpha$ and a reciprocal dispersion of 0.39 Å/pixel. The spatial scale was of 0.33 $^{\prime \prime }$ /pixel.

**Table 3:** Positions and sizes for the Mrk 86 neighborhood regions
# $\char93 '$	RA(2000)	DEC(2000)	$r_{\rm 1e}$ $r_{\rm 1e}'$	e	Cl.
(1) (2)	(3)	(4)	(5) (6)	(7)	(8)
01 -	08:13:16.08	+46:00:17.1	0.71 PLR	2	B
02 -	08:13:16.33	+46:00:12.1	0.85 0.45	1	B
03 -	08:13:15.91	+46:00:11.8	0.74 0.17	2	B
04 01	08:13:14.72	+46:00:04.1	1.13 0.87	2	F
05 -	08:13:13.38	+46:00:01.7	0.91 0.56	2	B
06 02	08:13:12.36	+45:59:57.6	1.07 0.79	2	E
07 -	08:13:15.70	+45:59:57.4	0.83 0.41	2	E
08 03	08:13:16.18	+45:59:56.0	0.96 0.63	2	E
09 $^{\dagger}$ -	08:13:15.28	+45:59:54.1	0.76 0.24	2	E
10 $^{\dagger}$ -	08:13:13.56	+45:59:53.4	0.68 PLR	1	E
11 04	08:13:12.20	+45:59:53.0	0.75 0.21	2	B
12 $^{\dagger}$ -	08:13:16.74	+45:59:52.9	1.06 0.78	2	E
13 05	08:13:14.36	+45:59:52.0	1.31 1.09	1	E
14 06	08:13:11.72	+45:59:50.7	1.36 1.15	2	S
15 08	08:13:13.85	+45:59:50.7	0.96 0.63	2	E
16 07	08:13:14.77	+45:59:50.6	1.72 1.56	1	S
17 -	08:13:09.23	+45:59:49.5	1.03 0.74	2	E
18 09	08:13:13.08	+45:59:48.2	0.93 0.59	2	S
19 10	08:13:15.02	+45:59:45.3	1.56 1.38	1	S
20 11	08:13:15.33	+45:59:45.2	2.29 2.17	1	S
21 12	08:13:15.78	+45:59:44.8	0.88 0.51	1	E
22 $^{\dagger}$ 13	08:13:13.52	+45:59:44.2	1.02 0.72	2	E
23 14	08:13:10.94	+45:59:43.0	0.97 0.65	2	E
24 -	08:13:08.52	+45:59:41.4	1.19 0.95	2	E
25 -	08:13:15.64	+45:59:41.2	1.39 1.19	2	N
26 15	08:13:13.13	+45:59:40.6	2.11 1.98	2	S
27 16	08:13:12.94	+45:59:38.4	1.32 1.11	1	S
28 $^{\dagger}$ 17	08:13:13.64	+45:59:36.9	0.84 0.43	2	E
29 18	08:13:15.51	+45:59:35.6	0.85 0.45	2	S
30 19	08:13:15.92	+45:59:35.6	0.93 0.59	2	E
31 20	08:13:12.87	+45:59:35.1	0.65 PLR	2	N
32 21	08:13:13.41	+45:59:34.3	0.69 PLR	2	S
33 22	08:13:16.11	+45:59:33.9	0.96 0.63	1	E
34 -	08:13:08.99	+45:59:33.0	1.32 1.11	2	E
35 23	08:13:13.69	+45:59:33.0	1.28 1.06	2	N
36 24	08:13:13.28	+45:59:32.4	0.71 PLR	2	N
37 26	08:13:15.88	+45:59:29.3	1.90 1.76	1	S
38 27	08:13:16.23	+45:59:29.2	0.94 0.60	2	N
39 28	08:13:17.76	+45:59:29.0	1.20 0.96	2	B
40 25	08:13:12.76	+45:59:28.8	1.62 1.45	2	S
41 29	08:13:14.04	+45:59:27.3	1.07 0.79	2	S
42 31	08:13:12.98	+45:59:26.5	1.29 1.07	1	S
43 32	08:13:16.77	+45:59:25.9	1.12 0.86	1	E
44 33	08:13:12.63	+45:59:24.6	0.79 0.33	2	N
45 -	08:13:14.57	+45:59:23.2	1.15 0.90	2	S
46 34	08:13:12.37	+45:59:22.3	1.03 0.74	2	N
47 36	08:13:16.76	+45:59:21.2	1.22 0.98	1	S
48 35	08:13:12.69	+45:59:21.2	1.18 0.93	1	E
49 -	08:13:15.36	+45:59:18.8	0.73 0.12	2	N
50 37	08:13:11.19	+45:59:18.8	1.42 1.22	2	E
51 38	08:13:12.72	+45:59:17.3	1.28 1.06	2	N
52 39	08:13:12.97	+45:59:17.2	1.32 1.11	2	S
53 $^{\dagger}$ -	08:13:13.64	+45:59:17.1	0.86 0.47	2	E
54 40	08:13:14.91	+45:59:15.9	2.01 1.88	1	S
55 $^{\dagger}$ -	08:13:16.12	+45:59:15.8	0.69 PLR	2	E
56 -	08:13:17.35	+45:59:14.6	0.81 0.37	2	E

**Table 3:** continued
# $\char93 '$	RA(2000)	DEC(2000)	$r_{\rm 1e}$ $r_{\rm 1e}'$	e	Cl.
(1) (2)	(3)	(4)	(5) (6)	(7)	(8)
57 $^{\dagger}$ -	08:13:16.05	+45:59:13.1	0.82 0.39	2	E
58 41	08:13:16.84	+45:59:12.3	1.00 0.69	2	E
59 44	08:13:15.25	+45:59:11.6	0.86 0.47	2	S
60 43	08:13:18.27	+45:59:11.4	0.86 0.47	2	E
61 42	08:13:14.39	+45:59:11.4	0.77 0.27	3	F
62 -	08:13:16.04	+45:59:09.8	1.28 1.06	1	S
63 -	08:13:12.48	+45:59:09.4	0.56 PLR	2	N
64 45	08:13:16.74	+45:59:08.7	1.30 1.08	1	E
65 -	08:13:17.60	+45:59:07.6	1.22 0.98	1	E
66 46	08:13:13.05	+45:59:07.0	1.47 1.28	2	S
67 47	08:13:15.94	+45:59:06.9	0.80 0.35	3	F
68 -	08:13:13.75	+45:59:05.9	0.62 PLR	2	E
69 48	08:13:11.94	+45:59:05.2	0.95 0.62	2	B
70 49	08:13:14.28	+45:59:02.8	1.43 1.24	2	S
71 -	08:13:15.81	+45:59:02.6	0.75 0.21	2	N
72 50	08:13:15.68	+45:59:02.0	1.05 0.76	2	N
73 -	08:13:09.67	+45:59:00.6	1.08 0.80	2	B
74 -	08:13:15.52	+45:58:59.0	1.08 0.80	2	E
75 -	08:13:15.07	+45:58:53.5	0.95 0.62	2	E
76 $^{\dagger}$ -	08:13:14.36	+45:58:52.3	1.11 0.84	1	E
77 -	08:13:17.80	+45:58:52.1	0.87 0.49	2	E
78 -	08:13:12.32	+45:58:48.3	1.03 0.74	1	E
79 $^{\dagger}$ -	08:13:09.05	+45:58:47.9	1.13 0.87	2	E
80 51	08:13:14.77	+45:58:46.0	1.09 0.82	2	E
81 $^{\dagger}$ -	08:13:08.64	+45:58:46.0	0.64 PLR	2	B
82 -	08:13:14.26	+45:58:45.1	0.85 0.45	2	B
83 -	08:13:16.92	+45:58:37.0	1.25 1.02	2	B
84 $^{\dagger}$ -	08:13:17.39	+45:58:36.6	0.91 0.56	2	E
85 55	08:13:13.69	+45:58:23.4	1.37 1.17	2	S

(1) Knot number (reverse DEC sorted).
(2) Old knot number as given in GZG.
(3) RA(J2000).
(4) DEC(J2000).
(5) Radius (in arcsec) at 1e-folding.
(6) Radius (in arcsec) corrected for the atmospheric seeing.
(7) Folding (e¹, e² or e³) where colors were measured.
(8) Classification (see Sect. 4.3).
PLR = Point-Like Region.
$^{\dagger}$ Knot sizes measured on the H $\alpha$ image (see Sect. 4.2).

These spectra were reduced using the FIGARO (January 1993) and IRAF (January 1998) software packages. After bias removal and flat-fielding, the frames were cleaned of cosmic rays. The sky was removed from each frame by subtracting a polynomial fitted to those regions free for object emission. Wavelength calibration was performed by using He-Ar lamps observed inmediately before and after the galaxy integration. The standard stars Hiltner 102 and Hiltner 600 were observed at different airmasses in order to correct for atmospheric extinction and to ensure absolute flux calibration.

Finally, we requested an UV spectrum of Mrk 86 from the International Ultraviolet Explorer (IUE) Final Archive (see Fig. 4). It was originally taken by Alloin and Duflot in January 1983 (see Bonatto et al. 1999). The total exposure time of this spectrum, SWP18927, was 24000s. It was obtained in low dispersion mode with the SWP camera. In the observing spectral range, 1150-1975 Å, the resolution power varies between 270-300 (Cassatella et al. 1985).

3 Observations and reduction

3.1 Optical imaging

3.2 Near-infrared imaging

3.3 Spectroscopy