astro-ph0007434/ms.tex
1: \documentstyle[psfig]{mn}
2: 
3: \def\refitem{\par\parskip 0pt\noindent\hangindent 20pt}
4: \def\lesssim{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$<$}}}}
5: \def\gtrsim{\mathrel{\hbox{\rlap{\hbox{\lower4pt\hbox{$\sim$}}}\hbox{$>$}}}}
6: 
7: \newcommand{\chisq}{\mbox{$\chi^2$}}
8: \newcommand{\nh}{\mbox{$N_H$}}
9: \newcommand{\ergs}{\mbox{erg~s$^{-1}$}}
10: \newcommand{\ergscm}{\mbox{erg~s$^{-1}$~cm$^{-2}$}}
11: 
12: \title[Spectral energy distributions of FR~I nuclei and
13: the FR~I/BL Lac unifying model]       
14: {Spectral energy distributions of FR~I nuclei  and
15: the FR~I/BL Lac unifying model}
16: 
17: \author[Alessandro Capetti et al.]
18: {A. Capetti$^1$, E. Trussoni$^1$, A. Celotti$^2$,
19: L. Feretti$^3$, M. Chiaberge$^2$\\
20: $^1$Osservatorio Astronomico di Torino, Strada Osservatorio 20,
21: I-10025 Pino Torinese, Italy\\
22: $^2$S.I.S.S.A., Via Beirut 2--4, I-34014 Trieste, Italy\\
23: $^3$Istituto di Radioastronomia del C.N.R., Via  Gobetti 101, I-40129 
24: Bologna, Italy}
25: \date{Received ***; in original form ***}       
26: \begin{document} 
27: \maketitle 
28: 
29: \begin{abstract}
30: 
31: We consider archival ROSAT and Hubble Space Telescope (HST)
32: observations of five FR~I radio galaxies and isolate their nuclear
33: emission from that of the host galaxy. This enable us to determine the
34: Spectral Energy Distributions (SED) of their nuclei spanning from the
35: radio to the X--ray band. They cannot be described as single
36: power--laws but require the presence of an emission peak located
37: between the IR and soft X-ray band. We found consistency between the
38: SED peak position and the values of the broad band spectral indices of
39: radio galaxies with those of BL Lac, once the effects of beaming are
40: properly taken into account. FR~I SED are thus qualitatively similar
41: to those of BL Lacs supporting the identification of FR~I sources as
42: their mis-oriented counterparts.  No dependence of the shape of the
43: SED on the FR~I orientation is found.
44: 
45: \end{abstract}
46: 
47: \begin{keywords} 
48: galaxies:active - jets - nuclei - photometry - BL Lacertae
49: objects:general
50: \end{keywords}
51: 
52: \section{Introduction} 
53: 
54: In the framework of the unifying models for AGN, the low luminosity
55: Fanaroff--Riley (1974) type I, FR~I, radio galaxies are identified as
56: the parent population of BL Lac objects: the non--thermal continuum   
57: of BL Lacs is dominated by beaming effects, resulting from
58: the relativistic motion of plasma at a small angle with respect to
59: the line of sight (Blandford \& Rees 1978), and FR~I radio galaxies
60: would represent their mis-oriented counterparts.  Support to this   
61: unification comes from different and independent observational  
62: pieces of evidence (see Urry \& Padovani 1995 for a review).
63: 
64: In this scenario the detection of any non--thermal jet emission       
65: in FR~I would allow a quantitative analysis of the
66: unified model based on the direct comparison of the beamed component 
67: in the two populations.  However, to pursue this goal we need to
68: isolate the jet emission in the cores of radio galaxies.
69: 
70: Recently, Chiaberge, Capetti \& Celotti (1999) have analyzed HST
71: images of a complete sample of 33 FR~I radio galaxies and found an
72: unresolved central optical source in the majority of them.  The
73: luminosity of these sources show a striking linear correlation with
74: that of the radio cores, strongly arguing for a non--thermal
75: synchrotron origin also for the optical emission.  Furthermore, the   
76: high detection rate suggests that obscuring tori in the nuclei of
77: FR~I must be either geometrically thin or present only in a minority
78: of the objects.
79: 
80: HST images also reveal the presence of extended nuclear disks in low
81: luminosity radio galaxies (e.g. Jaffe et al. 1993).  As shown by
82: Capetti \& Celotti (1999, hereafter CC99) from the analysis of 5
83: objects, such disks can be used as reasonably good indicators of the
84: orientation of the radio sources. All of the 5 FR~I also show the
85: optical central source. By comparing their intensity with the emission
86: of BL Lacs with similar extended (isotropic) characteristics they
87: found that: i) BL Lacs are $10^{2}$--$10^{5}$ brighter than the
88: corresponding radio galaxies and ii) the relative core brightness
89: decreases with increasing observing angle, quantitatively consistent
90: with the inference that also the FR~I nuclear component is dominated
91: by the beamed jet emission.
92: 
93: In this framework, multi--wavelengths observations of FR~I radio
94: galaxies clearly represent an alternative and more detailed way to
95: test the model, as their SED can be directly compared with those of BL
96: Lacs.  The latter ones are well described by two broad peaks (in $\nu
97: L_{\nu}$), the first one attributed to non--thermal synchrotron
98: emission while the second one is most commonly ascribed to inverse
99: Compton scattering.  Moreover, the whole BL Lac (and Flat Spectrum
100: Radio Quasar, FSRQ) broad band phenomenology seems to be accounted for
101: by the (continuous) distribution in the energy position of the two
102: peaks and their relative luminosity (Padovani \& Giommi 1995; Fossati
103: et al. 1998): an increase in the inverse Compton vs synchrotron
104: luminosity ratio corresponds to a decrease in the peak frequencies, 
105: describing the progressive trend from FSRQ, to High and Low Energy
106: Peaked BL Lacs (HBL and LBL, Giommi \& Padovani 1994). This
107: establishes a direct relation between the type of SED/object and the
108: relative contribution of the two components in the X--ray band (steep
109: synchrotron turnover vs flat Compton power law; Padovani \& Giommi
110: 1996).  These phenomenological trends seem also to be simply connected
111: with the source (bolometric and radio) power, increasing from HBL to
112: LBL (Ghisellini et al. 1998).
113: 
114: But how does orientation affect the SED ? While it has been rather
115: convincingly shown that orientation alone cannot account for the
116: differences in the SED of blazars (Sambruna, Maraschi \& Urry 1996),
117: the alternative interpretation just described has been - so far -
118: focused only on highly beamed sources (blazars), neglecting the
119: (comparatively smaller) effects due to orientation.
120: 
121: SED of radio-galaxies and in particular of FR~I have been derived by
122: several authors (e.g. Impey \& Gregorini 1993; Ho 1999). However, the
123: comparison with those of BL Lacs is hampered by the large aperture
124: from which data have been obtained, which often includes a dominant
125: contribution from the host galaxy.  We therefore gathered archival
126: observations of FR~I from which it is possible to separate the nuclear
127: from the extended emission, in particular ROSAT data and HST/NICMOS
128: infrared images. These can be combined with our analysis of the HST
129: optical images and radio core measurements from the
130: literature. Clearly, data in the infrared/optical and X--ray bands
131: could also provide an independent test for the role of absorption from
132: thick material (torus) in this unifying scheme.
133: 
134: As the orientation of the radio galaxies is a crucial parameter, in
135: this paper we focus on the sample of 5 objects studied in CC99 (see
136: Table 1), for which we have such information.
137: 
138: ROSAT archival pointed observations of 4 out of 5 galaxies are
139: available and are presented in Section 2.  HST/NICMOS infrared images
140: have been obtained for 3 of the selected objects and are described in
141: Section 3. In Section 4 we discuss the role of absorption with respect
142: to the inferred nuclear SED, which we assemble in Section 5. The
143: comparison with the SED of BL Lacs and the implications for the
144: FR~I/BL Lac unified scheme are discussed in Section 6, while in the
145: final Section 7 we present our conclusions and future perspectives.
146:   
147: Throughout this work, we adopt the following values for the
148: cosmological parameters: $H_{\rm 0} = 75$ km s$^{-1}$ Mpc$^{-1}$ and
149: $q_{\rm 0} = 0.5$.
150: 
151: 
152: \section{The ROSAT observations} 
153: 
154: ROSAT HRI and PSPC data are available for 3 of the selected sources,
155: namely 3C~31, 3C~270 and 3C~465, while for 3C~264 only PSPC data
156: exist. In Table 2 we list the available pointed observations. The
157: X--ray properties of these targets have been already discussed by
158: different authors (see last column of the table) but we decided to
159: re-analyze the data to follow a more homogeneous approach.
160: 
161: For the analysis we have used the EXSAS package (release of Jan.
162: 1996; Zimmermann et al. 1995) and the data have been corrected for
163: vignetting and dead time.  In the PSPC observations only the energy
164: range 0.15 - 2.1 keV has been considered and the spectra have been
165: rebinned according to the source count rate \footnote{The minimum
166: signal to noise ratio per energy channel is $5 \sigma$ for 3C~264 and
167: 3C~270, $4 \sigma$ for 3C~31 and $3 \sigma$ for 3C~465.}.
168:  
169: All targets are members of clusters or groups of galaxies, and this
170: implies that any central non--thermal X--ray emission must be
171: disentangled from that of a hot intergalactic medium, with
172: temperatures $\sim 3 - 8$ keV (Sarazin 1986), and of the extended hot
173: halos ($\sim 1$ keV; Donnelly et al. 1990; Trussoni et al. 1997).
174: Therefore for the spectral analysis of the PSPC data we used composite
175: models allowing for the presence of both a thermal and a non--thermal
176: (power-law) component.  The free physical parameters in the spectral
177: fits are, besides the two normalizations, the hydrogen column density
178: $N_H$, the photon index $\Gamma$ for the power--law spectrum, the
179: temperature $T$ and metallicity $\mu$ for the thermal component
180: (optically thin plasma, Raymond \& Smith 1977).  As count rates are
181: usually quite low, some parameters must be fixed in order to deduce
182: usefully constrained fits. In all cases (except for 3C~264) we set
183: $N_H$ to its galactic value $N_{H, gal}$, and $\mu =0.5$, as typical
184: for hot galactic halos (Fabbiano 1989). Fixing the metallicity does
185: not significantly affect our results as the parameters describing the
186: non-thermal component depend only weakly on $\mu$ -- at least if it
187: lies within the plausible range $0.2 \leq \mu \leq 1$.  To establish
188: the statistical significance of the non thermal component, we have
189: then checked through an F-test when its inclusion improves the fit.
190: 
191: The HRI data should allow in principle to spatially separate the
192: unresolved central component from the coronal emission.  However, the
193: spatial resolution of the HRI is insufficient to fully resolve the
194: cores of the galactic coronae, particularly for the more distant
195: sources. Furthermore, problems in the aspect solution associated with
196: the satellite wobble may lead to smearing of point-like sources
197: (algorithms proposed to overcome this problem work only for bright
198: sources, Morse 1994; Harris et al. 1998; Sarazin et al. 1999)
199: compromising the spatial analysis.  We therefore preferred to use the
200: HRI observations just to check their consistency with the count rates
201: predicted by the PSPC spectral decomposition.
202: 
203: No statistically significant variability has been found in the light
204: curves of the observations, and in the targets observed with PSPC and
205: HRI the count rates are consistent in both instruments. This is also
206: true for two very short HRI exposures of 3C~31 (Jan. 1992; see
207: Trussoni et al. 1997) and 3C~270 (Jul. 1994), that we have not
208: included in the present analysis.
209: 
210: \subsection{Individual sources}
211: 
212: In the following we discuss the X--ray properties of the targets and
213: in particular the results from the spectral fits with a two components
214: model (thermal + power-law). The results are summarized in Table 3.
215: 
216: \noindent
217: {\bf 3C~31}.  The host galaxy of 3C~31 (NGC 383) is the dominant
218: member of the group Arp 331. A fit to its PSPC spectrum including a
219: non-thermal source is more acceptable at $\approx 90$\% confidence
220: level than a single thermal component. We obtain a gas temperature of
221: $T \approx 0.6$ keV, but the spectral index of the non--thermal
222: component is essentially unconstrained, $\Gamma = 1.8^{+0.6}_{-1.1}$.
223: We derive the power-law flux adopting two values of the photon index
224: (namely $\Gamma=1.7$ and $\Gamma=2.3$ see Table 3): the resulting
225: estimates differ by only 20 \% indicating that in this case
226: uncertainties are dominated by photon statistics.  Our spectral
227: results are basically consistent with those of Komossa \& B\"ohringer
228: (1999).
229: 
230: The HRI brightness radial profile (Fig. 1a) shows a compact component
231: (broader than the PSF) and faint extended emission out to $40^{\prime
232: \prime}$.  From the PSPC analysis we expect a HRI photon flux of
233: $\approx 4.1 \times 10^{-3}$ cts s$^{-1}$ and $\approx 2.9 \times
234: 10^{-3}$ cts s$^{-1}$ from the thermal and non--thermal components,
235: respectively.  The total HRI count rate (within the same extraction
236: radius of the PSPC spectrum) is $(6.6 \pm 1.3)\times 10^{-3}$ cts
237: s$^{-1}$, in agreement with our estimates. Furthermore, the predicted
238: non--thermal photon flux cannot exceed that measured in the central
239: compact component. Indeed within $15 \arcsec$ we measure $(5.2 \pm
240: 0.5)\times 10^{-3}$ cts s$^{-1}$.
241:  
242: \noindent
243: {\bf 3C~264}. This radio galaxy lies at a projected distance of
244: $\approx 8^\prime$ from the center of the cluster A1367. Low
245: brightness emission from the cluster gas is detected around the
246: galaxy, but it is effectively removed after background subtraction.
247: The spectrum is remarkably well fitted with a single power law
248: ($\Gamma \approx 2.46$) absorbed by a column density $N_{\rm H} =
249: 2.6^{+0.4}_{-0.2} \times 10^{20}{\rm cm}^{-2}$, consistent with the
250: Galactic one ($2.55 \times 10^{20}{\rm cm}^{-2}$), in agreement with
251: the results of Edge \& R\"ottering (1995).  A thermal component might
252: be present at $\approx 1 \sigma$ level (assuming $T=1$ keV) at a
253: confidence level of $\approx 80$\%.
254: 
255: \noindent
256: {\bf 3C~270}. This source is associated to NGC~4261, main member of a
257: group of galaxies located behind the Virgo cluster. The emitted flux
258: is fitted better by a composite spectrum than by a single thermal law
259: (improving of the quality of the fit at a significance level of
260: $\approx 99.2$\%).  The two component spectral model provides
261: acceptable fits to the PSPC data with $\Gamma \approx 1.7$ and $T
262: \approx 0.8$ keV.  These results, and in particular the presence of a
263: flat non--thermal component, are confirmed by the analysis of the ASCA
264: observations (Sambruna et al. 1999).
265: 
266: The HRI brightness profile (Fig.~1b) is similar to that of 3C~31 but
267: the extended emission (detected up to a radius $\approx 1^{\prime}$)
268: is much more prominent. Again the central compact region is more
269: extended than the PSF.  In the HRI we detect a total count rate $(3.2
270: \pm 0.3) \times 10^{-2}$ cts s$^{-1}$, as expected from the PSPC data
271: which predict $2.2 \times 10^{-2}$ cts s$^{-1}$ and $1.1 \times
272: 10^{-2}$ cts s$^{-1}$ for the thermal and power--law components,
273: respectively. The photon flux of the compact component is $(1.9 \pm
274: 0.1) \times 10^{-2}$ cts s$^{-1}$ again consistent with the derived
275: intensity of the power--law one.
276: 
277: \noindent
278: {\bf 3C~465}. The galaxy NGC~7720 (3C~465) is the brightest member of
279: the cluster A2634. The PSPC emission from the central region is well
280: fitted by a thermal spectrum with $T \approx 1.2$ keV.  The addition
281: of a non--thermal component does not improve the fit.  We set an upper
282: limit to the non--thermal luminosity of $\lesssim 2 \times 10^{41}$
283: erg s$^{-1}$ (see Table 3).  This contrasts with the results obtained
284: by Hardcastle \& Worrall (1999) based on the spatial analysis of the
285: HRI brightness profile.
286: 
287: \noindent
288: {\bf NGC~7052}. No pointed observation is available for this source.
289: From the RASS an upper limit of the count rate $\lesssim 0.01$ cts
290: s$^{-1}$ is found (Brinkmann, private comm.). Assuming a power-law
291: spectrum with $\Gamma=2.3$ this corresponds to a total luminosity
292: $\lesssim 1.8 \times 10^{41}$ erg s$^{-1}$.
293:  
294: \subsection{Summary of the X-ray data}
295: Non--thermal emission is detected in 3 out of the 4 targets with
296: pointed X--ray observations: in 3C~264 it is the dominant source,
297: while in 3C~270 and 3C~31 the inclusion of a non--thermal component
298: significantly improves the fit.  Only an upper limit can be set for
299: the power-law emission from the nucleus of 3C~465.  The photon fluxes
300: predicted by the spectral decomposition agree with those observed in
301: the HRI images.
302: 
303: Worral \& Birkinshaw (1994) \& Worral (1997) found a correlation
304: between the luminosities of FR~I radio galaxies in the radio and soft
305: X-ray bands, recently confirmed by Canosa et al. (1999) and Hardcastle
306: and Worrall (1999) who analyzed the PSPC and HRI brightness profiles
307: of a large sample of objects.  This trend appears to hold also at
308: harder X-ray energies for brighter radio galaxies (Trussoni et
309: al. 1999a).  The non-thermal luminosity of the three detected sources
310: we found is consistent with this correlation.  Conversely, the upper
311: limit for 3C~465 is $\approx 4$ times lower than expected. Concerning
312: the thermal component, elliptical galaxies show a correlation between
313: the coronal X--ray luminosity and the optical magnitude in the $B$
314: band (Fabbiano et al. 1984, Donnelly et al. 1990, Eskridge et al.
315: 1995, Trussoni et al. 1997, Beuing et al. 1999).  The thermal
316: luminosities derived for 3C~31 and 3C~270 (as well as the upper limit
317: deduced for 3C~264) are again consistent with this X-ray/optical
318: relation.  Only the luminosity of 3C~465 is significantly lower (by a
319: factor $\approx 5$) than expected.
320: 
321: This overall agreement provides independent support to the validity of
322: our spectral decomposition.
323: 
324: \section{The HST/NICMOS observations} 
325: 
326: HST/NICMOS observations are available in the public archive for
327: 3C~264, 3C~270 and NGC~7052. For the first two objects we analyzed
328: images taken with the F110W, F160W and F205W wide band filters
329: (approximatively corresponding to the H, J and K bands, respectively),
330: while for NGC~7052 only F160W images exist.
331: 
332: The targets were observed using NICMOS Camera 2 (0\farcs075/pixel).
333: The field of view of this 256 $\times$ 256 pixels camera is 19\farcs4
334: $\times$ 19\farcs4.  All observations were carried out with a
335: MULTIACCUM sequence (MacKenty et al. 1997), i.e. the detector is read
336: out non-destructively several times during each integration to
337: facilitate removal of cosmic rays and saturated pixels.  The data were
338: re-calibrated using the pipeline software CALNICA v3.0 (Bushouse et
339: al. 1997) and the best reference files in the Hubble Data Archive to
340: produce flux calibrated images. Bad pixels were removed interpolating
341: from values of neighboring pixels.  A log of the observations is given
342: in Table 4.
343: 
344: The final images are presented in Fig.~2. While in NGC~7052 the
345: extended dusty disc seen in the optical images is clearly visible, we
346: do not find evidence for circum--nuclear structures in 3C~264 and
347: 3C~270.  This is somewhat expected at this longer wavelengths where
348: dust absorption is less efficient, confirming that these discs are
349: associated with dust with small optical depth (e.g. Jaffe et
350: al. 1993).  On the other hand, similarly to what found in the optical
351: band, in all targets it is detected a central IR source, whose
352: emission can be now easily isolated from that of the host galaxy. We
353: performed aperture photometry to measure their nuclear fluxes in the
354: different bands (reported in Table 4).  Photometric errors are
355: dominated by the accuracy of the NICMOS internal flux calibration,
356: which is within 5\%, except in the case of NGC~7052 where the relative
357: error is $\sim 30 $ \%.
358: 
359: \section{The role of absorption in FR~I nuclei}
360: 
361: As absorption could have significantly affected our measurements in
362: the optical and X--ray bands, and thus in principle alter the inferred
363: SED properties, in this section we carefully examine its possible
364: role.
365: 
366: Photoelectric absorption can be very effective on soft X-rays, and
367: thus measures in this band provide a powerful tool to estimate the
368: (equivalent) column density in neutral hydrogen $N_{\rm H}$.  In
369: 3C~264 the X-ray emission is dominated by a power--law component over
370: the whole ROSAT energy range and its spectrum is sufficiently well
371: determined that we could directly measure a column density which -- as
372: already mentioned -- is fully consistent with the Galactic one: no
373: significant local absorption is present in this source.  This is
374: somewhat expected as 3C~264 is likely to be seen at a small angle from
375: the jet axis (see Table 1) and thus along an unobscured line of sight
376: (within the nuclear region).  Conversely, in 3C~31, 3C~270 and 3C~465,
377: the observed X-ray flux is dominated by extended thermal emission
378: which is unaffected by nuclear absorption. The detection in 3C~31 and
379: 3C~270 of non--thermal emission above 1 keV only sets an upper limit
380: of $N_{\rm H} \lesssim 10^{23}$ cm$^{-2}$\footnote{This column density
381: corresponds to an attenuation of a factor 100 in the intensity at 2
382: keV}.  Note that these results are in agreement with the conclusions
383: of Chiaberge et al. (1999), namely that optically thick material
384: (torii) affects the nuclear emission at most in a minority of FR~I.
385: Clearly only X-ray observations at higher spatial resolution,
386: increasing the fraction of nuclear vs extended component, can provide
387: more stringent and valuable estimates of $N_H$ in these cases.
388: 
389: Obviously optical images do not directly lead to estimates of
390: absorption. However, the tight linear correlation which exists between
391: the optical (R band) and radio fluxes/luminosities of FR~I cores (see
392: Introduction) can set interesting upper limits on the nuclear
393: extinction. In fact, even if the entire dispersion of this correlation
394: (0.4 orders of magnitude) is due unically to foreground dust
395: absorption varying from source to source, the extinction can not
396: exceed $\sim$ 2 magnitudes.  This limit translates into a gas column
397: density of $N_{\rm H}\lesssim 5 \times 10^{21}{\rm cm}^{-2}$ (assuming a
398: standard gas-to-dust ratio and the extinction curve by Cardelli,
399: Clayton \& Mathis 1988), and thus the flux at 1 keV could be depressed
400: at most by a factor of 4 (photoelectric absorption values tabulated by
401: Morrison \& McCammon 1983).
402: 
403: Given these upper limits on the exctintion, what would be its effect
404: on the SED properties (and more specifically on the broad band
405: spectral indeces)?  As the fluxes in the V band and at 1 keV can be
406: underestimated by a factor $\lesssim 10$ and $\lesssim 4$
407: respectively, the corresponding (maximum) changes in the spectral
408: indeces are limited to $\Delta \alpha_{\rm RO} \lesssim -0.20$,
409: $\Delta \alpha_{\rm RX}\lesssim -0.08$ and $\Delta \alpha_{\rm OX}
410: \lesssim 0.15 $ ($F_{\nu} \propto \nu^{-\alpha}$ \footnote{they are
411: calculated between 5 GHz, 5500 \AA and 1 keV (rest frame)}). Such
412: uncertainties do not affect significantly our findings and
413: conclusions. 
414: 
415: \section{Spectral Energy Distributions of FR~I nuclei}
416: 
417: Combining the available data in different energy bands, it is possible
418: to build a SED for the FR~I nuclei, isolating for the first time this
419: emission from that of the host galaxy (see Fig.~3 and Table 5).
420: Besides the IR and X--ray data presented in the previous sections, we
421: used HST optical fluxes from CC99 and radio data taken from the
422: literature. In the following we describe the individual SED -- always
423: referring to a $\nu L_{\nu}$ representation -- and characterize them
424: through the broad band spectral indices, $\alpha_{\rm RO}$,
425: $\alpha_{\rm RX}$ and $\alpha_{\rm OX}$ (the potential effects
426: produced by the presence of local absorption are discussed in the
427: previous Section 4).
428: 
429: \noindent
430: $\bullet$ {\bf 3C~31}. The average radio-to-optical spectral index is
431: $\alpha_{\rm RO} = 0.79$ while the optical-to-X--ray one is
432: $\alpha_{\rm OX} = 1.03$.  These spectral indices imply the presence
433: of a peak in the energy distribution. A better energy coverage or the
434: determination of the optical (or X--ray) spectral slope are necessary
435: in order to better constraint its location.
436: 
437: \noindent
438: $\bullet$ {\bf 3C~264}. Its SED is the better determined in our sample
439: as there are data at radio, infrared and optical frequencies, the
440: X--ray spectrum is also well defined and the absorption is constrained
441: to a negligible value.  The energy distribution rises from the radio
442: toward higher frequencies (with an average index $\alpha_{\rm RO} =
443: 0.63$) and it is still rising in the IR/optical band, although with a
444: flatter slope.  At 1 keV the intensity is 20\% larger than in the
445: optical (in $\nu$L$_{\nu}$) but the X--ray spectrum is rather steep
446: ($\alpha_{\rm X}$= 1.45) indicating that the SED has already reached
447: its peak, which must be thus located in the interval $10^{15} -
448: 10^{16.5}$ Hz.
449: 
450: \noindent
451: $\bullet$ {\bf 3C~270}. This SED is characterized by the sharp decline
452: in the IR/optical range described by a spectral index $\alpha \sim
453: 3.5$.  Clearly dust absorption can play a significant role in
454: producing this rapid decline.  However its spectrum in the X--ray band, as
455: also measured by ASCA, indicates that the minimum observed at
456: optical energies is real, although possibly enhanced by the presence
457: of dust. Moreover, its representative point is not located
458: significantly below the radio-optical correlation described by
459: Chiaberge et al. (1999), arguing against an obscuration larger than
460: average.
461: 
462: \noindent
463: $\bullet$ {\bf 3C~465}. After rising from the radio to the optical
464: ($\alpha_{\rm RO} = 0.83$), the SED of 3C~465 is already falling at
465: higher energies as the X--ray upper limit lies below the optical
466: point. This indicates that also in this case a peak in the emission is
467: reached below the X--ray band, although the poorly determined SED does
468: not allow us to better constrain its location.
469: 
470: \noindent
471: $\bullet$ {\bf NGC~7052}. The average radio-to-optical spectral index
472: is $\alpha_{\rm RO} = 0.80$. Contrarely to what seen in 3C~270, the
473: nuclear source in NGC~7052 is (marginally) fainter in the infrared
474: than in the optical band. This indicates that, although the extended
475: disk is seen at the highest inclination ($\sim$ 70$^\circ$) and has a
476: very large optical depth, the nuclear emission is not significantly
477: affected by dust obscuration.  The upper limit derived from the RASS
478: requires a flattening of the SED slope at higher energies.
479: 
480: Summarizing, the nuclear SED of the selected FR~I present a wide
481: variety of behaviors.  However three common features can be
482: recognized: i) it always rises from the radio to the infrared/optical;
483: ii) no SED can be described as a single power law; iii) for all of the
484: sources with pointed X--ray data the presence of a peak in the
485: emission can be localized at energies lower than soft X--rays.
486: 
487: \section{Discussion}
488: 
489: \subsection{Comparison of the SED of FR~I and BL~Lacs}
490: 
491: Let us now compare the SED of FR~I with those of BL Lacs.  Despite of
492: the poor spectral coverage of the FR~I SED (somewhat reminiscent of
493: the SED coverage of BL Lac objects in the 80s), it is possible to
494: recognize a qualitative similarity with the SED of blazars, and in
495: particular the likely presence of one broad peak (see Fig.~3). In
496: fact, as discussed in Section 5, the spectra appear to raise from the
497: radio to the IR/optical band, and in all cases - except for NGC~7052 -
498: the X--ray information (flux and/or spectrum) suggest that a peak in
499: the SED occurs at lower energies.
500: 
501: In the specific case of 3C~270 the flat X--ray spectrum is plausibly
502: due to the dominance of a rising Compton component in this band.  And
503: interestingly its extremely steep IR-optical spectrum seems to
504: indicate the presence of a minimum, fully consistent with the
505: indications derived from the X--ray spectral shape.  Agreement between
506: the hints inferred from the IR/optical and X--ray spectra is also
507: found in the SED of 3C~264 in which the X--rays can be interpreted as
508: the steep high energy cut-off of the synchrotron component.
509: 
510: Clearly, the paucity of spectral information limits any detailed
511: analysis and in particular does not allow us to determine the location
512: of the energy peak precisely enough to identify the 'type' of blazar
513: putatively hosted in the radio galaxies, except possibly for 3C~270
514: and 3C~264, which resemble an LBL and HBL, respectively.
515: 
516: Nevertheless, it is possible to profitably use the available data to
517: compare the broad band spectral indices with those of BL Lacs, as
518: shown in Fig.~4 where complete samples of BL Lacs (and FSRQ) are
519: considered
520: \footnote{The separation between LBL and HBL does not univocally
521: correspond to the selection band of BL Lacs (i.e. Radio selected BL
522: Lacs and X-ray selected BL Lacs, RBL and XBL).  However, there is a
523: sufficiently large overlap of RBL with LBL (and of XBL with HBL) that
524: we consider the values of spectral indices for RBL and XBL as
525: representative of the respective classes as defined from the SED peak
526: location.}.  Three out of the four radio galaxies (3C~31, 3C~264 and
527: 3C~465) lie in the area occupied by FSRQ and RBL.  Particularly
528: puzzling is the case of 3C~264. In fact this source is observed at a
529: small angle with respect to the jet axis, re-enforcing the view that
530: its nucleus should be rather similar to that of a BL Lac. Furthermore
531: its SED suggests that the putative synchrotron peak is located between
532: the optical and X--ray bands, analogously to HBL, in contrast with
533: its position in the spectral index planes. The fourth source, 3C~270,
534: is located on its own in the upper left corner of the diagrams.
535: 
536: However, as discussed in detail by Chiaberge et al. (2000)
537: relativistic beaming can have strong effects on the position of the
538: source in the $\alpha_{\rm ro}$--$\alpha_{\rm ox}$ plane.  Due to the
539: shift in energy caused by beaming, these quantitatively depend on the
540: shape of the SED if there are changes in the spectral slope (even in
541: the simplest case in which the degree of beaming does not depend on
542: energy).  The transformation law for broad band spectral indices,
543: derived in Chiaberge et al. (2000), is \\
544: \begin{displaymath}
545: \alpha_{\rm BL Lac} - \alpha_{\rm FR~I} = (\alpha_1-\alpha_2)
546: 	\frac{\log (\delta_{\rm BL Lac}/\delta_{\rm FR~I})} 
547:                     {\log (\nu_2/\nu_1)} \, ,
548: \end{displaymath}
549: where $\alpha_1$, $\alpha_2$ are the radio, optical or X-ray spectral
550: indices and $\delta_{\rm BL Lac}$, $\delta_{\rm FR~I}$ are the
551: relativistic beaming factors.
552: 
553: For 3C~264 and 3C~270, for which $\alpha$ in both the optical
554: and X-ray bands are measured,
555: we can determine ``beaming tracks'', i.e. the changes of the broad
556: band spectral indices for increasing Doppler boosting (Fig.~4),
557: adopting $\alpha_r = 0$. For
558: 3C~264, since the SED slope in the X--ray is steeper than in the
559: IR/optical and radio bands, blue-shifting the SED has the effect of
560: reducing all spectral indices. Its representative point thus moves
561: toward the origin in both the $\alpha_{\rm ox}$--$\alpha_{\rm ro}$ and
562: $\alpha_{\rm ox}$--$\alpha_{\rm rx}$ planes and the beaming track
563: crosses the region in which HBL are located.  Similarly, the track of
564: 3C~270 intercepts the region typical of LBL
565: \footnote{As dust absorption can cause an enhancement of the minimum
566: in the IR/optical region of the SED of 3C~270, we also estimated its
567: `beaming track' allowing for the presence of extinction. This turns
568: out to be similar to the case of no absorption.}. For both objects, the
569: inconsistency between SED shape and the values of the spectral indices
570: might be thus resolved, once the effects of beaming are properly taken
571: into account.
572: 
573: We note, however, that the beaming factor required in the case of
574: 3C~264 is ${\delta_{\rm BL Lac}}/{\delta_{\rm FR~I}} \sim 10 - 100 $
575: (the beaming correction for 3C~270 is heavily dependent on possible
576: effects of dust extinction and thus is not quantitatively reliable).
577: As the nuclear luminosity increases as $\delta^{p+\alpha}$ (with $p= 2
578: - 3$), when seen pole-on 3C~264 would be 3 - 6 orders of magnitude
579: brighter in the optical band (where $\alpha \sim 1$ and for $p = 2$).
580: This is only marginally consistent with the estimates of the ratio
581: between its luminosity and that of BL Lacs of similar extended
582: (isotropic) power, which are only $\sim 10^2 - 10^3$ times brighter
583: (CC99).  Furthermore, as discussed in the Introduction
584: the radio luminosity of BL Lacs appears to be
585: univocally connected with their SED. As the nuclear radio luminosity
586: of 3C~264 is $\nu L_{\nu} \sim 10^{39.9}$ erg s$^{-1}$, if seen
587: pole-on this would increase by $\sim 10^2 - 10^4$ (again for $p = 2$)
588: reaching values which are typical of LBL and higher than those seen in
589: HBL (Fossati et al. 1998).  Chiaberge et al. (2000) found a similar
590: problem when comparing the luminosities of complete samples of FR~I
591: and BL Lacs: FR~I appear to be overluminous with respect to what is
592: expected by debeaming the BL Lacs emission, modeled with a single
593: emitting component.  They suggested that this can be ascribed to a
594: velocity structure within the relativistic jet and that in FR~I the
595: emission is dominated by components/regions moving slower than those
596: seen in BL Lacs.
597: 
598: Summarizing, the SED of FR~I are qualitatively similar to those of BL
599: Lac objects, supporting the identification of these sources as their
600: mis-oriented counterparts, but their intensity does not seem to be
601: compatible with the simplest model in which we are observing a single
602: beamed emitting component in both classes of objects.
603: 
604: \subsection{X-ray emission and radio-galaxy orientation}
605: 
606: Following the method developed in CC99, we compare the luminosity of
607: FR~I nuclei with that of BL Lac objects in the X-ray band. In order to
608: restrict the comparison to objects which potentially belong to the
609: same region of the luminosity function of the FR~I/BL Lacs population,
610: we identify BL Lacs whose isotropic properties are similar to those of
611: our radio-galaxies.  More specifically, we consider objects whose
612: extended radio luminosity (data taken from Kollgaard et al. 1996)
613: differs by less than a factor of two from each of our FR~I.
614: \footnote{ With respect to CC99, we dropped the further requirement of
615: similar host galaxy magnitude since this does not appear to strongly
616: correlate with the nuclear properties neither in BL Lacs nor in
617: FR~I. This less restrictive choice enables us to increase the number
618: of BL Lac counterparts for each FR~I.}  We thus selected between 6 and
619: 10 `relatives' for each FR~I. Their X--ray fluxes (1 keV rest frame
620: energy, corrected for absorption) and spectral indices in the ROSAT
621: band are taken from Lamer et al. (1996) and Urry et al. (1996).  The
622: X-ray nuclear emission of FR~I is fainter than that of BL Lacs by a
623: factor which ranges from $\sim$ 10 up to 3$\times$10$^4$.
624: 
625: For a given FR~I, the X-ray luminosity ratios with respect to the
626: different `relative' BL Lacs present a very large dispersion, of 2 --
627: 3 orders of magnitude, and therefore, contrarely to what is found in
628: the optical band (CC99) we cannot draw any conclusion on whether a
629: trend with orientation is present.
630: 
631: \section{Summary and future perspectives} 
632: 
633: We used ROSAT and HST observations to isolate the emission originating
634: from the nuclei of 5 FR~I sources with the aim of studying their
635: Spectral Energy Distributions.  For the ROSAT PSPC observations, in
636: order to disentangle the nuclear non-thermal emission from that of the
637: host galaxy halo, we performed fits allowing for the presence of both
638: a thermal and a power--law component: in the case of 3C~264 the
639: emission is dominated by a power law spectrum; in 3C~31 and 3C~270 the
640: inclusion of the non--thermal component improves the fit, while this
641: is not required in 3C~465. A nuclear source is clearly seen in the
642: HST/NICMOS infrared images, available for 3 objects.
643: 
644: Combining the measurements in the X-ray and IR band with optical
645: nuclear luminosities derived from HST/WFPC2 images and radio data
646: taken from the literature, we have been able to build for the first
647: time the SED of FR~I nuclei.
648: 
649: These indicate the presence of a peak between the IR and the soft
650: X--ray bands, and the X--ray slopes are consistent with the broad band
651: indications.  For the two best constrained SED, it is also possible to
652: tentatively classify the FR~I nuclei similarly to what is done for BL
653: Lacs, in High and Low Energy Peaked sources.  This turns out to agree
654: with that derived from the comparison of the broad band spectral
655: indices once the changes in the SED due to the beaming are properly
656: taken into account.  However, the intensity of the nucleus of 3C~264
657: does not appear to be compatible with the simplest model in which we
658: are observing a single beamed emitting component in both FR I and BL
659: Lacs.  The SED of FR~I are thus qualitatively (but probably not
660: quantitatively) similar to those of BL Lacs.
661: 
662: The proposed approach of comparing the SED of the parent and beamed
663: population appears to be promising. Better data and energy coverage
664: are clearly required, particularly in the X--ray band, as the spectral
665: shape at these energies is essential in giving clues on the frequency
666: range where most of the synchrotron emission is released and thus in
667: `classifying' FR~I sources similarly to what is done for BL Lacs.
668: Both the AXAF and XMM missions will soon provide us with X--ray
669: spectra for large, and especially complete, samples of FR~I.
670: 
671: \section*{Acknowledgments} 
672: 
673: The authors would like to thank the referee for their useful comments
674: and suggestions.
675: The authors acknowledge the Italian MURST for partial financial support
676: under grant Cofin 98-02-32 and from the Italian Space Agency (ASI).
677: This research was supported in part by the National Science Foundation
678: under Grant No. PHY94-07194 (A. Celotti). We wish to thank I. Lehmann for 
679: providing the inflight calibrated PSF of the HRI, and W. Brinkmann for
680: allowing us to access to the RASS data on NGC 7052.
681: 
682: \section*{References} 
683: 
684: %\refitem Abraham R.G., McHardy I.M., Crawford C.S., 1991, MNRAS, 252,
685: %482
686: 
687: %\refitem Antonucci R.R.J., Ulvestad J.S., 1985, ApJ, 294, 158
688: 
689: \refitem Beuing J., D\"obereiner S., B\"ohringer H., Bender R., 1999, MNRAS 302, 209
690: 
691: \refitem Blandford R.D., Rees M.J., 1978, in BL Lac Objects,
692: A.N. Wolfe, ed., Univ. Pitt. Press (Pittsburgh), p.~328
693: 
694: \refitem Buote D.A., Fabian A.C., 1998, MNRAS 296, 977    
695: 
696: \refitem Bushouse H., Skinner
697: C.J., MacKenty J.W, 1997, NICMOS Instrument Science Report, 97-28
698: (Baltimore STScI)
699: 
700: \refitem Canosa C.M., Worrall D.M., Hardcastle M.J., Birinshaw M.,  
701: 1999, MNRAS, 310, 30
702: 
703: \refitem Capetti A., Celotti A., 1999, MNRAS, 304, 434 (CC99)
704: 
705: \refitem 
706: Cardelli, J.A., Clayton, G.C., Mathis, J.S. 1988, ApJ, 329, L33 
707: 
708: %\refitem Celotti A., Maraschi L., Ghisellini G., Caccianiga A.,
709: %Maccacaro T., 1993, ApJ, 416, 118
710: 
711: \refitem Chiaberge, M., Capetti A., Celotti A.,
712: 1999, A\&A 349, 77
713: 
714: \refitem Chiaberge, M., Celotti A., Capetti A., Ghisellini G.,
715: 2000, A\&A 358, 104
716: 
717: \refitem Davis D.S., Mushotzky R.F., Mulchaey J.S., Worrall D.M., 
718: Birkinshaw M., Burstein D., 1995, ApJ 444, 582
719: 
720: %\refitem di Serego Alighieri S., Danziger I.J., Morganti R., Tadhunter
721: %C.N., 1994, MNRAS, 269, 998
722: 
723: \refitem Donnelly H.L., Faber S.M., O'Connell R.M., 1990, ApJ 354, 52 
724: 
725: \refitem Edge A.C., Rottgering H., 1995, MNRAS, 277, 1580
726: 
727: \refitem Eskridge P.B., Fabbiano G., Dong-Woo K., 1995, ApJS 97, 141
728: 
729: \refitem Fabbiano G., 1989, ARAA 27, 87
730: 
731: \refitem Fabbiano G., Miller L., Trinchieri G., Lomgair M., Elvis M., 1984, ApJ 277, 115 
732: 
733: %\refitem Falomo R., Urry C.M., Pesce J.E., Scarpa R., Giavalisco M.,
734: %Treves A., 1997, ApJ, 476, 113
735: 
736: \refitem Fossati G., Maraschi L., Celotti A., Comastri A., Ghisellini
737: G., 1998, MNRAS, 299, 433
738: 
739: \refitem Fanaroff B.L., Riley J.M., 1974, MNRAS, 167, 31p
740: 
741: %\refitem Ghisellini G., Padovani P., Celotti A., Maraschi L., 1993,
742: %ApJ, 407, 65
743: 
744: \refitem Ghisellini G., Celotti A., Fossati G., Maraschi L., Comastri, A.,
745: 1998, MNRAS 301, 451 
746: 
747: 
748: \refitem Giommi P., Padovani P. , 1994, MNRAS, 268, L51
749: 
750: \refitem Hardcastle M.J., Worrall D.M., 1999, MNRAS, 309, 969
751: 
752: \refitem Harris D.E., Silverman J.D., Hasinger G., Lehman I., 1998, A\&AS 133, 431
753: 
754: \refitem Impey C., Gregorini L., 1993, AJ, 105, 853
755: 
756: \refitem Jaffe W., Ford H.C., Ferrarese L., Van den Bosch F.,
757: O'Connell R.,W.  1993, Nat, 364, 213
758: 
759: \refitem Ho L.C. 1999, ApJ 516, 672 
760: 
761: %\refitem Kollgaard R.I., Wardle J.F.C., Roberts D.H., Gabuzda D.C.,
762: %1992, AJ, 104, 1687
763: 
764: \refitem Kollgaard R.I., Palma C., Laurent-Muehleisen 
765: S.A., Feigelson E.D. 1996, ApJ 465, 115 
766: 
767: \refitem Komossa S., B\"ohringer H., 1999, A\&A 344, 755
768: 
769: \refitem Lamer G., Brunner H., Staubert R., 1996, A\&A, 311, 384
770: 
771: \refitem MacKenty J.W., et al.,
772: 1997, NICMOS Instrument Handbook, Version 2.0 (Baltimore STScI)
773: 
774: \refitem Matsumoto H., Koyama K., Awaki H., Tsuru T., Loewenstein M., Matsushita K., 1997, ApJ 482, 133
775: 
776: \refitem Morrison R., McCammon D., 1983, ApJ, 270, 119
777:    
778: \refitem Morse J.A., 1994, PASP 106, 675
779: 
780: \refitem Padovani P. , Giommi P., 1995, ApJ, 444, 567
781: 
782: \refitem Padovani P. , Giommi P., 1996, MNRAS, 279, 526
783: 
784: %\refitem Padovani P., Urry C.M., 1990, ApJ, 356, 75
785: 
786: %\refitem Padovani P., Urry C.M., 1991, ApJ, 368, 373
787: 
788: \refitem Perlman E.S. , Stocke J.T., 1993, ApJ, 406, 430
789: 
790: \refitem Raymond J.C., Smith B.W., 1977, ApJS 35, 419 
791: 
792: \refitem Sakelliou I., Merrifield M.R., 1998, MNRAS 304, 434
793: 
794: \refitem Sambruna R.M., Maraschi L., Urry C.M., 1996, ApJ, 463, 444
795: 
796: \refitem Sambruna R.M., Eracleous M., Mushotzky R.F., 1999, ApJ, 526, 60
797: 
798: \refitem Sarazin C.L., 1986, Rev. Mod. Phys. 58, 1
799: 
800: \refitem Sarazin C.L., Koekemoer A.M., Baum S.A., O'Dea C.P., 
801: Owen F.N., Wise M.W., 1999, ApJ 510, 90
802: 
803: \refitem Schindler S., Prieto M.A., 1997, A\&A 327, 37
804:              
805: %\refitem Stickel M., K\"uhr H., 1993, A\&A, 98, 39
806: 
807: %\refitem Stickel M., Padovani P., Urry C.M., Fried J.W., K\"uhr H.,
808: %1991, ApJ, 374, 431
809: 
810: \refitem Tananbaum H., Tucker W., Prestwich A., Remillard R., 1997, ApJ 476, 83
811: 
812: \refitem Trussoni E., Massaglia S., Ferrari R., Fanti R., Feretti L.,
813: Parma P., Brinkmann W., 1997, A\&A 327, 27
814: 
815: \refitem
816: Trussoni E., Vagnetti F., Massaglia S., Feretti L., Parma P., Morganti R., 
817: Fanti R. Padovani P., 1999a, A\&A 348, 437
818: 
819: \refitem
820: Trussoni E., Feretti L., Capetti A., Celotti A. Chiaberge M., 1999, Ap. Lett.
821: Comm., in press 
822: 
823: %\refitem Ulrich M-H, 1989, in BL Lac Objects: 10 years after, Maraschi
824: %L., Maccacaro T. \& Ulrich M.-H. eds, Springer-Verlag, p. 45
825: 
826: \refitem Urry C.M., Padovani P., 1995, PASP, 107, 803
827: 
828: \refitem Urry, C.M., et al., 1996, ApJ, 463, 424
829: 
830: %\refitem Urry C.M., Padovani P., Stickel M., 1991, ApJ, 382, 501
831: 
832: \refitem Worrall D.M., Birkinshaw M., 1994, ApJ, 427, 134
833: 
834: \refitem Worral D.M., 1997, in Relativistic Jets in AGN's, M. Ostrowski, 
835: M. Sikora, G. Madjeski, M. Begelman eds., Astron. Obs. of the Jagiellonian Univ., Krakow, p. 20
836: 
837: \refitem Zimmermann H.U., Becker W., Belloni T., D\"obereiner S., Izzo C., 
838: Kahabka P., Schwentker O., 1995, EXSAS User's GUide, ed. 5, MPE Rep. n 257
839: 
840: 
841:  
842: \begin{table*}
843: \caption{Main optical parameters}
844: \begin{tabular}{|l c c c c c |} \hline
845: Source   & Alt. name & z & $M_{B_{T_o}}$ & Envir. & Orient.$^{\rm a}$\\
846: \hline
847: 3C~31      & NGC~383  & 0.0169  & -21.19   &   Arp~331  & 35$^\circ$ \\
848: 3C~264     & NGC~3862 & 0.0216  & -21.15   &   A~1367   & 15$^\circ$ \\
849: 3C~270     & NGC~4261 & 0.0074  & -21.02   &   Group    & 65$^\circ$ \\
850: 3C~465     & NGC~7720 & 0.0291  & -22.30   &   A~2634   & 45$^\circ$ \\
851: NGC~7052   & B2 2116+26  & 0.0164  & -21.41   &   Pair  & 72$^\circ$ \\
852: \hline 
853: \end{tabular} 
854: \vskip 0.1 true cm 
855: \noindent 
856: $^{\rm a}$ Radio source orientation taken from CC99.
857: \end{table*}
858: 
859: \begin{table*}
860: \caption{Log of ROSAT observations for the four FR~I radio galaxies}
861: \begin{tabular}{|l c c c c c c c |} \hline
862: Source   & Instrument & Obs. Date &  t$_{\rm exp}$ (s) & Counts$^{\rm a}$ 
863: & $r^{{extr}}_{source}$ & $r^{{extr}\,{\rm (c)}}_{bkg}$  &   Ref.$^{\rm d}$ 
864: \\  \hline
865: 3C~31    & PSPC  & Jul 91   & 29430 &   $552 \pm 38$  & $75^{\prime \prime}$
866:          &  $125^{\prime \prime}$ &   (1)(2)\\
867:          & HRI   & Apr 94   & 25007 &   $165 \pm 33$  &  
868:          &    &  (2)(3) \\
869: 3C~264   & PSPC  & Nov 91   & 18745 &  $5068 \pm 85$  & $100^{\prime \prime}$
870:          &  $175^{\prime \prime}$ &   (3)(4)(5)\\
871: 3C~270   & PSPC  & Dec 91   & 21863 &  $1935 \pm  65$  & $150^{\prime \prime}$
872:          &  $250^{\prime \prime}$ &   (6) \\
873:          & HRI   & Jun 95   & 18477 &   $584 \pm 51$   &  
874:          &  &      \\
875: 3C~465   & PSPC  & Jun 91   &  9866 &      &              
876:          &       &       \\
877:          & PSPC  & Dec 92   & 10522 &   $252 \pm 21^{\rm b}$ & 
878:            $35^{\prime \prime}$ &  $75^{\prime \prime}$ &  (7)  \\
879:          & HRI   & Jan 95   & 29063 &      &
880:          &       &        \\
881:          & HRI   & Jun 95   & 33392 &   $304 \pm 29^{\rm b}$  &
882:              &   &  (3)(8)  \\
883: \\
884: \hline
885: \end{tabular}
886: \vskip 0.1 true cm 
887: \noindent 
888: $^{\rm a}$ Background subtracted; the errors are at $1 \sigma$
889: 
890: \noindent
891: $^{\rm b}$ Merging the data of the two observations
892: 
893: \noindent
894: $^{\rm c}$ External radius, the inner radius $\equiv
895: r^{extr}_{source}$
896: 
897: \noindent
898: $^{\rm d}$ References: (1) Trussoni et al. (1997); (2) Komossa \&
899: B\"ohringer (1999); (3) Hardcastle \& Worrall (1999); (4) Edge \& R\"ottering 
900: (1995); (5)Tananbaum et al. (1997);
901: (6) Worral \& Birkinshaw (1994); (7) Schindler \& Prieto
902: (1997); (8) Sakelliou \& Merrifield (1997)
903: 
904: \end{table*}
905: 
906: \begin{table*}
907: \caption{Fits  of PSPC data with a thermal   +  power-law spectrum.
908: Errors are at $1 \sigma$. Upper and lower limits are at $2 \sigma$ 
909: confidence level. Quantities without errors are fixed. 
910: Metallicity is always set to $\mu=$ 0.5 solar.}
911: \begin{tabular}{|l c c c c c c  |} \hline
912: Source   & $N_{H}$ &  $T$  & 
913: $\Gamma$ & $\chi^2$ (d.o.f.) & $L_{\rm [0.1-2.4 \, keV],th}$   &  
914: $L_{\rm [0.1-2.4 \,keV],pl}$  \\
915:          &$\times 10^{20}$ cm$^{-2}$& keV  & 
916: &    &   $\times 10^{41}$ erg s$^{-1}$ 
917: &  $\times 10^{41}$ erg s$^{-1}$  \\
918: 
919: \hline
920: 3C~31    &  5.23 &  0.75$^{+0.17}_{-0.20}$   &    2.3 & 0.30 (15) 
921:          & $0.65^{+0.17}_{-0.16}$   &  $1.30^{+0.54}_{-0.58}$   \\
922:          &  5.23 &  0.59$^{+0.17}_{-0.20}$   &     1.7 & 0.30 (15) 
923:          &   $0.70^{+0.20}_{-0.22}$ &  $1.00^{+0.38}_{-0.23}$   \\
924: \hline
925: 3C~264   & $2.6^{+0.4}_{-0.2}$ & 1  &  $2.46^{+0.06}_{-0.06}$ 
926:          & 0.88 (81) & $1.00^{+1.17}_{-1.00}$ &  $47.1^{+0.9}_{-1.3}$   \\
927: \hline
928: 3C~270   & 1.63 &  $0.81^{+0.09}_{-0.13}$ &  $1.71^{+0.21}_{-0.18}$ 
929:          & 0.70  (33)   & $0.49^{+0.08}_{-0.06}$   &  $0.54^{+0.1}_{-0.1}$ \\
930: \hline
931: 3C~465   &  5.22 &  $1.16^{+0.20}_{-0.15}$ &    2.3  & 1.10 (17)
932:          &  $3.20^{+0.27}_{-0.76}$ ($\gtrsim 2.2$) &  0 ($\lesssim 1.3$)  \\
933:          &  5.22 &  $1.16^{+0.14}_{-0.18}$ &    1.7  & 1.09 (17)
934:          & $3.07^{+0.36}_{-0.49}$ ($\gtrsim 1.2$)  &  0 ($\lesssim 2.1$)   \\
935: \hline
936: \end{tabular}
937: \vskip 0.1 true cm
938: \noindent
939: \end{table*}
940: 
941: \begin{table}
942: \caption{Log of NICMOS observations}
943: \begin{tabular}{|l c c c c|} \hline
944: Source   & Obs. Date & Filter &  t$_{\rm exp}$ (s) & 
945: Flux$^{\rm a}$ \\ 
946: \hline
947: 3C~264     & May 98 & F110W & 112   & 6.7  \\
948:            &        & F160W & 112   & 4.3 \\
949:            &        & F205W & 112   & 2.6 \\
950: 3C~270     & Apr 98 & F110W & 192   & 0.69 \\
951:            &        & F160W & 192   & 1.2 \\
952:            &        & F205W & 192   & 1.6 \\
953: NGC~7052   & Aug 98 & F160W & 160   & 0.06  \\
954: \hline
955: \end{tabular}
956: \vskip 0.1 true cm
957: $^{\rm a}$Fluxes in units of 10$^{-17}$~erg~s$^{-1}$~cm$^{-2}$~\AA$^{-1}$ 
958: \end{table}
959: 
960: 
961: \begin{table*}
962: \caption{FR~I nuclear luminosities$^{\rm a}$ and spectral indeces}
963: \begin{tabular}{@{}lccccccc}\hline
964: Name  & L$_{\rm R}$ & L$_{\rm IR}$ & L$_{\rm O}$ & L$_{\rm X}$ 
965: & $\alpha_{\rm RO}$ & $\alpha_{\rm RX}$ & $\alpha_{\rm OX}$ \\
966: \hline
967: 3C~31     & 39.57  &  ---   & 40.63 &     40.49  
968: & 0.79 $\pm$ 0.03  &  0.88 $\pm$ 0.05 & 1.05 $\pm$ 0.12 \\ 
969: 3C~264    & 39.94 & 41.77 & 41.81 &     42.01  
970: & 0.63 $\pm$ 0.02 &  0.73 $\pm$ 0.02 &  0.93 $\pm$ 0.05 \\ 
971: 3C~270    & 38.96 & 40.30 & 39.08 &     40.32  
972: & 0.98 $\pm$ 0.04 &  0.82 $\pm$ 0.04 &  0.54 $\pm$ 0.09 \\ 
973: 3C~465    & 40.34 &  ---   & 41.21 & $<$40.91   
974: & 0.83 $\pm$ 0.02 &  $>$ 0.92 & $>$ 1.11  \\ 
975: NGC~7052  & 39.07 & 39.72 & 40.04 &       --   
976: & 0.81 $\pm$ 0.06 &  --- &  --- \\ 
977: \hline
978: \end{tabular}
979: \vskip 0.1 true cm
980: $^{\rm a}$ 
981: Luminosities, log ($\nu L_{\nu}$), in erg s$^{-1}$ measured
982: at 5 GHz, 1.6 $\mu$m, 0.5 $\mu$m and 1 keV. 
983: \end{table*}
984: 
985: %Figure 1a:
986: \begin{figure*}[t]
987: \centerline{\psfig{figure=3c31pr.ps,height=10.0truecm,angle=0}
988: \psfig{figure=3c270pr.ps,height=10.0truecm,angle=0}}
989: \caption{Radial profiles of 3C~31 and 3C~270 (left and right panel,
990: respectively), compared to the in--flight calibrated PSF (dotted line)
991: normalized to the source intensity and background level.}
992: \end{figure*}
993: 
994: %Figure 2
995: \begin{figure*}[t]
996: \centerline{\psfig{figure=fig.eps,height=9.0truecm,angle=270}}
997: \caption{HST/NICMOS F160W images of 3 FR~I radio galaxies. Note the
998: presence of the central nuclear sources and of the dusty disc in
999: NGC~7052. The arrows indicate the direction of the radio jet.}
1000: \end{figure*}
1001: 
1002: %Figure 3
1003: \begin{figure}[t]
1004: \centerline{\psfig{figure=sed.ps,height=7.0truecm,angle=0}}
1005: \caption{Spectral energy distribution of FR~I nuclei. 
1006: Errors are smaller than the symbol size, except when marked.}
1007: \end{figure}
1008: 
1009: %Figure 4
1010: \begin{figure}[t]
1011: \centerline{\psfig{figure=tracks.ps,height=7.0truecm,angle=0}}
1012: \caption{Broad band spectral indices of the FR~I nuclei (large stars)
1013: compared to those of blazars (circles are XBL, crosses
1014: and squares are RBL and FSRQ, respectively), as from Fossati et al. (1998). 
1015: Continuous lines
1016: represent the variations of spectral index due to the shift in energy
1017: of the Doppler boosting. Ticks on the lines correspond to a increase of
1018: $\delta$ by 10 and 100. The dashed line follows the beaming track of 3C~270
1019: after a correction for absorption corresponding to $A_R = 2$.}
1020: \end{figure}
1021: 
1022: \end{document} 
1023: 
1024:  
1025:               
1026: