astro-ph0109516/ms.tex
1:  \documentclass[11pt,preprint]{aastex}
2: %\singlespace
3:  %% manuscript produces a one-column, double-spaced document:
4: 
5:  % \documentclass[manuscript]{aastex}
6: 
7:  %% preprint2 produces a double-column, single-spaced document:
8: 
9:  %\documentclass[preprint2]{aastex}
10: 
11: 
12:  \newcommand{\feii}{\ion{Fe}{2}}
13:  \newcommand{\kms}{km\,s$^{-1}$}
14:  \newcommand{\lb}{$\lambda$}
15:  \newcommand{\hbeta}{H$\beta$}
16: 
17:  
18: 
19:  \slugcomment{Accepted to ApJ.}
20: 
21:  
22:  \shorttitle{Infrared Fe\,II Emission in NLS1 Galaxies}
23:  \shortauthors{Rodr\'{\i}guez-Ardila et al.}
24: 
25:  
26:  \begin{document}
27: 
28:  
29:  \title{Infrared \feii\ Emission in Narrow-Line Seyfert 1 Galaxies}
30: 
31: 
32: 
33:  \author{A. Rodr\'{\i}guez-Ardila\altaffilmark{1} and S. M. Viegas}
34:  \affil{Instituto Astron\^omico e Geof\'{\i}sico - Universidade de S\~ao
35:  Paulo,
36:  Av. Miguel Stefano 4200, CEP 04301-904, S\~ao Paulo, SP, Brazil}
37: 
38:  \author{M. G. Pastoriza}
39:  \affil{Departamento de Astronomia - UFRGS. Av. Bento Gon\c calves 9500,
40:  CEP 91501-970, Porto Alegre, RS, Brazil}
41:  
42: 
43:  \and
44: 
45:  \author{L. Prato\altaffilmark{1}}
46:  \affil{Department of Physics and Astronomy, UCLA, Los Angeles, CA
47:  90095-1562}
48: 
49:  \altaffiltext{1}{Visiting Astronomer at the Infrared Telescope facility, which
50: is operated by the University of Hawaii under contract from the National 
51: Aeronautics and Space Administration}
52: 
53: 
54:  
55:  \begin{abstract}
56:  We obtained 0.8-2.4 $\mu$m spectra at a 
57:  resolution of 320 km\,s$^{-1}$ of four narrow-line Seyfert
58:  1 galaxies in order to study the near-infrared properties of 
59:  these objects. We focus on the analysis of the \feii\ emission
60:  in that region and the kinematics of the low-ionization broad lines.
61:  We show that the 1$\mu$m \feii\ lines (\lb9997, 
62:  \lb10501, \lb10863 and \lb11126) are the strongest \feii\
63:  lines in the observed interval. 
64:  For the first time, primary cascade lines of \feii\ arising from 
65:  the decay of upper levels pumped by Ly$\alpha$ fluorescence are  
66:  resolved and identified in active galactic nuclei.
67:  Excitation mechanisms leading to the emission of the 1$\mu$m \feii\
68:  features are discussed. A combination of Ly$\alpha$ fluorescence
69:  and collisional excitation are found to be the main contributors. 
70:  The flux ratio between near-IR \feii\ 
71:  lines varies from object to object, in contrast to what 
72:  is observed in the optical region. A good correlation
73:  between the 1$\mu$m and optical \feii\
74:  emission is found. This suggests that the upper $z^{\rm 4}F^{\rm 0}$ and 
75:  $z^{\rm 4}D^{\rm 0}$ levels from which the bulk of the optical
76:  \feii\ lines descend are mainly populated by the transitions leading
77:  to the 1$\mu$m lines. The width and profile shape of \feii\ \lb11127, 
78:  \ion{Ca}{2} \lb8642 and \ion{O}{1} \lb8446 are very similar but 
79:  significantly narrower than Pa$\beta$, giving
80:  strong observational support to the hypothesis that the 
81:  region where \feii, \ion{Ca}{2} and \ion{O}{1}
82:  are produced are co-spatial, interrelated kinematically and most 
83:  probably located in the outermost portion of the BLR.  
84: 
85: 
86:  \end{abstract}
87: 
88:  
89:  \keywords{galaxies: active -- galaxies: Seyfert -- 
90:  infrared: galaxies }
91: 
92:  
93:  \section{Introduction}
94: 
95:  Narrow-line Seyfert 1 galaxies (hereafter NLS1s) are a 
96:  peculiar group of Seyfert 1 objects, recognized by 
97:  \citet{op85} for their optical properties: 
98:  (a) the broad components of the permitted lines have FWHM $<$ 
99:  2000 \kms, narrower that those of the classical Seyfert 
100:  1s (i.e $\sim$4000 \kms), (b) the [\ion{O}{3}]/\hbeta\ ratio is 
101:  less than 3, and (c) usually, strong \feii\ emission dominates 
102:  the optical and ultraviolet spectrum. In the 
103:  soft and hard X-ray bands NLS1s also seem to have 
104:  peculiar properties. In most cases they are 
105:  characterized by a very steep spectrum and extreme 
106:  and rapid variability \citep{bbf96, ley99}. Up to now, 
107:  no single model can explain all of the observed properties 
108:  of these objects and much effort has recently been made 
109:  in order to understand their central engines. Variability 
110:  studies \citep{pet2000} point out that the ultimate cause 
111:  for their peculiar behavior may be attributable to a smaller black 
112:  hole mass ($\sim 10^{6}$ M$\odot$) together with a  
113:  near- or super-Eddington accretion rate.
114: 
115:  NLS1 galaxies are ideal objects for studying many features 
116:  which are clearly separated in their spectra but are 
117:  heavily blended in classical Seyfert 1s because of the broadness 
118:  of their permitted lines. Such is the case with the \feii\ 
119:  emission lines, which produce a ubiquitous pseudo-continuum 
120:  from the ultraviolet through the near-infrared (near-IR).
121:  This pseudo-continuum 
122:  carries a large amount of energy from the BLR
123:  and is not yet well
124:  understood. In the optical and UV regions, the \feii\
125:  lines have been well studied \citep{wnw85, jol91, lip93}. 
126:  The main processes believed to contribute to the bulk of that
127:  emission are continuum fluorescence
128:  via UV resonance lines, self-fluorescence via overlapping
129:  \feii\ transitions and collisional excitation. Nonetheless,
130:  most models that include these processes cannot account 
131:  for the observed strength of the \feii\ lines. \citet{pen87},
132:  \citet{sp98} and \citet{ver99} considered that fluorescent 
133:  excitation by Ly$\alpha$ may be important
134:  in understanding the \feii\ 
135:  spectrum in active galactic nuclei (AGNs), from the 
136:  ultraviolet to the IR. Hovewer, the lack of suitable 
137:  data in the latter region has hindered tests of
138:  predictions. 
139:   
140:  In the near-IR region, \feii\ emission lines have been mostly
141:  observed in a variety of Galactic emission-line stars
142:  \citep{hp89, rudy91, ham94}.  At least in these sources,
143:  the \feii\ lines redward of 7000 \AA\ can be attributed
144:  to cascades from excited states which are pumped by
145:  resonant absorption of \ion{H}{1} Ly$\alpha$. In all
146:  cases, the strongest \feii\ lines are the ones located
147:  at $\lambda$9997, $\lambda$10171, $\lambda$10490,
148:  $\lambda$10501, $\lambda$10863 and $\lambda$11126,
149:  emitted after the decay of the common upper term ($b~^{\rm 4}G$). 
150:  Because of their close proximity in wavelength, they are termed
151:  the 1$\mu$m \feii\ lines. Other important \feii\ lines
152:  also detected in Galactic sources are located in
153:  the 8500$-$9300 \AA\ interval. They are primary cascade lines
154:  descending from the upper 5p levels to the lower
155:  $e^{\rm 4}D$ and $e^{\rm 6}D$ terms. With the advent of 
156:  a new generation of IR detectors with greater sensitivity 
157:  and resolution in this region, it is now possible to 
158:  search for these lines in AGNs, mainly in NLS1s, and
159:  test the validity of the models of \feii\ emission. This 
160:  is a crucial step in understanding the physical processes 
161:  at work in the central engines of these objects.
162: 
163:  The only detection to date of the 1$\mu$m 
164:  \feii\ lines in an AGN was reported by 
165:  \citet[hereafter RMPH]{rudy00} in their 0.8 -- 2.5$\mu$m 
166:  spectrophotometric observation of I\,Zw\,1, the 
167:  archetypical NLS1 galaxy. Their data clearly reveal \feii\
168:  $\lambda$9997, $\lambda$10501, $\lambda$10863 and 
169:  $\lambda$11126. Based on the absence of the crucial cascade 
170:  lines that feed the common upper state where the
171:  1$\mu$m \feii\ lines originate (assuming Ly$\alpha$ fluoresence as the
172:  dominant mechanism), as well as the relatively low energy of that
173:  state, RMPH suggest that the observed lines are collisionally 
174:  excited. No individual identification of \feii\ lines
175:  in the 8500$-$9300 interval has yet been done in Seyferts. 
176:  But it is believed that the strong broad feature, centered at
177:  \lb9220 \AA\ and observed in many AGNs is, in part,
178:  due to \feii\ \citep{morward89}.
179:  
180:  In this paper we present mid-resolution ($\sim$ 320 km\,s$^{-1}$),
181:  near-IR spectroscopy of four NLS1 galaxies (1H1934-063, Ark\,564, 
182:  Mrk\,335, and Mrk\,1044), not previously observed in
183:  this spectral region, whose optical spectra reveal
184:  strong to moderate \feii\ emission. In \S 2 we present
185:  the observations, data reduction, and a brief description of
186:  the most conspicuous spectroscopic features observed in each
187:  galaxy. The excitation mechanism
188:  of the \feii\ lines and the observed relative intensity are
189:  discussed in \S 3. The kinematics of the BLR derived from
190:  the low-ionization lines is treated in \S 4. Conclusions
191:  are presented in \S 5. 
192: 
193: 
194:  \section{Observations, Data Reduction and Results} \label{observa}
195: 
196:  \subsection{Observations}
197: 
198:  The data were obtained at the NASA 3m Infrared Telescope
199:  Facility (IRTF) on 2000 October 11 and 13 (UT) 
200:  with the SpeX facility spectrometer \citep{rayner98, rayner01}.  
201:  In the short wavelength
202:  cross-dispersed mode, SpeX provides nearly continuous coverage
203:  from 8200 \AA~ through 24000 \AA.  Ten minute integrations,
204:  nodding in an off-on-on-off source pattern, were combined to
205:  produce final spectra.  Typical total integration times ranged
206:  from 30$-$40 minutes.  A0V stars were observed
207:  near targets to provide telluric standards at similar airmasses.
208:  Seeing was $\sim$1.$''$0.  An 0.$''$8 slit yielded an
209:  instrumental resolution of 320 \kms. 
210: 
211:  \subsection{Data Reduction}
212: 
213:  The spectral extraction and wavelength calibration procedures were
214:  carried out using Spextool, the in-house software developed 
215:  and provided by the SpeX team for the IRTF 
216:  community \citep{cvr01}\footnote{Spextool is available from the
217:  IRTF website: http://irtf.ifa.hawaii.edu/Facility/spex/spex.html}.
218:  Each on-off source pair of observations was reduced individually
219:  and the results summed to provide a final spectrum.  A 1$''$ or
220:  3$''$ (in the case of 1H\,1934-063) aperture was used to extract the
221:  spectra. Observations of an argon arc lamp enabled wavelength calibration
222:  of the data; the RMS of the dispersion solution for the four target
223:  galaxies was $\sim$2$\times$10$^{-5}\mu$m.
224: 
225:  Wavelength calibrated final target spectra were divided
226:  by AOV stars observed at a similar airmasses.  However, even small
227:  differences in airmass between the target galaxies
228:  and standard stars resulted in residuals, primarily evident in the
229:  spectrum of Mrk 1044.  The spectra were each multiplied
230:  by a black body function corresponding to 10,000 K in order to
231:  restore the true continuum shape of the targets.
232: 
233:  Spectra were flux calibrated by normalizing to the $K$ band
234:  magnitude, 6.303 mag, of one of the observed A0V standard stars, 
235:  HD\,182761. We used the task STANDARD of the ONESPEC package of 
236:  IRAF\footnote{IRAF is distributed by NOAO, which is operated
237:  by AURA Inc., under contract to the NSF.} for this 
238:  purpose. In this process, a blackbody flux 
239:  distribution based on the magnitude and spectral
240:  type of the star was employed. The error in the flux calibration 
241:  is $\sim$10\%, estimated from comparison with an overlapping 
242:  visual wavelength, red spectrum of 1H1934-063 at 0.8$-$0.95 $\mu$m 
243:  and previously published IR spectrophotometry for 
244:  MRK\,335 \citep{rudy82}. Nonetheless, this calibration is
245:  not absolute. For this reason, the uncertainties quoted in this paper
246:  reflects the errors derived from the S/N of the spectra. 
247: 
248:  We correct the
249:  spectra for Galactic extinction, as determined from the 
250:  {\it COBE/IRAS} infrared maps of \citet{sfd98}.
251:  The value of the Galactic {\it E(B-V)} as well as basic 
252:  information for each object is listed in 
253:  Table~\ref{tabledata}. Finally, each spectrum
254:  was shifted to rest wavelength. The value of $z$ adopted
255:  was determined by averaging the redshift measured from the 
256:  strongest lines, usually \ion{O}{1} \lb8446, [\ion{S}{3}]
257:  \lb9531, \feii\ \lb9997, Pa$\delta$, \ion{He}{1} \lb10830,
258:  \ion{O}{1} \lb11287, Pa$\beta$, Pa$\alpha$ and Br$\gamma$.
259:  In all cases, the radial velocities were in very good 
260:  agreement with the values reported in the literature.
261: 
262: 
263:  \subsection{Results}
264: 
265:  The resulting extracted spectra in the wavelength 
266:  interval 0.8 -- 2.4$\mu$m, corrected for
267:  redshift, are displayed in Figures~\ref{fig1} -- \ref{fig4}.  
268:  Identification of the known emission lines are 
269:  given for each spectrum. Prominent lines of \feii,
270:  \ion{H}{1}, \ion{He}{1}, \ion{O}{1} and
271:  \ion{Ca}{2} are observed in the four galaxies,
272:  similar to those present in I\,Zw\,1. In addition, 
273:  \ion{He}{2} \lb10124, as well as lines of
274:  [\ion{S}{3}] $\lambda\lambda$9070, 9531, 
275:  [\ion{S}{8}] \lb9912, [\ion{S}{9}] \lb12523, 
276:  [\ion{Fe}{2}] \lb12567, [\ion{Si}{10}] \lb14305 
277:  and [\ion{Si}{6}] \lb19630, not present in I\,Zw\,1,
278:  are clearly detected in our data. A 
279:  detailed analysis of the narrow emission lines is 
280:  beyond the scope of this paper and will be discussed 
281:  in a future publication \citep{roar01a}. 
282: 
283:  In order to measure the flux of the emission lines
284:  we have assumed that they can be represented by a
285:  single or a sum of Gaussian profiles. The 
286:  LINER routine (Pogge \& Owen 1993), a $\chi^{2}$
287:  minimization algorithm that fits as many as eight
288:  Gaussians to a profile, was
289:  used for this purpose. Table~\ref{tableflux} list 
290:  fluxes and FWHM for the detected
291:  \feii\ and other permitted lines. The FWHM values were
292:  corrected for instrumental broadening ($\sim$ 320 \kms).
293:  Except for \feii\ \lb10501\ and
294:  \lb11127, which are isolated features, the 
295:  \feii\ lines are blendend with \ion{He}{1} 
296:  $\lambda$10830 plus Pa$\gamma$ (\feii\ $\lambda$10863) 
297:  or \ion{He}{2} $\lambda$10124 plus Pa$\delta$ 
298:  (\feii\ $\lambda$9956, $\lambda$9997 
299:  and $\lambda$10171). For consistency, in these latter
300:  cases the \feii\ and Paschen lines were constrained to 
301:  have the same width as those derived for \feii\ $\lambda$11127
302:  and Pa$\beta$, respectively. Nonetheless, it is interesting
303:  to note that even if no constraint was imposed, the
304:  resulting line widths were in agreement with those found
305:  for the isolated lines. 
306: 
307:  The above procedure also allowed us to separate the
308:  contribution of the NLR in those lines that are emitted 
309:  both in the BLR and the NLR. The results show that 
310:  in the Paschen lines, for instance, nearly half the total 
311:  flux is emitted by the NLR, very similar to what was found
312:  by \citet{ro00} for the Balmer lines. Table~\ref{tablefluxbn}
313:  list the fluxes and FWHM of the Pashen  and \ion{He}{1}
314:  \lb10839 lines that could be deblended into a narrow and
315:  broad component. All the values of FWHM were corrected for
316:  instrumental broading. In Mrk\,1044 it was not possible to find
317:  a consistent solution for Pa$\delta$, Pa$\gamma$ and \ion{He}{1}
318:  \lb10839 due to the lower S/N of the spectrum. For this
319:  reason, the total fluxes associated with each of the latter three
320:  lines are reported but must be understood as the sum of
321:  the contribution of both the NLR and BLR.
322: 
323:  The presence of emission from the NLR is demonstrated by
324:  the observation of strong [\ion{S}{3}] \lb9068, \lb9531.
325:  In order to test the validity of the Gaussian 
326:  approximation, we used the technique applied by
327:  \citet{ro00} to separate the narrow and broad flux in 
328:  Pa$\beta$, consisting of using
329:  the profile of a line known to be emitted exclusively in 
330:  the NLR as a template. In the current work, we assumed that the width
331:  of the narrow component of the permitted lines would be 
332:  similar to that of [\ion{S}{3}] \lb9531, the strongest
333:  forbidden line observed.  Based on this assumption, the
334:  [\ion{S}{3}] \lb9531 line was scaled to the peak
335:  intensity of a given permitted line and then subtracted off
336:  in different proportions until the residuals consist 
337:  of a pure broad component with no absorption superimposed.
338:  This procedure has the advantage of being free from
339:  any assumption regarding the form of the line profiles.
340:  The results were almost identical 
341:  to those obtained using the Gaussian approximation. 
342:  Figure~\ref{deblend} shows examples of the deblending
343:  of Pa$\beta$ and the region around Pa$\delta$ for 
344:  Ark\,564.
345: 
346:  In the rest of this section we will give 
347:  brief descriptions of the most important features
348:  observed in the NLS1 spectra.
349: 
350:  \subsubsection{1H\,1934-063}
351: 
352:  1H\,1934-063 is a relatively poorly known NLS1
353:  but an excellent laboratory to study many emission
354:  line processes. Its optical spectrum, described by
355:  \citet{ro00}, is full of both permitted and forbidden
356:  lines. High ionization lines such as [\ion{Fe}{10}]
357:  and [\ion{Fe}{14}] are very prominent. In the
358:  near-IR it also displays a very 
359:  rich spectrum, both in permitted and forbidden lines 
360:  (see Figure~\ref{fig1}). \ion{He}{1} \lb10830, 
361:  Pa$\alpha$, Pa$\beta$, Pa$\delta$, \ion{O}{1} \lb8446, 
362:  \lb11287, the \ion{Ca}{2} triplet in emission and
363:  the 1$\mu$m \feii\ lines are strong. In addition,
364:  Br$\gamma$ and Br$\delta$ are clearly observed. A very
365:  conspicous blend of low-ionization species, centered
366:  in 9220 \AA\ is present. We identified \feii\ \lb9177, \lb9202
367: and \lb9256 as well as Pa9 \lb9227 as contributors to this
368: feature. \ion{Mg}{2} \lb9218, \lb9244 are probably present
369: but heavily blended with the Pa9. The continuum emission is very steep, 
370:  and follows a simple power-law form.
371:  High-ionization, forbidden lines of [\ion{S}{8}],
372:  [\ion{S}{9}], [\ion{Si}{10}], [\ion{Si}{7}] and
373:  [\ion{Ca}{8}] are clearly visible as well as a
374:  blip at the expected position of the 
375:  H$_{2}$ (1,0)S(1) \lb2.122$\mu$m.  [\ion{S}{3}]
376:  \lb9531 is, by far, the strongest forbidden line
377:  detected.
378: 
379: 
380:  \subsubsection{Ark\,564}
381: 
382:  Ark\,564 is a widely known and well studied NLS1 galaxy,
383:  mainly in the UV and X-ray region \citep{bbf96, cren99, 
384:  com00, ball00}. It is the brightest NLS1 in the 2-10 keV range; the
385:  {\it ROSAT} and {\it ASCA} observations reveal a complex
386:  X-ray spectrum \citep{vaug99}. In the near-IR, this
387:  object is also very rich in emission lines 
388:  (see Figure~\ref{fig2}), similar 
389:  in intensity to those of 1H\,1934-063. The permitted lines
390: are rather narrow, just slightly broader than the forbidden lines.
391: The 1$\mu$m lines
392:  are strong and well resolved. Br$\gamma$ is clearly
393:  detected as well as the molecular H$_{2}$ line (1,0)S(1) 
394:  at \lb2.121$\mu$m. The \ion{He}{1} \lb10830,
395:  \ion{Ca}{2} triplet,
396:  and \ion{O}{1} \lb8446, \lb11287
397:  lines are all strong. The \lb9220 blend is 
398:  present with \feii\ \lb9177, \lb9202, \lb9256, Pa9 and 
399: \ion{Mg}{2} \lb9244 clearly resolved. The forbidden line 
400: spectrum is remarkable,
401:  displaying strong low and high ionization lines such
402:  as [\ion{S}{3}] \lb9068,9031, [\ion{C}{1}] \lb9850,
403:  [\ion{S}{8}] \lb9912, [\ion{S}{2}] \lb10286,10322,
404:  [\ion{S}{9}] \lb12523, [\ion{Si}{10}] \lb14305 and
405:  [\ion{Si}{6}] \lb19630. The continuum emission rises 
406:  steeply from Pa$\beta$ to shorter wavelengths and is flat 
407:  between 1.5 and 2.6$\mu$m.
408: 
409:  \subsubsection{Mrk\,335}
410: 
411:  This object is a strong UV and X-ray source
412:  \citep{bbf96, cren99, com00, ball00}. The near-IR
413:  spectrum is dominated by \ion{He}{1} \lb10830
414:  and the \ion{H}{1} Paschen lines (Figure~\ref{fig3}).
415:  The 1$\mu$m \feii\ feature is not as strong as in the former two 
416:  galaxies. In addition, \feii\ \lb11127 and Pa$\alpha$ are 
417:  affected by atmospheric absorption. The \ion{Ca}{2} 
418:  triplet in emission, although present, is weak.
419:  The \lb9220 feature is rather strong. From the Gaussian fitting, 
420: we identified \feii\ \lb9177, \lb9202, \lb9256 and Pa9
421: as individual components of this blend. The continuum 
422:  emission of this galaxy is particularly interesting:
423:  there is a broad minimum around 1.4$\mu$m and a rapid
424:  rise towards longer wavelengths. It
425:  probably represents the shift from a non-thermal
426:  continuum to the thermal dust emission regime.
427:  In their near-IR spectrophotometry
428:  of Mrk\,335, \citet{rudy82} pointed out that the colours of this
429:  object are among the reddest and most nonstellar
430:  reported for any Seyfert Galaxy. They also conclude
431:  that the small flux at 10$\mu$m and 20$\mu$m indicates that
432:  Mrk\,335 does not contain large amounts of cool dust
433:  and that the bulk of near-IR luminosity is produced
434:  by a relatively small amount of warm dust. 
435: 
436:  \subsubsection{Mrk\,1044} \label{mrk1044}
437: 
438:  The dominant feature in the spectrum
439:  of this galaxy is \ion{He}{1} \lb10830 
440:  (Figure~\ref{fig4}). The 1$\mu$m 
441:  \feii\ lines are present but only 
442:  \lb9997 and \lb10501 are easily identified. \feii\ 
443:  \lb11127 is severly affected by atmospheric absorption
444:  so it is only possible to derive a lower limit for
445:  the flux of this line. \ion{O}{1} \lb8446, \lb11287,
446:  \ion{Ca}{2} \lb8498, \lb8542, \lb8662 and the \lb9220
447:  feature are other important signatures of the BLR visible
448:  in this spectrum.  The NLR contribution, as in
449:  Mrk\,335, is rather weak, with only [\ion{S}{3}] \lb9531
450:  clearly observed. The continuum emission also shows
451:  a broad minimum around 1.4$\mu$m. It is very
452:  steep blueward and mostly flat redward of 1.4$\mu$m.
453: 
454:  \section{Discussion}
455: 
456:  The spectra presented here not only confirm the
457:  presence of the 1$\mu$m \feii\ lines in NLS1s, but also
458:  that they are the strongest and most prominent
459:  \feii\ lines observed in the 0.8$-$2.4 $\mu$m interval. 
460:  This suggests that the 1$\mu$m \feii\ lines 
461:  are a common feature in these objects, similar to
462:  the \feii\ bumps at both sides of H$\beta$  in the 
463:  optical region. In addition, fainter \feii\ lines 
464:  are clearly detected in the interval 8700$-$9300 \AA. 
465:  Our goal is to discuss the dominant excitation mechanism 
466:  for these lines and to confirm if, as in I\,Zw\,1, 
467:  collisional excitation is the most plausible one. For 
468:  this purpose, our data as well as information from other 
469:  wavebands will be used. It is also important to determine 
470:  if the relative intensities of the 1$\mu$m \feii\ lines 
471:  vary from object to object, or if, as in the optical region,
472:  they are approximately constant.
473:  
474:  
475:  \subsection{Excitation Mechanisms of the near-IR \feii\ lines}
476: 
477:  Excitation mechanisms invoked for the \feii\ emission
478:  include continuum fluorescence via the UV resonance lines,
479:  self-fluorescence via overlapping \feii\ transitions,
480:  and collisional excitation. The latter two are thought
481:  to contribute to the bulk of the \feii\ emission
482:  \citep{wnw85}. Nonetheless, most models that take into
483:  account these processes fail at reproducing the observed
484:  intensity of UV and optical \feii\ lines.
485:  
486:  \subsubsection{What we have learned from stars}
487: 
488:  Based on the observation of \feii\ lines in low resolution 
489:  {\it IUE} spectra of cool giants and supergiants, \citet{joh83} 
490:  and \citet{jojor84} proposed a large set of pumping channels
491:  for Ly$\alpha$ to excite \feii\ levels with
492:  excitation energy around 10 eV (5p levels). The decay route
493:  from these levels include $e^{\rm 4}D$ and $e^{\rm 6}D$ of 
494:  3$d^{\rm 6}(^{\rm 5}D)5s$, which lie at wavelengths 
495:  $\sim$8000$-$9600 \AA. These two levels
496:  decay giving rise to multiplets such as UV 399, 391, 380 ($\sim$
497:  2850\AA), 373 ($\sim$2770\AA) and 363 ($\sim$2530\AA), which supply
498:  approximately 20\% of the total energy in \feii\ lines
499:  between 2000 and 3000 \AA\ \citep{pen87}.
500:  In addition, \citet{jojor84} proposed Ly$\alpha$ 
501:  fluorescence to pump a further $^{\rm 4}G^{\rm 0}$ level at 
502:  $\sim$13 eV, whose primary decay to b$^{\rm 4}G$ produces the UV 
503:  lines at \lb1841, \lb1845, \lb1870 and \lb1873\footnote{In the \citet{jojor84}
504:  paper, the primary decay from $^{\rm 4}G^{\rm 0}$ to b$^{\rm 4}G$ 
505:  produce only \feii\ \lb1870,1873. Nonetheless, data from the Iron
506:  Project also predict lines at \lb1841 and \lb1845 due to the fine
507:  structure of that upper level \citep{nahar95, pra01}.
508:  The current notation for $^{\rm 4}G^{\rm 0}$ is $(u,t)^{\rm 4}G^{\rm 0}$.}.
509:  A  subsequent cascade to $z^{\rm 4}F^{\rm 0}$
510:  would produce the 1$\mu$m \feii\ lines, as is illustrated
511:  in Figure~\ref{levels}. 
512: 
513:  Near-infrared spectroscopy of young stellar objects \citep{hasim88,
514:  hp89, ham94, kerc94}
515:  has, in fact, shown most of the near-IR \feii\ emission features
516:  associated to L$\alpha$ fluorescence. The flux ratios measured 
517:  from our objects agree with this scenario. 
518: 
519:  \subsubsection{Is Ly$\alpha$ fluorescence also at work in AGNs?}
520: 
521:  In AGNs, a definitive test to confirm the presence of the
522:  Ly$\alpha$ pumping has been elusive due to both the blending
523:  of the broad lines and the lack of suitable data in the near-IR region. 
524:  Recently, \citet[hereafter SP98]{sp98}, 
525:  using theoretical \feii\ emission-line strengths, derived by 
526:  considering all of the above four mechanisms, demonstrated that a key 
527:  feature associated with Ly$\alpha$ pumping is significant \feii\ 
528:  emission in the wavelength interval 8500$-$9500 \AA. In particular, 
529:  their models predict strong  \feii\ lines at \lb8490, \lb8927,
530:  \lb9130, \lb9177 and \lb9203. Figure~\ref{feiilyalf} shows
531:  an enlarged portion of the \lb8900$-$\lb9300 region for each of 
532:  the object's spectra, clearly displaying the latter three \feii\ lines 
533:  and a blip at the expected position of \feii\ \lb8927.
534:  This result offers the first direct evidence of Ly$\alpha$ 
535:  pumping in AGNs and will be discussed in detail in \citet{roar01b}.
536:     
537:  If Ly$\alpha$ fluorescence is also responsible for the 
538:  production of the 1$\mu$m lines, the feed lines \lb1841, \lb1845,
539:  \lb1870 and \lb1873  
540:  should be observed (see Figure~\ref{levels}). We have searched 
541:  the {\it IUE} and {\it HST/FOS} public spectra available for the 
542:  objects studied here to look for emission lines at these four
543:  positions. No significant emission was found, as is illustrated
544:  in Figure~\ref{arkuv}, where the region around \lb1860 in 
545:  Ark\,564 and Mrk\,335 is shown.
546:  The spikes detected at the expected positions are almost
547:  at the noise level. Assuming that they actually correspond to
548:  \lb1840,1844 and \lb1870,1873, Table~\ref{balance} gives the flux 
549:  upper limit measured for these features (column 2) and the strength 
550:  that they should need to have (assuming zero internal reddening) 
551:  in order to produced the observed 1$\mu$m lines (column 3). 
552:  The expected values were derived assuming a one-to-one relation
553:  between the number of photons of the 1$\mu$m and UV \feii\ lines. 
554:  Column 4 of that same table lists the E(B-V) necessary to reduce
555:  the expected UV flux to the observed value. It was derived
556:  using the Galactic extinction law of
557:  \citet{ccm89}. Since that the amount 
558:  of reddening should conceal other UV features that
559:  are otherwise prominent, we discard this effect for explaining the
560:  weakness of the feed UV lines. Another argument against 
561:  strong reddening is the presence of a steep blue continuum 
562:  in most of these objects.  
563:     
564:  There is still the possibility that these lines are, in fact,
565:  emitted but then efficiently destroyed within the BLR. But 
566:  following the 
567:  same line of reasoning of RMPH,  other UV features should
568:  also be affected by such a destruction mechanism. Since
569:  Ly$\alpha$, \ion{He}{2} $\lambda$1640 and other important
570:  UV lines are clearly visible, we ruled out 
571:  that hypothesis. 
572:  
573:  We conclude that Ly$\alpha$ fluorescence cannot
574:  be the only process responsible for the emission of the
575:  1$\mu$m \feii\ lines. From the observed strength 
576:  of the UV lines, that mechanism appears to have a minimal
577:  contribution to the near-IR features.
578: 
579:  \subsubsection{Alternative mechanisms}
580: 
581:  An alternative process to the production of
582:  the near-IR \feii\ emission is recombination. However,
583:  in addition to the 1$\mu$m lines,
584:  a myriad of other transitions should also be present
585:  in the 0.8 -- 2.4$\mu$m interval, some of them as strong or
586:  stronger than the ones around 1$\mu$m \citep{kur81}. 
587:  Their absence clearly rules out this hypothesis.
588:  
589:  Continuum fluorescence, first suggested by \citet{waok67},
590:  has been a popular explanation for the presence of 
591:  strong \feii\ emission in Seyfert 1 galaxies and quasars.
592:  In this process, the 5 eV levels of \feii\ are excited by
593:  absorption of ultraviolet continuum photons in the 
594:  resonance-line transitions. Nonetheless, this process
595:  very unlikely contribute to populate the upper b$^{\rm 4}G$ 
596:  level because no resonant transitions connecting this level 
597:  to the ground levels have yet been identified.
598: 
599:  There are other two possible excitation mechanisms. The first 
600:  one takes advantage of the fact that the 1$\mu$m lines
601:  are, by far, the strongest \feii\ lines observed in the 
602:  0.8$-$2.4$\mu$m interval. Since they all descend from a common
603:  upper level ($b~^{4}G$, see Figure~\ref{levels}), their presence
604:  suggests that some sort of selective mechanism is operating
605:  to populate that particular level. Collisional pumping from 
606:  the metastable $a~^{4}G$ level is, at least 
607:  qualitatively, a viable process to explain the strength of 
608:  these features in AGNs. The small energy separation between the two
609:  levels involved ($\sim$3.57 eV), the high density of the
610:  BLR gas (N$_{\rm e} \leq 10^{9}$ cm$^{-3}$) and the 
611:  temperature predicted for the \feii\ region by most models 
612:  ($\sim$8000 K) favors this mechanims. In order to
613:  be efficient, the population of the level $a~^{4}G$ must
614:  be significant, i.e. above the Boltzman population.
615:  This can be done by the decay to $a~^{4}G$ from
616:  another set of levels such as $^{\rm 4}F^{\rm 0}$.
617:  A signature of this process is \feii\ emission
618:  in the 1535$-$1550 \AA\ interval. The \feii\ line at \lb1534 
619:  (5$p^{\rm 4}F^{\rm 0}-a~^{4}G$) is a particular example of this
620:  decay route. Unfortunately, its proximity to the strong 
621:  \ion{C}{4} \lb1548 feature may hamper the detection of that 
622:  emission. The $a~^{4}G$ levels can also be populated by decays 
623:  from $y^{\rm 4}H^{\rm 0}$. This route produces lines in 
624:  the 2430$-$2460 wavelength interval. It is also possible
625:  that $b^{\rm 4}G$ be collisionally populated from the lower
626:  $z^{\rm 4}F^{\rm 0}$. That requires high optical depths in 
627:  the strong optical and UV permitted transitions originating
628:  from this latter level in order to keep it significantly populated.
629:  This condition seems plausible because most optical and
630:  UV \feii\ multiplets are rather strong in AGNs.
631:   
632:  The last possible process is to populate $b^{\rm 4}G$ directly 
633:  by decays from other levels previously pumped by Ly$\alpha$ fluorescence.
634:  In particular, the upper $^{\rm 4}F^{\rm 0}$ level, after absorption
635:  of a Ly$\alpha$ photon, may cascade directly to $b^{\rm 4}G$, emitting
636:  \feii\ lines in the 2700$-$2770 \AA\ interval (\feii\ \lb2707,
637:  \lb2739, \lb2765, \lb2769). In this case, no collisional excitation
638:  of $b^{\rm 4}G$ is required. An inspection of the {\it IUE/HST} 
639:  spectra of Ark\,564, Mrk\,335 and Mrk\,1044 show evidence of
640:  the above lines at the expected position, suggesting that this
641:  process may also contribute to the 1$\mu$m \feii\ 
642:  line intensities.
643: 
644:  \subsubsection{Conclusions} 
645:  
646:  In summary, clear evidence of Ly$\alpha$ pumping of \feii\ 
647:  is found from the observation of the primary cascade lines in
648:  the 8900$-$9300 \AA\ interval, formed by the decay of the upper 5p 
649:  levels to the $e^{\rm 4}D$ and $e^{\rm 6}D$ levels. This result
650:  confirms that fluorescence excitation by Ly$\alpha$ has an
651:  active role in the \feii\ spectra of AGNs. Nonetheless,
652:  contrary to what was previously suggested, this same process seems 
653:  to be insufficient to explain the 
654:  strength of the 1$\mu$m lines, the strongest of all \feii\ lines 
655:  in the 8000$-$24000 \AA\ region. This conclusion is drawn from the 
656:  weakness of the primary cascade lines that feed the upper term from which
657:  the 1$\mu$m \feii\ emission originates. Our observations point 
658:  to a combination of collisional excitation and decays from 
659:  levels populated by Ly$\alpha$ photons to 
660:  produce the \feii\ spectrum in the near-IR. 
661:   
662:  \subsection{Relative Intensities of the 1$\mu$m \feii\ Lines}
663: 
664:  One advantage of the 1$\mu$m \feii\ lines over the
665:  corresponding emission in the optical region is that 
666:  most of the individual lines are clearly separated in
667:  wavelength, allowing their measurement without the need
668:  of constructing an \feii\ template to
669:  quantify them. Although in most cases emission from 
670:  \ion{H}{1}, \ion{He}{1} and \ion{He}{2} is near the \feii\ 
671:  lines, it is still possible to measure, at a good 
672:  confidence level, the flux of the individual components. 
673: 
674:  In all cases our spectra show that $\lambda$9997 
675:  is the strongest \feii\ transition in the near-IR. The 
676:  relative fluxes of other \feii\ lines compared 
677:  to that of $\lambda$9997 are listed in Table~\ref{feiirat}. 
678:  We have also included for comparison, data taken 
679:  from RMPH for I\,Zw\,1 (column 6), from \citet{rudy91} 
680:  for LkH$\alpha$\,101 (column 7), and from \citet{kerc94} for
681:  MWC 249 and R\,Mon (columns 8 and 9 respectively). The
682:  last three objects are Galactic sources, well known for 
683:  their strong \feii\ emission lines in the near-IR,
684:  which is believe to arise in a gas with physical
685:  conditions similar to those found in AGNs.
686: 
687:  The 1$\mu$m \feii\ ratios vary from 
688:  object to object, as is seen in  Table~\ref{feiirat}. Comparing 
689:  the values derived for our sample with those of I\,Zw\,1 
690:  (considered as the prototype NLS1), we found that, within errors, 
691:  1H\,1934-063 is the most I\,Zw\,1-like. On the other hand, Mrk\,1044
692:  deviates most significantly in regard to the \feii\
693:  emission. The situation is
694:  also highly discrepant when we compare the NLS1 ratios
695:  to those of Galactic objects. No agreement between these two
696:  groups can be seen. The ratios among
697:  the Galactic sources look rather constant, in contrast to
698:  the large scatter showed by the NLS1 data. \citet{rudy91}
699:  had already noted the similarity of line ratios between
700:  Lk\,H$\alpha$ 101 and $\eta$ Car, another Galactic source,
701:  and argued that this may suggest that the same process gives
702:  rise to the lines in both. They proposed Ly$\alpha$ fluorescence 
703:  as the dominant mechanism for the production of \feii\ 
704:  but did not discard collisional excitation. 
705: 
706:  It may be argued that the large scatter in \feii\ line 
707:  ratios is due to reddening, since none of the NLS1 
708:  spectra  have been corrected for internal extinction
709:  (the Galactic source line ratios all have been corrected). 
710:  However, the near-IR region is less affected by 
711:  reddening than the UV and optical regions  and the wavelength
712:  interval where the 1$\mu$m \feii\ features are located 
713:  is relatively small. Thus, one would expect that the 
714:  reddening is not important here. In addition, all the 
715:  objects have strong emission lines in the UV, 
716:  indicating that the reddening
717:  is small. In order to completely avoid any reddening effect,
718:  we have calculated line ratios that are very 
719:  close in wavelength and therefore insensitive to reddening.
720:  Such is the case for \feii\ \lb9997/\lb10171, \feii\ \lb11127/\lb10863 and 
721:  \feii\ \lb10501/\lb10863. The results, also listed in 
722:  Table~\ref{feiirat}, follow the same trend, in the 
723:  sense that the line ratios vary from galaxy to galaxy 
724:  and also differ from the line ratios derived for the Galactic 
725:  sources. This variation is more clearly seen in 
726:  Figure~\ref{ratiosfeii}, where the line ratios are plotted 
727:  for the objects listed in Table~\ref{feiirat}. The large scatter 
728:  of the NLS1 data and the constancy among Galactic sources
729:  is evident.  
730: 
731:  The above results strongly contrast to what is observed
732:  in the optical region, where the relative intensities of the
733:  \feii\ lines in AGNs seems rather similar from object to 
734:  object. Apparently, the only difference between the optical \feii\
735:  features of two given objects is the width of the individual 
736:  lines (unless a scale factor). This result was clearly illustrated 
737:  in \citet{phi78} by comparing the \feii\ emission in the region 
738:  4400$-$5300 \AA\ of several Seyfert 1 galaxies. The \feii\ 
739:  emission in that same interval for I\,Zw\,1,
740:  Ark\,564 and 1H\,1934-063 is shown in Figure~\ref{optfeii}.
741:  Note how similar are the \feii\ multiplets
742:  in these objects. For that reason, usually I\,Zw\,1, which has
743:  one of the narrowest permitted lines in Seyfert 1s, is used as a 
744:  template to determine the strength of the \feii\ emission 
745:  \citep{bogre91, roar00}.
746: 
747:  The question that arises now is if there 
748:  is a correlation between the strength of the \feii\ in 
749:  the optical and near-IR. Figure~\ref{feiicorr} shows 
750:  FeII/\hbeta\ versus \feii\ IR/Pa\,$\delta$. The former is 
751:  the flux of the \feii\ emission centered in $\lambda$4570, 
752:  which scales with the total optical \feii, normalized to 
753:  \hbeta, while the latter is the sum of the \feii\ lines
754:  \lb9997, \lb10171 and \lb10501, normalized 
755:  to the Pa$\delta$ flux. Except for Mrk\,335 (the galaxy with the
756:  smallest \feii/\hbeta), \feii\ \lb9997 seems to scale
757:  with the corresponding optical \feii. That correlation is
758:  very important, since most of the optical \feii\ lines 
759:  in the wavelength interval 4400$-$5300  arise from 
760:  the same upper two upper levels $z^{\rm 4}F^{\rm 0}$ and 
761:  $z^{\rm 4}D^{\rm 0}$. Although no definitive conclusions
762:  can be drawn due to the reduced number of points, 
763:  Figure~\ref{feiicorr} clearly suggest that the optical
764:  \feii\ emission is directly linked to the 1$\mu$m \feii\
765:  lines. In this scenario, decays from $b^{\rm 4}G^{\rm 0}$ largely 
766:  contribute to the population of those two levels. Although
767:  a detailed balance between the energy carried by the near-IR
768:  and optical \feii\ lines is far from the scope of this paper,
769:  this result is important for current \feii\ modeling.     
770:  
771:  Factors such as different optical depth effects and, possibly,
772:  different sensitivities to physical parameters between the
773:  near-IR and optical \feii\ lines can be invoked
774:  to explain why the line ratios in the former region present
775:  more scatter than those of the latter. Additional observations
776:  in a larger sample of objects and a more accurate modeling 
777:  of the 1$\mu$m \feii\ lines are needed to explain this trend. 
778:  It is also interesting 
779:  to understand why Mrk\,335 deviates so much from the other 
780:  NLS1s. While its optical \feii\ is rather low for a typical 
781:  NLS1s, its near-IR \feii\ ranks it as a strong 
782:  emitter. 
783: 
784:  \section{Kinematics of the BRL region} \label{kin} 
785: 
786:  The quality and spectral resolution of the SpeX data
787:  allow us to study and compare, for the first time, the 
788:  form and width of the different near-IR permitted emission 
789:  line profiles of AGNs, and in particular of NLS1s. This
790:  is important for at least two reasons. First,
791:  it has been argued (from both theoretical models and 
792:  observational grounds) that low-ionization lines such as 
793:  \feii, \ion{O}{1}, \ion{Mg}{2}, \ion{Ca}{2} and Balmer
794:  lines are formed in a distinct, separate zone (most
795:  probably the outer part of the BLR) from the place 
796:  where most high-ionization lines (i.e Ly$\alpha$, 
797:  \ion{He}{2}, \ion{Ca}{4}) are produced \citep{csdjp86, rms97}. 
798:  Differences in line widths between low 
799:  and high ionization lines are usually interpreted as 
800:  evidence of this stratification, provided the velocity 
801:  field does not differ strongly from gravitationally
802:  bound Keplerian or virial motion, as seems to be 
803:  the case for most AGNs \citep{pw99}. 
804:  Second, the presence of asymmetries in the line 
805:  profiles  helps to distinguish between radial and 
806:  rotational motions, setting constrains on the
807:  velocity field of the BLR. 
808: 
809:  Although the visible and UV spectral regions in AGNs 
810:  are full of permitted emission lines, usually they are 
811:  severely blended, affecting the placement of 
812:  the continuum level and the separation of the individual
813:  profiles. Such is the case with the \feii\ 
814:  lines, which form a pseudo-continuum even in NLS1s
815:  because of their close proximity. The near-IR region is
816:  promising in this respect (if the resolution is high enough).
817:  Lines such as \ion{Ca}{2} 
818:  \lb8664, \feii\ \lb11127, \ion{O}{1} \lb11287 and 
819:  Pa$\beta$ are completely isolated or mildly blended, 
820:  making their separation and the analysis of their 
821:  emission profiles easier.
822: 
823:  For the above reasons we have chosen \feii\ \lb11127, 
824:  \ion{O}{1} \lb11287, \ion{O}{1} \lb8446 and \ion{Ca}{2} 
825:  \lb8664 as representative of low-ionization BLR lines, and 
826:  Pa$\beta$ as an example of a BLR recombination line. 
827:  By comparing their emission line profiles, we expect to
828:  get clues about the structure and kinematics of  
829:  the emitting region. In addition,  
830:  we have used the [\ion{S}{3}] \lb9531 profile to
831:  represent the velocity field of the Narrow Line
832:  Region (NLR).
833: 
834:  The observed profiles of Pa$\beta$, \feii\ \lb11127,
835:  \ion{O}{1} \lb11286 and [\ion{S}{3}] \lb9531 are shown 
836:  in velocity space in Figure~\ref{profiles}. 
837:  In order to easily compare them, the peak 
838:  intensity of each line has been normalized to unity. 
839:  In all cases, Pa$\beta$ has been stripped of the 
840:  narrow component by subtracting the contribution 
841:  from the NLR according to the procedure explained 
842:  in \S \ref{observa}. Since the \feii\ and 
843:  the \ion{O}{1} emission are features restricted to 
844:  the BLR, there is no need to determine such a 
845:  contribution in these two lines.
846: 
847:  The results are surprising. Except for Mrk\,1044,
848:  \feii\ and \ion{O}{1} share almost the same velocity 
849:  field and display very similar line profiles. However, 
850:  they are significantly narrower than Pa$\beta$. In
851:  fact, they are just slightly broader than [\ion{S}{3}]
852:  \lb9531. Assuming that this latter line is formed
853:  in the inner portion of the NLR (its critical density,
854:  $N_{c}$, is $\sim 10^{7.5}$ cm$^{-3}$), 
855:  the small width of \feii\ and \ion{O}{1} indicates
856:  that we are probably looking at the outer boundaries
857:  of the BLR. The result obtained for Mrk\,1044 can
858:  be explained if we recall that in 
859:  Figure~\ref{profiles} the \feii\ profile plotted 
860:  corresponds to that of \lb10501 instead of \lb11127, because
861:  this latter line is severely affected by atmospheric
862:  absorption (see \S \ref{mrk1044}). 
863:  \feii\ \lb10501 is, in fact, a blend of two \feii\ lines 
864:  located at \lb10491 and \lb10501. Observations in
865:  stars \citep{rudy91} show that the latter line is about 5 times 
866:  stronger than the former (in which case the effect of \lb10491
867:  on the blend width would be negligible). A lower ratio may cause 
868:  \lb10501 to appear broader, as it seems to be the case 
869:  for Mrk\,1044 as well as for Ark\,564 
870:  (see Table~\ref{tableflux}).
871: 
872:  Assuming that the clouds emitting the broad lines 
873:  are gravitationally bound to the central mass 
874:  concentration and that the velocity field is Keplerian, 
875:  Figure~\ref{profiles} clearly supports the hypothesis 
876:  that \feii\ and \ion{O}{1} are emitted in the outer part 
877:  of the BLR while most of the Pa$\beta$ flux is formed 
878:  in the intermediate portion of this region. 
879:  This result is compatible with the physical conditions 
880:  needed to form permitted \feii\ and \ion{O}{1}, i.e. neutral 
881:  gas strongly shielded from the incident ionizing 
882:  radiation of the central source, conditions which can
883:  only be met deeper within the neutral zone of the
884:  outermost part of the BLR. 
885: 
886:  The above results seem to contradict that of \citet{morward89}, 
887:  who compared the H$\alpha$ and \ion{O}{1} \lb8446 line profiles 
888:  in a sample of 12 objects. They found no evidence for 
889:  differences between the profiles of these two lines. 
890:  Nonetheless, recall that Pa$\beta$ is less affected than 
891:  H$\alpha$ by radiation transport effects. When crossing the BLR,
892:  H$\alpha$ photons produced in inner higher velocity clouds 
893:  are more likely to be absorbed than Pa$\beta$ photons. Thus, 
894:  the H$\alpha$ line probes the outer low velocity clouds, while  
895:  the Pa$\beta$ profile shows the contribution from inner high
896:  velocity clouds. For the above reason, we expect that
897:  Pa$\alpha$ be broader than H$\alpha$ as well as broader
898:  than low ionization lines such as \ion{O}{1} and \feii.
899:  The most important point here is that Figure~\ref{profiles}
900:  shows that low-ionization BLR lines are emitted almost
901:  at the border between the BLR and NLR.
902: 
903:  Another interesting result that can be drawn from
904:  Figure~\ref{profiles} is the difference in the
905:  form of the profiles from object to object: in 
906:  1H\,1934-063 they are typically Gaussian while in Ark\,564 
907:  they are mostly Lorentzian, with Mrk\,335 and Mrk\,1044
908:  being intermediate cases. This is valid not only for Pa$\alpha$
909:  but also for \feii\ and \ion{O}{1}. Since the spectra were 
910:  taken at the same resolution, with the same telescope and setup, this
911:  clearly indicates differences in the velocity fields
912:  of the objects. 
913: 
914:  It is also possible to probe weakly ionized material within
915:  the BLR by means of \ion{Ca}{2} \lb8664, one of the Ca
916:  near-IR triplet lines that is perfectly isolated and has 
917:  good S/N in our data. The observed \ion{Ca}{2} \lb8664 
918:  is compared to that of \ion{O}{1} \lb8446 in Figure~\ref{caii}.
919:  Clearly, these two lines have very similar form and width, 
920:  indicating that they share the overall 
921:  kinematics of the region where they are
922:  emitted. This was already pointed out by \citet{pe88} in 
923:  his study of the \ion{Ca}{2} and \ion{O}{1} lines in
924:  14 out of 40 AGNs for which simultaneous emission of 
925:  these two ions was found. Here, our results not only
926:  agree with those of \citet{pe88}, but also extend
927:  that analysis to the \feii\ and show that those 
928:  three lines should have similar kinematics.
929:  Due to the high complexity the \feii\ atom, its spectrum 
930:  is far from being understood. Information derived for
931:  \ion{O}{1} or \ion{Ca}{2} (i.e. column density and ionization
932:  parameters) provide valuable constraints for modelling the \feii\
933:  atom, given that these species seem to form from
934:  the same ensemble of clouds.
935: 
936:  The importance of the \ion{Ca}{2} 
937:  emission as a valuable diagnostic of the the BLR is
938:  stressed by \citet{ferper89}. Using photoionization
939:  models that reproduce well both the high and low ionization
940:  line intensities, they found that clouds with extremely
941:  high column densities (N$_{\rm e} \sim$ 10$^{25}$ cm$^{-2}$)
942:  are needed to emit \ion{Ca}{2}. The large N$_{\rm e}$ suggests
943:  that the clouds might be a wind or corona above a star or
944:  accretion disk. It is well known that a wind scenario 
945:  usually produces asymmetries in the observed profiles.
946:  The \ion{Ca}{2} line profiles in Ark\,564 marginally
947:  show evidences of assymetries towards the blue (Figure~\ref{caii}, 
948:  upper right panel). High-resolution spectroscopy around
949:  this region is highly desirable to find additional clues 
950:  about the kinematics of the outer portion of the BLR.
951:  Our results stress the kinematic connection 
952:  between the low ionization species and confirm previous
953:  claims that the understanding of the mechanisms leading to 
954:  the formation of the \feii\ and \ion{Ca}{2} lines is crucial 
955:  for solving the puzzle associated with the geometry and 
956:  physical conditions of the outer parts of the BLR.  
957:    
958:  In summary, the comparison of the near-IR broad emission 
959:  line profiles definitively shows the presence of a stratification
960:  within the BLR. Low-ionization lines such as \feii, \ion{O}{1}
961:  and \ion{Ca}{2} are emitted in the outer edge of the BLR. 
962:  Pa$\beta$ emission, on the other hand, maps the velocity field 
963:  of the inner portions of the BLR. 
964:   
965:  \section{Conclusions}
966: 
967:  We have analyzed four NLS1 spectra in the wavelength range
968:  0.8 -- 2.4$\mu$m in order to study the \feii\ emission and
969:  the kinematics of the BLR. The results show that the \feii\
970:  emission in the 8500$-$9300 \AA\ interval and around 
971:  1$\mu$m are common features of this class of AGNs. The
972:  \feii\ lines in the former interval are, for the first time, 
973:  resolved in this type of object. Since they are primary cascades of the
974:  upper 5p levels pumped by Ly$\alpha$ fluorescence, our
975:  data offer clear observational evidence for the presence of 
976:  this excitation mechanism in AGNs. The \feii\ lines located
977:  at \lb9997, \lb10501, \lb10863 and \lb11127, which share
978:  the same upper and lower levels ($b^{\rm 4}G$ and $z^{\rm 4}F^{\rm 0}$,
979:  respectively), are very prominent and the strongest \feii\ lines 
980:  in the four spectra. Although Ly$\alpha$ fluorescence
981:  indirectly contributes to their production, that
982:  mechanism cannot explain their full strength because of the 
983:  weakness of the UV \feii\ lines that feed the upper $b^{\rm 4}G$
984:  level. Collisional ionization, mainly from the $a^{\rm 4}G$ and 
985:  $z^{\rm 4}F^{\rm 0}$ levels, is suggested as an additional
986:  mechanism that plays an active role in exciting the 1$\mu$m
987:  \feii\ lines.
988:     
989:  The near-IR \feii\ line ratios (i.e. \lb9997/\lb10501 and 
990:  \lb10501/\lb10863) show a large scatter from 
991:  object to object, contrary to what is observed in the 
992:  optical region. Nonetheless, the apparent good correlation 
993:  found between the \feii\ emission of these two wavelength 
994:  intervals supports a common origin. Since most of
995:  the optical \feii\ multiplets originate in transitions
996:  from $z^{\rm 4}F^{\rm 0}$, that level would be mainly
997:  populated by the decay from $b^{\rm 4}G$.  
998: 
999:  The emission line profiles of \feii,
1000:  \ion{O}{1} and \ion{Ca}{2} are very similar in form and
1001:  width but significantly narrower than that of Pa$\beta$. 
1002:  This strongly suggests that the \ion{Ca}{2}, \feii\ and
1003:  \ion{O}{1} regions are co-spatial and kinematically linked.
1004:  Moreover, because their profiles are just slightly broader
1005:  than that of [\ion{S}{3}] \lb9531, low ionization BLR lines
1006:  are most likely formed in the outermost portion of
1007:  the BLR. This is compatible with the physical conditions
1008:  required to form them: cool gas with high column 
1009:  densities that provides a shield from the incident 
1010:  ultraviolet continuum of the central source and a location
1011:  deep, within the neutral zone of the BLR. The larger
1012:  width of the Paschen lines indicates that they are
1013:  suitable for mapping the gas located at intermediate positions
1014:  in the BLR.  
1015:     
1016:  Given the high complexity of the \feii\ atom and its ubiquitous 
1017:  emission in AGNs, only detailed model calculations that take into 
1018:  account all possible excitation mechanisms and transitions among
1019:  the levels can shed some light on the dominant processes involved
1020:  in the excitation of this atom. The data presented here 
1021:  provide valuable constraints for understanding this complex emission.  
1022:   
1023: 
1024:  \acknowledgments
1025: 
1026: We thank the referee, Dr. Fred Hamann, for his useful and 
1027: thoughtful comments that helped to improve this manuscript. 
1028: We also wish to thank the IRTF staff, support scientist Bobby Bus, 
1029: and telescope operator Dave Griep for contributing to a productive 
1030: observing run. Mike Cushing provided patient assistance with the 
1031: xspextool software. This research has been supported by the 
1032: Funda\c c\~ao de Amparo a Pesquisa do Estado de S\~ao Paulo $-$
1033: FAPESP, under contract 00/01020-5 and PRONEX grants 662175/1996-4 
1034: and 7697100300. We acknowledge the use the Atomic
1035: Line List v2.04 (http://www.pa.uky.edu/~peter/atomic/).
1036: 
1037: 
1038:  \begin{thebibliography}{}
1039: 
1040:  \bibitem[Ballantyne, Iwasawa \& Fabian(2000)]{ball00} Ballantyne, D.
1041:  R.,
1042:      Iwasawa, K., and Fabian, A. C. 2000, Astro-ph/0011360.
1043:  \bibitem[Boller, Brandt \& Fink(1996)]{bbf96} Boller, Th., Brandt,
1044:      W. N. and Fink, H. 1996, \aap, 305, 53
1045:  \bibitem[Boroson \& Green(1991)]{bogre91} Boroson, T. A., and
1046:      Green, R. F. 1991, \apjs, 80, 109
1047:  \bibitem[Cardelli, Clayton \& Mathis(1989)]{ccm89} Cardelli, J. A.,
1048:  Clayton, G. C., and Mathis, J. S. 1989, \apj, 345, 245
1049:  \bibitem[Collin-Souffrin et al.(1986)]{csdjp86} Collin-Souffrin, S.,
1050:      Dumont, S., Joly, M., and P\'equignot, D. 1986, \aap, 166, 27
1051:  \bibitem[Comastri et al.(2001)]{com00} Comastri, A., Stirpe, G. M.,
1052:      Vignali, C., Brandt, W. N., Leighly, K. M., Fiore, F., 
1053:      Guainazzi, M., Matt, G., Nicastro, F., Puchnarewicz, E. M.,
1054:      and Siemiginowska, A. 2001, \aap, 365, 400
1055:  \bibitem[Crenshaw et al.(1999)]{cren99} Crenshaw, D. M., Kraemer,
1056:      S. B., Boggess, A., Maran, S. P., Mushotzky, R. F., and
1057:      Wu, C. 1999, \apj, 516, 750
1058:  \bibitem[Cushing, Vacca \& Rayner(2001)]{cvr01} Cushing, M. C., Vacca,
1059:   W. D., and Rayner, J. T. 2001, PASP, In prep.
1060:  \bibitem[Ferland \& Persson(1989)]{ferper89} Ferland, G. J., and Persson
1061:      S. E. 1989, \apj, 347, 656
1062:  \bibitem[Hamann \& Simon(1988)]{hasim88} Hamann, F., and Simon, M. 1988,
1063:      \apj, 327, 876
1064:  \bibitem[Hamann \& Persson(1989)]{hp89} Hamann, F., and Persson,
1065:      S. E. 1989, \apjs, 71, 931
1066:  \bibitem[Hamann et al.(1994)]{ham94} Hamann, F., DePoy, D. L.,
1067:      Johansson, S., and Elias, J. 1994, \apj, 422, 626
1068:  \bibitem[Johansson(1983)]{joh83} Johansson, S. 1983, \mnras, 205, 71P
1069:  \bibitem[Johansson \& Jordan(1984)]{jojor84} Johansson, S., and 
1070:      Jordan, C. 1984, \mnras, 210, 239      
1071:  \bibitem[Joly(1991)]{jol91} Joly, M. 1991, \aap, 242, 49
1072:  \bibitem[Kelly, Rieke \& Campbell(1994)]{kerc94} Kelly, D. M., Rieke, G. H.,
1073:      and Campbell, B. 1994, \apj, 425, 231
1074:  \bibitem[Kurucz(1981)]{kur81} Kurucz, R. 1981, SAO Spec. Rep., No. 390
1075:  \bibitem[Leighly(1999)]{ley99} Leighly, K. M. 1999, \apjs, 125, 317
1076:  \bibitem[Lipari, Terlevich \& Macchetto(1993)]{lip93} Lipari, S., 
1077:      Terlevich, R., and Macchetto, F. 1993, \apj, 406, 451
1078:  \bibitem[Morris \& Ward(1989)]{morward89} Morris, S. L., and Ward, M. J. 
1079:      \apj, 340, 713
1080:  \bibitem[Nahar(1995)]{nahar95} Nahar, S. N. 1995, \aap, 293, 967	 
1081:  \bibitem[Osterbrock \& Pogge(1985)]{op85} Osterbrock, D. E. and
1082:      Pogge, R. W. 1985, \apj, 297, 166
1083:  \bibitem[Penston(1987)]{pen87} Penston, M. V. 1987, \mnras, 229, 1p 
1084:  \bibitem[Persson(1988)]{pe88} Persson, S. E. 1988, \apj, 330, 751
1085:  \bibitem[Peterson \& Wandel(1999)]{pw99} Peterson, B., and Wandel, A.
1086:      1999, \apjl, 521, L95
1087:  \bibitem[Peterson et al.(2000)]{pet2000} Peterson, B., McHardy, I. M.,
1088:      Wilkes, B. J., Berlind, P., Bertram, R., Calkins, M., Collier, S.
1089:      J., Huchra, J. P., Mathur, S., Papadakis, I., Peters, J., Pogge, R. W.,
1090:      Romano, P., Tokarz, S., Uttley, P., Vestergaard, M., and Wagner,
1091:      R. M. 2000, \apj, 542, 161
1092:  \bibitem[Phillips(1978)]{phi78} Phillips, M. M. 1978, \apj, 226, 736
1093:  \bibitem[Pogge \& Owen(1993)]{poo93} Pogge, R. W., and Owen, J. M.
1094:      1993, OSU Internal Report 93-01
1095:  \bibitem[Pradhan(2001)]{pra01} Pradhan, A . K. 2001, Private  communication.	 
1096: \bibitem[Rayner et al.(1998)]{rayner98} Rayner, J. T., Toomey, D. W., 
1097: Onaka, P. M., Denault, A. J., Stahlberger, W. E., 
1098: Watanabe, D. Y., and Wang, S.-I. 1998, Infrared Astronomical Instrumentation, 
1099: ed. A. M. Fowler Proc. SPIE, 3354, 468
1100: \bibitem[Rayner et al.(2001)]{rayner01}Rayner, J. T., Toomey, D. W., 
1101: Onaka, P. M., Denault, A. J., Stahlberger, W. E., 
1102: Watanabe, D. Y., and Wang, S.-I.  2001, PASP, In prep.
1103: 
1104:  \bibitem[Rodr\'{\i}guez-Ardila, Pastoriza \& Donzelli(2000)]{roar00}
1105:      Rodr\'{\i}guez-Ardila, A., Pastoriza, M. G., and Donzelli, C. J.
1106:      2000, \apjs, 126, 63
1107:  \bibitem[Rodr\'{\i}guez-Ardila et al.(2000)]{ro00} Rodr\'{\i}guez-Ardila,
1108:     A., Binette, L., Pastoriza, M. G., and Donzelli, C. J. \apj, 538, 581
1109:  \bibitem[Rodr\'{\i}guez-Ardila et al.(2001a)]{roar01a} Rodr\'{\i}guez-Ardila,
1110:      A., Viegas, S., Pastoriza, M. G., and Prato, L. 2001a. {\it in Prep.}
1111:  \bibitem[Rodr\'{\i}guez-Ardila et al.(2001b)]{roar01b} Rodr\'{\i}guez-Ardila,
1112:      A., Viegas, S., Pastoriza, M. G., and Prato, L. 2001b. {\it in Prep.} 
1113: 
1114:  \bibitem[Rodr\'{\i}guez-Pascual, Mass-Hesse \& Santos-Ll\`eo(1997)]{rms97}
1115:      Rodr\'{\i}guez-Pascual, P. M., Mass-Hesse, J. M., and Santos-Ll\`eo,
1116:      M. 1997, \aap, 327, 72
1117:  \bibitem[Rudy et al.(1982)]{rudy82} Rudy, R. J., Jones, B., Le Van, P.
1118:  D.,
1119:      Puetter, R. C., Smith, H. E., Willner, S. P., and Tokunaga, A. T.
1120:      1982, \apj, 257, 570
1121:  \bibitem[Rudy et al.(1991)]{rudy91} Rudy, R. J., Erwin, P., 
1122:      Rossano, G. S., and Puetter, R. C. 1991, \apj, 383, 344 
1123:  \bibitem[Rudy et al.(2000)]{rudy00} Rudy, R. J., Mazuk, S., Puetter,
1124:      R. C., and Hamann, F. 2000, \apj, 539, 166 (RMPH)
1125:  \bibitem[Schlegel, Finkbeiner \& Davis(1998)]{sfd98} Schlegel, D. J.,
1126:      Finkbeiner, D. P., and Davis, M. 1998, \apj, 500, 525
1127:  \bibitem[Sigut \& Pradhan(1998)]{sp98} Sigut, T. A. A. and Pradhan,
1128:      A. K. 1998, \apjl, 499, L139 (SP98)
1129:  \bibitem[Vaughan et al.(1999)]{vaug99} Vaughan, S., Reeves, J., 
1130:      Warwick, R., Edelson, R. 1999, \mnras, 309, 113
1131:  \bibitem[Wampler \& Oke(1967)]{waok67} Wampler, E. J., and Oke, J. B.
1132:      1967, \apj, 148, 695
1133:  \bibitem[Verner et al.(1999)]{ver99} Verner, E. M., Verner, D. A., 
1134:      Korista, K. T., Ferguson, J. W., Hamann, F., and Ferland, G. J.
1135:      1999, \apjs, 120, 101
1136:  \bibitem[Wills, Netzer \& Wills(1985)]{wnw85} Wills, B. J., Netzer, H.,
1137:      and Wills, D. 1985, \apj, 288, 94 
1138: 
1139: 
1140: 
1141:  \end{thebibliography}
1142: 
1143:  \clearpage
1144: 
1145:  \begin{figure}
1146:  \plotone{f1.ps}
1147:  \caption{Near-IR spectrum of the NLS1 galaxy 1H\,1934-063 in rest
1148:  wavelengths. The upper panel shows the spectrum in the wavelength
1149:  interval 0.8--2.4$\mu$m, and the box in the upper right a zoom of 
1150:  the K-band spectrum, with the identification of the most important 
1151:  emission lines. The lower panel displays the spectrum around 
1152:  1$\mu$m and identifies the \feii\ and other permitted and forbidden 
1153:  lines present in this region. \label{fig1}}
1154:  \end{figure}
1155: 
1156:  \clearpage
1157: 
1158:  \begin{figure}
1159:  \plotone{f2.ps}
1160:  \caption{The same as Figure~\ref{fig1} for Ark\,564. \label{fig2}}
1161:  \end{figure}
1162: 
1163:  \clearpage
1164: 
1165:  \begin{figure}
1166:  \plotone{f3.ps}
1167:  \caption{The same as Figure~\ref{fig1} for Mrk\,335. \label{fig3}}
1168:  \end{figure}
1169: 
1170:  \clearpage
1171: 
1172:  \begin{figure}
1173:  \plotone{f4.ps}
1174:  \caption{The same as Figure~\ref{fig1} for Mrk\,1044. \label{fig4}}
1175:  \end{figure}
1176: 
1177: 
1178:  \clearpage
1179:  \begin{figure}
1180:  \plotone{f5.ps}
1181:  \caption{Example of the deblending procedure applied to Pa$\beta$ and
1182:  Pa$\delta$ for Ark\,564. The histogram is the data and the thick
1183:  line the fit to the observed blend. Individual Gaussian components are
1184:  in thin lines. The residuals of the fits are drawn in dashed
1185:  lines.\label{deblend}}
1186:  \end{figure}
1187: 
1188: 
1189:  \clearpage
1190: 
1191:  \begin{figure}
1192:  \plotone{f6.ps}
1193:  \caption{Energy level diagram for \feii\ illustrating the
1194:  transitions to $z~^{4}F^{0}$ and $z~^{4}D^{0}$
1195:  from the upper $b^{\rm 4}G$ level leading to the 1$\mu$m 
1196:  lines. The strongest lines observed are drawn in bold face. The
1197:  excitation via Ly$\alpha$ fluorescence is also shown. \label{levels}}
1198:  \end{figure}
1199: 
1200:  \clearpage
1201:  \begin{figure}
1202:  \plotone{f7.ps}
1203:  \caption{Zoom around the \lb8900$-$\lb9300 region showing the presence 
1204:  of the \feii\ lines at \lb8927, \lb9130, \lb9177 and \lb9202. These lines
1205:  are all primary cascades from L$\alpha$ pumped levels at $\sim$10 eV. 
1206:  \label{feiilyalf}}
1207:  \end{figure}
1208: 
1209: 
1210: 
1211: 
1212:  \clearpage
1213:  \begin{figure}
1214:  \plotone{f8.ps}
1215:  \caption{FOS/HST UV spectra for Ark\,564 (top) and Mrk\,335 (bottom) in the
1216:  region around \lb1870. The arrows mark the expected position of the 
1217:  \feii\ lines \lb1841, \lb1845, \lb1870 and \lb1873, resulting from the
1218:  primary decay of the upper $u~^{\rm 4}G^{\rm 0}$ and $t~^{\rm 4}G^{\rm 0}$
1219:  terms, pumped  Ly$\alpha$ fluorescence. The presence of these lines 
1220:  is crucial for determining if this mechanism contributes to the production of 
1221:  the 1$\mu$m \feii\ lines. \label{arkuv}}
1222:  \end{figure}
1223: 
1224: 
1225:  \clearpage
1226: 
1227:  \begin{figure}
1228:  \plotone{f9.ps}
1229:  \caption{Ratios between the most important \feii\ lines
1230:  observed in the galaxies and the Galactic sources 
1231:  Lk\,H$\alpha$101, MWC\,349, and R Mon. The solid line is the value reported
1232:  by RMPH for I\,Zw\,1 and the dotted line corresponds to the uncertainty.
1233:  The starred points correspond to the Galatic sources. 
1234:  \label{ratiosfeii}}
1235:  \end{figure}
1236: 
1237: 
1238:  \clearpage
1239:  \begin{figure}
1240:  \plotone{f10.ps}
1241:  \caption{Comparison of the optical \feii\ spectra of 1H\,1934-063
1242:  (upper), I\,Zw\,1 (middle) and Ark\,564 (lower). Except for the
1243:  width of the individual lines (narrower in Ark\,564), the three
1244:  spectra look similar. Note that the relative intensities of 
1245:  the \feii\ lines seem to be constant in the three objects.
1246:  The H$\beta$ and H$\gamma$ lines of I\,Zw\,1 were removed
1247:  for clarity.  
1248:  \label{optfeii}}
1249:  \end{figure}
1250: 
1251:  \clearpage
1252: 
1253:  \begin{figure}
1254:  \plotone{f11.ps}
1255:  \caption{Optical \feii\ emission (\feii \lb4750/\hbeta) versus the
1256:  sum of the flux of \feii\ \lb9997, \lb10171 and \lb10501 divided by Pa$\delta$
1257:  (\feii\ IR/Pa$\delta$). Values of \feii \lb4750/\hbeta\ are listed in Table~\ref{feiirat}.
1258:  Mrk 335 clearly departs from an apparent correlation 
1259:  between these two quantities.  \label{feiicorr}}
1260:  \end{figure}
1261: 
1262:  \clearpage
1263: 
1264:  \begin{figure}
1265:  \plotone{f12.ps}
1266:  \caption{Comparison of the line profiles of Pa$\beta$ (thick histogram), 
1267:  \ion{O}{1} \lb11287 (solid line), \feii\ \lb11127 (dashed histogram) and
1268:  [\ion{S}{3}] \lb9531 (dotted histogram) in velocity space, showing clearly 
1269:  that Pa$\beta$ is significantly wider than the lines emitted
1270:  in the partially ionized zone of the the BLR. This suggests that 
1271:  low-ionization lines such as \feii\ and \ion{O}{1} form in
1272:  the outermost part of the BLR.
1273:    \label{profiles}}
1274:  \end{figure}
1275: 
1276:  \clearpage
1277: 
1278:  \begin{figure}
1279:  \plotone{f13.ps}
1280:  \caption{Comparison of the emission line profiles of \ion{Ca}{2}
1281:  \lb8662 (thick line) and \ion{O}{1} \lb8446 (histogram), 
1282:  in velocity space. \label{caii}}
1283:  \end{figure}
1284: 
1285:  \clearpage 
1286: 
1287: 
1288: 
1289:  \begin{deluxetable}{lcccc}
1290:  \tablecaption{Basic information of the NLS1 galaxies \label{tabledata}}
1291:  \tablewidth{0pt}
1292:  \tablehead{
1293:  \colhead{Galaxy} & \colhead{z} & \colhead{M$_{v}$\tablenotemark{a}} &
1294: 
1295:  \colhead{A$_{v}$\tablenotemark{b}} & {Morphology}
1296:  }
1297:  \startdata
1298:  1H\,1934-063 & 0.01059 &  -19.04   & 0.972 & Sb  \\
1299:  Ark\,564     & 0.02468 &  -20.42   & 0.198 & SB  \\
1300:  Mrk\,335     & 0.02578 &  -21.32   & 0.118 & S0/a \\
1301:  Mrk\,1044    & 0.01645 &  -18.84   & 0.113 & SB0 \\
1302:  \enddata
1303:  \tablenotetext{a}{A value of H$_{\rm o}$ = 75 \kms Mpc$^{-1}$ was assumed}
1304:  \tablenotetext{b}{Galactic extinction}
1305: 
1306:  \end{deluxetable}
1307: 
1308: 
1309:  \clearpage
1310: 
1311:  \begin{deluxetable}{lcccccccc}
1312:  \tabletypesize{\scriptsize}
1313:  \tablecaption{Observed emission line fluxes and FWHM for the 
1314:  permitted lines\tablenotemark{a}. \label{tableflux}}
1315:  \tablewidth{0pt}
1316:  \tablehead{
1317:  \colhead{} & \multicolumn{2}{c}{1H\,1934-063} & \multicolumn{2}{c}{Ark\,564} &
1318:  \multicolumn{2}{c}{Mrk\,335}  & \multicolumn{2}{c}{Mrk\,1044} \\
1319:  \colhead{Line} & \colhead{Flux} & \colhead{FWHM} & \colhead{Flux} &
1320: \colhead{FWHM} & \colhead{Flux} & \colhead{FWHM} & \colhead{Flux} & \colhead{FWHM}
1321:  }
1322:  \startdata
1323:  \ion{Fe}{2} \lb8927  & \nodata  & \nodata  & 0.18$\pm$0.05 &580  & 0.34$\pm$0.13 & 700 & 0.22\tablenotemark{2} & \nodata \\
1324:  \ion{Fe}{2} \lb9127  & 2.13$\pm$0.41 & 950 & 0.59$\pm$0.11 &600  &0.72$\pm$0.23  & 900 & 0.91$\pm$0.29 & 1050  \\
1325:  \ion{Fe}{2} \lb9177  & 2.77$\pm$0.41 & 950 & 0.75$\pm$0.15 &600  & 1.83$\pm$0.23 & 900 & 0.38$\pm$0.25 & 1050 \\
1326:  \ion{Fe}{2} \lb9202  & 1.90$\pm$0.41 & 950 & 0.52$\pm$0.15 &600  & 1.38$\pm$0.23 & 900 & \nodata & \nodata \\
1327:  Pa9 \lb9229          & 5.00$\pm$0.42 & 950 & 1.91$\pm$0.20 &820  & 4.52$\pm$0.31 & 1250 & 2.60$\pm$0.47 & 1900  \\
1328:  \ion{Fe}{2} \lb9256  & 0.80$\pm$0.25 & 500 & 0.46$\pm$0.20 &800  & 0.66$\pm$0.21 & 900 & \nodata & \nodata \\
1329:  \ion{Fe}{2} \lb9956  & 1.24$\pm$0.28 & 1000 &0.54$\pm$0.07 &750  & \nodata & \nodata    & 0.63$\pm$0.07 & 1000  \\
1330:  \ion{Fe}{2} \lb9997  & 4.52$\pm$0.28 & 1000& 1.72$\pm$0.07 &750  & 4.36$\pm$0.12 & 1600 & 1.36$\pm$0.07 & 1000  \\
1331:  \ion{He}{2} \lb10124 & 3.44$\pm$0.30 & 1100& 1.72$\pm$0.13 &1650 & 3.62$\pm$0.14 & 1850 & 1.06$\pm$0.14 & 1800  \\
1332:  \ion{Fe}{2} \lb10171 & 0.85$\pm$0.28 & 1000& 0.38$\pm$0.07 &750  & 0.77$\pm$0.12 & 1600 & 0.22$\pm$0.07 & 1000  \\
1333:  \ion{Fe}{2} \lb10501 & 2.62$\pm$0.25 & 1000& 1.30$\pm$0.09 &1100 & 1.83$\pm$0.23 & 1600 & 1.16$\pm$0.28 & 1620   \\
1334:  \ion{Fe}{2} \lb10863 & 4.24$\pm$0.25 & 950 & 0.99$\pm$0.08 &700  & \nodata &\nodata     & 1.31$\pm$0.17 & 900  \\
1335:  \ion{Fe}{2} \lb11127 & 1.30$\pm$0.23 & 850 & 0.54$\pm$0.07 &600  & 0.55$\pm$0.14 & 1100 & 0.52\tablenotemark{3}      \\
1336:  \ion{O}{1} \lb11287 &  6.91$\pm$0.24 & 850 & 2.70$\pm$0.07 &600  & 3.87$\pm$0.17 & 1100 & 1.52$\pm$0.17 & 900  \\
1337:  Pa$\alpha$ \lb18761\tablenotemark{3} & 25.43 & 970 & 13.46 &600  & 20.76 & 1200& 4.21          & 1200  \\
1338:  Br$\gamma$ \lb21655 & 2.24$\pm$0.14  & 850 & 1.05$\pm$0.06 &600  & 2.67$\pm$0.31 & 1150 & 0.78$\pm$0.91 & 1000   \\
1339:  \enddata
1340:  \tablenotetext{a}{Fluxes in units of 10$^{-14}$ erg cm$^{-2}$ s$^{-1}$ and FWHM
1341: in \kms}
1342:  \tablenotetext{1}{lower limit}
1343:  \tablenotetext{2}{upper limit}
1344:  \tablenotetext{3}{Flux measurement affected by atmospheric absorption, so it represents a
1345:  lower limit}
1346: 
1347: \end{deluxetable}
1348: 
1349: \clearpage
1350: 
1351:  \begin{deluxetable}{lcccccccc}
1352: \tabletypesize{\scriptsize}
1353:  \tablecaption{Observed emission line fluxes and FWHM for the 
1354:  permitted lines with broad and narrow components\tablenotemark{a}. \label{tablefluxbn}}
1355:  \tablewidth{0pt}
1356:  \tablehead{
1357:  \colhead{} & \multicolumn{2}{c}{1H\,1934-063} & \multicolumn{2}{c}{Ark\,564} &
1358:  \multicolumn{2}{c}{Mrk\,335}  & \multicolumn{2}{c}{Mrk\,1044} \\
1359:  \colhead{Line} & \colhead{Flux} & \colhead{FWHM} & \colhead{Flux} &
1360: \colhead{FWHM} & \colhead{Flux} & \colhead{FWHM} & \colhead{Flux} & \colhead{FWHM}
1361:  }
1362: \startdata
1363:  Pa$\delta_{\rm n}$ \lb10049    & \nodata    & \nodata & 1.00$\pm$0.03 & 320  & \nodata   & \nodata   & \nodata   & \nodata \\
1364:  Pa$\delta_{\rm b}$ \lb10049    & 7.58$\pm$0.27 & 1000 & 3.16$\pm$0.10 & 1650 & 5.86$\pm$0.09& 1100 & 2.85$\pm$0.09  & 1250 \\
1365:  \ion{He}{1}$_{\rm n}$ \lb10830 & 7.53$\pm$0.20 & 410  & 4.90$\pm$0.08 & 370  & 5.49$\pm$0.14  & 550 &  \nodata   & \nodata \\
1366:  \ion{He}{1}$_{\rm b}$ \lb10830 & 19.6$\pm$0.33 & 1230 & 5.46$\pm$0.41 & 2400 & 24.10$\pm$0.44 & 2020 & 8.35$\pm$0.19 & 1000 \\
1367:  Pa$\gamma_{\rm n}$ \lb10937   & 4.29$\pm$0.25 & 570  & 1.84$\pm$0.07 & 320  & \nodata   & \nodata  &  \nodata   & \nodata \\
1368:  Pa$\gamma_{\rm b}$ \lb10937   & 9.22$\pm$0.41 & 1600 & 4.00$\pm$0.29 & 1700 & 11.60$\pm$0.24 & 1500 & 3.92$\pm$0.28 & 1500 \\
1369:  Pa$\beta_{\rm n}$ \lb12820   & 8.36$\pm$0.22 & 600  & 3.95$\pm$0.08 & 400  & 4.29$\pm$0.13  & 600  & 2.22$\pm$0.10 & 600 \\
1370:  Pa$\beta_{\rm b}$ \lb12820   & 9.64$\pm$0.39 & 1800 & 5.90$\pm$0.30 & 1800 & 12.73$\pm$0.36 & 2050 & 3.18$\pm$0.27 & 2300 \\
1371: 
1372: \enddata
1373: \tablenotetext{a}{Fluxes in units of 10$^{-14}$ erg cm$^{-2}$ s$^{-1}$ and FWHM
1374: in \kms. The subcripts n,b stand for narrow and broad components, respectively}
1375:  
1376:  \end{deluxetable}
1377: 
1378: \clearpage
1379: 
1380:  \begin{deluxetable}{lccc}
1381:  \tablecaption{Measured and expected flux for the \feii\ lines \lb1840,1844\tablenotemark{a,b} \label{balance}}
1382:  \tablewidth{0pt}
1383:  \tablehead{
1384:  \colhead{Galaxy} & \colhead{Observed\tablenotemark{c}} & \colhead{Expected} & \colhead{E(B-V)}}
1385:  \startdata
1386:  1H\,1934-063 &   13.2   &  158.0  &   0.61 \\
1387:  Ark\,564     &   0.7    &  58.0   &   1.10 \\
1388:  Mrk\,335     &   1.7    &  74.9   &   0.93 \\
1389:  Mrk\,1044    &   0.6    &  57.0   &   1.12 \\
1390:  \enddata
1391:  \tablenotetext{a}{In units of 10$^{-14}$ erg cm$^{-2}$ s$^{-1}$}
1392:  \tablenotetext{b}{The values of columns 2 and 3 correspond to the sum of the flux of
1393:  the individual \feii\ features}
1394:  \tablenotetext{c}{Upper limit}
1395:  \end{deluxetable}
1396: 
1397:  \clearpage
1398: 
1399:  \begin{deluxetable}{lcccccccc}
1400:  \tablecaption{\feii\ line ratios for the observed sample \label{feiirat}}
1401:  \tabletypesize{\scriptsize}
1402:  \tablewidth{0pt}
1403:  \tablehead{
1404:  \colhead{Ratio} & \colhead{1H\,1934-063} & \colhead{Ark\,564} &
1405:  \colhead{Mrk\,335}  & \colhead{Mrk\,1044}  & \colhead{I\,Zw\,1\tablenotemark{1}}
1406:  & \colhead{LkH$\alpha$\,101\tablenotemark{2}} & \colhead{MWC 349\tablenotemark{5}} &
1407:  \colhead{R Mon\tablenotemark{5}}
1408:  }
1409:  \startdata
1410:  \lb9997/\lb10171  & 5.32$\pm$1.78 & 4.53$\pm$0.85 & 5.66$\pm$0.90 & 6.18$\pm$2.00 & \nodata & 3.70 & 3.5  & \nodata \\
1411:  \lb9997/\lb10501  & 1.72$\pm$0.20 & 1.32$\pm$0.11 & 2.38$\pm$0.31 & 1.17$\pm$0.29 & 2.22$\pm$0.29 & 1.11 & 1.6 & 1.5 \\
1412:  \lb9997/\lb10863  & 1.07$\pm$0.10 & 1.74$\pm$0.16 & \nodata       & 1.04$\pm$0.15 & 0.90$\pm$0.15 & 1.26 & \nodata & 0.98 \\
1413:  \lb9997/\lb11127  & 3.47$\pm$0.65 & 3.19$\pm$0.43 & 7.92$\pm$1.05 & 2.51     &   4.59$\pm$0.76    & 1.59 & 2.00 & 1.9 \\
1414:  \lb11127/\lb10863 & 0.31$\pm$0.06 & 0.55$\pm$0.08 & \nodata       & \nodata     &   0.20$\pm$0.04 & 0.80 & \nodata & 0.55 \\
1415:  \lb10501/\lb10863 & 0.62$\pm$0.07 & 1.31$\pm$0.14 & \nodata       & 0.89$\pm$0.24 & 0.41$\pm$0.07 & 1.14 & \nodata & 0.70 \\
1416:  \lb10501/\lb11127 & 2.02$\pm$0.41 & 2.40$\pm$0.35 & 3.33$\pm$0.94 & 2.23     &  2.06$\pm$0.37     & 1.43 & 1.20  & 1.3  \\
1417:  \feii\ \lb4750/H$\beta$ & 0.89\tablenotemark{3} & 0.81\tablenotemark{4} & 0.47\tablenotemark{4} & 0.91\tablenotemark{4} & 1.34\tablenotemark{4}
1418:  & \nodata & \nodata & \nodata \\
1419: 
1420:  \enddata
1421: 
1422:  \tablerefs{
1423:  (1) Rudy et al. 2000; (2) Rudy et al. 1991 (3) Rodr\'{\i}guez-Ardila,
1424:  Pastoriza
1425:  \& Donzelli 2000; (4) Joly 1991; (5) Kelly, Rieke \& Campbell 1994}
1426: 
1427:  \end{deluxetable}
1428: 
1429: 
1430:  \end{document}
1431: 
1432: