astro-ph0109039/ms.tex
1: \documentstyle[12pt,amssym,aasms4]{article}
2: 
3: %\received{}
4: %\accepted{}
5: %\journalid{}{}
6: %\articleid{}{}
7: 
8: %\slugcomment{As Submitted to the Astrophysical Journal Supplement Series}
9: 
10: \lefthead{Fernandes, Ara\'ujo, Pereira \& Landaberry}
11: 
12: \righthead{A Spectral Atlas of the Pre-WN Candidate HD 326823}
13: 
14: \def\cl{\centerline}
15: 
16: \begin{document}
17: 
18: \title{A Spectral Atlas of the Pre-WN Candidate HD 326823}
19: \altaffiltext{1} {Based on observations mainly done with the 1.52m 
20: telescope at 
21: the European Southern Observatory (La Silla, Chile), under the agreement 
22: with 
23: the 
24: Observat\'orio Nacional-MCT (Brazil) and also with the 1.6m telescope at 
25: the 
26: Laborat\'orio Nacional de Astrof\' \i sica (Bras\'opolis, Brazil).}
27: 
28: 
29: \vskip 1truecm
30: \author{Marcelo Borges Fernandes}
31: \affil{borges@dage0.on.br \\} 
32: \author{Francisco Xavier de Ara\'ujo}
33: \affil {araujo@on.br \\} 
34: \author{Claudio Bastos Pereira}
35: \affil{claudio@on.br \\}
36: \author{Sayd J. Codina Landaberry}
37: \affil{sayd@on.br \\
38: Observat\'orio Nacional, Rua General Jos\'e Cristino 77 \\
39: CEP 20921-400, Rio de Janeiro  BRAZIL \\} 
40: 
41: 
42: \begin{abstract}
43: 
44: HD 326823 is a peculiar emission line object. In the literature it was 
45: considered 
46: as a pre-WN candidate. Based mainly on high resolution data, 
47: obtained 
48: with the FEROS spectrograph, we made an atlas covering 3800-9200 \AA \ region. 
49: This atlas confirms the presence of several strong He\,{\sc i} 
50: lines and numerous N lines. There is also a deficiency of H, indicating 
51: that 
52: this object is indeed in an evolved stage. In addition, with low resolution 
53: spectra, we obtained intensities of several lines.
54: 
55: \end{abstract}
56: 
57: \keywords{stars: HD 326823 --- stars: transition objects --- Lines: 
58: identifications, intensities, equivalent widths}
59: 
60: \eject
61: \section{Introduction}
62: 
63: 
64: The evolution of massive stars (M$_{ZAMS}$ $>$ 10 M$_\odot$) is not yet 
65: completely understood. We know that they start their lives as 
66: O or B 
67: types and finish them in supernova explosions. Phases like LBV, WN, WC and 
68: B[e]sg need a 
69: better comprehension. The existence of few objects in these classes is a 
70: problem. This problem becomes bigger for transition evolutionary phases  
71: due to even smaller number of members. The possibility that HD 326823 
72: (Hen 3-1330) 
73: is an 
74: object in a transition phase is the main reason that motivated the present 
75: study. 
76: 
77: HD 326823 (l$^{II}$ = 344.3;b$^{II}$ = -1.2) belongs to an association of 
78: distant 
79: OB supergiants in the direction of the galactic center at a distance R 
80: $\sim$ 2 
81: kpc (McGregor, Hyland, \& Hillier 1988). Feast et al. (1961) labelled 
82: it as a
83: peculiar 
84: emission object, with emission lines of H, He\,{\sc i}, Fe\,{\sc ii}, 
85: Mg\,{\sc ii} and interstellar absorption lines of
86: Ca\,{\sc ii}. Stephenson (1974) noted that the great strength of the 
87: He\,{\sc i} emission 
88: could be attributed to He overabundance and H deficiency.
89: 
90: McGregor et al. (1988) found that the continuum was essentially 
91: flat 
92: in flux between 1 and 2.5 $\mu$m, with several He\,{\sc i} emission lines 
93: and a 
94: few weak 
95: H\,{\sc i} emission lines present. They concluded that He overabundance in 
96: the 
97: envelope of this star would suggest that HD 326823 is becoming 
98: a WR 
99: star. In the UV, Shore et al. (1990) reported a B-type absorption 
100: spectrum 
101: consistent with a B1.5 spectral type. In addition, they found that the 
102: spectrum is 
103: similar to the one of LBV S128/LMC = R127. 
104: 
105: Lopes, Damineli, \& de Freitas Pacheco (1992), based on Coud\'e data 
106: limited to few spectral regions, confirmed the result of Carlson \& Henize 
107: (1979), finding complex 
108: He\,{\sc i} 
109: profiles with a double structure. 
110: This 
111: structure indicates that the lines are probably formed in different 
112: circumstellar regions 
113: with 
114: rotation. The red Ca\,{\sc ii} 
115: triplet and the Fe\,{\sc ii} 9997 \AA \ feature also showed 
116: similar 
117: two-peak profiles. 
118: The spectra of Lopes et al. (1992) also showed rather strong N lines, 
119: indicating 
120: a 
121: possible 
122: overabundance of 
123: this element. They suggested that HD 326823 may be the 
124: precursor 
125: of a WN star.
126: 
127: Sterken et al. (1995) made a spectroscopic and photometric study and 
128: confirmed 
129: the very peculiar nature of the star. They also suggested that HD 326823 
130: may be 
131: a LBV in a quiescent state, having had S Dor-type eruptions in the past, 
132: and 
133: that it is now possibly moving towards a WN configuration. On the other hand, 
134: Lamers et 
135: al. (1998) 
136: and de 
137: Winter \& P\'erez (1998) included it in the group of unclassified B[e] 
138: stars.
139: 
140: More recently Walborn \& Fitzpatrick (2000) presented a digital atlas of 20
141: high-luminosity, peculiar OB spectra in the 3800-4900 \AA \ range. They put HD
142: 326823 in the group of ``Iron Stars'', but they cited that it ``has perhaps the
143: most bizarre spectrum in this atlas''. Moreover they have emphasized ``the
144: extreme weakness of all its spectral features, including hydrogen, which must
145: be very deficient''.
146: 
147: In order to contribute to a better comprehension
148: of the evolutionary stage of  HD 
149: 326823 , we present in this paper a spectral atlas in the 
150: 3800-9200 \AA \ wavelength range. This atlas is based mainly on high 
151: resolution data obtained with the FEROS spectrograph (Kaufer et al. 1999) 
152: attached to the 1.52m ESO telescope.
153:   
154:   
155:   
156:   
157: \section{Observations \& Reductions}
158: 
159: 
160: 
161: High and low resolution observations were both obtained with the ESO 1.52m 
162: telescope in La Silla (Chile). 
163: 
164: HD 326823 was observed in high resolution mode on 1999 
165: June 24, with an exposure time of 1800 seconds, in order 
166: to present a detailed line list. These data were also used to deblend some 
167: features present in the low resolution spectra. The Fiber-fed Extended Range 
168: Optical Spectrograph (FEROS) is a bench-mounted Echelle spectrograph with the
169: fibers located at the Cassegrain focus with a spectral resolving power
170: of R = 48\,000 corresponding to 2.2 pixels of 15$\mu$m and with a
171: wavelength coverage from 3600\AA\, to 9200\AA. FEROS has a complete automatic 
172: on-line 
173: reduction. Equivalent widths have been measured using an IRAF task that 
174: computes the line area above the adopted continuum. The FEROS spectrum of HD 
175: 326823 of the present paper has, in the 5500 \AA \ region, a S/N ratio in the 
176: continuum of 
177: aproximately 45.
178: 
179: Low resolution spectra of HD 326823 were taken using a Boller \& 
180: Chivens 
181: spectrograph at the 
182: Cassegrain focus on 1998 March 14, 1998 July 12, and 2000 June 9. Spectra 
183: of five other similar objects, that are mentioned below, were also 
184: obtained on different dates. 
185: Table 1 shows the log of the observations. Two instrumental setups 
186: were 
187: employed. The first one made use of grating \#23 with 600 l mm$^{-1}$, 
188: providing a 
189: resolution of $\sim$ 4.6 \AA \ in the ranges 3600-7400 \AA \ and 
190: 3800-8700 \AA. The 
191: other one used grating \#32 with 2400 l mm$^{-1}$, 
192: providing a 
193: resolution of $\sim$ 1.0 \AA, in the range 3060-4060 \AA. The 
194: Cassegrain spectra (hereafter the data obtained with the Boller \& Chivens
195: spectrograph will be denoted as Cassegrain data) were reduced using standard
196: IRAF tasks, such as bias  subtraction,  flat-field normalization, and
197: wavelength calibration. We have done flux  calibrations, and extinction
198: corrections with the E(B-V) values taken from the  literature. 
199: Table 2 shows the object designations (Column 1), the E(B-V) values 
200: (Column 2), and their respective references (Column 3). Spectrophotometric 
201: standards 
202: from 
203: Hamuy et al. (1994) were observed, too.
204: 
205: In the linearized spectra the line intensities have 
206: been 
207: measured by the conventional method, adjusting a gaussian function to the 
208: profile. Uncertainties in the line intensities come mainly from the 
209: position 
210: of the 
211: underlying continuum. We estimate the errors in the measures to be about 
212: 20\% for the
213: weakest lines (line fluxes $\approx$ 10 on the scale of H$\beta$ = 100) 
214: and 
215: about 10\% for the strongest lines. 
216: 
217: In the 5500 \AA \ 
218: region, the S/N 
219: ratio in the continuum is aproximately 55 for the Cassegrain 
220: spectrum 
221: obtained on 1998 March 14 and 50 for the one obtained 
222: on 
223: 2000 June 9. In the 3600 \AA \ region, the S/N ratio for the Cassegrain 
224: spectrum 
225: is aproximately 35. Since there is no completely line free 
226: region in 
227: the spectrum, the S/N derived is an upper limit.
228: 
229: In addition to the FEROS and Cassegrain data, we also obtained spectra 
230: with medium resolution, using 
231: a Coud\'e spectrograph. These Coud\'e data came from the 1.6m telescope at the
232: Laborat\'orio Nacional de Astrof\' \i sica, Bras\'opolis 
233: (Brazil). We obtained spectra for HD 326823 and six other objects using 
234: an instrumental setup with a grating with 600 l mm$^{-1}$, providing a 
235: resolution of $\sim$ 0.8 \AA \ in the range 9800-10200 \AA. The 
236: data 
237: reduction has been performed 
238: using standard IRAF tasks, too. Table 3 shows the log of the observations.
239: 
240: 
241: 
242: \section{The Spectral Atlas}
243: 
244: 
245: In order to identify the lines, we have used the line lists provided by 
246: Moore 
247: (1945), 
248: Thackeray (1967), and Landaberry, Pereira, \& Ara\'ujo (2001). We also 
249: looked up 
250: two sites on the web: NIST Atomic Spectra Database Lines Form 
251: (URL physics.nist.gov/cgi-bin/AtData/lines\_form) and The Atomic 
252: Line List v2.04 (URL www.pa.uky.edu/~peter/atomic/). 
253: 
254: Tables 4 and 5 show the 
255: observed wavelength (Column 1), the proposed identification (Column 2), and 
256: the measured line equivalent width (W) (Column 3) for emission and 
257: absorption 
258: lines, respectively, present in the HD 326823 FEROS spectrum. 
259: There are 10 lines that remain unidentified and they are labelled as ''Uid`` 
260: in 
261: Column 2. Figures 1a, 1b, and 1c present 
262: this FEROS spectrum; the strongest lines (W $>$ 1 \AA) are marked.
263: 
264: Table 6 presents the observed wavelength (Column 1), the proposed 
265: transitions (Column 2), and 
266: intensities in ergs cm$^{-2}$ s$^{-1}$ (I($\lambda$)) (Column 3) for emission 
267: lines 
268: in the low resolution spectrum in the 
269: spectral range 3060-4060 \AA. We note that due to the low 
270: efficiency of the detector in the violet region, we were able to identify only 
271: a few lines between 3600 and 4060 \AA.
272: 
273: Tables 7 and 8 concern the emission and absorption lines respectively, 
274: present 
275: in the low resolution spectra with spectral 
276: ranges 3600-7400 \AA \ and 3800-8700 \AA. Columns 1 
277: and 2 in 
278: the tables show the observed wavelengths and the measured line intensities, 
279: respectively, for the spectrum taken on 1998 March 14. Columns 3 and 4 
280: show the 
281: same data for the spectrum taken on 2000 June 9. Finally, Column 5 
282: presents the 
283: proposed identifications. The intensities shown in these two tables are 
284: relative to 
285: H$\beta$ = 100 and as H$\beta$ is in emission, we decided to put the minus 
286: signal before the intensity values of absorption lines. 
287: We 
288: found that many emission lines are more intense than H$\beta$, especially 
289: He\,{\sc i} lines. 
290: In general, there is a rather significant difference between 
291: line 
292: intensities 
293: of the two dates. In most cases there was an increase between the 
294: spectrum taken on 1998 March 14 and that on 2000 June 9. On the contrary, 
295: no 
296: significant change is seen in the continuum (Figure 2). With the help of the
297: FEROS spectrum, we have been able to deblend, for instance, the feature 
298: observed at 5046.9 \AA, estimating the contributions of the
299: N\,{\sc ii} 
300: 5045.1 \AA, Si\,{\sc ii} 5047.3 \AA, and 
301: He\,{\sc i} 5047.7 \AA \ lines. These values are shown in Table 7.
302: 
303: The column labelled ``Identification'' in the tables has the element 
304: transition, the multiplet, and the rest wavelength of the transition. It 
305: is 
306: important to say that since differents sources were consulted, it is 
307: possible 
308: that more than one ion can be allocated to a single feature. In these 
309: cases, we 
310: give some possible alternative identifications. 
311: 
312: Our spectra confirmed, in the optical region, the predominance of He\,{\sc i}, 
313: Fe\,{\sc ii}, and N\,{\sc ii} lines, and the weakness of H transitions. In 
314: the next 
315: section we will comment on the most interesting features and measurements 
316: obtained 
317: with FEROS, Cassegrain, and Coud\'e data and discuss the nature of HD 326823.
318: 
319: 
320: \section {Discussion}
321: 
322: \subsection {FEROS Spectrum}
323: 
324: Thanks to its high resolution, the FEROS spectrum leads to the identification 
325: of a great number of 
326: lines and to a better definition of their profiles. It also allowed us to 
327: obtain individual equivalent widths of features blended in the Cassegrain 
328: spectra and so to evaluate the intensity of 
329: each line.
330: 
331: He\,{\sc i} 3889 \AA, 4922 \AA, 5016 \AA, 5876 \AA, 6678 \AA, 
332: 7065 \AA 
333: \ and 7281 \AA \ transitions are among those with the highest equivalent 
334: widths. 
335: These lines 
336: present a 
337: two-peak structure  with blue and red edges reaching 200 km s$^{-1}$ 
338: (Figure 3). The He\,{\sc i} 3889 \AA \ central absorption is very strong; it
339: suggests the effect of diluted radiation (see e.g. Niemela \& Sahade 1980).
340: 
341: Regarding H lines, we are only sure of the identification of H$\alpha$ 
342: and H$\beta$. The presence of P17 is uncertain. H$\alpha$ and H$\beta$ 
343: also reveal a two-peak structure (Figure 4). It is important to compare 
344: the 
345: strength of He and H lines. For instance, He\,{\sc i} 4713 \AA \ 
346: is 
347: stronger than H$\beta$ while He\,{\sc i} 6678 \AA \ is 
348: comparable to 
349: H$\alpha$. This behavior is confirmed by the low resolution spectra (see 
350: Table 7).
351: 
352: N lines are numerous and conspicuous. We have identified several permitted 
353: N\,{\sc i} emission lines as well as permitted and forbidden N\,{\sc ii} 
354: transitions. It 
355: is curious that the 
356: majority of the 
357: permitted N\,{\sc ii} lines are in absorption. This seems to indicate a 
358: photospheric origin for them. The forbidden lines are all in emission, 
359: indicating 
360: that they 
361: are formed in the star's envelope. In Figure 5 we can see some N\,{\sc 
362: ii} profiles.
363: 
364: We have found many permitted and forbidden emission lines attributed to 
365: low 
366: ionization 
367: metals, specially Fe\,{\sc ii}. However, these lines are not very strong. 
368: Others 
369: elements 
370: identified 
371: are Si\,{\sc ii}, O\,{\sc ii}, and Cr\,{\sc ii}. We also found that two lines 
372: of 
373: the red Ca\,{\sc ii} triplet have a high equivalent width and show the two-peak 
374: structure, confirming 
375: Lopes 
376: et al. (1992). Unfortunately, we cannot measure the third line's (8540 
377: \AA) equivalent width due to the existence of a gap in this spectral region. In 
378: addition, we 
379: identified some interstellar absorption bands (4430 
380: \AA, 
381: 5780 
382: \AA, 5797 \AA) and also IS lines due to Ca\,{\sc ii} (3934 \AA \ 
383: and 3968 \AA), 
384: CH$^+$ 
385: (3958 \AA, 4233 \AA) and Na I (5890 
386: \AA, 5896 \AA).
387: 
388: 
389: \subsection {Cassegrain Spectra}
390: 
391: The importance of the low resolution spectra is to show the continuum 
392: behavior 
393: and to allow the measurement of line intensities, thanks to the flux 
394: calibration. 
395: Below, the spectra in two 
396: different 
397: wavelength ranges are briefly described. A comparison with other stars 
398: observed 
399: by us is presented, too.   
400: 
401: \subsubsection{3060 \AA \ - 4060 \AA}
402: 
403: This spectrum has a wavelength range of $\sim$ 1000 \AA \ and it shows the 
404: behavior of 
405: the continuum from the UV to the blue region. We were able to identify only a 
406: few lines 
407: in this region due to the noise. The most remarkable feature is the He\,{\sc i} 
408: 3888.7 \AA. In Figure 6, we show the spectra of HD 326823, HD 316285, and HD 
409: 327083. The 
410: HD 326823 
411: spectrum in this figure is the one obtained on 1998 July 12. The HD 316285
412: spectrum shows many lines of the Balmer series, with  P-Cygni profiles. In 
413: HD 327823 these lines are in 
414: absorption. 
415: So, once more we note clearly a H deficiency in HD 326823. 
416: 
417: \subsubsection{3600 \AA \ - 7400 $\AA$ and 3800 \AA \ 
418: - 8700 \AA}
419: 
420: These spectra cover a wavelength range of $\sim$ 4000 \AA \ and they show 
421: the 
422: behavior 
423: of the continuum from the blue to the red region. 
424: 
425: Figure 7 presents a comparison between HD 326823 and five other 
426: stars: 
427: HD 87643, classified as B[e]sg (Miroshnichenko 1998; Oudmaijer 1998 \& 
428: Oudmaijer et al. 1998) and as Iron Star (Walborn \& Fitzpatrick 2000); HD
429: 316285, classified as LBV (Hillier et al. 1998) and as Iron Star 
430: (Walborn \& Fitzpatrick 2000); 
431: HD 327083, classified as LBV or B[e]sg (Machado, de Ara\'ujo, \& 
432: Lorenz-Martins 
433: 2001); $\eta$ Car, classified as LBV (Walborn 1995); and HR Car, classified 
434: as 
435: LBV, too (Hutsemekers \& van Drom 1991). The 
436: HD 326823 
437: spectrum in this figure is the one obtained on 1998 March 14. H$\alpha$, 
438: H$\beta$, and the stronger He\,{\sc i} lines are marked. We can see 
439: in this figure that only HD 326823 has He\,{\sc i} lines more 
440: intense than 
441: H$\alpha$ and 
442: H$\beta$. These data contribute to the assertion 
443: that 
444: HD 326823 is in an advanced evolutionary stage.
445: 
446: 
447: \subsection {Coud\'e Spectrum}
448: 
449: Figure 8 shows our Coud\'e spectra of HD 326823 and five other similar 
450: objects. 
451: HD 326823 is the only one
452: that shows the Fe\,{\sc ii} 9997 \AA \ transition (with two-peak structure) 
453: much 
454: more intense than P$\delta$. Moreover, the He\,{\sc i} 
455: 10031 \AA \ line is at least as strong as P$\delta$, confirming the H 
456: deficiency 
457: in HD 326823. 
458: 
459:  
460: \subsection{The nature of HD 326823}
461: 
462: 
463: Our data confirm that HD 326823 exhibts remarkably strong He\,{\sc i}
464: lines. On the contrary, the few H\,{\sc i} transitions seen are unusually
465: weak. The He/H line intensity ratios provide clear evidence for He
466: overabundance. Such a conclusion is reinforced by the comparison made in
467: previous subsections with some objects generally classified as B[e]
468: supergiants (B[e]sg), Luminous Blue Variables (LBV) or ``Iron Stars'' (Lamers
469: et al. 1998; Walborn \& Fitzpatrick 2000). Moreover our spectra reveal a
470: great number of conspicous N\,{\sc ii} transitions. N\,{\sc i} emissions and
471: (perhaps) few N\,{\sc iii} absorption lines may also be seen, but the dominant
472: ion of this species is surely N\,{\sc ii}. On the other hand, C lines are not seen at
473: all. Concerning oxygen, only the O\,{\sc i} 4661.2 \AA \ and the O\,{\sc ii}
474: 6046 \AA \ transitions seem to be present in emission. The identification of
475: O\,{\sc ii} 7319 \AA \ to the feature seen at 7321.8 \AA \ is rather doubtful, since
476: the contribution of the Fe\,{\sc ii} 7320.7 \AA \ (m7) is probably much more
477: important (we note that other Fe\,{\sc ii} lines of the same multiplet are
478: present). Thus, it is likely that N is overabundant too. These conclusions
479: point in favour of considering HD 326823 as an evolved object.
480: 
481: Let us compare now HD 326823 with other peculiar, evolved massive stars. Due
482: to the absence of the N\,{\sc iv} 4058 \AA, the N\,{\sc iii} 4630-4634 \AA, and the 
483: He\,{\sc ii} 4686 \AA \ emissions, it can not be considered as a WNL or WN9ha
484: star (Walborn \& Fitzpatrick 2000). For the same reason it may not be
485: classified as Ofpe/WN9 (Walborn 1982). As N lines are concerned, the spectra
486: of HD 326823 is more or less similar to those of some WNVL stars (Walborn \&
487: Fitzpatrick 2000). However, in such objects the H\,{\sc i} recombination lines
488: are the most prominents. Finally, the group of ``Iron Stars'', sometimes cited
489: in the literature, is not very useful since the objects generally put in this
490: category have diverse physical natures. We are tempted to conclude that HD
491: 326823 is an unique peculiar star. 
492: 
493: In the last years several works have been devoted to massive stars in
494: transition. From optical spectroscopy of Ofpe/WN9 and related stars, Nota et
495: al. (1996) suggested that Of, Ofpe/WN9, LBV, and B[e] stars must be
496: considered within the same evolutionary scenario. On the basis of infrared
497: spectra, Morris et al. (1996) concluded that Of supergiants, Ofpe/WN9, WNL,
498: B[e] supergiants, and LBV, are transitional in their morphological
499: classification and may be related in their evolution. Some years before
500: Schulte-Ladbeck et al. (1993) proposed that the distinction between the B[e]
501: supergiants and other transition objects is a stronger degree of nonspherical
502: symmetry in their envelopes, probably due to the higher rotational velocities.
503: We note the apparent disk line profiles seen in our data, which suggest that
504: HD 326823 may have been a rapid rotator.
505: 
506: From the theoretical point of view two different scenarios have been proposed
507: to describe the stages between O and WR stars. Crowther et al. (1995)
508: suggested that the WNL phase follows directly from the Of stars, at least for the 
509: objects with the highest initial masses. The WNL would then evolve to WNE and
510: finally WC stars. Langer et al. (1994) proposed the following sequence:
511: O $\rightarrow$ Of $\rightarrow$ H-rich WN $\rightarrow$ LBV $\rightarrow$ H-poor
512: WN $\rightarrow$ H-free WN $\rightarrow$ WC $\rightarrow$ SN. In this picture, HD
513: 326823 could be in a post-LBV stage , close to the H-poor WN phase. Such a
514: statement is hampered by the absence of an abundance analysis, that is beyond
515: the scope of the present paper.
516: 
517: In summary it is certain that HD 326823 is an indeed peculiar object, which is
518: in an evolved evolutionary phase of the massive stars. It might be the only
519: known example of a very precise, short-lived stage. However, its real nature
520: can not be ascribed until we have accurate determinations of some physical
521: parameters as the mass-loss rate and the helium and nitrogen abundances. 
522: 
523: The authors thank the useful comments and suggestions of an anonymous
524: referee, which improved the present paper. M.B.F. acknowledges financial
525: support from CAPES-Brazil (PhD studentship). 
526: 
527: 
528: \begin{thebibliography}{}
529: 
530: 
531: \bibitem{} Carlson, E. D., \& Henize, K. G. 1979, Vistas in Astronomy, 
532: 23, 213
533: 
534: \bibitem{} Crowther, P. A., Smith, L. J., Hillier, D. J., \& Schmutz, W. 1995,
535: A\&A, 293, 427
536: 
537: \bibitem{} de Winter, D., \& P\'erez, M. R. 1998, B[e] Stars (Dordrecht: 
538: Kluwer Academic Publishers)
539: 
540: \bibitem{} Feast, M. W., Stoy, R. H., Thackeray, A. D., \& Wesselink, A. 
541: J. 
542: 1961, MNRAS, 122, 239
543: 
544: \bibitem{} Feinstein, A., Marraco, H. G., \& Muzzio, J. C. 1973, A\&AS, 
545: 12, 331
546: 
547: \bibitem{} Hamuy, M., Suntzeff, N. B., Heathcote, S. R., Walker, A. R., 
548: Gigoux, P., \& Phillips, M. M. 1994, PASP, 106, 566
549: 
550: \bibitem{} Hillier, D. J., Crowther, P. A., Najarro, F., \& Fullerton, A. 
551: W. 1998, A\&A, 340, 483
552: 
553: \bibitem{} Hutsemekers, D., \& van Drom, E. 1991, A\&A, 248, 141
554: 
555: \bibitem{} Kaufer, A., Stahl, S., Tubbesing, S., Norregard, P., Avila, 
556: G., Francois, P., Pasquini, L., \& Pizzella 1999, The Messenger, 95, 8
557: 
558: \bibitem{} Kozok, J. R. 1985, A\&AS, 62, 7
559: 
560: \bibitem{} Lamers, H. J. G. L. M., Zickgraf, F.-J., de Winter, D., 
561: Houziaux, L., \& Zorec, J. 1998, A\&A, 340, 117 
562: 
563: \bibitem{} Landaberry, S. J. C., Pereira, C. B., \& de Ara\'ujo, F. X. 
564: 2001, A\&A, submitted
565: 
566: \bibitem{} Langer, N., Hamann, W. -R., Lennon, M., Najarro, F., Pauldrach, A .W. A., 
567: \& Puls, J. 1994, A\&A, 290, 819
568: 
569: \bibitem{} Lopes, D. F., Damineli, A., \& de Freitas Pacheco, J. A. 1992, 
570: A\&A, 261, 482
571: 
572: \bibitem{} Machado, M. A. D., de Ara\'ujo, F. X., \& Lorenz-Martins, S. 
573: 2001, 
574: A\&A, 368, L29
575: 
576: \bibitem{} McGregor, P. J., Hyland, A. R., \& Hillier, D. J. 1988, ApJ, 
577: 324, 1071
578: 
579: \bibitem {} Miroshnichenko, A. S. 1998, B[e] Stars (Dordrecht: Kluwer 
580: Academic Publishers)
581: 
582: \bibitem {} Moore, C. E. 1945, A Multiplet Table of Astrophysical 
583: Interest, Part I - Table of Multiplets (Revised Ed., Princeton, New 
584: Jersey: Princeton University Observatory)
585: 
586: \bibitem{} Morris, P. W., Eenens, P. R. J., \& Blum, R. D. 1996, ApJ, 470, 597
587: 
588: \bibitem{} Niemela, V. S., \& Sahade, J. 1980, ApJ, 238, 244
589: 
590: \bibitem {} Nota, A., Pasquali, A., Drissen, L., Leitherer, L., Robert, C.,
591: Moffat, A. F. J., \& Schmutz, W.  1996, ApJS, 102, 383
592: 
593: \bibitem {} Oudmaijer, R. D. 1998, B[e] Stars (Dordrecht: Kluwer Academic 
594: Publishers)
595: 
596: \bibitem{} Oudmaijer, R. D., Proga, D., Drew, J. E., \& de Winter, D. 
597: 1998, MNRAS, 300, 170 
598: 
599: \bibitem{} Schulte-Ladbeck, R. E., Leitherer, C., Clayton, G. C., Robert, C.,
600: Meade, M. R., Drissen, L., Nota, A., \& Schmutz, W. 1993, ApJ, 407, 723
601: 
602: \bibitem{} Sheikina, T. A., Miroshnichenko, A. S., \& Corporon, P. 2000, 
603: The Be 
604: Phenomenon in Early-Type Stars, IAU Colloquium, 175, in press
605: 
606: \bibitem{} Shore, S. N., Brown, D. N., Bopp, B. W., Robinson, C. R., 
607: Sanduleak, N., \& Feldman, P. D. 1990, ApJS, 73, 461
608: 
609: \bibitem{} Stephenson, C. B. 1974, ApJ, 191, 685
610: 
611: \bibitem{} Sterken, C., Stahl, O., Wolf, B., Th. Szeifert, \& Jones, A. 
612: 1995, A\&A, 303, 766
613: 
614: \bibitem{} Thackeray, A. D. 1967, MNRAS, 135, 51
615: 
616: \bibitem{} van Genderen, A. M., et al. 1990, A\&AS, 82, 189
617: 
618: \bibitem{} Walborn, N. R. 1982, ApJ, 256, 45
619: 
620: \bibitem{} Walborn, N. R. 1995, RevMexAA, 2, 51
621: 
622: \bibitem{} Walborn, N. R., \& Fitzpatrick, E. L. 2000, PASP, 112, 50
623: 
624: \bibitem{} Zorec, J. 1998, B[e] Stars (Dordrecht: Kluwer Academic 
625: Publishers)
626: 
627:                                                                         
628: 
629: 
630: \end{thebibliography}
631: 
632: \clearpage
633: 
634: \figcaption[f1a.eps]{The spectral atlas from 3800 to 5600 \AA \ 
635: obtained from the FEROS spectrum. The strongest lines (W $>$ 1 \AA) are marked.
636: \label{fig1a}}
637: 
638: \figcaption[f1b.eps]{The spectral atlas from 5600 to 7400 \AA \ 
639: obtained from the FEROS spectrum. The strongest lines (W $>$ 1 \AA) are marked.
640: \label{fig1b}}
641: 
642: \figcaption[f1c.eps]{The spectral atlas from 7400 to 9200 \AA \ 
643: obtained from the FEROS spectrum. The strongest lines (W $>$ 1 \AA) are marked.
644: \label{fig1c}}
645:  
646: \figcaption[f2.eps]{The HD 326823 low resolution spectra, obtained on 
647: 1998 March 14 and on 2000 June 9. These spectra were corrected for 
648: extinction, using E(B-V) values.
649: \label{fig2}} 
650: 
651: \figcaption[f3.eps]{Profiles of six He\,{\sc i} lines presented in 
652: the FEROS spectrum.   
653: \label{fig3}} 
654: 
655: \figcaption[f4.eps]{H$\alpha$ and H$\beta$ profiles shown in the 
656: FEROS spectrum.
657: \label{fig4}} 
658: 
659: \figcaption[f5.eps]{Profiles of permitted and forbidden N\,{\sc ii} 
660: lines obtained from FEROS spectrum.
661: \label{fig5}} 
662: 
663: \figcaption[f6.eps]{Low resolution spectra of HD 326823, HD 316285, 
664: and 
665: HD 327083 with $\lambda$$_{central}$ $\sim$ 3600\AA.
666: \label{fig6}} 
667: 
668: \figcaption[f7.eps]{Normalized low resolution spectra of HD 326823, 
669: HD 87643, HD 316285, HD 327083, HR Car, and $\eta$ Car.
670: \label{fig7}} 
671: 
672: \figcaption[f8.eps]{Normalized Coud\'e spectra of HD 326823, HD 
673: 87643, 
674: HD 316285, HD 327083, CPD-52$^o$9243, HR Car, and GG Car.
675: \label{fig8}}
676: 
677: 
678: \end{document}
679: