Cowley, A.J., Hutchings, I.B.: 1979, P.A.SP. 90,636. Debray, B., Dubout, R., L1ebaria, A., Petit, M.: 1988, Astron. Astrophys. (submitted). Feitzinger, J. V., Schlosser, W., Schmidt-Ka­ ler, Th., Winkler, C.: 1980, Astron. Astro­ + + phys. 84, 50. + . + Feitzinger, J. V., Isserstedt, J.: 1983, Astron. 4 + ~ + +. Astrophys. Supp. Ser., 51, 505. Henize, K. G.: 1956, Astrophys. J. Suppl. 2, 315. Heydari-Malayeri, M., Magain, P., Remy, M.: 1988, Astron. Astrophys. 201, L 41. Lortet, M. C., Heydari-Malayeri, M., Testor, 0 G.: 1986, in Luminous stars and associa­

2 A Brey 70a tions in galaxies, Proceed. lAU Symp. 116, (l) I A 401.

(l) + Lucke, P.B.: 1972, Ph.D. Thesis, University \J A of Washington. ::J ...... Lundström, 1., Stenholm, B.: 1984, Astron. C 0'1 Astrophys. Suppl. 58, 163. 0 + Morgan, D.H., Good, A.R.: 1987, ::::;E MNRAS. o Brey 73 224,435. 0 + Prevot-Burnichon, M. L., Prevot, L., Rebeirot, E., Rousseau, J., Martin, N.: 1981, Astron. + o Brey 71 Astrophys. 103, 83. Smith, L.F.: 1968, MNRAS. 140, 409. Stetson, P.B.: 1987, P.A.S.P. 99,191. Van der Hucht, K.A., Hidayat, B., Gunowan Admiranto, A., Supelli, K. R., Doom, C.: - 2 l.-----L_--"--_-'-----'_-'-_-'------''------'-_--'-_'------'-_--'-_''------L_--''-----' 1988 Preprint. -2 o 2 4 6 Walborn, N.R.: 1977, Astrophys. J. 215, 53. Magnitude Continuum Walborn, N. R.: 1980 (Private communica­ tion). Figure 6: Plot of the Hell relative magnitudes versus the continuum magnitudes of all stars of Weigelt, G., Baier, G.: 1985, Astron. Astro­ the frame (90 arcsec x 140 arcsec). The known W-R stars Brey 70a, Brey 71 and the Brey 73 phys. 150, L 18. association show a conspicuous Hell brightness excess. The stars belonging to the aggregate Westerlund, B. E.: 1974, Proceedings of the are represented by triangles. Most of the stars follow a close relationship v (A 4686 A) = 0.96 v first European Astronomical meeting in ()... 4794 A) + 0.34 and with a 0 rms of 0.22 magnitude after elimination of Brey 70a, Brey 71 and Athen, September 4-9, 1972, Volume 3. Brey 73. Star 12 is a faint star of the cluster, probably a red one which would need deep, good Springer, Berlin, Heldeberg, New York, seeing Band V photometry. p.39.

Spectroscopic Identification of White Dwarfs in Galactic Clusters O. REIMERS, Hamburger Sternwarte, Universität Hamburg, F. R. Germany 0. KOESTER*, Oepartment ofPhysics andAstronomy, Louisiana State University, USA

The End Products of Stellar fully as a white dwarf or undergoes a local SN 11 rate (e. g. Tammann, 1974) Evolution core collapse with a quasi instantane­ teils us that all stars more massive than ous energy release of - 1053 erg, visible 5 to 10 MG must become supernovae, The end products of the evolution of as a supernova explosion. and Kennicutt (1984) found that SN II's single stars are weil known: low mass Estimates of the maximum initial in Sc galaxies come from stars with stars leave white dwarf remnants, mas­ mass limit M for formation of white masses greater than 8 ± 1 MG' sive stars undergo a supernova explo­ WD dwarfs have been made by various Since the fate of intermediate mass sion with either a neutron star or a black techniques. stars is mainly determined by mass loss hole as a remnant. One possibility is to compare the in the red giant stage, the combination In a first approximation, it appears to supernova type I1 rate with the birth rate of stellar evolution tracks through the be only the initial mass which deter­ of massive stars in our galaxy. Since the red giant stages with empirical red giant mines whether a star ends its Iife peace- former rate is extremely uncertain and mass-Ioss rates also provides an esti­ can be observed only in external galax­ mate of MwD . The difficulty of this ap­ ies whereas the initial mass function can proach is that while mass-Ioss rates in be measured only in the solar neigh­ • Formerly at Institut für Theoretische Phy­ the normal red giant region are fairly weil sik und Sternwarte der Universität Kiel, F. R. bourhood, the constraints on SN 11 par­ known and can be parametrized by Germany ents are not stringent. An estimate of the semi-empirical interpolation formulae 47 .~ . .'0:' ...• .. •• • .. • .. .. •• " • • • ., .... • • • ' • .',: . . .' '. ',f. ; '. • 0•• '. #.. • '. .. " , , • ... •• . , ..: ,: • . ' . ~' • " .' .. ' .. , ' : .. ' .. .~ ... ' " ,.. •• • .~~ " ..... t. • • ... • " . "~\.. ',: .... :...• '7'7.:' . .'. . '.' .' . . ,10. .. .,.. .. '.

. ,- '.• . . " .. ,.'

Figure 1: Central part of NGC 3532 (ESO Schmidt, 11I a-J+ UG 1; 75 min) with marked positions of candidate objects.

(e. g. M- Ug . R, Reimers, 1975, 1987), ades. He found MWD = 4 MG' However, fields - in a number of intermediate age an extrapolation to advanced red giant with a turn-off mass of - 2 MG the Hy­ northern clusters. Based on their find­ phases (e. g. OH-IR stars) where most ades are not favourable for this method ings, they concluded MWD = 7 MG' mass is lost is certainly not allowed. On since according to the cluster luminosity However, the question remained the other hand, high mass-Ioss rates of function, their known six white dwarf whether the excess of blue objects in advanced red giants as observed in members must have had pragenitor the cluster fields were really white thermal GO lines or in OH maser lines masses only slightly higher than the dwarfs. Anticipating the results of our cannot easily be linked to stellar evolu­ turn-off mass. spectrascopic observations over the tion calculations, since for these stars years 1980-88 from both La Silla and masses and evolutionary stages are Galar Alto of all Romanishin and Angel's Faint Blue Objects in Open poorly known. candidates, it finally turned out that only Cluster Fields The only reliable method to determine 6 out of 17 WO candidates in the four Mwo seems to be the identification of A new attempt to determine MWD fram clusters NGG 2168, 2242, 2287 and white dwarf members of galactic clus­ cluster white dwarfs was made by 6633 are cluster members. This means ters with turn-off masses Mt :s MWD' Romanishin and Angel (1980) who that without time consuming spec­ The first attempt with this method was looked for a statistical excess of faint trophotometry of the faint 01 = 19 to 21) made by v. d. Heuvel (1975) far the Hy- blue objects - relative to comparison blue objects in the cluster fields, no safe 48 Confirmed white dwarf members of intermediate age clusters tions of WOs in cluster fields at such m faint magnitudes (-20 ) in the near fu­ 3 Cluster Age Turn-off White Te [10 K] V Mass Progenito ture. (Paper No.) logT Mass Dwarf [MG) Mass [MG) The Present Status

NGC 2516 8.0 4.75 -1 30 19.6 0.95 7.2 How many white dwarf members of (11) -2 36 19.55 1.14 7.6 intermediate age clusters (turn-off -5 34 20.1 1.17 9.8 masses ;;:: 4 MG) are known at present after the spectroscopic observations 4 NGC 2451 ) 7.8 5.3 -6 31 16.75 0.6: 5.9 with the 3.6-m + lOS in the years (111) 1980-84, the new start with EFOSC on NGC 3532 8.3 3.6 -1 28 19.55 0.9 4.1 fainter objects in 1988, and a few nights M -5 28 19.2 0.6 3.9 on the northern clusters NGC 2168 and -6 29 19.9 0.9 4.1 NGC 6633 with the Calar Alto 3.5-m telescope? Altogether we have iden­ NGC 2168 8.0 4.75 -3 38 .20.25 0.7 5.9 tified 12 (possibly 14) WO members of (IV) -4 44 20.05 0.7 5.9 intermediate age clusters (Papers I-VI).

1 Four were from Romanishin and Angel's NGC 2287 8.25 3.9 ) -2 25 20.1 0.6 (I) -5 25 20.1 0.6 (1980) candidate list (two in both NGC 2287 and NGC 2168). Eight (ten) hot OA NGC 6405 7.9: 5.3: -1 3) 82 18.1 1.18 6.5: white dwarfs (the hottest has 82,000 K!) (VI) have been discovered by ourselves on 2 Pie/ades 7.9 5.3 LB 1497 0.85 6.2 ) the ESO Schmidt plates taken for that particular purpose. Only one WO, the ') According to Romanishin and Angel (1980). 2) Weidemann (1987). Pleiades member LB 1497 (e. g. Green­ 3) Cluster membership uncertain, WO is _1 0 from cluster centre. stein, 1974). was known to be member .) NGC 2451 contains two more probable WO members (Paper 111), if conrirmed they have high masses of an intermediate age open cluster be­ and high progenitor masses; : denotes uncertain .or preliminary values fore we started this project in 1980. ESO's superb equipment and dark sky has thus made possible an important conclusions are possible. The results of 30,000 K) are extremely rare in the gen­ contribution to an understanding of the the first two observing years have been eral field and projection onto a cluster is final stages of stellar evolution. described briefly by Koester (1982). therefore highly improbable, two criteria Our first observing run with EFOSC Parallel to this activity we asked for are applied: (i) the theoretically calcu­ and the 3.6-m telescope in April 1988 deep ESO red and UV Schmidt plates of lated cooling time, which depends on conducted by O. K. was highly success­ suitable southern intermediate age clus­ mass M and temperature Te must be ful. We identified 3 WO members of ters with turn-off masses around 5 MG' smaller than the cluster age and (ii) the NGC 3532. Furthermore, we completed White dwarf members of clusters distance modulus V-Mv determined observations on several candidates in younger than about 2 . 108 years cannot from the observed apparent magnitude the clusters NGC 2422, 2516 and 2287 have cooled down below about V and from effective temperature Te and which had been too faint for an unam­ 25,000 K. Therefore one has to search radius R has to be consistent with the biguous identification with the lOS on the for very blue faint objects in the cluster observed cluster modulus. As Chan­ 3.6-m in earlier observing runs. The can­ fields. With typical distance moduli of drasekhar's work, which won him the didate in NGC 2422 could be proven to rich clusters between 8 and ~ 11, hot Nobel prize, has shown, degenerate be a subdwarf, as suspected earlier (Pa­ white dwarfs are expected in the range stars obey a mass-radius relation which per I). The same applies to NGC 2516-4. 1I3 V = 19 to 22. is approximately R - M- . The white A highlight of the April 88 observa­ The plates were blinked at Hamburger dwarf radius R which is needed for clus­ tions was the discovery of the extremely Sternwarte and we could identify further ter membership test is determined in the hot OA NGC 6405-1 (82,000 K, the suitable candidates in the clusters NGC ideal case from gravity which can be hottest known OA?). While the star is 2516,2451,3532,6405,6475,6087and measured from high-quality spectra by - 10 outside the cluster centre, it has IC 2391. Progress in follow up spectros­ comparison with model atmosphere cal­ the same distance according to the copy was slow, since due to the faint­ culations. This was possible, e. g., in model atmosphere analysis. One may ness of the objects only the 3.6 m + lOS cases of NGC 2516 and NGC 3532. If suspect that the near coincidence of was suitable in the years 1980-84, and gravity cannot be determined spectros­ such an extremely rare type of star with the faintest objects as weil as objects copically with sufficient accuracy, a typ­ a cluster is not just by chance. with relatively weak lines had to wait for ical WO mass is assumed in a first ap­ A further surprise was that one of our EFOSC with its improved sensitivity. proximation. In most cases adecision faint blue candidates in NGC 6087 Our first observing run with EFOSC was about cluster membership is then possi­ turned out to be a planetary nebula. It conducted in April 88. ble since for a given effective tempera­ has a hot featureless continuum and ture Te the absolute magnitude varies as nebular lines that extend - 14 arcsec in 2 2/3 Mv - R - M- with mass, and the direction of the spectrograph slit. A Cluster Membership assumptions about WO masses which rough estimate shows that at the dis­ Even if a faint blue object within the have a narrow range anyway are not too tance of the cluster (850 pe) this corres­ cluster field turns out to be a white crucial. As long as consistency with the ponds to a linear diameter of - 0.03 pe. dwarf, it is not necessarily a cluster cluster modulus is found for an This would imply a rather young PN with member. How can foreground and assumed WO mass in the range 0.5 MG a central star of Mv "" - 3 to -4, incon­ background white dwarfs be separated to 1.2 MG' cluster membership is highly sistent with the observed V magnitude from cluster members? Besides the fact probable for a hot WO. Unfortunately, it of 19.3. Conclusion: Unfortunately, the that extremely hot white dwarfs (Te > will be difficult to measure proper mo- PN is a background object. 49 M WD and the Initial-Final Mass I I I Relation NGC 3532 - 1 Oata on the individual cluster white 1.13 dwarfs identified by our spectroscopic observations and on the open clusters are compiled in the Table. A major uncertainty for the determina­ tion of cluster turn-off masses, cluster ~ ~~ 0.96 ~ ~fi~ ~ ages and white dwarf progenitor masses has been recognized in recent years: the amount of convective over­ shooting for intermediate mass main se­ ~ NGC 3532 - 5 quence stars is not known. We have 0.79 always assumed here an overshooting parameter Uc = 0.5. For a full discus­ 1.13 sion of this point, in particular the influ­ " ence of overshooting on the initial-final ~ mass relation, we refer to Weidemann (1987). 0.62 ~ ~~ The best case for determining MWD is - 0.96 the rich southern cluster NGC 2516

(Reimers and Koester, 1982). Using the Vl C observed cluster luminosity function, QJ the mass-interval above the cluster turn­ C off was estimated out of which the ob­ (11 0.79 NGC 3532-6 served number of evolved stars has .....> come. The result was MWD = 8~~ MG' o 4i This statistical method, however, de­ ll: mands that no mass segregation has 1.13 occurred within the cluster due to dy­ 0.62 namical evolution. In cases where it is observed that the massive stars are more concentrated to the cluster centre than the low mass stars as in the case of the Hyades or NGC 3532, the statistical 0.96 method can give only a lower limit to MWD (Paper V), since for a sufficiently short relaxation time the white dwarfs can have aspace distribution different from that of their progenitor stars or the 0.79 upper main sequence stars respectively. An alternative approach is to try to esti­ mate the WO progenitor masses by vir­ tually placing the white dwarfs back onto the main sequence taking into 0.62 account main sequence and red giant lifetimes and the WO cooling ages. For details we refer to Weidemann (1987). Thus obtained progenitor masses (Table) yield both an initial-final mass 0."5 5100 relation (Fig. 3) and an estimate of MWD' 3900 I. 1"0 1.380 1 1.620 1.860 Notice in Figure 3 that the initial-final A[Al mass relation may weil be a strip with a Figure 2: EFOSC spectra of DA white dwarf members of NGC 3532. finite width, which could be explained by differential mass-Ioss in the preced­ ing red giant stage. Clearly more stars and more accurate data are needed, so, called the "jeweIl box" because of its flows. Oue to the higher resolution of the e.g., the accuracy of WO masses (via concentration of blue main sequence CCO, EFOSC images look better. radii) depends on the accuracy of dis­ stars and red giants. It contains - 10 red Follow-up spectroscopy of 23m can­ tance moduli and reddening corrections giants and Cepheids. Oue to its extreme didates at a sufficient S/N for gravity of open clusters! richness and a turn-off mass of - 6 MG determination using the model atmo­ we expect about 15 hot white dwarfs in sphere technique will be a promising m the cluster, however, at V -23 . Our task for the VLT. The Future attempt to find candidates with deep We clearly need more WO members CCO exposures with the Oanish 1.5-m Acknowledgements of rich open clusters with turn-off telescope has so far been unsuccessful masses above 5 MG' One of the best mainly due to the concentration of bright This work would have been impos­ known examples is NGC 6067, also stars wh ich caused severe CCO over- sible without the efficient support of

50 staff astronomers and skillful night 2 4 6 8 10 assistants on La Silla. 1.4 I I II

I- References 1.2 I- Greenstein, J. L., 1974, Astron. J. 79, 964. C> I • Kennicutt, R.C., 1984, Astrophys. J. 277, I- 361. Koester, D., 1982, The Messenger28, 25. 1.0 I- / Koester, D., Reimers, 0.,1981, Astron. Astro­ / . phys. 99, L8 (I). Koester, 0., Reimers, 0., 1985, Astron. Astro­ /P phys. 153, 260 (111). 0.8 I- Koester, 0., Reimers, 0., 1989, Astron. Astro­ / phys. LeIters, in prep. (VI). I- 8 Reimers, 0.,1975, Mem. Soc. Roy. Sci. Liege + x+ o 8,369. 0.6 I- Reimers, 0., 1987, in Circumstellar Matter, I- lAU Symp. 122,307. Reimers, D., Koester, 0., 1982, Astron. Astro­ 0.4 I- - phys. 116, 341 (11). Reimers, 0., Koester, D., 1988, Astron. Astro­ - phys. 202, 77 (IV). Reimers, 0., Koester, 0., 1989, Astron. Astro­ 0.2 I- - phys. M. - Romanishin, W., Angel, J. R. P., 1980, Astro­ 0.1 I- phys. J. 235, 992. I I I I III II I II Tammann, G.A., 1974, in Supernovae and 2 4 6 8 Supernova Remnants (C. B. Cosmovici, M ed.) D. Reidel, p. 155. I v.d. Heuvel, E.P.J., 1975, Astrophys. J. 196, Figure 3: Initial-final mass relation for intermediate-mass stars according to cluster white L 121. dwarfs identified in the course of this programme. Symbols: 0 NGC 2516, + NGC 2287,

Weidemann, V., 1987, Astron. Astrophys. x NGC 3532, 0 NGC 2168, 0 NGC 2451, L'I NGC 6405, P Pleiades. The broken line is the 188,74. relation adopted by Weidemann (1987).

New Results About SBO Galaxies o. BEITONI, Osservatorio Astronomico di Padova, Italy, and G. GALLEITA, Oipartimento di Astronomia, Universita di Padova, Italy

We discuss here the preliminary re­ tics, stellar motions were in the past with good hope of progressing toward sults of a long term project on kinema­ very little studied. A systematic attempt its complete understanding. In recent tics and photometry of SBO galaxies to analyse the kinematics of SBOs (Kor­ years, therefore, new data on stellar begun at ESO in 1983 and not yet fully mendy 1982, 1983) was never com­ kinematics have become available for completed. Beginning this study, we pleted and practically only one galaxy, eight more SBOs (Galletta 1987, Bettoni were particularly interested in analysing NGC 936 (Kormendy 1983, 1984), was and Galletta 1988, Bettoni et al. 1988, the mark of triaxiality that the bar in­ studied in detail (photometry, stellar ve­ Jarvis et al. 1988, Bettoni 1988). The duces in otherwise symmetrical galax­ locity field and velocity dispersion field) observational techniques used in our ies, by perturbing and stretching out the before 1987. Similar projects at other observations reflect the improvement of stellar orbits. But we did not imagine observatories followed the same ESO instrumentation in these years: that, progressing in this almost unex­ course, with observations never com­ starting with the 3.6-m telescope with plored land, so many new and not yet pleted or results never published. Prob­ the Boiler & Chivens spectrograph and fully explained features would be dis­ able reasons for this were the difficulty the 3-stage EMI image tube, the obser­ covered. Now, we feel it would be inter­ of supporting for many years a project vations continued at the ESO-MPI esting to resume here before the com­ requiring many observing nights against 2.2-m telescope with RCA CCDs when pletion of the search the main results so the idea that in the melting pot of billions these detectors became available. Par­ far obtained. of stars moving within the bar it should allel to this study, the inner photometry be impossible to distinguish between of the selected galaxies has been per­ "families" of orbits and, last but not formed using the ESO-Danish 1.52-m 1. A Bit of History least, against the occasional misunder­ telescope. SBO galaxies are good candidates for standing of some (too human) time this study because of the low, but not commissions. But with some vi­ 2. Observations and negligible, quantity of gas and dust pre­ cissitudes and a little bit of luck, obser­ Data Reduction sent in them. But despite the large vations of SB Os continued at ESO tele­ number of theoretical models of bars, scopes, demonstrating that we can look Our sampie includes 15 SB 0 galaxies and the several works on gas kinema- inside the stellar and gas kinematics brighter than the 12 th magnitude and

51