TABLE 1 : -Earth-Comet during the observing periods of comet P/Tempel2 at ESO La Silla

Date r (A'-') A (Au) 73 (0) (0) Y (0) ESO telescope

4.5.1988 1.94 0.99 152 14 48 2.2 m 18. - 24.7.1988 1.53 - 1.50 0.79 115- 111 37 - 39 291 - 288 1.5 m Danish 30.10.-2.11.1988 1.46-1.47 1.22 - 1.24 81 42 258 1.5 tn ESO 3. - 10.11.1988 1.47 - 1.50 1.24 - 1.31 81 - 80 41 - 40 258 - 256 GPO

r = solar distance of the comet; A = Earth distance of the comet; R = Sun-Earth-Comet angle; a = Sun-Comet-Earth angle or phase angle; y = position angle of the Sun (measured east of north). spectral range of the Swan bands. How- as a-X) is unknown, but can be assumed ever, this would be fulfilled only if there as a free parameter. In Figure 4, the full is no exchange of populations between lines represent the dependence of the the lowest vibrational levels. Such tran- integrated flux ratio of (Av = + I)/(Av = 0) sitions are indeed forbidden, because Swan bands on the heliocentric dis- the C2 is homonuclear and one can ex- tance, theoretically predicted by Krishna pect that the ratio of the integrated band Swamy and O'Dell (1987) for various fluxes and consequently also the vibra- values of the transition moment (Fie)" tional temperature would be constant expressed in atomic units. The filled cir- and independent on the heliocentric dis- cles indicate the most precise values of tance. But the observational results ob- the flux ratios obtained by O'Dell et al. tained for many comets indicate that Tvib (1988) for comet P/Halley. Symbols V HELIOCENTRIC DISTANCE (AU) is systematically lower than T,. This indicate data obtained from comet P/ Figure 4: The dependence of flux ratio C2 (Av effect can be explained by a sophisti- Halley with high spatial resolution on- = il)/(Av = 0) on the heliocentric distance. cated model developed by Krishna board the spacecraft VEGA 2 (Vanysek Full lines represent theoretical model calcula- Swamy and O'Dell in the past decades et al., 1988). V, stands for the impact tions for various values of the electronic mo- (1987). They propose a mechanism parameter (i.e. minimal distance of the ment, derived by Krishna Swamy and O'Dell. Full circles are measurements of comet which allows transitions between trip- line of sight from the nucleus) Halley, symbol T for comet Pflempel 2. V, let state and adjacent lowest singlet < 1000 km and V the same for 2500 km. and V represent the average value obtained state via a cascade-like radiation pro- Our preliminary result for the (Av = from Vega 2 spectrograms of P/Halley (for cess. +l)/(Av = 0) flux ratio of C, in comet P/ details, see text). Due to this process low values of Tvib Tempel 2, marked by symbol T in Fig- should be derived from the observed ure 4, falls between data for comet P/ flux ratios of the Swan bands. Since the Halley. Hence the most probable value probability of the downward sponta- of the transition moment seems to be confirmed by VEGA 2 data obtained for neous transitions is independent of the about 2.5 and the corresponding comet P/Halley. radiation density, while the upward in- Einstein coefficient for the downward duced transitions are proportional to the spontaneous transition to be about References photon flux, the effect of the downward 7 Io-~ S-l . A'Hearn, M.F., Campins, H., Schleicher, D.: transitions into singlet states becomes Since the C, molecules can be formed 1988, IAU Circ. 4614. more dominant if the radiation field de- either in the singlet and/or in the triplet Krishna Swamy, K.S., O'Dell, C.R.: 1987, Astrophys. J. 317, 543. creases. Thus the apparent Tvib de- state, the time required for establishing Luu, J., Jewitt, D.: 1988a, IAU Circ. 4582. creases with increasing heliocentric dis- the triplet/singlet equilibrium would be Luu, J., Jewitt, D.: 1988b, Astrophys. J. 328, tance. Because the relative strength of several hundred seconds and the vibra- 974. the (Av = +I) band increases with Tvib, tional temperature would be time-de- O'Dell, C.R., Robinson, R.P., Krishna the flux ratio of (Av = +I)/(Av = 0) is a pendent. Immediately after the C2 for- Swamy, K.S., McCarthy, P. J., Spinrad, H.: function of the heliocentric distance. mation which occurs in the innermost 1989, Astrophys. J., in press. Ratios of the integrated fluxes of part of the coma, the vibrational tempe- Sekanina, Z.: 1987, ESA SP-278, 323. these bands are primarily determined by rature should be high and close to the Sekanina, Z.: 1988, IAU Circ. 4624. the rate at which transitions occur be- colour temperature of the Sun, but dur- Spinrad, H., Stauffer, J., Newburn, R.L.: 1979, Publ. Astron. Soc. Pacific 91, 707. tween the lowest electronic triplet state ing the expansion of the C, molecules Vanysek, V., Valnicek, B., Sudova, J.: 1988, forming the Swan bands and the lowest into space, the populations in the lower Nature 333, 435. singlet level. states become redistributed and T,,, as West, R.: 1988a, IAU Circ. 4599. The transition probability between well as the relative flux of the (Av = + 1) West, R.: 1988 b, The Messenger 54, 55. these two states (denoted for simplicity band decreases. This effect seems to be Wisniewski, W.: 1988, IAU Circ. 4603.

Spectral Analysis of A-F Giant S. BERTHETand 5. HAUCK, lnstitut d'Astronomie de I'Universite de Lausanne, Switzerland

1. Introduction metallic-line stars, magnetic stars, h etry. Two photometric systems are Bootis, . . . especially well adapted to this effect: A very large variety of stars of spectral These stars are best distinguished the uvbyo and Geneva systems. A de- type A-F display abundance anomalies: from normal stars by means of photom- scription of the latter has been given in Arb. unlls belong to higher classes (IV Are these giants, with a high Am,, and Ill), and this has been clearly con- evolved Am stars or are they SDelphini firmed by a number of studies (Burkhart stars that have not been noticed by et al. 1980, Van't Veer et al. 1985, Smith spectroscopists? Is there an evolutio- 1971, Faraggiana and Van't Veer 1971, nary sequence between all these Stickland 1972). Furthermore, another categories? kind of with a metallic line spectrum A first series of spectra was obtained similar to that of Am stars belongs to the at Kitt Peak by Abt and Hauck in 1985 luminosity classes IV and Ill: the SDel- and Berthet (1989) has recently pub- phini stars. lished the results of an analysis of five A systematic study of A-F spectral observed stars. type stars with luminosity class Ill was wsvelengrh lmlcrornerersl recently carried out by Hauck (1986). By Figure 1 : Responses for the seven colours of 3. Observations and Data using an additional photometric crite- the Geneva photometric system to an Reduction equiphotonic flux. The determination of these rion, Ad 2 0.10, 132 stars could be profile are from Rufener and Nicolet (1988). considered as giants on a spectros- At Kitt Peak, Abt and Hauck used the copic and photometric basis. During the coude spectrograph of the 2.1-m tele- study of their photometric properties, a scope. The data were obtained with a very high proportion (36 %) was found to CCD camera at a reciprocal dispersion the Messenger by Rufener (1983). The have a A m2 value of 2 0.015. However, of 9.9 kmm in the spectral region spectral response of the seven filters before concluding that nearly one third 4400-5100 A. The analysis of these ob- used (Rufener and Nicolet, 1988) is giv- of the giants exhibit a high [FeIH] value, servations was very interesting, and at en in Figure 1. it should be ascertained that there also the same time, at the end of 1987, the The metallic-line stars, or Am stars, exists a relation between [FeIH] and 1.52-m ESO telescope was being up- were the first to be studied at depth in Am, values for these stars. Unfortu- graded in having the Echelec spectro- the Geneva system and all their pho- nately, only six of them belong to the graph equipped with a CCD detector at tometric properties are to be found in Cayrel et al. (1988) catalogue, and even the coude focus. The announced perfor- Hauck and Curchod (1980). The temper- though their [Fe/H] and Am2 values mances (Alloin et al. 1987 and Gilliote et ature parameter used for the Am stars is seem to be in good agreement with the al. 1987) of this new equipment and the the B2-V1 index, while d = (U-B1) - relation established for the dwarf stars, results from the Kitt Peak spectra in- 1.430 (Bl-B2) and m2 = (BI-B2) - this number is too low to draw a valid duced us to undertake an observing 0.457 (B2-V1) are the parameters for conclusion. It is for this reason that a programme to increase our sample of A- luminosity and blanketing respectively. campaign for the abundance determina- F spectra. In September 1988, An important parameter is that of tions of giant A-F stars has been under- we obtained for our programme of , Am,, being the difference taken. This study should not only enable spectroscopic study of A-F giant stars between m2 observed and m, for the a better understanding of the signifi- five nights with this new equipment at Hyades at the same B2-V1 value. The cance of A m2 for the A-F giant stars, but La Silla. Thus observations were carried average A m2 values are - 0.009 + 0.01 1 also attempt to understand the surface out with the 1.52-m telescope and the for the AV stars, 0.002 +- 0.01 1 for the abundance pattern of the evolved stars. Echelec spectrograph working with a mild Am and 0.013 ? 0.019 for the classical Am stars. We have known for more than a quarter of a century (Van't Veer, 1963) that the Am stars show an Fig. 2 HR 6492 (ESO Echelec, ord 26) overabundance of iron and a and deficiency. Those stars for which [Fe/H] values have been deter- mined confirm the relation between [Fel HI and Am2 obtained by Hauck et al. (1985) for stars. If A m2 is considered to be a parame- ter for segregating the Am from the nor- mal stars, an attempt may be made to use it for detecting stars of this type which would not have been considered by spectroscopists. Only six main se- quence stars were found by Hauck and Curchod (1980) with a Am, value 2 0.015 and classified as main sequence stars. On the other hand, several stars in luminosity class Ill showed a high Am, value.

2. Metallicity Among A and F Giant Stars 4 -11 Both spectroscopic and photometric 4500 4510 4520 4530 4540 observations confirm that the Am stars WAVELENGTH (A) are considered to be main sequence Figure 2: Portion of the spectrum of star HR 6492 in the spectrum order 26 (4495-4550 A) stars. However, a small number of them observed on September 1988 with the Echelec. normalized to the Fe abundance. This normalization minimizes the effect of errors in equivalent width scale and in and metallic-line star abundances are best represented in this form. An interesting point is to com- pare these abundances with those of a classical Am star (63 Tau), an evolved Am star (HRI 78) and a 6Delphini star (bDel). This comparison (Fig. 4) points out the fact that the abundance of the stars presented in this paper does not look like that of an evolved Am star or of an Am star. We do not observe the typical underabundance of Ca and Sc of Am stars but on the other hand the -5 - - configuration of the abundance distribu- - tion is close to that of the 6 Del star. The - n HR2927 0 HR1676 6Delphini stars with a or giant - luminosity are defined as stars having 0 HR2880 a HR4042 -1 - abundance anomalies comparable with - o 11R4090 0 HRG492 those of the main sequence classical - I I I I I I I Am stars, except for Ca and Sc which C Mg Ca Sc TI Cr hln Fe NI Y ZI. Ba are normal or overabundant. Figure 3: Logarithmic abundance ratio with respect to the sun normalized to iron. 5. Discussion The comparison of abundances nor- high-resolution CCD detector (ESO pressure and density at each depth and malized to Fe of a classical Am star, an CCD # 13). The choice of the in- continuous opacities at selected evolved Am and a bDelphini star raises strumentation was dictated by the need wavelengths. We then use the model the question of knowing if there is a link for observing spectra with a linear dis- prepared to determine the abundance or an evolutionary sequence between persion as small as possible and a good from the equivalent width by iteration on these three types of star. Can we con- SIN ratio (100). In its actual configura- the solar abundances; this process sider the assumption of an evolutionary tion the Ecbelec allows a linear disper- assumes a homogeneous plane-parallel sequence in the variation of the Am sion of 3.1 A/mm at 4000 A to 4.5 kmm atmosphere, LTE line formation and characteristics? Can we suppose that at 6000 A. The Echelec resolving power Voigt profile for the line absorption coef- during the main sequence phase Ca and falls between that of the CES (with its ficient. Sc are underabundant and when the short camera R = 60,000) and that of star reaches the giant phase the abun- Caspec (R = 20,000), moreover the wide dance of Ca and Sc increases to be- separation of orders permits some bin- come normal or overabundant? Thus an ning in the direction perpendicular to the The results are summarized in Fig- Am star leaving the main sequence, dispersion. This improves the SIN ratio ure 3 which displays the logarithmic would become first an evolved Am star without any loss of resolution. The abundance ratio with respect to the sun (Ca and Sc are not so underabundant as meteorological conditions were not so good during the observing run and ex- posure times were 20 to 60 minutes for stars of m, = 4.3 to 7.4 with a binning of 2 pixels to have at least a SIN ratio r 70 (Fig. 2). 15 stars were measured during these 5 nights in the spectral region 4280-4900 A. The reduction of these spectra is now being carried out accord- ing to MlDAS procedures. The first re- sults we obtained with the new config- uration of the Echelec are very en- C hlg Ca Sc Ti Cr hln Fc NI Y Zr Ba couraging, the quality of spectra will L-..-.-J allow detailed abundance analysis. In this paper we present the first re- sults. The determination of the atmospheric parameters is presented in detail in a paper to be published in Astronomy and Astrophysics. From a temperature dis- Figure 4a, b, c: Comparison of 6 A-F giants whose spectra have been taken at Kitt Peak tribution with depth (T(.c~~~~))for a given Observatory (KP)and at ESO (A- HR 1676, 0 effective temperature, surface HR2880, )2 HR2927, A HR4042, U HR4090, and metallicity, we recompute the ioni- 0 HR6492) with 63 Tau (M) an Am classic zation equilibrium and hydrostatic star, HR 178 (A) an evolved Am star and 0 Del equilibrium to determine the total (@) a A Delphini star. the SDel star is more evolved than the evolved Am star and that, of course, the evolved Am star is more evolved than the classical Am star. This fact is inter- esting but the test would be better by using these three kinds of star (Am, evolved Am and 6 Delphini) belonging to a cluster, thus the evolutionary state could be determined without ambiguity. To explain Ca and Sc anomalies, the most useful mechanism is diffusion. The efficiency of this process is strongly de- pendent on the location of the H-con- vective zone which is dependent on the temperature, i.e. on the evolutionary stage. Thus observed surface anomalies at a given time mostly depend on the way the diffusion works at this period...... [re I!] = b927b,lrI I 0"" lo8 Y The way the abundance stratifications ilsurk IiQOG) evolve with time depends on the ele- -1 ment, and it implies that the star can present different anomalies according 11111111111111111-1 to its age. - 04 - 02 0 02 04 0G 08 a 111, After these considerations about the possible chemical evolution of the Am Figure 5: [Fe/H] vs Am, for A-F giant stars. stars it is interesting to consider the relation [FeIH] vs Am, for A-F giants, evolved Am stars and hDel stars. Stars in the main-sequence phase) and then these three types of star (Am, evolved of Table 1 have been plotted in Figure 5 becomes a bDelphini star, Ca and Sc Am, hDelphini). But is the disappear- and it points out a correlation between being normal or overabundant. ance of these Am characteristics really the spectroscopic value [FeIH] and the Indeed, we observe that the deficien- attached to the evolution stage of these photometric index A m2. cy of Ca and Sc decreases through kinds of star? The dashed line in Figure 5 shows the With the help of the evolutionary mod- relation obtained by Hauck et al. (1985) el (Maeder and Meynet 1988) we plot for the FV stars. The straight line repre- evolutionary tracks of mass stars sents the relation, mentioned above, for TABLE 1 : A-F Giants, Am evolved and b Del 1.5-2.0 Ma in the HR diagram. The lo- the A-F giants, the evolved Am and 6 Del stars plotted in Figure 5. cation of the Am star (63 Tau), the stars. In the present situation the two evolved Am star (HR 178) and the 6 Del- relations between Am, and [FeIH] are phini star (h Del) points out the fact that relatively close. But a question arises: is

[Fe /HI = 5 486Am2 + 0 064 for V and I11 -to 350 +O 01G I

- 15 -.l - 05 0 05 References: (1) Berthet (in prep.), (2) Cayrel et al. am2 (1988), (3) Burkhart et al. (1980), (4) Hauck et al. (1985), (5) Lyubimkov et al. (1985). Figure 6: [Fe/H] vs Am, for A-F giant stars anf F-V stars. there one relation for A-F giants and photometric parameter to estimate the Burkhart, C., Van't Veer, C. and Coupry, evolved Am stars and another for A-F metal content of these kinds of stars. M. F.: 1980, Astron. Astrophys. 92, 132. main-sequence stars? In Figure 6, re- And perhaps more than that, because it Cayrel de Strobel, G., Hauck, B., Thevenin, producing our data plus data that Hauck appears that only one relation could be P., Francois, P. and Mermilliod, M.: to be published. et al. (1985) used to determine his rela- sufficient for the A-F stars. This study Faraggiana, R. and Van't Veer, C.: 1971, As- tion, the general trend seems to show points out that it is important to examine tron. Astrophys. 12, 258. that one relation could be sufficient for the evolutionary stage of the different Gilliote, A. and Magain, P.: 1987, The A-F main-sequence stars, A-F giant kinds of stars encountered. Messenger No. 50. stars, evolved Am stars and 6Del stars So, with the new configuration of the Hauck, B.: 1986, Astron. Astrophys. 155, and this assumption was already sug- Echelec spectrograph on the ESO 371. gested by Hauck et al. (1985). 1.52 m telescope at La Silla we have a Hauck, B., Foy, R. and Proust, D.: 1985, very interesting instrument for stellar Astron. Astrophys. 149, 167. spectroscopy, working with a Hauck, B. and Curchod, A.: 1980, Astron. Astrophys. 92, 289. wavelength range between 3500 and 6. Conclusion Lyubimkov, L.S. and Rachkovskaya, T.M.: 5500 A and a good linear dispersion (3.1 1985, lzvestiya Krymstoi Astrofizicheskoi The abundance analysis has shown to 4.5 &mm). The spectra obtained dur- Observatorii vol. 71, pp. 127. that Ca and Sc abundance of our stars ing our observing run allowed detailed Rufener, F.: 1983, The Messenger 31, p. 17. is normal: no deficiency has been found. spectroscopic analysis of good quality Rufener, F. and Nicolet, B.: 1988, Astron. Thus the answer to one of the introduc- and encourage us to pursue our pro- Astrophys. 206, 357. tory questions comes immediately: the gramme on the spectroscopic study of Smith, M.A.: 1971, Astron. Astrophys. 11, A-F giants with a blanketing parameter A-F giant star atmospheres. 325. Stickland, D.J.: 1972 Mon. Not. R. Astr. Soc. 2 Am2 0.015 are not necessarily 159, 29. evolved Am stars, they could also be Van't Veer C.: 1963, Ann. Astrophys. 26,289. GDelphini stars. The relation found be- References Van't Veer-Menneret, C., Coupry, M.F. and tween Am2 and [Fe/H] is encouraging Alloin, D. and Hofstadt, D.: 1987, The Burkhart, C.: 1985, Astron. Astrophys. 146, and supports the possibility of using a Messenger 49, p. 18. 139.

T Tauri Stars Make Us Wonder What They Are G. F. GAHM, Stockholm Observatory, Sweden

Once upon a time, some 4.6 billion ly symmetric forms. One could specu- Messenger by Bouvier and Bertout, years ago, the sun formed in an inter- late that it is generally regarded as more 1985). It is certainly interesting, also stellar cloud of gas and dust. The scientific (or maybe just simpler) to from a cosmogonic point of view, that we see in the sky all move close choose a spherical form when there is for many of the youngest T Tauri stars to the ecliptic. It is therefore clear that at no evidence, other than intuition, of a the angular momentum is smaller than some phase the young sun was flat geometry. The cosmogonist, of the total angular momentum of our solar surrounded by a flat dusty disk, in which course, always faces a flat geometry. system! In other words, when the T Tauri these planets agglomerated. By 1980 it was evident that for many stars contract out of interstellar cloud- In cosmogony one is concerned of these stars the excess emission ex- lets, they have found ways to get rid of about how the came into tends over the far-ultraviolet into the angular momentum very rapidly indeed. being. A standing question has been X-ray region. This high-energy tail of Last, but not least, a number of inde- that of universality: do planetary sys- circumstellar origin was usually tied to pendent indications that many of the T tems form also around other young spherically symmetric / Tauri stars are surrounded by flat disks stars, like the T Tauri stars, which are coronae. During the next years it was of dimensions had located in interstellar clouds? also realized that some of these stars been found. Flat molecular disks, appar- From what we know, these clouds of drive powerful molecular flows into the ently rotating, had been discovered at today are not very different from those surrounding interstellar cloud, often only some stars. The spherical geometry is present 4.6 Gyr ago with regard to in two opposite directions (bipolar gone for ever. chemical composition and galactic dis- flows). I left astronomy in 1983. When I The concept of disks, in tribution. I would think that ever since came back a few years later I was sur- which material is transported towards the fifties, when the T Tauri stars were prised to see how much is accom- the stars, has replaced earlier views. recognized as very young objects, most plished in science over such a short Even the continuous excess emission is scientists involved have been thinking of period of time. The sky had been now explained as a disk phenomenon. the objects as being similar to the sun, painted with numerous bipolar flows, Here, the ultraviolet excess originates in only much more active, and that they some extending over tens of minutes of a warm boundary layer between the have circumstellar regions with a flat arc. Plasma jets were seen to shoot out and the star while the geometry. There is an early paper by from the stars. A major break-through infrared excess originates further out in Poveda (1965) with respect to this latter was the mapping of rotational velocities the disk. Some stars lack the charac- point. Nevertheless, during the late six- of the stars and the discovery that some teristics of disks, and on these so-called ties when the infrared excess emission of the light variability present was of "naked" T Tauri stars the X-ray variabili- (from circumstellar dust) and the ul- periodic nature and related to rotational ty, for instance, could have its cause in traviolet excess emission were discov- modulation of bright and dark areas on enhanced surface activity. Such activity ered, the models always took spherical- the stellar surface (see the article in the may exist also on the "clothed" stars but