Santiago. The agreement dates from 29April 1965 [31]. ESO provided the as­trolabe with chronograph equipmentand a building to house the instrument,and the University of Chile itschronometrie facilities. But most impor­tant: the observations would be con­ducted and supervised by the staft ofCerro Calan. After overhaul in Paris, theinstrument was installed on Cerro Calanin November and Oecember 1965 withthe collaboration of Guinot. Since then ithas made, under the supervision of F.Noel, solid contributions to the Funda­mental Reference System in the South­ern Hemisphere and to research on theEarth's rotation; a first demonstration ofthe appreciable systematic errors in thesouthern FK4 declinations was pub­lished by Anguita and Noel in 1969 [32].In ESO Bulletin No. 4 of June 1968 Noeldescribes the nature of the project andthe first years of operation.

ESO Chooses its Emblem

Not only heavy tasks kept the ESOCommittee busy. After the Conventionhad been signed, it acquired its emblemfor which at the October 1962 EC meet­ing Bannier presented some designs bythe artist Mrs. G. M. Pot. The Committeehad no problem in making up their mind;according to the minutes it chose thedesign "in which the stars show at theirbest". The emblem's stars - the South­ern Cross - still show weil, as is appar­ent from the front page of thisMessenger.

References and NotesAbbreviations used:EC = ESO Committee (the Committee pre-

ceding U,e ESO Council).ECM = ESO Committee Meeting.IC = Instrumentation Committee.EHA = ESO Historical Archives (see the arti­

eie in the Messenger of December 1988).FHA = Files Head of Administration at ESO

Headquarters.

[1] Circular letter by Oort to EC memberspreparatory to the ECM of May 1959, inEHA-I.A.1.9., and minutes of that meet­ing.

[2] See letters of Oort to Danjon and Funkeof May 30, 1962, in EHA-1. C. 1.1.c.

[3] In EHA-1. C. 1.1.d.[4] In EHA-1. C. 2.1.g.[5] See correspondence between ZWO and

University of Groningen in the years1962 and 1963 in EHA-1. C. 2.1.e.

[6] Information provided by the PersonnelDepartment of ESO; also: minutes of theECM of July 1963, p. 12.

[7] Pub!. Astron. Soc. of the Pacific 72, 225,1960.

[8] I. S. Bowen, Pub/. Astron. Soc. of thePacific 62, 95, 1950.

[9] I.S. Bowen, Pub!. Astron. Soc. of thePacific 61, 243, 1949.

[10] Jahresberichte Hamburger Sternwarte1954 and 1955; Sky and Te!escope 15,Nov. 1955, p. 10.

[11] See, for instance, minutes ECM of Oct.1957, June 1961, Oct. 1962, Nov. 1963,Council Meetings of May 1964 and April1966 and correspondence betweenFehrenbach, Heckmann and Oort ofJune 1964 in EHA-I.A. 2.9. and I.A. 2.10.

[12] Minutes ECM of April and Oct. 1957; theEHA do not contain the written report.

[13] In EHA-1. C. 1.9.c.[14] In EHA-1. C. 1.9.a., Visite des Obser­

vatoires Americains.[15] Minutes ECM of November 1961.[16] See, for instance, the letter by Blaauw to

Fehrenbach of April 6, 1961 in EHA-1. C.1.9.c.

[17] EHA-1. C. 1.9.c.[18] See letter by Minnaert to Oort and

Blaauw of May 1, 1961 in EHA-1. C.1.9.c.

[19] See Minnaert's letter to Van Geelen of10 October 1961 in EHA-1. C. 1.9.c.

[20] Maps EHA-1. C. 1.9.flk contain prepara­tory correspondence, technical descrip­tions, and the tender of Rademakers.

[21] Minutes IC of November 1961.[22] See ref. No. 18.[23] Minutes ECM of June 1961.[24] Minutes ECM of November 1961.[25] EHA-1. C. 1.9.e. contains the Cahier de

Charges with drawings and the Marchede Gre a Gre of REOSC of May 20,1963.

[26] In EHA-1. C. 1.1.c. See also corre­spondence between Fehrenbach andOort of October 1958 in EHA-1. A. 2.1.

[27] The Yale and U.S. Naval Observatoriesplanned an instrument in Argentina andthe Pulkovo Observatory one in Chile,whereas Greenwich Observatory con­templated a collaborative project withthe Cape Observatory and HamburgObservatory one with Perth.

[28] Minutes Council Meeting of May 1964,p. 10.

[29] Letter by Guinot to Blaauw and follow­up correspondence with Van Geelen inEHA-I. C. 1.9.d.

[30] EHA-I.A. 1.19. and I.A. 2.6.[31] ESOAnn. Report 1965, p.10.[32] Astron. Journa/74, 954. 1969.

Field Strömgren Photometry with a CCDJ. KNUOE, H. J0NCH-S0RENSEN, Copenhagen University Observatory, Denmark

Introduction

The chemical evolution of the Galaxyis somehow coupled to its formation.The location of stars with a certainmetalIicity may therefore also dependon the Galaxy's dynamical history.

Laws describing the ga/actic distribu­tion of the various stellar populationsintroduced to understand the construc­tion of the Galaxy are often based ondetailed studies of the solar vicinity. Wehave been interested in studying par­ticularly the F stars in a few galacticdirections of interest, e. g. the SGP, tosearch for [Fe/H] gradients in space andtime. Such studies are mostly based onaccurate photoelectric photometry butthe cry for data in more remote volumeshas been acute lately and as large tele­scopes are not available for extended

46

photometrie surveys we have tried touse a medium sized telescope with aCCO instead.

Stars with a metal content down by afactor of 2.5 relative to the Sun havebeen suggested to form a spheroid witha local scale height in the range from600 to 1000 pe and the stars with [Fe/H]:s - 0.8 another system with ascaleheight of several kpc. Strämgren photo­metry of F stars seems weil suited totrace the metal variation with age anddistance. The intermediate band photo­metry thus permits computation of dis­tances based on individual absolutemagnitudes. Oistances based on a col­our - absolute magnitude relation as(b-y)o - Mv may be quite uncertain. Foran F star with (b-y)o = 0.3 the width ofthe main sequence band is observed to

be 2 mag at least. The scale height ofthe most metal poor stars thus suggeststhat observations of objects several kpcfrom the plane should be performed.

According to current models of theGalaxy, it is only several kpc from theplane that extreme population 11 starswill dominate. Our observing para­meters are set by the detection of an F9star 5 kpc from the plane. The V mag­nitude is about 18 at this distance. Themost critical colour is, however, the u·band, partly because the F stars are cooland partly because this band falls in thewavelength range where the CCO'sR.Q.E. is smallest, only 10 to 20 %. Thestellar metallicity may be computedwithout u but the band is required forestimating Mv. An F9 star has (b-y) = 0.4and (u-b) = 1.5, V = 18 then implies that

11.75

.j. 12.25

Background Subtraetion

After correcting for the sensitivityvariation across the frame, we noticedthat the background depended onthe brightness of the star and that thestellar magnitudes did not convergewith aperture.

From the seeing conditions during ourruns and the scale of the Danish1.5-m telescope we expected stellar im­ages of 5 pixels or smaller.

Figure 2 shows the variation of thestellar magnitude with aperture when weuse the background suggested by theDAOPHOT photometry package. Thestar brightens with the aperture. Obvi­ously we don't correct for all the signalin the background. DAOPHOT derivesthe background as, mode = 3 x median- 2 x mean. A good background es­timator in crowded regions like a globu­lar cluster, but apparently not in thesparsely populated general fjeld. We re­placed the mode by the simple meanand the result is shown in Figure 3where we obtain a good convergenceafter - 10 pixels. For the faint pro­gramme stars we thus use a stellarradius of 12 pixels and not the 2-3suggested by the seeing measure­ments.

12

Figure 3: As Figure 2, but the mode is nowreplaced by mean and a convergence is es­tablished.

12.75

12.5

13O"-'-"-'-'-"'-:--"-'--"-'--,.LO

..................L,5....................-l.20..........~25

Aperture radius (plxelsJ

Transformation to the StandardSystem

The instrument magnitudes resultingfrom the aperture photometry are thencorrected for extinction and trans­formed to our secondary standard sys­tem by means of our standards. We areusing the transformations for the wholerange of apparent magnitude of our pro­gramme sampie. We have approximate­Iy three stars per frame at the SGP, alsoindicating that our limiting magnitude isabout V = 20 mag, so the transformationis used 6 magnitudes beyond the fain-

Flat Fielding

Correct flat fielding is of the outmostimportance when an accuracy of0.02 mag or - 2 % is required. It wouldof course be most convenient if a scien­tific frame with its astronomical objectsand background could be flat fieldedwith a single, well-defined, responseframe. Considering the possible intensi­ty range in a frame bracketed by thebackground and a source, the CCD'sresponse surely must be linear. How­ever, we may have seen indications thatthis is not quite the case. The responseseems to obey apower law, response_1,·03, valid for intensities from a fewhundred to several thousands. Figure 1shows the ratio of two u flat fields at alow and a high intensity. A three per centvariation is noted. A non-linearity meansthat we cannot use identical flat fieldsinside a stellar image and for thebackground. Using a flat field pertainingto the background level or to someintermediate level leaves the starsslightly too bright. For a y = 17 mag starwe make an error in the range I;o,.y =

0.01 -0.02 mag.

the images still did not saturate. As asecond approach we also tried to useopen clusters with deep uvby photome­try allowing several stars in a singleframe but this may result in difficulttransformations of the m, index be­cause of the cluster's narrow metallicityrange. About 20 standards in each col­our is preferable per night. With the ex­tinction determination, one third of thenight is spent on the photometric cali­bration of the system.

Back ~ "ode11.5

11.75

12

~ 12.25~

12.5

12.75

130 10 15 20 25

Aperture radius (pixels)

Figure 2: Aper/ure-magnitude versus aper­ture. The star brightens with aper/ure. The ticmarks on the curve indicate a one-pixel step.The magnitudes are computed with abackground estimated as: mode = 3x me­dian - 2 x mean. Oue to the large stellarimages the background is measured in anannulus with radii 35 and 45 pixels.

1.1

u ~ 20 mag. So the problem concen­trates on how one obtains high-preci­SIOn Strömgren photometry for starswith u - 20. Depending on the actualChip available, the necessary integrationtlmes may be estimated. We used ESOCCD # 8 which required 70 to 80m.inutes to go down to the 20 th mag in uwlth a S/N of 30.

Photometrie Proeedure

As an objective of our study is tomake possible a distinction between Fstars with (Fe/H] = 0.0, -OA and lessthan - 0.8 representing the disk, theIntermediate population II and the ex­treme population 1I respectively, we re­~uire a standard error of 0.02 (or better)In the colours v, band y to have a one­sigma difference in (Fe/H] only. A similaraccuracy of u and thus of c, gives arelative distance error of 20 %.

For the CCD observations, we try toadopt the procedure established forphoto-electric measurements with ex­tinction determination in all four colours~nd with copious standard star observa­tions each night.

.9-~~:-<--'--'....,-'-ll..L-LL.L-,-L-.L..L--'---'--'--'-...L....1..J100 200 300 400 500

~igure 1: The ratio of a u flat field with originalmtensity about 650 AOU to a u flat withoriginal intensity 75 AOU. 80th flat fields arenormalized to a unit median sensitivity. Theabscissa indicates the row number and theaverage is taken through column 30 to 300.The ratio of the two flat fields is seen to showa variation of -3%.

uf 1hduf 11, co 1uon, 30 to 3001.2rr-.-".--,,-,-.---.-,.,---.,,.--,,-,-.-....-n

Standard Stars

No primary standard stars are faintenough to be used with a CCD on a1.S-m telescope; instead we have beenUSlng secondary standards down to the14th

mag from the literature. Standardswere exposed with a defocused tele­SCOpe and with sufficiently long expo­sure times so the uncertainty in theshutter timing was of no importance and

47

test standard stars. Our results are sure­Iy depending on the detector linearity.m" c" and (b-y) have almost linear

Figure 4: uy versus y for three frames inSA 168. The magnitudes are in the instru­mental system, but the transformation coeffi­Gient is about unity.

51168

Ox X

>Xx X x~ ~@o

plane seems within reach. We haveGlFe/Hj = 0.4 dex and D/D - 20 % orbetter. However, we do not see how theerror may be improved to better than- 0.01 mag or -1 % implying that thebest obtainable error is 0.3 dex in themetal content [Fe/H]. Regarding F stars,observations are just feasible at 5 kpcfrom the plane with a 1.5-m telescope.

It will be particularly interesting to seethe relative population shift with dis­tance, but also to see if there exist starswith solar metallicity at these remotedistances.

As our general results are not tooencouraging concerning the obtainableerrors we want to stress that it seemspossible to do CCD photometry - also inthe u region - in the field without havingto establish standards in each frame.

We should mention that the reductionof the several thousand frames formingthe basis of this note have been per­formed with the MIDAS, IRAF andDAOPHOT packages.

Results

transformations whereas V includes, asexpected, a more significant colourterm.

Figure 4 shows as an example thevariation of Gy with y for three frames inthe selected area SA 168 and appar­ently Gy stays below the maximumacceptable error 0.02 mag down toabout the 19th mag. The three othercolours have an identical behaviour.

Towards the SGP we have so faridentified about 39 F stars, 0.2 < (b-y) <0.4 mag, in the whole magnitude rangedown to V = 20 and in a solid angle only- one tenth of a square degree. 33 ofthese stars also have good u measure­ments, so the sampie already is of somesignificance.

When u, v, band y are obtainable withan error 0.02 mag, the study's objectiveto investigate the [Fe/H] variation of theF stars beyond D = 5000 pc from the

2220I

I

°xo J

00

~ 0

18I

16I

I I

o

14

.1

.2

ö-Scuti Stars in NGC 6134H. KJELOSEN and S. FRANOSEN, Institute ofAstronomy, University ofAarhus, Denmark

The CCD camera on the Danish 1.5 mtelescope has been used to obtain ex­posure time series of small areas inopen clusters. The purpose is to studythe frequencies of different types of pul­sating variables. Very low noise levelshave been reached by the use of differ­ential photometry carefully consideringthe error sources.

cent. Consequently, for the bright stars,the change of seeing introduces a varia­tion due to the non-linearity. We wereable to correct for this effect using thelarge number of exposures and largenumber of stars we have.

All time strings were transformed intopower amplitude spectra. Figure 1 pre­sents the mean amplitude in three fre­quency intervals for stars over a rangeof 7 magnitudes. For high frequenciesthe amplitudes scatter very little about a

B-magFigure 1: The noise level for different frequency intervals. Squares correspond to periods in therange 3-10 min, triangles 10-60 min and diamonds 1-2 h. The abscissa is the B magnituderelative to a set of reference stars on the frame.

o

NGC

o 9

o Il D

o 11 /j~ 0

o !i'I ij

6.04.02.00.0

Il ce 0

0 0 0

• a0

0 \l SOb oe ~

9 \l o 0

5

2

-310

W 5tf1 0

0 Bz 2

-q10

5

2

-510

-2.0

Noise Levels

To illustrate the high precision one canobtain with CCD's, we present the datafrom one night in late May 1988 on NGC6192. Exposure times were 20 secondsand exposures were collected each mi­nute for nearly 7 hours. The time serieshas some gaps, when tapes had to bechanged or the seeing and the trackingchecked. The resulting 370 frames werereduced with the DAOPHOT packageand relative magnitudes determined forall reasonably isolated stars. A small setof not too bright, weil isolated stars de­fine the reference.

Two corrections turned out to be ofcritical importance. A colour correctionto eliminate differential extinction effectson stars of different colours. And acorrection for non-linearity of the CCD.The CCD (# 8) turned out to have a non­linear response at high exposures be­fore saturation of the order of two per

48