1

No. 87 – March 1997

The Minister of Foreign Affairs of the Republic of Chile, Mr. Miguel Insulza, and the DirectorGeneral of ESO, Prof. Riccardo Giacconi, signing the instruments of ratification on December 2,1996, in the Ministry of Foreign Affairs in Santiago.

Important Events in ChileR. GIACCONI, Director General of ESO

The political events foreseen in the December 1996 issue of The Messenger did take place in Chile in the earlypart of December 1996. On December 2, the Minister of Foreign Affairs of the Republic of Chile, Mr. Miguel Insulza,and the Director General of ESO, Professor Riccardo Giacconi, exchanged in Santiago Instruments of Ratificationof the new “Interpretative, Supplementary and Amending Agreement” to the 1963 Convention between theGovernment of Chile and the European Southern Observatory. This agreement opens a new era of co-operationbetween Chilean and European Astronomers.

On December 4, 1996, the “Foundation Ceremony” for the Paranal Observatory took place on Cerro Paranal, inthe presence of the President of Chile, Mr. Eduardo Frei Ruiz-Tagle, the Royal couple of Sweden, King Carl XVIGustaf and Queen Silvia, the Foreign Minister of the Republic of Chile, Mr. José Miguel Insulza, the Ambassadors ofthe Member States, members of theof the ESO Executive, ESO staff andthe Paranal contractors’ workers.

The approximately 250 guestsheard addresses by Dr. Peter Creo-la, President of the ESO Council,Professor Riccardo Giacconi, Direc-tor General of ESO, Foreign Minis-ter José Miguel Insulza and Presi-dent Eduardo Frei Ruiz-Tagle. Theoriginal language version of the fouraddresses follows this introduction.(A translation in English of the Span-ish text is given on pages 58 and 59in this issue of The Messenger.)

A time capsule whose contentsare described in Dr. Richard West'sarticle was then deposited by Presi-dent Frei with the works being bless-ed by the Archbishop of Antofagas-ta, Monsignor Patricio Infante.

Both the formal and informalparts of the Ceremonies were de-

122

noted by great warmth. The event is seen as initiating a new period in the relations between ESO and its host countryChile and the harbinger of closer scientific co-operation between Chilean and European astronomers.

The ESO Council meeting took place in Santiago on December 5 and 6, 1996. Mr. Henrik Grage was electedPresident of Council as of January 1, 1997, and Dr. Bernard Fort as Vice-President as of January 1, 1998, at the endof Dr. Jean Pierre Swings' term.

The Council approved the ESO 1997 budget as recommended by the ESO Finance Committee and discussed inthe December 1996 issue of The Messenger. More detailed discussion of this and other Council actions will appearin the June 1997 issue of The Messenger.

Ruiz-Tagle, your Majesties, King andQueen of Sweden, Carl XVI Gustaf andQueen Silvia, Minister of Foreign Af-fairs, Don José Miguel Insulza, Señor

Intendente of the II Region, DonCesar Castillo, honourable Congress-men, Senators, Ambassadors, au-thorities, Members of Council, friends

Paranal, December 4, 1996

Your Excellency President of theRepublic of Chile, Don Eduardo Frei

Paranal, December 4, 1996

Your Excellency, President of the Re-public of Chile, Don Eduardo Frei Ruiz-Tagle, your Majesties the King andQueen of Sweden, Carl XVI Gustaf andQueen Silvia, your Excellency, Ministerof Foreign Affairs, Don José Miguel In-sulza, Señor Intendente of the II Re-gion, Don Cesar Castillo, honourableSenators, Ambassadors, authorities,colleagues of the ESO Executive andthe ESO Council; friends, ladies andgentlemen,

It gives me great pleasure to welcomeall of you on Cerro Paranal, the site ofthe world’s largest optical observatory.

Our organisation’s life in Chile started33 years ago when the original agree-ment between ESO and its host countrywas signed under the government ofPresident Jorge Alessandri Rodriguez,agreement which was to be supplement-ed later under President Eduardo FreiMontalva, who inaugurated in 1969 theLa Silla Observatory with his son, nowPresident of the Republic. For us it isthen a special honour to welcome todaymembers of both presidential families.

Nine years ago, in a historical meetingat the ESO headquarters in Garching,Germany, the 8 Member States of ESOunited in the Council took the momen-tous decision to construct the VeryLarge Telescope. This decision followeda period of several years of intensediscussions in the European ScientificCommunity; as well as in the variousministries and science committees ofthe ESO Member Countries.

The decision was not an easy one.Never before had a ground-based astro-nomical project of this size been pro-

posed. Never before had the scientificcommunity of Europe, together with thefunding authorities, dared to consider aground-based research facility to studythe cosmos of this complexity and cost.The positive decision to engage in thisproject for the 21st century started ahighly elaborate process on many fronts.Within ESO a gifted group of scientistsand engineers began to work on thetechnical implementation, invoking thelatest technologies available anywhere inthe world. In many cities all over Europe,industrial engineers and technicians be-gan the construction of the complex partswhich would later come together as theVLT at the Paranal Observatory. In theministries in the Member Countries, webegan the long process of ensuring thesteady funding and political support forthis European flagship project.

Now, 9 years later, we are close to thegoal of our common dreams. Despitemany obstacles, we have succeeded inkeeping to the plan and in just a yearfrom now, the first of the four 8.2-metreVLT telescopes will open its eye towardsthe sky.

This project would not have beenpossible without the close and contin-ued, extremely positive collaboration,not only between the individual ESOMember States but, in particular, withthe host country of this organisation, theRepublic of Chile.

Mr. President, thanks to your personalsupport, the continued efforts of yourgovernment and of many other authori-ties in Chile, we have succeeded intransforming this mountain into what willsoon become the world’s most modernoptical observatory. This process hasnot been easy, many areas have been

interlinked, and there were some stoneson the road, pebbles as well as rocks,but we are now certain that soon Chileanscientists, together with their Europeancolleagues, will reap the fruits of our jointlabours. Thanks to this unique project, atthe limit of today’s technologies fron-tiers, Europe and Chile have come clos-er together than ever before.

The presence of ESO in Chile hasgiven European scientists access to theclearest skies in the world and hasprovided Chilean scientists with the pos-sibility to interact continuously with theircolleagues in the front line of researchand technology. We now look forward tothe first exciting results from this collab-oration at the Paranal Observatory. As Ihear, there are already many discus-sions going on between European andChilean researchers about how to bestuse this unique facility for the benefit ofall involved.

The Members of the ESO Council andI are happy to be here today and tosense the enthusiasm which is evidentin all quarters. We have no doubt that itwas a wise decision by our countries tobuild the VLT and support this greatproject. The fruits will be not only in thefields of science and technology but,equally important, the understandingamong peoples and nations will be fur-thered. It is a uniting aspect of all cul-tures to look up from our home planettowards the universe to admire its won-ders and to grasp its origin and destiny.Our countries may be far apart in ageographical sense, but within thisproject we will labour together andthereby open new horizons for all ofhumanity.

Muchas gracias.

Speech by the President of the ESO Council,Dr. Peter Creola

Speech by the Director General of ESO,Prof. Riccardo Giacconi

3

and colleagues, ladies and gentle-men,

It is a great honour for us to welcomeyour Majesties on this very special occa-sion: There are certain events in thedevelopment of science, which markimportant achievements, or a new prom-ising departure towards, as yet, unex-plored regions.

More than 30 years ago, I had theprivilege of being part of an astronomicalproject that soon thereafter succeededin opening the X-ray sky for scientists.

More recently, I was fortunate enoughto be associated with the developmentand launching of the Hubble SpaceTelescope.

Today, I have a feeling of excitementequal to that on those occasions, as theESO Very Large Telescope becomes areality. No other project in ground-basedastronomy has been more ambitious,more complex and, indeed, more de-manding in resources. Few otherprojects in the history of astronomy havehad a scientific potential of similar dimen-sions. Just one year from now, the firstunit telescope of the VLT goes intooperation and astronomers of the worldwill begin to open new vistas in thisfundamental science. The VLT will deliv-er sharper images than any other opticaltelescope and, thanks to the enormousarea of its mirrors, it will collect morephotons and, therefore, reach fainter andmore distant objects than any existing

facility, either on the ground or in space.As we are now approaching the end of

the construction phase, astronomers inEurope, Chile and elsewhere, are pre-paring the exciting research projectsthey will soon undertake with their newobservatory. Many of these will take usway beyond current horizons and willenable us to search for the answers tosome of the deepest questions mankindhas ever posed. The VLT has the capa-bility of looking so far out in space and,therefore, so far back in time that we willultimately reach the period, soon afterthe big bang explosion, when the matterin the universe had just begun to con-dense in the space islands we nowobserve as galaxies. The VLT will make itpossible for us to look into the mysteriouscentres of galaxies where processes ofunimaginable violence take place. TheVLT will help us to understand the birth ofstars, deep inside dense and, otherwise,impenetrable interstellar nebulae, ena-bling us to watch the processes that werethe base of our own distant origins, 4.5billion years ago. The VLT has the bestpotential of any telescope to search for,hitherto, unknown planets around otherstars and, if they exist, to help us todiscover other abodes of life in space.

The science of astronomy is a never-ending process which has drawn on theexperience and ingenuity of countlessindividuals during the past millennia.Throughout the ages, scientists and en-

gineers have put their faith in the latesttechnology and the VLT is no exceptionfrom this. We, as scientists, are deeplythankful to all those in- and outside ESOwho have helped to realise this projectand, in particular, to those authoritieswho have provided political and financialsupport for this project. Without theirforesight, this moment would not havebeen possible.

During the coming years, it will be ourprivilege to share with them and the restof mankind the excitement of new dis-coveries which will be made with theVery Large Telescope. The comprehen-sion of our cosmic surroundings is oneof the noblest goals of the human race. Itenables us to understand and appreci-ate our niche in space and time, and itopens our minds towards fundamentaltruths which unite us all.

Las circunstancias me han llevado aleer el idioma de Chile, y hoy voy aaventurar algunas palabras en cas-tellano.

Quiero decirles que nuestra organiza-ción no solo quiere hacer ciencia sinotambién participar en la vida cultural deChile. Queremos apoyar el gran esfuer-zo del gobierno del Presidente Frei parael desarrollo de la educación – en par-ticular en la Segunda Región. Regióndonde la Cordillera, el desierto y sugente vive en una relación mas íntimacon nuestro universo.

Muchas gracias.

Speech by the Minister of Foreign Affairs,Mr. José Miguel InsulzaParanal, December 4, 1996

Señor Presidente de la República,don Eduardo Frei Ruiz-Tagle, sus Ma-jestades los Reyes de Suecia, Karl XVIGustaf y Reina Silvia, Señor Presiden-te del Consejo de la Organización parala Investigación Astronómica en el He-misferio Austral, Señor Director Gene-ral de ESO, Señores Senadores,autoridades civiles, militares y ecle-siásticas, Señores miembros de la de-legación de Suecia, Señores miembrosdel Consejo de la ESO, Señoras ySeñores,

Para mi constituye un alto honor re-presentar al Gobierno de Chile en estaceremonia y compartir con todos uste-des la satisfacción y la esperanza quenos produce estar aquí participando enesta significativa ocasión por medio dela cual se inaugura el centro de obser-vación de la ESO en Cerro Paranal.Esta satisfacción se origina, en primerlugar, en la importancia del proyectoque hoy inauguramos. El telescopio

VLT/VLTI ya descrito por el Presidentedel Consejo y el Director de ESO cons-tituye no sólo una expresión de la másmoderna tecnología puesta al serviciode las ciencias astronómicas, sino tam-bién una oportunidad de selección paraprofundizar en el conocimiento del Uni-verso y responder así a las interrogan-tes que han preocupado a la humani-dad desde sus orígenes. En este senti-do, más allá del considerable valor eco-nómico de la inversión que hoy inaugu-ramos, es evidente que nos encontra-mos participando en un hito en el desa-rrollo mundial y nacional de la astrono-mía.

Este sentimiento de satisfacción des-de luego se acrecienta si consideramosel largo, y no siempre fácil, camino por elque debimos transitar en los últimosaños para llegar hasta este lugar y hastaesta ocasión. Todos conocemos las difi-cultades heredadas de las incompren-siones hoy felizmente superadas que enun momento amenazaron la concreciónde este proyecto.

Sin embargo, creo que es importanteseñalar que en todo este proceso, elGobierno mantuvo, permanentemente,su apoyo a la ESO y a la posibilidad deque esta Organización continuara des-arrollando y expandiendo sus activida-des en Chile. Lo hicimos no sólo por-que se trataba de un compromiso inter-nacionalmente asumido por el Estadoen lo cual nuestro país tiene una tradi-ción de respeto riguroso a las obliga-ciones emanadas de tratados interna-cionales que comprometen el honor dela República, sino también porqueapreciábamos que el desarrollo de lasactividades de la ESO en nuestro país,en un marco jurídico claro, no sólobeneficiaría a la ciencia mundial, sinotambién al desarrollo científico de Chiley de importantes regiones de nuestropaís.

La veracidad de este análisis quedahoy demostrada, pues el Centro deObservación de Paranal servirá no sóloa las actividades de la ESO, sino queredundará también en beneficio de los

144

científicos chilenos, quienes dispon-drán así de un instrumento de trabajoque, de otra manera, les sería absolu-tamente inaccesible.

Por ello, esta ceremonia, que es con-secuencia del intercambio de Instru-mentos de Ratificación del Acuerdo In-terpretativo, Suplementario y Modifica-torio del Convenio de 1963, que realizá-ramos el lunes pasado, es el punto finalde un proceso difícil, y también es elinicio de una nueva relación de coopera-ción y entendimiento entre Chile y laESO, que es una relación madura, con-solidada, equitativa y de cooperación enmutuo provecho.

La consolidación de la relación jurí-dico-política de Chile con la ESO garan-tiza a esta Organización el desarrollopacífico de sus actividades en un marcode mutuo provecho y constituye un avalimportante para asegurar la expansiónde las mismas en el futuro.

Al mismo tiempo, al otorgar a lossectores nacionales más directamenteconcernidos con la actividad de laESO, un reconocimiento a sus legíti-mas aspiraciones, se consagra una re-lación equitativa, pues los derechos re-conocidos a Chile y sus ciudadanos enmateria científica y laboral, constituyenuna adecuada contrapartida al aporte

chileno a las actividades de la ESO enChile.

Al establecer instancias permanentesde cooperación y de resolución de con-troversias entre Chile y la ESO, estaOrganización se asocia estrechamenteal desarrollo científico y tecnológico na-cional.

Por último, esta relación es tambiénpionera en numerosos aspectos, comola protección medioambiental, los dere-chos de trabajadores chilenos en enti-dades públicas extranjeras y los dere-chos de la comunidad científica nacio-nal, y esperamos que constituirá unejemplo para futuras acciones en con-venios astronómicos que el país subs-criba en los próximos años.

Dentro de este marco, deseo expre-sar la más cálida acogida del Gobiernoa la iniciativa de creación de la “Fun-dación para el Desarrollo de la Astro-nomía”, iniciativa originada en los as-trónomos chilenos que han obtenidoCátedras Presidenciales en ciencias yque ha contado con el apoyo del Di-rector General de la ESO, Dr. RiccardoGiacconi. Creemos que esta funda-ción, que se complementará con la ac-ción de otras entidades que vinculan aChile con la ESO, está llamada a jugarun papel de gran importancia en la

administración de recursos destinadosal desarrollo astronómico de Chile y enla promoción de la cooperación inter-nacional en esta importante ciencia.

Señor Presidente, Majestades, nosencontramos en este imponente lugarpara participar en una ceremonia queha sido el fruto de la decidida y visiona-ria cooperación entre el Gobierno delPresidente Frei y la ESO. Agradece-mos también, esta es la ocasión dehacerlo, a los muchos científicos quehan participado con su esfuerzo eneste proyecto, a los trabajadores que lohan realizado y también al CongresoChileno, representado aquí por los se-nadores de la Región por la compren-sión que manifestó siempre hacianuestro esfuerzo.

Esperamos que el Centro de Obser-vación de Cerro Paranal entre pronto enfuncionamiento, y que los frutos del tra-bajo que chilenos y extranjeros realicenen sus instalaciones, no sólo redundenen un mejor conocimiento del Universo,sino también en beneficios concretospara la ESO, para sus Estados miem-bros, entre los que se cuenta el Reino deSuecia, y para Chile. En esta esperanzaexpreso mis mejores deseos de éxito aesta nueva aventura de la ciencia.

Muchas gracias.

The President of Chile, Mr. Eduardo Frei Ruiz-Tagle, addressing the audience.

5

tica de destruir todos los obstáculos,porque no íbamos a permitir que unaobra como ésta no se concretara ennuestro país.

Estamos felices además de contar hoydía con la presencia de sus Majestadeslos Reyes de Suecia. Queremos simboli-zar en ellos también nuestro reconoci-miento y expresarles también que nues-tra comunidad académica tiene unadeuda de solidaridad con su país por elapoyo prestado a centenares de investi-gadores y profesores Chilenos que vivie-ron durante largos años el exilio enSuecia, y el apoyo también dado a unimportante número de instituciones na-cionales que se han desarrollado graciasa la colaboración y cooperación interna-cional de esta generosa nación.

Por último, quiero hacer un recuerdoespecial ya que el año 1969 le corres-pondió a mi padre, el entonces Presi-dente Eduardo Frei Montalva, inaugurarel primer centro astronómico en el Ob-servatorio de La Silla. Lo hizo en compa-ñía de un destacado hombre de Suecia,el Ex-Primer Ministro Olaf Palme, enesa época Ministro de Educación.

El mundo es pequeño, las historias serepiten, con gran alegría estamos hoydía dando el vamos a esta gran obraque significa, a futuro, desarrollo paranuestro país.

Muchas gracias.

Speech by H.E. President Eduardo Frei

Paranal, December 4, 1996.

Sus Majestades los Reyes de Suecia,Señor Ministro de Relaciones Exterio-res, Señores Senadores de la 2a Re-gión, Señor Intendente, Señores Emba-jadores, Señores Miembros del Consejode la ESO, y un saludo muy especialpara el Presidente del Consejo del Ob-servatorio Europeo Austral, ESO, DonPeter Creola, y también para el ProfesorRiccardo Giacconi, Director General,estimado amigo.

Al contemplar esta obra en medio deldesierto chileno, hecha con el esfuerzoconjunto de un grupo de países euro-peos, de su comunidad científica deastrónomos, y de Chile, sus trabajado-res, sus ingenieros y sus técnicos, sien-to no solo satisfacción, sino que tambiénun legítimo orgullo.

Esta es una concreta señal de comoestamos internacionalizando nuestropaís e insertándolo cada día más en elmundo. Y no es solamente en materiaeconómica.

Todos sabemos los logros que hemosobtenido este año al asociarnos con laUnión Europea, con Mercosur, al llegara un acuerdo de libre comercio conCanadá.

Pero hay también otros aspectos enque nos estamos integrando positiva-mente al mundo. Hoy, nuestra relaciónpolítica con la comunidad internacionalse haya a un altísimo nivel, impensablehace algunos años. La democracia nosolamente nos ha aproximado al mun-do, sino que nos ha dado un prestigioque hoy es respetado en todo el globo.

Y también, en el plano de la coopera-ción científica hemos dado importantespasos hacia una mayor cooperacióninternacional. El establecimiento de esteobservatorio es una muestra clara deello: aquí vendrán astrónomos de Euro-pa y de muchas naciones del mundo, ytendrán la oportunidad de llevar a cabosus investigaciones los más destacadosastrónomos Chilenos.

Esto abre múltiples oportunidades decolaboración y crea también una plata-forma de gran valor para la formaciónde los jóvenes astrónomos chilenos.Nuestras principales universidades severán beneficiadas, así como el país ensu conjunto, que obtendrá el provechode ser sede de uno de los lugares deobservación astronómica de mayorconcentración y más avanzado delmundo.

Otorgo vital importancia a estas acti-vidades conjuntas con países industrialy científicamente desarrollados, porquenosotros enfrentamos como el reto denuestra modernización, el desafío de

construir nuestras propias capacidadesen ciencia y en tecnología.

Chile no podrá proyectarse creativa-mente hacia el futuro, ni vamos a podercrear una sociedad más moderna, másjusta, más equitativa, si junto al desarro-llo y al crecimiento económico no gene-ramos las oportunidades para sus ciu-dadanos, y no logramos simultánea-mente producir y aplicar estos conoci-mientos avanzados en la industria, en laagricultura, en los servicios, en la ges-tión del Estado.

Quizás sea este uno de los desafíosmás grandes que tenemos; esto requieredel esfuerzo de toda la comunidad na-cional, del gobierno, de las empresas,de las universidades. Agradecemos elapoyo al esfuerzo de modernización dela educación. Sabemos que ese es elfundamento de todo el edificio que esta-mos construyendo: sin una educaciónde mayor calidad y más equitativa, novamos a construir un país de oportuni-dades y no vamos a terminar con lamarginación y la pobreza.

Quisiera al terminar, agradecer y darun público reconocimiento a las autori-dades que participan en la ESO, a susejecutivos, y decir que, poco despuésde asumir la Presidencia, como decíael Presidente, no solamente habíanpiedras sino que habían rocas en elcamino. Pero tuvimos la voluntad polí-

The Time CapsuleDuring the December 4, 1996 event at Paranal, President Frei placed a

Time Capsule with selected materials in the wall of the UT1 building. Thiscapsule, an aluminium cylinder of 15 cm diameter and 45 cm long, was filledwith nitrogen gas and sealed hermetically. A commemorative plaque was fixedin front of the cavity. In addition to the various papers listed below, the capsulealso contains a list of contents etched on a metal plate which will survivevirtually indefinitely in this environment, while it cannot be excluded that thepapers may deteriorate with time.

Contents of the Time Capsule:

1. Document signed by President Frei on the occasion;2. Copy of the Acuerdo between ESO and Chile;3. One outstanding scientific paper from each of ESO's member countries (a

list of the titles is available at the ESO Web-site)4. Copy of December 4 issues of the Chilean newspapers El Mercurio and La

Epoca;5. Copies of the written version of the speeches delivered on the occasion;6. Viewgraph and description of VLT, as presented at the exhibition;7. List of VLT Team members;8. Photos of Paranal before and after construction;9. Photo of the VLT model at exhibition.

R.M. West

166

Figure 1: The enclosures for telescopes1, 2 and 3 show the completion of theexternal cladding and the series of open-ings (louvers and doors) that were de-signed in the enclosure to be able tocontrol ventilation and minimise domeseeing. The excavations in the fore-ground are for the foundations of one ofthe tracks that will be used by the auxil-iary telescopes for the VLTI.

Figure 2: The enclosure of UT1 is shownin this picture as it will be used duringobservation. The slits and louvers areopen, the large box in the lower part isone of the air conditioning/air treatmentunits. Above the small door in the con-crete structure one can see the largesliding door 10 metre wide that will beused to integrate the 8.2 m diameter mir-ror with its cell.

Figure 3: This aerial picture taken earlyin the morning from the west side showsclearly the different phases of erection of

Paranal, December 1996M. TARENGHI, ESO

1

2

7

3 4

5

the four enclosures and the steel frame of the control building. In addition, it is possible to see the roof of the delay-line tunnel andthe laboratory for interferometry located in the middle of the flat summit area.

Figure 4: A view of the interior of enclosure No. 1 during functional tests of the rotation of the upper part. The long exposure timeused to take this picture shows clearly the movement of the rotating part. Visible above the electronic cabinets controlling thecomplex functions of the enclosure are some of the ducts and vents of the air conditioning system which are used during the dayto control the temperature of the telescope.

Figure 5: This aerial view shows the three camps utilised by ESO and the contractors in this phase of construction. To the far rightis the old ESO camp. In the centre is the camp erected by Skanska-Belfi and now used by SOIMI and ESO. At the far left is theinitial SOIMI camp. Clearly visible in the background are the excavations for the construction of the Mirror Maintenance Building,the power generation building, warehouse and all the other structures necessary for the operation of the observatory.

(Photographer: H.-H. Heyer)

188

A New Plan for the VLTIO. VON DER LÜHE, D. BONACCINI, F. DERIE, B. KOEHLER, S. LÉVÊQUE, E. MANIL,A. MICHEL, M. VEROLA, ESO Garching

Figure 1: Resolution of an 8-m telescope (left) and a 1.8-m telescope (right) with the seeing characteristics of Paranal. Median seeing refers tor0,550 = 15.6 cm, best 10% of seeing to r0,550 = 22.4 cm, and worst 10% seeing to r0,550 = 9.8 cm.

1. Introduction

Since the decision by the ESO Coun-cil in December 1993 to postpone theimplementation of the VLTI (see TheMessenger No. 74, December 1993),ESO has pursued a programme with theaim to develop all necessary designs toassure a fast implementation as soon asfunds would become available. As a partof this programme, the InterferometryScience Advisory Committee (ISAC), acommittee of astronomers who have abackground in high angular resolutionastrophysics and interferometry, was es-tablished. Its charge is to review thedevelopment of, to define a key scienceprogramme for, and to recommend nec-essary conceptual changes to, the VLTI.It has convened in May and October1995 and in April 1996. The ISAC founda wealth of scientific topics which couldbe addressed favourably or uniquelywith the VLTI. Some of them have beendescribed in more detail (see The Mes-senger No. 83, March 1996). The top-level technical requirements which re-sult from the ISAC scientific require-ments are:

• The VLTI should acquire first fringesaround the turn of the century in order toensure competitiveness with other inter-ferometry programmes.

• There should be a phased approachto the implementation of the VLTI. Thefirst phase shall include the 8-m UnitTelescopes from the start, and shallfocus on spectral regimes where tele-scopes can perform near the diffractionlimit with a minimum of adaptive control.The capability, within the budget, toopen the spectral ranges towards short-er wavelengths for all telescopes, inparticular to the near-infrared regimewith Unit Telescopes using higher-orderadaptive optics systems, shall be inves-tigated.

• Auxiliary Telescopes should have adiameter of order 1.8 m as a compro-mise between cost and sensitivity.

• A field of view for the science targetof 2 arcseconds (“primary beam”) issufficient. In order to enable phase refer-encing as well as narrow-angle astrome-try, the capability to do interferometry ata second field position (“secondarybeam”) within 1 arcminute radius fromthe science beam shall be included.

• VLTI shall have the capability tosimultaneously combine four tele-scopes, including at least two 8-m UnitTelescopes and up to three AuxiliaryTelescopes.

• Beam combination instruments shalloperate in “single mode” (fringe detec-tion on fields of the size of the Airy disk)

in the red, near-infrared, and in the mid-infrared.

• VLTI should have narrow-angle as-trometric capability with the a precisioncomparable to the atmospheric limit.

In view of the progress made in thescientific definition as well as of thegrowing competition, ESO decided inDecember 1995 that the time was rightto reintroduce the implementation of theVLTI into the VLT programme, withinfunds available at ESO as well as in thecommunity. We have developed a “NewPlan” with a technical scope which isadapted to a limited budget, and a signif-icantly modified technical concept whichmeets better the scientific needs. Themain goal of the New Plan is to introducethe original VLTI capability, but in sever-al phases with a modified schedule andsequence off supplies. However, therewill be significant changes to the earlierconcept (see J.M. Beckers, “Planningthe VLT Interferometer”, in The Messen-ger No. 60, June 1990). The new planwas endorsed by the Scientific-Tech-nical Committee, Council has subse-quently confirmed its authorisation to theDirector General to proceed with parts ofthe VLTI programme. Also, additionalfunds of 10 million DM which were madeavailable through an agreement on theenhancement of the interferometric

T E L E S C O P E S A N D I N S T R U M E N T A T I O N

9

Figure 2: Layout of the VLT Interferometer on Paranal.

mode of the VLT, and which was signedby the CNRS, the MPG and ESO in1992 (see The Messenger No. 71,March 1993), could be secured for theNew Plan through an update which wassigned by the three partners end of1996. This article describes the essen-tial elements of the New Plan and recentchanges, as well as the currentprogress.

2. Consequences on the VLTIConcept

2.1. Wavelength range for earlyphases

ISAC recommended to focus for theearliest operational phases on the nearand thermal infrared spectral regimeswhere Unit and Auxiliary Telescopeswould be diffraction limited with tip-tiltcompensation. Figure 1 shows the re-solution attained by 8-m and 1.8-m tele-scopes on Paranal without and with fasttip-tilt only compensation. An 8-m UnitTelescope achieves the best resolution

for wavelengths above 5 µm under me-dian seeing conditions without anycompensation, and is close to the dif-fraction limit for these wavelengthswhen image motion is controlled. A1.8-m Auxiliary Telescope operates atthe diffraction limit for wavelengthsabove 2 µm. Therefore, operation with-out higher-order adaptive optics will beoptimum in the thermal IR (5 µm andlonger) with the Unit Telescopes and inthe near IR (1 µm … 2.4 µm) with theAuxiliary Telescopes. Without adaptiveoptics beyond tip-tilt compensation, theUnit Telescopes would not be signifi-cantly more sensitive in the near-IRthan the Auxiliary Telescopes. ESOtherefore currently pursues the devel-opment of low-cost adaptive-optics sys-tems specifically tailored for interferom-etry with the Unit Telescopes in thenear infrared.

Diffraction within the collimatedbeams inside the delay-line tunnel limitsthe suitability of the Auxiliary Telescopesfor use in the N and Q bands. Theimplications of doing interferometry with

Unit Telescopes in the thermal IR needfurther detailed investigations consider-ing that beam combination occurs aftermore than 20 reflections at room tem-perature. However, fringes appear atprecisely-known frequencies and theirposition can be modulated in time bysmall variations of optical delay. Thisshould substantially facilitate their de-tection against the background.

2.2. Reduction of the field of view

The reduction of the field of view to 2arcseconds makes a substantial reduc-tion of the optics in the delay linespossible. The beam diameter which thedelay line primary optics must accom-modate is now at most 15 cm. Thedimensions of the cat’s eye primarymirror is reduced from a size of aboutone metre to 60 cm. Another conse-quence of the field reduction is thatsevere requirements on the lateral sta-bility of the delay-line carriage motioncould be relaxed substantially, makingsimpler drive concepts viable.

11010

Auxiliary Telescope stations observingbright stars until the time Auxiliary Tele-scopes become available. Figure 2shows the general layout of the VLTI andits major components for Phase A (forPhase B in parentheses). We discusssome major elements of the VLTI devel-opment in the following sections.

VLTI will consist of the following sub-systems after the completion of PhaseA:

1. Coudé optical trains on two Unit Tele-scopes

2. Two delay lines3. Two test siderostats, to be replaced

by4. Two Auxiliary Telescopes, relocata-

ble between 30 stations5. Control system6. Beam combination instruments

The following subsystems will be add-ed during Phase B:

1. Third Auxiliary Telescope2. Coudé optical trains on remaining

Unit Telescopes3. Two additional delay lines

Phase A has just begun and is expect-ed to last until the end of 2001 whenmore or less regular science observa-tions will begin. “First fringes” by thecoherent combination of two Unit Tele-scopes would be observed early 2000 ifthe Phase A coudé trains are installedon Unit Telescopes 1 and 2, the delaylines are in place, and the first instru-ment is completed. First fringes withAuxiliary Telescopes are expected in themiddle of 2001 after installation of Auxil-iary Telescopes 1 and 2. The kick-off ofPhase B, which represents about 1/

4 of

the total cost, depends on the availabilityof additional funds. If started early 2000,interferometry with all four Unit Tele-scopes and three Auxiliary Telescopeswould begin in 2003.

3.1. Unit Telescope Coudé Trains

The early use of 8-m Unit Telescopeswith VLTI requires that coudé opticaltrains, which also were delayed in

Figure 3: On-axis Strehl ratio as a function of wavelength for new UT coudé trains.

A reconstructed field of view of theVLTI is most likely limited by incompleteUV coverage, in particular when thenumber of baselines is initially small.The imaging beam combiner in the labo-ratory which was conceived for a co-phased field of view of 8 arcsec willtherefore not be part of the early VLTI.Phase referencing, for which the fieldshould be actually larger than 8 arcsec,is better implemented with a dual feed(see below). Beam combination will oc-cur in the first phases within the instru-ments.

2.3. Dual feed

The desire for including narrow-angleastrometric capability (Shao et al., 1992)into VLTI (Léger et al., 1995, Quirren-bach, 1995, von der Lühe et al., 1995)eventually resulted in the decision toreplace the original 8-arcsec continuousfield with dual feeds. These allow select-ing two positions at the coudé focus tobe directed towards the delay lines andthe laboratory. One beam will propagatethe light from the science target, theother one will propagate the beam of anearby phase reference (UTs and ATs)or an astrometric reference (ATs only).The concept behind the phase referenceis similar to a reference star for adaptiveoptics, it serves for image and fringetracking should the science target be toofaint. There are limits to the field anglewhich come from atmospheric anisopla-natism.

The coudé foci of both Unit and Aux-iliary Telescopes will be equipped withdual feeds. One position will be on theoptical axis. The other position will bewithin a range of 5 to 60 arcsec fromthe axis anywhere within the coudéfoci. The two beam lines will share thedelay lines for any given telescope,making differential delay lines in thelaboratory necessary. The design con-cept for the dual feed is still in theworks. Although designs for dual feedsexist elsewhere, they are not easilyadapted to VLTI because the require-ments on spectral coverage and astro-

Figure 4: Field dependence of Strehl ratio of new UT coudé trains in two orthogonal directions at1 µm.

metric precision essentially excludetransmissive optics.

The requirements and engineeringimplications of narrow-angle astrometryare very severe and need thoroughstudies. The atmospheric limits to astro-metric precision depend linearly on fieldangle and amount to about 10 µas with a100-m baseline and 30 minutes of inte-gration. Achieving this precision in prac-tice requires the knowledge of the instru-mental differential delay between theprimary and secondary beam to 5 nmaccuracy. At this time, the technical lim-its to determine the differential delay arenot known.

3. VLTI Development

Budget limitations require an imple-mentation of the VLTI in two phases(Phases A and B). The first and mostexpensive Phase A is covered by ESO’sfinancial projections. Since Auxiliary Tel-escopes make up a large part of thecost, it was decided to build only two ofthem in Phase A and to defer the thirdone to Phase B. Because of the longlead times for the Auxiliary Telescopes,first fringes will be observed first withUnit Telescopes which we expect to behighly subscribed. To be able to exerciseand test the interferometry subsystemswith little impact on Unit telescope time,we intend to use simple siderostats on

11

Figure 5: Fractional fringe contrast loss as a function of field position for a representativedistribution of turbulence for Paranal, with 8-m and 1.8-m telescopes.

1993, be re-introduced into the tech-nical and financial concept. Since theuse of Unit Telescopes from Hα to-wards longer wavelengths is empha-sised, we have investigated a simplerand cheaper optical design where largeoff-axis elliptical mirrors are replaced byspherical mirrors. The astigmatism andfield scale anisotropy resulting from thischange are compensated on-axis byreplacing two flats with weak cylindricalmirrors. Figure 3 shows the on-axisStrehl ratio as a function of wavelength.The main contributing aberration isastigmatism. Figure 4 shows the Strehlratio as a function of field angle in twodirections for a wavelength of 1 µm.

The on-axis performance of themodified coudé optics is very satisfac-tory over a large spectral regime, in-cluding the visible, when atmosphericdistortions are taken into account. TheStrehl ratio degrades with increasingfield angle but the resulting fringe con-trast loss will always be small com-pared to the loss due to atmosphericpiston anisoplanatism. This is shown inFigure 5, where the fringe contrast lossdue to the differential piston variationbetween two field positions is shown.This effect is the interferometry ana-logue to the “isoplanatic patch” inadaptive optics. These curves werecalculated based on measurements ofthe turbulence above Paranal by radiosondes in 1994.

Two out of the four Unit Telescopeswill be equipped with coudé trains duringthe first phase. The decision which tele-scopes to equip will be influenced byscientific, schedule, and technical con-siderations.

3.2. Auxiliary Telescopes andStations

The Conceptual Design of the Auxil-iary Telescope System (ATS) was es-tablished in 1992 after a study per-formed by IRAM under ESO contract. Itfeatures 1.8-m telescopes in Alt-Azmount with mechanical bearings. Theaxes are controlled through optical en-coders and friction coupled or directdrives. The optical layout is similar tothat of the VLT 8-m telescope. It in-cludes a coudé train with an intermedi-ate pupil image to enable fast tip-tiltcompensation (and possibly later adap-tive optics). The coudé beam is collimat-ed and sent horizontally to the DelayLine Tunnel through underground lightducts. The telescope is self-protectedagainst wind and adverse weather con-ditions by built-in wind shield protectivecover. The telescope is movable on arail network and can be located on anyof the 30 observing stations through akinematic interface. Each telescope isequipped with a transporter which han-dles the telescope for the relocation andhouses a number of auxiliary equip-ment. During observation, the transport-er is anchored on separate foundationsisolated from the ground to avoid trans-mission of vibrations from the windshield and auxiliary equipment.

ESO has thoroughly reviewed during1995 alternatives to the original designand development concept in an attemptto trade off technical scope and cost withscientific performance. This concernsmainly the diameter of the primary mir-rors, as well as a trade between in-house design and industrial design. Theconclusion was that the original ap-

Figure 6: Variable curvature mirror head.

11212

Figure 8: General interferometry laboratory layout.

proach represents the best compromise.Based on the original concept, ESO willcontract to industry the final design,manufacturing and testing of the ATSincluding optics, mechanics and controlfor the telescope and its transporter in1997. Integration and testing of the Aux-iliary Telescopes at the observatory willbe performed by ESO staff.

3.3. Delay Lines

The purpose of the delay lines istwofold, to equalise optical path lengthdifferences (OPD) and to transfer a pupilat a fixed location inside the interferom-etry laboratory.

The delay lines equalise optical pathlength differences caused by the staticgeometric path length difference be-tween telescopes in any given configura-tion, by the diurnal motion of the astro-nomical source during observation (side-real motion of a star), and by the rapidfluctuations due to atmospheric distur-bances and/or mechanical vibrations. Toperform the compensation of the OPD,the delay lines include an optical retro-reflector (cat’s eye) which moves duringobservation with high precision. Therange required for OPD equalisation is120 m which implies a mechanical strokeof up to 60 m for each delay line. Thismotion must be very accurate andsmooth in order not to distort the interfer-ence fringe pattern formed at the interfer-ometry focus. The longitudinal position ofthe cat’s eye must be continuously mon-itored and controlled by a fast metrologysystem; high dynamic stability and ade-quate drives and bearings are required toavoid vibrations which otherwise couldblur the fringe contrast.

The second purpose of the delay linesis to image the pupil of the telescope at afixed position inside the interferometrylaboratory for beam management. A mir-ror with a variable curvature is used atthe focus of the cat’s eye which hasbeen developed for this purpose by theLaboratoire d’Optique de l’Observatoirede Marseille (Figure 6).

Figure 7 shows a model of the delayline cat’s eye according to a designstudy performed by ESO. The lower partshows a section of the 69 m long foun-dation. The cat’s eye consists of a com-pact Cassegrain optical configurationwith a 60 cm diameter primary mirrorand an effective focal length of 3.5 m.The optics support the beams from twofield positions which enter through theupper two holes and leave the cat’s eyethrough the lower two holes on oppositesides. The delay line covers half of thesupport bench, the other half providesspace for a second one. ESO will con-tract to industry the design, manufactur-ing and testing of the delay lines includ-ing optics, mechanics, control and me-trology in 1997. The integration andtesting of the delay lines at the observa-tory will be performed by ESO staff.

3.4. Beam Combination and FringeDetection

The beam combination laboratoryarea provides the following functionsduring Phases A and B: beam align-ment (image and pupil position align-ment), calibration, fringe tracking, beamcombination and fringe detection. Fig-ure 8 shows an overview of the layoutfor Phase A. Stellar light beams enterthe laboratory from below. A switch-yard, which consists of dichroic andreflective mirrors, directs the light to-wards the instruments and to the imageand pupil alignment and fringe sensorsystems. A calibration unit which pro-

vides a number of sources in the visibleand infrared feeds the alignment unitsand the beam combiner/fringe detectorinstruments and assures their intercali-bration. Wide-band visible and laserlight will be available for this purpose,as well as for the alignment of thebeam combination/fringe detection in-struments. Laser metrology beams feedthe fringe tracking sensor for internalcalibration. Another metrology beamtravels back to the telescopes and de-termines the delay line zero positionfor the internal path during a set-upphase.

A prototype of the fringe sensor iscurrently under development at the Ob-

Figure 7: A model of the VLTI delay line. See text..

13

ratory in the upper half. The switchyard isnear the centre of the layout. Instru-ments are on the top and right. Fringeand image sensors are to the left. Thestructure to the lower right is the metrolo-gy beam launching system.

The two boxes represent the finalbeam combination and fringe detectionstage for the near and thermal infrared.At this time, these areas are not yetdefined. We expect to populate themwith the help of the community. Thefunctions needed here are beam combi-nation (two beams in Phase A, three orfour in Phase B), any dispersion re-quired, detectors and associated read-out electronics, control electronics formechanical functions as needed, cryo-genics, and metrology sensors con-tained in the instrument. The interfacesto which the equipment should fit includea mechanical interface (optics bench),an optical interface (size and position ofincoming stellar and metrology opticalbeams), and a control and data inter-face. The VLT Programme provides anextensive set of standards for electronicequipment whose use in encouraged, aswell as general-purpose software pack-ages which facilitate the integration intothe VLT Observatory. ESO currently de-velops a computer model of the essentialVLTI elements for engineering purposes(Fig. 11). It is planned to also use thesimulated data to model in detail thefringe generation and detection with real-istic astronomical sources and spectralband within the instruments to assessthe performance of VLTI.

ESO’s partners to the agreement onthe enhancement of the VLTI, the CNRS

Figure 11: Sample output from the VLTI end-to-end model under development at ESO.Top left: ray tracing of an individual telescope,top right: resulting OPD map near the point ofbeam combination, bottom left: fringes in acombined pupil image, bottom right: time var-iation of fringe contrast with uncontrolled at-mospheric, seismic and mechanical distur-bances.

Figure 9: Allocation of spectral band usage for VLTI.

servatoire de Côte d’Azur in Nice. Thissensor uses a long stroke synchronousdemodulation technique to detect boththe presence and the position of stellarfringes in the H band. The error signal isfed back to the delay line control systemto provide real-time fringe stabilisation.Figure 9 shows a provisional allocationof spectral bands for metrology, sensing,and science fringe detection purposes.We are developing a test bed for thismetrology system in a nearby laboratoryto validate the concept and its perform-ance with an optical path of severalhundred metres.

Figure 10 shows the central part ofthe tunnel and the laboratory with somemore detail. Primary and secondarybeams emerging from the delay linecat’s eyes are directed towards the labo-

Figure 10: Preliminary configuration of interferometric laboratory.

11414

and the MPG, have already agreed toalso contribute to the beam combinationand fringe sensing instruments in theform of equipment and manpower. ESOinvites the wider community to partici-pate in this effort, and to help definingand developing the beam combinationand fringe detection equipment.

4. Conclusions

We have described the new imple-mentation plan for the interferometricmode of the ESO Very Large Telescope.ESO proceeds with VLTI with a largereffort in terms of manpower than everbefore and involves a significant fractionof the ESO astronomical community intothis endeavour. The development planfor the VLTI has been adapted to thedifficult financial situation by substantialmodification of the concept and by grad-ually increasing the capability. We feelthat the modifications simplify the

project in many areas while strengthen-ing the scientific potential. However, in-troducing the astrometric capability is anew requirement which will certainly bedifficult. When Phases A and B arecompleted, a large subset of the initialgoals for VLTI will have been put intoexistence.

At this time, the VLTI will be far fromcomplete, and one should prepare forfurther upgrades. An important capabil-ity will be the near-infrared capability ofUnit Telescopes through adaptive op-tics. Another important capability wouldbe extending the spectral regime to-wards the visible, opening a widerange of spectral diagnostics and evenhigher resolution. Similarly importantwould be adding a fourth Auxiliary Tele-scope and more delay lines to bringthe number of beam lines up to anumber of eight. VLTI will then unveilits full potential as an optical aperturesynthesis array.

ReferencesLéger, A., Mariotti, J.-M., Rouan, D., Schnei-

der, J., “How to search for extra-solar plan-ets with the VLT/VISA?” in Science with theVLT, Proceedings of the ESO Workshop 28June–4 July 1994, J.R. Walsh and I.J. Dan-ziger (eds.), Springer 1995, pp. 21–32, 1995.

Shao, M., Colavita, M.M., “Potential of long-baseline infrared interferometry for narrow-angle astrometry”, Astron. Astrophys. 262,pp. 353–358, 1992.

Quirrenbach, A., “Astrometric detection and in-vestigation of planetary systems with theVLT Interferometer”, in Science with the VLT,Proceedings of the ESO Workshop 28 June–4 July 1994, J.R. Walsh and I.J. Danziger(eds.), Springer 1995, pp. 33–37, 1995.

von der Lühe, O., Quirrenbach, A., and Koeh-ler, B., “Narrow-angle Astrometry with theVLT Interferometer”, in Science with the VLT,Proceedings of the ESO Workshop 28 June–4 July 1994, J.R. Walsh and I.J. Danziger(eds.), Springer 1995, pp. 445–450, 1995.

The NTT upgrade project has the following goals:1. Establish a robust operating procedure for the telescope to minimise

down time and maximise the scientific output.2. Test the VLT control system in real operations prior to installation on

UT1.3. Test the VLT operations scheme and the data flow from proposal

preparation to final product.

J. SPYROMILIO, ESOLast time I wrote for The Messenger,

the NTT upgrade was progressing well.We had been able to point and track withthe telescope but still had some prob-lems with telescope oscillations. Thetelescope realignment was about to startand the instruments had not as yet beenused. Since then a huge amount of workand progress has taken place.

The alignment of the telescope, apractice run for the VLT Unit Telescopes,was successful. The telescope altitudeaxis was redetermined and the adapter-rotators were re-aligned to ensure thatthe mechanical and optical axes of thetelescope were as close to each other aspossible. We found that the telescopewas within the specifications but never-theless we decided to compensate forthe small misalignment present. Theadapter-rotators were moved by approx-imately 100 microns. In addition, thesetting angle for the tertiary mirror wasalso redetermined. Here the changewas somewhat larger, possibly explain-ing an effect found in the old NTTwhereby when tracking in an unguidedmode across the meridian the starsmoved. This would not have affectedguided exposures. In addition, the posi-tion of the secondary mirror was re-determined. M2 is within the specifica-tions as well, although close to the limit.

During the next months we shall beusing the active optics results to deter-mine if we should undertake to move thesecondary mirror.

In December and January, an intenseperiod of commissioning and integratingthe software, the telescope was broughtfrom an engineering mode to an opera-tional state. The telescope oscillationsproved very hard to fix. A combination ofproblems contributed to causing the tel-escope to misbehave at random inter-vals. Only in the very last days of Janu-ary were the most critical problems real-ly solved. The telescope now tracks verywell in most areas of the sky. However,there seems to be significant friction,especially in the altitude axis, causingsome problems as the telescope movespast the meridian (where the altitudespeed slows down to zero before chan-ging sign). The azimuth axis seems tobehave very well but we have not beenrunning the telescope long enough to besure that all is well.

The telescope pointing is excellent,with the exception of a zone of avoid-ance around the zenith. This of course isa problem all alt-az telescopes have andis not new to the NTT. However, we hopewith more detailed pointing models wewill be able to make the zone of avoid-ance small enough so as not to impactany scientific programme. On otherfronts the telescope control system is

also improving. Image analysis is beingrun regularly and although as yet it is notrobust, it is functioning and the tele-scope image quality is back to the excel-lent values we have come to expect fromthe NTT. Autoguiding is now workingreasonably well, and automatic guidestar selection by the control system isavailable either in blind mode (let thesystem select for you) or manual mode(let the system offer you a choice).Selection of the guide star to be used isa matter of clicking with the mouse onany suitable star seen on the guidecamera output.

We believe that the new NTT is atelescope that is easy to use, with mostof the complexity hidden from the userby the graphical user interfaces. Nocryptic commands need any longer beissued and a fair amount of work hasgone into the on-line help available.

On the instrumentation front, the newACE controllers have now been com-missioned and we have an improvementof a factor of 2 in readout time over theold VME controllers for the same noisefigures. In addition, the re-aluminisationof the telescope mirrors and the refur-bishment of the instruments have givenus an improvement in throughput of afactor of 2 to 3 in most bands. In thecase of SUSI we also replaced thedetector with one which has high UVsensitivity. The early estimates suggest

Oskar von der Lüheovdluhe@eso.org

15

we may have improvements as high as afactor of 20 in the U band in SUSI. TheSUSI control software is going throughfinal testing and seems to work well,although some parts of the system, es-pecially linked to the interaction of SUSIwith the telescope, are still not fullytested. EMMI underwent a completeclean-up and has had a new grismwheel, the refurbished red camera, andthe new RILD mirror re-installed. Thenew EMMI software, which is almostidentical to that running in SUSI is alsobeing tested.

The plan for the NTT upgrade whichwas approved by the STC and ESOmanagement had the telescope returningto partial operations on the 1st of Febru-ary. Some of you who submitted applica-tions have already been through ESOand completed the Phase 2 ProposalPreparation process. After seven monthsof keeping to the daily NTT upgradeschedule religiously, I am sorry to say thatthe project has slipped. The first execu-tion of a service programme took place 4days later than planned for, on February5th. The telescope is still in shared-risksmode and we have quite some way to gobefore the requirements set on the tele-scope and instrumentation are met. How-ever, we believe we are close enough to

be exercising the telescope and trying todo real science. I should emphasise thatwe are not guaranteeing any perform-ance figures before the 27th of Junewhen the NTT returns to full operations.However, we shall do out best to deliverscience quality data to as many users aspossible.

On the operations front, an enormousamount of progress has been made.The output of P2PP is now beingpassed to the control system in a trans-parent fashion and provides the param-eters for the automatic execution of theprogrammes. The observing and targetacquisition templates are running andsome calibration templates are alsoavailable (e.g. automatic focusing andsky flat fields). In this way we canensure to be doing exactly what theusers specified. In December/January,we took delivery of the archive andpipeline data reduction workstations.They have been integrated into the NTTcontrol system and while they are stillrunning prototype software, the data doflow across the complete system and doget automatically reduced. The reduc-tions right now are pretty rudimentarybut there is no reason why this workcannot be expanded to provide theusers with a first cut at their data.

As mentioned earlier, the NTT up-grade still has 5 months to run. For thetime being we have been finding andsolving the 1 in 1 bugs. Over the nextmonths we have a lot of work to do. Notonly do we have to acquire the neces-sary experience with service observing,but also find the 1 in 10 and hopefully the1 in 100 bugs. We also have somemaintenance we wish to do, which wedid not wish to perform while the firstphase of the upgrade was taking place.This was a decision based on variousplanning and technical reasons. Weplan that the NTT shall continue over thenext months to improve in reliability andexpand the functionality of the tele-scope. In April we plan to start with somelimited service observing using EMMI.

I would like to take the opportunity tothank all our colleagues in all the ESOdivisions that have helped to make thenew NTT a reality.

Staff Movements

I would like to welcome to the NTTteam our new instrument operator Nor-ma Hurtado.

Jason Spyromiliojspyromi@eso.org

1. Introduction

The IQ of a telescope degrades asthe telescope moves from zenith. For amechanically perfect telescope, thisdegradation is due only to the naturalseeing which worsens with increasingairmass. However, for real telescopes,mechanical flexures accelerate thedegradation. These flexures affect the

The La Silla News Page

The editors of the La Silla News Page would like to welcome readers of the sixth edition of a page devoted toreporting on technical updates and observational achievements at La Silla. We would like this page to inform theastronomical community of changes made to telescopes, instruments, operations, and of instrumental perform-ances that cannot be reported conveniently elsewhere. Contributions and inquiries to this page from thecommunity are most welcome. (P. Bouchet, R. Gredel, C. Lidman)

Earlier articles about the Image Quality(IQ) at the 3.6-m telescope (part I to partIV, see previous Messenger issues) con-cerned the IQ near zenith. Final conclu-sions and improvements have been writ-ten already [1]. In this article we willpresent the results of the study sinceSeptember, 1996, when we started toanalyse the telescope behaviour far fromzenith.

The Image Quality of the 3.6-m Telescope (Part V)What Happens far from Zenith?

S. GUISARD, ESO-La Silla

structure of the telescope itself and themirror supports.

Flexures of the Serrurier struss mis-align the two mirrors of the telescopeand cause mainly decentring coma ab-erration. Malfunctionning of the mirrorcell with zenith distance is more likelyto create astigmatism and triangularcoma. We shall see that the 3.6-m is noexception to this rule.

11616

The triangular aberration is in fact afixed defect of the mirror after it waspolished. But in order to compensate forthis aberration, the polisher, twentyyears ago, deliberately modulated theaxial force on the astatic levers belowthe mirrors. The aberration is thereforecompensated at zenith but reappears asa cosine function of zenith distancewhen the telescope is inclined. This isexactly what we measure.

Part of the astigmatism has the samecause; the other part is due to thepneumatic lateral supports of the mirrorwhich do not work properly.

3. Solutions

3.1 “Semi-activation” of M2

Decentring coma can be corrected bymoving the secondary mirror. In thecase of the 3.6-m, we call it ‘semi-activation of M2’, as, unlike the NTT, themovements of the secondary mirror willde-point the telescope. Therefore, comacorrection can only take place betweenexposures.

Successful tests of semi-activationhave been done during test nights.Coma could be completely corrected forzenith distances of 45 and 60 degrees;however, this correction did not reducesignificantly the image size because oth-er terms, like astigmatism and triangularaberration, are as important as coma(see Table 1). Furthermore, the outsideseeing depends on zenith distance andthis makes the correction of coma less

TABLE 1

Zenithal distance (deg) 0 15 30 45 60

Coma (d80%) 0.20″ 0.20″ 0.35″ 0.40″ 0.45″Astigmatism (d80%) 0.15″ 0.15″ 0.20″ 0.30″ 0.45″Triangular (d80%) 0.15″ 0.15″ 0.15″ 0.20″ 0.30″Other terms (d80%) 0.25″ 0.25″ 0.25″ 0.25″ 0.25″

Total (d80%) 0.40″ 0.40″ 0.50″ 0.60″ 0.75″

impressive than expected. For example,the IQ at 60 degrees zenith distance witha natural seeing of 0.60″ at zenith will be1.22″ without coma correction and 1.15″with the correction. Without degradationof the image by astigmatism and trian-gular aberration, this would be 1.05″.

Figure 1 gives the gain in IQ in % as afunction of zenith distance and outsideseeing (given at zenith) when only comais corrected. On this graph, the otheraberrations have the values given byTable 1. If we set the criteria for correct-ing the decentring coma as an improve-ment of 10% in the image size (whichcorresponds to a gain of 1.2 in exposuretime), we see that it is worth semi-activating M2 only for outside seeingbetter than 0.4″ and zenith distanceslarger than 45 degrees. This situationmay happen for a few hours in a yearonly!

Figure 2 also gives the gain inducedby coma correction but assuming thatastigmatism and triangular keep theirvalues of 0.15″ for any position of thetelescope. In that case semi-activatingM2 becomes more attractive.

3.2 Eliminating the dependence ofastigmatism and triangularaberrations on zenith distance

As we saw above, this correction is anecessity as it will improve images andmake the correction of coma useful. Itrequires significant work on the primarymirror cell, mainly the axial supports.Solutions have been found already. Theidea is to change the force distributionon the astatic levers below the mirror asif the mirror were perfect. This would ofcourse introduce at zenith the constanttriangular pattern of the mirror. Thisaberration has to be compensated byaxial forces independent of zenith dis-tance like springs for example.

Part of the astigmatism should also becorrected in this way, the other part ofthe astigmatism will be removed byimprovements of the lateral pneumaticmirror support system.

Technical time has been requestedfor April 1997 to install load cells on allthe mirror supports (33 axial and 21lateral supports). This change involvesthe manufacturing of nearly 200 me-chanical pieces, which has started al-ready in the La Silla workshop, and 2weeks of telescope time to install theaxial load cells in April. The installation

2. Measuring the Optical Qualityat Large Zenith Distances

This work started in September 1996and is continuing. It is only because of arigorous test scheme, involving manytest nights, that the telescope aberra-tions are now understood. The aberra-tions are calculated from extra-focal andintra-focal images using the curvaturesensing method.

Since September 1996, more than1300 defocused images have been tak-en and analysed, with the distance fromzenith ranging from 0 to 60 degrees andwith the telescope pointing in differentpositions. Movement cycles were donealso to study the hysteresis of theaberrations. Two detailed technical re-ports have been written [2], [3]. Here wewill only summarise the main results.

Table 1 summarises the variation ofeach of the aberrations with zenith dis-tance. The aberrations that change mostare coma, astigmatism and triangular.The spherical aberrations (at the newfocal plane [1]) and quadratic astigma-tism are included in the row ‘other terms’and do not change much with zenithdistance. The values in this table areaveraged over many azimuth positions.In practice, the telescope behaviourchanges if it points South, East, West orNorth. The aberrations are given in arc-sec d80% (diameter of the circle con-taining 80% of the light).

The aberration that was expected tochange most was coma; however, thechanges are much smaller than whathad been reported several years ago.Indeed, the flexure of the telescope ismuch smaller than stated in the oldreports. Gerardo Ihle (Mechanics Sup-port Team, La Silla) confirmed this by afinite-element analysis of the telescopestructure.

The surprise came from the astigma-tism and triangular aberrations whichhave large changes with zenith dis-tance. These aberrations are due to theprimary mirror and its support.

Figure 1.

17

of the lateral load cells will follow, and weshould be able to have full informationon the support forces by the end ofAugust this year. Time will be requestedbefore the end of the year to change theforce distribution below and around themirror, and to install springs on the axialastatic levers.

A detailed planning of the interventionon the primary mirror support has beenprepared by Roland Gredel. We hopethat all the necessary work on the mirrorcell can be done within the year.

4. Conclusions

The behaviour of the telescope atlarge zenith distance in terms of opticalquality has been investigated carefullysince September last year. Improve-ment plans have been proposed andwork has started already. The phase weare entering now is very delicate as itinvolves the intervention on the mirrorsupport itself. Everything will be donenot to degrade the optical quality atzenith while changes are made. Moretechnical time will be needed before theend of the year to decrease the aberra-tions for all telescope positions. The3.6-m is getting better; however, muchwork still has to be done.

Figure 2.

During October 1997, EFOSC2, theimaging spectrograph now at the 2.2-mtelescope, will be moved from the2.2-m to the 3.6-m. The current spec-trograph on the 3.6-m, EFOSC1, willbe de-commissioned. EFOSC2 on the3.6-m will have higher throughput, alarger field of view and significantly

better image quality than what waspossible with EFOSC1. The significant-ly smaller pixel scale of the EFOSC2CCD, 0.18 arcsec per pixel at the3.6-m telescope, compared to 0.61 arc-second per pixel of EFOSC1, will allowobservers to fully exploit the recentprogress in the improvement of the

From the 3.6-m and 2.2-m Teams 3.6-m image quality (see S. Guisard’sreports in The Messenger, December1996, March 1997). Multi-object spec-troscopy will not be avaible with EFO-SC2 during Period 60 but only in Peri-od 61 and thereafter.

For ESO time during period 60, the2.2-m will be dedicated to observationswith the two infrared cameras, IRAC1and IRAC2b.

With the 3.6-m upgrade in progress,the 2.2-m telescope will be the only majortelescope on La Silla which still runs offan HP-1000 computer. To make sure thatthe 2.2-m telescope will be maintainable

2.2-m Telescope Upgrade Planinto the next decade, we are preparingan upgrade plan for the telescope whichwill be presented to the STC in thebeginning of May. The upgrade plan willdiscuss both the possible replacement of

much of the electronics and computersas well as possible modifications to thedrive system and possible improvementsof the image quality. This will be a goodopportunity to address some long-stand-ing problems with this otherwise excel-lent telescope. J. Storm

News from the Danish 1.54-m TelescopeJ. BREWER and J. STORM

TCS User Interface Upgrade

A new TCS graphical user interface(GUI), written by Gaetano Andreoni us-ing the VLT panel editor, is now in useat the Danish 1.54-m telescope. Ob-servers will find that the frequentlyused telescope and adapter controlsare now contained within a single win-dow, while lesser-used functions are

within a dismissable pop-up window. A‘virtual handset’ can also be enabledfrom the main control window. Thenew interface retains the same func-tionality as the old interface, though itis simpler and more user friendly. Thenew interface also offers the advantag-es that it is significantly more robustthan the old system and is easily mod-ifiable.

DAISY

A new instrument GUI, based on theGUI at the Dutch telescope, is now inuse at the Danish 1.54-m telescope.DAISY (Data Acquisition IntegratedSYstem), written by Eduardo Robledo,combines the control of the CCD Cam-era, the DFOSC (Danish Faint ObjectSpectroscopic Camera), the FASU (Fil-ter And Shutter Unit), and the telescopefocus control all into one package. Ob-servers will find that DAISY is very easyto use; the operation is intuitive andthere is little to remember. The DAISY

References

[1] S. Guisard: The Image Quality of the ESO3.6-m Telescope (Part IV): Better than 0.6″,The Messenger No. 86, December 1996.

[2] S. Guisard: 3.6-m + CAT Upgrade, “Re-port on the test nights 7th and 8th Octo-ber 1996”, 3P6-TRE-ESO-032-009, Oc-tober 1996.

[3] S. Guisard: 3.6-m + CAT Upgrade, “Re-port on the test nights 18/10/96 and 19,20, 25, 26, 27/11/96”, 3P6-TRE-ESO-032-010, December 1996.

Stephane Guisarde-mail: sguisard@eso.org

11818

interface allows observers to define asequence of exposures with each expo-sure having its own filter/grism/slit com-bination. In addition, it is possible todefine a sequence of sequences. Bycombining the various controls into onepackage, DAISY both optimises andsimplifies observations.

Focus Pyramid

During Danish time at the beginning ofJanuary, a new focusing device wastested and installed in DFOSC by PerKjaergaard Rasmussen and Michael An-dersen from Copenhagen University Ob-servatory. The prism works in the samemanner as focus wedges, which havebeen in use in the focal reducers at LaSilla for many years. However, instead ofsplitting the telescope pupil into twoimages, as is done with a wedge, thefocus pyramid splits it into four compo-nents. The advantage is that alignment isnot critical for the resulting focus esti-mate, which makes it simpler to maintain.

A new observing batch has beendeveloped for analysing the images and

it is now possible to check the focus inless than a minute. This will enable amuch more frequent refocusing of thetelescope than has previously been feasi-ble and thus help to improve the imagequality.

Image Quality Improvements

Following the promising results whichhave been achieved at the 3.6-m tele-scope (see e.g. S. Guisard, The Mes-senger 86, p. 21), an investigation of theimage quality of the Danish 1.54-m isalso in progress in close collaborationwith the group at Copenhagen Univer-sity Observatory. It is clearly a complexproblem and we must proceed one stepat a time to achieve consistent results.The first step is to improve the monitor-ing of the current performance and am-bient conditions to determine which arethe major sources of the seeing. Apartfrom the atmospheric seeing there areprobably significant contributions to theseeing from the dome, the telescopeitself as well as from the finite size of thepixels of the detector.

echelle grating and the long camera forboth the RED and BLUE cross-dis-persers. Additionally, S/N estimates aregiven for the two standard configura-tions (31.6 lines/mm echelle + RED(BLUE) cross-disperser + Long Camera+ CCD #37); new bright, 1.6 nm resolu-tion flux standard stars have been add-ed to the table of standards stars; refer-ence exposure times for Quartz andThorium-Argon lamps exposures withthe different neutral density filters andcolour filters are given; and, finally, anew Long Slit filter centred at 671.1 nm,which arrived very recently, have beenadded to the list of Long Slit filters.

The structure of the manual differswith the 1989 version. For example, amore comprehensive introductory chap-ter is given. It summarises the proper-ties of CASPEC and compares it withthe two other high-resolution spec-trographs at La Silla, namely EMMI atthe NTT and the CES at the 1.4-m CATtelescope. In this chapter, the CASPECobserving modes are listed and all thebasic characteristics of the instrumentare summarised. This first chaptershould be read before writing the Ob-serving Proposal; in particular, it shouldallow the observer to discern whetherCASPEC is suitable or not for yourscientific programme. Chapter 2 de-scribes, in detail, the characteristics ofthe different components, and shouldallow one to choose the most suitableconfiguration for a given project. Chap-

ter 3 (Instrument performance) shouldalso be read when writing the proposalin order to ascertain the feasibility of theproject and to estimate the number ofnights needed to carry out the project.This Chapter is complemented by asimulator (see below). Finally, Chapters4 and 5, may be skipped when prepar-ing the proposal, but should be readbefore carrying out the observations.

CASPEC Exposure TimeCalculator

The ESO 3.6-m CASPEC ExposureTime Calculator (Version 1.0) has beencompleted and is available on the3.6-m Team CASPEC Homepage.

The calculator can simulate the mostcommon observing set-ups for CAS-PEC; future versions may include LongSlit mode and the Zeeman Analyzer. Itis based on both observed andtheoretical parameters. The simulatorwill be used as a starting point for theCASPEC Physical Model.

This calculator takes into accountmany observing parameters not includ-ed in other simulators or S/N estimators(e.g. Chapter 3 of the CASPEC man-ual). For example, the phase of themoon, seeing slit losses, airmass cor-rections and basic colour terms in theassumed input magnitude are included.The calculator can either be used toestimate the S/N in the resulting spec-tra for a given exposure time or to

New CASPEC Manual and SimulatorS. RANDICH, ESO-Garching, and M. SHETRONE, ESO-Chile

1. CASPEC Operating Manual

A new operating manual for theCassegrain Echelle Spectrograph(CASPEC) at the 3.6-m telescope isnow available.

The manual can be retrieved fromthe 3.6-m & CAT WWW page:

(h t tp : / /www. l s .eso .o rg / l as i l l a /Telescopes/360cat/html/CASPEC/caspec.html).

Since 1989, when the last operatingmanual was written (ESO OperatingManual #2), CASPEC has undergoneseveral modifications; the main onesbeing: the installation of two new (REDand BLUE) cross-dispersers; CCD up-grades; a new clamping system, whichreduced the flexure in the spectro-graph; and the addition of new colourfilters to the CASPEC set. An update tothe manual was written in 1993 byL. Pasquini. The major changes sincethen are the installation of a new high-efficiency CCD (ESO CCD #37) and thefact that only one grating (31.6 lines/mm) and only the Long Camera arepresently offered.

The new manual describes CASPECin its present status. This includesall the information reported in sepa-rate documents since 1989 (e.g.,new cross-disperser efficiencies, 3.6 +CASPEC + CCD #37 overall efficiency,CCD #37 characteristics), plus new in-formation; in particular, the inter-orderseparations for the 31.6 lines/mm

E N D O F L A S I L L A N E W S P A G E

J. Brewere-mail: jbrewer@eso.org

The physical size of the pixels corre-sponds to 0.39 arcsec on the sky, but asthe overthinned LORAL-CCD smearsthe charge, the effective pixel size issignificantly larger, especially in theblue. Still, images with a seeing of0.9 arcsec have been obtained withDFOSC so the potential for sub-arc-second images is definitely there.

To investigate the effect of mirrorseeing, the mirror cover has been liftedsome 10 cm at the beginning of Janu-ary and a couple of fans have beenmounted following the example of the3.6-m. The first tests suggest that theforced mirror ventilation reduces thetypical FWHM of images by 0.2–0.3arcsec. More tests under a wider rangeof external conditions will have to becarried out to derive a clearer picture.

The next step will be to assess theamount of dome seeing and the ways inwhich this contribution can be reduced.

19

Figure 1: Exposure Time Calculator.

The Eclipse SoftwareN. DEVILLARD, ESO

algorithms to most standard softwarepackages. The result had to work simul-taneously under most other data-processing software packages, andmodifications should be easy to quicklyadd new methods as they come out ofthe Working Group.

The decision was taken to develop inANSI C, which acts as a lingua franca forUnix-based workstations. The set ofsupported machines is: any machinerunning a Unix Operating System. In thisway, we ensure portability, code reusa-bility, and integration into all major data-reduction software packages.

Overview

eclipse offers an open environmentfor data-reduction algorithm develop-ment and simple pipeline processings.At the highest level, the user will find amanual describing how to observe andcalibrate data using Adonis in its differ-ent modes, with associated examplereduction scripts ready to be called fromwithin standard software or directly fromthe Unix command line. So far, coveredreduction aspects are:

• pixel gain map creation• dead pixel detection and correction• simple data classification from FITS

information• sky-subtraction and flat-field division• shift-and-add• data extraction and merging• statistics computation• cube arithmetic• spatial filtering• Fourier transform• image resampling

All of these high-level procedures arerunning without user interaction, whichis typical of a pipeline, number-crunch-ing approach.

software development to produce theeclipse software.

When we took the decision to developeclipse, we had to set up how far itshould go when reducing the data, andwhich language/platform should be host.By questioning Adonis users, we foundout that writing the Data Reduction Soft-ware within a standard reduction pack-age such as MIDAS, IRAF, or IDL, wouldcertainly reduce the number of potentialusers. Furthermore, high-resolution datareduction techniques are constantlyevolving, and it is not always easy oreven possible to add up brand new

A Brief History

Adonis, the Adaptive Optics instru-ment on use on the 3.6-m telescope onLa Silla, is available for the astronomicalcommunity since April 1993. Experiencehas been acquired by many users sincethen, and what Adonis needed most wasto gather this experience, and then offera set of data-reduction facilities andguidelines for observation and calibra-tion procedures. To get all the expertise,a working group was created, dealingwith High-Resolution Infra-Red Data Re-duction. Its output was directly fed into

estimate the required exposure to ob-tain a spectrum with the given S/N.

For example, input parameters mightinclude: a 15th V-magnitude star withan effective temperature of 5000K. Theobservations are to be taken duringbright time at 1.5 airmasses and with1.0 arcsec seeing. The wavelength ofinterest is 6000 Å employing the Redcross-disperser and a 1.2 arcsec slit(see Fig. 1).

With these input parameters S/N of15 per pixel (in the summed spectra)can be achieved in an exposure ofapproximately 2600 seconds.

The simulator can be used to plan anobserving run but should not be substi-tuted for actual observing!

S. Randich; e-mail: srandich@eso.org

12020

Going one layer down reveals thateclipse is made of 32 Unix commands,each of them delivered with a manualpage containing explanations about al-gorithms, input parameters and files,and several examples. By combiningthese commands into scripts, it is possi-ble to create simple pipelines handlingbasic calibration and cosmetic proce-dures in a fast and efficient way. Noticethat commands are scriptable from anyhigh-level programming language, e.g.MIDAS, IRAF, IDL, perl, tcl, or standardUnix shells. Most of all, feedback fromthe working group allows a fast develop-ment of the latest algorithms, specific toAdonis.

Going to the very heart of eclipsemeans getting into 30,000 lines of docu-mented ANSI C code. Written in anobject-oriented way, it provides the pro-grammer with a set of typical objects(images, cubes, pixel maps . . .) todevelop with. From this level, it is pos-sible to embed validated algorithms,which allows strong optimisations bothin terms of speed and disk space, andencapsulation of these algorithms forhigher-level processes. Notice that opti-misation in the case of Adonis is a majorconcern: with data cubes reaching sizesof several hundred megabytes, it is easyto overflow disks, saturate memory, or inthe best case monopolise computingpower for several hours, preventing anyother computer work! And the situationis likely to get worse on these aspects,with ever increasing detector sizes…Nearly all algorithms included in theeclipse code have been especially de-signed to handle these aspects in thebest possible way, making use of auto-matic disk swapping and delicate mem-ory handling routines.

Notice that a link to off-line data-processing software is essential. eclipsedoes not provide any complex post-processing algorithm such as deconvo-lution, nor does it contain any image ordata displayer. Global data analyserssuch as MIDAS provide the full range offunctionality needed for evolved post-processing and analyses; eclipse is tobe used as a pre-processor, a signal-processing engine.

Figure 1: From left to right: Figure (a) is a raw frame, Figure (b) is sky-substracted, Figure (c) is bad pixel corrected and flat-fielded. The wholeprocess used 1.26 seconds CPU time on the La Silla main server, including flat-field creation from a set of sky images and bad pixel detection.The total amount of involved data is about 5 megabytes.

The Adonis Pipeline

A set of scripts and Unix commandshas been especially designed to takecare of most basic data reductions forAdonis. This has been set up duringAugust 1996 on La Silla, with help fromobservers, telescope and Adonis teams.

The following set of operations hasbeen automated:

• flat-field creation• bad pixel detection• sky extraction from data cube, aver-

aging, and subtraction from objectframes

• average of the resultThere is a preparation phase, during

which the observers have to identify theirfiles according to their logbook, and sortthem out, preferably in separate directo-ries. It is then possible to design quickly aUnix script to launch a unique reductioncommand on all directories, and getcleaned data in a very short time.

Let us browse through an example:Figure 1(a) shows a raw image of an

object taken in thermal infrared. Noise isdominant, the object is barely visibleamong the noisy background. Figure (b)shows the same image after subtractionof the closest (in time) averaged sky.On Figure (c), bad pixels have beencleaned, and a flat-field division hasbeen applied. The star is now cleaned,ready for deeper analysis.

The whole process applied on the LaSilla main server (kila) took 1.26 sec-onds CPU time, and 12.73 seconds usertime with 70 active users, for a totalamount of more than 5 megabytes ofdata. The given figure includes process-ing time for flat-field creation and badpixel detection. Such processing speedimplies that data cleaning takes place inless time than needed for acquisition,which makes such a pipeline a goodcandidate for on-line display.

Future Developments

For the moment, eclipse handles di-rectly all imaging modes in all wave-lengths for Adonis. By combining itslow-level functionalities, it is certainly

possible to develop new modules, specif-ic to other instrumental modes (coron-ography, spectro-imaging with Fabry-Perot, polarimetry, astrometry, . . .).Christian Drouet D’Aubigny (ESO-Garching) is right now working on theFabry-Perot extension for Adonis, to-gether with Patrice Corporon (ESO-Chile), both coopérants.

Julian Christou’s deconvolution soft-ware: idac, has been recently linked toeclipse. In this way, the user also has aniterative blind deconvolution algorithmreadily available in La Silla’s machines.Though it would probably not be used ason-line software, idac may need some ofeclipse functionalities to provide defaultstart-up estimation images, create badpixel maps, or apply cube arithmetics.

eclipse is covered by the MIDAS GNUpublic general license, which allows us todistribute it on the World Wide Web. Itshould be included in the future in MIDASdistribution as a contribution for Adonis.MIDAS has been enhanced to supportdata cube manipulations, viewing, anddirect FITS access, which ensures that acomplete toolbox is available to theAdonis user for data reduction.

Finally, portability made it possible toinstall eclipse in many other observato-ries and institutes, for data reduction ofeventually other instruments on othertelescopes.

I would like to thank all people partici-pating in the High-Resolution Data Re-duction Working Group for their kind helpand patience in explaining to me thesecrets of infrared data processing. Notethat for 16 months the development teamconsisted of a single person (N.D.), but,in fact, many people worked on thealgorithms, and it would be hard toacknowledge them all here. The goal hasbeen reached: Adonis is now provided tothe astronomical community togetherwith a documented and dedicated datareduction package.

Most eclipse documentation is availa-ble on-line on the ESO server at thefollowing address: http://www.eso.org/~ndevilla/adonis/eclipse.

Nicolas Devillarde-mail: ndevil@eso.org

21

S C I E N C E W I T H T H E V L T / V L T I

The VLT Science Cases: a Test for the VLTA. RENZINI and B. LEIBUNDGUT, ESO

Introduction

In preparation for VLT operations,ESO is gearing up its interaction with itsastronomical community. It is importantfor ESO to understand the needs andrequirements of future users and toaccomplish as high a standard of serv-ice as possible. In addition, we have gotto know the most demanding needs ofthe community in order to verify at theearliest possible stage that the VLT willactually meet such expectations. To thisend, we have been actively seekinginput from future observers.

We asked several astronomers toprovide us with ‘VLT Science Cases’(SC), potential programmes of high sci-entific interest and especially demand-ing in terms of required performance. Inparallel, a set of ‘Reference ObservingProposals’ (ROP) have been created atESO to cover some more technical as-pects. The SC and ROP are used as aconceptual test of the VLT, of its first-generation instrumentation (ESO Tech-

nical Preprint No. 72), and of the VLTscience data flow (P. Quinn, The Mes-senger No. 84). They are meant to helpus to achieve a quantitative understand-ing of the expectations concerning thescientific performance of the VLT, andidentify possible improvements of theproject. This should provide an earlyappreciation of potential shortcomingsthat could be eliminated before the VLTis offered to the community, and/or anearly identification of needs that may besatisfied by appropriate upgrades in theinstrumentation plan and VLT opera-tions. The ROP were also establishedfor tests of the VLT operational system(cf. P. Quinn, The Messenger No. 84,and J. Spyromilio, The Messenger No.85). The SC were analysed by the ‘Sci-ence Performance Group’ co-ordinatedby the VLT Programme Scientist (A.Renzini) with the goal to identify actionswhich could improve the scientific per-formance of the VLT. This approach iscomplementing the other user inter-action channels through the standing

oversight ESO Committees, and theVLT/VLTI workshops held over the lastfew years.

Up to now 28 SC and 7 ROP havebeen completed (Tables 1 and 2). Thecollection of VLT Science Cases illus-trates the wide involvement of ESO’suser community. More than 90 astrono-mers from all ESO member countrieshave contributed in one form or anoth-er. To this are added over 20 ESO/ECFstaff astronomers in Garching and inChile. The effect is to strengthen theties between ESO and its scientificcommunity on a rather informal, butdirect, level. In effect, through the SC a‘network’ has started to be establishedwithin the community which can be rap-idly accessed for specific scientificquestions and to explore the relativeeffectiveness of various options con-cerning the VLT. We are very grateful toall the astronomers that have participat-ed in the SC, and believe they deservethe gratitude of the community for hav-ing spent some of their time for thecommon benefit. There will be indeedno privilege or tangible reward for them,with the exclusion perhaps of what maycome from having started thinkingabout VLT observations in a more pre-cise and practical way. With VLT firstlight being already so close, the VLThas to become a much stronger realityin many European astronomers’ mind inthis and the coming year. The SC werealso meant to stimulate and assist thisprocess.

The size of this SC network has beenso far limited only by our capacity tomaintain interactions with it and to fruit-fully analyse its input. Yet, the exercise isnot over and we plan to “call” for a newset of SC this year. The selection of thetopics of the SC was done to cover someof the most active areas of astronomicalresearch for which we believe the VLTwill become a major contributor. To thiswe added observational techniques we

TABLE 2: VLT Reference ObservingProposals

• Quasar absorption lines• Gravitational arcs in clusters of galaxies• The nature of AGNs• Dynamics of dwarf galaxies• Stellar winds on the AGB• Comet impact on Jupiter• Gravitational shear in clusters of

galaxies

TABLE 1: VLT Science Cases

• Evolution of galaxies from z = 0.6 to z = 4.3• Testing the redshift evolution of potential wells and scaling relations of galaxies• The evolution of cluster galaxies• A search for binaries in globular cluster cores• High redshift radio galaxies:

(a) Their stellar content and surrounding clusters(b) Gas kinematics and jet/cloud interactions(c) Imaging- and spectro-polarimetry to identify and separate the scattered component(d) UV nebular diagnostics of the extended gas(e) Absorption studies of Lyα and C IV as probes of the circumnuclear gas particularly

the cold component• Distant cluster of galaxies• Measuring Ω with weak gravitational lensing• Dark matter searches with weak gravitational lensing from a drift-scan image• Measuring ΩΛ from (weak and strong) lensing modelling of rich clusters• The star formation history of ultra-low surface-brightness galaxies• Extending extragalactic PN as probes of galaxies out to 50 Mpc• Astronomy of isolated neutron stars• Spectroscopy and photometry of very distant supernovae• Chemical evolution and star formation history in nearby galaxies• A complete sample of 1000 active galactic nuclei to R = 23.5• Dynamics of the Carina dwarf spheroidal galaxy• Properties of compact emission line galaxies up to z = 1.2:

velocity dispersions and emission line ratios• Chemical abundances of stars in globular clusters• Physics of main-sequence and slightly evolved stars in young open clusters• RR Lyrae stars in the LMC: tracers of the structure and the metallicity of the old

population• The galaxy population in the redshift interval 0.5 < z < 5• Redshift evolution of chemical abundances in damped Lyman-α system• Dynamics of galaxies at the VLT with FUEGOS/ARGUS• Optical identification of gamma-ray burst sources

12222

expect to be essential for the success ofVLT observation, i.e. demanding pro-grammes in terms of preparations andobservational skills. We are fully awarethat the current set of SC does not coverall of astronomy, but it is the purpose ofthe next rounds to start filling in thegaps. The scientific contents of the SCare treated confidentially.

The VLTI has been excluded from thecurrent science cases. The Interferom-etry Science Advisory Group has indeedprovided very valuable input to the defi-nition of the VLTI programme (cf. Pa-resce et al., The Messenger No. 83),and further science cases at this stagewould have unnecessarily duplicatedthis effort. One emerging field of astron-omy, extra-solar planets, has also beenset aside as this area has been judgedso rapidly expanding and crucial for theVLT/VLTI that a special working groupwas created at the beginning of 1996.The reports of this working group arecomplementary to the SC.

While several useful indications forfurther improvements have been pro-duced by the SC exercise, it is reassur-ing to note that no major drawback hasemerged, while most SC can be servedadequately with the telescopes and in-struments as they are specified at thispoint. In a few cases with very specialrequirements (e.g. fast reaction time totargets of opportunity – possibly throughdirect outside access) the proposed ex-periment does not lie within the specifi-cations for the telescope. Such casesare studied further to explore the possi-bility of future implementation.

Almost all SC demonstrate that VLTscience will not be possible without thepreparatory and supporting work at oth-er telescopes. This is true for targetselection as well as target characterisa-tion. A direct response to this need is theESO Imaging Survey which is to becarried out at the NTT (cf. Renzini andda Costa article in this Messenger, seealso the ESO WEB site).

The extraordinary advantages of theconnection with HST has been amplydemonstrated by results combiningKeck and HST data. The presence of theECF at ESO/Garching offers the ESOcommunity a unique opportunity for pro-ductive combination of VLT and HSTdata, and several SC are directly built upon such combinations.

Perhaps not surprisingly, it has beenfound that most demands from the SChave been placed on the instrumenta-tion. This may be due to a self-imposedlimitation of many astronomers to cur-rently available telescope hardware.The VLT science cases offer the possi-bility to ask for more than just the pro-posed instrumentation and possibly re-fine the instrumentation programme. It isclearly understandable that most peoplerestrict themselves to the possible ratherthan develop dreams; however, it is alsoevident that only by asking for the ‘ideal’

experiment can we try to push hardwareto the limit and learn what to do next.

Many SC need massive multiplexingcapabilities for the VLT to be competitivein the corresponding scientific areas.Besides this, capabilities to be includedin the instrumentation programme high-lighted by the SC are adaptive opticssupported by artificial guide stars, inte-gral field spectroscopy in the optical andIR, high-speed CCD photometry, andcoronography. There are other parts ofthe instrumental parameter space whichare not covered by the VLT instrumenta-tion. However, no SC was designed todemand a capability which could not beserved within the current instrumenta-tion plan.

The Science CasesPseudo-proposals

The VLT Science Cases solicited bythe SPG were meant to address funda-mental problems in different areas ofastronomical research. By no means is itexpected that these SC cover all of themost relevant current astrophysical is-sues, but they are rather indicative of thetype of scientific operations the VLT willbe asked to perform. Each PI was askedto fill out a standard VLT Science Caseform to obtain relevant and detailedinformation about the VLT requirementsfor the proposed research. The form –nicknamed pseudo-proposal – was splitinto five main sections.

(1) The scientific rationale, also ad-dressing the possible impact of currentresearch before the VLT will start operat-ing; (2) a description of the proposedobservations, including an estimate ofthe total observing time required toachieve the scientific goal (preparatoryLa Silla observations were also men-tioned and quantified); (3) a detailed listof the technical requirements to accom-plish the scientific goal (e.g., pointing,tracking, image quality, throughput,etc.); (4) identification and a quantitativediscussion of the performance of theVLT and the instruments required toachieve the science goal and the associ-ated calibration requirements; and final-ly (5) proposers were asked to brieflyidentify the limits of first-generation in-struments for the specific SC, all the wayfrom simple items (e.g., the filter list) tothe whole instrumentation plan.

In a parallel development, severalROP were created in house. Their pur-pose was specifically to test the opera-tional concepts proposed for the VLT. Toprovide a semirealistic scenario, theyalso contain a scientific rationale appro-priate for an 8-m project, but the empha-sis was put on the technical aspects ofthe proposals. Since they were plannedfor realistic checks of the data flow andthe operational system, they assumedthe approved instrumentation and weredeveloped to cover as wide a parameterspace as possible. Operationally simple

observations are included as well ascomplicated and delicate projects whichrequire a very high level of planning andagility of the system, even the combina-tion of individual telescopes for the si-multaneous coverage of particularevents. There is some overlap betweenVLT science cases and the referenceproposals. We included the referenceproposals as they describe the opera-tional requirements in greater detail.

Each SC and ROP was analysed by amember of the SPG. The main observa-tional elements and their requirementswere addressed. All proposed observa-tions were assessed with a particularview on technological and operationalaspects. The results are now currentlyused in discussions on the scientificneeds of the VLT project, and to quanti-tatively investigate instrumental options,optimisation, and telescope/instrumentcombinations.

Discussion

All VLT science cases were brokendown to the level of individual types ofobservation and the relative instru-ment(s) (imaging, spectroscopy, optical,near-IR, high-resolution, low-resolution,etc.). We have concentrated the investi-gation on some specific topics, since weare interested on the impact of therealistic programmes on the observingmodes and the operations. All observa-tions were assessed for the image qual-ity requirements, the importance of tem-poral resolution, spectral resolution andcoverage, and typical flux levels to judgethe amount of observing time.

On the operations side we checkedfrom where the target catalogues for theproposal would be drawn. The VLT willwork at flux levels for which no wholesky survey is available. Preparation isthe key to the success of VLT obser-vations. Almost all SC are based onextensions of current data sets (e.g. theESO Imaging Survey) and only very fewprojects will build their catalogues di-rectly from VLT imaging data. The re-quirement for the logistic support of theVLT in terms of object catalogues andsupporting observations at other (small-er) telescopes is evident. The need tofind faint and/or rare but interesting ob-jects for study with the VLT in statistical-ly significant quantities has spawned theESO Imaging Survey, and many otherobservatories are gearing up for large-field-of-view instruments. The wide-fieldcapabilities planned at ESO will becomecrucial for the VLT. Multi-object spec-troscopy depends on very good posi-tional information which has to be pro-vided ahead of the VLT observations (toset up slitlet arrays in FORS, fiber posi-tioning for FUEGOS, or provide masksfor VMOS and NIRMOS). This informa-tion has to be available either fromobservations with smaller telescopes orimaging with the VLT itself.

23

There are many different factorswhich determine the detailed schedul-ing of an observation. All SC were as-sessed for their constraints on theschedule. Exceptional atmospheric con-ditions (e.g. seeing, low IR background)can be critical as well as the absolutetiming (co-ordination with other observ-atories), or the dependence on earlierobservations. Most importantly for VLToperations, such factors directly influ-ence whether a programme can be car-ried out by means of service observa-tions, or whether the presence of theastronomer at the telescope is required.This will still be possible in the ‘classi-cal’ observing mode. Note that theremight be programmes which require ex-cellent conditions, yet are delicateenough to demand the astronomer’sdirect interaction. Situations like thesewill have to be resolved and it was oneof the aims to learn how many suchprogrammes might emerge in a realisticschedule. Another aspect of importancefor the VLT operations is the data ratewhich will have to be handled at theobservatory. Estimates of typical datavolumes were derived for each SC. Arelated topic is the format in which dataare archived. For certain observations itwill be impossible to maintain all rawdata and some preliminary reductionprocedures will have to be applied (typi-cally IR data will have to be combined‘on the fly’). Data storage and handlingwill not pose a critical problem with thepossible exception of some special pro-grammes (speckle, high-speed photom-etry). The current data-flow schemesshould be able to cope with the amountof data delivered by the instruments.The SC provide ground examples fordecisions to be taken soon. A few SCrequire specialised observing tech-niques, e.g. drift scans offer flatfieldquality which may be essential for someapplications.

The available VLT science casesspan a wide range in project size. Thereare a few SC which, if implemented atthe proposed scale, would take up amajor fraction of the available time,while other programmes are estimatedto take only a few nights for completion.Interestingly, the majority of the projectswould request more than 100 hoursobserving time. Thus it should not be

expected that observing projects at theVLT will be of smaller scale or be com-pleted in less time than current pro-grammes at La Silla.

Many projects are suitable for serviceobserving, some completely depend onit to catch the excellent image qualityrequired for the experiment. The optionto make full use of co-ordinated obser-vations, however, is not explored yet.Only few projects try to combine variousobserving techniques for a more com-plete picture of the science object.

There are certainly limitations in thecurrent set of SC. There is an obviousbias towards observational cosmology.The large statistical samples of faintobjects required for this type of researchdrives the demand for high multiplexingfacilities at the large telescopes. Anoth-er often requested facility is outstandingimaging quality stable for very weakobjects over a large wide-field of view(weak gravitational lensing, statisticallensing).

A few science areas are missing com-pletely, e.g. there are no SC for thethermal IR or adaptive optics with IRwavefront sensor. No studies of theinterstellar medium or of gas in generalhave been provided. High-resolutionstudies of stars, mapping of stellar sur-faces (e.g. through Doppler mapping),stellar outflow, stellar environments, orthe whole area of star formation andpre-main-sequence evolution have notyet been considered. Other stellar top-ics, like initial mass functions, low-metal-licity stars, stars in nearby galaxies, willhave to be addressed in the future. Noscience case for the solar system (e.g.trans-Neptunian objects) is available.

Summary

The VLT Science Cases have openeda new channel of user interaction withESO. They were meant to have atwo-way effect; first to raise the aware-ness within the community by stimulat-ing European astronomers to thinkabout the forthcoming capabilities of theVLT 8-m telescopes and entice them toprepare for the exciting opportunitiesthey will provide. Besides this, they weremeant to set elements for a concretescientific platform for future develop-ments in the VLT project. They also were

meant to better identify what the astro-nomical community is expecting fromESO.

It is gratifying to see that there seemto be no major shortcomings in the VLTprogramme, and the instrumental re-sources for most planned observationswill become available in due time. Yetwe have identified improvements in theinstrumentation plan like the massivemultiplexing spectroscopy needs or thepossible fiber feed boosting the multi-plex capability of the high-resolution in-struments (e.g. UVES).

With the experience gained from thisfirst set of SC we now feel confident toask interested astronomers to submit ascience case to ESO. While we cannotguarantee that every pseudo-proposalcan be included into our list, we doencourage all ESO astronomers to thinkabout their plans for VLT observing,whether they submit a science case tous or not. Anybody interested in provid-ing a VLT Science Case should contactone of the authors of this article. We areparticularly interested in projects whichcover research areas not represented inthe current sample, and instrumentalcapabilities which are poorly represent-ed in the first set of SC. These includethe mid-infrared spectral range, near-infrared high resolution spectroscopy,and projects which require adaptive op-tics with and without artificial referencestar.

We believe it is timely to think aboutthe capabilities of the VLT now and askfor the necessary preparatory observa-tions with La Silla telescopes. The OPChas specifically set guidelines to devotesome of the available time to preparato-ry programmes (cf. Call for Proposals).The work on a Science Case may alsoprovide stimulus to check out the devel-opments in the VLT project (see theWWW VLT page http://www.eso.org/vlt/and the descriptions of the instruments).It certainly will prepare you for the occa-sions when the VLT will be ‘your’ tele-scope. On the other hand, soon realVLT proposals will take the place of SCpseudo-proposals, and some of thequestions in the pseudo-proposals maybe transferred to real ones, for ESO tokeep active an ongoing monitoring ofthe needs and expectations of the com-munity.

1. Introduction

During 1997 (July–November), ESOplans to carry out a wide-angle, multicol-our imaging survey using EMMI on theNTT, followed by a deeper, narrow-anglesurvey using SUSI2 and SOFI, early in

The ESO Imaging SurveyA. RENZINI and L.N. DA COSTA, ESO

1998. A summary of the expected char-acteristics of the survey is shown inTable 1. This project, hereafter referredto as the ESO Imaging Survey (EIS), ismeant to generate moderate-size statis-tical samples for a variety of astronomi-cal applications, ranging from candidate

objects at the outer edge of the SolarSystem all the way to galaxies andquasars at extremely high redshift. EISdata should provide a suitable databasefrom which targets can be drawn forobservations with the VLT, in its earlyphase of scientific operation (from the

12424

third quarter of 1998 to approximatelythe end of 2000). EIS has been con-ceived as a service to the ESO commu-nity, with all the data becoming publicimmediately after its completion. Herewe provide information that may helpastronomers in the ESO community plantheir initial VLT programmes, taking ad-vantage of this survey.

The main motivation for EIS was thegeneral recognition of the urgent need toprepare suitable target lists for the VLT,essential for it to soon play a leading rolein ground-based optical/IR astronomy.By the year 2000, the competition will befierce with many 8-m-class telescopesin operation. The main goal of EIS is,therefore, to foster the scientific produc-tivity of the early VLT, and maximise itsscientific impact. There was also theunderstanding that a co-ordinated effortwas required in order to optimise thepreparatory work at the La Silla tele-scopes as well as to guarantee thehomogeneity and quality of the data.The ambitiously tight timetable and thelarge volume of data to be processedand analysed also implied the need forextensive collaboration between ESOand the community. Finally, there wasthe perception that no individual teamwould be able to fully exploit a surveysimilar to EIS in a timely fashion, werethe data to remain proprietary, even ifonly for a limited period. Instead, withthis co-ordinated effort most of the short-term survey needs will be met in a moreefficient way, while the competition forVLT observing time will take place at theappropriate moment.

In order to insure the involvement ofthe community, a Working Group (WG)was created to design and supervise thesurvey. Following the recommendationof the OPC, the WG members wereselected on the basis of their expertise indifferent research areas that could mostdirectly benefit from the survey. The EISWorking Group has not concluded itsactivity with the submission of the EISproposal and its endorsement by theOPC. It will, instead, continue to overseeall EIS activities, from the final selectionof some of the survey parameters to thedistribution of the data.

With the parameters listed in Table 1,the survey is expected to include p 150candidate clusters with z > 0.6, up to 50candidate quasars at z > 3 brighter thanV = 21 and about 200 down to I = 22.5.These candidates will represent naturaltargets for follow-up work with ISAAC,FORS and UVES, as soon as they comeinto operation at the VLT. The deep,narrow-angle part of the survey, cover-ing 250 square arcminutes, should con-tain about 10,000 objects to I P 26, withat least 30% of them being expected tolie at z Lp 1, and at least 200 at z L 3.The 25 square arcminutes imaged alsoin the J band should lead to the identifi-cation of a similar number of galaxieswith 1 lp z lp 2.8. For more details and

updates please check the EIS web site“http://www.eso.org/vlt/eis/”.

One aspect worth emphasising is theexperimental nature of EIS, which intro-duces a novel approach to large sur-veys. With EIS both ESO and the ESOcommunity will gain additional experi-ence which will be valuable for thescientific planning and operation of fu-ture surveys, such as those that arelikely to take place using the ESO/MPIA2.2-m telescope with the new wide-fieldcamera scheduled for the second se-mester of 1998.

Finally, it is worth pointing out that EISis not meant to be the only preparatorywork for the VLT, nor to meet the long-term needs of the community. Other set-ups may be required for scientific goalsnot addressed by EIS. Other opticalsurveys are also likely to be needed inthe future, in support of VLT researchand of various space missions. The aimof EIS is just to bridge the gap betweennow and the early VLT era, focusing ontopics that are likely to be mainstream inthe time frame considered.

2. Scientific Rationale

The WG has agreed on the necessityof identifying a set of science driversused to optimise the survey. The studyof objects in the high-redshift universehas then been singled out as the fieldmost in need of a preparatory survey.Correspondingly, the main science driv-ers used to optimise the survey havebeen identified to be the searches for:(1) Distant Clusters of Galaxies; (2)Quasars; (3) High-Redshift Galaxies.With this selection, the WG has paidattention to include all classes ofhigh-redshift objects, so as to place allresearch teams on a condition of paritywhile fostering the overall productivity ofthe VLT. Although the WG felt that it wasessential to optimise the survey for alimited number of goals, this should notovershadow the fact that the survey willhave almost countless applications invirtually every field of astronomy. Inparticular, it will provide a unique stellardatabase for studies of Galactic struc-ture and stellar populations, by detectingseveral hundred very low metallicitystars, M dwarfs, white dwarfs and bysetting strong limits on the local densityof brown dwarfs (possibly detectingsome of them).

The urgency of the proposed surveycomes from the necessity of providingsuitable scientific targets to be observedby the VLT as early as the secondsemester of 1998. In designing the sur-vey, the following timetable for the avail-ability of VLT instruments was taken intoaccount: ISAAC (1998/Q3); FORS(1999/Q1); CONICA (1999/Q2); UVES(1999/Q3). Adopting this specific timeta-ble as a constraint proved to be ex-tremely useful in designing the size ofthe survey which requires a balance

between the need for enough targets togenerate statistical samples and theavailability of time to allow them to beobserved with the early VLT, between1998 (Q4) and 2000 (Q4).

While some of the targets will beimmediately used in VLT observationsas early as 1998, others will requirefurther observations at the NTT or the3.6-m to provide “clean” samples for theVLT (e.g. to confirm candidate high-redshift quasars). This lead time has tobe taken into account in order to insurethe immediate competitiveness of theVLT in some research areas in a waythat it is consistent with the existinginstrumentation plans for the VLT.

In order to achieve the main scientificgoals and to reconcile the need forwide-angle coverage and depth, the sur-vey will consist of two parts. Awide-angle survey to search for clustersand quasars, complemented by deep-pointed observations of a smaller areato search for high-redshift galaxies. Forall the scientific goals the parameters ofthe survey were set to generate samplesof about 200 targets each, the minimalsize suitable for statistical analysis.

3. Background

In the summer of 1995, a small Panelwas set up at the ESO headquarters toinvestigate and make recommendationsabout the scientific needs and the tech-nical requirements for wide-field imagingcapabilities in support of the VLT. ThePanel was composed by ESO staff,visiting scientists and scientists in theGarching area, including: T. Broadhurst.S. Cristiani, L. da Costa, R. Gilmozzi,B. Leibundgut, R. Mendez, G. Monnet,A. Renzini (chair), P. Schneider andJ. Villumsen. The Panel first identified aset of topics that required extensiveimaging observations either in prepara-tion for either VLT or stand-alone pro-grammes. Among the topics consideredwere: search for primeval galaxies; high-redshift QSOs; distant clusters; low-sur-face-brightness galaxies; weak gravita-tional lensing by galaxies, clusters, andlarge-scale structures; galaxy inventoryof low redshift clusters; search for high-redshift supernovae; search and studyof individual objects in nearby galaxies(globular clusters, planetary nebulae,massive stars, HII regions, etc.); stellar-population studies in the Milky Way, theMagellanic Clouds and the dwarf sphe-roidal galaxies; and finally the search forexo-planets by gravitational microlens-ing. The Panel reviewed all these sci-ence cases thoroughly, as well as all thepossible options for wide-field imaging.The Panel concluded that to ensure acompetitive use of the VLT in severalresearch areas required both a verywide field imager at a 2–4-m-class tele-scope, and a moderately-wide field im-ager at the VLT.

Since then, several positive develop-

25

Table 1. EIS General Characteristics

Science Driver Area tint Filters Hours mlim(5σ)

Wide-Angle Survey

Distant clusters 18 sq. degree 300s V 96 24.7— 300s I 96 23.8

Quasars — 150s B 48 24.510 sq. degree 50s Gun-z 16 22.3

Narrow-Angle Survey

High-z galaxies 250 sq. arcmin. UGrIK 104 26, K = 21.525 sq. arcmin. J 24 24

Total 384

ments have taken place thanks to thecontinuous effort of ESO as well as theinitiative of individual institutes in theESO Community. The Max-Planck-Insti-tut für Astronomie (Heidelberg, MPIA)has offered financial and technical sup-port to equip the ESO-MPIA 2.2-mtelescope with a wide-field camera(35′ × 35′), while the Osservatorio Astro-nomico di Capodimonte (Napoli) hasoffered financial and technical supportfor the development of the 8K × 8K CCDdetector for the same camera. While thisinitiative is now well underway, the Os-servatorio di Capodimonte has also of-fered to ESO a 2.5-m telescope opti-mised for wide-field imaging, a proposalnow under evaluation at ESO. In addi-tion, the STC has recently approved theconstruction of two new, highly-competi-tive instruments for the VLT, the opticaland near-infrared multiobject spectro-graphs (VMOS and NIRMOS, respec-tively), each with imaging cameras witha field of view 14′ × 14′.

Despite these important develop-ments, none of these facilities will beavailable before the VLT is offered to thecommunity. In order to cope with theshort-term needs, a group of ESO as-tronomers was set up (L. da Costa(chair), A. Baker, J. Beletic, D. Clem-ents, S. Coté, W. Freudling, E. Huizinga,R. Mendez, and J. Ronnback) to investi-gate other alternatives. During severalmonths, this group investigated the pos-sibility of conducting an imaging surveyat the NTT during the second semesterof 1997 (about one year before thebeginning of the VLT scientific opera-tions) and making the data available tothe ESO community. On 12 May 1996,the Group released its final documentdescribing the conclusions of this feasi-bility study, which included a carefulanalysis of the science drivers of thesurvey and its technical, observational,and operational aspects. The documentalso included contributions on specifictopics by H. Böringer, S. Cristiani, P.Schneider, and J. Villumsen.

On 31 May 1996, the ESO DirectorGeneral and the VLT Programme Scien-

tist illustrated to the OPC the benefitsthat an imaging survey would have forthe early VLT observations. The OPCfully endorsed the notion, and recom-mended the formation of a WorkingGroup to elaborate a formal proposal forthe imaging survey to be discussed atthe following meeting of the OPC on29 November 1996. Being recognisedas an integral part of the VLT Pro-gramme, the VLT Programme Scientistwas asked to organise and chair theWorking Group. On 10/11 July 1996, theWG had its first meeting in Garching,with the presence of G. Chincarini, S.Cristiani, J. Krautter, K. Kuijken, Y. Melli-er, D. Mera, H. Röttgering and P. Schnei-der. From ESO, besides the authors ofthe present article, several other astron-omers attended the meeting, includingJ. Bergeron, S. D’Odorico, W. Freudlingand R. Gilmozzi. As the starting point ofthe discussions, the WG adopted the daCosta et al. document, complementedby additional input from S. D’Odorico onthe deep observations. The WG soonendorsed the concept, goals, and gener-al design of the proposed survey. Thiswas followed by a very productive dis-cussion about the various trade-offs,such as depth versus sky coverage,length versus width and filter selection,among others. As a result of the discus-sion, important improvements weremade to the original design, and theESO astronomers were requested toprepare a revised and more conciseproposal to be circulated in advance ofthe next meeting.

The draft of the EIS proposal wasdistributed to the WG members in Sep-tember 1996, and on 4 October 1996 theWG met again, this time with the pres-ence of S. Charlot and R. Saglia, andthoroughly discussed the proposal.While the wide-angle part of the EISproposal was unanimously endorsed byall the WG members, a few among themexpressed some reservation concerningthe pointed observations. Although fullyacknowledging the outstanding scientificvalue of this part of the proposal, someargued that individual teams from the

community could have submitted a reg-ular proposal on their own to conductthe corresponding observations, whilemaintaining proprietary data rights forsome time. The WG finally endorsedboth parts of the survey, and issued aset of recommendations to the OPC( “ h t t p : / / w w w . e s o . o r g / v l t / e i s /eis_wgr.html”).

An additional and independent en-dorsement to EIS came from the ESOVLT Key Programme Working Group inExtragalactic Astronomy that stated inits final report: “The working groupstrongly endorses the OPC policy toreserve substantial amounts of NTTtime for programmes that are preparato-ry to VLT programmes. In particular, theworking group expresses very strongsupport for the important initiative of theESO Imaging Survey on the NTT as avital preparation for the first few years ofthe VLT”.

In the letter to the OPC accompanyingthe EIS proposal and the WG recom-mendations, the VLT Programme Scien-tist stressed the following additionalpoint:

“The ESO scientists appearing as PIand Co-I of the EIS proposal committhemselves to avoid any personal use ofthe survey data before they becomepublic. If appropriate, the time limit ofthis commitment can be extended at thediscretion of the OPC. A similar commit-ment will be asked to all those membersof the community that may come to ESOto handle the data.”

The EIS proposal was formally sub-mitted to the OPC on 31 October 1996as planned, with the VLT ProgrammeScientist as PI, and L. da Costa,W. Freudling, S. D’Odorico, P. Quinn,J. Spyromilio, and J. Beletic as Co-Is.Since the proposal was not intended toprovide data rights to the team, the teamcomposition reflected exclusively thefunctional aspects of the Survey. In prac-tice, a much larger number of peoplehas been and will be directly or indirectlyinvolved in the EIS project.

On 29 November 1996, the OPCreviewed the EIS proposal and as ofFebruary 1997, the wide-angle sectionof the survey has been scheduled forperiod 59. Funds have been allocated tothe project and a visitors programmehas been established to sponsor short-and long-term visits of scientists fromthe community that can contribute withspecific skills to the preparation andexecution of the survey. The survey hasalso received high priority within all otherESO divisions, thus securing the neces-sary support.

4. Recent Developments andFuture Work

In early February 1997, it becameapparent that a FIERA controller couldnot be implemented on EMMI in time for

12626

the survey to use this facility. This hasled to the decision to abandon thedrift-scan mode of observations, thatwas originally recommended by the WG,in favour of the shift-stare mode with theEMMI-ACE system. Using recent esti-mates of the readout time, we expect anoverhead of about 100 seconds perexposure which will imply an efficiencyof p 60% relative to that expected forthe drift-scan. This has already beentaken into account in Table 1. The impli-cation is a reduction of the area coveredduring the Chilean winter-spring period,unless more time is allocated to theproject. On the positive side, the EMMI-ACE system will be fully tested by thetime EIS observations get underway.The system is also VLT-compliant, andas so EIS will serve as a prototype forthe service-mode observations and thedevelopment of the archive researchenvironment. Another advantage of theshift-stare mode is that it gives moreflexibility in the choice of the surveyregions.

Given these changes, the WG is cur-rently evaluating the possibility of stillcovering a wide range of right ascen-sions, thus allowing for an even distribu-tion of VLT targets. For example, thesurvey may be conducted in 3 patchesof about 6 square degrees each, sepa-rated by 3h and spread over the rightascension range 21h to 3h. The exactpointings and the format of each patch(e.g., 1° × 6° or 0.5° × 12°) are still underconsideration. The EIS home page willbe updated as soon as the WG makesits final recommendation.

Since receiving the endorsement ofthe OPC, the EIS team has initiated itswork on several fronts:

• Preparation and distribution of theAnnouncement of Opportunities for theEIS visitor programme.

• The preparation of the EIS homepage (“http://www.eso.org/vlt/eis/”) which

so far has been accessed over 1000times.

• Orders have been placed for a set ofwide-band filters and computer hard-ware.

• ESO fellows are already at work in avariety of tasks related to the survey.

• Interviews are being conducted toselect the first group of visitors who willwork on the development of the data-processing pipeline.

• The general plan for EIS is beingelaborated in collaboration with otherESO Divisions to establish the neces-sary interfaces between EIS and theUser Support Group, the NTT team, theData Management Division and theECF.

Some of the main tasks of the EISteam are:

• Selection of the survey regions to bepresented to the WG for final approval.

• Phase II Proposal Preparation(P2PP) and production of the Observa-tional Blocks.

• Implementation of data-processingpipeline.

• Implementation of EIS database,data archive, and procedures for datadistribution.

• Development of algorithms for theproduction of catalogues of candidateclusters, quasars, high-redshift galaxiesand galactic objects.

The EIS team is also following closelythe progress of SUSI2 and SOFI, withwhich the deep observations will beconducted.

Given the large amount of work aheadand the ambitious timetable of the sur-vey the involvement of the community inEIS is an essential ingredient for itssuccess.

5. Participation of the Community

The involvement of the ESO commu-nity is essential for two reasons. First, itis appropriate that this “service to thecommunity” is carried out with the directcontrol and participation of the commu-nity itself. Second, the tight timetable ofthe survey and the fact that the ESOstaff already have other commitmentsmake it necessary to rely on the partici-pation of scientists from the community.ESO is committed to provide the neces-sary support to make this participationpossible.

The involvement of the community isalready present through the EIS Work-ing Group. However, to expand thisparticipation, we urge interested scien-tists to interact either directly with theEIS team at ESO or with the members ofthe EIS Working Group which will re-main active until the completion of thesurvey.

ESO also envisions a more activeparticipation of the community via thedirect involvement of scientists andstudents in the observations, data re-duction and analysis. For this purposea special (“http://www.eso.org/adm/personnel/ann/eis.html”) visitors pro-gramme has been established to spon-sor short- and long-term visits of scien-tists/programmers/data assistants will-ing to contribute to EIS. As EIS willoperate in a very tight schedule, thepriority will be determined by the mostpressing needs of the survey which willevolve in time.

So far we have received over 70applications for the visitors programmefrom all over Europe and abroad. Weare currently selecting visitors who willhelp us implement the data reductionpipeline.

Alvio Renzini; e-mail: arenzini@eso.orgLuiz N. da Costa; e-mail: ldacosta@eso.org

27

R E P O R T S F R O M O B S E R V E R S

The Deep Near-Infrared Southern Sky Survey(DENIS)N. EPCHTEIN, B. DE BATZ, L. CAPOANI, L. CHEVALLIER, E. COPET, P. FOUQUÉ,

F. LACOMBE, T. LE BERTRE, S. PAU, D. ROUAN, S. RUPHY, G. SIMON, D. TIPHÈNE,Paris Observatory, France

W.B. BURTON, E. BERTIN, E. DEUL, H. HABING, Leiden Observatory, NetherlandsJ. BORSENBERGER, M. DENNEFELD, F. GUGLIELMO, C. LOUP, G. MAMON, Y. NG,

A. OMONT, L. PROVOST, J.-C. RENAULT, F. TANGUY, Institut d’Astrophysique de Paris,France

S. KIMESWENGER and C. KIENEL, University of Innsbruck, AustriaF. GARZON, Instituto de Astrofísica de Canarias, SpainP. PERSI and M. FERRARI-TONIOLO, Istituto di Astrofisica Spaziale, Frascati, ItalyA. ROBIN, Besançon Observatory, FranceG. PATUREL and I. VAUGLIN, Lyons Observatory, FranceT. FORVEILLE and X. DELFOSSE, Grenoble Observatory, FranceJ. HRON and M. SCHULTHEIS, Vienna Observatory, AustriaI. APPENZELLER AND S. WAGNER, Landessternwarte, Heidelberg, GermanyL. BALAZS and A. HOLL, Konkoly Observatory, Budapest, HungaryJ. LÉPINE, P. BOSCOLO, E. PICAZZIO, University of São Paulo, BrazilP.-A. DUC, European Southern Observatory, Garching, GermanyM.-O. MENNESSIER, University of Montpellier, France

1. The DENIS Project

Since the middle of 1994, the ESO1-metre telescope has been dedicatedon a full-time basis to a long-term projectwhose main objective is to survey theentire southern sky in 3 bands of thenear-infrared regime, namely I, J, andKs . The routine observations of theDENIS project began in December1995, after a long period of tests of theinstruments and telescope optics, ob-servational pilot programmes, and anunexpected delay due to the implemen-tation of the new read-out electronics ofthe I channel.

Extensive catalogues of stars, galax-ies, and other celestial objects consti-tute a fundamental tool of astronomy.Statistics of any class of objects re-quires data as complete and as well-calibrated as possible. Before makingany deep investigation of a particularregion, a large-scale map of the sur-rounding area is essential to establishthe general context. For these simplereasons, the sky has been surveyed atalmost all wavelengths accessible fromthe Earth’s surface and for which theappropriate technology is available; andsoon after man had escaped theEarth’s atmosphere, or discovered newways of detecting radiation thus unveil-ing new spectral ranges, explorationsquickly commenced in quest of new

celestial objects and unknown physicalprocesses.

The 2.2-micron window is of particularastrophysical interest. It is the longestwavelength window not much hamperedby extinction in the Earth’s atmosphere,and is also a band quite free of back-ground thermal emission. The interstel-lar medium is quite transparent at 2.2microns. The astrophysical importanceof the band lies in the fact that thiswavelength corresponds to the maxi-mum in the emission spectrum ofevolved stars.

At the end of the 1980’s, it had be-come clear that an all-sky survey wastechnically possible which would yieldan improvement by more than 4 ordersof magnitude in the K band, at 2.2microns, over what had been achievedby the extraordinarily valuable Two-Mi-cron Sky Surveys (TMSS) (Neugebauerand Leighton, 1969). Obviously, crucialwas the release from military classifica-tion of infrared detectors suitable forastronomical work; also important werethe advances in computer technologywhich would allow the enormousamount of data to be processed anddistributed in a feasible manner.

Two major efforts to map the infraredsky began more or less simultaneously.In Europe, several laboratories, underthe leadership of the Paris Observatoryin Meudon, started to consider pooling

efforts which led to a proposal for theDENIS project, aimed at covering theentire southern sky from the ESO site atLa Silla, making full-time use of the ESO1-metre telescope. In the US, followingthe initiatives of the University of Massa-chusetts, the 2MASS project was putforward, which proposed to use twocustom-built telescopes to observe bothhemispheres of the sky. Operating withsomewhat different wavelength bands,and in particular without the CCD-cam-era coverage of the I band, the 2MASSproject is in some ways complementaryand in other ways competitive with theDENIS survey of the southern sky.

Although not originally designed forinfrared observations, the ESO 1-metretelescope has been used successfullyover a two-decade period to carry outphotometric observations from 1 to 20microns. In fact, many of the most well-known “infrared objects” populating thesouthern sky, including very young mas-sive stars, extreme AGB stars, activegalactic nuclei, and actively star-formingsystems, were discovered with this in-strument.

The DENIS efforts started in earnestin 1990 to generate the necessary fundsand manpower. The French EducationMinistry provided the initial fundingwhich supported the preliminary feasibil-ity studies. The Département de Recher-che Spatiale of Paris Observatory at

12828

Meudon had established expertise inbuilding IR cameras during the CIRCUSand ISOCAM short-wavelength projects,and this allowed a prompt design of thetwo DENIS infrared cameras. At thesame time, several European laborato-ries pooled their manpower resourcesand technical expertise to propose aproject which was first granted fundingwithin the context of the Science plan ofthe European Commission, then subse-quently as a network in the HumanCapital and Mobility programme.

The ambition of the DENIS project is toproduce the first directly-digitised astro-nomical sky survey, and the largest infra-red star catalogue ever produced. It is along-range project that will extend overthe full decade of the 1990’s, from theinitiation of the design studies to the finalexploitation of the data. The basic DENISobjective is to explore the entire sky atδ < +2° in three spectral bands of thenear-infrared, namely the I, J, and Ks ,bands centred at 0.85, 1.25, and2.15 µm, respectively. DENIS will thusproduce and deliver to the community aset of fundamental astronomical docu-ments, including a point-source cata-logue of almost 1 billion stars in the Iband, an extended-source cataloguecontaining some 250,000 galaxies, andan atlas of 1 million elementary imagesof 12′ × 12′ size, in three colours. Thismaterial, which will amount to some 4Terabytes of data, will be of direct valueand will also support statistical investiga-tions (star and galaxy counts) in a newlyexplored spectral range; it will also sup-port the preparation and subsequentexploration of the planned and on-goingspace missions such as the InfraredSpace Observatory (ISO) mission of theEuropean Space Agency, the Near-Infra-red Camera and Multi-Object Spectrom-eter (NICMOS) project on the HubbleSpace Telescope, and the Space Infra-red Telescope Facility (SIRTF), as wellas ground-based projects including inparticular some of those to be carried outwith the ESO Very Large Telescope.

The DENIS material will support arange of investigations in several im-portant astronomical fields, includingthose dealing with the structure of ourgalaxy, specifically with the distribution ofvarious stellar populations, the in-vestigation of the missing mass in ourgalaxy, the formation and evolution ofstars, and the local structure of theUniverse.

The DENIS project incorporatesstate-of-the-art technologies in severaldifferent areas, including IR-array de-tectors, electronics, real-time comput-ing, and data-base management, eachof which require a different sort of ex-pertise. For this reason, the design andimplementation of the DENIS projecthas required the pooling of fundingand human resources in a number ofdifferent European institutes. No lessthan 20 institutes, located in 8 coun-

Figure 1: Sketch of the DENIS focal instrument (from L. Capoani).

tries1, are involved in the DENISproject, together with the EuropeanSouthern Observatory (ESO). The tech-nological requirements and the realitiesof the funding problems have led to asomewhat cumbersome organisationalstructure, but has had the advantage ofstimulating new collaborations for thescientific exploitation of the data.

1 Contributing countries and institutes are Austria(Universities of Innsbruck and Vienna) Brazil (Uni-versity of São Paulo), France (Paris, Lyon, Grenobleand Besançon Observatories and Institutd’Astrophysique de Paris), Germany (Landesstern-warte, Heidelberg), Hungary (Konkoly Observatory,Budapest), Netherlands (Leiden Observatory), Italy(Istituto di Astrofisica Spaziale, Frascati) and Spain(Instituto Astrofísica de Canarias).

Figure 2: The DENIS focal instrument installed at the Cassegrain focus of the 1-metre telescopeat La Silla. The stainless steel dewar seen on the left of the main structure contains the CCDarray. The infrared NICMOS arrays are set up in 2 cryostats, one of which (K band) is seen onthe right of the picture (blue vertical cylinder), the other is hidden by the structure. Part of theread-out electronics is hanged on the fork of the telescope.

2. The DENIS Instrument and theObserving Strategy

The DENIS focal instrument is a com-pletely new and specially designed 3-channel camera including its dedicateddata-handling system (Fig. 1, Fig. 2). Itconsists of 3 independent cameras at-tached to a main structure that is set upat the Cassegrain focus of the 1-metretelescope. The I-band camera isequipped with a Tektronix array of 1024× 1024 pixels; the J- and Ks-band cam-eras are each outfitted with RockwellNICMOS-3 detector arrays of 256 × 256pixels. Installing such a substantial in-strument at the focus of a relatively small

29

troller or FIC) controls the hardware(including filter-wheel rotation on the IRchannels, the movement of the CCDshutter and microscanning mirror, etc.)as well as general housekeeping (in-cluding temperature probes, etc.) andthe sequence of observations. Finally,an HP 9720 workstation (NRTU) is usedas the observer interface; it managesthe survey strategy automatically, bychoosing the right strip to be observed ata given time, and displays all necessaryinformation regarding telescope controland the status of the cameras andprogress of the survey. This workstationis also interfaced with the HP 1000 of theTelescope Control System (TCS) (seeFig. 3).

The focal instrument was designedand constructed under the leadership ofthe Département de Recherche Spa-tiale of Paris Observatory at Meudon,where D. Rouan, D. Tiphène, F. La-combe, B. de Batz, N. Epchtein, L.Capoani, S. Pau and E. Copet contrib-uted to the design and implementa-tion of the hardware and software, andin partnership with the Institut d’Astro-physique in Paris, where J.C. Renaultserved as project manager. The Istitutodi Astrofisica Spaziale in Frascati, Italy,the Instituto Astrofísica de Canarias inSpain, and the Observatories of Lyonand Haute-Provence in France, eachcontributed optical and mechanicalparts for the focal instrument. Thereal-time data-handling software wasimplemented by a collaboration be-tween F. Lacombe at the Meudon Ob-servatory, S. Kimeswenger at the Uni-versity of Innsbruck, and T. Forveille atthe Grenoble Observatory.

A preliminary version of the DENISinstrument, equipped only with the J andKs channels, was mounted on the ESOtelescope for the first time in December1993. The I-band CCD channel wasdelivered to La Silla in July 1995, andput in full operation soon after. Theperformance characteristics (see Ta-ble 1) of the cameras are in agreementwith the design specifications. During

Figure 3: The DENIS computer network (from F. Lacombe).

telescope was quite challenging; for ex-ample, due to weight limitations, themain instrumental structure had to bemade of magnesium. The f/15 beamcoming from the telescope is split intothree beams, corresponding to the threesurvey wavelengths, using dichroicbeam splitters. Three focal reducers,each optimised for the respective wave-lengths, transform the f/15 focal ratio tof/3. They provide a field of view of 12′ ×12′ on the sky for the three arrays and ascale of 3″ per pixel on the J and Kschannels and 1″ on the I channel. Amicroscanning mirror is inserted in theinfrared beams which can be tilted bypiezoelectric actuators in the right as-cension and declination directions insmall steps corresponding to fractions ofpixels.

The purpose of this microscanningmirror is to optimise the sampling of theJ and Ks images and to minimise theeffect of bad pixels on the NICMOSarrays. An elementary DENIS image inthe J and Ks bands results from combin-ing and interlacing 9 sub-images ob-tained by moving the microscanningmirror by 1/3 of a pixel in α and 21/3pixels in δ, in both negative and positivedirections. Because of this optimisedsampling of the DENIS images, each ofthe three channels is characterised bysimilar astrometric accuracy, of the orderof 1″.

The DENIS project will survey theentire sky between δ = +2° and –88°.The sky-coverage strategy involves di-viding the sky into three zones, eachcovering 30° of declination; each zone isitself divided into narrow strips, 12′ widein α and 30° long in δ. Each strip isscanned by moving the telescope in theso-called step and stare mode. The

integration time on each position of thetelescope is approximately 10 seconds;after this duration, the telescope ismoved 10′ in declination in order to allowfor an overlap of 2′ between two consec-utive images. One strip contains 180images of 12′ × 12′ and takes approxi-mately 1.5 hours to be completed. Thisincludes the time spent on photometricand astrometric calibrations and theoverhead delays required by the move-ment and stabilisation of the telescope,by the data read-out of the cameras, bydata transfer between computers, aswell as by the actions of shutters andfilter wheels. Flat-fielding observationsare performed in the three colours atdusk and at dawn.

The data handling at the telescope iscontrolled by three “Real-Time Units”(RTU), one per channel, each consistingof a pair of 68040 processors. A specialprocessor (the Focal Instrument Con-

Channel I J Ks

Central wavelength (µm) 0.80 1.25 2.15Array manufacturer Tektronix Rockwell RockwellNumber of pixels 1024 × 1024 256 × 256 256 × 256Array quantum efficiency 0.65 0.81 0.61Pixel size (µm, arcsec) 24, 1 40, 3 40, 3Number of bad pixels a few 216 254Largest defect (pixels) none 16 73Read-out noise (e–) 6.7 38 39Read-out time (second) 2.98 0.13 0.13Exposure time (second) 9 8.8 8.8Storage capacity (e–) 1.5105 2.0105 2.0105

Achieved limiting magnitude(point source 5σ) 18.0 16.0 13.5Magnitude of saturation 10.3 8.0 6.5

TABLE 1: Characteristics of the DENIS Channels.

13030

3. The Off-Site Data-AnalysisCentres in Leiden and in Paris

The data processing and archiving isperformed as a joint task between theParis Data Analysis Centre (PDAC), op-erating out of the Institut d’Astrophy-sique and the Paris Observatory underthe supervision of J. Borsenberger, andthe Leiden Data Analysis Centre (LDAC)operating at the Leiden Observatory un-der the supervision of E. Deul. The datapipeline is schematised in Figure 4.

The PDAC is responsible for archiv-ing and preprocessing the raw data inorder to provide a homogeneous set ofimages suitable for the subsequentdata-analysis streams in both Leidenand Paris. The LDAC is extracting ob-jects, ranging from point sources tosmall extended sources, and then pa-rameterising them and entering the re-sults into an archive constituting a pre-liminary point source catalogue (PPSC).The PDAC also extracts and archives

the commissioning phase, hardwaresubsystems and software were testedand improved in order to optimise theimage quality, the data reduction proce-dures, and the survey strategy. Becauseno major change in the instrumentalconfiguration nor in the survey strategywill be possible during the lifetime of thesurvey, in order to ensure homogeneityof the data, all the parameters thatdefine this strategy were carefully re-viewed during the commissioning phaseof the project.

The series of pilot-programme obser-vations (protosurvey) of selected areasof the sky chosen for specific scientificreasons, has been fully completed andhas resulted in a sky coverage of about2000 square degrees, in either two orthree colours, depending on whether thepilot programme in question was com-pleted before or after the CCD channelhad been installed. Scientific analysis ofthese data is now underway (see sec-tion 5).

Figure 4: The DENIS data pipeline. Users will access the image archives (PROFDA) that will be installed on a jukebox of DLTs at PDAC, and thesource catalogues in the SODA and LODA databases. The FOURBI database already provides the consortium members with information on thecurrent status of observations (observed areas). The connections with the CDS in Strasbourg and ROE are under discussion.

images for those sources found to beextended, and will thus create a cata-logue of galaxies. Both DAC’s are work-ing in close collaboration, supplement-ing each other, to perform a coherentand complete data reduction analysistask.

3.1 The Paris Data Analysis Centre(PDAC)

The PDAC performs a number ofsteps necessary to prepare the data forfurther processing and analysis. Thesteps involve the flat-fielding and sky-emission corrections of the strip imagesusing flat-fields derived during thereal-time processing on the mountainand through sunrise/sunset observa-tions. Subsequently, a sorting of theframes is done to obtain colour-groupedsets of data for each pointing position. AtPDAC, the incoming data stream isinspected thoroughly to assess the qual-ity of each and every image of the strips.

31

been compiled from those in use atCTIO, ESO, MSSSO, SAAO, andUKIRT. At LDAC, a working databasestores all the raw and derived extractionparameters along with the most recentcalibration parameters. At regular inter-vals, the raw information is transformedinto astronomically meaningful parame-ters using the best available calibration.

3.3 Quality estimates and finaldatabases

The photometric accuracy of theDENIS data can be assessed using thephotometric standard stars as a localreference frame. The standard deviationin the photometric parameter derivedfrom the protosurvey data shows thathigh accuracy is achieved over a largerange in magnitudes. The photometricaccuracy of the object extraction (rmserror) has been shown to be about 0.03mag, just below the saturation limit, and0.02 mag at the faint cut-off. The DENISinstrument has so far been very stable intime. As a result of the large amount ofdata produced by DENIS, we will defineour own photometric system and pro-vide transformation formulae to mostother standard photometric systems.Our system, however, will increase inaccuracy with time and can only befinalised at the end of the survey.

The positional accuracy is influencedmainly by the systematic errors of theGSC, which on average are about 1″. Infact, the DENIS input catalogue usedis a combination of several catalogues,the GSC v1.2, the PPM catalogue, theHIPPARCOS input catalogue and even-tually also the Tycho catalogue, eachappropriately weighted in the astromet-ric solution. The relative rms errors(within a DENIS frame) are of the orderof 0.1″. After the TYCHO cataloguebecomes available (we expect oneTYCHO star per DENIS frame), thepositional errors can be reduced to a fewtenths of arcseconds. Current positionalerrors are about 0.3″. Once the datahave passed the quality tests, they willbe gathered into final databases, onecontaining the small sources (SODA)the other the extended sources (LODA).These databases will result from acollaboration between both DACs andthe CDS in Strasbourg. Cross-identifica-tions with the optical SUPERCOSMOSdata are also envisaged to eventuallyprovide 5-colour photometric cata-logues.

4. DENIS Operations in Chile

The survey operations in Chile in-volve a major effort, requiring the pres-ence of at least one person every night,all the year round, at the observing sitefor an expected total duration of P 4.5years. The project maintains a dedicatedteam, resident in Chile and now fullyoperational. The team involves a survey

Figure 5: DENIS image of the molecular cloud OMC2 as seen in J and Ks (false colours, blue =J, red = Ks, processed by E. Copet).

This information is relayed back to LaSilla to allow re-observations when nec-essary. This procedure ensures a homo-geneous survey.

The processed frames (strips) arestored in the Processed Frame Data-base (PROFDA) and also sent to theLDAC for further processing. ThePROFDA will allow mosaicking the in-dividual images. Access to the PROFDAis provided through a database pro-duced and maintained at the PDAC.This database allows verification of theobserving strategy goals aimed at fullmapping of the southern sky, and alsokeeps track of the data quality andstorage administration.

Upon return of source and calibrationinformation from LDAC, the PDAC ex-amines the processed frames for ex-tended objects. These non-stellar ob-jects are parametrised and archived inthe extended-source database. Foreach extended object a postage stampimage is extracted and stored.

3.2 The Leiden Data AnalysisCentre (LDAC)

The LDAC concentrates on the ex-traction of “small objects” from theDENIS images, at all three wavelengthbands. Production of a catalogue isdone during the course of the surveydata acquisition and should result in an

incremental first-order data product trac-ing the observations with a delay periodof approximately one year. Access to the“protosurvey” data is open to the DENISScience Working Groups.

The derivation of object parameters isbased solely on image properties; noa priori astronomical information willenter this catalogue. Object parametersare corrected for instrumental effectsin the final catalogue. Apart from theusual extraction parameters, informationabout the local image environment andextraction characteristics are also re-tained. All objects above the 5-σ noiselevel are extracted. The objects arede-blended and their positions calculat-ed using the frame centres and plateparameters computed with the help ofcross-identifications with the DENIS In-put Catalogue sources. In order to im-prove the positional accuracy, strip wideastrometric solutions are computed uti-lising the overlap information inherent tothe observing strategy. The small ob-jects are photometrically calibrated us-ing a set of DENIS standards widelydistributed over the southern sky. Thelist of standard stars used for DENIS hasbeen divided into two parts, one for the Iband and the other for the J and Kbands. The I-band standards come fromthe set of UBVRI standards used at ESOand at CTIO (Landolt, 1992). For the Jand K bands, a list of standards has

13232

astronomer, P. Fouqué, who supervisesthe observations; an operations engi-neer, formerly B. Jansson; and a youngscientist, generally a French “coopé-rant”, E. Bertin for the 1994–95 year,P.-A. Duc for the 1995–96 year, andpresently R. Moliton. In addition, visitingastronomers and engineers come fromEurope to reinforce the local staff, tomake urgent hardware or software mod-ifications, and to provide the necessarylink between the operations in Chile andthe data analysis centres in Paris andLeiden.

The mountain operations are con-ducted under the responsibility of thesurvey astronomer. All commissioningtests have been done under his super-vision in close interaction, via dailye-mail connections, with the instrumentteam based in Meudon and with the twodata analysis centres. The primary re-sponsibilities of the operations engineerinvolve maintenance of all aspects of theinstrumentation, mainly the electronicsand hardware devices; as with any newand complex instrument, maintenanceof the complete DENIS facility is animportant and demanding task.

5. First Scientific Results fromthe DENIS Project

5.1 Organisation of the scientificdata analysis

The scientific interpretation of the dataobtained during the pilot-survey periodhas now started, under the responsibilityof five dedicated Science WorkingGroups set up for this purpose. Theyinvolve astronomers from the differentinstitutes of the consortium, including anumber of graduate students and post-doctoral fellows, several of whom aresupported by the European Communitythrough a network of the Human Capitaland Mobility programme, as well asthrough other individual and institutionalfellowships. The five working groups arededicated to the principal areas to whichDENIS is expected to contribute: (i)late-type and AGB stars, chaired by H.Habing, Leiden; (ii) low-mass stars,chaired by T. Forveille, Grenoble; (iii)galactic structure, chaired by A. Robin,Besançon; (iv) star formation, chaired byP. Persi, Frascati; and (v) extragalacticsystems and cosmology, chaired by G.Mamon, IAP, Paris.

A data-release policy group (DRPG)defines and regulates the access to thedata and the authorship of the publica-tions. The main rule is in accordance withESO policy, that once the data have beenprocessed and certified regarding quali-ty, they are put at the disposal of theconsortium members for the duration ofone year after archival in the DACs;subsequently the data are released tothe community at large. The first suchgeneral release of DENIS data is expect-ed in spring 1997. The DENIS consorti-

Figure 6: DENIS two-colour (J, Ks) image of the Tarantula Nebula in the Large Magellanic Cloud(false colours, blue =J, yellow = Ks, processed by E. Copet).

um has been actively pursuing its goals.Twice a year progress review meetingsare held, involving the entire consortiumand chaired by the PI. More frequently, inpractice at least five times a year, thedata processing specialists meet in DataAnalysis Meetings, or DAMs, chaired byE. Deul, in order to discuss and optimisethe data processing techniques andmethods. The DENIS project has beenthe focus of several national and interna-tional conferences. A Les HouchesSchool was held on the subject of Sci-ence with Astronomical Near-InfraredSurveys (Epchtein et al., 1994). TwoEuroconferences, supported by a grantfrom the EC, have been dedicated topresenting and discussing the first scien-tific results from DENIS. The first of thesewas organised in Italy (Persi et al., 1995),and the second in Spain (Garzon et al.,1997). Several scientists from 2MASSand IPAC have participated in theseworkshops, thus maintaining contactsand collaborations between the two in-frared survey projects.

A couple of DENIS images obtainedduring the now-routine period of obser-vations are shown in Figure 5 andFigure 6. These images illustrate thequality of the data and the results of thereduction pipeline; the images involvethe nominal integration times of 10 sec-onds, and each represents a field of12′ × 12′ area.

Several Ph.D. theses based on theexploitation of DENIS data acquiredduring the protosurvey period havebeen recently defended or are currentlyin progress. Eric Copet (Meudon,DESPA), the first graduate student in-volved in DENIS since its beginning, inaddition to contributing to the imple-mentation of parts of the real-time datahandling software, studied the luminosi-ty function of young stars in the Orionregion based on the early DENIS ob-servations. He investigated the natureof some 2000 objects detected in J andKs in an area covering 2° × 0.5° aroundthe Trapezium (Copet, 1996). Most ofthem are likely to be associated withthe Orion complex. Stephanie Ruphy(Meudon, DESPA), has performedlarge-scale DENIS star counts, mainlyin J and Ks, in various directions of theGalaxy, and compared them with thepredictions of currently available mod-els of stellar populations synthesis. Sheis particularly involved in a study of theanticentre direction. In these directions,all current existing models disagreewith DENIS star counts in the sensethat there are fewer stars observedthan the models predict. At low galacticlatitudes, colour diagrams are success-fully used to separate giant and dwarfstellar populations (Fig. 7) (Ruphy,1996). New revised values of the scalelength and of the cut-off distance of the

33

Figure 7: DENIS colour-colour diagram of P 3800 point sources detected in an area of one halfsquare degree at b = 1° and l = 238° at K < 13.5. This diagram shows two main concentrationsof sources, the blue one corresponds mainly to dwarf stars and the redder one mainly to giants.At low latitude, the interstellar extinction efficiently separates the two populations.

I -J

stellar distribution in the Galaxy havebeen derived (Ruphy et al., 1996).

Another study in progress involves asearch for low-mass stars at high galac-tic latitudes. Xavier Delfosse, at Greno-ble, is preparing a Ph.D. thesis basedon the analysis of DENIS images takenfar from the galactic disk. A first analysisof several tens of strips at high galacticlatitude resulted in the discovery of sev-eral very late (M8 to M10) red dwarfs.Other results have been already ob-tained by Y. Dabrowski who has com-bined DENIS images with optical datacoming from Schmidt plates digitisedwith the COSMOS machine (Fig. 8).Most of the DENIS source extractionswere done using the sextractor softwaredeveloped by Emmanuel Bertin.

Extensive investigations involvingthe combination of DENIS and ISO im-ages have already started. A. Omont(IAP, Paris) is leading a large-scaleISO open-time programme, entitled“ISOGAL” and aimed at probing thegalactic bulge populations at 7 and15 µm. Several areas located in therange +45° > l > –45° and b ≤ 0.5° arecurrently being observed by ISO and byDENIS almost simultaneously, giving awide spectral coverage and allowingmuch easier population separation incolour-colour diagrams (Perault et al.,1996). Other DENIS and ISO collabo-rations have begun, aimed e.g. at ex-ploring nearby star formation regions,and the evolved giant and supergiantpopulations of the Magellanic Clouds.P. Persi (IAS, Frascati) and collabora-tors are studying the star-formation ratein several nearby molecular clouds suchas the Chameleon clouds. L. Cambresy

DENIS colours of late-type giants (e.g.,O-rich vs. C-rich) and identifying newsuch stars in the Magellanic Clouds.

Gary Mamon (IAP, Paris) and hiscollaborators have made simulations forstar-galaxy separation and have esti-mated that the number of galaxies thatwill be detected by DENIS in the Ks,band would be around 250,000, while inthe I band, this number could be closeto 1 million. As an illustration of thecapability of detecting reddened galax-ies in the Zone of Avoidance, anedge-on galaxy, IRAS 07395-2224, asseen by DENIS in its three bands isshown in Figure 9. Although the galaxyis barely seen on Schmidt plates(SRCJ) and is still rather faint in I, itappears prominently in Ks , where thedensity of field stars is much lower. Along-range programme to detect highly-obscured galaxies in the Zone of Avoid-ance is planned in collaboration withRenée Kraan-Korteweg (Meudon). G.Paturel and I. Vauglin from Lyon Ob-servatory are currently cross-identifyingDENIS sources with the galaxies of theLyon-Meudon Extragalactic Data Base(LEDA) using the deep I images, andare drawing up a list of still uncata-logued galaxies detected by DENIS.Thousands of new galaxies have beenalready identified on the DENIS images.

Although DENIS has been operatingin its routine mode only for one year,covering about one quarter of the south-ern sky, the project has already resulted

is drawing out a high spatial resolutionand large-scale mapping of the extinc-tion towards this nearby dust and mo-lecular complex on the basis of starcounts and [I – K] reddening. C. Loupand collaborators are characterising the

Figure 8: [B-J] vs. [J-K] colour diagram obtained by cross-identifying hundreds of objectsdetected by DENIS with optical sources extracted from a Schmidt SRCJ plate with theCOSMOS machine at high galactic latitudes. This diagram illustrates the extremely promisingbreakthrough that the combination of optical and near-IR data at large scale will provide in thequest for very low mass stars. According to the recent models of VLM stars, (e.g. Allard andHauschildt), they should appear much bluer in [J-K] than the normal main-sequence dwarfs,and thus good candidates could be easily recognised in this diagram.

13434

Figure 9: The galaxy IRAS 07395-2224, located at b = 0.2° as seen in the 3 DENIS bands from left to right: I, J, Ks. Notice that this edge-on spiralappears prominently in the Ks band. DENIS is expected to pick up many such galaxies in the zone of avoidance.

in a number of promising results. A smallfraction of the archived data has beenexploited, so far. Several fruitful collabo-rations have started within the consor-tium institutes and with ISO teams. Ow-ing to the huge amount of data to ana-lyse, we expect that the interest in DENISproducts in the astronomical communityat large, and in particular within the ESOcommunity, will grow as the surveyprogresses; we also expect that the2-micron point source catalogue will beavailable, at least partly, for the first lightof the VLT.

Acknowledgements

The DENIS project is partly fundedby the European Commission throughSCIENCE and Human Capital and Mobil-ity plan grants. It is also supported inFrance by the Institut National des Scien-ces de l’Univers, the Education Ministryand the Centre National de la RechercheScientifique, in Germany by the State of

Baden-Württemberg, in Spain by theDGICYT, in Italy by the Consiglio Nazio-nale delle Ricerche, in Austria by theFonds zur Förderung der wissenschaft-lichen Forschung und Bundesministe-rium für Wissenschaft und Forschung, inBrazil by the Foundation for the develop-ment of Scientific Research of the Stateof São Paulo (FAPESP), and in Hungaryby an OTKA grant and an ESO C&EEgrant. The European Southern Obser-vatory is greatly contributing to the suc-cess of the project by providing a long-term access and maintenance of the1-metre telescope and free subsistencefor the DENIS operations team at LaSilla. We are also greatly indebted tothe graduate and postdoctoral studentswho have contributed to the observa-tions at La Silla, namely, L. Cambresy(Meudon), P. Le Sidaner (Paris), M. Lo-pez-Corredoia (IAC, Tenerife), P. Herau-deau (Lyon), M. J. Sartori (São Paulo),R. Alvarez (Montpellier), V. de Heijs(Leiden).

References

Copet E., 1996, Thèse de Doctorat, Paris 6.Epchtein N., Omont A., Burton W.B., Persi P.

(eds.), 1994, Science with AstronomicalNear-lnfrared Sky Surveys, Proc. Les Hou-ches School, Sept. 1993, Kluwer Ac. Pub.,(reprinted from Astrophys. Sp. Sci., 217).

Garzon F., Epchtein N., Burton W.B., Omont A.,Persi P. (eds.), 1996, Proc. of the 2nd Euro-conference on the impact of Near-InfraredSky Surveys, Puerto de la Cruz, Spain, April1996, Kluwer Ac. Pub., in press.

Landolt A.U., 1992, AJ 104, 340.Neugebauer G., Leighton R.B., 1969, Two Mi-

cron Sky Survey, NASA SP 3047.Perault M., Omont A., Simon G., et al., 1996, AA

Lett. 314, L165.Persi P., Burton W.B., Epchtein N., Omont A.

(eds.), 1995, Proc. of the 1st Euroconfer-ence on near-infrared sky surveys, San Min-iato, Italy, Jan. 1995, reprinted from Mem. S.A. It., 66.

Ruphy S., Robin A.C., Epchtein N., Copet E.,Bertin E., Fouqué P., Guglielmo F., 1996, AALett, 313, L21.

Ruphy S., 1996, Thèse de Doctorat, Paris 6.

What is the biggest star in the sky?Here we are not speaking of absolutesize in kilometres, but in apparent sizeas seen from the Earth. The obvious

answer is the Sun, which has an angulardiameter of half a degree. But which starcomes next? The answer has long beenassumed to be Betelgeuse (alpha Ori-

onis), which appears to be about 35,000times smaller than the Sun (dependingon the wavelength at which you ob-serve). In fact, Betelgeuse was the first

R Doradus: the Biggest Star in the SkyT.R. BEDDING, J. G. ROBERTSON and R.G. MARSON,

School of Physics, University of Sydney, AustraliaA.A. ZIJLSTRA and O. VON DER LÜHE, ESO

35

with an annular mask (see Bedding etal., 1993). The purpose of the mask is toreduce the effects of seeing on theaccuracy of the measurement, with thebeneficial side effect of attenuating theflux to a level which can be managed bythe detector. The moduli squared of theFourier transforms of 300 frames wereaveraged for each star. Calculating theratio of the mean spectra from R Dor andthe unresolved stars calibrates for in-strumental and atmospheric attenua-tion of the mean spectra. The result isthe two-dimensional visibility functionsquared of R Dor: it shows a decreasetowards larger angular frequencieswhich indicate that the source is re-solved. Figure 1 shows a sample cali-brated visibility function (top-left panel).A non-linear least-squares fit of a modelof the visibility function for a uniformlyilluminated disk, with the apparent diam-eter as the only parameter, is applied.The top-right panel shows the modelfitted to the data shown in the top leftpanel. The difference between observa-tions and model is seen in the lower-leftpanel: the curved features are instru-mental residuals from the spiders of thetelescope, the remainder being thenoise. For comparison, the bottom-rightpanel shows the visibility of one of thereference stars calibrated by the otherone.

The mean diameter of R Dor we ob-tain from these data is 0.0587 ± 0.0026(1 sigma) arcsec, uniform disk equiva-lent. This is the average of 11 model fits

star apart from the Sun to have itsangular diameter measured (Michelson& Pease, 1921).

Betelgeuse is a red supergiant and itslarge apparent size is due both to itsextremely large physical diameter(about 700 times that of the Sun) and toits relative closeness (about 200 pc).Several other nearby red giant and su-pergiant stars are resolvable with dif-fraction-limited 4-m-class telescopes.Not only can the angular diameter ofthese stars be measured directly (e.g.,Tuthill et al., 1994b), but in some cases(e.g., Betelgeuse) features on the stellarsurface have been detected (Buscher etal., 1990). In the case of some Mirastars, significant elongation has beenfound (Wilson et al., 1992, Haniff et al.,1992, Tuthill et al., 1994a).

Here we report observations of thestar R Doradus, which has not previous-ly been observed at high angular resolu-tion. R Dor is an M8 giant and is thebrightest, and presumably closest, starwith such a late spectral type. Based onits infrared brightness, Wing (1971) pre-dicted that R Dor (HR 1492) could belarger than Betelgeuse. We observedR Dor with SHARP on the NTT (infrared)using the technique of aperture mask-ing. Other results, not presented here,were obtained with MAPPIT on the AAT.We indeed find that R Dor exceedsBetelgeuse in angular size.

Sequences of short (0.1s) exposureswere taken of R Dor and two near-byunresolved stars, α and γ Ret (HD 27256

and HD 27304), using the SHARP cam-era on the NTT in the Johnson J band.The NTT entrance aperture was covered

Figure 1.

Figure 2.

13636

sured to be 44 mas (Dyck et al., 1992).Thus, R Dor appears to be the largeststar in the sky.

Despite its classification as semi-regular, R Dor is in many ways closer tothe Miras than to other SRb stars. Itsperiod is near the peak of the Miraperiod distribution function (250–350days), while SRb stars almost alwayshave much shorter periods (e.g.,Kerschbaum & Hron, 1992). Its latespectral type and large J–K are alsomore typical of Miras than other M-typestars (Feast, 1996). Using a bolometricmagnitude at the epoch of our measure-ment of –0.96, we derive an effectivetemperature of 2740 ± 190 K from themeasured diameter. Assuming thatR Dor is closely related to the Miravariables, as seems likely, we can applythe period-luminosity relations for Mirasin the LMC given by Feast 1996. Weobtain a distance of 61 ± 7 pc and aluminosity for R Dor of 6500 ± 400 LA.Our distance for R Dor agrees with esti-mates of 60 pc by Judge & Stencel,1991, and 51 pc by Celis, 1995. Thisdistance, together with our observedangular diameter, implies a stellar radiusof 370 ± 50 LA. From the pulsationequation Q = P (M/MA)1/2 (R/RA)–3/2 andassuming Q = 0.04 days (appropriate forfirst overtone pulsation), we derive amass of 0.7 ± 0.3 MA. All the derivedparameters are consistent with a classi-fication of this star as Mira-like, with theeffective temperature being slightlyhigher than the average for Miras.

All previous measurements of the ra-dii of Miras fall in the range 400–500 RA,which is taken by Haniff et al., 1995, asevidence that Miras are associated with

a well-defined instability strip. The factthat R Dor shows a more irregular pulsa-tion behaviour but with many character-istics of a Mira is consistent with it lyingnear the edge of such a strip. Miras havebeen found to show surface structuresand/or ellipticity: if R Dor is related to theMiras, it may be expected that it alsoshows these effects. MAPPIT/AAT ob-servations of R Dor indeed show non-zero closure phases, indicative of asym-metries or surface structure. R Dor isclearly an excellent candidate for moredetailed observations with the VLT andVLTI.

References

Bedding T.R., von der Lühe O., Zijlstra A.A.,Eckart A., Tacconi-Garman L.E., 1993, TheMessenger, 74, 2.

Buscher D.F., Haniff C.-A., Baldwin J.-E.,Warner P.J., 1990, MNRAS, 245, 7P.

Celis S.L., 1995, ApJS, 98, 701.Dyck H.M., Benson J.A., Ridgway S.T., Dixon

D.J., 1992, AJ, 104, 1982.Feast M.W., 1996, MNRAS, 278, 11.Haniff C.A., 1994, in Robertson J.G., Tango

W.J. (eds.), IAU Symposium 158: VeryHigh Angular Resolution Imaging, p. 317,Kluwer: Dordrecht.

Judge P.G., Stencel R.E., 1991, ApJ, 371,357.

Kerschbaum F., Hron J., 1992, A&A, 263, 97.Michelson Pease, 1921, ApJ, 53, 249.Tuthill P.G., Haniff C.A., Baldwin J.E., 1994a,

in Robertson J.G., Tango W.J. (eds.), IAUSymposium 158: Very High Angular Reso-lution Imaging, p. 395, Kluwer: Dordrecht.

Tuthill P.G., Haniff C.A., Baldwin J.E., FeastM.W., 1994b, MNRAS, 266, 745.

Wilson R.W., Baldwin J.E., Buscher D.F.,Warner P.J., 1992, MNRAS, 257, 369.

Wing R.F., 1971, PASP, 83, 301.

to the corrected mean power spectrawhich result from all but one combina-tions between 3 sets of observations ofR Dor and 4 sets of reference-star ob-servations. There is no significant indi-cation of a deviation from circular sym-metry of more than 2 per cent. Smallerdeviations are easily explained by resid-ual artifacts in the visibility data.

Figure 2 shows radial plots (averagesover the azimuth) of the compensatedspectra. The black, heavy dots are theazimuth averages of the R Dor spectra.The origin of the dip near 1 lp/arcsecmay be related to seeing variations; thisregion was excluded from the fittingprocess. The red curve shows the aver-age and the variation of the non-linearfits to the data. The green, open circlesshow azimuth averages of inter-calibrat-ed reference data to demonstrate thesignificance of the R Dor measure-ments.

The observed angular diameter of anM star can be strongly affected by theatmospheric extensions. Within bandsof high opacity (such as the TiO bands)the star can appear 50 per cent largerthan in continuum regions of the spec-trum. It is therefore important to com-pare diameters for different stars only formeasurements made in the continuum.A limb-darkening correction may alsohave to be applied, but this correction issomewhat ad-hoc due to the lack ofreliable model atmospheres. Opacity ef-fects and limb darkening are smallest inthe infrared and stellar diameter mea-surements should preferably be taken inthis region of the spectrum.

The uniform-disk diameter of Betel-geuse at 2.2 microns has been mea-

1Note that Figure 1 in van Langevelde et al.(1994) is labelled incorrectly. The correct image sizeis p 26″, rather than p 52″.

1. Introduction

Although T Tauri is the prototype of alarge class of pre-main-sequence ob-jects (the T Tauri stars), it is actuallyvery peculiar. First, the optical primary(T Tau N) has an infrared companion(T Tau S) at a separation of p 0.7″,which is completely obscured at visiblewavelengths, but may dominate the bo-lometric luminosity of the system (Dycket al., 1982, Ghez et al., 1991). Infraredcompanions have been found near anumber of other pre-main-sequencestars, but their presence seems to bethe exception rather than the rule (Zin-necker and Wilking, 1992). Second,

and even more surprisingly, strong ex-tended 2.121 µm H2 (v = 1 – 0) S(1) ro-vibrational emission was found fromT Tau (Beckwith et al., 1978). Despiteextensive searches, H2 emission withcomparable strength has been found invery few other pre-main-sequence stars(e.g. Carr, 1990). Finally, there are twoHerbig-Haro objects at right angles as-sociated with T Tau (Schwartz, 1975,Bührke et al., 1986), giving rise to thesuspicion that two misaligned pairs ofjets emanate from the two componentsof the binary.

Taking a closer look at the environ-ment of T Tau, and at the relation of thegas to the two stellar components re-

quires spectral line imaging with thehighest possible angular resolution. VanLangevelde et al. (1994) present animage of T Tau in the H2 (v = 1 – 0) S(1)line with a resolution of 0.8″, but thecentral region of this image had to beblanked because of saturation1. Morerecently, Herbst et al. (1996) observedT Tau with the MPE imaging spectrome-ter 3D; these observations cover a fieldof 8″ × 8″ with a resolution of p 0.7″. Acomplicated structure was detected onthis scale, which was interpreted as an

Molecular Hydrogen Towards T TauriObserved with Adaptive OpticsA. QUIRRENBACH, Max-Planck-Institut für Extraterrestrische Physik, GarchingH. ZINNECKER, Astrophysikalisches Institut, Potsdam

37

day in substantially poorer seeing (1.4″to 1.9″) were discarded.

The raw data were corrected for badpixels and flatfielded with exposures ofthe sky taken before sunset. (Becauseof strong wavelength-dependent fringescaused by interference in the detectorsubstrate and the circular variable filter(CVF), which is used as an order sorter,we took sky flats at exactly the samewavelengths that we used for the astro-nomical observations.) All images werethen resampled on a 0.05″ grid andrecentred, before the continuum frameswere subtracted from the line expo-sures. It turned out that this proceduredid not only remove the continuum lightfrom the stars, but it also resulted in anexcellent cancellation of the sky back-ground, because of the small wave-length difference between the line andcontinuum channels. The resulting im-ages were coadded to form the finalimage of the v = 1 – 0 S(1) transition ofH2 towards T Tauri (see Figs. 1 and 2).

3. The Morphologyof the 2.121 µm H2 EmissionTowards T Tauri

Perhaps the most striking property ofthe H2 emission from the T Tau system isits unusual complexity. This may bepartly due to the presumed nearlypole-on orientation, which would causepolar outflows and material in the equa-torial plane to be projected on top ofeach other, and bends and misalign-ments to be exaggerated by projectioneffects. Therefore, we start with a purelymorphological description of the resultsof our observations, before we proceedwith the interpretation.

First, we are interested in the questionwhether most of the bright H2 emissionfrom the innermost star is associatedwith the visual primary or with the infra-red companion. While observations witha scanning Fabry-Perot, in which lineand continuum data are not taken simul-taneously, are not optimum for the de-tection of line emission close to brightcontinuum sources, Figure 2 suggeststhat most of the compact H2 emissioncomes from the infrared companion andthat the emission region is resolved inthe east-west direction. The next obvi-ous features are a very bright knot 2– 3″NW of the primary and an elongatedfilament at a position angle of p 290°(measured from N through E) extendingfrom this knot. A second filament withnearly the same orientation is seen 4.5″further south, apparently pointing to-wards a position 0.7″ south of the IRcompanion. A third filament at positionangle 60° is found in the east, connect-ing another bright knot 1.5″ east of thestars with an extended region of en-hanced emission 9″ to the east. It shouldbe pointed out that none of the filamentsappears to be pointing back to either ofthe stellar components. A complex re-

Figure 1: 2.121 µm H2 (v = 1 – 0) S(1) ro-vibrational emission towards T Tauri. The field size is25″ × 25″, North is up, and East is left. The positions of the two components of the T Tauri binaryare marked with crosses. The continuum emission has been subtracted. The colour table hasbeen chosen to emphasise low surface brightness emission. The most prominent features arethe central “Z”-shaped emission region, and the complex structure in the south towardsBurnham’s nebula. In the north, a faint arc is seen at a position nearly mirror-symmetric toBurnham’s nebula. In addition, three filaments are seen extending from the NW tip of the “Z”,from its SW corner, and from its eastern side.

indication for two outflow systems, oneof them oriented roughly N-S and asso-ciated with T Tau N, the other at anearly perpendicular direction and ema-nating from the infrared companion. Totest this picture, and to get even moredetailed information, higher angular res-olution is required, which provided themotivation for our adaptive optics ob-servations.

2. Observationsand Data Analysis

T Tauri was observed in the (v = 1 – 0)S(1) transition of molecular hydrogen at2.121 µm with the SHARP II cameraattached to the ADONIS adaptive opticssystem on January 15, 1996. To obtainoptimum contrast between the faint lineemission and the bright backgroundfrom the stellar continuum and from thesky, it is necessary to use high spectralresolution. Therefore, we did not use anH2 line filter, but rather SHARP II’shigh-resolution Fabry-Perot etalon,which provides a spectral resolution Rp 2500 (i.e. p 120 km/s). With an imagescale of 0.1″ per pixel, we could cover a

field of 25″ in each exposure. Position-ing T Tauri successively in the fourquadrants of the detector, we obtained asmall 35″ × 35″ mosaic centred on thestar.

Each data set contains a short se-quence of two observations at thewavelength of the emission line, brack-eted by two frames in the adjacentcontinuum at –600 km/s and +600 km/s,respectively. On the one hand, it isdesirable to execute these Fabry-Perotsequences as quickly as possible, tominimise the seeing variations betweenthe line and continuum images. On theother hand, with the relatively smallpixel scale and the high spectral resolu-tion of our observations, detector noisedominates over the sky background forintegration times up to several minutes,which calls for long exposures. As acompromise between these conflictingrequirements, we settled for a frameintegration time of 60 seconds. On Jan-uary 15, 1996, we spent a total of 20minutes integrating on the 2.121 µmline in p 0.7″ seeing. The results fromthese observations are presented here;additional data taken on the previous

13838

gion of fairly bright hydrogen emissionlies 4″ to 10″ south of the stars, towardsBurnham’s nebula. This emission regioncontains several knots and filaments; itappears to end abruptly in a brightsouthern rim. A corresponding, albeitmuch weaker, structure can be seenfaintly in the north. It is worth noting thatthese features are all present at lowerresolution in the lowest contours of Fig-ure 1 in van Langevelde et al. (1994);this gives us confidence that they are allreal. The bright NW knot, the westernfilament, and the northernmost knots ofthe southern emission region were alsodetected by Herbst et al. (1996).

4. Interpretation

The H2 data are suggestive of at leastone collimated outflow in the T Tau sys-tem, roughly in the north-south direc-tion: Burnham’s nebula a few arcsec tothe south of T Tau appears as a brightbow-shock projected close to the line ofsight, while its arc-like counterpart sym-metrically placed a few arcsec to thenorth of T Tau appears much fainter.This flow direction is about the same asthe direction of the superjet seen in Hαas a series of 3 bow-shocks north andsouth of T Tau up to a distance of 20′(1 pc) from the source (Reipurth, Bally,& Devine, in preparation). It is not clearwhether the collimated flow traced in H2originates from T Tau N (the optical star)or T Tau S (the IR companion). On onehand, Solf et al. (1988) have found ajump in velocity in [SII] emission on theoptical star, interpreted as a jet-like fea-ture. On the other hand, the generallypresumed connection between accre-tion and ejection tends to favour the IRcompanion as the jet source, as IRcompanions are considered as tempo-rary accretors (Koresko et al., 1997).Herbst et al. (1996) argue that there aretwo outflow systems, one of them com-prised by the NW knot and the southernemission region, and the other one bythe SW and E filaments. Our findingthat none of the filaments seems toemanate from either star, and the detec-tion of an apparent counter-structure toBurnham’s nebula, appear to contradictthis picture. However, it must be kept inmind that precession and motion of thejet footpoints in the binary system mightlead to strong bends and misalign-ments. Furthermore, the near-IR emis-sion traces shocked hydrogen and notthe bulk of the material, which mayrender the observed morphology mis-leading.

The fact that it is the IR companionrather than the optical star on which theH2 emission is concentrated is taken toimply that matter is falling onto thecircumstellar disk of the IR companionwhereby it is shock-excited. Infall veloci-ties of the order of 5–10 km/s are need-ed to produce the H2 v = 1 – 0 excitationaround 2000 K (Smith & Brand, 1990).

This infall could arise from the inneredge of a circumbinary disk (Artymowicz& Lubow, 1996), which is detected bothin CO interferometric observations(Weintraub et al., 1989, Momose et al.,1996) and in scattered IR light (Wein-traub et al., 1992). There is no questionthat a fair amount of dusty disk andenvelope gas (of the order of 0.01 MA)likely to be bound to T Tau is present.

There are more H2 features in ourmap than can be explained by H2 jets.The filaments or fingers (one to the eastand two to the west) may have nothingto do with jets whatsoever but rathercome from material in the equatorial(disk) plane. We suggest two possiblealternative origins for these features:

(1) Tidal tails caused by the inter-action of the two star-disk systems(T Tau N and T Tau S), much like ininteracting disk galaxy pairs. In thiscase, the expected shock velocities areof the order of the Keplerian orbitalvelocity of the binary system (around5 km/s).

(2) Spiral shocks in the circumbinarydisk of the T Tau binary system. Thesewould arise as the binary drives spiralwaves into the circumbinary disk with apattern speed of the order of the orbitalspeed (5 km/s). Although the gas in thecircumbinary disk has only a small orbit-al velocity (less than 1 km/s at radialdistances in excess of 300 AU), the gas

could be shock excited to emit in H2,while running into the externally excitedspiral potential well with a differencespeed of a few km/s. These spiralshocks could be associated with accre-tion streams from the inner edge of thecircumbinary disk, which also cause theshock excitation of the H2 in the circum-stellar disks.

Finally let us discuss the origin of thebright H2 knot 2–3″ NW of T Tau N. Wespeculate that it might be related to aspiral shock which might be strongestright at the inner edge of the circumbi-nary disk. It is likely a density enhance-ment, because this knot is also visible incoronographic I-band images (Nakajima& Golimowski, 1995) as a reflectionnebula (also seen in Hα but not in [SII],Robberto et al., 1995). It is noteworthythat the whole H2 emission in the imme-diate vicinity of the binary system (in-cluding the NW knot and the IR compan-ion and its respective H2 fingers) are “Z”shaped as is the associated optical re-flection nebula. Thus there is a strongsimilarity between dense gas and dustseen in reflected optical light andshocked H2 emission, which has notbeen noticed before. Perhaps this couldindicate some shock interaction of adiffuse outflow with the remnant materialin the circumbinary environment.

In summary, we stress that the H2features seen in the T Tau system very

Figure 2: Same asFigure 1, but 12.5″ ×18″ field and differentcolour table. Artefactsfrom the non-perfectcontinuum subtractionare seen at the posi-tions of the stars, butit is obvious that mostof the compact H2emission is associat-ed with the southern(IR) component. Twobright knots, which arewell detached from thestars, are visible in theNW and E, at the foot-points of filaments 1and 3. The complexstructure in the southends abruptly in abright rim.

39

map the gas temperature at higher spa-tial resolution, and to search for fluores-cent H2, whose presence is expected onthe basis of IUE observations of UVfluorescence (Brown et al., 1981). Evenmore efficiently, such observations couldbe performed with the 3D integral fieldspectrometer, which will be coupled tothe ALFA adaptive optics system onCalar Alto in the course of 1997. Inaddition to the multiplex advantage of3D, the strictly simultaneous spectra willallow a much better continuum subtrac-tion, and thus a better definition of theemission close to the two bright stars.

Acknowledgements

We thank the ESO staff on La Silla fortheir excellent support during our ob-serving run. We are particularly indebtedto Frank Eisenhauer, who built theSHARP II camera and calibrated itsscanning Fabry-Perot etalon, for his ex-pert advice on observing strategy anddata reduction.

References

Artymowicz, P., & Lubow, S.H. (1996), ApJ467, L77.

Beckwith, S., Gatley, I., Matthews, K., & Neu-gebauer, G. (1978), ApJ 223, L41.

Brown, A., Jordan, C., Millar, T.J., Gondhale-kar, P., & Wilson, R. (1981), Nature 290,34.

Bührke, T., Brugel, E.W., & Mundt, R. (1986),A&A 163, 83.

Carr, J.S. (1990), AJ 100, 1244.Dyck, H.M., Simon, T., & Zuckerman, B.

(1982), ApJ 255, L103.Ghez, A.M., Neugebauer, G., Gorham, P.W.,

Haniff, C.A., Kulkarni, S.R., Matthews, K.,Koresko, C., & Beckwith, S. (1991), AJ102, 2066.

Herbst, T.M., Beckwith, S.V.W., Glindemann,A., Tacconi-Garman, L.E., Kroker, H., &Krabbe, A. (1996), AJ 111, 2403.

Koresko, C.D., Herbst, T.M., & Leinert, Ch.(1997), ApJ, in press.

Momose, M., Ohashi, N., Kawabe, R., Hay-ashi, M., & Nakano, T. (1996), ApJ 470,1001.

Nakajima, T., & Golimowski, D.A. (1995), AJ109, 1181.

Robberto, M., Clampin, M., Ligori, S., Paresce,F., Sacca, V., & Staude, H.J. (1995), A&A296, 431.

Schwartz, R.D. (1975), ApJ 195, 631.Smith, M. D., & Brand, P. W. J. L. (1990),

MNRAS 242, 495.Solf, J., Böhm, K.-H., & Raga, A. (1988), ApJ

334, 229.van Langevelde, H.J., van Dieshoeck, E.F.,

van der Werf, P.P., & Blake, G.A. (1994),A&A 287, L25.

Weintraub, D.A., Kastner, J.H., Zuckerman,B., & Gatley, I. (1992), ApJ 391, 784.

Weintraub, D.A., Masson, C.R., & Zucker-man, B. (1989), ApJ 344, 915.

Zinnecker, H., & Wilking, B.A. (1992), in Bina-ries as tracers of Stellar Formation, Eds.Duquennoy, A., & Mayor, M., CambridgeUniv. Press, p. 269.

likely have different origins. The fact thatthe T Tau system is almost unique inshowing extended H2 emission could(a) be related to the presence of a jetinteracting with the local environment(Burnham’s Nebula and its northerncounterpart), and (b) be related to arelatively massive circumbinary disk anda binary system with a component orbit-al speed of the same order as is neces-sary to shock excite molecular hydrogen(5 km/s) in this disk.

5. Future Prospects

The high angular resolution affordedby SHARP II and ADONIS has enabledus to study the distribution of warmmolecular hydrogen in the T Tauri sys-tem in considerably more detail than hasbeen possible before. With multi-lineobservations it is possible to determinethe excitation mechanism. The ratio ofthe (v = 1 – 0) S(1), (v = 2 – 1) S(1), ande.g. (v = 9 – 7) 0(3) near-IR lines can betaken to distinguish between UV fluores-cence and shock excitation, and to de-termine the gas temperature in the lattercase. From their integral field spectros-copy with 3D, Herbst et al. (1996) con-clude that the bulk of the material isshock excited with a typical temperatureof p 2000 K. Future observations of the(v = 2 – 1) S(1) and (v = 9 – 7) O(3) withSHARP II and ADONIS could be used to

1. Introduction

High angular resolution polarisationand surface brightness images ofR Monocerotis (R Mon) and its asso-ciated reflection nebulae NGC 2261(Hubble’s variable nebula), have beenobtained. Ground-based optical imagingand near-infrared polarimetric data aswell as HST optical polarisation mea-surements are presented as examplesof current high-resolution imaging capa-bilities.

R Mon is a variable semi-stellar ob-ject, which has never been resolved intoa single star. It is presumed to be a veryyoung star in the process of emergingfrom its parent cloud and the source ofillumination of the nebula. For decadesthe variability of both R Mon, andNGC 2261, has been observed both inbrightness and polarisation (Hubble,1916; Lightfoot, 1989). The apparentcometary nebula NGC 2261 extends, inthe optical P 3′ northward of R Mon. COmapping (Canto et al., 1981) reveals

that it is indeed one lobe of a bipolarnebula; the southern lobe being ob-scured by the presumed tilted diskaround R Mon. Some 7′ north of R Monthere are a group of associated Herbig-Haro objects (HH39). A faint loop ex-tending between HH39 and the easternextremity of NGC 2261 has been inter-preted by Walsh & Malin (1985) asevidence of a stellar wind driven flowbetween R Mon and HH39. Imhoff &Mendoza (1974) derived a luminosity of660 LA for R Mon, assuming a distanceof 800 pc (Walker, 1964), which is atypi-cal of the low luminosity normally obser-ved for Herbig-Haro exciting sources.There is also a faint jet-like feature southof R Mon (Walsh & Malin, 1985); this isnot a true jet but dust illuminated byradiation escaping from the dusty diskaround R Mon (Warren-Smith et al.,1987).

In this article we present new obser-vational data at high spatial resolutionand discuss some possible explanationsof the features revealed. These data

serve to illustrate the state of the art inhigh angular resolution polarimetry. Theobservations have been acquired bothfrom the ground with SUSI and ADONISat ESO-Chile and from space with theHubble Space Telescope. In section 2we present in some detail the observa-tional procedures. Some results areshown in section 3 and we end by ashort discussion on R Mon and NGC2261 together with some comments onthe potential of high-resolution polari-metry.

2. Observations

NGC 2261 has been adopted as anHST polarisation calibrator (Turnshek etal., 1990), since it is extended, has highpolarisation (Lp 10%) and has been wellobserved from the ground (see e.g.Aspin et al., 1985 and Minchin et al.,1991). The ADONIS and HST polari-metric data of R Mon and NGC 2261presented here were acquired in orderto calibrate polarimetric data of other

Examples of High-Resolution Imaging andPolarimetry of R Monocerotis and NGC 2261N. AGEORGES, J. R. WALSH, ESO

14040

ESO adaptive optics instrument, ADON-IS, on the 3.6-m telescope. The camerais a 256 × 256 pixels NICMOS3 arrayand a pixel scale of 0.05″/pixel wasused for the observations, giving a fieldof 25″ × 25″ (see the ADONIS homepagehttp://www.ls.eso.org/lasilla/Telescopes/360cat/html/ADONIS/for details). In the case of these obser-vations, the resulting images are onlypartially affected by the seeing effects.The data presented are the first exam-ple of scientific use of polarisationmeasurements coupled with the ADON-IS adaptive optics system.

As point spread function (PSF) andpolarisation calibrator, the nearby un-polarised star HD 64299 was em-ployed. Table 1 indicates the number offrames observed, per position of thepolariser, for a given wavelength andintegration time for both the referencestar and R Mon. Nine positions of thepolariser were used for each series ofobservations: from 0° to 157.5° in stepsof 22.5° and 0° again to finish thesequence. The first and last imageswere compared in order to check thephotometric stability. The basic data re-duction (dead pixel correction, flat field-ing, sky subtraction, stacking of imag-es) were performed with ECLIPSE – animage-processing engine developed atESO for astronomical data reduction ingeneral and ADONIS data in particular.For further information on ECLIPSEsee: http://www.eso.org/~ndevilla/adonis/eclipse. At each pixel, or set of binnedpixels, a cosine function was fitted tothe values as a function of polariserposition angle, and maps of the meansignal, linear polarisation and polarisa-tion position angle were computed forboth the R Mon and HD 64299 datausing a dedicated MIDAS programme.A value of the instrumental polarisationof 2.7% ± 0.2% at 158 ± 2° was derivedat J, with the errors based on the imagestatistics. We consider this as a prelimi-nary value until all the calibration dataon unpolarised and polarised standardstars have been analysed. The po-larisation and position angle of the po-larisation vectors shown in the mapswill change slightly upon adoption of animproved instrumental polarisation.

2.3 Optical HST observations

The WFPC2 imager aboard HST con-sists of four 8002 CCD chips – threeWide Field (0.1″ pixels) and one Plane-tary Camera (0.046″ pixels). There is apolariser filter which is a quad of fourpolarisers oriented at different angles(see the WFPC2 Instrument Handbookfor details). Data can be taken by eitherrotating the filter over a single chip or bytaking images in successive chips. Theadvantage of space observations is thatthe PSF is constant in time, but, in thecase of WFPC2, it is spatially varying

(on account of the correction for thespherical aberration of the telescope).The resolution is therefore high but notdiffraction limited since the PC pixelsundersample the PSF.

As part of the calibration programmefor this instrument, images of R Monand NGC 2261 were taken in February1995 in both modes, together with anunpolarised standard (G191B2B) and apolarised standard (BD+64°106). ThePI was W. Sparks (Programme ID5574). Images taken with the POLQfilter (in which each chip is imaged bythe polariser at a different rotation 0,45,90 and 135°) were studied. A set ofF555W (V band) and F675W (R band)filter images were analysed, all with300s exposure. The field of the threeWFC chips is 80 × 80″ but the field ofthe PC chip is 35 × 35″ and, since thedata from all four chips were used, theimage area was restricted. The polari-sation in R Mon and the brightest por-tions of NGC 2261 was mapped. How-ever, the HST polarisation maps do notgive adequate data over R Mon itselfsince R Mon occurs close to the edgeof the PC chip in a region where thereis cross-talk between the polariserquad filters.

3. Results

3.1 SUSI data

Figure 1 shows an Hα–V colour mapof NGC 2261 from the SUSI images.The map has been convolved by a 3pixel wide Gaussian to smooth the ob-served structures. The colour mapshows a region of redder colour (codedin black in Fig. 1) just to the south ofR Mon and two spikes of bluer colourextending NE and NW, with the bluerone to the NE. There are some smallspatial colour changes in the NGC 2261nebula in addition.

3.2 ADONIS data

On account of their short integrationtimes, the total flux ADONIS images, atJ, H and K, do not show much detail ofNGC 2261. Some extension north-eastof the bright source appears with de-

R Mon HD 64299

# Frames τ (s) # Frames τ (s)

J 10 1 10 5H 20 0.4 10 3K 10 0.1 10 3

Figure 1: Hα – V SUSI colour map. In thisfigure, white corresponds to blue physical col-our and black to redder colour. North is to thetop and east to the left. The size of the imageis p 2′ × 2′. North is to the top and east to theleft.

TABLE 1: Observational details of the ADONISdata taken for R Mon and the unpolarised ref-erence star HD 64299. Columns 2 and 4 indi-cate the total number of frames acquired perposition of the polariser; columns 3 and 5 therespective integration time per frame.

astronomical sources; however, weshow that the high resolution of thedata reveals new details of this fasci-nating source.

2.1 Ground-based opticalobservations

Ground-based optical observationswere made with the SUperb SeeingImager (SUSI) of ESO, at the NTT inApril 1996. The Tek CCD chip of SUSIhas 1024 × 1024 pixels, each with asize of 0.13″, giving a 2′ field of view.These observations are typical of ‘pas-sive’ imaging. An advantage is that alarge field can be observed, but thedisadvantage is that the spatial resolu-tion is at the mercy of the seeing varia-tions. Indeed the V and Hα band datapresented here have been acquiredunder the worst seeing conditions (av-erage 1.2″) compared to the other dataand thus have the lowest resolution.

Images in V and a narrow-band (60 Åwide) Hα filter with exposure times of 60seconds were obtained. The data havebeen flat fielded and then flux calibratedin MIDAS. In the absence of a real fluxcalibrator observed during these obser-vations, and in order to be able to createthe colour (Hα–V) map, the integratedflux in a 5″ aperture centred on R Monwas compared with the equivalent mag-nitudes from Aspin, McLean & Coyne(1991). This calibration does not allowaccurate photometric information to bederived, since R Mon is known to bevariable (e.g. 4 magnitudes in 1 year –Matsumara & Seki, 1995) on short andlong time scales. The calibration is how-ever sufficient for the purposes of scal-ing the colour map.

2.2 Near-infrared observations

The near-infrared polarimetric datawere acquired in March 1996 with the

41

Figure 3: ADONIS J band restored image with the polarisation vectorplot superposed. For this plot the polarisation vectors have beencalculated in square apertures of 0.4″ radius.

Figure 4: Same as Figure 3 but in H band.

correction level and is not repeat-able.

In order to increase the effective res-olution of the images, deconvolutionwith the star HD 64299 used as PSFcalibrator, by the Lucy-Richardson algo-rithm (Lucy, 1974) was applied. Afteronly 10 iterations, the nebular structurearound R Mon is apparent at a level of0.14% of the peak intensity in J Bandfor example. Figures 3, 4, 5 show col-our plots of the restored J, H and K

creasing wavelength. In the K bandimage (see Fig. 2) the spider of thetelescope is clearly recognisable at alevel of up to 0.09% of the peak intensi-ty. Two other optical effects can benoticed in this image: an elongationalong the X-axis, which results from avibration in this direction, whose originis not clear; the triangular shape of thePSF, which is a residual triangularcoma. This last effect, linked to aservo-loop problem, appears at poor

Figure 2: Intensity contour plot of the observed ADONIS K bandaveraged image of R Mon. North is to the top and left to the east. Thescale is in pixels (0.05″/pixel). The trace of the telescope spider iswell seen at the lowest plotted contour level. The central contourspoint out the residual coma in our data as well as a slight East-Westelongation.

images. Two distinct ‘knots’ north-eastof R Mon are revealed in all IR imagesand another to the NW in the J bandimage. These knots are also seen inthe HST V and R images; the NW knotis the same as the NW spike mentionedin section 3.1.

3.3 HST data

The WFPC2 data for R-Mon had to berecalibrated since the original data were

Figure 5: Same as Figure 3 but in K band.

14242

taken before flat fields with the polariserfilters were available. The WFPC2 Webpages give full details of flat fieldingWFPC2 polariser images and under-standing the data(http: //www.stsci.edu/ftp/instrument_news/WFPC2/Wfpc2_pol/wfpc2_pol_cal.html).

The images were corrected for thegeometrical distortion and difference inpixel size for the different chips usingthe standard WFPC2 software in IRAF/STSDAS (wmosaic). The data werehandled in the conventional manner:Stokes Q and U parameters were cal-culated from the difference of the 0 and90° frames and the 45 and 135° framesrespectively. The total flux (sum of po-larised and unpolarised), linear polari-sation and polarisation position anglewere calculated from the Stokes pa-rameters. The errors on the Stokes pa-rameters were calculated based on theimage statistics and propagated to thepolarisation and position angle. Thedata for the polarized and unpolarisedstandards were treated in the sameway. The deduced instrumental polari-sation is about 2.5% for F555W. Dedi-cated software (written in MIDAS) pro-duces the polarisation maps allowingthe data to be binned to increase thesignal-to-noise per pixel and cosmicrays to be rejected. Figures 5 and 6,

Figure 6: HST V band polarisation map, con-volved to a 0.8″ resolution, overplotted on anintensity map of 0.1″ resolution. The orienta-tion in this image is tilted at position angle 15°compared to the previous images, where northis to the top and east to the left.

respectively, show the V and large Rband total flux image and the deducedpolarisation map as a vector plot.

4. Discussion

R Mon is known to have Hα emission(e.g. Stockton et al., 1975) which is thenscattered by dust in the reflection neb-ula; there is no convincing evidence forintrinsic line emission in the nebula andthe polarisation is very similar at Hα andV (Aspin et al., 1985). The SUSI Hα – Vcolour map (Fig. 1) then traces smalldifferences in extinction through thenebula, differences in scattered flux aris-ing from different scattering angles andpossibly also multiple scattering in thenear vicinity of R Mon itself. Interpreta-tion of these colour changes is complexsince R Mon itself also shows temporalcolour changes and these then propa-gate through reflection over the nebula.The ‘structural’ changes observed in

Figure 7: Same as Figure 6 but in large R band.

43

NGC 2261 are considered to be due tothe movement of obscuring clouds in thenear vicinity of the embedded stellarsource which modulate the illuminationescaping in the direction of NGC 2261(“shadowplay” – Lightfoot, 1989).

The most prominent features on theHST images are the lack of a resolvedstellar source to R Mon and the twospikes emanating from it. The easternspike is present at all wavelengths in theimages presented here from V to Kband. It is clearly a dust feature or atunnel of lower obscuration from R Moninto the nebula. The polarisation behav-iour of these two spikes could proveuseful in studying their origin. The NEspike, extended both in length andwidth, shows higher linear polarisation(25% at V integrated over its area) witha trend to increased polarisation out-wards from R Mon. The NW spike, haslower polarisation (12% at V, and fairlyconstant along its length), is narrow(0.3″) and not visible in the H and KADONIS images; Figs. 4, 5). The polari-sation position angles of both featuresare quite consistent with illumination byR Mon. An interpretation of the lowerpolarisation (of the NW knot) could be interms of the (single) scattering angle ofthe radiation from R Mon, reflected offthe dust features towards the observer.The NE spike then is more tilted awayfrom the observer than the NW spike.The features could be either ‘tubes’relatively free of dust within the extend-ed dust disk about R Mon or individualclumps. An extended cloud would reflectlight at different angles giving rise to

differing polarisation along its lengthwhilst a radial feature should producesimilar polarisation along its wholelength (assuming single Mie scattering).However, Scarrott et al. (1989) havestudied the polarisation behaviour of thedisk around R Mon and suggest thatscattering of polarised light and perhapsmagnetically aligned grains may play arole. Detailed modelling is required toresolve these differing interpretationsbut the long wavelength coverage andcomparable high resolution (0.2″) polari-metry offered by HST and ADONISbring critical data to this problem.

Polarisation mapping has been ap-plied to a great variety of astronomicalsources: reflection nebulae around youngstars, AGB stars, proto-planetary andplanetary nebulae; to galactic scale ex-tended emission-line regions in radiogalaxies and normal and dusty galaxies;synchrotron sources in supernova rem-nants and quasars and AGN. Most stud-ies have been at modest spatial resolu-tion dictated by ground-based seeing.However, the application of adaptiveoptics to polarimetry brings a powerfultool to study the nearby regions of dustysources enabling the study of dust struc-tures around AGN, non-axially symmetricoutflows near AGB stars and dust disksaround young stars for example. NIC-MOS will also enable high-resolutionpolarimetry to be achieved in J and Hbands. HST with WFPC2 allows high-resolution optical imaging in the opticaland with the Advanced Camera for Sur-veys (ACS, see http://jhufos.phajhu.edu/for details) installed into HST in 1999,

even higher resolution polarimetry over alonger wavelength range (2500 to 8200Å) will be achievable. Polarimetry bringscritical data related to orientation thatcannot be gained in other ways, and istherefore an important tool for deeperunderstanding of structures surroundedby dust.

5. Acknowledgemen ts

We would like to thank the ADONISteam for their excellent assistance at thetelescope and for the smooth workingof the instrument. We also thank JohnBiretta, at STScI, for helpful discussionson WFPC2 polarimetry.

References

Aspin C., McLean l.S., Coyne G.V., 1985, AA149, 158.

Canto J., Rodriguez L.F., Barral J.F., CarralP., 1981, ApJ 244, 102.

Hubble, E. P., 1916, ApJ 44, 19.Imhoff C.L., Mendoza, V.E.E., 1974, Rev. Mex-

icana Astr. Ap. 1, 25.Lightfoot, J.F., 1989, MNRAS 239, 665.Lucy, L.B., 1974, AJ, 79, 745.Matsumara M., Seki M., 1995, Astroph. and

Space Sci. 224, 513.Minchin N.R. et al., 1991, MNRAS 249, 707.Scarrott, S.D.M., Draper, P.W., Warren-Smith,

R. F., 1989, MNRAS 237, 621.Stockton, A., Chesley, D., Chesley, S., 1975,

ApJ 199, 406.Turnshek, D.A., Bohlin, R.C., Williamson, R.L.,

Lupie, O.L., Koornneef, J., Morgan,D.H., AJ, 99, 1243.Walsh J.R., Malin D.F., 1985, MNRAS 217, 31.Warren-Smith, R.F., Draper, P.W., Scarrott,

S.M., 1987, ApJ, 315, 500.

Proper Motion as a Tool to Identify the OpticalCounterparts of Pulsars: the Case of PSR0656+14R. MIGNANI1, 3, P.A. CARAVEO

3 AND G.F. BIGNAMI 2, 3

1 Max-Plack-Institut fur Extraterrestrische Physik, Garching2 Dipartimento di Ingegneria Industriale, Università di Cassino3 Istituto di Fisica Cosmica del CNR, Milan

1. Introduction

Up to now, about 1% of IsolatedNeutron Stars (INS) have been ob-served in the optical domain (Caraveo,1996). From an observational point ofview, INSs are challenging targets owingto their intrinsic faintness. Young (τ lp104 yrs) pulsars like the Crab, PSR0540-69 and Vela are relatively bright,powered by magnetospheric processes,and easily identified through the detec-tion of optical pulsations. However, suchmagnetospheric emission fades awayrapidly in the optical domain, giving wayto the star’s surface thermal emission. Itis the case of the so-called “Middle-

Aged” (τ p105 yrs) Isolated NeutronStars (MINSs) that, with a temperature TP 105–106 K, can be detected at opticalwavelengths as thermal emitters. How-ever, since neutron stars are extremelycompact objects (10 km in radius), theoptical luminosity of MINSs turns out tobe very faint, preventing the timing oftheir optical emission.

Thus, for MINSs the optical identifica-tion generally relies on the positionalcoincidence with a candidate as well ason its peculiar colours. This is how thecandidate counterparts of Geminga(Bignami et al, 1987), PSR 1509-58(Caraveo et al., 1994b), PSR 0656 +14(Caraveo et al., 1994a) and PSR 1055-

52 (Mignani et al., 1997a; Mignani et al.,1997b) were singled out. However, forsome of the closer objects (d ≤ 500 pc)other, independent, pieces of evidencecan be collected.

Isolated Neutron Stars are known tobe high-velocity objects (Lyne andLorimer, 1994; Caraveo, 1993), movingin the sky with transverse velocities vT p100–400 km/s. A tentative optical identi-fication can thus be confirmed, or, atleast made much more compelling, bymeasuring the proper motion, if any, ofthe optical counterpart. Indeed, this ishow the optical identification of Gemingawas confirmed by Bignami et al. (1993).In the case of a radio pulsar, with known

14444

proper motion, this would indeed securethe optical identification with a degree ofconfidence comparable to that obtainedthrough detection of pulsed emission.

This strategy gains much more powerwhen exploiting the high precision posi-tioning of the target, presently achieva-ble with the imaging facilities onboardHST and, in the near future, with theVLT.

2. PSR 0656 +14: A Middle-AgedNeutron Star

PSR 0656 +14 is one of the INSs stillwaiting for a confirmation of the opticalidentification. It is a “Middle-Aged” (τ p105 yrs) radio (P = 385 ms) as well as asoft X-ray pulsar (Finley et al., 1992)with a thermal spectrum consistent withthe surface temperature (T ≤ 8 × 105 K)expected on the basis of its age (Becker,1996). Recently, PSR 0656 +14 hasbeen tentatively observed as γ-ray pul-sars by EGRET (Ramanamurthy et al.,1996). It is thus, with PSR1055-52, apulsar of the “Geminga class” (Mignaniet al., 1997a). PSR 0656 +14 is alsocharacterised by a p 70 mas/yr propermotion (Thompson and Cordova, 1994)towards SE (position angle p 114°),corresponding to a transverse velocityVT p 250 km/s at the assumed radiodistance (760 pc).

Images of the field were obtained atESO in 1989 and 1991 and a probableoptical identification of PSR 0656 +14was proposed by Caraveo et al.(1994a). A faint (mV p 25) point sourcewas detected in both observations, coin-cident, within a few tenths of arcsec, withthe pulsar’s radio co-ordinates (seeFig. 1a). This object was later tentative-

ly associated by Pavlov et al. (1996) witha point source detected with the HST/FOC. However, the data presently avail-able do not provide a firm optical identifi-cation of PSR 0656 +14. Since the ob-ject’s faintness renders extremely diffi-cult the search for optical pulsations, theoptical identification must rely on thea priori knowledge of the pulsar’s propermotion. We, thus, needed a new, high-resolution image, containing enough ref-erence stars to allow very accuraterelative astrometry to unveil, if at allpresent, its expected angular displace-ment (p 0.5 arcsec in 7 years). Asshown for the measure of Geminga’sparallax (Caraveo et al., 1996) theWFPC2, operated in the PC mode, ispresently the ideal instrument to mea-sure tiny angular displacements sinceit offers a high angular resolution(0.0455″/px) coupled with a 35 × 35arcsec field of view.

3. The Observations

As a starting point to measure theobject’s proper motion, we used our1989 and 1991 ESO images, datingback sufficiently to provide the requiredtime baseline. The observations were

Figure 1: (a) Image of the field of PSR 0656 +14 taken in 1991 with the NTT/EMMI (from Caraveo et al., 1994a); the proposed optical counterpartis marked by a cross. (b) Image of the same region, taken in January 1996 with the WFPC2 onboard HST. The proposed optical counterpart of thepulsar is indicated by an arrow. The image size is about 35 × 35 arcsec.

TABLE 1: Ground-based plus HST observations of PSR 0656 +14 used for the measure of theproper motion.

Date Telescope Filter Pixel size Seeing Exposure time

1989.01 3.6-m/EFOSC2 V 0.675″ 1.5″ 60 min

1991.01 NTT/EMMI-RED V 0.44″ 1.0″ 70 min

1996.01 HST/WFPC F555W 0.0455″ — 103 min

performed with the 3.6-m and with theNTT, respectively, using EFOSC2 andEMMI (see Table 1 for details). Theseimages also provided a useful referenceto plan the upcoming HST observa-tions.

The target was observed in January1996 with the PC, equipped with thewide band F555W filter (λ = 5252 Å;∆λ = 1222.5 Å) i.e. the one closer to theJohnson’s V which was used to obtainour 1989/1991 images. The final image(Fig. 1b) clearly confirms the previous3.6-m/NTT detection of Caraveo et al.(1994a). The object magnitude wascomputed to be 25.1 ± 0.1, consistentwith our previous ground-based meas-urements, but with a considerablysmaller error. We then computed thecentroids for our target and for thereference stars in the field, fitting theirprofiles with a 3-D gaussian using dif-ferent centring areas. This procedureyielded the stars’ relative co-ordinateswith a precision of a few hundredths ofa PC pixel (0.0455 arcsec). We havethen compared the present position ofPSR 0656 +14 with the ones corre-sponding to our ESO 1989 and 1991observations rebinned and tilted tomatch the PC reference frame. While

45

the position of the reference stars doesnot change noticeably, a trend is recog-nisable for the proposed candidate.

Unfortunately, the factor of ten differ-ence in the pixel size between the ESOimages and the PC one translates into alarge error of the object relative positionas computed in the 1989 and 1991frames, thus making it difficult to reliablycompute its displacement. After fittingthe object positions at different epochs,we obtain a value µ = 0.107 ± 0.044arcsec/yr with a corresponding positionangle of 112° ± 9° to be compared withthe radio values obtained by Thomp-son and Cordova (1994). Althoughnot statistically compelling, a hint ofproper motion is certainly present in ourdata.

4. Conclusions

Proper motion appears to be a power-ful tool to secure optical identifications ofcandidate counterparts of INSs, as suc-

Figure 2: Colours ofPSR 0656 +14 com-pared with theRayleigh-Jeans ex-trapolation of the softX-ray spectrum(T p 8 •105 K) as ob-served by theROSAT/PSPC(Finley et al., 1992).The optical/UV fluxeshave been correctedfor the interstellarextinction using theNH best fitting theROSAT data (NHp10 20 cm–2) andcorrespond to theoriginal detection inthe V band (Caraveoet al, 1994a), to theHST observation inthe 555W filter (seetext), to the FOC ob-servation of Pavlov etal. (1996) in the130LP filter and tothe B photometry ofShrearer et al.(1996). The pointsare at least 2 magni-tudes above the ex-pectations.

patible with the Rayleigh-Jeans extrapo-lation of the soft X-ray Planckian, evenconsidering the spectral distortion in-duced by an atmosphere around theneutron star (Meyer et al., 1994). New,as yet unpublished FOC observations inthe B/UV (Pavlov, 1996, private commu-nication) seem to confirm this findingwhich appears to be further substan-tiated by the possible detection of opticalpulsations (Shrearer et al., 1996), ac-counting for almost 100% of the flux inthe B band (B p 25.9) More colours ofthe counterpart are needed to under-stand the optical behaviour of thesource. Now that the FOC has exploredthe B side of the spectrum, it is manda-tory to concentrate on the Red partseeking for an R or I detection. NTTtime has been granted for this projectand the observations are currently un-derway.

Acknowledgements. R. Mignani isglad to thank J. Trümper and M.-H.Ulrich for the critical reading of thepaper.

References

Becker, W., 1996, in Proc. of Würzburg Con-ference Roentgenstrahlung from the Uni-verse, MPE Report 263,103.

Bignami, G.F. et al., 1987 Ap.J. 319, 358.Bignami, G.F., Caraveo P.A. and Mereghetti,

S., 1993 Nature 361, 704.Caraveo, P.A, 1993 Ap.J. 415, L111.Caraveo, P.A., Bignami, G.F. and Mereghetti,

S., 1994a Ap.J. Lett. 422, L87.Caraveo, P.A., Bignami, G.F. and Mereghetti,

S., 1994b Ap.J. Lett. 423, L125.Caraveo, P.A., Bignami, G.F, Mignani, R. and

Taff, L., 1996, Ap.J. Lett. 461, L91.Caraveo, P.A., 1996, Adv. Sp. Res. in press. Finley, J.P., Ögelman, H. and Kiziloglu, U,

1992 Ap.J. 394, L21.Lyne A. and Lorimer D., Nature 369, 127.Meyer, R.D. et al, 1994 Ap.J. 433, 265.Mignani, R., Caraveo, P.A. and Bignami, G.F.,

1997a, Adv. Sp. Res. in press.Mignani, R., Caraveo, P.A. and Bignami, G.F.,

1997b, Ap.J. Lett, 474, L51. Pavlov, G.G., Strigfellow, G.S. & Cordova,

F.A., 1996, Ap.J. 467, 370 . Ramanamurthy P.V. et al 1996 Ap.J. 458, 755.Shrearer A. et al 1996 IAUC 6502.Thompson, R.J. and Cordova, F.A., 1994 Ap.J.

421, L13.

cessfully shown for Geminga (Bignamiet al., 1993). Its application is straightfor-ward, just relying on deep, high-qualityimaging.

Following this strategy, we have usedour ESO 3.6-m/NTT images plus a morerecent HST/PC one, to measure theproper motion of the candidate counter-part of PSR 0656 +14. Although promis-ing, this case is not settled yet. Thesignificant uncertainty makes the pres-ent result tantalising but certainly farfrom being conclusive. Of course, newhigh-resolution observations, couldeasily say the final word. Again, thesecould be obtained with the PC or, fromthe ground, using the VLT.

A firm optical identification ofPSR 0656 +14 would be important toestablish its emission mechanism in theoptical domain and, thus, to completethe multiwavelength phenomenology ofthis source. The multicolour data pres-ently available (Pavlov et al., 1996; Mig-nani et al., 1997a), are not easily com- rmignani@rosat.mpe-garching.mpg.de

One of the most debated argumentsof recent astronomy is the understand-ing of the physical conditions of theIntergalactic Medium (IGM) at high red-shift since it represents a unique probeof the young Universe. The state of the

diffuse matter, the formation and distri-bution of the first collapsed objects, theirnature and evolution with redshift are allopen questions strongly related to theorigin of the Universe. The most distantand powerful sources, the quasars, offer

On the Nature of the High-Redshift UniverseS. SAVAGLIO, ESO

the opportunity to explore this side of theUniverse.

The optical part of high redshift qua-sar spectra (z > 2) manifests the pres-ence of a high number of Lyα absorptionlines due to neutral hydrogen. For very

14646

distant objects, suitable observationsare guaranteed only by 4-m-class orlarger telescopes. In this paper, newresults on the z p 4 Universe obtainedwith EMMI observations (the high-reso-lution echelle spectrograph mounted atthe ESO New Technology Telescope) ofthe high-redshift quasar Q0000-26 willbe presented. These data are part of asample of high-redshift QSO observa-tions obtained during the ESO Key Pro-gramme 2-013-49K devoted to study theIntergalactic Medium at high redshift(P.I. S. D’Odorico).

The light emitted by Q0000-26 (at aredshift of z = 4.12) scans the Universestarting at an epoch when it was only10% of its present age (for a flat Uni-verse with q0 = 0.5). Since it is one of themost distant and intrinsically brightestobjects, it has been observed severaltimes in the past. Recently, new obser-vations at high resolution and signal-to-noise ratio have been presented al-most simultaneously by two differentgroups of astronomers: by Lu et al.(1996) using the 10-m Keck Telescopeand by Savaglio et al. (1996) with theESO 3.5-m NTT. In Figure 1, an interest-ing comparison between the two spectrais shown. In that spectral range, theKeck spectrum has a resolution ofFWHM = 6.6 km s–1 for a total exposuretime of 12,000 seconds. The NTT obser-vations in the same spectral range havea resolution of FWHM = 13 km s–1 and atotal exposure time of 27,600 seconds.One may notice that all the informationcontained in the Keck observations canbe found in the NTT ones, even if theresolution is considerably better in theformer. This means that higher resolu-tion does not help in overcoming anintrinsic limitation of the high-redshiftQSO observations due the confusion ofthe absorption lines. The signal-to-noiseratio is similar in the two spectra, whichis what one would expect comparing theexposure times and the different collect-ing areas. Remarkable differences in theinterpretations of the data are due to theline-fitting procedure used.

The majority of the very numerouslines seen in QSO spectra has HI col-umn densities below NHI p 1015 cm–2.Higher HI column density systems (NHILp 1016 cm–2) frequently show associat-ed metal lines which probably haveorigin in protogalactic halo or diskclouds. In the past, astronomers haveraised a still unresolved question on thenature of the low HI column density (NHIlp 1015 cm–2) clouds, commonly referredto as Lyα forest. In particular, they havebeen thought for a long time to beintergaIactic gas clouds of primordialchemical composition not associatedwith galaxies. Only the recent perform-ances of optical telescopes allowed tosee in more detail the QSO spectra atthe expected positions of the metal linesassociated with the Lyα lines and to givenew constraints on the chemical compo-

Figure 1: Spectrum of the high redshift quasar Q0000-26 (z = 4.12). The two upper panels havebeen obtained with the 10-m Keck telescope (Lu et al., 1996) at a resolution of FWHM =6.6 km s–1 and an exposure time of 12,000 seconds. The two lower panels are NTT data(Savaglio et al., 1996) in the same spectral range at a resolution of FWHM = 13 km s–1 and anexposure time of 27,600 seconds.

sition and evolutionary state of thoseclouds. The NTT observations revealedthe presence of C IV and Si IV in three

optically thin HI clouds in the redshiftrange 3.5 l z l 3.8 (Savaglio et al.,1996). The age of these clouds, assum-

47

Figure 2: Models of the UV ionising background as a function of the frequency. The red, greenand blue curves indicate the models with log SL = 2, 2.9 and 3.5, respectively. For each value ofSL three different shapes of Jν in the frequency interval 1 m ν/νLL m 4 are shown. The verticaldotted lines indicate the values of ionisation potential for species of interest.

ing that they are connected to galaxyformation and the redshift of galaxyformation is not beyond zf = 10, is lowerthan 1.4 Gyr (for H0 = 75 km s–1 Mpc–1

and q0 = 0.1) or 0.9 Gyr (for H0 =50 km s–1 Mpc–1 and q0 = 0.5).

To investigate the physical state ofthese systems, namely their ionisationstate, density and size, it is possible tomodel the observed parameters of theclouds. Optically thin systems at highredshift with metal absorption lines areparticularly interesting by themselves,not only because they are related to theearly phases of galaxy evolution, butalso because their ionisation state isprobably dominated by the UV back-ground flux, thus providing indirect infor-mation on the IGM.

The UV background (UVB) radiationfield at high redshift is the result of theintegrated light of the collapsed objectsat early epoch once the absorption dueto both diffuse IGM and discrete cloudsis taken into account. Its shape, intensityand redshift dependence (Jν(z)) is asubject of controversy both from thetheoretical and observational point ofview. Direct estimates at the Lyman limitwavelength λLL = 912 Å as a function ofredshift (J912(z)) can be given by com-paring the ionisation state of the cloudsclose to the quasar with those far fromthe quasar and presumably ionised bythe UVB. Modelling of Jv(z) has beenmostly done considering the quasar in-tegrated light only (Haardt & Madau,1996). The effects of the presence of apopulation of young galaxies have beendiscussed by Miralda-Escude & Ostriker(1990). The soft spectrum of young gal-axies would cause an increase of thejump of the UVB at the two edges of theLyman limit wavelengths for H I andHe II, namely SL = J912/J228. The sameeffect would be obtained if an absorptionof He II in the diffuse IGM were consid-ered. The amount of He II in the IGM hasnot been firmly established yet and it willnecessitate further efforts with HST ob-servations for more significant esti-mates.

Figure 2 shows the details of possibleUVB spectra. For three different valuesof log SL = 2, 2.9, 3.5, three differentspectral shapes in the range 1 m ν/νLL m4 (912 Å M λ M 228 Å) can be consid-ered. Also indicated as dotted verticallines are the ionisation potential of themost interesting species. It is clear thatthe abundances of these species de-pends not only on SL but also on theshape of Jν for 4 m ν/νLL m 1.

Given the shape and intensity of theUVB, it is possible to explore the par-ameter space using the observed ionabundances. Figure 3 shows the resultsof photoionisation calculations usingCLOUDY (Ferland, 1993) applied to oneof the optically thin systems in Q0000-26at z = 3.8190 observed with the NTT.The UVB spectra used as ionisingsources are those indicated in Figure 2.

The values of silicon-to-carbon ratiosare shown as a function of the carbonabundances [C/H] for different values ofthe gas density obtained forcing thecode to reproduce the observed H I andC IV column densities.

In the study of the heavy elementabundances of the Interstellar Medium(ISM) and stars, the abundance mea-surements are usually reported as afunction of the iron-to-hydrogen ratiosince the iron content is usually consid-ered an age estimator (Timmes et al.,1995). The evolution of carbon in diskand halo stars follows that of iron. In fact,iron is mainly produced by type IaSupernovae, while carbon is supposedto be mainly produced by intermediate-mass stars and these two processeshave similar time scales. The siliconhistory is different being higher than inthe sun for low metallicity, since theintermediate-mass α-chain elementsare formed early by type II Supernovaevents. In QSO absorption lines, carbonis easier to detect than iron, therefore,normally, in absence of iron, carbon isconsidered for metallicity evaluationsassuming that carbon abundance evolu-tion follows that of iron. This is justifiedby models of chemical evolution, eventhough observations show a large

spread with no particular trend (Timmeset al., 1995). Comparisons with the localISM abundances are complicated by thefact that iron is strongly depleted intodust grains, both in warm and cold gas(–1.4 dex in the first and –2.2 dex in thesecond, Lauroesch et al., 1996) where-as carbon is much less hidden into dustgrains, its depletion being around –0.3dex. Therefore, considering the carbonabundance as metallicity indicator hasthe advantage that depletion in low-density gas is negligible, especially athigh redshift when the dust formation isprobably in an early phase. The red linesin Figure 3 represent the silicon over-abundance with respect to carbon as afunction of metallicity as found by chem-ical evolution models.

Another interesting parameter whichcan be explored by photoionisation cal-culations is the physical size of theabsorbing cloud. The regions with R L50 kpc and R l 1 kpc are representedby the left and right shaded regions,respectively, in Figure 3. The green dotin each panel represents the result for agas density of nH = 10–3.1 cm–3. Dotstowards lower metallicity represent gasdensities decreasing by steps of 0.2dex, and vice versa for dots towardshigher metallicities.

14848

Figure 3: Silicon-to-carbon ratio as a function of the carbon-to-hydrogen ratio for the absorb-ing system at z = 3.8190 observed in Q0000-26 ([X/Y] = log(X/Y)A – log(X/Y)). Both valuesare compared to the solar values. The dots are results of CLOUDY calculations obtainedvarying the gas density nH with a 0.2 dex step. The green dot indicates the result for log nH =–3.1. Lower densities have higher [C/H] values, while higher densities have lower [C/H]values. The shaded areas indicate the regions where the cloud sizes become smaller than 1kpc (right side) or larger than 50 kpc (left side). The results are shown for three differentmodels of Jν and three different SL values.

the calculations indicates that the cloudis smaller than typical clouds of the Lyαforest for which lower limits to the sizehave been found to be 70 kpc (for H0 =70 km s–1 Mpc–1) at redshift z p 2(Smette et al., 1995). The difference isalso in the gas density which is alwayslower than the typical value of nH = 10–4

cm–3 found in the Lyα forest clouds.Thus we are still far from concluding thatall the optically thin Lyα lines have thesame intergalactic origin.

Metal systems optically thin to the HIionising photons are an important probeof the shape of the UV backgroundradiation. If this is the main ionisingsource, the quasar light contribution onlyabsorbed by H I of the discrete Lyαclouds and Lyman limit systems (whichimplies log Sl p 1.5) cannot explain theobservations. This means that either anadditional source of ionisation is neces-sary, or intergalactic diffuse He II is re-sponsible for the spectral shape of Jν atwavelengths lower than 228 Å, or both.Further detailed analysis of NHI p 1015

cm–2 absorption systems with metallines will hopefully clarify this situation.Challenging results can be obtained bydedicated observations with 4-m-classtelescopes and do not need to wait forthe VLT to be operational.

References

Ferland G.J., 1993, University of Kentucky,Department of Physics and Astronomy In-ternal Report.

Haardt F., Madau P., 1996, ApJ, 461, 20.Lauroesch J.T., Truran J.W., Welty D.E., York

D.G., 1996, PASP, 726, 641.Lu L., Sargent W.L.W., Womble D.S.,

Takada-Hidai M., 1996, ApJ, 472, 509.Miralda-Escude J., Ostriker J.P., 1990, ApJ,

350, 1.Savaglio S., Cristiani S., D’Odorico S., Fon-

tana A., Giallongo E., 1996 A&A, in press.Smette A., Robertson J.G., Shaver P.A., Rei-

mers D., Wisotzki L., Köhler T., 1995, A&AS,113, 199.

Timmes F.X., Woosley S.E., Weaver T.A.,1995, ApJS, 98, 617.

The expected values of [Si/C] can bereproduced by the models with log SL =2 only assuming a cloud size by farsmaller than 1 kpc and this would give ametallicity for this cloud larger than 1/10of solar. Therefore, observations can bebetter explained only if log SL L 2. Themetallicity in log SL = 2.9 models is in the

range s2.2 lp [C/H] lp s1, while thedensity spread is s3.5 lp log nH lp s2.9.A similar spread in metallicity is obtainedfor the log SL = 3.5 models, whereas theupper limit to the density is log nH lps3.7. Metallicities lower than 1/10 ofsolar found in this cloud, suggest ageslower than 1 Gyr. The size inferred by Sandra Savaglio, e-mail: ssavagli@eso.org

1. Introduction

Isolated Neutron Stars (INSs) move inthe sky with high (≥ 100 km s–1) tangen-

tial velocities (Lyne & Lorimer, 1994).Indeed, proper motions have beenmeasured for several INSs, mainly in theradio domain (Harrison et al., 1993). In

addition, for two (possibly three) INSsproper motions have been measured inthe optical domain through relative as-trometry of their optical counterparts.

Optical Observations Provide a New Measureof the Vela Pulsar’s Proper MotionF.P. NASUTI1, R. MIGNANI2,1, P.A. CARAVEO1 and G.F. BIGNAMI1, 3

1 Istituto di Fisica Cosmica del CNR, Milano, Italy2 Max-Plack-Institut für Extraterrestrische Physik, Garching3 Dipartimento di Ingegneria Industriale, Università di Cassino, Italy

49

TABLE 1: This table summarises the published values of the Vela pulsar proper motion obtained from radio as well as optical observations.

Optical Radio

Bignami & Caraveo Ögelman et al. Markwardt & Ögelman Bailes et al.* Fomalont et al. 1988 1989 1994 1989 1992

µ < 60 mas 38 ± 8 mas 49 ± 4 mas 59 ± 5 mas 116 ± 62 masµα –26 ± 6 mas –41 ± 3 mas –48 ± 4 mas –67 ± 20 masµδ 28 ± 6 mas 26 ± 3 mas 34 ± 2 mas –95 ± 75 mas

*After correction for the motion of the Sun and for the rotation of the Galaxy they obtain µα= −40 ± 4 mas and µδ = 28 ± 2 mas.

Epoch Telescope Pixel size Exp. time ∆α ∆δ

1975.2 CTIO 4-m — 45 min. 0 ± 0.05 0 ± 0.051987.1 ESO 3.6-m 0.67″ 30 min. −0.46 ± 0.06 0.21 ± 0.061988.1 ESO 3.6-m 0.67″ 30 min. −0.52 ± 0.09 0.31 ± 0.051990.1 ESO NTT 0.26″ 85 min. −0.73 ± 0.05 0.40 ± 0.051995.1 ESO NTT 0.36″ 30 min. −0.92 ± 0.02 0.45 ± 0.02

Table 2: Data set used for the proper-motion measure. The displacements of Vela are computedwith respect to the 1975 frame. The quoted uncertainties include stars’ centring errors as wellas errors in the rebinning/rotation procedures.

These NSs are Geminga (Bignami et al.,1993), the Vela pulsar (Bignami andCaraveo, 1988; Ögelman et al., 1989;Markwardt and Ögelman, 1994) and,possibly, PSR 0656 +14 (Mignani et al.,1997). In the case of Geminga, becauseof its radio quietness, optical observa-tions are the only way to measure suchproper motion. In particular, proper-mo-tion measures in the radio and opticalbands are a valuable strategy to confirmoptical identifications of pulsars that, inthe optical domain, are far too faint forfast photometry observations (Mignaniet al., 1997).

2. The Vela Pulsar

Optical observations represent a realalternative to measure proper motionssince they are independent from timingirregularities. This is specially true inthe case of the Vela pulsar, well knownfor its gigantic glitches. Indeed, a com-parison between published values ofthe Vela pulsar proper motion showssignificant differences between radioand optical measurements (see Table 1for a summary) with angular displace-ments going from a minimum of 38mas yr–1 (Ögelman et al., 1989) to amaximum of 116 mas yr–1(Fomalont etal., 1992). Thus, although the Vela pul-sar moves in the sky, the actual valueof its displacement is still to be mea-sured.

3. The Observations

For the measure of the proper mo-tion we have used a set of four imagescollected over several years with theESO telescopes, 3.6 m and NTT (seeTable 2). To obtain a reliable measureof the pulsar’s proper motion, we haveselected only those images obtained in

Figure 1: Images of the Vela pulsar field taken in 1975 with the CTIO 4-m and in 1995 withNTT/EMMI.

the Johnson’s B filter. This overcomesshifts of the object’s centroid inducedby differential refraction of the Earth’satmosphere (Filippenko, 1982 and ref-erences therein). Our data set was laterintegrated by the original image, takenin 1975, which lead to the discovery ofthe Vela optical counterpart (Lasker,1976). Since the 1975 image is record-ed on a broad-band blue IIIa-J plate, ithas been digitised in order to allow animmediate comparison with other, morerecent images, taken with CCD detec-tors.

To compute the pulsar’s proper mo-tion, the different images have beenrebinned and rotated in order to matchexactly the same scale and orienta-tion.

As a reference frame we used themost recent image of the field which isalso the one obtained under the bestseeing conditions (p 0.8 arcsec). Re-binning and rotating were applied usingthe standard alghorithms available in

MIDAS. After this procedure, all thereference stars were seen to overlapwithin a few hundredths of the refer-ence pixel.

4. The Proper Motion

Contour plots of all the images werelater prepared in order to provide animmediate guess on the source’s dis-placement (Fig. 2). As it is clearly seenfrom the figure, while the position of thereference stars does not change, thepulsar exhibits a displacement of about1 arcsec to NE. The angular displace-ment of Vela was then computed usingthe 1975 position as starting point andfitting linearly the pulsar’s position at thedifferent epochs.

The resulting proper motion is

µα = −47 ± 3 mas yr−1

µδ = 22 ± 3 mas yr–1

for a total annual displacement of 52 ±5 mas yr−1 with a corresponding posi-tion angle p 295°, which appears com-pletely unrelated to the soft X-ray jetobserved by Markwardt and Ögelman(1995) to protrude south from thepulsar.

5. Conclusions

Our measurements make a finalcase for the Vela pulsar proper motionto be around 50 mas/yr. Since ourproper-motion study covers a period of20 years, the present result can be

15050

sessment will only be possible throughaccurate HST astrometry, currently inprogress, which could also yield ameasurable annual parallax, endingthe current debate on the Vela pulsar/SNR distance. Indeed, a reanalysis ofROSAT data by Page et al. (1996)

considered reliable. In particular, ourvalue is compatible with the previousmeasurements of Markwardt and Ögel-man (1994) and Bailes et al. (1989),while it is clearly incompatible with theones of Ögelman et al. (1989) andFomalont et al. (1992). A better as-

Figure 2: Contour plots of the five images available. While the position of the reference starsdoes not change, the overall displacement (~ 1 arcsec) of the Vela Pulsar over a time span of20 yr is evident.

suggests that the pulsar could be asclose as 285 pc, while radio measure-ments (Taylor et al., 1993) provide adistance of p 500 pc.

Acknowledgements

We wish to thank C. Markwardt andH. Ögelman, who kindly made availablea copy of the poster presented at the1994 AAS meeting. We would also liketo thank B. Lasker for providing the1975 plates.

References

Bailes M. et al., 1989, ApJ, 343, L53.Bignami G.F. & Caraveo P.A. 1988, ApJ, 325,

L5.Bignami, G.F., Caraveo P.A. and Mereghetti,

S., 1993 Nature 361, 704.Filippenko A.V.,1982, PASP, 94, 15F.Fomalont E.B. et al., 1992, MNRAS, 258,

497.Harrison et al. 1993, MNRAS 261,113.Lasker B.M, 1976, ApJ, 203, 193.Lyne A.G. & Lorimer D.R., 1994, Nature 369,

127.Markwardt C. & Ögelman H., 1994, BAAS,

26,871.Markwardt C. & Ögelman H., 1995, Nature,

375, 40.Mignani R. et al., 1997, The Messenger 87,

43.Ögelman H. et al., 1989, ApJ, 342, L83 .Page D. et al., 1996, in Proc. of the Interna-

tional Conference on X-ray Astronomy andAstrophysics Röntgenstrahlung from theUniverse. Würzburg, MPE Report 263, 173.

Taylor J.H. et al, 1993, ApJS 88, 529.

F.P. Nasuti; e-mail: nasuti@ifctr.mi.cnr.it

BavarianPrimeMinisterat La SillaIn the afternoonof March 9, 1997,the Bavarian PrimeMinister, Dr. EdmundStoiber, on the invita-tion of the DirectorGeneral of ESO,Professor RiccardoGiacconi, visited theESO La Silla Obser-vatory.The photographshows the PrimeMinister andMrs. Karin Stoiberon the platform nearthe ESO 3.6-mtelescope with thelower observatoryarea in the back-ground.

51

O T H E R A S T R O N O M I C A L N E W S

The Astronomy On-Line ProjectR. WEST and C. MADSEN, ESO

latter would be responsible for all As-tronomy On-Line-related activities, in-cluding the establishment and subse-quent operation of national AstronomyOn-Line Web sites.

The Programme

The basic Web structure was availa-ble at the beginning of August. This

Throughout the summer of 1996, dis-cussions were held between the in-volved parties at EAAE and ESO, inparticular during an intensive one-daymeeting that took place at the ESOHeadquarters in mid-June. At that time itwas decided to set up an InternationalSteering Committee of eight key per-sons and also National Steering Com-mittees in all participating countries. The

Figure 1: The virtual marketplace was the focal point of Astronomy On-Line.

A major Web-based educational pro-gramme, known as Astronomy On-Line,has just taken place in close collabora-tion between the European Associationfor Astronomy Education (EAAE), theEuropean Southern Observatory andthe European Union.

During a period of two months, fromearly October to late November 1996, acomprehensive network of astronomy-oriented educational Web-pages wasbuilt up at various European sites, in-cluding the ESO Headquarters inGarching. Throughout this period, as-tronomy-interested groups of mostlyyoung people from all over Europe reg-istered with Astronomy On-Line; in theend, 720 groups with approximately5000 participants from 39 countriestook part.

The Astronomy On-Line Web-site atESO received up to 100,000 hits perday. All pages were mirrored once perday or more frequently to about 25mirror sites in other European countries.No accurate statistics are available forthe number of entries at these sites, butthere is little doubt that Astronomy On-Line quickly developed into what theorganisers early claimed: the World’sBiggest Astronomy Event on the World-Wide Web.

Background

Astronomy On-Line took place underthe auspices of the European Unionwithin the yearly European Week forScientific and Technological Culture.This Programme was first conceived inApril 1996 and presented to the Europe-an Commission during a special meet-ing in Brussels early that month. Fromthe beginning, it was carried out underthe direct responsibility of the newlyestablished European Association forAstronomy Education, with technicalsupport by ESO and financial support bythe European Commission.

The main goal was to create a systemwhich would attract young people, inparticular students in Europe’s highschools, to participate in an exciting, well-structured “astronomical” Web event.The benefits from such a Programmewould be many-fold: interest in the sci-ence of astronomy and astrophysics,useful knowledge about the efficient useof the nearly unlimited possibilities on theWeb and, not the least, the establishmentof personal contacts across Europe’sborders. There is no doubt that all threegoals were amply fulfilled.

15252

ence available. Thus, it was natural thatvarious modifications took place duringits implementation. The organisers werealso pleased to obtain advice from manysides.

In view of the relatively short timeavailable, a simple and efficient strategywas adopted. It took the form of a“Market Place”. According to this con-cept, the participants were able to ac-cess a central Web page with ninedifferent “Shops”, each representing dif-ferent activities. The corresponding URLis: http://www.eso.org/astronomyonline/market/

For instance, one much visited “Shop”enabled the students to join a series of“Collaborative Projects”. This includedobservations of the Lunar Eclipse on 27

Figure 2: Activities around the partial solar eclipse on October 12 became a prelude to themany collaborative projects, carried out during the Astronomy On-Line peak phase.

The ESO AstronomyOn-Line Team

James Brewer, Jacques Breysacher,Fernando Comeron, Vanessa Doublier,Christian Drouet d’Aubigny,Gregory Dudziak, Hans-Hermann Heyer,Edmund Janssen, Lex Kaper,Kurt Kjär, Jacco van Loon,Claus Madsen, Karen Müller,Michael Naumann, Resy de Ruijsscher,Elisabeth Völk, Rein Warmels,R. West, Albert Zijlstra, Herbert Zodet

involved a very welcome and effectivesupport by the ESO Data ManagementDivision staff. From then on, many Eu-ropean astronomy teachers contributedactively to the development with infor-mation, exercises, images, texts, etc.which were progressively inserted intothe Astronomy On-Line structure. Infact, this process continued until thelast few days in November, providing asteady inflow of new possibilities forthe participants and thereby ensuringthat their interest continued at a highlevel.

No other project of this scope andvolume had ever been attempted any-where in the world, and Astronomy On-Line, as a pilot project, had to be set upwith comparatively little previous experi- September and the Solar Eclipse on 12

October. These two events could befollowed live on the Web, thanks todiligent observers among the partici-pants who made their images availablein real time. In the follow-up activities, themeasurements made by the studentswere among others used to calculate thedistance to the Moon and thereby thephysical size of our satellite. Consideringthe comparative simple observationaltools employed, the achieved accuracywas quite impressive.

In another Collaborative Project, theparticipants made measurements of thelength of a gnomon shadow at localnoon, repeating the historical experi-ment of Eratosthenes and, when thesevalues were compared, the circumfer-ence of the Earth was calculated towithin 5 percent of the correct value.

Another Shop permitted participantsto submit requests for observations fromseveral professional observatories inBulgaria, Denmark, France, Germany,Great Britain, Portugal, Slovenia, Spainand La Silla in Chile. These had kindlymade a substantial number of nightsavailable for this purpose. Many excel-lent observing proposals were receivedand, although the weather – as expect-ed at this time in Europe! – was notequally good in all places, most of therequests could be satisfied. The datafiles were transmitted through the Webto the proposers, and groups of studentsat many schools have since been busyreducing their observations of galaxies,stellar clusters, comets, etc. A group ofESO fellows provided great support forthis part of the Astronomy On-Line Pro-gramme.

In order to promote knowledge aboutWeb techniques, specific informationwas inserted into the Astronomy On-Line structure with the appropriate Weblinks, e.g. about how to search on theWeb. “Treasure Hunts” were establishedin support, with visits to various astrono-my related-sites, for instance the ESOVLT pages. And, not the least, the stu-dents had the opportunity to communi-cate via e-mail and during the last fewdays of the project, also by “Whiteboard”and “Chat”.

53

In many observatories professionalastronomers agreed to answer ques-tions. Reports to the Steering Commit-tee reveal they were positively im-pressed by the quality and depth ofmany of the inquiries submitted by theyoung participants.

An Astronomy On-Line “Newspaper”kept the participants informed about thelatest developments. This not only in-cluded information by the observatories,but also reports of the related activitiesfrom the participants themselves.

Reaching the Public

The success of a programme such asAstronomy On-Line obviously lies in theeffective organisation of the work andtasks. Equally important, however, is toreach as many potential participants aspossible.

This was achieved through a com-bination of information meetings forteachers (at the national level), the dis-tribution to secondary schools of attrac-tive posters as well as press releasesand video news reels to the media bothby ESO and the National Steering Com-mittees. Thus Astronomy On-Line foundits way into newspapers, magazines,radio and TV.

Also a big effort on the part of the Na-tional Steering Committees went intofinding suitable sponsors, e.g. networkproviders and hardware companies, aswell as maintaining close contact to thenational ministries of education, who inmany cases provided extremely valua-ble help.

Astronomy On-Line maintained a veryhigh visibility during the peak phase ofthe programme. At the same time it wasa pilot programme and as such, theexperience from it is of great interest toexperts in network-based learning. Forthis reason Astronomy On-Line was fol-lowed very closely by Dr. Jari Multisiltafrom Tampere University of Technology.At the end of the programme (22 No-vember 1996), all participants were invit-ed to fill in a detailed questionnaireabout their own background and person-al experience with Astronomy On-Line.Following the evaluation of the incomingreplies, it is the intention to publish theresults in the journal Education & Infor-mation Technologies. A provisional re-port, however, is already available onthe Internet (http://www.eso.org/astron-o m y o n l i n e / m a r k e t / n e w s p a p e r /news1703.html)

Current Status

The entire Astronomy On-Line struc-ture is still available at the Web, forinstance from the central ESO site URL:http://www.eso.org/aol/. Although thedaily number of hits is considerablylower than during the “Hot Phase” in

November, it is still obvious that it isbeing used extensively.

The EAAE organisers of AstronomyOn-Line early decided that it would bedesirable to continue this programmeand to establish it on a more permanentbasis. Indeed, it has already become animportant tool for the exchange of infor-mation among Europe’s astronomy edu-cators. Few doubt that it will again beused extensively when particular eventsas Solar and Lunar Eclipses offer theopportunity of joint projects among as-tronomy-interested students in Europe’sschools.

Early 1997, the EAAE made a formalapproach to the EU for continued sup-port of Astronomy On-Line. If granted,this would ensure the maintenance andfurther development of the Astronomy

Figure 3: The “Try your skills” shop had a wide range of exciting exercises on offer.

On-Line Web sites, thus enhancing theusefulness of this unique educationaltool. Meanwhile, the Astronomy On-Linesite at ESO will be kept running.

Conclusion

Apart from its educational aspects,Astronomy On-Line proved to be apowerful vehicle for stimulating the in-terest in astronomy. Reports filed bythe National Steering Committeesshow the huge number of national ac-tivities, that were prompted by the veryexistence of the programme. These ac-tivities ranged from popular talks byprofessional astronomers, visits to localobservatories, joint (and public) obser-vations – in particular in connection

15454

with the lunar eclipse and the partialsolar eclipse – to dedicated, often high-ly successful, efforts to accelerateschool plans for getting an Internetconnection.

We at ESO have been pleased to beinvolved in this project from the begin-ning and thereby to contribute to thethree goals mentioned above. We arealso confident that Astronomy On-Linewill provide a stimulus to young peopleto become more interested in naturalsciences, especially at a time whenworries have been expressed at manyEuropean Universities about the dwin-dling number of young people who areconsidering to pursue a career in thesefields.

Acknowledgements

Astronomy On-Line was made possi-ble by the active personal involvementof a large number of dedicated people allover Europe and, indeed, on several

The European Week for Scientific andTechnological Culture

The objective for the European Week forScientific and Technological Culture is toimprove European citizens’ knowledge andunderstanding of science and technology,particularly in their European dimension: itaddresses both pan-European scientific andtechnological co-operation, as well as sci-ence and technology in each Europeancountry.

“The Week” is organised every autumnby the European Commission in collabora-tion with national and international researchorganisations, universities, museums, TV,etc.

Introduction

The range of astronomical computinghas become so broad that there is noone single data-processing systemwhich covers all possible aspects. Thus,in addition to the indigenous data analy-sis system MIDAS (Munich Image DataAnalysis System) ESO supports anumber of other analysis systems. Thetwo most important ones are, at thistime, IRAF (Interactive Reduction andAnalysis Facility), and IDL (InteractiveData Language).

For many years now, IDL has played asignificant role in astronomical dataanalysis. Originally developed by DaveStern of Research Systems Inc. in thelate seventies, IDL has survived manytransitions of hardware platforms andoperating systems, all the time improv-ing the functionality while retaining fullbackward compatibility. This is a trulyremarkable achievement.

For Space Telescope data analysis,IDL has been an important tool since thevery early days. Since it had been usedfor the analysis of IUE data, and staff ofthe IUE centre at GSFC was heavilyinvolved in the development of the God-dard High Resolution Spectrograph(GHRS), a large amount of spectroscop-ic analysis software was available inIDL. In addition, IDL provided the possi-bility to read the HST native data format,GEIS (Generic Edited Information Set).

Because of its importance for HST dataanalysis, ST-ECF has been supportingIDL since 1984. More recently, IDL hasbeen made available for the full ESOcommunity.

An often heard objection to the use ofIDL is its commercial nature, i.e. licensefees have to be paid. However, there areseveral aspects to consider: the costsare quite moderate in comparison withthe other computer system related costs(hardware, operating system, utility soft-ware), and buying an IDL license opensup an enormous amount of astronomicalapplication software which has beendeveloped over the years, representingan effort of hundreds of person-years;and, IDL has the potential of considera-ble cost savings by increasing the effi-ciency and productivity of software de-velopers.

IDL is available on the ESO local areanetwork to staff and visiting astrono-mers. Copies of the most importantastronomical libraries are available lo-cally, or they can be downloadedthrough the Internet.

Application ProgrammeDevelopment

Experience shows that it is very diffi-cult to incorporate contributed softwarewritten by astronomers for their ownresearch into systems like IRAF andMIDS. The reason for this is that the

requirements for adherence to stand-ards, level of documentation, testing,quality control and configuration man-agement have to be so strict that re-searchers are unwilling and/or incapableof following them. On the other hand,such requirements are absolutely nec-essary in order to keep the systemtransportable, maintainable, and relia-ble. As a consequence, we can hardlyever incorporate “contributed” softwareinto MIDAS or IRAF as is. Instead, wehave to take the research software(which may or may not run under one ofthose systems), re-design and re-code itin the proper manner, and then incorpo-rate it into the target system.

We have found this process to be verysmooth when IDL is used. IDL allowsquick and easy, step-by-step prototypingby the researcher or by the s/w develop-er. A large function library encouragesmodularity. The code is eminently reada-ble, making the process of re-casting itinto a different target language veryeasy. In fact, IDL can be considered apowerful detailed level design languagewith the advantage that it actually exe-cutes.

Of course, there is a penalty for allthis: although IDL is very fast as far asinteractive languages go, and it canactually be pre-compiled to speed up theexecution, it cannot compete with For-tran or C when it comes to pure numbercrunching involving a large number of

Availability of IDL at ESO/ST-ECFR. ALBRECHT, W. FREUDLING, R. THOMAS,Space Telescope – European Coordinating Facility

continents. Thanks to their unremittingefforts, it became possible to implementin an extremely short time, a projectwhich will undoubtedly serve as a very

useful example for similar events withinother subject areas.

Richard West, e-mail: rwest@eso.org

55

large data sets. This is one of the rea-sons why it is necessary to re-codeapplication programmes. Another rea-son is the fact that sophisticated dataanalysis systems offer enormous func-tionality which, however, can only beused if the application programme iscompatible within the system.

Example: NICMOSLook andNICMOS grism extraction pipeline

A recent example of the use of IDL atESO/ST-ECF is the development of apipeline for the extraction and calibrationof NICMOS grism spectra. NICMOS(Near Infrared Camera and Multi-ObjectSpectrometer) was installed in theHubble Space Telescope (HST) duringthe 2nd Servicing Mission in February1997. Software is being developed at theSpace Telescope Science Institute (STS-cI) to ensure the proper calibration ofdata obtained with NICMOS. Software isalso being developed within the NICMOSInvestigation Definition Team (IDT, Prin-cipal Investigator Roger Thompson) tosupport the IDT science programme.

Among the operating modes of NIC-MOS is a grism spectrum mode. Whilethis mode is not being considered one of

the primary operating modes, it none-theless provides a very important scien-tific capability and has the highest po-tential for serendipitous discovery.

NicmosLook is an interactive tool toextract spectra for individual sources ona NICMOS grism image. A matchingdirect image of the same field is used todetermine the location of sources andtherefore the zero point for the wave-length scale for each individual spec-trum.

The tool is implemented as an IDLwidget. After loading both the grismimage and the direct image, both imag-es can individually be displayed andmanipulated. Objects can be located bythe user using the cursor, by supplyingobject co-ordinates or by supplying athreshold for an automatic objectsearch.

Spectra can be extracted for anyuser-selected object or for all objects atonce. The dispersion and distortionspectra are read from a database. Thereare several options for the spatialweighting of the spectra. The output ofNICMOSLook is wavelength calibratedspectra, which are plotted on the screenand can be saved as postscript files orASCII data files (Fig. 1).

NICMOSLook is a convenient tool toextract spectra from small numbers ofgrism images. However, the routine ex-traction of spectra from large numbers ofNICMOS grism images requires a toolwhich extracts spectra without humanintervention. Establishing the require-ments for such a software package, wequickly found that we did not have theresources to write it from scratch. On theother hand, it soon became evident thatmost of the individual steps needed forthe reduction have been coded before inone way or the other. The challenge wasto harvest as much as possible of thisavailable functionality and to combine itinto a package which would meet therequirements of minimum human inter-vention and operational resilience.

The resulting IDL programme, caln-icc, uses the SExtractor programme toidentify objects and classifies them asstars or galaxies using a neural networkapproach (Bertin & Arnouts, 1996). Thewavelength calibration of the extractedspectra is performed using the positionof the objects as determined by theSExtractor programme as the zeropoint, and using parameterised disper-sion relations. After background subtrac-tion and extractions of the spectra, they

Figure 1: NICMOSLook screen shot, showing the graphical user interface, display panels, and on-line documentation implemented in html.

15656

are corrected for the wavelength de-pendence of the quantum efficiency ofthe detector. The flux scale is thencomputed using the standard NICMOSflux calibration data. The extractedspectra are automatically checked forartifacts from bad data and contamina-tion from nearby objects. An attempt ismade to correct spectra for the contami-nation. All spectra are automatically

searched for emission and absorptionlines. In addition, the continuum emis-sion is determined. The final data prod-uct consists of binary FITS tables withthe spectra, error estimates, object pa-rameters derived from the direct imagingand details of the spectrum extractionprocess (Fig. 2).

IDL allowed the quick merging ofcomponents written by different authors,

including components written in differentlanguages. This was an important as-pect of our project as the object extrac-tion software is only available in C.

The software will be used for theextraction and calibration of NICMOSgrism spectra during the orbital verifica-tion and initial science programme ofNICMOS. We expect to have to doconsiderable changes in response to in-orbit performance. Once the software isstable and the data volume increases,we are in a position to speed up theprocessing by re-casting the software asan IRAF task and executing it in the HSTcalibration pipeline system, using theexisting software as the detailed design.

Conclusions

IDL in many ways combines the ad-vantages of an interactive and a com-piled language. The speed penalty men-tioned above is not significant consider-ing the speeds afforded today even bylow-cost hardware.

IDL is a traditional language as far assyntax and usage mode is concerned. Itis thus usable by astronomers, which isto say by users with a minimum compu-ter science background, who do noteasily adapt to state-of-the-art objectoriented paradigms. However, recentadd-ons to IDL make it possible for theprofessional programmer to use object-oriented concepts.

By using IDL, it was possible to devel-op a software package consisting ofabout 5000 lines of IDL and C code inabout 3 person-months. The total effort,including requirements definition, re-viewing, testing, documentation, and in-tegration was about twice that. This fast-track development, which was madepossible by the use of IDL, enabled us toproduce the required software on time.The portability of the code made deliver-ing it to the NICMOS IDT and the STScIvery easy. It was mainly due to thedemonstrated availability and capabilityof this software package that a majorNICMOS grism survey proposal wasapproved by the HST Time AllocationCommittee.

Reference

Bertin, E., Arnouts, S., 1995, A&A Suppl.,117, 393.

Figure 2: NICMOS grism spectrum extraction pipeline processing steps.

Rudi Albrechte-mail: ralbrech@eso.org

57

Gerhard Bachmann (1931–1996)Gerhard Bachmann, who for many years as Head of Administration played an essential role in the

development of ESO, died on 9 December 1996. Through his wisdom, dedication, integrity andpersistence many problems were solved, ESO’s relations with the member countries were excellentand ESO’s internal organisation functioned smoothly. He always had a willing ear for complaints fromthe staff and, if justified, would see to it that the necessary corrective action was taken.

Born in Berlin on 24 May 1931, he held various positions in the German civil service, before joiningthe NATO Maintenance and Supply Agency in France and Luxembourg (1963–1970). In 1970 hejoined the ESO Administration which he headed since 1972. He retired in May 1996. During this time,ESO has developed into a major European research organisation. Mr. Bachmann’s leadership hasdone much to inspire confidence among member state governments that ESO could be entrusted withsubstantially increased funds, that it had the management ability to deal with large projects, and that itcould conduct diplomatic negotiations at the highest level. He successfully established the PersonnelStatutes for International and Local Staff, the financial and contractual procedures of ESO, theadherence of ESO to many of the procedures of the Co-ordinated Organisations, even though thefickleness of the European governments led to deviations from these soon thereafter. All ESO’spensioners can be grateful to him for the way he integrated ESO into the CERN pension fund. Heplayed an essential role in the delicate negotiations about the increase of the membership of ESO,which led to Italy and Switzerland joining the Organisation. His diplomatic abilities in smoothing out

problems with the member states avoided many difficulties. More than anywhere this diplomacy was required in Chile, where he managedto maintain appropriate relations with governments ranging over the whole political spectrum. Unfortunately, he could not be presentanymore at the signing of the Acuerdo that put an end to a period of some difficulties between Chile and ESO.

Mr. Bachmann had a strong personality and at the same time a great loyalty to ESO. He strongly defended his point of view, but it wasalways possible to have a rational discussion and to come to a positive conclusion. Once arrived at, there was no doubt that it would be fullyrespected, even in the rare cases where his personal preference might have been slightly different.

Gerhard Bachmann always had time for what was needed, whatever the hour of the day or the day of the week. In the course of time webecame very good friends, and I remember with particular pleasure the many travels we made together, most of all the first discovery trip toParanal which turned out to have such far-reaching consequences. Between his various activities, Mr. Bachmann could perfectly relax, andI have fond memories of the many evenings we spent at the Casa Schuster at La Silla.

Mr. Bachmann was looking forward to spending more time at his home near Bordeaux that he had spent much effort to develop.Unfortunately, fate was different. His many friends in Europe and in Chile will miss him. All feel that he had merited better. Our sympathygoes to Mrs. Bachmann and his two daughters.

L. Woltjer

A N N O U N C E M E N T S

(7215) Gerhard = 1977 FSAsteroid discovered 1977 March 16 by H.-E. Schuster at the European Southern Observatory.Named in memory of Gerhard Bachmann (1931–1996), who came to the European Southern Observatory in 1970 and was head of

administration at the organisation from 1972 to 1996. Throughout this time he greatly contributed to ESO’s success. His unusual diplomaticabilities and thorough knowledge of the interaction of science and politics ensured the smooth running of the observatory and efficientinteraction with the member countries and the European Union. The naming marks his retirement from ESO and subsequent untimelydeath. Name proposed by the discoverer. Citation prepared by R.M. West.

ESO Studentship ProgrammeThe European Southern Observatory research student programme aims at providing the opportunities and the facilities to enhance the

post-graduate programmes of the member-state universities by bringing young scientists in close contact with the instruments, activities,and people at one of the world’s foremost observatories.

Students in the programme work on an advanced research degree under the formal tutelage of their home university and department,but come to either Garching or Vitacura/La Silla to do their studies under the supervision of an ESO Staff Astronomer. Candidates and theirsupervisors should agree on a research project together with the potential ESO local supervisor.

ESO has positions available for 12 to 14 research students. Students normally stay at ESO up to two years, so that each year a total of6–7 new studentships are available either at the ESO Headquarters in Garching or in Chile at the Vitacura Quarters and the La SillaObservatory. These positions are open to students enrolled in a Ph.D. programme in the ESO member states and exceptionally at auniversity outside the ESO member states.

The closing date for applications is June 15, 1997.For further information on the programme, the scientific interests of the ESO astronomers, and application forms please contact:

European Southern ObservatoryStudentship ProgrammeKarl-Schwarzschild-Str. 2D-85748 Garching bei München, Germanye-mail: ksteiner@eso.org

or look up: http://www.eso.org/adm/personnel/students

15858

F I R S T A N N O U N C E M E N T

ESO Workshop on

Cyclical Variabilityin Stellar Winds –

recent developments and future applications

14–17 October 1997

ESO HeadquartersGarching bei München, Germany

Variability is a fundamental property of stellar winds. In recentyears it has become clear that in many cases the observedvariations show a cyclical behaviour. This is a property that hot-and cool-star winds seem to have in common, although thephysical mechanism driving the wind is different.

Topics to be covered include:

• Wind acceleration mechanisms• Observations of cyclical wind variability (hot and cool stars)• Latest solar wind results• Variability in pre-main-sequence winds• Processes affecting the emergence of the wind• Modelling time-dependent behaviour stellar winds• MUSICOS 1996 results• Future developments

Scientific Organising Committee: T. Böhm (Germany), A.Cameron (UK), C. Catala (France), L. Hartmann (USA), H. Hen-richs (The Netherlands), L. Kaper (Germany), H. Lamers (TheNetherlands), K. MacGregor (Chair, USA), S. Owocki (USA), J.Puls (Germany), O. Stahl (Germany) Local Organising Com-mittee: A. Fullerton, L. Kaper (Chair), C. Stoffer.

European Southern ObservatoryKarl-Schwarzschild-Str. 2D-85748 Garching bei München, GermanyTel: +49-89-320060; FAX: +49-89-32006480Please contact: windvar@eso.org

Your Excellency, Don Eduardo Frei Ruiz-Tagle, Your MajestiesKing Karl XVI Gustaf and Queen Silvia of Sweden, Mr. President ofthe ESO Council, Mr. Director General of ESO, Senators, civil, mili-tary and church officials, Members of the Swedish delegation, Mem-bers of the ESO Council, ladies and gentlemen,

It is a great honour for me to represent the Chilean Governmentin this ceremony and to share with you my feeling of excitement andhope as we take part in this foundation ceremony of ESO’s at Cerro

1202. P. Ferruit, L. Binette, R.S. Sutherland and E. Pécontal: Mode-ling Extragalactic Bowshocks. I. The Model. AA.

1203. G. Mathys and S. Hubrig: Spectropolarimetry of Magnetic Stars.VI. Longitudinal Field, Crossover and Quadratic Field: NewMeasurements. AA.

1204. M. Victoria Alonso and D. Minniti: Infrared Photometry of 487Sources in the Inner Regions of NGC 5128 (Cen A). ApJ Suppl.

1205. I.J. Danziger and R. Gilmozzi: The Final Optical IdentificationContent of the Einstein Deep X-Ray Field in Pavo. AA.

1206. G. De Marchi and F. Paresce: The IMF of Low Mass Stars inGlobular Clusters. ApJ Letters.

1207. G. Mathys and T. Lanz: The Variations of the Bp Star HD137509. AA.

1208. P.-A. Duc, I.-F. Mirabel and J. Maza: Southern UltraluminousInfrared Galaxies: an Optical and Infrared Database. AA.

1209. A.A. Zijlstra, A. Acker and J.R. Walsh: Radial Velocities ofPlanetary Nebulae Towards the Galactic Bulge. AA.

1210. T.R. Bedding, A.A. Zijlstra, O. von der Lühe, J.G. Robertson,R.G. Marson, J.R. Barton and B.S. Carter: The Angular Diame-ter of R Doradus: a Nearby Mira-like Star. M.N.R.A.S.

1211. S. Randich, N. Aharpour, R. Pallavicini, C.F. Prosser and J.R.Stauffer: Lithium Abundances in the Young Open Cluster IC2602. AA.

1212. D. Baade and H. Kjeldsen: A Spectroscopic Search for HighAzimuthal-Order Pulsation in Broad-Lined Late F- and Early G-Stars. AA.

New ESO Preprints(November 1996 – February 1997)

Scientific Preprints

1198. J. Manfroid and G. Mathys: Variations of the Ap Star HD 208217.AA.

1199. L. Kaper, J.Th. van Loon, T. Augusteijn, P. Goudfrooij, F. Patat,L.B.F.M. Waters, A.A. Zijlstra: Discovery of a Bow Shock AroundVela X-1. ApJ Letters.

1200. R. Gredel: Interstellar CH+ in Southern OB Associations. AA.1201. J. Rönnback and P.A. Shaver: A Distant Elliptical Galaxy Seen

Through a Foreground Spiral. AA.

Paranal. This feeling is, first of all, based upon the importance of theproject that we inaugurate today. The VLT/VLTI telescope alreadydescribed by the President of the ESO Council and by the DirectorGeneral of ESO, is not only an expression of the most modern tech-nology ever used in astronomy. It is also an opportunity to deepenour knowledge of the Universe and thus be able to answer the ques-tions that have occupied mankind since its origin. In this sense, overand above the significant cost entailed by the project, it will be a

PERSONNEL MOVEMENTSInternational Staff (1 January – 31 March 1997)

ARRIVALS

EUROPEWICENEC, Andreas (D), Archive Information Designer &

EngineerWOLFF, Norbert (D), Control EngineerSILVA, David (USA), Head of User Support GroupBARZIV, Orly (GR), StudentROSATI, Piero (I), FellowDUDZIAK, Gregory (F), CoopérantPIEPER, Holger (D), UpA Optical Detector Team

CHILE

CANNON, Russell (GB/AUS), Senior Visitor

DEPARTURES

EUROPEVAN DEN BRENK, John, Detector EngineerVON DER LÜHE, Oskar (D), Experimental Physicist/

AstronomerSACRÉ, Philippe (F), Mechanical EngineerDUDZIAK, Gregory (F), StudentVAN LOON, Jacobus (NL), StudentPIEPER, Holger (D), StudentCÔTÉ, Stéphanie (CDN), Fellow

Translation of Speech by Foreign MinisterMr. José Miguel Insulza

59

milestone for the development of astronomy at a national and globallevel.

My satisfaction becomes even greater, when considering the longand winding path that Chile and ESO had to follow these last years, toreach this place and this occasion. We all know of the difficultiesencountered and the misunderstandings, which have now been happi-ly overcome, but which at some point threatened the completion of thisproject.

However, I believe it is important to say that during that process theGovernment permanently supported ESO, in order to enable the Or-ganisation to continue developing and extending its activities in Chile.We did this not only because of the international commitment under-taken by the State and to follow Chilean tradition in honouring engage-ments of international treaties but also because we thought that thedevelopment of ESO’s activities in the country, within a clear legalframe, would not only benefit world science, but also Chile’s scientificdevelopment in its most important regions.

Our point of view is validated today, since the Paranal Observatorywill not only benefit ESO’s activities but will also redound to the benefitof Chilean scientists who will thus have access to a facility which wouldotherwise have been completely beyond their reach.

For this reason, this ceremony, which follows the signing of theInstruments of Ratification and Approval of the Interpretative, Supple-mentary and Amending Agreement of the 1963 Convention, consti-tutes not only the end of a difficult process, but also the beginning of anew relationship of co-operation and understanding between Chileand ESO, which is a mature, consolidated and equal relationship inwhich both parties benefit.

The consolidation of the legal-political relationship between Chileand ESO warrants a steady development for ESO’s activities, within aframework of mutual benefit and also warrants the extension of theseactivities in the future.

Furthermore, through the recognition of the legitimate aspirationsof the national sectors more directly involved in ESO’s activities, anequal relationship is established, because the acknowledgement of

the rights of Chile and its citizens in labour and scientific mattersconstitute an adequate counterpart to Chile’s contribution to ESO.

By establishing permanent bodies of co-operation and for theresolution of controversies between Chile and ESO, the Organisationbecomes closely linked to the scientific and technological develop-ment of the country.

Finally, this relationship has pilot character in several aspects, suchas environmental protection, rights of Chilean workers in foreign publicorganisations and the rights of the national scientific community. Wetrust that it will set an example for future agreements related toastronomy, subscribed by the country in the next years.

In this context, I would like to express that the Government wel-comes the establishment of the „Foundation for the Advancement ofAstronomy”. This initiative was started by Chilean astronomers whohave obtained awards in sciences, and it is supported by the DirectorGeneral of ESO, Prof. Riccardo Giacconi. We believe that this founda-tion, complemented by the action of other organisations that bringChile and ESO together, will play an important role in administeringresources for the advancement of astronomy in Chile and will promoteinternational co-operation in this field.

Mr. President, Your Majesties, we are gathered in this impressivescenery to take part in a ceremony which has resulted from the strongand visionary co-operation between the Government of President Freiand ESO. We also take this opportunity to thank the many scientistswho have participated in this project with all their effort, the workerswho have accomplished it and also the Chilean Congress, represent-ed here by Senators of the Region, for the continuous support. Wehope that the Cerro Paranal Observation Centre may soon becomeoperational and that the work carried out on these premises, both byChilean and foreign guests, may redound to a greater knowledge ofthe Universe, as well as to clearer benefits for ESO, for its MemberStates, one of which is the Kingdom of Sweden, and for Chile. In thisspirit I would like to express my best wishes for the success of this newadventure in science.

Thank you very much.

Translation of Speech by H.E. President Eduardo FreiYour Majesties, Excellencies, Ambassadors, authorities, and a very

special greeting to the President of the ESO Council, Dr. Peter Creola,and the Director General, Prof. Giacconi, esteemed friend.

As I contemplate this enterprise, in the middle of the desert, madepossible through the joint effort of a group of European countries, theirscientific communities and astronomers, on one hand, and of theChilean nation, its workers, engineers and scientists, on the other, I feelnot only deeply gratified but, also, extremely proud.

This is a concrete signal of how our country is becoming increasinglyinserted in the international community. And not only in economicterms.

It is certainly true that our country is, nowadays, closely linked tointernational trade. This year we have completed our association withthe MERCOSUR, we have achieved strong ties with the EuropeanUnion and, for the first time in history, we have signed a Free TradeAgreement with an industrialised country: Canada.

But there are other aspects as well, where we are becomingpositively integrated to the rest of the world. Today, our political relationswith the community of nations are at a very high level, which would haveseemed unfeasible a few years ago. Democracy has brought us closerto the rest of the world and has restored the prestige historically enjoyedby the nation.

In the field of scientific co-operation we have also made progress.The establishment of this observatory is a good example of it.Astronomers from Europe and other countries will come and work here,and the most prominent Chilean astronomers will be given theopportunity of carrying out their research projects.

This opens multiple opportunities for collaboration and creates anextremely valuable platform for training young Chilean astronomers.Our main universities will benefit, as well as the country as a whole,since we will become the seat of one of the most advancedastronomical observatories in the world.

I believe that all these joint activities with industrially and scientificallyadvanced countries are vitally important, since our own challenge, asfar as modernisation is concerned, is the challenge of upgrading ourown capabilities in the fields of science and technology.

Chile would not be able to project itself creatively towards the future,nor to create a modern society, if economic growth and the generationof opportunities for people are not simultaneously coupled with theapplication of frontier technologies in production, in industry, inagriculture, in services and State management.

Perhaps this is one of the greatest challenges we face. Thesuccessful achievement of this goal calls for the joint effort ofuniversities, enterprises and the Government. We are working everyday towards this end, through the modernisation of education, which isthe very basis of the building we must construct. Without education,there would be no future, nor democracy, nor growth, nor competitive-ness and we would be unable to create a fairer and more equitablesociety.

Finally, I would like to take this occasion to publicly express ourappreciation to the ESO Member States, to the Organisation itself, itsexecutives and to all those in Chile who have contributed in this taskand to say that when I took office there were not only stones, but rockson the path, but there was also the political will to overcome them, sincewe could not allow hindrances in the construction of this greatenterprise.

At the same time, I would like to avail myself of the presence of TheirMajesties, the King and Queen of Sweden, to express our gratitude tothem. And also, to tell them that our academic community is deeplyindebted to their country, for the support provided to hundreds ofChilean researchers and professors, during their years of exile inSweden, as well as for the support given to a great number of nationalinstitutions that have sustained and developed in Chile, thanks tointernational collaboration, which was particularly generous and active,in the case of Sweden.

Finally, a special remembrance to the fact that in 1969 my father,President Eduardo Frei Montalva, inaugurated the La Silla Observa-tory, in the presence of a distinguished personality, former PrimeMinister Olaf Palme. The world is small and history is repeating itself.Today, with great joy we are saying “go ahead” to this great enterprisewhich points to the future and will foster development in our country.

Thank you very much.

16060

ESO, the European Southern Observatory,was created in 1962 to . . . establish andoperate an astronomical observatory in thesouthern hemisphere, equipped with pow-erful instruments, with the aim of furtheringand organising collaboration in astrono-my . . . It is supported by eight countries:Belgium, Denmark, France, Germany, Ita-ly, the Netherlands, Sweden and Switzer-land. It operates the La Silla observatory inthe Atacama desert, 600 km north of San-tiago de Chile, at 2,400 m altitude, wherefourteen optical telescopes with diametersup to 3.6 m and a 15-m submillimetre radiotelescope (SEST) are now in operation.The 3.5-m New Technology Telescope(NTT) became operational in 1990, and agiant telescope (VLT = Very Large Tele-scope), consisting of four 8-m telescopes(equivalent aperture = 16 m) is under con-struction. It is being erected on Paranal, a2,600 m high mountain in northern Chile,approximately 130 km south of Antofagas-ta. Eight hundred scientists make propos-als each year for the use of the telescopesat La Silla. The ESO Headquarters arelocated in Garching, near Munich, Ger-many. It is the scientific, technical and ad-ministrative centre of ESO where technicaldevelopment programmes are carried outto provide the La Silla observatory with themost advanced instruments. There are alsoextensive facilities which enable the scien-tists to analyse their data. In Europe ESOemploys about 200 international Staff mem-bers, Fellows and Associates; at La Sillaabout 50 and, in addition, 150 local Staffmembers.

The ESO MESSENGER is published fourtimes a year: normally in March, June, Sep-tember and December. ESO also pub-lishes Conference Proceedings, Preprints,Technical Notes and other material con-nected to its activities. Press Releases in-form the media about particular events.For further information, contact the ESOInformation Service at the following ad-dress:

EUROPEANSOUTHERN OBSERVATORYKarl-Schwarzschild-Str. 2D-85748 Garching bei MünchenGermanyTel. (089) 320 06-0Telex 5-28282-0 eo dTelefax (089) 3202362ips@eso.org (internet)ESO::IPS (decnet)

The ESO Messenger:Editor: Marie-Hélène DemoulinTechnical editor: Kurt Kjär

Printed byDruckbetriebe Lettner KGGeorgenstr. 84D-80799 MünchenGermany

ISSN 0722-6691

Contents

On January 12,1997, his HolinessPope John Paul II.graciously receivedin private audiencethe participants tothe internationalConference “theThree Galileos: theMan, the Space-craft and theTelescope”. ThePope was intro-duced to theastronomers andtheir families, andamong others tothe ESO DirectorGeneral, ProfessorRiccardo Giacconiand his wifeMirella.

R. Giacconi: Important Events in Chile .............................................................. 1Speech by the President of the ESO Council, Dr. Peter Creola ........................ 2Speech by the Director General of ESO, Prof. Riccardo Giacconi .................... 2Speech by the Minister of Foreign Affairs, Mr. José Miguel Insulza ................... 3Speech by H.E. President Eduardo Frei ............................................................ 5R.M. West: The Time Capsule ........................................................................... 5M. Tarenghi: Paranal, December 1996 .............................................................. 6

TELESCOPES AND INSTRUMENTATIONO. von der Lühe et al.: A New Plan for the VLTI ................................................. 8NEWS FROM THE NTT: J. Spyromilio ...................................................................... 14THE LA SILLA NEWS PAGE

S. Guisard: The Image Quality of the 3.6-m Telescope (Part V).What Happens far from Zenith? ............................................................. 15

From the 3.6-m and 2.2-m Teams ............................................................... 17J. Storm: 2.2-m Upgrade Plan ..................................................................... 17J. Brewer and J. Storm: News from the Danish 1.54-m Telescope ............. 17

S. Randich and M. Shetrone: New CASPEC Manual and Simulator ................. 18N. Devillard: The Eclipse Software ..................................................................... 19

SCIENCE WITH THE VLT/VLTIA. Renzini and B. Leibundgut: The VLT Science Case for the VLT ................... 21A. Renzini and L.N. da Costa: The ESO Imaging Survey .................................. 23

REPORTS FROM OBSERVERSN. Epchtein et al.: The Deep Near-Infrared Southern Sky Survey (DENIS) ...... 27T.R. Bedding et al.: R Doradus: the Biggest Star in the Sky .............................. 34A. Quirrenbach and H. Zinnecker: Molecular Hydrogen Towards T Tauri

Observed with Adaptive Optics ................................................................... 36N. Ageorges and J.R. Walsh: Examples of High-Resolution Imaging

and Polarimetry of R. Monocerotis and NGC 2261 ..................................... 39R. Mignani et al.: Proper Motion as a Tool to Identify the Optical

Counterparts of Pulsars: the Case of PSR0656+14 .................................... 43S. Savaglio: On the Nature of the High-Redshift Universe ................................ 45F.P. Nasuti et al.: Optical Observations Provide a New Measure of the

Vela Pulsar’s Proper Motion ........................................................................ 48Bavarian Prime Minister at La Silla .................................................................... 50

OTHER ASTRONOMICAL NEWSR. West and C. Madsen: The Astronomy On-Line Project ................................. 51R. Albrecht et al.: Availability of IDL at ESO/ST-ECF ......................................... 54

ANNOUNCEMENTSL. Woltjer: Gerhard Bachmann (1931–1996) ..................................................... 57(7215) Gerhard = 1977 FS ................................................................................. 57ESO Studentship Programme ............................................................................ 57First Announcement: ESO Workshop on Cyclical Variability in Stellar

Winds – recent developments and future applications ................................ 58Personnel Movements ....................................................................................... 58List of Preprints (November 1996 – February 1997) .......................................... 58Translation of Speech by Foreign Minister Mr. José Miguel Insulza .................. 59Translation of Speech by H.E. President Eduardo Frei ..................................... 59