arXiv:astro-ph/9812275v1 15 Dec 1998 al 1. Table 12 GGNCVCTDef T VCC NGC CGCG 07 4392Sc07 .314. G123 oB95 HG81 HH84 Co HH84 HI HH84 2.62 HI 2.38 HI GK83 2.60 HI 2.44 72.5 HG81 2.44 HG81 48.1 HH84 1 2.35 78.8 76.6 HH84 1 2.43 HH84 1 77.7 1 2.33 78.3 2.60 1 2.44 1.05 1 2.44 0.76 2.35 0.83 Sa 0.90 Sbc 958 1.09 Sab 912 1.24 4419 Irr 836 4413 Sab 630 99054 4388 Sab 570 70076 4330 559 70068 4313 70039 4312 70034 99029 09 4814 b12 .326. K32.67 GK83 67.4 2 2.43 1.25 Sb 1043 4438 70097 99 8013 c07 .324. G622 IHG86 T89 HL89 HI HH84 Co HI 2.68 2.23 HI 2.50 CG90 HG86 2.35 62.8 40.9 GK83 2.66 34.3 HL89 1 2 HH84 63.9 2 2.62 44.7 2.23 1 2.48 1 2.35 1.20 0.72 2.66 0.83 Sab 1.05 Sc 1690 0.87 Sb 1532 4569 1145 Sb I800 Sab 70192 1126 4457 99090 1110 I3392 42124 4450 99065 99062 28 5616 a10 .417. H424 IHH84 CG90 HI HI 2.52 2.44 HH84 2.84 HH84 77.8 72.5 CG90 3 36.5 1 2.60 1 2.44 2.84 1.35 1.08 Sbc 0.81 524 Sa Sab 1760 4307 1727 4586 70029 4579 42187 70197 06 3072Sb08 .515. H42.63 HH84 54.5 1 2.55 0.89 Sab 792 4380 70061 00 4518 .623 00H8 2.41 HL89 90.0 4 2.37 1.26 S 1086 4445 70104 06 5211 b .623 38H8 2.34 HH84 73.8 1 2.37 0.76 Sbc 1516 4522 70168 02 4919 Sa 1190 4469 70121 04 3673S .424 30H8 2.30 HG86 83.0 3 2.47 1.24 Sc 713 4356 70048 09 4498S .919 19G8 1.75 GK83 61.9 2 1.95 1.09 Sa 978 4424 70090 01 6615 a16 .635. H42.09 HH84 58.2 3 2.36 1.63 Sa 1859 4606 70213 1 2 3 4 5 6 7 8 9 1)(1 1)(13) (12) (11) (10) (9) (8) (7) (6) (5) (4) (3) (2) (1) yaia aaeeso Idfiin galaxies deficient HI of parameters Dynamical > .6-07. K32.59 GK83 71.9 0 - 1.76 HI o Wc log HI uliRflgW log Ref i Qual c H < H < rmRfComment Ref from H H H H H H H H H α α
o> Co o> Co α α α α α α α α α c 00RS MNRAS RAS, 0000 S95-KY88 CS85-K95 CS85 T.W. T.W. T.W. T.W. T.W. ..(o used) (not T.W. ..(o used) (not T.W. ..(o used) (not T.W. 000 000–000 ,
arXiv:astro-ph/9812275v1 15 Dec 1998 2. Table c GGNCVCTMm Def Memb T VCC NGC CGCG 93 3156S .967 A717 .312. G13.2Ai A s s B 30.82 A A B 31.70 30.34 HG81 A 29.84 32.05 N A 27.1 HG86 30.78 HH84 81.4 30.86 1 HH84 30.53 S W T.W. 77.7 78.3 1 77.8 2.73 HH84 N 1 32.19 30.94 HH84 1 78.8 2.52 B90 N - 65.4 1575 N 2.44 31.48 66.0 32.0 2.35 1 CA97 HH84 HH87 N 1254 2.52 30.65 1 31.04 N 6.79 5 47.8 1443 31.2 CA97 2.58 153 HH84 31.08 2.37 HL89 CA97 1092 9.32 0.29 2.61 1 W CG90 2 46.5 M 31.29 CA97 36.0 8.58 1150 HH84 N 0.17 78.5 T95 357 8.79 2.59 CA97 2.06 46.5 HU89 1 31.91 1.09 32.64 948 W 1 N 9.67 A HH84 30.36 7.92 1 CA97 57.1 1.24 1 2.55 CA96 B 1.35 2378 77.8 2.33 1524 32.72 9.65 M 30.86 0.29 HH84 2.74 HH84 1 A 7.44 Sc CA97 M CA97 2.57 CG90 -0.08 81.0 1 A Sbc 863 58.0 596 733 9.26 32.92 12.17 0.56 2.40 B 73.4 Sab HL89 140 576 32.85 A W86 CA97 -0.17 1 2.46 -83 4321 Sab 1 CA97 -0.11 570 68.1 A CA97 1 67.3 8.86 4316 Sbc 596 559 11.19 HH84 99030 A 2.45 CA97 2.61 4313 Sc 6.54 702 524 2 HH84 W -0.03 70035 47.2 0.46 W 2.68 CA96 1 8.26 4312 Sc 497 35.8 70034 0.46 CA97 4307 Sa 2064 9.64 2.65 465 2466 1 99029 0.48 Irr 2.46 4302 Sc 9.38 -132 p 460 CA97 1 A CA97 70029 0.01 324 4294 382 2.54 CA96 2082 10.08 99027 A 0.30 9.55 4293 369 2.60 Sbc A 70024 - 6.72 0.21 CA97 4273 Sc 0.14 m 99023 267 2207 CA96 9.31 Sb 226 0.13 m 42028 2168 CA97 8.95 Sc 167 I3115 0.20 m 42008 CA97 4237 Sc 10.38 M 157 41069 -0.17 9.84 4216 Sc A 152 0.05 98130 4212 Sc 145 W 69112 -0.28 Sc 4207 m 120 69110 Sb 97 4206 M 69107 Sb 92 4197 M 69104 Sc 73 4193 Sb 41052 66 4192 69091 Sc 58 4180 98108 25 4178 41048 I769 69088 4152 69083 98077 07 4392ScA07 .2C9 0 .8-4. 9 10 se A 30.2 S 31.55 CS85-K95 A 31.01 A 30. 30.93 67.4 30.29 S95-KY88 HL89 31.00 - 72.5 84.4 s HR89 A B B95 HH84 2.67 58.8 2 32.09 - 48.1 31.13 B 72.0 31.58 S90 1 B 2.24 2.62 70 - 71.1 31.19 1 SA82 31.73 A 2.37 HG81 HL89 31.89 2.38 CA96 1120 37.8 -273 1 2.45 B 78.8 83.3 A 7.17 CA96 HH84 CA97 30.48 1688 T.W. 1 2.31 105 11.36 57.1 1 S 31.71 7.80 1.25 234 1 CA97 54.5 31.31 S90 -0.05 2.39 CA96 9.72 1.05 HG81 1 2.60 CA96 -124 54.3 2.48 31.19 - 9.72 HH84 45.2 8.46 A -0.13 CA97 HH84 1103 m 2.42 1 75.0 2515 0.76 2.63 10.26 992 A 76.6 0.63 CA97 1 HW89 CA96 Sb 0.45 W 2.32 10.73 1 1710 64.4 Sc CA97 1 8.12 971 2.22 1043 A Sa 0.21 CA97 971 9.51 A 2.55 1 4438 Sc 1136 0.83 958 CA97 2.44 9.59 A 0.11 2324 4423 Sbc 957 CA97 8.48 70097 2.55 A -0.69 4419 Sc 1014 CA96 912 11.08 42107 1564 A 4420 Sc 0.89 873 CA97 10.45 99054 0.18 B CA97 4413 Sbc 1670 865 42106 9.09 0.38 A 4402 Sab 9.55 849 CA97 70076 B 4396 0.46 836 Sc m 70071 8.57 0.90 4390 Irr A 99049 827 4388 Sab -0.07 801 70067 W’ Sc 792 I3322 70068 A 4383 Sc 787 42095 W 4380 Sb 692 99044 4376 Irr 656 70061 4351 Sa 630 42093 4343 613 70045 4330 42070 4324 70039 42063 00RS MNRAS RAS, 0000 1 2 3 4 5 6 7 8 9 1)(1 1)(3 1)(15) (14) (13) (12) (11) (10) (9) (8) (7) (6) (5) (4) (3) (2) (1) :drvdprmtr flt-yegalaxies late-type of parameters derived a: 000 000–000 , HI H c e V Ref hel o W log c uliRef i Qual µ 7AseTab.1 see A 97 B 8 e Tab.1 see A 6 o einComments Region A B B 13
< Tab.1 see
e Tab.1 see Tab.1 see eTab.1 ee Tab.1 ee eTab.1 ee 0 Ceph. 30, Tab.1 e (16) 14
Table 2 – continued
CGCG NGC VCC T Memb DefHI Hc Ref Vhel log Wc Qual i Ref µo Region Comments (1) (2) (3) (4) (5) (6) (7) (8) (9) (10) (11) (12) (13) (14) (15) (16)
70104 4445 1086 S B 1.26 9.74 CA97 167 2.41 - 90.0 T.W. 31.26 A see Tab.1 99062 4450 1110 Sab A 0.87 7.03 CA96 1954 2.66 1 44.7 HH84 30.51 A see Tab.1 70111 4451 1118 Sc B 0.44 10.01 CA97 862 2.49 1 45.2 GK83 32.16 B 99065 I3392 1126 Sb A 1.05 9.31 CA97 1687 2.35 1 63.9 HL89 30.36 A see Tab.1 42124 4457 1145 Sb S 0.83 7.92 CA97 884 2.50 - 34.3 T89 30.15 S see Tab.1 70121 4469 1190 Sa B >1.76 8.19 CA97 508 2.59 - 71.9 T.W. 31.12 A see Tab.1 42132 4470 1205 Sc p -0.16 10.21 CA97 2339 2.35 1 40.9 HH84 31.26 S 42139 4480 1290 Sb W -0.02 9.78 T96 2438 2.58 1 58.1 HH84 32.64 S 99075 4498 1379 Sc A 0.35 9.78 CA96 1505 2.37 1 58.0 HH84 30.99 A 99077 I797 1393 Sc p 0.03 11.02 CA97 2091 2.32 1 46.5 HR89 31.83 A 99076 4501 1401 Sbc A 0.43 6.29 CA97 2284 2.80 1 56.6 HH84 30.87 A 70167 4519 1508 Sc B -0.18 9.65 CA97 1212 2.43 1 43.9 HH84 31.33 S 70168 4522 1516 Sbc B 0.76 9.74 CA97 2330 2.34 - 73.8 T.W. 30.71 S see Tab.1 99090 I800 1532 Sc p 0.72 10.82 CA97 2335 2.23 2 40.9 HG86 30.93 A see Tab.1 42156 4527 1540 Sb W -0.23 7.02 CA97 1736 2.60 1 71.1 HW89 30.03 S 42158 4532 1554 Irr B -0.16 9.49 CA96 2021 2.31 1 64.7 HH84 30.23 S 42159 4535 1555 Sc B 0.13 6.97 CA97 1962 2.62 1 42.4 HH84 30.67 S 14068 4536 1562 Sc W 0.07 7.64 CA97 1807 2.60 1 62.7 P79 30.65 S 42162 I3521 1575 Irr B 0.47 11.04 CA96 597 2.24 2 40.9 HL89 31.23 S 99093 4540 1588 Sc A 0.51 9.46 CA97 1288 2.35 1 35.8 S90 30.51 A 99096 4548 1615 Sb A 0.63 7.40 CA96 483 2.66 1 36.3 CG90 30.88 A 42168 4544 1624 Sc S 0.67 10.20 CA97 1151 2.34 1 68.4 HW89 31.17 S 70191 I3583 1686 Sm A 0.65 11.36 CA96 1121 2.13 1 52.3 H96 30.68 A 70192 4569 1690 Sab A 1.20 6.73 CA96 -212 2.68 - 62.8 CS85 30.37 A see Tab.1 70194 4571 1696 Sc A 0.55 8.71 CA97 342 2.56 1 27.1 HH84 31.41 A i<30, Ceph. 70197 4579 1727 Sab A 0.81 6.71 CA96 1512 2.84 1 36.5 CG90 31.61 A see Tab.1 42187 4586 1760 Sa S 1.08 8.64 CA97 792 2.44 1 72.5 HH84 30.40 S see Tab.1 99106 4595 1811 Sc A 0.11 10.20 CA96 632 2.31 1 47.8 HH84 30.94 E 42208 4630 1923 Sbc S 0.17 9.94 T96 742 2.34 1 42.7 HL89 30.91 S 99111 4633 1929 Sc A 0.51 10.67 CA96 291 2.32 1 64.7 S90 31.48 E 70230 4639 1943 Sb A 0.10 8.83 CA97 1048 2.62 1 44.4 HG81 32.00 E 71015 4647 1972 Sc A 0.41 8.65 CA96 1422 2.54 1 35.8 HH84 31.19 E 71019 4654 1987 Sc A -0.23 7.74 CA96 1039 2.60 1 51.4 HG81 30.75 E 71032 I3742 2023 Sc A 0.58 11.52 CA96 967 2.31 1 60.4 HL89 32.25 E 71045 4698 2070 Sa m -0.33 7.65 CA96 1008 2.71 1 57.2 HH84 31.53 E
B90: Bottinelli et al. 1990; B95: Boselli et al. 1995; CS85: Chincarini & de Souza 1985; CG90: Cayatte et al. 1990; GK83: Giovanardi, Krumm & Salpeter 1983; H96: Hoffman et al. 1996; HH84: Helou, Hoffman & Salpeter 1984; HH87: Hoffman et al. 1987; HG81: Helou et al. 1981; HG86: Haynes & Giovanelli 1986; HL89: Hoffman et al. 1989a; GK83: Giovanelli & Haynes 1983; K95: Kenney et al. 1995; KY88: Kenney & Young 1988; P79: Peterson 1979; T89: Thronson et al. 1989; SA82: Sulentic & Arp 1982; S90: Schneider et al. 1990; S95: Speraindio et al. 1995 W86: Warmels 1986;
c 0000 RAS, MNRAS 000, 000–000 Table c CGCG RS MNRAS RAS drvdprmtr ferytpegalaxies early typ e of parameters derived b NGC VCC E Sa S S E S E S S E E E Sa S S E Sa S E S S S S E S S E S S S E S a S S S E S S S S S E T Memb W W W W W W W M M A A A B A B A A A A B B A B A A A A B A A A B A A A A A A A B A A V hel T W M M M T W M T W M T W M M M T W M M T W M M T W M M T W M M M M M M M M M M T W M M M M T W T W T W M M Ref o r log eH e Seeing C T T T T T T C T T T C T T C T C C T T T T T T T T T T T T T T T C T T T T T T Ref 15 o A A A A A A B A A A S S A A A A S A A A A A A A A A A A A B S N A W W W W W M M Region M M al Table 16 CGCG continued NGC VCC S S S E S E S E E S S S S S E S E S S T Memb m m A A A A A A A A A A A A A A B B S V hel MEry McElroy M M M M M M M M M M T W M M M M M M M M M Ref o r log eH e c RS MNRAS RAS Seeing T T C T C T T T T T C T T T T T T T T Ref o Region A A A A E E E E E E E E E E E S S S S
arXiv:astro-ph/9812275v1 15 Dec 1998
al 3. Table c A 00RS MNRAS RAS, 0000 lu 723 4477 458 23 .3-5 275 -151 0.23 32.38 83 2435 8 75 437 94 2439 27 W Cloud lu 323 7 6 2 578 27 .1(76 210 (-736) 0.11 32.77 85 2507 4 126 666 173 2333 23 M Cloud A F T lu 56211638 5 5 09 .2-6 126 -768 0.12 30.94 157 659 9 89 653 121 642 35 N Cloud lu 816 8725 510 03.301 14139 -124 0.16 31.23 90 1304 15 56 702 88 1466 58 E Cloud lu 219 5674 317 4 09 .0+0 105 +202 0.10 30.91 140 1677 23 43 607 75 1595 62 S Cloud ig 6 345 6 55 39183.400 66 0 0.06 30.84 108 1369 51 35 762 54 1354 166 A Virgo ig 219 4575 218 13.401 7295 -762 0.10 31.84 51 1282 12 55 507 74 1293 62 B Virgo F T lu N Cloud F–––––2 35203.300 – – 0.09 30.83 200 1315 28 – – – – – TF A P–––––2 389 08 .8–– – 0.08 30.86 91 1388 23 – – – – – FP A nodef def 1 2 3 4 5 6 7 8 9 1)(1 1)(13) (12) (11) (10) (9) (8) (7) (6) (5) (4) (3) (2) (1) eie aaeeso h ig clouds Virgo the of parameters Derived VCC 414 3 08 .0–– – 0.10 30.82 235 1442 14 – – – – – 418 8 08 .0–– – 0.20 30.80 480 1180 14 – – – – – 000 V LG 000–000 , ± σ v ± N us V LG ± µ o ± V pec A ± 17
arXiv:astro-ph/9812275v1 15 Dec 1998 o.Nt .Ato.Soc. Astron. R. Not. Mon. .Gavazzi, G. band H from distance determinations. cluster Tully-Fisher Virgo and the Plane Fundamental of structure 3-D The 1 4 3 2 h pns ainlCmiso o Astronomy. for Commission National by Spanish operate the operated is Observatoire joi Alto is f¨ur (Heidelberg) the Max-Planck-Institut Astronomie Calar TIRGO the Italy. at (France). Firenze, and Arcetri, (CNRS), Observatory CAISMI-CNR, Provence Alto Haute Calar de the at land), at lie to ⋆ suggested were they and cluster, from M87 units main independent as the identified were substructure. the galaxies significant of (1961) contain northe Clouds to the Vaucouleurs in known cluster de was galaxy hemisphere, rich of nearest the work cluster, Virgo pioneering the Since INTRODUCTION 1 c nvri` el td iMln,dpriet iFisica, di dipartimento Milano, di Studi Universit`a degli a-lnkIsiu ¨rKrpyi,Psfc 090 D- 103980, Postfach f¨ur Kernphysik, Max-Planck-Institut Karl-Schwarzschild-Str. Syphon du Observatory, Traverse Southern European BP8, Spatiale Astronomie d’ Laboratoire ae nosrain ae tTRO(onrrt Switzer- (Gornergrat, TIRGO at taken observations on Based 00RAS 0000 1 .Boselli, A. 000 0–0 00)Pitd1 coe 08(NL (MN 2018 October 12 Printed (0000) 000–000 , sdrvdfo Isetafo h ieaue opeetdwith complemented literature, the rota from maximum dete spectra the is use HI galaxies we from late-type purpose derived Fun 75 this the as For another using method. galaxies of (TF) type distance Tully-Fisher early The 59 spec method. of new distances (FP) derive with we literature galaxies the 11 from measurements dispersion velocity eudrokasraepooer uvyo 0 aaisi h V the in galaxies 200 of survey to photometry plete surface a undertook We ABSTRACT uvso ihyH ecetglxe.Tezr-on fteF and Virgo FP of the modulus of distance zero-point the The assuming galaxies. calibrated deficient are HI lations highly 8 of curves µ a-ieada 200 at and far-side netite f03 T) .5(P a)w n httedistance the that find we mag) (FP) is M87, 0.45 determinations with distance (TF), associated 134 0.35 these of Using uncertainties method. Cepheids the with erdfo h euirmtosidcdo t iiiyi ossetw consistent is words: vicinity Key its on induced motions peculiar expectation. the from ferred ugsigta hyaefligot lse tapoiaey770 approximately at A cluster onto falling veloc different are and of significantly they North-West exclusively with that the almost but suggesting composed on A, clouds with Galaxies two consistent A. to distances of with belong distance A cluster the main twice the at are M and lses niiul Virgo. individual: clusters: o ⋆ 31 = 2 .Scodeggio, M. . 84 < B ± 0 aais udmna aaees–Glxe:selrcnet–Ga – content stellar Galaxies: – parameters fundamental Galaxies: 14 . 0 hssblse sfligot taot750 about at A onto falling is subcluster This 10. . a)cridoti h erIfae NR ad Combining band. H (NIR) near-Infrared the in out carried mag) 0 tywith ntly i eoi,1,213Mln,Italy Milano, 20133 16, Celoria, via 91 edleg Germany Heidelberg, 69117 ,D878Grhn e M¨unchen, Germany bei Garching D-85748 2, ms km by d rn -37 asil,France Marseille, F-13376 , µ 3 o − 30 = 1 .Pierini, D. rmtena-iersetvl.Tems fVroin- Virgo of mass The respectively. near-side the from rteHT(errs ta.19a a e eg 1996; Bergh den van 1996a; al. et (Ferrarese HST Cepheids 1994 the al. the galaxies et or with (Pierce 5 measured observations ground-based to been using have distance method cluster the Virgo proximity, the relative in their entities S to separate and are W’ Due a W, designations) Vaucouleurs that M, de cloud and namely (using M49, clouds, by individual dominated M87, of B, number containing cluster These A, from cluster. Cluster separated Virgo is that the establishing of helped structure works the toward the contributions of major understanding were Popes (1993) Binggeli, Tammann scale by and high follow-up spectroscopic with the (1985) and Tammann plates, ou and morphologi- carried Sandage visual survey Binggeli, of photographic by deep basis The the classification. on cal purely distances, larger . 84 ± 0 . 6 lse ,o-e otesuh sfudat found is south, the to off-set B, Cluster 06. 4 .Belsole E. A T E tl l v1.4) file style X 1 µ o 3.,a determined as =31.0, ms km rocutr(com- cluster irgo rsoi aafor data troscopic mnduigthe using rmined new ms km t distributions, ity wt individual (with Ftmlt re- template TF aetlPlane damental inlvelocity, tional − prlgalaxies spiral ot-atof South-East t h virial the ith fcutrA, cluster of 1 H − lusW Clouds . 1 α rmthe from rotation laxies: . cu ) t 2 G. Gavazzi et al.
Saha et al. 1997). The distance of 4 of these (N4321, appearance in the NIR, and this makes it easier to fit pho- 4496, 4536, 4571) was found consistent with 16 Mpc (cor- tometric profiles to derive the galaxy total magnitude, and responding to a distance modulus µo=31.0), while the re- the disk inclination. Both these factors should contribute to maining one (N4639) was found at a larger distance of 25 a reduction in the overall uncertainty with which distance Mpc (µo=32.0)(Sandage et al. 1996, Saha et al. 1997). Dis- determinations are obtained. tances determinations based on the Tully-Fisher (TF) re- The remainder of this paper is organized as follows. In Sec- lation (Tully & Fisher 1977) were used by a number of tion 2 we illustrate the sample selected for the present anal- groups to infer the 3-D structure of this cluster. Tully & ysis. The H band imaging observations and the new spec- Shaya (1984) and Pierce & Tully (1988) found evidence for troscopy are briefly summarized in Section 3. The cosmo- a significant sample depth along the line of sight, interpret- logical assumptions, and the derivation of the TF and FP ing their evidence as a presence of infall. Fukugita et al. templates are described in Section 4. Our new distance es- (1993) used their photographic photometry to confirm that timates, and a comparison with other estimates from the the cluster has a significant depth. More recently Yasuda literature are presented in Section 5, together with a discus- et al. (1997) re-addressed the issue, and concluded that the sion on the structure of the Virgo cluster as delineated by B, M and W clouds are at larger distance than Virgo itself. these new measurements. Our main conclusions are summa- Federspiel, Tammann & Sandage (1998) used new B band rized in Section 6. photometry to pursue the issue. They agree with Yasuda et al. (1997) in determining that W and M are distant clouds. They confirmed that cluster B is 0.5 mag further away than Virgo itself, although the mean recessional velocity of this 2 THE SAMPLE sub-cluster is identical to that of the dominant cluster A (Binggeli, Popescu & Tammann 1993). Galaxies analyzed in this work are selected from the 220 Beside the Cepheids and the TF relation, other methods Virgo Cluster Catalog (VCC; Binggeli, Sandage & Tam- have been used to estimate distances to the Virgo spirals. mann. 1985) objects with B < 14.0. We add to this sample These include the peak of the luminosity vs. declining ratio 13 objects, slightly fainter than this limit, that were taken of Novae (Ferrarese et al. 1996b), the expanding photosphere from the CGCG catalogue (Zwicky et al. 1961-68), because method for the type II supernovae (Schmidt et al. 1994), the distance estimates are readily available for them. For each peak brightness of the Type Ia supernovae (Schank 1997) galaxy a membership estimate is given in the VCC (revised and the bright stars as standard candels (Pierce et al. 1992) by Binggeli, Popescu & Tammann 1993). Cluster A coincides methods. The mean distance modulus derived from these with the X-ray cluster associated with M87. The subcluster methods is 30.92 ± 0.28. B is centered on M49. Other aggregates include the M, W Distance estimates to early-type galaxies have been based and W’ clouds and the Southern extension. Galaxies with −1 on the surface brightness fluctuation method (Tonry et al. Vhel > 3000 km s are treated as background objects, and 1990; Jensen et al. 1996; Morris & Shanks 1998; Ajhar et are excluded from the present analysis. al. 1997), on Novae luminosities (Della Valle & Livio 1995), Both the availability of the photometric and spectroscopic on the globular clusters luminosity function (Secker & Har- data necessary to build the TF and FP relation, and inclina- ris 1993; Whitmore et al. 1995; van den Bergh 1996), and tion restrictions that apply to any TF sample play a role in on the planetary nebulae luminosity function (Jacoby et al. limiting the number of galaxies for which we can obtain dis- 1990; Ciardullo et al. 1998). These methods give distances in tance estimates. We have obtained NIR H or K’ band images the range 30.31-32.13. M87 itself has been found at µo=30.8- of 200 out of the 233 selected galaxies (these observations 31.3. M49 poses a problem, since its distance determination are described in Section 3). Of the 89 early-type galaxies (µo=31.0) is significantly smaller than the distance estimate imaged, 48 have central velocity dispersion measurements to cluster B (µo=31.8, Federspiel et al. 1998) to which it is available in the literature (McElroy 1995). To these we add supposed to belong. 11 new measurements of the central velocity dispersion ob- Altogether there appears to be unanimous consent that tained at OHP (see Section 3.4). galaxies in the Virgo region are not at a unique distance. Of the 111 spiral galaxies imaged, 75 match the criteria for The W, M and B clouds are found further away than A, the inclusion in the TF sample, i.e.: a) have the correct in- and evidence for infall has been reported. clination (i> 30◦) (see Section 4.2); b) have high signal-to- In this paper we make a further step in the direction of noise 21cm HI line profiles (or Hα rotation curves) and, c) unveiling the structure of the Virgo cluster, with two main meet the additional requirement that the line widths cor- improvements over previous works. Firstly, we combine dis- rected for inclination exceed 100 km s−1). The HI line pro- tances to spirals obtained with the TF method with dis- files in the literature were individually inspected. If more tances to E/S0s obtained with the Fundamental Plane (FP) than one measurement was available we choosed the one method (Djorgovski & Davis 1987; Dressler et al. 1987), with higher S/N ratio (references to these measurements are thus increasing the statistical significance of the determina- listed in Table 2a). tion. Secondly, for both of these methods we rely on H band NGC 4496, one of the 5 galaxies with Cepheids measure- surface photometry. The advantages of using near-infrared ments, is not included in the present analysis because an- (NIR) photometry over the B band one are that NIR mag- other galaxy is superimposed on its disk, making the ex- nitudes are less sensitive to recent episodes of star forma- traction of its photometric parameters uncertain. Conversely tion, and they are less affected by internal extinction, thus two other galaxies with Cepheids (NGC 4321 and 4571) are they better trace the luminous matter in galaxies (Gavazzi, included in the present work in spite of their low inclination Pierini & Boselli 1996b). Also, spiral disks have a smoother (27 degrees).
c 0000 RAS, MNRAS 000, 000–000 The 3-D structure of the Virgo cluster 3
3 OBSERVATIONS The observations were calibrated and the fluxes transformed into the H band photometric system using the standard stars 3.1 NIR imaging observations in Elias et al. (1982), observed hourly throughout the night. In various runs from 1994 to 1997 we took NIR H (1.65 The typical uncertainty of the measurements is 0.05 mag. µm) or K’ (2.1 µm) band images of 206 of the 233 galax- Details on the image analysis and extraction of the photo- ies selected from the VCC and CGCG, using the TIRGO metric parameters can also be found in Boselli et al. (1997). 1.5 m and the Calar Alto 2.2 m telescopes. Eight galaxies Here it is sufficient to say that the H magnitudes used were not observed because they have velocity larger than in the present analysis were derived by simulating aper- − 3000 km s 1, thus they do not belong to the cluster, and ture photometry, following the method of Gavazzi & Boselli four because their angular size (a> 4 arcmin) exceeds that (1996): the counts are integrated in concentric circular rings of the available NICMOS3 detectors. The remaining 15 ob- around the galaxy centres to provide curves of growth up jects were not observed due to time limitations. Images of to the galaxy optical diameter (determined in the B at the th −2 6 galaxies are not usable, because the frames were ruined 25 mag arcsec ). The H magnitudes used in this work by stray light from the moon. Of the 200 images available, are corrected for internal extinction using the prescriptions 166 were taken in H band, and 34 in K’ band. The K’ data of Gavazzi & Boselli (1996). For 1053 galaxies (not necessar- ′ are converted into H band using
c 0000 RAS, MNRAS 000, 000–000 4 G. Gavazzi et al.
The basic image reduction was performed using standard was instead detected in absorption along with the Ca[I] routines in the IRAF†-STSDAS‡-PROS environment. The (λ 6492.5 A),˚ providing us with a useful velocity dispersion bias-subtracted, flat-fielded, combined, and calibrated im- measurement. ages were analyzed using the package GAPLHOT (devel- The blue grism was selected to give a spectral coverage in oped for IRAF- STSDAS mainly by W. Freudling, J. Salzer, the region 5050 − 5505 A,˚ containing the redshifted Mg[I] and M. Haynes, and adapted by one of us (M.S.) to per- triplet (λλ 5167.3−5172.7−5183.6 A)˚ and the CaFe doublet form the light decomposition of early-type galaxies). For (λ 5269.0 A).˚ With the blue grism we obtained dispersion each frame the sky background was determined as the mean measurements of 10 early-type galaxies. Each galaxy was number of counts measured in regions of “empty” sky, and it observed with the slit positioned E-W, with an integration was subtracted from the frame. Sky-subtracted frames were time varying from 20 to 45 min, depending on the brightness inspected individually and the light of unwanted superposed of the galaxy, enough to produce a signal-to-noise ratio of ∼ or nearby stars and galaxies was masked. The 2-dimensional 15 per pixel at Mg[I]. Each observation was preceded by an light distribution of each galaxy was fitted with elliptical exposure of a Ar-lamp to ensure the wavelength calibration. isophotes, using a modified version of the STSDAS isophote The calibration was checked using known sky lines which package. The fit maintains as free parameters the ellipse cen- were found within 0.1 A˚ from their nominal wavelengths. −1 ter, ellipticity and position angle, and the ellipse semi-major The estimated redshift uncertainty is 10 km s . axis is incremented by a fixed fraction of its value at each The rotation curves of 8 galaxies were obtained by deter- step of the fitting procedure. Using the fitted parameters mining the central wavelength of a gaussian fit, pixel per a model of the galaxy light distribution is obtained, which pixel along the slit, to the redshifted Hα. These are given in is used to compute integrated magnitudes as a function of Fig. 1 and their parameters are listed in Table 1. semi-major axis. The velocity dispersion measurements of 13 galaxies were The effective radius re and effective surface brightness µe obtained using the Tonry & Davis (1979) method. This (the mean surface brightness within re) of each galaxy were method is based on a ”comparison” between the spectrum of obtained by fitting its radial surface brightness profile with a a galaxy whose velocity dispersion is to be determined, and de Vaucouleurs r1/4 law (de Vaucouleurs 1948). The fit was a fiducial spectral template of a star of appropriate spec- performed from a radius equal to twice the seeing radius, tral type to contain the wanted absorption lines. The basic out to the outermost isophotes for E galaxies; for S0 and assumption behind this method is that the spectrum of an S0a galaxies only the central core was fitted. The median elliptical galaxy (and also of the bulge of a disk galaxy) is uncertainty on the determination of log re and µe is 0.05 well approximated by the spectrum of its most luminous and 0.16 mag., respectively. stars (K0-K1 giants), modified only by the effects of the stellar motions inside the galaxy. We observed three stars of K type: HD26162, HD132737 and HR5361 with integration 3.4 Long-slit spectroscopy time of 60 sec, to secure a spectrum with sufficient signal- to-noise ratio. The heliocentric velocities of these stars are − To obtain high dispersion spectra of 21 target galaxies we known with few km s 1 accuracy. The quantitative mea- used the 1.93 m telescope of the Observatoire de Haute surements of the spectra have been obtained using the cross- Provence (OHP), equipped with the Carelec spectrograph correlation technique of Tonry & Davis (1979), implemented (Lemaitre et al. 1990) coupled with a 512×512 pixel Tek- in the IRAF task fxcor (see more details on the application tronix CCD. of this technique in Scodeggio 1997). Since the velocity dis- The observations were carried out in the nights of Febru- persion inside an elliptical galaxy veries as a function of ary 26 - March, 5, 1998 in 2-2.5 arcsec seeing conditions. radial distance from the center, we extracted from our data The spatial resolution is 1.17 arcsec per pixel. The slit monodimensional spectra of the central 6 arcsec. The mea- was 1.83 arcsec wide. We used two grisms, both with surements were then corrected to the value that one would ˚ a spectral resolution of 33 A/mm. The red grism was obtain at the distance of the Coma cluster, using the rela- chosen to give a spectral coverage in the region 6475 − tion derived by Jørgensen, Franx & Kjærgaard (1996) 6930 A,˚ containing the redshifted Hα (λ 6562.8 A),˚ the logσ/σ6” = + 0.04 log dV irgo/dComa [NII] doublet (λλ 6548.1, 6583.4 A)˚ and the [SII] doublet where σ6” and σ are the measured and corrected velocity (λλ 6717.0, 6731.3 A).˚ This setup was selected to obtain Hα dispersion, dV irgo and dComa are the respective distances, rotation curves of 11 spiral galaxies with large (> 0.7) HI assumed 16 and 88.3 Mpc. The velocity dispersion measure- deficiency. ments obtained in this work are listed in Column 7 of Table Each galaxy was observed with the slit parallel to the major 2b along with their internal errors. The total uncertainty axis, with an integration time of 15 min, ensuring a signal- can be obtained by adding in quadrature the systematic un- to-noise ratio of ∼ 10-15 per pixel at Hα. certainty of 10 km s−1. Unfortunately two of the newly ob- For three of these objects (70099, 70115, 70149) the Hα tained dispersion measurements (42026 and 70100) are not used in the present analysis because these galaxies have no H band surface photometry. † IRAF (Image Reduction and Analysis Facility) is distributed by NOAO, which is operated by the Association of Universities for Research in Astronomy, Inc. (AURA), under cooperative agree- 3.5 The T.F. parameters of HI deficient galaxies ment with the National Science Foundation. ‡ STSDAS(Space Telescope Science Data Analysis System) is dis- Some 22 spiral galaxies included in the present analysis have tributed by STScI, which is operated by AURA, under contract exceedingly large (> 0.7) HI deficiency that could produce to the National Aeronautics and Space Administration. systematic underestimates of their widths, thus of their TF
c 0000 RAS, MNRAS 000, 000–000 The 3-D structure of the Virgo cluster 5 distances. We inspected these HI profiles individually and The infall velocity of the Local Group (LG) toward Virgo is − we searched in the literature for alternative maximum rota- taken to be 220 km s 1 (see Federspiel et al. 1998); veloc- tional velocity measurements, either from CO observations ities with respect to LG centroid, VLG, are thus computed (Boselli et al. 1995; Thronson et al. 1989; Kenney et al. 1995; as Vhel+220. Kenney & Young 1988), or from Hα rotation curves (Chin- Distances to Virgo cluster early- and late-type galaxies are carini & de Souza 1985; Speraindio et al. 1995; this work). computed using the FP and TF relation, respectively. Tem- Table 1 reports detailed information on the rotational prop- plates for these relations are obtained from other well stud- erties of these galaxies, as follows: ied clusters, scaled according to the relative distances on the Columns 1-3: CGCG, NGC, VCC designations. basis of the above assumptions. Column 4: morphological type as given in the VCC. Column 5: HI deficiency parameter as defined by Haynes & Giovanelli (1984). 4.1 The Fundamental Plane relation Column 6: logarithm of the maximum rotational velocity, An H band FP template was derived from a sample of 74 corrected for inclination, as derived from HI observations (average of 20% and 50% of the peak value). galaxies in the Coma cluster by Scodeggio et al. (1998a). Column 7: a quality mark given to each HI profile after in- The best fit to this relation, obtained assuming that Coma is at rest in the CMB reference frame (see, for example, dividual inspection: 1 are two-horn, high S/N profiles, 2 are one-horn, high S/N profiles, 3-4 are profiles of insufficient Giovanelli et al. 1997, Scodeggio et al. 1997a), is: logRe = − quality for T.F. work. 8.354 + 1.52logσ + 0.32µe (with the zero-point adapted to H =81.35). The scatter of the template relation, thus the Column 8: galaxy inclination in the plane of the sky (deter- o mined following Haynes & Giovanelli 1984). uncertainty on the distance modulus determination of a sin- Column 9: reference to the HI data. gle galaxy is 0.45 mag. The fit is obtained minimizing the weighted sum of the or- Column 10-12: logarithm of the maximum rotational veloc- ity, corrected for inclination, as derived from CO observa- thogonal distances of the data points from the plane. This is a generalization to 3 dimensions of the maximum likeli- tions or Hα rotation curves, with references. hood method of Press et al. (1992) (their “fitexy” routine), If the three available measurements are found in agreeement we use the HI maximum line widths. Alternatively we use with a modification introduced to take into account the high degree of covariance shown by the uncertainties on the deter- combinations of CO and Hα measurements (when both are mination of log Re and µe (see Scodeggio 1997 for details). available, as for 99054 and 70097) or we adopt the Hα val- ues alone, if the data extend sufficiently to ensure that the Uncertainties on the FP parameters are determined using plateau of the rotation curve is reached. the statistical jackknife: N sub-samples, each one composed of N–1 data-points, are extracted from the original sample of For three galaxies (70048, 70090 and 70213) observed in this N data-points, rejecting in turn one of the data-points. The work the Hα data do not trace the rotation curve far enough to see the classical S shape, indicative that the plateau of distribution of a certain statistical parameter among those N sub-samples is then used to estimate the uncertainty in the rotation curve has been reached. Their Hα maximum the value of that same parameter, without having to assume rotational velocity is smaller than that obtained from the HI line (although of poor quality). These galaxies have been an a-priori statistical distribution for the parent population omitted from the analysis. of the data-set under examination (see for example Tukey 1958, and Efron 1987). One object (70104) has a rotaton curve with marginal change of concavity. However its maximum velocity determi- The FP relation of early-type Virgo galaxies is given in Fig. nation exceeds that available from the HI profile. We decided 2a superposed to the fit of the Coma template relation. The points are coded according to wether σ is larger or smaller to rely on the Hα measurement. The remaning 4 objects −1 observed at OHP show a clear S shaped rotation curve. For than 100 km s . It is known that uncertainties in the mea- surement of such low velocity dispersions are larger than these we adopt the maximum velocity obtained from our Hα in the case of galaxies with higher velocity dispersion (e.g. measurement. Scodeggio et al. 1998b). In fact the dispersion of Virgo galax- ies around the template relation is 0.75 mag if σ < 100 are excluded, significantly larger than that of the template rela- 4 THE METHOD tion. If these 12 galaxies are included, the dispersion rises to 0.85 mag, enough to suggest to drop galaxies with σ < 100 We assume that the average distance of the Virgo cluster A from the following analysis. (the membership is taken from Binggeli, Popescu and Tam- mann 1993) is 16 Mpc (µo=31.02), as derived from recent primary distance determinations of 4 HST galaxies with 4.2 The Tully-Fisher relation Cepheids (van den Bergh 1996). We also assume that the difference of distance moduli between Virgo cluster A and The TF template was derived combining 73 galaxies from Coma is 3.71 mag; thus µo coma=34.73 or Dcoma=88.3 Mpc the clusters A262, Cancer, Coma and A1367. The clusters (van den Bergh 1996). were considered to be at rest in the CMB reference frame. Given that the average recessional velocity of Coma cor- The galaxies were selected from the CGCG (Zwicky et al. rected for the motion with respect to the Cosmic Mi- 1961-68) (thus with mp ≤ 15.7) according to slightly more −1 crowave Background (CMB) is
c 0000 RAS, MNRAS 000, 000–000 6 G. Gavazzi et al. large number of sources. It is important to stress here that Column 11: a quality mark given to each HI profile after in- the method used for determining the H magnitudes of the dividual inspection: 1 are two-horn, high S/N profiles, 2 are template galaxies is identical to that used for the Virgo one-horn, high S/N profiles, 5 is for one profile which was galaxies (see Section 3.2). Galaxies with DefHI > 0.5 were not published, but was considered of good quality given the not used, because the HI line-width in this case could un- high S/N ratio. derestimate the galaxy true rotation velocity. Also galaxies Column 12: galaxy inclination in the plane of the sky (de- with logWc < 2.2 were not used, to avoid objects whose termined following Haynes & Giovanelli 1984). correction for turbulent motion is relevant. We assume that Column 13: reference to the adopted rotational velocity. the uncertainty on the measurements of the line width is Column 14: distance modulus as obtained in this work using 10 km s−1 and that on H band magnitudes is of 0.15 mag. the TF method. The TF relation for the 73 galaxies used in the derivation of Column 15: revised region of membership (see Section 5.3 the template is shown in Fig. 2c. The best fitting template, for definition). obtained using the bivariate method (Giovanelli et al. 1997) A comment is given in Column 16 for the deficient galax- is given by: H = −2.60 − 7.85 LogWc. The scatter of the ies whose maximum rotational velocity is derived from CO TF template, thus the uncertainty in the distance modulus profiles or Hα rotation curves (see Tab. 1). of the individual galaxies is 0.35 mag. The the zero point and slope of the TF template relation Tab. 2b lists the FP parameters of 59 early-type galaxies are found consistent with the parameters of the local TF as follows: calibrators. Columns 1 to 5 contain the same information as Tab. 1a. The TF relation of Virgo late-type galaxies is given in Fig Column 6: heliocentric recessional velocity. 2b, with the template relation superposed. The scatter of Column 7-9: central velocity dispersion with error (given Virgo galaxies is 0.70 mag, thus significantly larger than for the OHP measurements only) and reference. McElroy that of the template relation. We have checked if to this (1995) (M95) did a compilation of all dispersion measure- larger scatter might contribute a not enough conservative ments available at the time. Not only the values are averages threshold on galaxy inclination (i > 30 deg). However we of various sources, but they are also corrected for aperture found that for galaxies with inclination i> 45 deg the scat- according to the prescriptions of Jørgensen, Franx & Kjær- ter remains 0.70 mag. Meanwhile we determined that the gaard (1996). Our data (T.W.) are taken through a 2”x6” scatter is not affected by the inclusion of galaxies with HI aperture, corrected similarly (see Section 3.4). deficiency larger than 0.7. In Figs 2a and 2c all Virgo cluster Columns 10 to 15 contain the H band parameters (see Sec- galaxies are assumed to have a distance modulus equal to tion 3.3). the average distance modulus of Virgo cluster A (µo=31.02). Column 10-11: Log of the H band effective radius re, cor- Templates shifted by ± 1 mag are also given. The deviations rected for seeing according to the prescriptions of Saglia et of the individual objects from the template relations can be al. (1993), with uncertainty (in arcsec). converted into distances to the individual galaxies. Column 12-13: corrected H band effective surface brightness − with uncertainty (in mag arcsec 2). The correction includes the cosmological expansion (1+z)4 and K-correction (taken to be proportional to 1+z) terms, and the seeing correction, 5 RESULTS according to Saglia et al. (1993). No galactic absorption cor- rection was applied since AH = 0.085 AB (Pahre et al. 1995), 5.1 Derived distances with AB ≤ 0.1 mag in the direction of Virgo. The distances to 134 galaxies in the Virgo cluster are listed Column 14: seeing during the H band observations. in Tab. 2a and b. Column 15: NIR observing run: T97 refers to the 1997 run at Tab. 2a lists the TF parameters of 75 late-type galaxies as TIRGO (this work), 9 objects were serendipitously observed follows: at Calar Alto (CA94-96) in K’ band and are found in Boselli Columns 1-3: CGCG, NGC, VCC designations. et al. (1997); one of the T97 objects was also serendipitously Column 4: morphological type as given in the VCC. observed at Calar Alto in 1997 in H band. Column 5: membership according to the VCC (revised by Column 16: distance modulus as obtained in this work using Binggeli, Popescu and Tammann 1993). the FP method. Column 6: HI deficiency parameter as defined by Haynes & Column 17: revised region of membership (see Section 5.3 Giovanelli (1984). for definition). Column 7: total H band magnitude corrected for internal extinction (see Section 3.1). Column 8: NIR observing run: CA96 refers to Boselli et al. 5.2 Comparison with independent distance (1997), CA97 are the observations taken in 1997 at Calar estimates Alto (see Section 3.1), T95 and T96 are found in Gavazzi et al. (1996a). Because many different methods have been applied to derive Column 9: heliocentric recessional velocity from the litera- distance estimates to Virgo galaxies, a number of consis- ture. tency checks can be made between our TF and FP distance Column 10: adopted logarithm of the maximum rotational determinations and those obtained independently by other velocity, corrected for inclination, as derived from HI obser- groups, either with the same or with different methods. Fig vations (average of 20% and 50% of the peak value) or from 3a shows the comparison of our own distance estimates with CO observations or Hα rotation curves (see Tab. 1). those in the literature, obtained with methods other than
c 0000 RAS, MNRAS 000, 000–000 The 3-D structure of the Virgo cluster 7 the TF or FP relation: Cepheids, type I and II supernovae, frame (approx. 4 arcmin), thus affecting the estimate of the planetary nebulae luminosity function, novae, globular clus- sky background. Conversely the distance of M49 found in ters luminosity function, and surface brightness fluctuations. this work µo=31.18 is in excellent agreement with Tonry et The rms scatter in the distance modulus differences, when al. (1990) (µo=31.18) and consistent with the most recent measured for E/S0 galaxies alone, is larger (0.40 mag) than determination by Morris & Shanks (1998) (µo=30.97). that measured for the spiral galaxies (0.33 mag), even if we exclude objects with σ < 100 km s−1. This latter compari- son is based on distance estimates obtained using Cepheids 5.3 The 3-D Structure of the Virgo Cluster and SNI and II methods. Assuming that these methods have The newly obtained distance moduli are plotted against VLG an average intrinsic accuracy of 0.2 mag, this implies that in Fig. 4, where small dots mark the early type galaxies (FP) our TF uncertainty is ∼ 0.25 mag. separated into σ < 100 km s−1 and σ > 100 km s−1 and The comparison for E/S0 galaxies is based on distance esti- larger circles mark the spirals (TF), separated into normal mates obtained using the following methods: surface bright- and HI deficient. The dashed line represents the Hubble flow, ness fluctuation (Tonry et al. 1990; Jensen et al. 1996; Morris drawn for Ho=81.35. Distance moduli derived with the TF & Shanks 1998; Ajhar et al. 1997), Novae luminosities (Della and FP relation have similar distributions and mean values. Valle & Livio 1995), the globular clusters luminosity func- HI deficient objects have TF distances not different from tion (Secker & Harris 1993; Whitmore et al. 1995; van den galaxies with normal HI content. Conversely the FP plane − Bergh 1996), and the planetary nebulae luminosity function distances of galaxies with σ < 100 km s 1 are significantly (Jacoby et al. 1990; Ciardullo et al. 1998). Assuming once smaller and with higher rms than those with larger disper- again that these methods have an average intrinsic accuracy sion. Notice that the 4 galaxies with the smallest distance of 0.2 mag, this implies that our FP accuracy is ∼ 0.35 mag. estimates are among the objects with the smallest velocity − Fig. 3b compares our TF distances with the recent B band dispersions (σ ∼ 50−80 km s 1). Accordingly their distance TF compilation of Yasuda et al. (1997). The rms scatter is estimate uncertainties must be quite large, up to the point 0.47 mag, which would correspond to an accuracy of each that we cannot trust them. As anticipated in Section 4.1, separate measurement (assuming equal contributions to the these galaxies are thus excluded from the rest of the analy- scatter) of roughly 0.3 mag. This is in good agreement with sis. the internal accuracy estimate of Yasuda et al. (1997), but Three arbitrary regions are evidenced in the figure: it is surprisingly larger than the accuracy estimate derived µo < 31.5 and µo > 31.5. The latter is split in two velocity − above from the comparison with non TF-based distance es- regions: above and below 1900 km s 1. The region delim- timates of S galaxies, which should be contributed to by ited by 30 < µo < 31.5 contains the majority of galaxies at both measurement uncertainties and intrinsic variations in the distance to the Virgo cluster. This region contains the galaxy properties. majority of HI deficient galaxies. Unfortunately a similar comparison cannot be made with A better insight into the structure of the cluster is ob- the Federspiel et al. (1998) distances, which were not pub- tained with Fig. 5, which reports the same data as Fig. lished individually. 4, but with symbols whose size decreases with increasing Distances obtained in this work with the FP method (σ > distance, and whose shape refers to 7 regions of the Virgo −1 100 km s ) are compared in Fig. 3c with those obtained cluster, represented in Fig. 6 with dotted lines (galaxies with − in the B band by Dressler et al. (1987) and Dressler, (1987) σ < 100 km s 1 are omitted). Galaxies are assigned to these with the Dn-σ method. The agreement is not very satis- regions with a criterion that combines their position on the factory, with a sistematic offset of 0.10 mag and an rms sky with their distance and redshift. scatter of 0.40 mag. Since the accuracy of the B band Dn-σ Distant (µo > 31.5) galaxies with VLG > 1900 belong al- method is expected to be slightly worse than that of the FP most exclusively to the regions marked M and W. These method, we conclude that the observational scatter of our correspond to the M and W clouds, which thus are found to FP distance determination is at most 0.3 mag. be in Hubble flow. A test for the internal consistency between our TF and FP Distant (µo > 31.5) objects with VLG < 1900 mostly be- distance determinations can be done using the interacting long to the region marked B. This corresponds to Cluster system NGC 4435 (S0) + NGC 4438 (S). The distance mod- B of Binggeli et al. (1993), but with a smaller extent than uli of the two galaxies (30.59 and 30.26 respectively) agree in the original definition. The mean distance modulus for within the errors. Similarly, the pair NGC 4649 (S0) + NGC this structure is found to be 31.84, in good agreement with 4647 (S) has consistent distance moduli (31.37 and 31.19 re- the determination by Federspiel et al. (1998) (µo = 31.8). spectively). A test of the reproducibility of our results can To the East of approx α = 12h25m (M49) all galaxies have be done with the galaxy NGC 4261, which was observed distances not dissimilar from those of Cluster A (µo ∼ 31). both at TIRGO in 1997 and, serendipitously, at Calar Alto For example 4 spiral galaxies (NGC 4519, 4332, and 4535, in 1997; the difference between the distance moduli based and IC 3521) which are assigned to Cluster B by Binggeli on the two separate observations is 0.15 mag. et al. (1985) are not confirmed to lie at significantly larger A discrepancy is found between our FP-based determina- distance than Cluster A, since they are found at µo=30.2- tion of the distance of M87, and other determinations found 31.3. For these objects our TF distance determinations are in the literature. Our value µo=31.62 is larger than 31.30 in agreement with those of Yasuda et al. (1997). Another 3 (van den Bergh 1996), 31.12 (Whitmore et al. 1995) and E/S0 Cluster B candidates according to Binggeli et al. (NGC 30.79 (Ciardullo et al. 1998). The discrepancy most likely 4472 (M49), 4526, and 4570) have µo < 31.2, as determined arises from an underestimate of the total H magnitude of using the FP relation. Our low distance estimate of M49 M87 on our part, due the fact that M87 fills the TIRGO agrees with independent estimates obtained using all meth-
c 0000 RAS, MNRAS 000, 000–000 8 G. Gavazzi et al. ods quoted above. Unfortunately this is the only galaxy with sented in Tab. 3, and is also shown in Fig. 10, where the av- independent distance estimate among this group of early- erage velocities and distance moduli are plotted with error type objects. We propose that M49 belongs to cloud S. Its bars indicating the statistical uncertainties on the determi- −1 −1 −1 redshift (VLG = 1200 km s ) is 400 km s lower than the nation of the two quantities (galaxies with σ < 100 km s mean, but it is well within the distribution of VLG in this are excluded). Estimates of cluster A are also given sepa- region. rately for early-type and late-type galaxies, subdivided into The remaining galaxies have µo < 31.5. They form the main HI deficient and HI normal. Mean quantities and the asso- body of the cluster, indicated here with region A (M87), N, ciated uncertainties were computed using the so-called bi- to the NW, E, to the East and S, to the South. weight estimators (see Beers et al. 1990 for details), that are Region A coincides with Cluster A and with the X-ray po- known to provide a robust parameter estimation for sam- sition (see Fig. 7 adapted from B¨ohringer et al. 1994, repro- ples covering a wide interval in size. Statistical uncertain- duced on the same scale as Fig. 6). As expected, most HI ties on the distance modulus determinations are obtained deficient galaxies belong to this region (see Fig. 8). Two ex- adding in quadrature the uncertainty on the FP template ceptions are surprisingly found at the southern edge of the zero-point determination and the uncertainty with which a cluster in the region of NGC 4636, a strong and extended given sample distance is determined (taking into account the X-ray source (Trinchieri et al. 1994). We argue that a signif- dispersion that characterizes the FP relation). These should icant quantity of extended gas must be associated with this be conservative estimates, since the uncertainties associated galaxy. with the TF relation, that is used to derive 60% of the dis- The distribution of µo in the cluster A itself is centered at tance estimates, are somewhat smaller than those associated § µo = 30.84 with a dispersion of 0.45 mag. This is com- with the FP. The last column of Table 3 lists the peculiar ve- parable to the nominal uncertainty of the distance determi- locities of the Virgo clouds with respect to Cluster A (com- nation methods (0.35 and 0.45 mag for TF and FP), thus puted as the difference between the actual velocity of the the depth along the line of sight of this aggregate cannot be various entities (Column 3) and their Hubble flow velocity, −1 determined. minus 156 km s which represents the deviation of cluster One of the most interesting results of the present analysis A itself from the Hubble flow) , with statistical uncertain- is that, among galaxies at the main cluster distance, those ties. belonging to clouds N and S show a significant velocity seg- It appears that the main Virgo cluster (cluster A), cloud E regation. This is illustrated in Fig. 9. The two rightmost and the background cloud W are in near Hubble flow (the panels of this figure give histograms of VLG and of µo de- deviation of cloud M can be assesed with low significance rived from this work for the 7 considered regions. While with only 4 distance measurements). Cluster B lies at sig- the distances of cluster A and clouds S, N, and E are in nificantly larger distance than predicted from its velocity agreement (µ = 30.84 − 31.23) the only significantly more if it was in Hubble flow. Thus it is falling into Virgo from o −1 distant structures are cluster B and clouds W and M. Clus- the far-side at ∼ -760 km s . The two structures N and ters A and B and cloud E have similar velocity distributions S have distances similar to the main cluster. However their −1 peaked at the standard VLG ∼ 1350 km s . Clouds W and velocities are smaller and larger respectively than those of M have instead a higher velocity. Clouds S and N have signif- galaxies in dynamic equilibrium with the main body of the icantly different distributions. The N one contains galaxies Virgo cluster. Cloud N, falling from the far-side and cloud S, −1 with VLG < 1300 km s , thus blueshifted with respect to falling from the near-side, have reached the distance of the Virgo. Similar evidence was pointed out by Hoffmann et al. main cluster. We remark that these clouds are composed al- (1989b) and was extensively analyzed by Tully & Shaya, most entirely of spirals (spiral fraction ∼80%), while in all (1984) to model the infall of galaxies on the Virgo cluster. other regions the fraction of these galaxies is ∼ 50%. On the contrary, cloud S contains mainly redshifted galax- The derived parameters of the individual clouds can be com- −1 ies, with 750 < VLG < 2700 km s . To check if the lat- pared with similar determinations carried out by previous ter result is not due to the limited statistics of the sample investigators. Yasuda et al. (1997) found 30.7 < µo < 31.5 used in this work, and since the velocity distribution can be for Cluster A, µo ∼ 31.7 for Cluster B and 32.4 < µo < 33.0 derived from a larger body of velocity measurements than for Clouds W and M, in good agreement with ours. How- the one represented in our sample of galaxies with distance ever these authors claim evidence of infall of clouds W and estimates, we determine the velocity distributions in the 7 M, that we do not confirm. Federspiel et al. (1998) found studied regions using the whole VCC (which contains over µo = 31.35 for Cluster A, larger than our estimate, and 400 galaxies with redshift estimates) (see left panel of Fig. µo ∼ 31.81 for Cluster B, in excellent agreement with the 9). The difference between regions S and N, noticed in our present determination. smaller sample with distance estimates, is equally present in the larger VCC sample, and we conclude it represents a real 5.4 The mass estimate of the Virgo cluster difference between the two clouds. A summary of the mean velocity, velocity dispersion, and Determining the mass of the Virgo cluster is beyond the distance modulus determinations for all seven regions is pre- scope of the present investigation (see Tully & Shaya 1984 for a comprehensive analysis). However a simple consistency check can be done comparing the virial mass of the cluster § It is important to stress once more that the the absolute dis- with the mass required to perturb the Hubble flow in the tance of cluster A is not an independent result of this work, since vicinity of Virgo in order to produce the peculiar velocities a distance of Virgo = 16 Mpc has been assumed to calibrate the observed in the various Virgo clouds and in the Local Group. zero point of both the TF and FP template relations The virial mass Mvir of the Virgo cluster was estimated by
c 0000 RAS, MNRAS 000, 000–000 The 3-D structure of the Virgo cluster 9
15 8 Tully & Shaya (1984) to lie at ∼ 10 M⊙ (assuming Ho=80 (less than few × 10 yrs, the time scale for gas depletion) − − km s 1 Mpc 1). taken place (see also Tully & Shaya 1984). Two more galaxy The mass required to perturb the Hubble flow expansion can clouds, M and W, projected on the Western edge of Virgo, be estimated from δv/v = Ω0.6/3×δρ/ρ, valid in a spherical have distances consistent with B. W is in near Hubble flow, geometry for density enhancements over the mean density while the velocity of M indicates infall. However its devi- of the universe δρ/ρ ∼ 1 (linear regime) (Davis & Peebles ation from the Hubble expansion can be assesed with low 1983). For larger density enhancements (quasi-linear regime) statistical significance. we use the analytic correction scheme proposed by Nusser The mass of the Virgo cluster inferred from the peculiar mo- at al. (1991) (see their equation 34). The deviations from tions induced on its surroundings (Local Group, clouds N, the Hubble flow due to the Virgo cluster at the distance of B and W) ranges between 2 and 3 ×1015M⊙, a factor of 2-3 the Local Group, of cluster B, of clouds N and W are re- larger than the virial expectation. spectively: δv/v = 220/1354, 762/701, 768/56 and 150/1236. We conclude with a note of caution: due to the large grav- These correspond to density enhancement δρ/ρ ∼ 1.0, 6.7, itational perturbation induced by cluster A, galaxies with 85 and 0.75 respectively, assuming Ω=0.3. Clearly large cor- large peculiar velocities (regions B, N and S) must be cau- rection for non-linear regime are required for clouds N and B. tiously rejected to obtain an umbiased determination of Using ρ = 2 × 10−29 , and given the distance of 15 Mpc from the Hubble constant in the Virgo region. Moreover N4639, the Local Group to Virgo, of 8.6 Mpc from cluster B to A, whose µo = 32.00 from the present work coincides with the of 0.7 Mpc from cloud N to A and of 15 Mpc from cloud W Cepheyds distance obtained with the HST, should not be to A, the resulting mass enhancement ∆M due to the Virgo used to represent the distance of Virgo: in fact it belongs to cluster ranges between 2.1 and 3.1 ×1015M⊙. Not surpris- a region (E), containing other galaxies with similarly large ingly, this is a factor of 2-3 larger than the virial expectation. distances. In fact, a significant fraction of the mass is expected to lie outside the main cluster, given the Virgo complex structure, one that is far from being a single, dinamically relaxed en- ACKNOWLEDGMENTS tity. We thank Bruno Binggeli for sending us the VCC in elec- tronic form. We are grateful to the TAC of the TIRGO and Calar Alto Observatories for the generous time allocation to 6 SUMMARY AND CONCLUSIONS this project. We thank C. Baffa, A. Borriello, V. Calamai, B. Catinella, I. Randone, P. Ranfagni, M. Sozzi, P. Strambio We have carried out a H band surface photometry survey for assistance during the observations at TIRGO. A special of 200 galaxies in the Virgo cluster brigther than B < 14.0 thank to V. Gavriusev for software assistance at TIRGO, mag. These data, complemented with dynamical measure- and to T. Beers for permission to use his ROSTAT fortran ments taken from the literature and with new spectroscopic code. We also wish to thank an unknown referee whose crit- data for 19 objects, allowed us to derive distances to 134 icism helped improving this work. galaxies in this cluster with either the TF or the FP meth- ods. The individual distance uncertainties are estimated ∼ 0.35 mag and 0.45 mag, for the TF and FP methods respec- tively. The Virgo cluster is confirmed to be anything but a viri- alized cluster. Its main potential well (cluster A) coincides in position with M87, as marked by the distribution of the extended hot gas. An equal mixture of E+S0 and of very HI deficient Spiral galaxies populates this cluster. The cluster appears elongated in the N-S direction. The accuracy of the present distance determinations is still not sufficient to con- vincingly assess a possible extension in depth. The second major galaxy aggregate, cluster B, is found to have a smaller spatial extension than originally proposed by Binggeli, Sandage & Tammann (1985), it lies one magnitude further away than Virgo, falling onto the main cluster with -760 km s−1 peculiar velocity. M49, projected onto cluster B, is found at significantely smaller distance than B, thus it is proposed to belong to cloud S. It remains unexplored whether the hot gas found in this region is associated with M49 or with cluster B itself. Two other clouds, almost exclusively composed of spiral galaxies, are falling onto the cluster (they are at the same distance as A): the one on the North-West is falling from − behind (with a peculiar velocity of -768 km s 1), the other, to the South of A, is falling from the near-side at about − +200 km s 1. None of these galaxies suffers from severe HI deficiency, indicating that phenomenon has only recently
c 0000 RAS, MNRAS 000, 000–000 10 G. Gavazzi et al.
REFERENCES Lisi, F., Baffa, C., Biliotti, V., et al., 1996, PASP, 108, 364 McElroy, D., 1995, ApJS, 100, 105 Ajhar, E., Lauer, T., Tonry, J., et al., 1997, AJ, 114, 626 Morris, P. & Shanks, T., 1998, MNRAS, 298, 451 Beers, T.C., Flynn, K., Gebhardt, K. 1990, AJ, 100, 32 Nusser, A., Dekel, A., Bertschinger, E., & Blumenthal, G., 1991, Binggeli, B., Sandage, A., & Tammann, G., 1985, AJ, 90, 1681 ApJ, 379, 6 (VCC) Pahre, M., Djorgovski, G., & de Carvalho, R., 1995, ApJ, 453, Binggeli, B., Popescu, C., & Tammann, G., 1993, A&AS, 98, 275 L17 B¨ohringer, H., Briel, U. G., Schwarz, R. A., Voges, W., Hartner, Peterson, S., 1979, ApJS, 40, 527 G., & Trumper, J. 1994, Nature, 368, 828 Pierce, M., Mcclure, R., & Racine, R., 1992, ApJ, 393, 523 Boselli, A., Casoli, F., Lequeux, J., 1995, A&AS, 110, 521 Pierce, M., Welch, D., Mcclure, R., van den Bergh, S., Racine, Boselli, A., Tuffs, R., Gavazzi, G., Hippelein, H., & Pierini, D., R., & Stetson, P., 1994, Nat, 371, 385 1997, A&AS, 121, 507 Pierce, M. J., & Tully, R. B. 1988, ApJ, 330, 579 Bottinelli, L., Gouguenheim, L., Fouque, P., & Paturel, G., 1990, Press, W., Teukolsky, S., Vetterling, W., & Flannery, B. 1992 A&As, 82, 391 “Numerical Recipes” 2nd edition (Cambridge: Cambridge Chincarini, G., & de Souza, R., 1985, A&A, 153, 218 University Press) Cayatte, V., van Gorkom, J., Balkowski, C., & Kotanyi, C., 1990, Saglia, R., Bertschinger, E., Baggley, G., Burstein, D., Colless, AJ, 100, 604 M., Davies, R., McMahan, R., & Wegner, G., 1993, MNRAS, Ciardullo, R., Jacoby, G., Feldmeier J., Bartlett, R., 1998, ApJ, 264, 961 492, 62 Saha, A., Sandage, A., Labhardt, L., Tammann, G., Macchetto, Davis, M., & Peebles, P., 1983, ARA&A, 21, 109 F., & Panagia, N., 1997, ApJ, 486, 1 Della Valle, M., & Livio, M., 1995, ApJ, 452, 704 Sandage, A., Saha, A., Tammann, G., Labhardt, L., Panagia, N., de Vaucouleurs, G. 1948, Ann. Astrophys., 11, 247 Macchetto, F., 1996, ApJ, 460, L15 de Vaucouleurs, G. 1961, ApJS, 6, 213 Schank, T., 1997, MNRAS, 290, L77 Djorgovski G., Davis M., 1987, ApJ, 313, 59 Schmidt, B., et al., 1994, ApJ, 432, 42 Dressler, A., 1987, ApJ, 317, 1 Schneider, S., Thuan, T., Magri, C., & Wadiak, J., 1990 ApjS, Dressler, A., Lynden-Bell, D., Burstein, D., Davies, R., Faber, S., 72, 245 Terlevich, R., & Wegner, G., 1987, ApJ, 313, 42 Scodeggio, M., 1997, PhD Thesis, Cornell University Efron B. 1987, Journal Am. Stat. Assoc., 82, 171 Scodeggio, M., Giovanelli, R., & Haynes, M., 1997a, AJ, 113, 101 Elias, J., Frogel, J., Matthews, K., & Neugebauer, G., 1982, AJ, Scodeggio, M., Gavazzi, G., Belsole, E., Pierini, D., & Boselli, A., 87, 1029 1998a, MNRAS (in press) Ferrarese, L., Freedman, W., Hill, R., et al., 1996a, ApJ, 464, 568 Scodeggio, M., Giovanelli, R., & Haynes, M., 1998b, submitted to Ferrarese, L., Livio, M., Freedman, W., et al. 1996b, ApJ, 468, AJ L95 Secker, J., & Harris, W., 1993, AJ, 105, 1358 Federspiel, M., Tammann, G., & Sandage, A., 1998, ApJ, 495, Speraindio, M., Chincarini, G., Rampazzo, R., Molinari, E., 1995, 115 A&AS, 110, 279 Fukugita, M., Okamura, S., & Yasuda, N. 1993, ApJ, 412, L13 Sulentic, J., Arp, H., 1982, AJ, 88, 489 Gavazzi, G. & Boselli, A., 1996, Astrophys. Lett. & Communicat. Thronson, H., Tacconi, L., Kenney, J., Greenhouse, M., Margulis, 35, 1 M., Tacconi-Garman, L., Young, J., 1989, ApJ, 344, 747 Gavazzi, G., Pierini, D., Baffa, C., Lisi, F., Hunt, L., & Boselli, Tonry, J., & Davis. M., 1979, AJ, 84, 1511 A., 1996a, A&AS, 120, 521. Tonry, J., Ajhar, E., & Luppino, G. 1990, AJ, 100, 1416 Gavazzi, G., Pierini, D., & Boselli, A., 1996b, A&A, 312, 397 Trinchieri, G., Kim, D-W. Fabbiano, G., & Canizares, C., 1994, Giovanardi, C., Krumm, N., Salpeter, E., 1983, AJ, 88, 1719 ApJ, 428, 555 Giovanelli, R., & Haynes, M., 1983, AJ, 88, 881 Tukey J., 1958, Ann. Math. Stat., 29, 614 Giovanelli, R., Haynes, M., Herter, T., Vogt, N., da Costa, L., Tully, B., & Fisher, R., 1977, A&A, 54, 661 Freudling, W., Salzer, J., & Wegner, G., 1997, AJ, 113, 53 Tully, B., & Shaya, E. 1984, ApJ, 281, 31 Haynes, M., & Giovanelli, R., 1984, AJ, 89, 758 van den Bergh, S., 1996, PASP, 108, 1091 Haynes, M., & Giovanelli, R., 1986, ApJ, 306, 466 Yasuda, N., Fukugita, M., & Okamura, S., 1997, ApJS, 108, 417 Helou, G., Giovanardi, C., Salpeter, E., & Krumm, N., 1981, Warmels, R., 1986, Ph.d. Thesis, Univeristy of Groningen ApJS, 46, 267 Whitmore, B., Sparks, W., Lucas, R., & Biretta, J., 1995, ApJ, Helou, G., Hoffman, L., & Salpeter, E., 1984, ApJS, 55, 433 454, L73 Herbst, T., Beckwith, S., Birk, C., et al. 1993, SPIE 1946 Zwicky, F., Herzog, E., Karpoviwicz, M., Kowal, C., Wild, P., Hoffman, L., Helou, G., Salpeter, E., Glosson, J., & Sandage, A., 1961-68, “Catalogue of Galaxies and Clusters of Galaxies” 1987, ApJS, 63, 247 (Pasadena: California Institute of Technology)(CGCG) Hoffman, L., Lewis, M., Helou, G., Salpeter, & E., Williams, B., 1989a, ApJS, 69, 65 Hoffman, L., Helou, G., Salpeter, & Lewis, M., 1989b, ApJ, 339, 812 Hoffman, L., Salpeter, E., Farhat, B., Roos, T., Williams, H., Helou, G., 1996, ApJS, 105, 296 Jacoby, G., Ciardullo, R., & Ford, H., 1990, ApJ, 356, 332 Jensen, J., Luppino, G.,& Tonry, J., 1996, ApJ, 468, 519 Jørgensen, I., Franx, M., Kjærgaard, P., 1996, MNRAS, 280, 167 Kenney, J., Rubin, V., Planesas, P., Young, J., 1995, ApJ, 438, 135 Kenney, J.,& Young. J., 1988, ApJS, 66, 261 Lemaitre, G., Kohler, D., Lacroix, D., Meunier, J., Vin., A., 1990, A&A, 228, 540 Lisi, F., Baffa, C., & Hunt, L.K. 1993, SPIE, 1495, 594
c 0000 RAS, MNRAS 000, 000–000 The 3-D structure of the Virgo cluster 11
Figure 1. The rotation curves of 8 HI deficient galaxies observed Figure 10. The average distances obtained in this work are plot- at the OHP. The extent of the horizontal scale is up to the galaxy ted against the average VLG for each of the 7 regions. Error bars optical diameter. Galaxies 70048, 70090 and 70213 are not used indicate one-sigma statistical uncertainties on the measured quan- in the following analysis. tities. The broken line represents the Hubble flow, for Ho=81.35 km s−1 Mpc−1.
Figure 2. a): the FP relation for 59 Virgo galaxies (dots) sepa- rated into σ < 100 and σ > 100 km s−1. The template relation obtained on the Coma cluster is given using µo=31.0 and ±1 mag. b): the TF relation for 75 Virgo galaxies (dots) separated into DefHI < 0.7 and DefHI > 0.7. The template relation ob- tained on 4 clusters is given using µo=31.0 and ± 1 mag. c): The 73 galaxies in the A262, Cancer, Coma and A1367 clusters used to derive a template TF relation.
Figure 3. a): the comparison between the distance moduli ob- tained in this work with those available in the literature, obtained with other than Tully-Fisher methods. b): the comparison be- tween the distance moduli obtained in this work with the H band TF method, with the B band ones obtained by Yasuda et al. (1997). c): the comparison between the distance moduli obtained in this work (for σ > 100 km s−1) and those obtained with the Dn-σ method in the B band. The dotted line is drawn to indicate the one-to-one relationship.
Figure 4. The distances obtained in this work are plotted against VLG. Small dots mark the early type galaxies (FP), separated into σ < 100 and σ > 100 km s−1. Larger circles mark the spirals (TF), separated into normal and HI deficient. The broken line −1 −1 represents the Hubble flow, for Ho=81.35 km s Mpc .
Figure 5. Same as Fig. 4 with symbols size decreasing with in- creasing distance and whose shape refers to 7 regions of the Virgo cluster, represented in Fig. 6 with dotted lines.
Figure 6. The distribution in celestial coordinates of 134 galaxies with distance estimates obtained in this work in the region cov- ered by the VCC (solid line). The 7 regions of revised membership are given contoured by broken lines. Symbol sizes decrease with increasing distance (as in Fig. 5). The positions of M49 and M87 are marked.
Figure 7. The ROSAT all sky map of Virgo, adapted from B¨ohringer et al. (1994) on the same scale of Fig. 6.
Figure 8. The distribution in celestial coordinates of 134 galaxies with distance estimates. Small dots mark the early type galaxies (FP) (for σ > 100 km s−1) and larger circles mark the spirals (TF), separated into normal and HI deficient (as in Fig. 4).The positions of M49 and M87 are marked.
Figure 9. Histograms of VLG in the 7 cluster regions. a): using 433 velocities in the VCC; b): using 134 galaxies with distance estimates obtained in this work. c): histograms of the distance moduli obtained in this work in the 7 regions of revised member- ship.
c 0000 RAS, MNRAS 000, 000–000