A Search for Active Galactic Nuclei in Sc Galaxies with H II Spectra

arXiv:astro-ph/0208460v1 26 Aug 2002 ohv uemsiebakhlsa hi etr seKor- these (see of masses centers the ( and their review), holes a at black for 2001 holes Gebhardt & black mendy supermassive have to 1 n ugls aay h erysia 3,weea upper an where M33, least spiral at 1500–3000 of in nearby limit confirmed the appears galaxy, mass bulgeless bulge rela to one The mass 2001). hole al. black et ma of Ferrarese reverberation 2000b; al. by derived et (Gebhardt the as with ping and hole consistent lines black emission the are broad from relation of distances widths 2002). this the al. from from derived et masses those Tremaine hole 2000a; al. (Ferrar black et luminosity The Gebhardt bulge 2000; the Merritt from f & derived with correlation that m a than hole is black scatter dispersion luminosities of velocity relation bulge stellar the bulge fact, the to In from 1998). al. inferred et as (Magorrian masses bulge galaxy w efr aaiswr on nta uvy n 2 fthe of Fifty 82% and survey, universe. that in nearby of found the population were galaxies the in Seyfert of (AGNs) two census nuclei a Tam- conduct galactic & to active Revised order (Sandage the in Galaxies 1981), from sur- Bright mann selected spectroscopic of galaxies a Catalog nearby Shapley-Ames perform bright, 486 to of telescope vey inch 200 Palomar 2001). Joseph & Ferrarese, Merritt, 2001; al. et (Gebhardt rpittpstuigL using typeset Preprint T PERIN APPEAR O h ainlRdoAtooyOsraoyi aiiyo th of facility a is Observatory Astronomy Radio National The otnab omladatv aaiswt ugsappear bulges with galaxies active and normal nearby Most o iipno agn 19,19a 97)ue the used 1997b) 1997a, (1995, Sargent & Filippenko, Ho, n lotcranyi eeae ytemleiso rmg predictio to from comparable emission are H powers thermal radio by 6-cm The generated galaxies. is nu extended certainly the almost that shows and galaxies detected the for results hs aaishv ek6c ai oesof powers radio 6-cm we sources peak radio have objects nuclear galaxies three star-forming These Only as present. classified be might optically nuclei are that Mpc 30 o-eetdglxe yial ag rm10 from range typically galaxies non-detected L k uvyipisglx-iesa omto ae f0 of rates formation star galaxy-wide implies Survey Sky VLA ujc headings: p Subject the H out, with ruled galaxies definitively late-type be these in cannot re holes star-forming black central massive of by over-resolution to due possibly icmula trfrainhsbe eetdi hs thr these in detected been has formation star circumnuclear r eovdadapa ohv ifs opoois ihl with morphologies, diffuse have to appear and resolved are fteudtce aais h pe iist h ai pow radio the to limits upper the galaxies, undetected the of h srpyia Journal Astrophysical The α ehv erhdfrncerrdoeiso rmastatistic a from emission radio nuclear for searched have We ∼ uioiis n ml ula trfrainrtso 0 of rates formation star nuclear imply and luminosities, 10 A T E 6 tl mltajv 14/09/00 v. emulateapj style X M ERHFRATV AATCNCE NS AAISWT H WITH GALAXIES SC IN NUCLEI GALACTIC ACTIVE FOR SEARCH A M ⊙ ⊙ 1. –10 h bevtre fteCrei nttto fWashingto of Institution Carnegie the of Observatories The a enfudfrtebakhl mass hole black the for found been has INTRODUCTION 9 M aais cie—glxe:nce aais efr ga — Seyfert galaxies: — nuclei galaxies: — active galaxies: otnu:galaxies continuum: . ⊙ emwl orltdwt the with correlated well seem ) ainlRdoAtooyObservatory Astronomy Radio National II spectra. oapa nTeAtohsclJournal. Astrophysical The in appear To ainlSineFudto,oeae ne cooperative under operated Foundation, Science National e J 18 rless ar AMES ∼ . 4 [email protected] tion ABSTRACT ese ass 10 L [email protected] Hz W p- ir UIS - 20 .U S. AND Hz W .H C. 1 − in.Atog h rsneo cience powered nuclei active of presence the Although gions. LVESTAD 1 eH ee 1 la ai msinhsafltsetu ntoobjects, two in spectrum flat a has emission radio clear ..BxO oor,N 87801 NM Socorro, O, Box P.O. , h rsneo opc oreo adXry rradio or X-rays hard of source includ compact blac AGN a an central of of presence massive presence the the contain for discriminators galaxies ques- Good the these holes. to whether equivalent H of be whether may tion of then AGNs question observable the contain gas, of plies rlbakhl n upyo a o“ed htbakhole. black that “feed” to gas of H supply with cen a galaxies massive Since and a hole constituents: black two tral requires typically AGN an of r lsie sH as classified are later. or Sb type of galaxies in found are omto,i loi osbeta hycnanwa AGNs the weak by H overpowered nuclear contain from been they signal simply dominant that more has possible signature optical is whose also it formation, H the while Sbc), than (earlier foun are galaxies (thoug LINERs) early-type and general (Seyferts in a AGNs As the rule, stars. universal) massive not young, by ionization with 0 aaisfo h aoa apewr eotdt con- to reported were sample Palomar H tain the from galaxies 206 nuclei. active the t powering implying holes AGNs, black massive the low-luminosit of with these presence associated in are sources & typically radio Ulvestad Seyferts detected 2001; Ulvestad The & (Ho 2001). length Ho d observa- in with were min cm, Seyferts 15–18 6 of at these (VLA) tions of Array Large 45 Very the of using sample tected statistical a in objects o10 to nodrt eemn hte ekatv galactic active weak whether determine to order in , edtce,i G 6,NC42,adNC4535. NGC and 4123, NGC 864, NGC in detected, re O sfrtemleiso htaebsdo h nuclear the on based are that emission thermal for ns raelwrta rdce rmteH the from predicted than lower are er na ie of sizes inear − ,83SnaBraaS. aaea A91011 CA Pasadena, St., Barbara Santa 813 n, . eetrslssgetta hyaelkl ob rare be to likely are they that suggest results resent lhuhtelt-yeglxe ntePlmrsml that sample Palomar the in galaxies late-type the Although nadto ote5 efrsietfidb oe l (1997a), al. et Ho by identified Seyferts 52 the to addition In 08 1 II sinzdb on tr ntecneso those of centers the in stars young by ionized as taceodrslto,wieuprlmt fthe of limits upper while resolution, arcsecond at aais oprsnwt rvos2-mVLA 20-cm previous with Comparison galaxies. 20 . lycmlt apeo 0S aaiswithin galaxies Sc 40 of sample complete ally − 3 II 0 − Hz W uli aaynce ihotclln aisconsistent ratios line optical with nuclei galaxy nuclei: . 8 1 . M 0 M ⊙ − 1 ⊙ ∼ yr h he ula ai ore all sources radio nuclear three The . − yr 0 c hssrnl niae that indicates strongly This pc. 300 1 II hl h o-eouinNRAO low-resolution the while , − ais trus radio — starburst laxies: 1 aaisapa ob oiae ystar by dominated be to appear galaxies nsasaoe5 above stars in II pcr eeal abrcpossup- copious harbor generally spectra gemn yAscae nvriis Inc. Universities, Associated by agreement II SPECTRA II M α ein.Tepresence The regions. ⊙ II luminosity, nafew a In . dmntdobjects -dominated II nuclei he e- k d h y e -

2 Ulvestad & Ho

TABLE

VLA Observations of H II Galaxies

a a b

Galaxy Date Array Integration J J Distance rms

0 00

min h m s Mp c Jy b eam

NGC B

NGC BA

IC Oct A

NGC Nov BnA

NGC Oct A

NGC B

NGC Oct A

NGC Jun A

NGC Oct A

NGC BBnA

NGC Oct A

NGC Dec A

NGC Oct A

NGC Oct B

NGC Oct A

NGC AB

NGC Oct A

NGC ABBnA

NGC Oct A

NGC May B

NGC Oct A

NGC Aug A

NGC Oct A

NGC A

NGC Dec A

NGC Oct A

NGC B

Observed p ositions are quoted for undetected galaxies images always were large enough to incorp orate the

currently accepted galaxy co ordinates well within the elds of view

Distances are taken from Tully

The quoted p osition refers to the strongest detected source asso ciated with the galaxy

emission identiﬁed with the galaxy nucleus. Here, we report 2. SAMPLE SELECTION AND OBSERVATIONS on a study of a subsample of H II nuclei in Sc galaxies from the In orderto providea uniformsample, we chose the 57 objects Palomar bright galaxy sample, in which we test for the presence classiﬁed as Sc (T = +5) in the Ho et al. (1997a)catalog of H II of weak AGNs and central black holes by searching for radio galaxies. Due to the constraints of limited observing time, we emission from the galaxy nuclei. selected a distance-limited sample of 41 of these galaxies lying within 30 Mpc. In this sample, 15 objects had 6-cm data AGNs in H II Galaxies 3

NGC 864 4.9 GHz NGC 4123 4.9 GHz NGC 4535 4.9 GHz 3 8 (a) (b) (c) 4 200 pc 200 pc 6 2 200 pc

4 2 1 2

0 0 0 Arcsec Arcsec Arcsec

-2 -1 -2 -4

-2 -6 -4

-3 -8 4 2 0 -2 -4 3 2 1 0 -1 -2 -3 8 6 4 2 0 -2 -4 -6 -8 Arcsec Arcsec Arcsec

FIG. 1.— Images at 4.9 GHz of the three detected nuclear radio sources from the H II galaxy sample. All contour levels start at 3 times the rms noise in the image, and increase by factors of √2. Scale bars indicating linear sizes of 200 pc are indicated on all images. (a) NGC 864: rms noise of 49 µJy beam−1, peak flux density of 0.55 mJy beam−1, beam size 1.′′77 1.′′48 at PA = −35◦. (b) NGC 4123: rms noise of 45 µJy beam−1, peak flux density of 1.39 mJy beam−1, beam size 0.′′57 0.′′44 at PA = 5◦. (c) NGC 4535: rms× noise of 115 µJy beam−1, peak flux density of 1.66 mJy beam−1, beam size 2.′′71 1.′′55 at PA = 87◦. × × at .1′′ resolution in the VLA archive, from observations made from the archive for a given galaxy, each data set was calibrated in the high-resolution A and B configurations, or the hybrid individually and then combined. Bad data were identified and BnA configuration with a long north arm (see Thompson et al. flagged, and deconvolution and imaging then was performed on 1980). Data for these objects were extracted from the archive. the calibrated data sets. All data sets were imaged with “natu- Twenty-five sample galaxies, as well as one galaxy at a distance ral” weighting in the (u,v) plane in order to provide the best slightly greater than 30 Mpc (IC 467), were observed by us on possible sensitivity. For data sets with excess noise or obvious 16 October 2000, using the VLA A configuration. All new ob- stripes in the images, confusing sources were imaged and self- servations were centered at a frequency of 4.86 GHz, with a calibrated in order to reduce the noise level. The 1σ rms noise total bandwidth of 100 MHz in each of two circular polariza- was 100 µJy beam−1 for all the new observations, and varied tions. One galaxy that met our selection criteria, NGC 3430, depending on data quality and quantity, as well as array config- was not observed because of the apparent presence of data in uration for the archival data (see Table 1). Synthesized beam the VLA archive, but we later found that there were no usable sizes were 0.′′4–0.′′5 for the new A configuration data, and 0.′′4 archive data. to as much as 2′′ for the archival data. Approximately 3.5 minutes were spent integrating on each of the newly observed galaxies, providing a detection threshold 3. RESULTS about 2.2 times higher than for the Seyfert galaxies observed by Ho & Ulvestad (2001). For the archive data, total integra- The range of resolutions and galaxy distances (typically 15– tion times ranged from a few minutes to more than two hours. 30 Mpc) provideslinear resolutionsin the rangeof 40 to 300 pc, Table 1 summarizes the new and archival observations of the sufficient to isolate the galaxy center from the rest of the galaxy. None of the 26 newly observed galaxies was detected, with a galaxies, including the nominal positions for the galaxy nuclei −1 that were used for imaging. For the seven detected galaxies 5σ upper limit of 0.5 mJy beam for each. For the 15 galax- (see Section 3), the listed positions are those of the peaks of the ies whose archival data were analyzed, seven were detected: radio sources apparently associated with the galaxies. All im- NGC 864, 1073, 3359, 3877, 4123, 4535, and 5907. The flux ages covered fields of at least 100′′ in both right ascension and densities and powers of the detected galaxies are summarized declination, usually centered at the nominal galaxy positions at in Table 2, while the positions of their radio peaks were given the times of the observations. In two galaxies, NGC 3666 and in Table 1. NGC 5907, the currently accepted positions for the galaxy nu- The noise levels for the archival data often were lower than clei are more than 15′′ from the original pointing positions, so for the new data due to longer integration times, but only one the data for these sources were phase-shifted to the more accu- of the detections was weaker than the upper limit for the new rate galaxy positions in order to center the images on the nuclei. data. However, the archival galaxies often contain significant Processing of both new and archival data sets was similar, amounts of data from the larger B configuration, sensitive to and was performed in the NRAO Astronomical Image Process- more extended structures of lower surface brightness than for ing System (van Moorsel, Kemball, & Greisen 1996). The flux- the new A configuration data. The beam area in the larger density scales were set by observations of 3C 48 and 3C 286, configuration typically is 10 times larger than the beam area using the scale of Baars et al. (1977), with slight adjustments in the smaller configuration; the respective beam diameters are ∼ ∼ for variability that have been found by VLA staff. A local cal- 130 pc and 40 pc at the median galaxy distance of 17 Mpc. ibrator was observed for each galaxy, in order to transfer the If the radio sources tend to be resolved on scales of 100–200pc, amplitude scale from the primary flux calibrators and to pro- an increased peak flux density would be seen in the lower res- vide a local phase calibration, removing the effects of electron- olution data, perhaps moving the galaxies above our detection ics and atmosphere. If more than one data set was available threshold. This possibility is supported by the fact that the detected nuclear radio sources in three galaxies are partly resolved

4 Ulvestad & Ho

TABLE

Detections of H II Galaxies

Galaxy Peak Total Log P Log P

;p eak ;total

mJy b eam mJy W Hz W Hz

NGC

a;c

NGC

The detected radio source is not asso ciated with the nucleus of the galaxy

The detected radio source is asso ciated with SN S

The source app ears slightly smeared due to the nite bandwidth and the large distance from the

eld center by the VLA (see Section 4.1 for more details). sources detected in the field, one associated with a background quasar identified by Colbert & Ptak (2002); the other detection 4. ORIGINSOFTHERADIOEMISSION may be associated with the galaxy, but is located approximately ′′ ◦ 4.1. Locations of the Detected Radio Sources 30 from the nucleus at position angle 37 , and is not coinci- dent with either of the intermediate-luminosity X-ray objects In this section, we consider primarily the H II galaxies that shown by Colbert & Ptak (2002). NGC 3359 has a detection were detected by the VLA. Only three of those galaxies have about 9.′′5 south of the nucleus, and NGC 5907 has a detec- radio detections consistent with the positions of their nuclei tion nearly an arcminute away from the nucleus, possibly as- as given in the NASA Extragalactic Database (NED) and con- sociated with the disk of the edge-on host galaxy. Finally, the firmed with the measurements of Cotton, Condon, & Arbizzani only source detected in NGC 3877 is at the location of the Type (1999). The three nuclear detections are in the galaxies IIn supernova SN 1998S, whose radio detection using the same data was reported previously by Van Dyk et al. (1999). NGC 253 at 20 Mpc distance 4.9 GHz Figure 1 shows the radio images of the three nuclear radio sources found in our survey. All three are resolved and appear 3 somewhat diffuse (as opposed to “jet-like”), with diameters of ∼ 200 pc 300 parsecs. This radio emission may be associated with starbursts, but the unresolved components of the core radio emis- 2 sion conceivably may be associated with AGNs. For these three nuclear radio sources, NGC 864 and NGC 4535 contain only 1 data from the B configuration of the VLA, whereas NGC 4123 has data from both the A and B configurations. To test the effects of resolution on galaxy detectability, we have made im- 0 ages of NGC 4123 containing only the A configuration data

Arcsec and only the B configuration data. The peak flux density of the nuclear source is a factor of 3 higher in the B configuration -1 image than in the A configuration image (4.4 mJy beam−1 vs. 1.4 mJy beam−1), as expected from inspection of Figure 1. If the same ratio had held for NGC 864 and NGC 4535, NGC 864 -2 would not have been detected in our short A configuration inte- grations, whereas NGC 4535 would have been marginally de- -3 tected. Therefore, the generally higher sensitivity to extended structures in the archival galaxies accounts for at least some of 3 2 1 0 -1 -2 -3 the increased detection probability for these objects. In fact, it Arcsec may be that a significant fraction of the galaxies that were not detected in the October 2000 observing run would have been detected using a more compact VLA configuration. FIG. 2.— VLA image of NGC 253, from Ulvestad & Antonucci (1997), degraded in resolution and flux density to indicate its appearance viewed at a distance of 20 Mpc. Contours increase by √2 from a lowest level of 4.2. Radio Emission Associated with Star Formation − ′′ ′′ 150 µJy beam 1. At 20 Mpc distance, the beam size is 0. 66 0. 36 at PA = A young starburst typically can give rise to thermal radio 12◦, and the peak flux density is 8.22 mJy beam−1. × emission from collections of H II regions or super star clusters, NGC 864, NGC 4123, and NGC 4535. NGC 1073 has two or to nonthermal radio emission from collections of supernova remnants. Such radio emission, with varying ratios of thermal

AGNs in H II Galaxies 5

TABLE

Supernova Rates Required for Radio Emission from H II Galaxies

Galaxy Center Entire Galaxy

a b

Galaxy Log P Log P

;nucleus SN ;galaxy SN

W Hz yr W Hz yr

NGC

The nuclear cm radio p owers are taken from Hummel et al

The galaxy cm radio p owers are derived from ux densities measured from the NRAO VLA

Sky Survey Condon et al and nonthermal components, is observed in starburst galaxies density of 2.4 Jy integrated over the entire galaxy (Griffith et such as NGC 253(Antonucci& Ulvestad 1988; Ulvestad& An- al. 1994). However, the flux density found by integrating over tonucci 1997; Johnson et al. 2001), M82 (Kronberg, Biermann, the central NGC 253 starburst in the 4.9-GHz image shown by & Schwab 1985; Muxlow et al. 1994), Henize 2-10 (Kobul- Ulvestad & Antonucci (1997) is only 1.3 Jy, corresponding to a − nicky & Johnson 1999; Beck, Turner & Gorjian 2001), and radio power of 9.6 × 1020 W Hz 1. The three H II nuclei whose NGC 5253 (Turner, Beck, & Ho 2000; Gorjian, Turner, & Beck detections are reported here have radio powers ranging from 2001), or in merger galaxies such as NGC 4038/9 (Neff & Ul- a factor of 2 (for NGC 4123) to a factor of 15 (for NGC 864) vestad 2000). The three H II galaxies with nuclear radio detec- lower than this central starburst in NGC 253. Thus, it is entirely tions that are shown in Figure 1 have radio sizes of a few hun- possible that the detected H II galaxies harbor nuclear star for- dred parsecs, indicating that any intense star formation must be mation somewhat weaker than that in NGC 253. Supporting confined to the near-nuclear regions of the galaxies. this possibility, Figure 2 shows the 6-cm image of NGC 253 (Ulvestad & Antonucci 1997), as it would appear at a distance of 20 Mpc instead of at 2.5 Mpc, having the flux density re- duced by a factor of 64, and angular resolution degraded by a factor of 8 in each dimension. The radio morphology of this degraded NGC 253 image looks similar to the H II galaxies in Figure 1, except that the respective peak and total flux densities are somewhat higher, at 8.2 mJy beam−1 and 20 mJy. The archival 6-cm data for NGC 864 and NGC 4535 were taken in the B configuration, and we also have measured a flux density from the B configuration 6-cm data for NGC 4123 (total flux density given in Table 2). Comparing the 6-cm values with the 20-cm A configuration flux densities (at similar angular resolution) published by Hummel et al. (1987) indicates that the resolved nuclear sources in NGC 4123 and NGC 4535 have +α respective two-point spectral indices (using Sν ∝ ν ) of −0.51 and −0.02; due to its non-detection at 20 cm, the spectral index of NGC 864 is limited to a value greater than −0.27. The rela- tively flat spectra of at least two of the objects may be indicative of radio emission dominated by optically thin thermal emission from star formation regions. For a thermal radio source, both the hydrogen emission lines and the free-free radio emission are powered by ionizing pho- tons from hot young stars. The radio power and Hα luminosity FIG. 3.— Plot of radio power vs. Hα luminosity for 40 Sc galaxies are related by (Filho, Barthel, & Ho 2002) with H II nuclei in the statistical sample. Radio detections and upper limits are distinguished by different symbols. The diagonal line is a prediction −1 of thermal radio emission based on the Hα luminosity, using P6 (W Hz ) −1 −20 −1 −20 −1 ≈ P (W Hz ) ≈ 10 L α (ergs s ) . (1) 10 LHα (ergs s ) (Filho et al. 2002). 6 H The 4.9-GHz radio sizes of NGC 864, NGC 4123, and NGC 4535 are quite similar to the 500–600 pc nuclear starburst For all 40 H II galaxies observed in our distance-limited sample in the nearby Sc galaxy NGC 253 (Antonucci & Ulvestad 1988; (those in Table 1 except for IC 467), Figure 3 plots the radio Ulvestad & Antonucci 1997). NGC 253 has a total 5-GHz flux power against the Hα luminosity given by Ho et al. (1997a)2. This figure shows that three of the five galaxies with the highest 2 As explained in Ho, Filippenko, & Sargent (2002), some of the Hα luminosities published in Ho et al. (1997a) have been updated with more accurate values from the literature. These are listed in Ho et al. (2002). 6 Ulvestad & Ho

the formalism for an object whose total radio emission is dominated by cosmic rays accelerated in supernova remnants, but not identiﬁable with speciﬁc remnants. In this case, for a spectral index of −0.5, Equation (18) of Condon (1992) becomes

P20 ∼ νSN −1 11 −1 . (3) 1022 W Hz ! yr ! For the nuclear radio sources, we can use the 20-cm radio emission found by Hummel et al. (1987) to estimate the nuclear supernova rates if the nuclear emission were entirely nonthermal (although the spectra appear thermal in two cases), and find values less than 0.01 yr−1 for all three galaxies. We also can compute the total supernova rates for the entire galaxies by using the galaxy flux densities determined from the NRAO VLA Sky Survey (Condon et al. 1998); this implies required supernova rates for the galaxies ranging from 0.01 yr−1 (NGC 864) to 0.03 yr−1 (NGC 4535). Results of the two different com- putations are summarized in Table 3. The star formation rates in stars above 5M⊙ (Condon 1992) required to account for the radio emission (presumed nonthermal) for the entire galaxies −1 −1 range from ∼ 0.3M⊙ yr to ∼ 1M⊙ yr . These values are roughly an order of magnitude higher than those required for the nuclear sources, whether they are dominated by thermal or nonthermal emission. However, they are consistent with predictions for the entire galaxies that are based on the far-infrared fluxes measured by the Infrared Astronomical Satellite (Soifer et al. 1989; Moshir et al. 1990), converted to star formation rates using the formulation summarized by Kennicutt (1998). FIG. 4.— Histogram of radio powers of 40 Sc galaxies with H II nuclei within 30 Mpc, compared to Seyfert galaxies from the same Palomar galaxy 4.3. Radio Emission Associated with Active Galactic Nuclei sample. Powers for the Seyfert galaxies have been taken from Ulvestad & Ho (2001). Upper limits are shown hatched. The other possibility for the radio sources in the H II nuclei of Sc galaxies is that they have radio cores associated with AGNs Hα luminosities harbor detected nuclear radio sources. Fur- and supermassive black holes. There has been considerable thermore, these three detected galaxies fall near the relation discussion in the literature about whether the core radio emis- given by Equation 1, supporting the inference (from their spec- sion might be correlated with black hole masses in these cases tra) that their radio emission has a significant thermal compo- (e.g., Franceschini, Vercellone, & Fabian 1998; Laor 2000). In nent. We note that the radio upper limits for some of the unde- fact, Ho (2002) contends that any correlation for weakly active tected galaxies actually are below the predictions based on the galaxies is an indirect consequence of a more fundamental rela- nuclear Hα luminosity, with the largest discrepancy existing for tionship between radio luminosity and optical bulge luminosity NGC 3877. Since the Hα luminosities were measured with a (or mass) as well as the established relationship between bulge 2′′ by 4′′ aperture, it may be that nuclear thermal radio sources luminosity and black hole mass. have been over-resolvedby the higher resolution VLA imaging. Since Sc galaxies such as those in our H II sample generally If the radio emission is thermal and related to starbursts, the have no well-defined bulges whose luminosities can be mea- inferred star formation rate also can be computed. Kennicutt sured, and no known black holes, direct predictions of radio (1998) gives the following formula for the star formation rate power cannot be made based on such properties. Instead, as a inferred from the Hα luminosity, assuming a Salpeter mass comparison, we have plotted a histogram of 6-cm nuclear ra- function for 0.1–100M⊙ and solar metallicity: dio powers for the 40 H II galaxies in our statistical sample in −1 −42 −1 SFR (M⊙ yr )=7.9 × 10 L(Hα) (ergs s ) . (2) Figure 4, and compared the results to a similar plot for the sta- For our three detected nuclei, the Hα luminosities range from tistical sample of 45 Seyfert nuclei (39 with galaxy types earlier 1040 ergs s−1 to 1041 ergs s−1 so the inferred nuclear star forma- than Sc) that was published by Ulvestad & Ho (2001). Applica- tion rates that could account for the Hα luminosity would be tion of the Gehan’s Generalized Wilcoxon Test (Isobe & Feigel- −1 0.08 − 0.8M⊙ yr . Alternatively, using the formalism given by son 1990; LaValley, Isobe, & Feigelson 1992) indicates that the Condon (1992), the inferred star formation rates in stars above probability that the Seyferts and H II galaxies were drawn from −1 −4 5M⊙ are only 0.02 − 0.2M⊙ yr . the same parent population in radio luminosity is less than 10 . Another possible result of a circumnuclear starburst would This result would be expected even if the H II galaxies contain be a compact radio source with a steep spectrum, made up of AGNs, since they have faint or non-existent bulges, and there complexes of supernova remnants that are unresolved by the are no active galaxies with faint bulges that have high radio VLA. Examples of these have been imaged at higher resolution powers (Ho 2002). by the Very Long Baseline Array (VLBA), in Arp 220 (Smith In general, the radio sources in low-luminosity AGNs are et al. 1998) and Mrk 231 (Taylor et al. 1999). The total radio quite compact. For the Seyferts from the Palomar sample, Ho powers can be used to estimate the associated supernova rates & Ulvestad (2001) showed that about half the objects have di- that would be required to account for the radio sources. We use ameters of ∼ 100 pc or less. In the case of LINERs, Nagar et AGNs in H II Galaxies 7

−1 al. (2000) demonstrated that those with ﬂat radio spectra typ- emission for star formation rates of only 0.08 − 0.8M⊙ yr . ically have AGN radio cores with diameters less than 25 pc, However, nonthermal emission from supernova remnants also and Falcke et al. (2000) used the VLBA to show that these could contribute to the detected nuclear sources, requiring su- cores typically are parsec-scale or smaller. The radio peaks of pernova rates of well under 0.01 yr−1 and star formation rates −1 NGC 864, NGC 4123, and NGC 4535, having size limits less of ∼ 0.1M⊙ yr . than 50–150 pc, are consistent with the presence of AGNs but To the extent that radio cores are a common feature in low- are not required to be black-hole powered. Instead, the trend luminosity AGNs, the nondetection of radio cores in H II nuclei of decreasing peak ﬂux density with higher resolution seen in strongly suggests that they intrinsically lack AGNs, and, by in- NGC 4123 (see Section 4.1) indicates that those peaks probably ference, massive black holes. This is not to say that late-type would be resolved if somewhat smaller spatial scales (longer galaxies never host AGNs. Indeed, ∼15% of the Seyfert nu- baselines) were sampled. If so, AGNs are unlikely to contribute clei and ∼10% of all AGNs in the Palomar survey are hosted in much to the currently unresolved radio emission. If there really galaxies with Hubble types Sc or later. In terms of AGN demo- are weak AGNs causing the most compact radio emission, they graphics, the present results imply that the existing statistics of should be detectable with the VLBA, as are many other low- AGNs in late-type galaxies based on optical searches (Ho et al. luminosity AGNs (Falcke et al. 2000; Nagar et al. 2002). 1997b)are likely to be robust. Thereis not a large populationof AGNs hidden among late-type galaxies that have H II spectra. 5. SUMMARY We have searched for nuclear radio sources in a statistical sample of 40 Sc galaxies hosting H II nuclei, using imaging of We thank the anonymous referee for extremely useful com- new and archival VLA data at 6 cm. Only three such sources ments that helped clarify and focus the paper. This research were detected, in galaxies among those with the largest Hα lu- has made use of the NASA/IPAC ExtragalacticDatabase (NED) minosities in the sample. The radio powers and morphologies which is operated by the Jet Propulsion Laboratory, California are consistent with the association of the detected radio sources Institute of Technology, under contract with the National Aero- with nuclear starbursts similar to that in NGC 253, rather than nautics and Space Administration. In addition, this research has being caused by active galaxies powered by massive black made use of NASA’s Astrophysics Data System Abstract Ser- holes. The ratio of radio power to Hα luminosity is generally vice. The work of L. C. H. is partly funded by NASA grants consistent with thermal radio emission from nuclear starbursts, awarded by the Space Telescope Science Institute, which is op- and the detected central sources could produce such thermal erated by AURA, Inc., under NASA contract NAS5-26555.

REFERENCES

Antonucci, R. R. J., & Ulvestad, J. S. 1988, ApJ, 330, L97 Kormendy, J., & Gebhardt, K. 2001, in 20th Texas Symposium on Relativistic Baars, J. W. M., Genzel, R., Pauliny-Toth, I. I. K., & Witzel, A. 1977, A&A, Astrophysics, AIP Conf. Proceedings 586, ed. J. C. Wheeler & H. Martel 61, 99 (New York: American Institute of Physics), 363 Beck, S. C., Turner, J. L., & Gorjian, V. 2001, AJ, 122, 1365 Kronberg, P. P., Biermann, P., & Schwab, F. R. 1985, ApJ, 291, 693 Colbert, E. J. M., & Ptak, A. F. 2002, ApJS, in press (astro-ph/0204002) Laor, A. 2000, ApJ, 543, L111 Condon, J. J. 1992, ARA&A, 20, 575 LaValley, M., Isobe, T., & Feigelson, E. D. 1992, in ASP Conf. Ser. 25, Condon, J. J., Cotton, W. D., Greisen, E. W., Yin, Q. F., Perley, R. A., Taylor, Astronomical Data Analysis Software and Systems I, ed. D. M. Worrall, C. G. B., & Broderick, J. J. 1998, AJ, 115, 1693 Biemesderfer & J. Barnes (San Francisco: ASP), 245 Cotton, W. D., Condon, J. J., & Arbizzani, E. 1999, ApJS, 125, 409 Magorrian, J., et al. 1998, AJ, 115, 2285 Falcke, H., Nagar, N. M., Wilson, A. S., & Ulvestad, J. S. 2000, ApJ, 542, 197 Merritt, D., Ferrarese, L., & Joseph, C. L. 2001, Science, 293, 1116 Ferrarese, L. & Merritt, D. 2000, ApJ, 539, L9 Moshir, M., et al. 1990, IRAS Faint Source Catalog, Version 2.0 Ferrarese, L., Pogge, R. W., Peterson, B. M., Merritt, D., Wandel, A., & Muxlow, T. W. B., Pedlar, A., Wilkinson, P. N., Axon, D. J., Sanders, E. M., & Joseph, C. L. 2001, ApJ, 555, L79 de Bruyn, A. G. 1994, MNRAS, 266, 455 Filho, M. E., Barthel, P. D., & Ho, L. C. 2002, ApJS, in press Nagar, N. M., Falcke, H., Wilson, A. S., & Ho, L. C. 2000, ApJ, 542, 186 (astro-ph/0205196) Nagar, N. M., Falcke, H., Wilson, A. S., & Ulvestad, J. S. 2002, A&A, in press Franceschini, A., Vercellone, S., & Fabian, A. C. 1998, MNRAS, 297, 817 (astro-ph/0207176) Gebhardt, K., et al. 2000a, ApJ, 539, L13 Neff, S. G., & Ulvestad, J. S. 2000, AJ, 120, 670 ——. 2000b, ApJ, 543, L5 Sandage, A. R., & Tammann, G. A. 1981, A Revised Shapley-Ames Catalog ——. 2001, AJ, 122, 2469 of Bright Galaxies (Washington, DC: Carnegie Inst. of Washington) Gorjian, V., Turner, J. L., & Beck, S. C. 2001, ApJ, 554, L29 Smith, H. E., Lonsdale, C. J., Lonsdale, C. J., & Diamond, P. J. 1998, ApJ, Grifﬁth, M. R., Wright, A. E., Burke, B. F., & Ekers, R. D. 1994, ApJS, 90, 493, L17 179 Soifer, B. T., Boehmer, L., Neugebauer, G., & Sanders, D. B. 1989, AJ, 98, Ho, L. C. 2002, ApJ, 564, 120 766 Ho, L. C., Filippenko, A. V., & Sargent, W. L. W. 1995, ApJS, 98, 477 Taylor, G. B., Silver, C. S., Ulvestad, J. S., & Carilli, C. L. 1999, ApJ, 519, 185 ——. 1997a, ApJS, 112, 315 Thompson, A. R., Clark, B. G., Wade, C. M., & Napier, P. J. 1980, ApJS, 44, ——. 1997b, ApJ, 487, 568 151 ——. 2002, ApJ, submitted Tremaine, S., et al. 2002, ApJ, 574, 740 Ho, L. C., & Ulvestad, J. S. 2001, ApJS, 133, 77 Tully, R. B. 1988, Nearby Galaxies Catalogue (Cambridge: Cambridge Univ. Hummel, E., van der Hulst, J. M., Keel, W. C., & Kennicutt, R. C., Jr. 1987, Press) A&AS, 70, 517 Turner, J. L., Beck, S. C., & Ho, P. T. P. 2000, ApJ, 532, L109 Isobe, T., & Feigelson, E. D. 1990, BAAS, 22, 917 Ulvestad, J. S., & Antonucci, R. R. J. 1997, ApJ, 488, 621 Johnson, K. E., Kobulnicky, H. A., Massey, P., & Conti, P. S. 2001, ApJ, 559, Ulvestad, J. S., & Ho, L. C. 2001, ApJ, 558, 561 864 Van Dyk, S. D., Lacey, C. K., Sramek, R. A., & Weiler, K. W. 1999, IAUC Kennicutt, R. C., Jr. 1998, ARA&A, 36, 189 7322 Kobulnicky, H. A., & Johnson, K. E. 1999, ApJ, 527, 154 van Moorsel, G., Kemball, A., & Greisen, E. 1996, in Astronomical Data Analysis Software and Systems V, ed. G. H. Jacoby & J. Barnes (San Francisco: ASP), 37