CEPHEUS OB2: DISK EVOLUTION AND ACCRETION AT 3 10 Myr 1

Transcription

1 The Astronomical Journal, 130: , 2005 July # The American Astronomical Society. All rights reserved. Printed in U.S.A. A CEPHEUS OB2: DISK EVOLUTION AND ACCRETION AT 3 10 Myr 1 Aurora Sicilia-Aguilar, 2 Lee W. Hartmann, 2 Jesús Hernández, 2,3,4 César Briceño, 3 and Nuria Calvet 2,4 Receivved 2004 December 20; accepted 2005 March 23 ABSTRACT We present the results of MMT observations of young stars for our study of protoplanetary disks at ages 1 10 Myr in two young clusters located in the Cepheus OB2 region: Trumpler 37 (embedded in the H ii region IC 1396) and NGC Using low-resolution optical spectra from the Hectospec multifiber spectrograph, we have tripled the number of known low-mass cluster members, identifying 130 new members in Tr 37 and 30 in NGC We use indicators of youth ( Li absorption at ) and accretion/chromospheric activity ( H emission) to identify and classify the low-mass cluster members. We derive spectral types for all the low-mass candidates and calculate the individual extinctions and the average over the clusters. With the extended member samples, we estimate the disk fraction in the clusters, finding that 40% of the low-mass stars in Tr 37 are actively accreting, whereas only 1 of the 55 low-mass stars in NGC 7160 shows indications of accretion. Optical photometry and theoretical isochrones are used to determine the age of the cluster members, confirming the estimates of 4 Myr for Tr 37 and 10 Myr for NGC Accretion rates in Tr 37 (10 8 M yr 1 on average) are derived from U-band photometry. We find that only 50% of the accreting stars have near-ir excesses (from 2MASS), which could be due to the geometry of their disks or be an indication dust of settling/grain growth. Finally, we study the high- and intermediate-mass members of the clusters. With the extended member list, we revise the spatial distribution of stars with disks. Our results are crucial for interpreting Spitzer Space Telescope studies of accretion disks at the ages of planet formation (3 10 Myr). Key words: accretion, accretion disks planetary systems: protoplanetary disks stars: pre main-sequence Online material: machine-readable tables 1. INTRODUCTION The evolution of protoplanetary disks between the ages 3 and 10 Myr is not well understood. Whereas accretion disks are very common at an age of 1 Myr, they become rare around 10 Myr old stars (Calvet et al. 2002), and no optically thick disks have been found around solar-type stars older than 30 Myr (Strom et al. 1993; Hillenbrand et al. 1998; Lada et al. 2000; Muzerolle et al. 2000; Briceño et al. 2001). Therefore, most disks seem to dissipate, presumably after forming planets, during the epoch from 3 to 10 Myr, which is also consistent with evidence from the study of meteorites (Podosek & Cassen 1994). Because of the fact that the long survival time of some protoplanetary disks suggests different disk/angular momentum properties ( R. Hueso & T. Guillot 2005, in preparation), and taking into account the probable dependence of disk evolution with mass ( Hartmann et al. 1998), the study of disk characteristics and disk removal requires statistically significant samples of stars in different mass ranges. The best places to obtain these samples are young open clusters, which have the advantage of having a small scatter in distance, age, and (in clusters where most of the gas has been dissipated) extinction. The main drawback of this approach is that there are no nearby clusters aged 3 10 Myr, so a sound method and powerful instruments are required in order to identify members and study their properties. We have been carrying out a study of two clusters in the Cepheus OB2 region, Trumpler 37 and NGC 7160 (Sicilia- 1 Observations reported here were obtained at the MMT Observatory, a joint facility of the Smithsonian Institution and the University of Arizona. 2 Center for Astrophysics, 60 Garden Street, Cambridge, MA 02138; asicilia@ cfa.harvard.edu, hartmann@cfa.harvard.edu, ncalvet@cfa.harvard.edu. 3 Centro de Investigaciones de Astronomía, Apartado Postal 264, Mérida 5101-A, Venezuela; jesush@cida.ve, briceno@cida.ve. 4 Postgrado de Física Fundamental, Universidad de Los Andes, Mérida 5101-A, Venezuela. 188 Aguilar et al. 2004, hereafter Paper I), using optical photometry and spectroscopy to find the young, low-mass members. The Cep OB2 region with its two clusters, Tr 37 and NGC 7160 (Platais et al. 1998), is a prime candidate for searching for young stars aged 3 10 Myr. Located at a distance of 900 pc (Contreras et al. 2002), Cep OB2 is a bubble-shaped structure of atomic and molecular gas (Simonson & van Someren Greve 1976; Patel et al. 1994, 1998) containing the cluster NGC 7160 near its center and Tr 37 (with the O6 star HD exciting the H ii region IC 1396) near the rim (Garrison & Kormendy 1976; Ábrahám et al. 2000). The presence of several B stars in NGC 7160 suggests that the structure was probably formed a few million years ago by one or more supernovae exploding inside a bubble blown away by stellar winds (Balázs & Kun 1989; Duvert et al. 1990; Patel et al. 1995, 1998; Codella et al. 2001). Kinematic and stellar ages (Patel et al. 1998) support this picture, with an expansion age for the Cep OB2 bubble of about 7 10 Myr, which would result in an age of 10 Myr for NGC 7160 and around 3 Myr for Tr 37. In Paper I, we identified 40 stars in Tr 37 and 15 in NGC 7160, confirming the age estimates suggested by the kinematics. We determined the membership using young star features present in optical spectra (Li k6707 absorption and H emission lines) and extinction. We found that accretion disks with accretion rates around 10 8 M yr 1 are a frequent phenomenon in Tr 37, whereas no accreting stars were found in NGC 7160, suggesting substantial disk evolution. Here we present a major extension of our work in Tr 37 and NGC Using the 300 fibers of the multifiber optical spectrograph Hectospec (Fabricant et al. 1994), we were able to observe most of the stars with V ¼ above the main sequence (MS) in the V versus V I color-magnitude diagram (according to the isochrones from Siess et al. 2000) in Tr 37 and NGC Observations are detailed in x 2. Section 3 describes the identification of cluster members. Following the method we used in Paper I, we identify a total of 165 members in Tr 37, of

2 CEP OB2: DISK EVOLUTION AND ACCRETION 189 Fig. 2. Observations in NGC The 4-Shooter fields are indicated by large rectangles. Hectospec targets are shown as crosses. BAF-type cluster members and the most likely members are displayed as small stars; G-type stars are represented by pentagons (confirmed members, filled; likely members, open). For the low-mass members, CTTSs are represented by circles and WTTSs are represented by triangles; filled symbols denote confirmed members or those where Li absorption is resolved in the Hectospec/Hydra/FAST spectra, and open symbols denote likely members where Li is not detectable. Fig. 1. Observations in Tr 37. The 4-Shooter fields are indicated by large rectangles. Hectospec targets are shown as crosses. The O6 star HD is indicated as a large star, and high- and intermediate-mass ( BAF-type) cluster members and the most likely members are displayed as small stars. For the lowmass members, CTTSs are represented by circles and WTTSs are represented by triangles; filled symbols denote confirmed members or those where Li absorption is detected in the Hectospec/ Hydra/ FAST spectra, and open symbols denote likely members where Li is not detectable because of poor S/ N. which 140 have been observed with Hectospec, and 55 members in NGC 7160, of which 35 came from the Hectospec observations. We include here all the members found in Paper I and revise the information for those that were observed again with Hectospec. We also include here a preliminary study of the high- and intermediate-mass stars in the two clusters, expanding the work done by Contreras et al. (2002) on Tr 37. Similarly, we study the early-type candidates (spectral types B, A, and F) in NGC With the new identifications, we are able to do a statistical analysis of the clusters in x 4. We estimate the ages of the clusters using isochrones by Siess et al. (2000), confirming the values of 3 5 Myr for Tr 37 and 10 Myr for NGC 7160 and finding some evidence of a possible younger (1 Myr) population in Tr 37 (x 4.1). In x 4.2, we revise the spatial distribution of the cluster members in Tr 37 and explore the hypothesis of a younger, triggered population. We also consider the possibility of disk evaporation being responsible for the lower disk fraction in the center and to the east of the cluster. Disk fractions and accretion rates are estimated in x 4.3 (40% disk frequency in Tr 37, in contrast to only one active accretor in NGC 7160 out of 55 identified members). The membership determination is a necessary step in our study of disk properties (of both accretion and debris disks) in the range AU, using photometry in JHK, and from 3.6 to 24 m to study the dust emission in a forthcoming paper (A. Sicilia-Aguilar et al. 2005, in preparation, hereafter Paper III), in which we will present the results of the Spitzer Space Telescope IRAC (Fazio et al. 2004) and MIPS (Rieke et al. 2004) observations in Cep OB2. 2. OBSERVATIONS AND DATA REDUCTION 2.1. UVRI Photometry We carried out our optical observations at the Fred Lawrence Whipple Observatory (FLWO) in Mount Hopkins, Arizona, between 2000 and We obtained optical VR C I C (hereafter VRI ) and ultraviolet U photometry using the 4-Shooter CCD camera mounted on the 1.2 m telescope (Caldwell & Falco 1993). The 4-Shooter camera is an array of four CCD chips of 1048 ; 1048 pixels and an approximate field of view of 23 0 ; VRI photometry was done in several runs between 2000 and 2002, as described in Paper I, and was used for selecting the candidates for spectroscopic follow-up. The RI imaging was repeated over several nights (4 7) in this period in order to include a variability study of stars above the MS in our candidate selection, as described in Paper I. As we stated there, variability is an unbiased criterion for finding young stars with and without disks, since both accreting and nonaccreting stars are variable because of accretion processes and stellar spots (Hartmann 1998; Briceño et al. 2001, 2005; Rhode et al. 2001). We took sets of three short (3 ; 30 s in VR, 3; 45 s in I ) and three long (3 ; 300 s) exposures using the VRI filters in order to effectively remove cosmic rays. We reduced the data using IRAF 5 tasks for mosaic reduction available in the package MSCRED, and we obtained instrumental magnitudes by performing aperture photometry with an aperture of 6 pixels, corrected to 10 pixels, using IRAF tasks in the package APPHOT. The data were calibrated photometrically using standard stars in the Landolt catalog ( Landolt 1992). The astrometry was performed using IRAF tasks in the IMCOORD package, producing coordinates with an accuracy better than 0B3. In total, we observed eight fields in Tr 37, covering an approximate area of 1 diameter (see Fig. 1), and three fields in NGC 7160 (see Fig. 2). We took 3 ; 1800 s exposures with the U filter, covering all the fields in Tr 37 and the two flanking fields in NGC Because of the presence of the bright O6 star HD in Tr 37, as well as some B stars, which can saturate a large surface of the chips during the 1800 s exposures, we obtained several fields by moving the telescope to place the bright stars at the edges and between the chips, so the saturated stars would cover a small surface. Reduction, photometry, and calibration were performed in a similar way to the VRI data. We also obtained coordinates for the U-band detections and matched them with their optical counterparts, creating a UVRI catalog of objects in the fields in Tr 37 and NGC The optical data for the Tr 37 and NGC 7160 members are listed in Tables 1 and 2, respectively Hectospec/MMT The spectroscopic follow-up of the candidates was done using the Hectospec multifiber spectrograph mounted on the 6.5 m MMT telescope in Mount Hopkins (Fabricant et al. 5 IRAF is distributed by the National Optical Astronomy Observatory, which is operated by the Association of Universities for Research in Astronomy, Inc., under cooperative agreement with the National Science Foundation.

3 TABLE 1 Summary of 4-Shooter Optical and 2MASS Near-IR Observations for Low-Mass Stars in Tr 37 ID1 2MASS V V R V I J H K Variability ::: ::: RI RI RI RI RI No RI RI RI No No RI RI RI No RI RI RI RI RI RI RI ::: ::: No ::: ::: RI No RI No I No No No RI RI RI RI I No RI RI RI No 190

4 TABLE 1 Continued ID1 2MASS V V R V I J H K Variability No No No No ::: ::: No RI No No RI RI No No RI RI RI RI No No RI No No RI ::: ::: No RI RI RI No RI RI RI RI RI RI RI RI RI RI No RI RI No RI No No RI

5 192 SICILIA-AGUILAR ET AL. Vol. 130 TABLE 1 Continued ID1 2MASS V V R V I J H K Variability ::: I ::: ::: RI RI No RI RI No RI No RI RI RI RI RI RI No RI RI RI No RI Notes. Table 1 is also available in machine-readable form in the electronic edition of the Astronomical Journal. ID1 corresponds to the 4-Shooter ID, and 2MASS is the 2MASS identification for each star. Errors are given for the 2MASS magnitudes; the lack of definite errors in 2MASS is an indication of possible problems in the photometry. Photometric errors in the 4-Shooter visual magnitudes are very small; the larger deviations are produced by the intrinsic variability of the stars. In the column headed Variability (see text), RI ¼ observed significant variability (over 3 ) in both the R and I bands, I ¼ observed significant variability in the I band, and No ¼ no significant (over 3 ) variations detected in our variability study. 1994). Hectospec is a low-resolution optical spectrograph with a total of 300 fibers that can be placed within a 1 diameter field. We used the 270 groove mm 1 grating, obtaining spectra in the range , with a resolution of The field position is fixed using 2 to 3 guide stars. Very accurate coordinates (better than 0B3) are required for fiber positioning; the fibers subtend a circle of 1B5 diameter. Since guide stars are drawn from the Two Micron All Sky Survey (2MASS) catalog (Cutri et al. 2003), we used the 2MASS coordinates for our optical candidates. Using Hectospec/MMT, we improved significantly the signal-to-noise ratio (S/N) of the spectra; therefore, Li k6707 absorption, an indicator of youth (Randich et al. 2001; Hartmann 2003), was detectable in about 90% of the low-mass, young stars in the clusters. Moreover, the 300 fibers (compared to the ~90usefulfibersinHydraortheonestaratatimeofFAST) allowed us to observe a very numerous sample, covering most of our candidate list in 1.5 nights observing time, in 2004 June and November. The selection of candidates for the spectroscopic follow-up was based on VRI photometry, RI variability, and U-band and 2MASS colors. For the two clusters, we selected all the stars with V ¼ that appeared over the 100 Myr isochrone in the V versus V I color-magnitude diagram, using the isochrones from Siess et al. (2000). We applied an average extinction correction using A V ¼ 1:5, consistent with the average extinction obtained in Contreras et al. (2002) and Paper I. Individual extinction variations do not much affect the selection, since the extinction vector runs nearly parallel to the isochrones for the low-mass stars. For the run in 2004 June, we gave higher priority to RI variable stars (see Briceño et al. 2001, 2005; in Paper I we found that 40% of our detected variables are young) and to the objects showing near-ir excess in 2MASS colors and U excess in the V I versus U V diagram. The selection basedon2massandu excesses is obviously biased toward accreting stars and so could affect the disk fraction estimates. Although both accreting and nonaccreting stars show variability (Herbst et al. 1994; Briceño et al. 2001; Rhode et al. 2001), variations in accreting stars show larger amplitudes, and this may result in a bias as well. Before the run in 2004 November, nearly 95% of the objects with higher priority had been observed. We included the remaining low-priority objects, together with a sample of stars showing IRAC excesses, 2MASS counterparts (for coordinate verification), and no optical counterparts. Most of these stars presented JHK magnitudes consistent with

6 No. 1, 2005 CEP OB2: DISK EVOLUTION AND ACCRETION 193 TABLE 2 Summary of 4-Shooter Optical and 2MASS Near-IR Observations for Low-Mass Stars in NGC 7160 ID1 2MASS V V R V I J H K Variability No No No R No No No RI No No No No I RI R No R R R No R R R Notes. Table 2 is also available in machine-readable form in the electronic edition of the Astronomical Journal. See notes for Table 1. spectral types later than M1 (which is at the detection limit of our VRI photometry). The sample of stars drawn from our IRAC catalog is obviously biased toward stars with disks; therefore, we must be cautious when estimating accretion disk fractions. Since most of the stars from the IRAC-selected sample are below our visual magnitude limit or between the 4-Shooter chips and they are all stars with disks, we do not consider them for disk fraction estimates, as we explain later on. The disk fraction estimates (see x 3.4) are obtained for stars of spectral types late G to M2 in the central fields, and we calculate the errors depending on the number of potential objects that may be missing and whether they are probably classical T Tauri stars (CTTSs) or weak-lined T Tauri stars (WTTSs). In 2004 June, we observed three configurations in Tr 37, repeating 3 ; 900 or 720 s integrations, plus two configurations in NGC 7160, repeating two integrations of 900 s. For the run in 2004 November, we observed one field in each cluster, obtaining 3 ; 900 s integrations. The objects and regions covered in both clusters are shown in Figures 1 and 2. In order to make an unbiased estimation of the disk fraction in Tr 37, we gave priority to obtaining spectra of the objects within the two central fields in this cluster, achieving very complete coverage of the objects above the MS in this region (see Fig. 1). In order to avoid the difficulties encountered in Paper I when subtracting background emission lines from the highly variable H ii region, we took 600 s exposures offset by 5 00 after each one of the Tr 37 configurations. This way, we obtained a sky spectrum very close to each star, taken with the same fiber. In the case of NGC 7160, because of the lack of strong nebular emission found in Paper I, we simply selected a large number of sky fibers (50) and subtracted an averaged sky for each field. The data reduction was performed by S. Tokarz through the CfA Telescope Data Center, using IRAF tasks and other customized reduction scripts. The reduction procedure was the standard for Hectospec data, except for the special individual sky subtraction in Tr 37. The background subtraction for NGC 7160 objects was performed using the IRAF program SKYSUB. For the Tr 37 fields, the background subtraction was done separately using IRAF task sarith to remove the wavelengthcalibrated and normalized sky spectrum from the corresponding wavelength-calibrated and normalized object spectrum. A sample of spectra of Tr 37 and NGC 7160 members is displayed in Figure FAST/1.5 m We also selected a sample of bright stars in NGC 7160 for spectroscopic follow-up in order to complete the region of BAF

7 194 SICILIA-AGUILAR ET AL. Vol. 130 Fig. 3. Sample of Hectospec spectra from stars in Tr 37 and NGC The hydrogen Balmer series in emission ( H, H, andh) is labeled. Insets reveal the Li i k6707 absorption line. stars in the H-R diagram, checking the distance estimate in Contreras et al. (2002) and producing a sample to study debris disks with Spitzer. Since our 4-Shooter/1.2 m photometry saturates for magnitudes brighter than V ¼ 14, we selected our sample of candidates from the literature. We chose those stars consistent with the MS in the V versus B V diagram, using the colors given in De Graeve (1983) and the Siess et al. (2000) isochrones. We also revised the Tr 37 sample from Contreras et al. (2002; obtained from Marschall & van Altena 1987; Kun et al. 1987). The spectra of the early- and intermediate-type candidates were obtained using the FAST slit spectrograph mounted on the 1.5 m telescope in FLWO (Fabricant 1994; Fabricant et al. 1998), equipped with the Loral 512 ; 2688 CCD. We used the standard configuration intended for FAST COMBO projects, a 300 groove mm 1 grating and a 3 00 wide slit. This combination offers of spectral coverage centered at , witha resolution of about 6 8, which is comparable to the Hectospec spectra. The candidates were observed during several queued runs from 2001 July until 2003 July (observers: P. Berlind & M. Calkins), using the positions available in the literature. Exposure times ranged from 60 to 300 s for these stars with magnitudes V ¼ The spectra were reduced at the CfA (S. Tokarz) using software developed specifically for the FAST COMBO observations. 3. IDENTIFICATION OF CLUSTERS MEMBERS 3.1. Low-Mass Stars As we described in Paper I, the membership of young, low-mass (G M) stars can be established using low-resolution, optical spectra. There are two main features at optical wavelengths that can be used for the identification of the spectra of young stars: H emission at and lithium absorption at H emission (and, in general, all the hydrogen Balmer series emission lines, which are not always detectable) is produced in the accreting columns of accreting CTTSs (see Calvet & Hartmann1992; Hartmann et al. 1994; Hartmann 1998; Muzerolle et al. 1998). Weaker H emission is produced in the chromospheres of nonaccreting WTTSs. In the case of H emission from accretion processes, the large velocities of the accreting material produce significant velocity wings (which can be up to 200 km s 1 ) and, in general, the equivalent widths (EWs) of the lines are larger than what is observed in WTTSs.

8 No. 1, 2005 CEP OB2: DISK EVOLUTION AND ACCRETION 195 Several criteria have been established to distinguish accreting and nonaccreting stars based on their H EWs (Herbig & Bell 1988; Hartmann 1998; Muzerolle et al. 2001). Here we use the criteria of White & Basri (2003), considering as accreting CTTSs those with H emission EW > 20 8 for stars later than M2, EW > 10 8 for spectral types M2 K5, and EW > 3 8 for spectral types K5 G. The main problem in the detection of H is that very frequently, young stellar clusters are embedded in inhomogeneous H ii regions, which produce highly variable emission in H superposed on the emission of the stars. As we described in Paper I, we require a very careful background subtraction to avoid the over- and undersubtraction of nebular H. This would lead to confusion, affecting primarily the WTTSs, and is an important problem in Tr 37, where the nebular emission is very strong and inhomogeneous, as we observed in Paper I. In contrast, there is no significant nebular emission in NGC 7160, where we used a standard average sky spectra. Using Hectospec spectra and their corresponding sky offset of 5 00 (producing sky spectra from a location very close to the star and with the same fiber), we achieve a very clean background subtraction, improving our identification of WTTSs in the cluster. Since we observe that the multiple background and night-sky lines are properly subtracted using this method, we can assume that the error in the subtraction of nebular H is minimal, being able to detect clearly EW as low as and confirming the youth of the candidates with very low H EW via the presence of Li absorption. The second indicator of the youth of low-mass stars is the Li absorption line at , since Li is depleted from low-mass, partially convective stars in timescales 30 Myr (Strom et al. 1989; Briceño et al. 1997, 1998; Randich et al. 2001; Hartmann 2003). Although Li EWs are usually in absorption, requiring good S/N in the region, Li absorption is a very good criterion for determining the youth of the star, since it does not depend on the quality of the background subtraction as H does. It is also very efficient at distinguishing contaminating dme stars, of which we found a couple of dozen among the Hectospec spectra. These stars are foreground and background M-type (sometimes late K) stars with weak H emission due to their chromospheric activity, but they are older and do not show Li absorption. We consider all the low-mass star identifications where Li was detected to be safe. Using Li, we could also establish the authenticity of some of our weakest H measurements, concluding that H EWs as low as are true detections and not background contamination. Because of the much better S/N of the Hectospec spectra compared to the Hydra and FAST spectra used in Paper I, we are able to detect Li absorption (or to determine the absence of it) in more than 90% of the stars, so both the contamination of our final member list and the rejection of members among the observed stars are very small. In the few cases where some doubts arose, we used, in addition, the extinction values as an extra criterion for membership and, as we did in Paper I, we preferred to be conservative in the identification of members rather than include low-probability members in our study. The values of H and Li EWs for the low-mass members in Tr 37 and NGC 7160 are listed in Tables 3 and 4, respectively, including the derived membership based on these detections. For completeness, and since they will be used for statistical purposes, we include the members found in Paper I. Plots of the 4-Shooter V versus V I colors of the cluster members, compared to the isochrones by Siess et al. (2000), are shown in Figure 4. The J H versus H K diagrams constructed using the 2MASS colors (Cutri et al. 2003) are displayed in Figure Spectral Types and Extinction We classified our objects following the spectral classification scheme of Hernández et al. (2004). This scheme is optimized for the FAST wavelength range and resolution, which are similar to those of Hectospec. We used two different sets of indices, the first for stars with spectral types in the range from O to early G and the second for stars later than G5. The sets of indices were calibrated as a function of the spectral type, using a collection of FAST spectra from O8 M6 MS standard stars (García 1989; Gray et al. 2001; Keenan & Barnbaum 1999; Jaschek 1978; Buscombe 2001). For each index, the width and centers of the blue continuum band (BCB), red continuum band (RCB), and central band (FB) were optimized. The scheme for early stars (O8 early G) is based on 33 spectral indices that are sensitive to changes in T eff but insensitive to reddening, stellar rotation, luminosity class, and S/N. This scheme is designed to largely avoid problems caused by nonphotospheric contributions (mostly originating from material surrounding the star). We require that the various spectral types calculated from each index agree with the others; wildly discrepant values are rejected and a weighted mean spectral type is obtained. The errors in our estimated spectral types have two contributions, the error from the fit of each index to the standard MS and the error in the measurement of each index (Hernández et al. 2004). In order to classify stars with spectral type later than G5, we extended the scheme, incorporating and optimizing spectral indices in the spectral type ranges G0 K5 and K5 M6 (see Hernández et al. 2004; Briceño et al. 2005). Following Hillenbrand (1995), the EW for each spectral feature is obtained by measuring the decrease in flux in FB due to line absorption, considering the continuum expected when interpolating between the two adjacent bands, BCB and RCB. Using the V I color predicted by Kenyon & Hartmann (1995) for stars with different spectral types, our VI photometry, and the relation between A V and E(V I ) from Bessell & Brett (1988), A V ¼ 2:5E(V I ), we calculated the individual extinction of each star. These values are listed in Tables 3 and 4. We used the confirmed members (indicated by Y in Tables 3 and 4) to calculate the average extinction in each cluster, as well as the standard deviation of the extinction within the cluster, obtaining A V ¼ 1:56 (with a standard deviation of 0.55 over the entire cluster) for Tr 37 (using the 135 confirmed members) and A V ¼ 1:17 (with a standard deviation equal to 0.45) for NGC 7160 (using the 39 confirmed members). The values of the extinction of the probable members (P in the tables) are consistent with those of the confirmed members and are very close to those obtained in Paper I from a much reduced sample of stars. Except for stars very close to the globules to the west and north in Tr 37, we do not find an evident spatial variation of the extinction within the clusters. Therefore, we assume that extinction is mostly foreground and that the standard deviation of A V is produced by both individual extinction variations and the error derived from spectral typing, colors, and photometry. An analysis of the statistical distribution of A V for the confirmed members in both clusters shows that it can be considered as pseudo-gaussian (as for the probabilities, note the absence of negative A V values), with 76% of the stars in Tr 37 in the 1 region, 96% of stars in the 2 region, and 98% of them in the 3 region. (Only three stars, related to the globules, have higher extinctions beyond 3.) For NGC 7160, the percentages

9 TABLE 3 Summary of Spectroscopy, Extinction, and Membership for Low-Mass Stars in Tr 37 ID1 Instrument Spectral Type H EW (8) Li EW Age (8) TTS Type A V ( Myr) Member Comments Hecto K w: Y IR Hecto K5.0? w: P Hecto M c Y IR Hecto K c Y IR Hecto K c Y L, IR Hecto M c Y L, IR Hecto K c Y IR Hecto M c Y IR Hecto K w P Hecto G c Y IR Hecto M c P Hecto K w: P Hecto c P Hecto K c P Hydra K c Y Yg, L, IR Hecto K w Y Hecto K c Y Yg, L, IR Hydra Hydra K w Y Yg, L, IR Hecto M c P Hecto K w Y Hecto M w Y Hecto M c Y Yg, L, IR Hydra Hecto M c: Y FAST K w Y Yg, IR Hecto M c Y Yg, IR Hecto K c Y L, IR Hydra Hecto M w Y Hecto M c P Hecto M w Y Yg, IR Hecto K c Y IR Hydra Hecto K w Y Hecto K c Y IR Hydra Hecto K c P Hecto K w Y Hydra K c Y IR Hecto K w Y Hecto K w Y Hecto M c P Hecto K c Y IR FAST Hecto M w Y FAST G w: P Hydra M c Y IR Hecto K c Y L, IR Hydra Hydra G K c Y Hecto M w Y Yg Hecto K c Y Yg, IR Hydra Hecto M w Y Hecto F c P Hecto K w Y Hecto K c Y IR Hydra Hydra M w Y Hecto K w: Y Hydra Hecto K w Y Hecto K c Y IR 196

10 TABLE 3 Continued ID1 Instrument Spectral Type H EW (8) Li EW Age (8) TTS Type A V ( Myr) Member Comments Hecto K c Y IR Hydra M c Y Hecto G c Y IR Hydra Hecto K w Y Hydra M w Y Yg Hydra K w: P Hecto M c Y Yg, IR Hydra Hecto M w Y Hecto G c Y IR Hecto M w Y Hydra Hecto M w Y Hydra Hydra M c P Hydra M c P IR Hydra K c Y IR Hydra M w P Hecto K w: Y IR Hecto M c Y Hecto K c Y Yg, L, IR Hydra G c Y Yg, L, IR Hecto K w: Y Hecto K w Y Hecto K c Y Hecto K w Y Hydra Hydra K c Y L, IR Hecto M c Y L, IR FAST Hecto K c Y L, IR Hecto M w Y Hecto M c P Hecto K w Y IR Hydra M c Y Hecto M w Y IR Hecto K w Y Hydra M c P IR Hecto K w: Y Hecto K w Y Hecto M c P Hecto K c Y L, IR Hecto Hydra Hecto M w Y Yg Hydra K c Y IR Hecto K w Y Hecto M c P L, IR Hecto M w Y Hecto Hecto K w Y Hecto K w PN Hecto K c Y L, IR Hecto M w Y Hydra M c Y Hecto K w Y Hecto K c Y Hecto K w: Y Hecto M c Y L, IR Hecto Hydra Hecto M c Y Hecto M w Y Hecto K w Y IR 197