THE ASTROPHYSICAL JOURNAL, 490:L59L63, 1997 November 20
© 1997. The American Astronomical Society. All rights reserved. Printed in U.S.A.

---

Up:

Is the Large Magellanic Cloud Microlensing Due to an Intervening Dwarf Galaxy?

C. ALCOCK, ^{1, 2} R. A. ALLSMAN, ³ D. R. ALVES, ^{1, 4} T. S. AXELROD, ^{1, 3} A. C. BECKER, ^{2, 5} D. P. BENNETT, ^{1, 2, 6} K. H. COOK, ^{1, 2}

K. C. FREEMAN, ³ K. GRIEST, ^{2, 7} M. J. LEHNER, ⁷ S. L. MARSHALL, ¹ D. MINNITI, ¹ B. A. PETERSON,

³ M. R. PRATT, ^{5, 8} P. J. QUINN, ^{3, 9} A. W. RODGERS, ³ A. RORABECK, ¹⁰ C. W. STUBBS,

^{2, 5, 8} W. SUTHERLAND, ¹¹ A. B. TOMANEY, ⁵ T. VANDEHEI, ⁷ AND D. L. WELCH ¹⁰

(The MACHO Collaboration)

Received 1997 July 24; accepted 1997 September 17; published 1997 October 23

ABSTRACT

     The recent suggestion that the microlensing events observed toward the Large Magellanic Cloud are due to an intervening Sgr-like dwarf galaxy is examined. A search for foreground RR Lyrae in the MACHO photometry database yields 20 stars whose distance distribution follow the expected halo density profile. Cepheid and red giant branch clump stars in the MACHO database are consistent with membership in the LMC. There is also no evidence in the literature for a distinct kinematic population, for intervening gas, or for the turnoff of such a hypothetical galaxy. We conclude that if the lenses are in a foreground galaxy, it must be a particularly dark galaxy.

Subject heading: galaxies: individual (Large Magellanic Cloud, Sagittarius)Galaxy: halostars: individual (RR Lyrae)stars: variables: other

CONTENTS

• 1.  INTRODUCTION
• 2.  SEARCH FOR A DWARF GALAXY IN THE MACHO DATABASE
  • 2.1.  The RR Lyrae Stars
  • 2.2.  Other Possible Evidence
• 3.  SUMMARY
• ACKNOWLEDGMENTS
• REFERENCES
• FIGURES
• TABLES
• REFERENCES TO THIS ARTICLE

FOOTNOTES

     ¹ Lawrence Livermore National Laboratory, Livermore, CA 94550; alcock@llnl.gov, alves@llnl.gov, kcook@llnl.gov, stuart@llnl.gov, dminniti@llnl.gov.

     ² Center for Particle Astrophysics, University of California, Berkeley, Berkeley, CA 94720.

     ³ Mount Stromlo and Siding Spring Observatories, Australian National University, Weston, ACT 2611, Australia; robyn@merlin.anu.edu.au, tsa@merlin.anu.edu.au, kcf@merlin.anu.edu.au, peterson@merlin.anu.edu.au, alex@merlin.anu.edu.au, pjq@merlin.anu.edu.au.

     ⁴ Department of Physics, University of California, Davis, Davis, CA 95616.

     ⁵ Department of Astronomy, University of Washington, Seattle, WA 98195; becker@astro.washington.edu, stubbs@astro.washington.edu, austin@astro.washington.edu.

     ⁶ Department of Physics, University of Notre Dame, Notre Dame, IN 46556; dbennett@phys.nd.edu.

     ⁷ Department of Physics, University of California, San Diego, La Jolla, CA 92093; griest@astrophys.ucsd.edu, matt@astrophys.ucsd.edu, vandehei@astrophys.ucsd.edu.

     ⁸ Department of Physics, University of California, Santa Barbara, Santa Barbara, CA 93106; mrp@lensing.physics.ucsb.edu.

     ⁹ European Southern Observatory, Karl Schwarzschild Strasse 2, D-85748 Garching bei München, Germany; pquinn@eso.org.

     ¹⁰ Department of Physics and Astronomy, McMaster University, Hamilton, ON L8S 4M1, Canada; rorabeck@physics.mcmaster.ca, welch@physics.mcmaster.ca.

     ¹¹ Department of Physics, University of Oxford, Oxford OX1 3RH, England, UK; wjs@oxds02.astro.ox.ac.uk.

§1.  INTRODUCTION

     The MACHO and EROS projects have detected microlensing toward the Large Magellanic Cloud (LMC) (Alcock et al. 1996, 1997; Aubourg et al. 1995) and interpreted these events as due to lenses located in the Milky Way halo. There have been, however, other interpretations of the LMC microlensing events.

     Sahu (1994) argued that the lenses may be located in the disk of the LMC itself. This is unlikely because the spatial distribution of the events is not clustered toward the LMC bar and because this model predicts only one event in the 2 yr data of Alcock et al. (1997), while six to eight bonafide events are observed. We note, however, that the binary event LMC 9 appears likely to be located in the LMC, if the LMC velocity dispersion (and microlensing optical depth) is small (Bennett et al. 1996).

     Gould, Bahcall, & Flynn (1996) argue that one of the MACHO events (LMC 5) could be due to an M-star lens located in the Milky Way disk. This is possible, as Alcock et al. (1997) estimate that about one event out of the eight detected should be due to a thin- or thick-disk lens.

     Zhao (1997) has suggested that the LMC microlensing could be due to an intervening dwarf galaxy like the Sgr dwarf (Ibata, Gilmore, & Irwin 1995). The suggestion is supported by an apparent peak in the RR Lyrae distribution of Payne-Gaposchkin (1971) at 1625 kpc. However, Connolly (1985) suggests that these faint RR Lyrae may be associated with the LMC and thus may not be RR Lyrae at all. These stars would have periods similar to RR Lyrae but absolute magnitudes that are about 2 mag brighter, if they belong to the LMC. This issue can be readily explored with the MACHO photometric database and other recent observations reported in the literature. In this Letter we examine in detail the RR Lyrae distance distribution, concluding that there is no evidence of the existence of such a galaxy. Zhao (1997) also proposed that a significant fraction of the microlensing events may be caused by lenses within an extended tidal tail in front of the LMC. This very interesting suggestion may be addressed in the future when variable stars and microlensing toward new lines of sight, such as the Small Magellanic Cloud, are explored.

§2.  SEARCH FOR A DWARF GALAXY IN THE MACHO DATABASE

     We will look for evidence of an intervening galaxy between us and the LMC in the MACHO photometric database. Dwarf galaxies of the Local Group have complex star formation histories, containing at present old and intermediate-age stellar populations (see reviews by Hodge 1994; Mateo 1996). Thus, we will search for these populations using known tracers: RR Lyrae, Cepheid, and clump giant stars. These tracers are good distance indicators, and they would be brighter if they belong to an intervening galaxy located in front of the LMC. Zhao (1997) uses the specific example of a galaxy like the Sgr dwarf discovered by Ibata et al. (1995). This hypothetical Sgr-like dwarf galaxy would be located between the outskirts of the Galactic disk and the LMC, with distance modulus constrained to the range 11.5<(m-M)<18.5 mag. We also assume that the line-of-sight depth of this galaxy is small, R<3 kpc (i.e., comparable to the observed depth of the Sgr dwarf; Alcock et al. 1997), a priori discarding a "finger of God" effect.

§2.1.  The RR Lyrae Stars

     The RR Lyrae stars are easily detected tracers of an intervening galaxy, because they are easily identified from their light curves, and they are excellent distance indicators. RR Lyrae belonging to a dwarf galaxy that is not extended along the line of sight will show as a peak in the luminosity function, while RR Lyrae belonging to the Milky Way halo would be more or less uniformly distributed between the Sun and the LMC. For example, the Sgr galaxy shows up clearly as an extra peak in the RR Lyrae luminosity function of bulge fields (Alard 1996; Mateo et al. 1996; Alcock et al. 1997).

     Upon inspection of the data presented by Payne-Gaposchkin (1971), Zhao (1997) suggests that there are a number of RR Lyrae clustered at 1625 kpc, arguing that they trace an overdense region, with significant consequences for the interpretation of microlensing toward the LMC. The spectroscopic follow-up of eight of these stars by Connolly (1985) and Smith (1985) confirmed that three of them are not RR Lyrae and that five of them are galactic halo RR Lyrae stars, not members of the LMC. The radial velocities also prove that these five stars are not related to each other: =147 km s (for comparison, =11 km s; Ibata et al. 1995).

     We have studied the line-of-sight distribution of the RR Lyrae stars in front of the LMC using the year 1 MACHO photometry database, which contains 5×10 variable stars in the LMC bar region. The period-amplitude diagram for 10 short-period LMC variable stars is presented in Figure 1, showing the cuts applied to select candidate RR Lyrae types ab and c.

Fig. 1

     Figure 2 shows the color-magnitude diagram of RR Lyrae star candidates from the cuts applied in the period-amplitude plane of Figure 1. Magnitudes are approximate Kron-Cousins V and R, adopting a global calibration solution and neglecting small field-to-field zero point differences. The left-hand panel shows the vast majority of LMC RRab (N=6700) centered at about V=19.1, V-R=0.25. In order to select foreground RR Lyrae stars, we will consider only stars with V<18. There is a large group of stars (N200) centered at V=18.3, V-R=0.35. These are LMC RRab blended with clump giant stars, according to their brighter magnitudes, redder colors, and smaller amplitudes. There is an extended group of blue stars (N600) with V-R<0.1, which include some RRab blended with main-sequence stars and many eclipsing binaries. Finally, there is a group of stars centered at about V=17.0, V-R=0.25, which are mostly short-period type I Cepheids in the LMC (Alcock et al. 1997). They are a major contaminating source when selecting foreground RRab stars, because they have periods, amplitudes, and colors similar to those stars. Foreground RR Lyrae and type I Cepheids were distinguished with the shape of their light curves. We also checked the location of these stars in the period-magnitude diagram and compared them with the MACHO type I Cepheid sample. The short-period Cepheids clearly lie along the first-harmonic (overtone) sequence. Some of these Cepheids turned out to be newly discovered first-harmonic/second-harmonic beat Cepheids, identified via excess photometric scatter at maximum brightness and confirmed by discrete Fourier transforms (Alcock et al. 1995). The right-hand panel of Figure 2 shows the RR Lyrae type c stars (N=1800), with about N200 eclipsing binaries and RRc stars blended with clump giants and main-sequence stars.

Fig. 2

     We find 16 foreground RRab stars with amplitudes A>0.25, periods 0.4<P<0.7 days, colors 0.2<V-R<0.35, and magnitudes V<18 and four foreground RRc stars with A>0.25, 0.2<P<0.4 days, 0.1<V-R<0.25, and V<18. These cuts are optimized to discriminate foreground RR Lyrae from LMC RR Lyrae, blends, and other variable stars. The stars were additionally identified in the 4 yr MACHO database. Calibrated light curves confirm their identification as RR Lyrae and do not change the distance distribution. The coordinates, magnitudes, colors, and periods of the present sample are listed in Table 1.

     We measure distances to the foreground RR Lyrae stars adopting M_V = +0.4 (Reid 1997; Gratton et al. 1997). The uncertainty in these individual distances is estimated to be about 10%, with the caveat that the zero point in RR Lyrae absolute magnitudes may be more uncertain (e.g., M>0.53; Gould 1995), which does not have any effect in the results of this Letter. The distance distribution of the 20 foreground RR Lyrae stars in the direction of the LMC is shown in Figure 3. The lines illustrate a r density halo with spherical flattened (b/a = 0.6) distributions. The flattened halo is normalized arbitrarily to give the same total number of stars out to 40 kpc. Note that there is no significant concentration at any distance. Furthermore, the distribution of these RR Lyrae stars in the sky is uniform across the LMC, not concentrated or clumped.

Fig. 3

     The observation of RR Lyrae stars in the Sgr dwarf by Alcock et al. (1997) allows us to give a direct comparison with what would be expected if there were a dwarf galaxy in the foreground. We place this set of RRab stars (N=34) in the direction of the LMC, given that the total areas surveyed toward the LMC and bulge are similar (neglecting incompleteness corrections, which would be worse in the bulge fields, and neglecting RRc stars in Sgr, which would increase this sample to N50, but which are incomplete). The distance distribution of the Sgr RRab, equivalent to about 0.5% of all the MACHO RRab found in the LMC, is overplotted in Figure 3 (we have actually used the reddening independent magnitudes to compute their individual distances because of the heavy differential reddening of the bulge fields). This magnitude distribution is very different from that of the observed foreground RR Lyrae stars, which argues against the existence of a Sgr-like galaxy in front of the LMC. More stringent limits are possible once radial velocities can be obtained for the stars in our sample.

§2.2.  Other Possible Evidence

     RR Lyrae trace old and metal-poor stellar populations. Classical Cepheid variable stars arise from a different parent population, with intermediate masses and ages. There is no evidence for a population of foreground objects among the 1500 Cepheids found in the inner 22 fields of the LMC (Alcock et al. 1995). Note that every Cepheid would be accompanied by about 100 times their number in main-sequence B-type stars, which are not present.

     RR Lyrae and Cepheid stars represent rare stages of stellar evolution. It may then be possible that only a few of them belonging to an intervening dwarf galaxy would be present in the fields explored here. Core He-burning clump giants should be much more numerous, since stellar evolution theory predicts that many low-mass stars go through this phase of evolution.

     We have assembled MACHO project photometry, properly calibrated to V and R, for 9 million stars in the LMC bar region into a composite color-magnitude diagram. The large number of stars in this diagram reveal many low-level features tracing short-lived stages of stellar evolution. The clump giants peak near V-R=0.45 mag and R=18.6 mag. There is an extra "bump," seen near V-R=0.55 mag and R=17.7 mag, superposed on the red giant branch. While some of the stars in this region of the color-magnitude diagram are likely clump-clump blends, the naturally occurring stellar populations of the LMC are the most plausible explanation for this feature. However, Zaritsky & Lin (1997) report evidence for a distinct population of clump giants in front of the LMC. A discussion of this evidence, and a detailed analysis of the 9 million star color-magnitude diagram will be presented elsewhere (Alves 1998).

     It is very difficult to detect overlapping nearby galaxies, as proven by the recent discovery of the Sgr dwarf behind the Milky Way bulge. Effective ways of searching for such systems would be distance indicators (RR Lyrae stars, Cepheid stars, and clump giants, discussed above), radial velocity surveys, and star counts. Extensive star counts and surface brightness work by de Vaucouleurs (1955) and more recently by Bothun & Thompson (1988) shows no evidence for anything that is not LMC-centric.

     If there were such a galaxy, we should be able to see a distinctive old main-sequence turnoff in deep Hubble Space Telescope (HST) photometry. This is because main-sequence turnoff stars are much more numerous than the tracers discussed above. A clear example of this is the detection of the main-sequence turnoff of the Sgr galaxy behind the bulge by Fahlman et al. (1996) using WFPC2 photometry. Olszewski, Suntzeff, & Mateo (1996) review the recent color-magnitude diagrams of the LMC. We note that none of the recent deep HST color-magnitude diagrams of the LMC (e.g., Elson, Forbes, & Gilmore 1994; Holtzman et al. 1997) show such an extra stellar component.

     Different authors have published radial velocities of giant stars in LMC fields (see Olszewski et al. 1991). No evidence is found for grouping in velocity space of stars that may belong to a dwarf galaxy in front of the LMC.

     Also, the Sgr dwarf has four globular clusters sharing the same kinematics and spatial distribution (DaCosta & Armandroff 1995). However, the old globular clusters located in the direction of the LMC have kinematics and distances consistent with membership to this galaxy (Suntzeff et al. 1992), and there is no evidence for clusters belonging to an unknown Sgr-like dwarf galaxy.

     The presence of a gas-rich dwarf galaxy (dIrr) is also ruled out because we do not see large amounts of gas in front of the LMC. Extensive searches have revealed only H I detection belonging to gas in the Magellanic Stream (Rohlfs et al. 1984; Heller & Rohlfs 1994). Even though there is evidence for interstellar lines at many velocities in high-resolution spectra of SN 1987A (Savage et al. 1989), absorption-line studies of LMC stars do not reveal any significant gaseous component other than that of the Milky Way or the LMC disk (Westerlund 1990; Roth & Blades 1997). To our knowledge, there is no other evidence reported in the literature regarding a galaxy in front of the LMC.

§3.  SUMMARY

     We have tested the idea of Zhao (1997) that the microlensing events observed in LMC fields may be due to an intervening dwarf galaxy like the Sgr dwarf. Searching for foreground RR Lyrae in the year 1 MACHO photometry database yielded 20 stars whose distribution follows the expected halo density law. These foreground RR Lyrae represent 0.3% of the total number of LMC RR Lyrae stars with similar colors detected in these fields. Classical Cepheid variable stars and clump giants seen in the MACHO database are consistent with membership in the LMC. We do not find any evidence for the stellar component of such a dwarf galaxy in the MACHO database, from which we conclude that if the lenses are indeed in a foreground galaxy, this must be a particularly dark galaxy. Alternatively, some of the microlensing events may be due to an extended tidal tail (Zhao 1997). This is more difficult to test with the present data but could be addressed when other lines of sight are studied. The RR Lyrae variable stars in front of the SMC from the MACHO database would give a complementary test to this idea.

ACKNOWLEDGMENTS

     We are also very grateful for the skilled support given to our project by the technical staff at the Mount Stromlo Observatory. Work performed at LLNL is supported by the DOE under contract W7405-ENG-48. Work performed by the Center for Particle Astrophysics on the University of California campuses is supported in part by the Office of Science and Technology Centers of NSF under cooperative agreement AST-8809616. Work performed at MSSSO is supported by the Bilateral Science and Technology Program of the Australian Department of Industry, Technology and Regional Development. W. J. S. is supported by a PPARC Advanced Fellowship. K. G. and M. J. L. acknowledge support from DOE, Alfred P. Sloan, and Cottrell awards. C. W. S. thanks the Sloan, Packard, and Seaver Foundations for their support. D. M. thanks the Aspen Center for Physics for their kind hospitality, as well as useful comments and suggestions from A. Gould, H.-S. Zhao, and D. Zaritsky.

REFERENCES

• Alard, C. 1996, ApJ, 458, L17 First citation in article | Full Text | NASA ADS
• Alcock, C., et al. 1995, AJ, 109, 1653 First citation in article | NASA ADS
• . 1996, ApJ, 461, 84 First citation in article | NASA ADS
• . 1997, ApJ, 474, 217 First citation in article | Full Text | NASA ADS
• Alves, D. R. 1998, Ph.D. dissertation, Univ. California, Davis, in preparation First citation in article
• Aubourg, E., et al. 1995, A&A, 301, 1 First citation in article | NASA ADS
• Bennett, D., et al. 1996, Nucl. Phys. Proc. Suppl., 51B, 152 First citation in article
• Bothun, G. D., & Thompson, I. B. 1988, AJ, 96, 877 First citation in article | NASA ADS
• Connolly, L. P. 1985, ApJ, 299, 728 First citation in article | NASA ADS
• DaCosta, G. S., & Armandroff, T. E. 1995, AJ, 109, 1533 First citation in article
• de Vaucouleurs, G. 1955, AJ, 60, 126 First citation in article | NASA ADS
• Elson, R. W., Forbes, D. A., & Gilmore, G. F. 1994, PASP, 106, 632 First citation in article | NASA ADS
• Fahlman, G. G., Mandushev, G., Richer, H. B., Thompson, H. B., & Sivaramakrishnan, A. 1996, ApJ, 459, L65 First citation in article | Full Text | NASA ADS
• Gould, A. 1995, ApJ, 452, 189 First citation in article | NASA ADS
• Gould, A., Bahcall, J. N., & Flynn, C. 1996, ApJ, 465, 759 First citation in article | NASA ADS
• Gratton, R., Fusi Pecci, F., Carretta, E., Clementini, G., Corsi, C. E., & Lattanzi, M. 1997, ApJ, in press First citation in article
• Heller, P., & Rohlfs, K. 1994, A&A, 291, 743 First citation in article | NASA ADS
• Hodge, P. 1994, in The Local Group: Comparative and Global Properties, ed. A. Layden, R. C. Smith, & J. Storm (Garching: ESO), 57 First citation in article
• Holtzman, J., et al. 1997, AJ, 113, 656 First citation in article | NASA ADS
• Ibata, R., Gilmore, G., & Irwin, M. 1995, MNRAS, 277, 781 First citation in article | NASA ADS
• Mateo, M. 1996, in ASP Conf. Ser. 92, Formation of the Galactic HaloInside and Out, ed. H. Morison & A. Sarajedini (San Francisco: ASP), 434 First citation in article
• Mateo, M., Mirabal, N., Udalski, A., Szymanski, M., Kaluzni, J., Kubiak, M., Krzeminski, W., & Stanek, K. Z. 1996, ApJ, 458, L13 First citation in article | Full Text | NASA ADS
• Olszewski, E. W., Schommer, R. A., Suntzeff, N. B., & Harris, H. C. 1991, AJ, 101, 515 First citation in article | NASA ADS
• Olszewski, E. W., Suntzeff, N. B., & Mateo, M. 1996, ARA&A, 34, 511 First citation in article | NASA ADS
• Payne-Gaposchkin, C. 1971, Smithsonian Contrib. Astrophys., 13 First citation in article
• Reid, I. N. 1997, AJ, in press First citation in article
• Rohlfs, K., Kreitschann, J., Siegman, B. C., & Feitzinger, J. V. 1984, A&A, 137, 343 First citation in article | NASA ADS
• Roth, K. C., & Blades, J. C. 1997, ApJ, 474, L95 First citation in article | Full Text | NASA ADS
• Sahu, K. C. 1994, Nature, 370, 275 First citation in article | NASA ADS
• Savage, B. D., Jenkins, E. B., Joseph, C. L., & de Boer, K. S. 1989, ApJ, 345, 393 First citation in article | NASA ADS
• Smith, H. A. 1985, PASP, 97, 1053 First citation in article | NASA ADS
• Suntzeff, N. B., Schommer, R. A., Olszewski, E. W., & Walker, A. R. 1992, AJ, 104, 1743 First citation in article | NASA ADS
• Westerlund, B. E. 1990, A&A Rev., 2, 29 First citation in article | NASA ADS
• Zaritsky, D., & Lin, D. N. C. 1997, AJ, in press First citation in article
• Zhao, H.-S. 1997, MNRAS, submitted First citation in article

FIGURES

Full image (116kb)

Full image (81kb)

Full image (53kb)

TABLES

TABLE 1
FOREGROUND RR LYRAE STARS TOWARD THE LMC

MACHO ID	P (day)	V	R	Type	R.A.2000	Decl.2000	Comments
1.4297.33...	0.68819	15.05	14.86	ab	05 06 50.396	-68 49 38.06
15.10307.80...	0.51864	17.64	17.29	ab	05 44 18.167	-71 26 15.89
19.4302.382...	0.55915	16.20	15.99	ab	05 06 48.759	-68 26 40.18
6.5722.3...	0.55342	13.73	13.55	ab	05 15 50.055	-70 36 46.56	Sat. at max, also 13.5722.2833
11.9107.197...	0.60315	18.13	17.79	ab	05 36 13.510	-70 48 44.40
5.5125.3195...	0.65617	18.23	17.93	ab	05 11 51.262	-70 01 43.98
5.5495.70...	0.61461	16.61	16.36	ab	05 14 33.476	-69 37 06.10
81.9001.2087...	0.55881	17.17	16.97	ab	05 35 55.304	-69 48 27.85
82.8410.55...	0.51445	16.27	16.08	ab	05 32 18.963	-68 51 52.15	Blazkho
82.7928.95...	0.47585	17.06	16.88	ab	05 29 08.602	-68 42 34.57
10.3187.111...	0.54498	17.78	17.44	ab	05 00 09.002	-70 09 59.20
14.8374.25...	0.62844	15.70	15.48	ab	05 32 11.964	-71 13 50.90
18.2598.10...	0.63913	15.21	14.95	ab	04 56 29.268	-69 08 25.32	Blazkho?
18.3208.40...	0.66432	15.83	15.64	ab	05 00 21.175	-68 46 21.00
19.3825.18...	0.58675	15.62	15.42	ab	05 04 09.837	-68 00 21.30
80.7071.3069...	0.55312	13.88	13.67	ab	05 24 06.126	-69 25 11.93	Also 77.7071.12
3.6718.38...	0.33555	15.89	15.72	c	05 21 53.228	-68 42 37.02	Also 80.6718.4092
14.8371.68...	0.32683	16.93	16.79	c	05 32 24.384	-71 29 03.81
18.2485.17...	0.30745	14.88	14.75	c	04 55 29.721	-68 35 36.87
5.4644.8...	0.32799	15.14	15.04	c	05 09 18.712	-69 50 15.31	Period changes

     NOTE. Units of right ascension are hours, minutes, and seconds, and units of declination are degrees, arcminutes, and arcseconds.