SHOCKED AND SCORCHED: THE TAIL OF A TADPOLE IN AN INTERSTELLAR POND

The presence of strong radiation fields from massive O and B stars can be detrimental to nearby star formation, when compared with star formation in the absence of nearby high-mass stars. Much has been published about this phenomenon; the ionizing radiation from massive stars can evaporate circumstellar disks around nearby low-mass stars (partially ionized globules, i.e., PIGS; see, e.g., Laques & Vidal 1979; Garay 1987, ⁴) or can progressively ionize dense blobs of nearby molecular material, forming evaporating gaseous globules or EGGs, defined explicitly by Hester et al. (1996) as "globules of dense gas that are being photoevaporated more slowly than their lower density surroundings, and so are left behind as the gas around them is driven off." These terms, which describe the ionization/evaporation of gas in disks around low-mass stars, or surrounding molecular material, are rather ill-defined otherwise. The term "proplyd" (photoevaporating protoplanetary disk) is probably the best defined and is generally accepted as referring to evaporating circumstellar disks around young, low-mass stars. An extensive literature can be found about proplyds, especially those in the Orion Nebula Cluster (e.g., O'Dell et al. 1993; Ricci et al. 2008) and their interpretation (e.g., Störzer & Hollenbach 1999; Johnstone et al. 1998). Sizes and mass-loss rates of proplyds in the Orion Nebula Cluster range from 40–350 AU and (0.7–1.5) × 10⁻⁶ M_☉ yr⁻¹ (Henney & O'Dell 1999). The EGGs, first discovered in the Eagle Nebula, appear to be connected to ongoing star formation; near-infrared, high angular resolution observations show that ∼15% of the EGGs in the "elephant trunks" of the Eagle show evidence for associated young low-mass stars, and that there is also evidence for relatively massive young stellar objects (YSOs) at the tips of the elephant trunks (McCaughrean & Andersen 2002). With the advent of Spitzer and the Wide-field Infrared Survey Explorer, new imaging surveys have revealed the wide-spread presence of elephant-trunk structures (also referred to as "pillars") and EGGs toward many massive star-forming regions (e.g., Smith et al. 2010; Koenig et al. 2012).

We have serendipitously discovered an extended, limb-brightened, tadpole-shaped nebula associated with the IRAS source IRAS 20324+4057 (I 20324) near the Cygnus OB2 No. 8 cluster. This paper presents Hubble Space Telescope (HST) imaging, Palomar 5 m spectroscopy, Spitzer IRAC and MIPS imaging, Very Large Array (VLA) radio continuum imaging, and (sub) millimeter-wave spectroscopy using the Arizona Radio Observatory (ARO) SMT and 12 m telescope in order to elucidate the nature of this object and its relationship to proplyds, PIGs, or EGGs.

I 20324 was observed in an HST imaging survey of a list of candidate long-lived preplanetary nebulae (PPNs) selected from the IRAS Point Source Catalog using a color criterion (F₆₀/ F₂₅ > 1) chosen to select objects that are dominated by relatively cool dust located at some distance from a stellar source or where any hot dust is highly obscured (Sahai et al. 2007). All objects were required to have point-source Two Micron All Sky Survey (2MASS) (JHK) and MSX (A band/8 μm) counterparts, implying the presence of local stellar sources that have heated the dust revealed in the HST and IRAS data. However, we found that a large fraction of the objects resolved in our survey reveal morphologies quite different than those found for PPNs (e.g., Sahai et al. 2007). I 20324 is one of the most prominent of these. It was first detected as an emission-line nebula together with two other nebulous objects within a 31 × 31 field in a ground-based Hα + [N ii] survey of candidate post-asymptotic giant branch objects (Pereira & Miranda 2007, PM07). PM07 concluded that all three objects are "related YSOs" and that the emission-line spectrum of I 20324 is indicative of photoionization rather than shock excitation.

Independently, I 20324 has been noted recently by Wright et al. (2012, Wetal12), who included it among a group of 10 "proplyd-like objects" associated with the Cygnus OB2 association. They considered both the proplyd and EGG hypotheses, and concluded that neither scenario adequately explains their observations. They suggest that these and the other objects they identified are "a unique class of photoevaporating partially embedded YSOs." In contrast, we find that our data favor the EGG hypothesis.

2.1. HST Imaging

We obtained optical images of I 20324 on 2006 July 22 using HST's Wide-Field Camera (WFC) of the Advanced Camera for Surveys (ACS), which has a plate scale of 0050 pixel⁻¹, using the F606W (λ = 0.60 μm, Δλ = 0.123 μm) and F814W (λ = 0.80 μm, Δλ = 0.149 μm) filters. A two-point dither was used for the imaging with a total exposure time of 2 × 347 s per filter. These images show an extended nebula with a limb-brightened "tadpole" shape (hereafter, we often use the label "Tadpole" to refer to I 20324; Figure 1(a)) oriented roughly east–west (E–W); the periphery of the nebula is distinctly brighter and more defined on the N side and shows prominent ripples. A bright compact knot is located interior to the tadpole head (Figure 2). The intensity of the knot, which itself has a cometary shape, peaks at J2000 α = 20:34:13.23, δ = 41:08:14.6 (in the F814W image), coinciding with the location of luminous point sources in the 2MASS (20341326+4108140: α = 20:34:13.26, δ = 41:08:14.07) and MSX6c catalogs (G080.1909+00.5353: α = 20:34:13.4, δ = 41:08:14). The axis of the cometary knot is aligned at a P.A. of −43° (north through east). From the F606W image, we find that the Tadpole has a length of 547 (76,600 AU) and a maximum width of 141 (19750 AU). The cometary knot, seen more sensitively in the F814W image, is about 24 × 13 (3350 AU×1800 AU). The knot has a conical shape near the apex with an "inner" opening angle of 55°, but then narrows toward a more cylindrical shape at offsets greater than about 075, implying an "outer" opening angle of ∼0.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The cometary knot has a diametrically opposed faint counterpart seen only in the F814W filter—the simplest interpretation of these features together is that they represent scattered light from the lobes of a collimated bipolar outflow directed along the polar axis of a flared disk (or dense equatorial region) tilted such that the near-side of the disk lies to the nortwest (NW) of the bright apex of the knot. Extinction by the tilted disk then explains the faintness of the NW lobe of the bipolar outflow. The NW outflow extends toward three faint red star-like objects (a, b, and c in Figure 2): high-resolution (015) H- and K_s-band images obtained with the Laser Guide Star Adaptive Optics system on the Palomar 200-in using the High Angular Resolution Observer NIR camera show point sources at the locations of a, b, and c (R. Sahai et al. 2012, in preparation), supporting the idea that these are stars, and not simply compact disks illuminated by starlight from the Tadpole's central star. The stars a and b show small "tail" structures emanating from them (Figure 2, inset); the inner part of star a's tail is roughly aligned with the bipolar outflow axis, whereas in the outer parts, the tail starts curving backward.

I 20324 falls within the region imaged by the Spitzer Cygnus X Legacy Survey program (PI: J. Hora) with the IRAC (four bands at 3.6, 4.5, 5.8, and 8.0 μm) and MIPS instruments (two bands at 24 and 70 μm)—since some of these data have been recently reported by Wetal12, we do not discuss them in detail. Inspection of the Legacy Cygnus-X Spitzer images reveals a bright point source in all four IRAC bands and the MIPS 24 μm band, together with a tail structure. The coordinates of the point source in the Cygnus-X catalog are R.A. = 20:34:13.23, decl. = 41:08:14.1, virtually identical with the location of the cometary knot apex in the HST images. The 8 μm image is notable in showing the ripples on the south periphery of the Tadpole (Figure 1(b)), which are also seen in the HST image but less clearly. Although the angular resolution is rather limited relative to the object's size at 70 μm, the MIPS image shows a cometary shape with a bright head which is likely dominated by flux from the central source in the Tadpole. We assume that I 20324 is associated with the Cygnus OB2 association and adopt a distance, D = 1.4 kpc (Rygl et al. 2012).⁵

2.2. Additional Multi-wavelength Observations

Following the HST observations, we obtained supporting ground-based observations with (1) the Palomar 5 m, (2) the VLA of the NRAO,⁶ and (3) the 10 m (SMT, Mt. Graham, AZ) and 12 m (Kitt Peak, AZ) (sub)millimeter-wave telescopes of the ARO.

The TripleSpec Spectrograph at Palomar was used to observe I 20324 with its 1'' × 30'' slit aligned N–S, so as to cut laterally across the Tadpole body, at two locations (1) passing through the cometary knot (Slit 1) and (2) 12'' east of the knot (Slit 2). The seeing was about 1'' at K band. Strong emission from H₂ S(1), v = 1–0 and other H₂ lines was detected in both slits; for Slit 2, the emission is clearly limb-brightened (Figure 1(b)). The Slit 1 spectra also show a strong continuum at the position of the cometary knot, and extended emission in the H i Brγ and Paβ lines—the latter lines display a local peak centered on the knot.

The VLA was used in the C configuration to observe the field near the Tadpole nebula in 2009 August, as part of program AS980. Observations were made at frequencies of 8.5 and 22.5 GHz (λ = 3.6 and 1.3 cm, respectively). Standard data editing, calibration, imaging, and deconvolution were performed using the Astronomical Imaging Processing System for both frequencies. For the 8.5 GHz image, observed on August 5 (Figure 3), the restoring beam was 32 × 28 at P.A. = −65°. The rms noise in the final image was ∼65 μJy beam⁻¹. For the 22.5 GHz image (observed on August 10), a 70 kλ taper was applied in the imaging step to provide a similar angular resolution to the 8.5 GHz image. The VLA 8.5 GHz image shows a tadpole-shaped structure at the location of I 20324 and two additional sources, which were also detected in the PM07 emission-line survey (objects B and C in their Figure 1; object A is I 20324). IRAC 8 μm images (insets in Figure 3) of objects B and C show similarly elongated morphologies: given its resemblance to a goldfish, we dub object B the "Goldfish." We also searched for the water maser emission line at 22.235 GHz toward the Tadpole, but did not detect any emission over a total V_lsr range of ±41.85 km s⁻¹ to a 1σ sensitivity of 25 mJy beam⁻¹ per 48.83 kHz wide channel.

Zoom In Zoom Out Reset image size

The total flux densities at 8.5 GHz for the Tadpole, Goldfish, and object C are 55 ± 1.2, 10 ± 0.4, and 6 ± 0.7 mJy. At 22.5 GHz, only the Tadpole is detected with a flux density of 30 ± 1.4 mJy. The significantly lower flux density at 22.5 GHz compared to that at 8.5 GHz for the Tadpole implies the presence of non-thermal emission in this source. The radio emission peaks strongly along the shock/ionization front at the head of the Tadpole, possibly as a result of a compressed magnetic layer in this front that is interacting with cosmic rays (CRs) associated with the Cyg OB2 association—we note that the Fermi Large Area Telescope has recently revealed a 50 pc wide cocoon of freshly accelerated CRs that fill the cavities carved in the Cygnus star-forming region by stellar winds and ionization fronts (Ackermann et al. 2011). Further observations at different radio wavelengths are needed to confirm and explore the nature of non-thermal emission from the Tadpole.

We estimate a total mass of ionized gas in the Tadpole of ≲ 10⁻² M_☉, assuming optically thin free–free emission at 22.5 GHz, and approximating the emitting region with a triangular slab (of area 375 arcsec²), which fits the Tadpole's projected shape in the 8.5 GHz image. Our mass estimate is an upper limit since there may be a non-thermal contribution at 22.5 GHz.

The ARO data were taken during 2009 January–May. The ¹²CO J = 1–0, ¹³CO J = 2–1 lines, and the HCO⁺ J = 3–2 line were observed with the 1 mm dual-polarization receiver employing ALMA Band 6 SBS mixers on the SMT 10 m, and filterbanks with 1 MHz resolution. Typical system temperatures were 240 K (770 K) for the CO (HCO⁺) observations. The beam size was θ_b = 32'' at 1.3 mm, and pointing accuracy was estimated to be about ±4''. The ¹²CO and CS J = 2–1 lines were observed with the 12 m telescope at Kitt Peak using the dual-polarization superconductorâ insulatorâ superconductor receivers. Typical system temperatures were 430 K (240 K) for the CO (CS) observations, and data were recorded with the Millimeter Autocorrelator configured to provide 48.8 kHz resolution. The beam size was θ_b = 60'' at 2.6 mm, and pointing accuracy was estimated to be about ±6''. Observations were conducted in position-switching mode using an off position 15' south of I 20324 which was tested to be free of significant emission.

The spectra of the high-density tracer lines, HCO⁺ J = 3–2 and CS J = 2–1 (Figure 4), show strong emission peaking at V_lsr = 10.4 km s⁻¹, indicating that the peak density in the emitting region is quite high,⁷ ∼10⁶ cm⁻³. The map (5' × 5') of the CO and ¹³CO J = 2–1 emission (with both lines being observed simultaneously) using the "on-the-fly" mapping technique shows a compact source peaking at the same V_lsr (Figure 5) centered on the Tadpole (object A). A second nearby peak coincides with the Goldfish. The CO emission from I 20324 is extended along the E–W direction with an FWHM of about 45'' and it is unresolved in the N–S direction; source B is unresolved. Sparse mapping of the HCO⁺ J = 3–2 emission shows that it is also extended E–W. There is also extended, structured emission in the velocity range V_lsr = −5 to 8 km s⁻¹ that is not connected spatially or kinematically with the Tadpole or Goldfish, but likely comes from the same complex of progenitor molecular clouds that spawned them.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

We derive an excitation temperature, T_ex = 11 K from the peak intensity (T_R = 6.2 K) of the CO J = 2–1 line, assuming it to be optically thick. Assuming LTE conditions, we find (using the RATRAN online code; Van der Tak et al. 2007) that the peak ¹³CO J = 2–1 line intensity (T_R = 2.5 K) and observed line width (2 km s⁻¹) imply a total column density of N(H₂) = 0.4 × 10²² cm⁻², assuming a standard interstellar ¹³CO/H₂ abundance ratio of 10⁻⁶ (e.g., Hayashi et al. 1993). This gives a total molecular mass of 3.7 M_☉ for I 20324. Assuming a similar kinetic temperature and abundance for Source B, we find its mass is about 1.3 M_☉ (its peak ¹³CO J = 2–1 line intensity is about half that of I 20324). Our mass estimates are likely lower limits because the "standard" value adopted for the ¹³CO/H₂ abundance ratio may be too large, since a significant fraction of CO (and ¹³CO) may have been photodissociated in a photodissociation region that occupies much of the volume of the tadpoles.

The Tadpole is detected at all wavelengths from the near- to the far-infrared, with point-source counterparts in the 2MASS, MSX, IRAS, and Akari catalogs: the spectral energy distribution (SED) is shown in Figure 6. We have also included the IRAS Low Resolution Spectrum (LRS) for the Tadpole (with a correction of the absolute calibration applied as described by Cohen et al. 1992). Since the LRS data were rather noisy, we increased the signal-to-noise ratio (S/N) by binning the spectrum to a factor of three lower resolution: a weak 10 μm silicate dust feature can be seen in absorption. The IRAC photometry was extracted from the Cygnus-X catalog, whereas the MIPS photometry was derived from the Legacy Cygnus-X images using aperture photometry. The core of the 24 μm image of I 20324 is partly saturated, so we used the bright Airy ring at a radius of 265 from the center (fourth ring) for measuring the flux; our estimate of 37 Jy is consistent with the IRAS flux value. Comparing the Spitzer photometry with that from missions having lower angular resolution (red symbols) at similar wavelengths (24 and 70 μm), we find that the former (hereafter, the "small aperture" fluxes) falls systematically below the latter at λ ∼ 60–70 μm. This indicates the presence of an extended source of warm dust at these wavelengths heated by the external radiation field, in addition to the compact central source, which looks pointlike out to a wavelength of at least 24 μm. The large mid- and far-IR fluxes of this point source imply the presence of a substantial mass of circumstellar dust around the Tadpole's central star. The total source luminosity of the Tadpole derived from integrating the "small aperture" SED is about 500 L_☉, implying that the Tadpole's central star is not a low-mass star. From an inspection of pre-main-sequence stellar evolutionary tracks by Bernasconi & Maeder (1996), we find that our estimated luminosity implies a stellar mass of around 5 M_☉ if the star has just begun deuterium burning or somewhat lower if it is approaching the main sequence.

Zoom In Zoom Out Reset image size

We fitted I 20324's "small aperture" SED using the online tool provided by Robitaille et al. (2007), which computes least-squares fits of pre-computed models of YSOs having disks and rotationally flattened infall envelopes (with biconical outflow cavities), to user-defined SEDs. We set the input distance range to D = 1.3–1.6 kpc. The input interstellar extinction range was set to A_v = 0–2, based on our estimate of A_v = 0.7 ± 0.7 using the numerical algorithm provided by Hakkila et al. (1997), which computes the three-dimensional visual interstellar extinction and its error from inputs of Galactic longitude, latitude, and distance, from a synthesis of several published studies. The best-fit model has D = 1.4 kpc, A_v = 1.5, and a central star mass of M_* = 4.6 M_☉ and an effective temperature of T_eff = 4140 K; the model spectrum shows the presence of a weak silicate absorption feature as observed (Figure 6). In this model, the full opening angle of the bipolar outflow cavity is θ_c = 7°, and the outflow cavity axis is inclined at a modest angle of i = 18°, to the line of sight. The next best model has D = 1.4  kpc, A_v = 2, M_* = 6.5 M_☉, T_eff = 4200 K, θ_c = 5°, and i = 18°, however the depth of the silicate feature relative to the continuum is twice that observed. In both models, almost all of the emitting mass is associated with the envelope. These models are qualitatively consistent with the optical appearance of the luminous central source with scattered light from the inner regions of the near-side bipolar cavity producing the cometary knot seen in the HST image, and its far-side counterpart being obscured by the dense inner equatorial region of the envelope. The value of the projected opening angle in the best-fit model, θ_proj = tan ⁻¹(tan θ_c/sin i) = 23°, lies in the range covered by the inner and outer opening angles of the knot (0° − −55°). We note that the minimum inclination angle used in the pre-computed model library is i = 18°, so it is possible that a better fit might have been obtained with a smaller inclination angle which would result in a larger value of θ_proj. However, a better 10 μm spectrum covering the silicate feature (i.e., with higher S/N and a small aperture) is needed before the derived parameters from the above modeling can be put on a firm footing.

We note that there are several tadpole-shaped objects in the vicinity of I 20324 that share with the latter a common E–W head–tail orientation. These are (1) the Goldfish (outside the field of view in our HST images; No. 8 in Wetal12); (2) "object C" of PM07 seen faintly in the 8.5 GHz, HST, and Spitzer images, which has a relatively small tail (No. 9 in Wetal12); (3) and an object at R.A. = 20:34:45.9, decl. = 41:14:46.5 in the IRAC images of this region (No. 10 in Wetal12). The common E–W orientation strongly suggests these structures are shaped by the ram pressure of a passing wind from a distant source or sources located to the West of these objects, with material being ablated from their heads and blown eastward. Noting that the tails in both the Tadpole and Goldfish are bright on their northern sides (both in the optical and in the radio), we conclude that this is due to asymmetric irradiation of these objects by the ultraviolet radiation field of the Cyg OB2 No. 8 cluster that lies to the NW of the Tadpole and Goldfish, at a separation of about 900'' (∼6 pc).

The Tadpole and its associates are most likely small, dense molecular cores near the Cyg OB2 complex that are being subjected to photoevaporation by the strong UV radiation field due to the large number of O stars in this massive star formation region. The presence of additional young stars (objects a, b, and c; R. Sahai et al. 2012, in preparation) in close physical proximity to I 20324's cometary knot indicates that several young stars have formed within the dense molecular core inside the head of the Tadpole. The small tails associated with a and b may represent (1) winds from these stars interacting with the NW outflow of or (2) possibly photoevaporation of their circumstellar disks by radiation from the Tadpole's central star. The curvature of a's tail away from the Tadpole's head may be due to interaction with the compressed "wall" of gas in the latter (note that the wall curves around both the back and the front, so a could be located quite close to the rear Tadpole wall).

The physical processes responsible for producing the radio-bright peripheries of the Tadpole and its associates are likely the same ones that operate in the vicinity of the M16 (Eagle Nebula) elephant trunks (lateral width ∼(0.5–1) × 10¹⁸ cm) where dense, dusty molecular clouds are being eroded by photoevaporation due to radiation from O stars at a distance of 6 × 10¹⁸ cm from the trunks, producing a photoevaporative flow. The heads of these trunks appear in emission-line images as limb-brightened arcs at the interface between the molecular cloud and the ambient H ii region. The bright heads (size ∼3 × 10¹⁷ cm) and bodies of the Tadpole and its associates likely represent similar interfaces.

We interpret the ripples evident in the Tadpole body as resulting from Kelvin–Helmholtz (K–H) instabilities. Such instabilities are expected to occur at the periphery of a dense cloud embedded in diffuse gas when there is relative motion between the two components (e.g., Fleck 1984; Kamaya 1998). The spatial wavelength of the ripples (∼15'' = 3.2 × 10¹⁷ cm) is very similar to the wavelengths of the "ripples" found on the surfaces of some molecular clouds in the Orion Nebula by Berné et al. (2010) that have also been interpreted as resulting from K–H instabilities. We consider (but discard) the possibility that the ripples could be caused by the orbital motion of two bound stars separated by a distance comparable to the ripple amplitude (≳2'' = 2800 AU), since the ripple wavelength implies a timescale of 1050 yr (100 km s⁻¹/V_w) (V_w is an assumed speed for the exterior wind impacting the Tadpole), which is much shorter than the expected orbital time for such a binary system, 10⁵ yr (M₁ + M₂)^−1/2 (taking the masses of the stars in the binary to be M₁ ∼ M₂ ∼ 1 M_☉).

Although the bow-shock morphology of the head of the tadpole structure and the cometary shape of the central knot in I 20324 are very similar to the shapes of structures seen toward proplyds in the Orion star-forming region (e.g., O'Dell & Wong 1996; Bally et al. 1998), we do not think I 20234 or object B are proplyds because the associated molecular masses (greater than 3.7 and 1.3 M_☉) are much too large compared to what one might expect for circumstellar material being evaporated from protoplanetary disks: disk masses of (0.003–0.07) M_☉ have been inferred from a submillimeter array continuum survey for the Orion proplyds (Mann & Williams 2010). Indeed, no dense molecular medium has yet been found around the central disks in any known proplyd. Furthermore, the minimum separation between the bright periphery of the Tadpole head (which would represent the wind shock in the proplyd hypothesis) and the cometary knot in I 20324 is 47 = 6600 AU, which is much larger than the corresponding typical separations for the Orion proplyds (∼500–1000 AU; Bally et al. 1998). However, this argument needs to be strengthened by a quantitative assessment of the competing effects of (1) the larger number of ionizing stars in Cyg OB2 and (2) their larger distance from the location of I 20324 and its associates compared to Orion, since the physical scales of structures associated with proplyds scale inversely with the ionizing EUV flux.

We conclude that I 20324 and its associates are EGGs. EGGs are likely to be the surviving dense portions of their parent molecular clouds, and as such, they should be predominantly molecular, as observed in our objects. While the original concept of EGGs shows them as having elongated tails that connect them continuously with the elephant trunks in M16, this is due to the ionizing sources in M16 being concentrated in one direction relative to the EGGs. However, the presence of multiple ionizing sources distributed over a large solid angle relative to the EGGs, as in Cyg OB2, results in a diffuse radiation field that eventually pinches off their tails—a possibility noted by Hester et al. (1996) and supported by numerical simulations (e.g., Ercolano & Gritschneder 2011)—creating "free-floating" EGGs, such as the Tadpole and its associates.

Their relatively high internal density has not only prolonged their survival in the hostile radiative and windy environment of the Cyg OB2 cluster, but it makes these globules a likely place for stars to have begun forming, especially as the overpressure of the surrounding H ii region would have aided gravity by compressing the globules. It therefore seems to be no accident that the Tadpole and the Goldfish have embedded stars. Wetal12 point out that a higher fraction of the globules in this region (70%) appear to have stars than those in the Eagle Nebula (15%), and they use this fact to argue that their objects in Cygnus are therefore unlikely to be EGGs. However, several variables can affect this fraction. The parent cloud of the Eagle Nebula may have been considerably less dense and massive than the parent cloud of the globules in Cygnus, and the Cygnus OB2 association could provide a much stronger radiation field and more powerful winds than the cluster powering the Eagle Nebula. Both of these factors would affect both the size distribution of the resulting EGGs and the compression to which they are subjected.

In summary, it is quite plausible that not only the Tadpole and the Goldfish, but all of the "proplyd-like" objects listed in Wetal12 are EGGs, considering their rather similar physical sizes and shapes. Molecular-line observations of these objects, like the ones reported in this paper, will help in determining their masses and thus confirming their true nature.

We thank Anna Rosen for her valuable help in reducing the ARO on-the-fly data presented in this paper, as part of her Spring 2009 NASA/USRP student internship at JPL. We thank the staff of the Arizona Radio Observatory for granting us observing time. The National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc. R.S.'s contribution to the research described here was carried out at the Jet Propulsion Laboratory, California Institute of Technology, under a contract with NASA. Financial support was provided by NASA through a Long Term Space Astrophysics award to R.S. and M.M., and HST GO award 10536 to R.S.