A NEAR-INFRARED SPECTROSCOPIC STUDY OF YOUNG FIELD ULTRACOOL DWARFS

We first compared our spectra of young objects with field dwarf near-IR spectral standards from Kirkpatrick et al. (2010). Not surprisingly, the entire 0.8–2.5 μm spectra of young objects are not well matched by older field dwarfs, in part due to the redder near-IR colors of young objects (Kirkpatrick et al. 2008). In some spectral regions, however, the continuum shape is sensitive to spectral type with little dependence on gravity. In particular, we used the 1.07–1.40 μm and 1.90–2.20 μm wavelength regions to determine J-band and K-band spectral types by visual comparison to field dwarf standards (Figure 3). For each spectrum in our sample, we over-plotted the spectra of field dwarf standards (normalized over the comparison wavelength region) and qualitatively determined which standard best matched the continuum shape of our object. If two standard spectra provided similar matches, we assigned the spectral type intermediate to the two standards, e.g., if the L0 and L1 standards provided equally good fits, we assigned a spectral type of L0.5. For the majority of our objects, selecting a standard with 1 subtype difference compared to the best-fitting standard provided a noticeably poorer fit. Thus, we assign an uncertainty of 1 subtype to our visual classifications. For objects which had either particularly noisy or peculiar spectra, several field dwarf standards provided equally good fits, thus uncertainties of ±2 subtypes were assigned to their visual classification.

Zoom In Zoom Out Reset image size

To test for any gravity dependence of our visual spectral typing, we compared the near-IR spectral types of ∼10 Myr old objects in our sample (TWA members and objects with optical gravity classifications of γ) to their optical spectral types. On average, the J-band spectral types of ∼10 Myr old objects are 1.3 subtypes later than their corresponding optical spectral types. The K-band spectral types for low-gravity objects are, on average, 0.1 subtypes earlier than their corresponding optical spectral types. Determining near-IR spectral types for young objects based on visual comparison to field dwarfs will lead to near-IR types that are slightly later than their optical spectral types. In addition to visually comparing our spectra to the IR standards, we also computed reduced χ² for the 1.07–1.40 μm and 1.90–2.20 μm wavelength regions to assign spectral types based on the best matching standard. The spectral types determined from the minimum χ² value agreed to within the uncertainties with the spectral types determined by visual classification (±1 subtype). Minimum reduced χ² values for fitting spectra of field standards to ∼10 Myr old objects were typically ∼30 for J-band fits and ∼7 for K-band fits. The large values for reduced χ² are a result of the poor match between field dwarf standards and young field objects. Figures 4 and 5 show the differences between our J- and K-band visual spectral types and published optical spectral types.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Another method of determining spectral type is using spectral indices. This method has the potential advantage of measuring spectral features which are well correlated with the overall spectral morphology (type) of the object, while avoiding wavelength regions containing broad, gravity-sensitive features. We calculated many published spectral type-sensitive indices (Tokunaga & Kobayashi 1999; Cushing et al. 2000; Testi et al. 2001; Geballe et al. 2002; McLean et al. 2003; Slesnick et al. 2004; Allers et al. 2007; Weights et al. 2009; Covey et al. 2010; Scholz et al. 2012) for the objects in our sample. Most of these spectral-type sensitive indices were developed to correlate with optical spectral types. The majority of spectral type-sensitive indices were either found to be gravity-sensitive (i.e., young dwarfs and field dwarfs having very discrepant index–SpT relations) or were only sensitive over a narrow range in spectral type. Overall, we find that the H₂O (Allers et al. 2007), H₂O-1, H₂O-2 (Slesnick et al. 2004), and H₂OD⁴ (McLean et al. 2003) indices are spectral type sensitive and gravity-insensitive over a broad range in spectral type. For these four indices, we fit index versus optical spectral type for field dwarfs with a third-degree polynomial and use the scatter about the fit as the uncertainty in the index–SpT relation (Figure 6). The polynomial fits and scatter in the index–SpT relations, along with the range of spectral type sensitivity, are presented for each index in Table 3. Table 1 presents the spectral types calculated from these indices for our sample of young objects. Uncertainties in the index-derived spectral types were determined using a Monte Carlo approach to account for uncertainties in the index calculations from each spectrum. These were then added in quadrature to the rms SpT scatter in the index–SpT relations (Table 3).

Zoom In Zoom Out Reset image size

Table 3. Spectral Type Sensitive Indices

Index Definition	Index Ref.	SpT	Coefficients of Polynomial Fitsa				rms
			c0	c1	c2	c3	SpT
${\rm H_2O} = \frac{\langle F_{\lambda =1.550\hbox{--}1.560} \rangle }{\langle F_{\lambda =1.492\hbox{--}1.502} \rangle }$	Allers et al. (2007)	M5–L4	−83.5437	169.388	−104.424	24.0476	0.390
${\rm H_2OD} = \frac{\langle F_{\lambda =1.951\hbox{--}1.977} \rangle }{\langle F_{\lambda =2.062\hbox{--}2.088} \rangle }$	McLean et al. (2003)	L0–L8	79.4477	−202.245	229.884	−97.230	0.757
H2O-1$= \frac{\langle F_{\lambda =1.335\hbox{--}1.345} \rangle }{\langle F_{\lambda =1.295\hbox{--}1.305} \rangle }$	Slesnick et al. (2004)	M4–L5	12.1927	39.3513	−80.7404	28.5982	1.097
H2O-2$= \frac{\langle F_{\lambda =2.035\hbox{--}2.045} \rangle }{\langle F_{\lambda =2.145\hbox{--}2.155} \rangle }$	Slesnick et al. (2004)	M4–L2	10.8822	55.4580	−97.8144	37.5013	0.501

Note. ^aThe spectral type is calculated from the polynomial fits as: SpT = c₀ + c₁(index) + c₂(index)² + c₃(index)³.

Download table as:  ASCIITypeset image

To arrive at final near-IR spectral types, we take the weighted mean of all of the spectral types determined using indices and visual comparison. For many of our objects, this results in the visual spectral types having little effect on the final spectral type determination. We round the final spectral type to the nearest integer subtype. Uncertainties in the weighted mean spectral type were 0.3–0.9 subtypes, thus we adopt a conservative uncertainty of 1 subtype for our near-IR spectral types. The near-IR spectral types of our sample are, on average, 0.07 subtypes earlier than their published optical spectral types. For the ∼10 Myr old objects in our sample, their near-IR spectral types are, on average, identical to their published optical spectral types. Figure 7 shows the differences between our near-IR spectral types and published optical spectral types. The vast majority (55/64) of near-IR and optical spectral types agree to within 1 subtype. The nine objects having discrepant (by more than 1 subtype) near-IR and optical spectral types do not show a preference for particular optical spectral types, gravities, or near-IR colors. Overall, our method for determining IR spectral types yields results that are consistent with optical spectral types. An additional benefit of our method of spectral typing is that we are not biased by an object's J − K color, which has a large dispersion among the L dwarfs (e.g., Knapp et al. 2004), inhibiting the use of the entire spectrum for visual classification. Figure 8 shows the spectra of objects classified as L3 that have J − K colors ranging from 1.6 to 3.1 mag.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The near-IR spectra of late-M and L dwarfs contain a wealth of absorption features attributed to FeH bands (McLean et al. 2003; Cushing et al. 2005), which also provide significant atmospheric opacity (Rice et al. 2010a). The most prominent FeH features seen at low resolution are bandheads at 0.99, 1.20, and 1.55 μm. The depth of the FeH absorption features increases steadily through the late-M to early-L spectral types, with spectral types of L2–L3 having the strongest FeH features. For spectral types later than L3, the strengths of the FeH features decrease. By spectral types of L7, very little FeH is discernable. Low-gravity M and L dwarfs display much weaker FeH bands than field dwarfs of the same spectral type.

We establish the FeH_z index (Table 4 and Figure 16) which measures the depth of the 0.99 μm FeH feature. We optimized the index to be sensitive to gravity for both moderate- and low-resolution spectra by making the wavelength windows for the index as wide as a single resolution element for our lowest resolution (R ∼ 75) spectra. The index is calculated as follows:

$$\begin{equation} {\rm index} = \left(\frac{\lambda _{{\rm line}}-\lambda _{{\rm cont1}}}{\lambda _{{\rm cont2}}-\lambda _{{\rm cont1}}} F_{{\rm cont2}} + \frac{\lambda _{{\rm cont2}}-\lambda _{{\rm line}}}{\lambda _{{\rm cont2}}-\lambda _{{\rm cont1}}}F_{{\rm cont1}} \right) / F_{{\rm line}}. \end{equation} \tag{ 1 }$$

The central wavelengths and widths of the line and continuum regions are listed in Table 4. The numerator of the equation gives the expected flux at the line wavelength (based on a linear interpolation of flux in the continuum windows) if no absorption or emission were present. F_cont1 is the average of the spectrum (in F_λ units) over a window as wide as the bandwidth listed in Table 4 and centered on λ_cont1. F_cont2 and F_line are calculated similarly. Indices calculated using Equation (1) will have values of one for spectra showing no FeH absorption, with higher index values indicating deeper absorption features. Uncertainties in F_λ per pixel were estimated from the rms scatter about a linear fit to wavelength versus F_λ in the continuum windows. Uncertainties in the index value were calculated assuming that the line region has the same flux uncertainty per pixel as the continuum. We established the expected field dwarf index (black line in Figure 20) and its uncertainty (gray shaded region in Figure 20) from the mean and standard deviation of the index values of all field dwarfs in ± 1 subtype bins (e.g., the field dwarf index value and uncertainty for L0 are the average and standard deviation of indices for the M9–L1 field dwarfs). Table 5 and Figure 20 present the FeH_z indices calculated for our sample. This index is sensitive to gravity for M6–L7 spectral types.

Zoom In Zoom Out Reset image size

Many of our moderate-resolution spectra do not extend to short enough wavelength to calculate the FeH_z index. To measure the 1.20 μm FeH feature, we created the FeH_J index. At low spectral resolution, the 1.20 μm FeH feature is blended with Fe i, Mg i, and K i, thus the FeH_J index is only appropriate for moderate-resolution spectra. Table 6 and Figure 21 display the FeH_J indices calculated from our moderate-resolution spectra. The 1.55 μm FeH feature contributes to the measured values of the H-cont index and is discussed in Section 4.2.4.

Zoom In Zoom Out Reset image size

Condensation effects and higher metal hydride opacities contribute to the weaker strength of vanadium oxide (VO) bands in field dwarfs compared to low-gravity (young) ultracool dwarfs. We established a gravity-sensitive VO_z index (Table 4 and Figure 17), similar to the z–VO index presented in Cushing et al. (2005), but optimized for low-resolution spectra. A larger value of the VO_z index corresponds to deeper 1.06 μm VO absorption. Figure 20 shows the index calculated for our sample, using Equation (1). We established the expected field dwarf index and its uncertainty in the same manner as for the FeH_z index. We find that the 1.06 μm VO feature is an excellent gravity indicator for L0–L4 dwarfs, with the index values for optically classified Lγ dwarfs lying well above the field dwarf sequence. Young, late-M dwarfs also show enhanced VO absorption, but the difference between young M dwarfs and field dwarfs is more subtle. We do not use the other notable VO band (at ∼1.17 μm) in our analysis as this feature is blended with H₂O, FeH, and K i features.

Na i and K i alkali lines are the most prominent features in the J-band spectra of late-M and L field dwarfs. Pressure broadening and the condensation of opacity sources contribute to the large EWs (∼10 Å; Cushing et al. 2005) measured for these features at spectral types of L0–L6. The K i and Na i lines are blended with FeH, Fe i, and H₂O features in low-resolution spectra of late-M to mid-L dwarfs, which limits the reliability of these features as age indicators at low spectral resolution. Allers et al. (2007) established a gravity-sensitive index from the 1.14 μm Na i line, which is appropriate for R ≳ 300 spectra. We tested this index on the moderate-resolution spectra in our sample and found that while the index does an excellent job of distinguishing low-gravity M dwarfs from field M dwarfs, it is not as effective at determining the gravity of L dwarfs.

We have established a K i_J index to measure the depth of the 1.244 and 1.253 μm K i feature at low resolution. The index is calculated using Equation (1). The index is contaminated by an FeH feature at 1.239 μm. Fortunately, the trend of FeH strength with gravity is similar to the dependence of the K i line depth with age. Figure 22 shows the K i_J index, which is sensitive to gravity for spectral types of M5–L6 for low-resolution spectra.

Zoom In Zoom Out Reset image size

From moderate-resolution spectra, we can calculate the pseudo-EWs of the K i and Na i lines. Table 7 and Figure 18 show the line and continuum wavelengths for our calculation of EWs. We use line windows of 0.006 μm and continuum windows of 0.002 μm (both of which are much larger than a single resolution element for R = 750). We approximate the continuum in the line region from a linear fit to the flux in the continuum windows, and use the rms scatter about the fit in the continuum windows as the flux uncertainty per pixel. We then propagate uncertainties using a Monte Carlo method to compute the uncertainty in the EW. We calculated EWs for the Na i lines centered at 1.1396 μm, and the K i lines centered at 1.1692, 1.1778, 1.2437, and 1.2529 μm, which are presented in Table 8 and Figures 21 and 23. We do not use the 1.2437 μm K i EW in our youth analysis, as it is blended with an FeH feature, which results in large uncertainties (see Figure 23). The J-band Na i and K i EWs show similar dependence on spectral type and gravity, with TWA members having substantially weaker absorption than field dwarfs.

Zoom In Zoom Out Reset image size

A hallmark of youth seen in the near-IR spectra of late-M and L dwarfs at low spectral resolution is a triangular H-band continuum shape (Figure 10). Though both very low and intermediate-gravity objects have a triangular shape compared to field dwarfs, intermediate-gravity objects appear to have a "shoulder" at ∼1.57 μm, likely due to a combination of increased FeH absorption and H₂ collision induced absorption (Borysow et al. 1997). As noted by Bowler et al. (2012), the H-band continuum shape for intermediate-age (∼50–150 Myr) M9 dwarfs is less triangular than that of a ∼12 Myr old TWA M9 dwarf. We note that the K-band continuum shape also appears to be sensitive to gravity (Figures 10 and 11), with low-gravity objects having slightly more positive 2.15–2.25 μm spectral slopes than field dwarfs. The differences in K-band spectral shape with gravity, however, are subtle, and dusty objects have K-band continuum shapes that are similar to young objects (Figure 15), so we do not use K-band continuum shape in our analysis.

We establish the H-cont index to measure the triangular shape of the H-band spectrum (Table 4 and Figure 19). Though calculated using Equation (1), the H-cont index does not measure the depth of an absorption feature, but rather measures how much the shape of the blue end of the H-band deviates from a straight line. For very low gravity objects the blue side of the H band is nearly a straight line, corresponding to an H-cont index value of ∼1.0. Higher gravity objects have lower H-cont indices. The H-cont index is sensitive to gravity for spectral types of M6–L7. Spectra for objects earlier than M6 tend to have a fairly flat H-band continuum shape, so the index loses its efficacy for early- to mid-M dwarfs. The H-cont indices for our sample are displayed in Figure 24. Though objects classified as low gravity in the optical (γ and β) appear to have low gravity using the H-cont index, older "dusty" field dwarfs also show a hint of low gravity in this index (e.g., Figure 15). Thus the continuum shape of the H band is not the most reliable gravity indicator, particularly for intermediate gravities, and should be used in combination with other gravity-sensitive features.

Zoom In Zoom Out Reset image size

A number of objects in our sample have been tied to young kinematic groups, which allows us to examine our gravity classifications for objects of known age. Based on their calculated index values and gravity scores, the TWA objects in our sample are all classified as vl-g, including the L3 planetary-mass companion TWA 27B (a.k.a. 2M1207b). This is in good agreement with the young age of TWA (∼8–12 Myr; Torres et al. 2008; Mentuch et al. 2008). One object in our sample, 2M 0608−27 (Rice et al. 2010b), has been recently linked to the ∼10 Myr old β Pictoris (Torres et al. 2008) moving group. 2M 0608−27 is characterized by our indices as L0 vl-g, which agrees with its possible membership in β Pictoris. Two objects in our sample, GSC 08047B (Chauvin et al. 2005a) and AB PicB (Chauvin et al. 2005b), are companions to members of the ∼30 Myr old Tuc–Hor Association. Both GSC 08047B (L1) and the AB PicB (L0) are classified as vl-g, as is 2M 0141−46 (L0 vl-g), a candidate Tuc–Hor member (Kirkpatrick et al. 2010). Complicating the possible age for our vl-g classifications, 2M 0355+11 (L3 vl-g) has a measured space motion that is consistent with membership in the ∼100 Myr old AB Doradus moving group (Liu et al. 2013). Interestingly, CD−35 2722B, also an AB Dor member (Wahhaj et al. 2011), has only very subtle signatures of low gravity and is classified as an L3 int-g. Figure 27 compares the spectra for these two objects. Overall, it appears that an infrared classification of vl-g corresponds to an age of ∼10–30 Myr, in agreement with the age estimate for optically classified γ objects (log(age(yr)) ≈ 7; Kirkpatrick et al. 2010). It is important to note, however, that older objects (e.g., 2M 0355+11) can also display very low gravity spectral signatures.

Zoom In Zoom Out Reset image size

2M 0314+16 and 2M 1705−05 are candidate members of the ∼500 Myr old (King et al. 2003) Ursa Majoris moving group (Seifahrt et al. 2010), and neither shows signatures of low gravity. As mentioned in Section 4.3, LP 944−20 is classified as L0 fld-g but is on the borderline for int-g classification. LP 944−20 is thought to be a member of the ∼200 Myr old Castor moving group (Ribas 2003). Thus, though an absolute upper limit to the ages of our int-g sources is difficult to determine from young moving group members, we appear to be sensitive to ages ≲200 Myr.

A number of our sources have published lithium (Li i) detections from optical spectroscopy. The detection of lithium in the spectrum of a low-mass star or brown dwarf can imply two things: (1) the object has too little mass (⩽65 M_Jup) to ever have burned lithium or (2) the object is young and has not yet depleted its lithium. For M dwarfs, a Li i detection provides a useful upper limit on the ages, with detections at later spectral types corresponding to an older upper limit on the age (Chabrier et al. 1996). For L dwarfs, however, young objects can have weaker Li i than field dwarfs (Kirkpatrick et al. 2008) due to lower surface gravity (i.e., a similar effect as seen for Na i and K i). For late L dwarfs, lithium is expected to form molecular species and Li i should not be seen in their spectra. Shkolnik et al. (2009) detected strong Li i absorption in the optical spectra of 2M 0557−13 (M7 vl-g) and 2M 0335+23 (M7 vl-g) and determined upper limits to their ages of 10 Myr, consistent with our near-IR classification. The detection of Li i in the optical spectrum of SERC 296A (M6 vl-g) has a limit on its age of ≲200 Myr (Thackrah et al. 1997), which is consistent with its classification. Both 2M 0019+46 and 2M 1411−21 are M8 int-g objects with Li i detections (Reiners & Basri 2009), which sets an age limit for these sources of ≲300 Myr.

Six of our sources are companions to stars. Four of these objects are discussed in Section 4.5.1, as their stellar companions are known members of young kinematic groups. Two of our sources are companions to stars for which ages can be approximated. G 196−3B is classified as L3 vl-g based on its near-IR indices, and has an estimated age of 20–300 Myr (Rebolo et al. 1998; Kirkpatrick et al. 2001). The vl-g classification of G 196–3B and spectral similarity to other vl-g L3 objects (Figure 8) argue that the G 196−3 system likely falls at the low end of its estimated age range. Gl 417B is an L4.5 companion to a G0 star. We classify Gl 417B as L5 fld-g but note that its spectrum shows very subtle signatures of low gravity. Based on a variety of indicators, Kirkpatrick et al. (2001) estimate an age of 80–300 Myr for Gl 417B. Gyrochronology of Gl 417A, however, gives an age estimate of 750 Myr (Allers et al. 2010).

With a J − K color of 2.45 mag, 2M 2244+20 is one of the reddest known L dwarfs. Its optical spectrum does not show obvious signs of peculiarity (Kirkpatrick et al. 2008), but its very peculiar IR spectrum has been attributed to low gravity and/or low metallicity (McLean et al. 2003). Low metallicity seems an unlikely explanation for 2M 2244+20's peculiar spectrum, as low-metallicity objects tend to have much bluer J − K colors (e.g., the sdL7 2MASS J05325346+8246465 with J − K = 0.26 mag; Burgasser et al. 2007). Figure 15 shows the spectrum of 2M 2244+20 (L6 vl-g) compared to 2M 0103+19, an optical L6β we classify in the infrared as an L6 int-g, as well as the dusty L6.5 dwarf 2M 2148+40 (Looper et al. 2008). Despite its J − K color being similar to 2M 2148+40 (J − K = 2.38 mag), the spectrum of 2M 2244+20 more closely resembles 2M 0103+29 (J − K = 2.14 mag). The continuum shape of 2M 2244+20's spectrum is remarkably similar to 2M 0103+29 (despite the difference in J − K color for the two objects), but the K i and FeH features in 2M 2244+20 are weaker, implying that 2M 2244+20 has a lower gravity.

As seen in Figure 1, three of the M dwarfs in our sample (2M 0422+15, 2M 0435−14, and 2M 0619−29) have particularly red near-IR spectra for their spectral types. We have estimated the reddening to each of these sources by comparing their 2MASS J − K colors to the photospheric colors for young objects of their spectral types (Luhman et al. 2010). We find reddenings of A_v = 4.6, 7.4, and 6.5 mag for 2M 0422+15, 2M 0435−14, and 2M 0619−29, respectively. We dereddened their near-IR spectra and found that these levels of reddening do not change the spectral type or gravity determinations for these objects.

2M 0422+15 lies ∼10° south of the Taurus star-forming region in an area that has diffuse extinction of A_v ≃ 0.8–1.4 mag (Lombardi et al. 2010) and CO emission (Dame et al. 2001), thus it seems plausible that it is either behind or embedded in interstellar material associated with the Taurus–Auriga region. 2M 0422+15 displays excess emission in the mid-IR (J. Lyons et al., in preparation) indicative of a circumstellar disk, which is not particularly surprising given our classification of vl-g.

The diffuse extinction measured toward 2M 0435−14 is A_v ∼1.5 mag (Schlegel et al. 1998), and no CO emission is detected in the immediate vicinity. As noted by Cruz et al. (2003), 2M 0435−14 lies in the direction of MBM20, a molecular cloud 112–161 pc away (Hearty et al. 2000). Cruz et al. (2003) calculated a spectrophotometric distance for 2M 0435−14 of 30 pc and concluded that it could not be a member of MBM20. 2M 0435−14 does not display excess emission from a disk (J. Lyons et al., in preparation) and is in front of the MBM20 cloud. Thus, the source of its reddening remains unknown.

The dust map toward 2M 0619−29 indicates very low extinction (A_V ≲ 0.2) in the region (Schlegel et al. 1998). 2M 0619−29 has mid-IR excess emission indicative of a circumstellar disk (J. Lyons et al., in preparation). With a spectral type of M5, we could not assign a gravity classification to 2M 0619−29. Qualitatively, 2M 0619−29 has low-gravity features compared to a field M5 dwarf spectrum. This, combined with the detection of a circumstellar disk for this source, makes it likely to be ≲10 Myr old, and possibly reddened by its disk.

2M 0355+11 is one of the more interesting objects in our sample. It has one of the reddest J − K colors (2.52 mag) of any known L dwarf. It was classified as an L5γ by Cruz et al. (2009), and it is a mere 9.1 ± 0.1 pc away (Liu et al. 2013). Based on its kinematics and sky position, Liu et al. (2013) link 2M 0355+11 to the ∼100 Myr old AB Doradus moving group. In contrast, Faherty et al. (2013) determine that this object is unlikely to be an AB Dor member, but based on a lower precision parallax measurement. Figure 27 shows the spectrum of 2M 0355+11 compared to CD−35 2722B, a young companion to an AB Dor member (Wahhaj et al. 2011). Despite having the same infrared spectral type (L3) and nominally the same (∼100 Myr) age, the spectra of 2M 0355+11 and CD−35 2722B are quite different. The spectrum of CD−35 2722 has very subtle hints of youth, and we classify this object as int-g, consistent with the ∼100 Myr age of AB Dor. 2M 0355+11, on the other hand, has very distinct low-gravity features, and is classified as vl-g. We note that even if 2M 0355+11 were assigned an IR spectral type of L5 (its optical spectral type), we would still classify this object as vl-g. The discrepancies between the spectra of 2M 0355+11 and CD−35 2722B are puzzling. Perhaps 2M 0355+11 and CD−35 2722 are not the same age (i.e., one of them is not a member of AB Dor)? If 2M 0355+11 and CD−35 2722 are indeed coeval, their spectra indicate that objects of the same age may have very different spectral signatures of youth.