arXiv:0909.1921v1 [astro-ph.SR] 10 Sep 2009
Mon. Not. R. Astron. Soc. 000, 1–?? (2008) Printed 10 September 2009(MN LATEX style file v2.2)
NSVS06507557; a low-mass double-lined eclipsing
¨O. C ¸akırlı1‡ and C.˙Ibanoˇ glu1
1Ege University, Science Faculty, Astronomy and Space Sciences Dept., 35100 Bornova,˙Izmir, Turkey
Released 2009 Xxxxx XX
In this paper we present the results of a detailed spectroscopic and photomet-
ric analysis of the V=13m.4 low-mass eclipsing binary NSVS06507557 with
an orbital period of 0.515 d. We obtained a series of mid-resolution spectra
covering nearly entire orbit of the system. In addition we obtained simul-
taneous VRI broadband photometry using a small aperture telescope. From
these spectroscopic and photometric data we have derived the system’s orbital
parameters and determined the fundamental stellar parameters of the two
components. Our results indicate that NSVS06507557 consists of a K9 and
an M3 pre-main-sequence stars with masses of 0.66±0.09 M⊙and 0.28±0.05
M⊙and radii of 0.60±0.03 and 0.44±0.02 R⊙, located at a distance of 111±9
pc. The radius of the less massive secondary component is larger than that of
the zero-age main-sequnce star having the same mass. While the radius of the
primary component is in agreement with ZAMS the secondary component
appers to be larger by about 35 % with respect to its ZAMS counterpart.
Night-to-night intrinsic light variations up to 0m.2 have been observed. In
addition, the Hα, Hβlines and the forbidden line of [Oi] are seen in emission.
The Lii 6708˚ A absorption line is seen in most of the spectra. These features
are taken to be the signs of the classic T Tauri stars’ characteristics. The
parameters we derived are consistent with an age of about 20 Myr according
to the stellar evolutionary models. The spectroscopic and photometric results
are in agreement with those obtained by theoretical predictions.
Key words: stars:activity-stars:fundamentalparameters-stars:lowmass-stars
Low-mass stars constitute the majority of stars by number in our Galaxy. Since their lower
masses with respect to the Sun they have also very low intrinsic brightness. Although the
intrinsic faintness of these stars many low-luminosity stars were discovered particularly by
the near-infrared sky surveys: Deep Near Infrared Survey (Delfosse et al. 1997), Two Micron
All Sky Survey (Skrutskie et al. 1997), Sloan Digital Sky Survey (York et al. 2000), Northern
Sky Variability Survey (Wozniak et al. 2004). Since their main-sequence lifetimes are consid-
erably longer than the age of the universe both the young and old low-mass stars are located
on the lower right part of the the HR diagram. Low-mass stars surround many important
regions of stellar parameter space which include the onset of complete convection in the
stellar interior, the onset of electron degeneracy in the core, and the formation of dust and
depletion metals onto dust grains in the stellar atmosphere (West et al. 2004). Recent studies
have shown that while the observed radii of the low-mass stars are significantly larger than
those predicted by current stellar models, in contrast their effective temperatures are cooler
(Ribas et al. 2008, Lopez-Morales and Ribas 2005). Chabrier, Gallardo & Baraffe (2007)
have put forward the hypothesis that the observed radius and temperature discrepancies are
consequences of the convection due to rotation and/or magnetic field and the presence of
large surface magnetic spots. Therefore low-mass stars are key interest in studies of both
formation of the stars in star-forming regions and comparison their parameters with those
predicted from theoretical stellar models.
The fundamental parameters such as mass, radius, effective temperature and luminos-
ity, all in a distance-independent manner, of a star could be determined empirically from
eclipsing binary stars. Precise masses and radii can be determined from multi-wavelength
photometry and spectroscopy, obtained with current technology, of double-lined close binary
systems. However, the number of well-studied eclipsing binaries with low-mass components is
rather small because of their low intrinsic brightness. Furthermore, most of their light curves
are undergone strong distortion due to magnetic activity. Therefore, multi-passband photo-
metric and spectroscopic observations of additional low-mass binaries would be extremely
The binary nature of the star known as NSVS 06507557 (=2MASS J01582387+2521196,
⋆Based on photometric and spectroscopic observations collected at T¨UB˙ITAK National Observatory (Turkey).
† Table 1 is only avaliable in electronic form at the CDS via ftp to http://www.blackwell-syngery.com/doi....
NSVS06507557; a low-mass double-lined eclipsing binary
hereafter NSVS 0650) was discovered by Shaw and Lopez-Morales (2006) using the database
NSVS (Wozniak et al. 2004). The eclipse period was determined to be 0.515 d. Later on the
first VRI light curves and preliminary models are presented by Coughlin and Shaw (2007).
Taking the BVRI magnitudes from the USNO NOMAD catalog and JHK from 2MASS
catalog they estimated the effective temperature of 3860 K for the primary, corresponding
to a spectral type of M0V. As they have noted a difficulty encountered in modeling was the
high-level spot activity of the components. Not only the radii and effective temperatures of
the component stars were determined but also the rough masses estimated by them. We have
conducted a photometric and spectroscopic monitoring program of several low-mass eclipsing
binaries. In this paper we present, the results of multi-wavelength optical photometry and
spectroscopy for double-lined eclipsing binary NSVS 0650.
NSVS0650757 was first identified in the Northern Sky Variability Survey (NSVS;
Wozniak et al. 2004) as a detached eclipsing binary system with a maximum, out-of-eclipse
V-bandpass magnitude V =13m.05 and a period of P=0.51509 day. The data from the NSVS,
obtained with the Robotic Optical Transient Search Experiment telescopes (ROTSE), con-
tains positions, light curves and V magnitudes for about 14 million objects ranging in mag-
nitudes from 8 to 15.5. The B, V, R, and I magnitudes for NSVS 0650 were listed in the
USNO NOMAD catalog as ( Naval Observatory Merged Astronomical Dataset,
NOMAD-1.0, Zacharias et al. 2004), B=14m.53, V =13m.37, R=12m.47; on the other hand
the infra-red magnitudes in three bandpasses were given as J=10m.918 H=10m.267, and
K=10m.092 in the 2MASS catalog (Cutri et al. 2003).
In the NSVS survey, 262 V-bandpass measurements of the variable were obtained during
the period June 1999 - March 2000 with a median sampling rate of 0.25−1. The resulting
light curve exhibits periodic eclipses with a depth of ∼0m.7 in the deeper eclipse and the
mean standard deviation in the out-of-eclipse phases was about 0m.073.
The photometric observations of NSVS 0650 were carried out with the 0.4 m telescope
at the Ege University Observatory. The 0.4 m telescope equipped with an Apogee 1kx1k
CCD camera and standard Bessel VRI bandpasses. The observations were performed on
seven nights between September 01 and November 30, 2008. To get the higher accuracy
Table 1. Differential photometric measurements of NSVS0650 in the V, R and I bandpasses.
HJD(2400000+) ∆V HJD(2400000+)∆R HJD(2400000+)∆I
the target NSVS0650 was placed near to the center of the CCD and three nearby stars
located on the same frame were taken for comparison. The stars GSC 01760-01860 and
USNO A2.0 1125 638990 were selected as comparison and check, respectively. Therefore
the target and comparison stars could be observed simultaneously with an exposure time
of 10 seconds. Since the variable is very cool, red star the signal-to-noise ratio was highest
in I- and lowest in the V-bandpass. The differential observations of the comparison stars
showed that they are stable during time span of our observations. The data were processed
with standard data reduction procedures including bias and over scan subtraction, flat-
fielding, and aperture photometry. A total of 743, 812 and 612 photometric measurements
were obtained in each V, R and I bandpasses, respectively. The average uncertainity of
each differential measurement was less than 0m.030. The V-, R- and I-bandpass magnitude
differences, in the sense of variable minus comparison, are listed in Table 1 (available in the
electronic form at the CDS).
The light curve shows a deep primary eclipse with an amount of 0m.70 in the V-bandpass
and a shallow secondary eclipse with an amount of 0m.23 which are clearly separated in
phase, as is typical of fully detached binaries. The primary and secondary eclipses occur
almost 0.5 phase interval, indicating nearly circular orbit. An inspection of the nightly light
curves presented in Fig. 1 clearly indicates considerable out-of-eclipse light variations up to
0m.2. This intrinsic variation of the binary system manifests itself in the deeper primary
2.2 Orbital period and ephemeris
The first orbital period for NSVS0650 was determined as P=0.51509957 d by Shaw and
Lopez-Morales (2006) from the NSVS database. Later on Coughlin & Shaw (2007) observed
seven low-mass detached systems, including NSVS0650, with the Southeastern Association
for Research in Astronomy (SARA) 0.9 m telescope. An orbital period of P=0.5150895±0.0000008
days, and an initial epoch T0(HJD)=2453312.3722±0.0005 for the mid-primary eclipse were
calculated using a least square fit. Partial primary and secondary eclipses which were de-
NSVS06507557; a low-mass double-lined eclipsing binary
Figure 1. The V-, R-, and I-bandpass nightly light curves for NSVS0650 from top to bottom. The V-, R-, and I-bandpass light
curves clearly show that the brightness of the variable significantly varies from night to night, particularly in out of eclipse.
tected in the time series photometric data were used in combination with the NSVS pho-
tometry to derive this ephemeris for the system.
We obtained three times of mid-primary and one secondary eclipse during our observing
run. The mid-eclipse timings and their standard deviations are caculated using the method
of Kwee & van Woerden (1956). These timings of the eclipses were listed in Table 1 together
with two primary and a secondary eclipse collected from literature. The times for mid-eclipses
are the average of times obtained in three bandpasses. We define the epoch of the system,
T0, to be the midpoint of the most complete primary eclipse. For this reason we use the V-,
R-, and I-bandpass data obtained on JD=2454 746 which cover almost the whole primary
eclipse. A linear least square fit to the data listed in Table 1 yields the new ephemeris as,
MinI (HJD) = 2454746.3801(5)+ 0.51508836(9)× E,(1)
where E corresponds to the cycle number. The residuals in the last column of Table 2 are
computed with the new ephemeris. While the orbital period is nearly the same with that
determined by Coughlin & Shaw (2007) its uncertainty is now very smaller than estimated
by them. In the computation of the orbital phase for individual observations we used this
Table 2. Times of minima measured from the V RI-bandpass light curves.
∗∗From the NSVS database.
†Coughlin & Shaw (2007).
2.3Intrinsic light variations
The light curve of NSVS 0650 shows two well-separated eclipses, as a typical of detached
eclipsing binaries. The phase difference between the eclipses is about 0.5 which indicates a
nearly circular orbit. Since the depths of the eclipses are very different, indicating that the
components have unequal effective temperatures. The light variation both in primary and
out-of-eclipse is clearly seen in all bandpasses. This light variation of about 0.2 mag peak-
to-peak in the out-of-eclipse portions of the light curve reveals that there is an intrinsic
variation in one or both components of the system. The amplitude of the intrinsic variations
seems to larger with longer wavelengths. The light variations observed on JD 2454746 with
long duration, just between primary and secondary eclipses, and also on JD 2454767 with
very short duration, resemble a flare-like event which is common in M-type dwarf stars.
The data obtained by us are concentrated on seven nights ranging a time span of 56
days. The stars having masses smaller than that of the Sun are known to be heavely spotted.
Therefore the out-of-eclipse light variations may be attributed to large spots on the surface of
one or both component stars. In addition, flares on the less massive star cannot be ignored.
However, it should be noted that the intrinsic light variations do not resemble to those
observed in the spotted stars. A spot or spot groups on one or both components produces
usually wave-like distortion on their light curves. However, the out-of-eclipse light variations
in NSVS 0650 seem to not correlated with the orbital period.
Optical spectroscopic observations of NSVS0650 were obtained with the Turkish Faint Ob-
ject Spectrograph Camera (TFOSC) attached to the 1.5 m telescope on 3 nights (September
15, 16, and 17, 2008) under good seeing conditions. Further details on the telescope and the
spectrograph can be found at http://www.tug.tubitak.gov.tr. The wavelength coverage of
NSVS06507557; a low-mass double-lined eclipsing binary
each spectrum was 4100-8100˚ A in 11 orders, with a resolving power of λ/∆λ 7000 at 6563
˚ A and an average signal-to-noise ratio (S/N) was ∼120. We also obtained a high S/N spec-
trum of the M dwarf GJ740 (M0 V) and GJ623 (M1.5 V) for use as templates in derivation
of the radial velocities (Nidever et al. 2002).
The electronic bias was removed from each image and we used the ’crreject’ option for
cosmic ray removal. Thus, the resulting spectra were largely cleaned from the cosmic rays.
The echelle spectra were extracted and wavelength calibrated by using Fe-Ar lamp source
with help of the IRAF echelle package.
The stability of the instrument was checked by cross correlating the spectra of the stan-
dard star against each other using the fxcor task in IRAF. The standard deviation of the
differences between the velocities measured using fxcor and the velocities in Nidever et al.
(2002) was about 1.1 km s−1.
We have used our spectra to reveal the spectral type of the primary component of NSVS0650.
For this purpose we have degraded the spectral resolution from 7000 to 3000, by convolving
them with a Gaussian kernel of the appropriate width, and we have measured the equivalent
widths (EW) of photospheric absorption lines for the spectral classification. We have fol-
lowed the procedures of Hern´ andez et al. (2004), choosing helium lines in the blue-wavelength
region, where the contribution of the secondary component to the observed spectrum is al-
most negligible. From several spectra we measured EWHeI+FeIλ4922= 1.18 ± 0.12˚ A.
From the calibration relations EW–Spectral-Type of Hern´ andez et al. (2004), we have
derived a spectral type of K8 with an uncertainty of about 1 spectral subclass. The effec-
tive temperature deduced from the calibrations of Drilling & Landolt (2000) or de Jager &
Nieuwenhuijzen (1987) is about 4050K. The spectral-type uncertainty leads to a tempera-
ture error of ∆Teff≈ 300K.
The catalogs USNO, NOMAD and 2MASS provides BVRIJHK magnitudes for NSVS
0650. Using the observed colors of B-V=1.36±0.02 and V-I=2.13±0.02 mag and color-
temperature relationships given by Drilling & Landolt (2000) for the main sequence stars we
estimate a spectral type K9±1 with an effective temperature of 3930±50 K for the primary
star. The observed infrared colors of J-H=0.651±0.043 and H-K=0.175±0.038 given in the
2MASS catalog (Cutri et al. 2003) correspond to a spectral type of K9±2 is in a good agree-
Shift (km s
C o r r e l a t i o n
Figure 2. Sample of Cross Correlation Functions (CCFs) between NSVS0650 and the radial velocity template spectrum around
the first and second quadrature.
ment with that we derived by wide-band B-V and V-I photometric colors. We estimated a
temperature of 3920±175 K from the calibrations of Tokunaga (2000). Temperature uncer-
tainty of the primary component results from considerations of spectral type uncertainties,
and calibration differences. The weighted mean of the effective temperature of the primary
star is 3960±80 K. The effective temperature of the primary star what we derived from the
photometric measurements is an a good agreement with that we estimated from the spectra
3.1 Radial velocity curve
To derive the radial velocities for the components of binary system, the 16 TFOSC spectraof
the eclipsing binary were cross-correlated against the spectrum of GJ740, a single-lined
M0V star, on an order-by-order basis using the fxcor package in IRAF. The majority
NSVS06507557; a low-mass double-lined eclipsing binary
Figure 3. Radial velocity curve folded on a period of 0.51508836 days, where phase zero is defined to be at primary mid-eclipse.
Symbols with error bars show the RV measurements for two components of the system (primary: open circles, secondary: open
of the spectra showed two distinct cross-correlation peaks in the quadrature, one for each
component of the binary. Thus, both peaks were fit independently in the quadrature with a
Gaussian profile to measure the velocity and errors of the individual components. If the two
peaks appear blended, a double Gaussian was applied to the combined profile using de-blend
function in the task. For each of the 16 observations we then determined a weighted-average
radial velocity for each star from all orders without significant contamination by telluric
absorption features. Here we used as weights the inverse of the variance of the radial velocity
measurements in each order, as reported by fxcor. In these data, we find no evidence for
a third component, since the cross-correlation function showed only two distinct peaks.
We adopted a two-Gaussian fit algorithm to resolve cross-correlation peaks near the first
and second quadratures when spectral lines are visible separately. Figure2 shows examples of
cros-correlations obtained by using the largest FWHM at nearly first and second quadratures.
The two peaks, non-blended, correspond to each component of NSVS0650. The stronger
peaks in each CCF correspond to the more luminous component which has a larger weight
into the observed spectrum.
The heliocentric RVs for the primary (Vp) and the secondary (Vs) components are listed
in Table3, along with the dates of observation and the corresponding orbital phases com-
Table 3. Heliocentric radial velocities of NSVS0650. The columns
give the heliocentric Julian date, the orbital phase (according to
the ephemeris in Eq. 1), the radial velocities of the two components
with the corresponding standard deviations.
HJD 2400000+PhaseStar 1
puted with the new ephemeris given in §2.2. The velocities in this table have been corrected
to the heliocentric reference system by adopting a radial velocity of 9.5 km s−1for the tem-
plate star GJ740. The RVs listed in Table3 are the weighted averages of the values obtained
from the cross-correlation of orders #4, #5, #6 and #7 of the target spectra with the cor-
responding order of the standard star spectrum. The weight Wi= 1/σ2
ihas been given to
each measurement. The standard errors of the weighted means have been calculated on the
basis of the errors (σi) in the RV values for each order according to the usual formula (e.g.
Topping 1972). The σivalues are computed by fxcor according to the fitted peak height,
as described by Tonry & Davis (1979).
First we analysed the radial velocities for the initial orbital parameters. We used the
orbital period held fixed and computed the eccentricity of the orbit, systemic velocity and
semi-amplitudes of the RVs. The results of the analysis are as follows: e=0.002±0.001, i.e. for-
mally consistent with a circular orbit, γ=44±6 km s−1, K1=77±3 and K2=181±12 km s−1.
Using these values we estimate the projected orbital semi-major axis and mass ratio as:
asini=2.63±0.12 R⊙ and q =M2
3.2 Light curve modeling
As we noted in Section 2.3 the light curve of the system is considerably distorted due to light
fluctuations both at maxima and in the deeper primary minimum. The largest distortion
with longest duration was observed on JD 2454746. Neither the amplitude nor the period
NSVS06507557; a low-mass double-lined eclipsing binary
Table 4. Results of the V-, R-, and I-bandpass light curve analysis for NSVS0650. The adopted values are the weighted means
of the values determined from the individual light curves.
or cycle of these intrinsic variations are known at this step. Therefore, we take all the
available V-, R- and I-bandpass data for the orbital parameter analysis. The differential
magnitudes of 743 in V-, 812 in R- and 612 in I-bandpass were converted to intensities using
the differential magnitudes at out-of-eclipses as ∆V=1m.648±0m.003, ∆R=1m.198±0m.001,
We used the most recent version of the eclipsing binary light curve modeling algorithm
of Wilson & Devinney (1971) (with updates), as implemented in the phoebe code of Prˇ sa
& Zwitter (2005). The code needs some input parameters, which depend upon the physical
structures of the component stars. In the light curve solution we fixed some parameters
whose values can be estimated from global stellar properties, such as effective temperature
and mass of the star. Therefore we adopted the linear limb-darkening coefficients from Van
Hamme (1993) as 0.39 and 0.28 for the primary and secondary components, respectively; the
bolometric albedos from Lucy (1967) as 0.5, typical for a fully convective stellar envelope, the
gravity brightening coefficients as 0.32 for the both components. The rotation of components
is assumed to be synchronous with the orbital one. The mass-ratio of 0.425 was adopted
from the semi-amplitudes of the radial velocities. We started the light curve analysis with
an effective temperature of 3960 K for the primary star of NSVS 0650. The adjustable
parameters in the light curves fitting were the orbital inclination, the surface potentials, the
effective temperature of secondary, the luminosity of the primary.
Using a trial-and-error method we obtained a set of parameters, which represented the
observed light curves. A detached configuration, Mode 2, with coupling between luminosity
and temperature was chosen for solution. The iterations were carried out automatically until
convergence, and solution was defined as the set of parameters for which the differential
corrections were smaller than the probable errors. The orbital and stellar parameters from
Figure 4. The phase folded VRI light curves for NSVS0650. The best fitting solutions represented by the solid lines are also
plotted for comparison (see text).
the V-, R- and I-bandpass light and radial velocity curves analysis are listed in Table 4. The
uncertainties given in this table are taken directly from the out-put of the program. The
computed light and velocity curves corresponding to the individual light-velocity solutions
are compared with the observations in Figs. 3 and 4.
NSVS0650 has a complex spectrum over the wavelength interval from ∼4100 to 8100˚ A. The
spectrum is dominated by forbidden lines and to a smaller degree, permitted emission lines
of neutral metals. Strong and broad double-peaked Hα, Hβand [Oi] lines are present, with
the peak separation in Hαlarger than the higher Balmer lines. The presence of the strong Li
NSVS06507557; a low-mass double-lined eclipsing binary
Figure 5. The composite spectrum in the spectral region containing the Li I 6708˚ A line observed on JD2454727.4836.
6708˚ A absorption line can serve as a reliable youth indicator of a star, as evidenced in the
case of NSVS 0650. Young, low-mass pre-main-sequence stars are called T Tauri stars (TTS).
They present the following characteristics: 1) Emission line spectra, 2) Presence of forbidden
narrow-lines such as [Oi], [Nii] and [Siii], 3) Photospheric continuum excesses (Barrado y
Navascues and Martin, 2003). TTSs are classified into two sub-groups, the classical T Tauri
stars (cTTSs) and the weak-lined T Tauri stars (wTTSs). A cTTS is surrounded by an
optically thick disk from which it accretes material. Whereas a wTTS represents the final
stage of accretion and disc-clearing processes (Bertout et al. 2007, Schisano et al. 2009).
The equivalent width of Hαemission is used as an empirical criterion to distinguish between
cTTS and wTTS, being smaller in the latters. Due to possible varability, no clean cut can
be defined between the cTTS and wTTS based on the Hαemission alone.
Spectral and photometric properties and night-to-night light variability of NSVS0650
indicate that the active star in the system resembles many characteristics of the TTSs as
given above and discussed by Alcala et al. (1993), Covino et al. (1996), and Alencar & Basri
(2000). As it is known the optical emission lines are definite characteristics of the many late
type, main-sequence systems, including NSVS0650. Another fundamental characteristic of
TTSs is the variations of H-line profiles (Ferro & Giridhar 2003). NSVS0650 is composed of
low-mass stars which cover most of the properties of the T Tauri stars.
The most conspicuous line with dramatic profile variations in the system’s spectrum appears
to be the Hα. The Hαline is the most prominent feature in the spectra of TTSs. The presence
of the Li absorption line at 6708˚ A (see Fig. 5, for an example) and weak Hαin emission
leads us to classify the star as weak-lined T Tauri star. In Figure 6 we display the Hαline
Figure 6. Variation of the Hα line profiles of NSVS0650. The normalized spectrum at the Hα ordered with the orbital
phase. The vertical thick and thin lines show the rest wavelenghts corresponding to the primary and secondary component
region observed at various orbital phases in three consecutive nights. Each spectrum has been
normalized to the continuum. Julian date and the orbital phase for each observation are given
in each panel. On JD2454725 the Hαline appears to be a single, shallow absorption, i.e.,
filled-in by emission, at orbital phase of about 0.3058. At orbital phases of 0.4242, 0.5079
and 0.5939 the same line becomes single, emission above the continuum and at phase of
0.6770 it turns to be an absorption again. On JD2454726, the following night, the Hαline
is seen as single absorption at phases of 0.1404 and 0.2237, whereas double-peaked emission
profiles at phases of 0.3162 and 0.3996, but it turns to absorption in a short time interval
NSVS06507557; a low-mass double-lined eclipsing binary Download full-text
Figure 7. Variation of the forbidden emission line profiles of the [Oi] at 6300˚ A.
at phases of 0.4829 and 0.5973. The Hαemission line profile at orbital phase of 0.3996 has
an unexpected shape because it is very resemblance of inverse P Cygni profile, most similar
to UXTau A (see Reipurth et al. 1996). The dramatic changes in the shape of the Hαline,
collected on JD2454727, are clearly seen in the last five panels of Fig. 6. The Hαline in
the spectra of the NSVS0650 taken at phases of about 0.2114, 0.3141 and 0.3772 displays
blue-shifted absorption, similar to wTTS GGTau (Folha & Emerson, 2001). It turns to be
single absorption at orbital phases of 0.4610 and 0.5410.
The higher Balmer series, Hβ and Hγ lines of NSVS 0650 generally appear to be in
emission at all orbital phases. Again, dramatic line profile changes are evident. Inverse P
Cygni profiles are also visible at some orbital phases.
The existence of blue-shifted absorption components in the Balmer lines of TTSs’ spectra
was first noted by Herbig (1962), who suggested that these absorption components are
evidence for strong stellar winds. On the other hand Walker (1972) drew attention to the
wTTSs which have red-shifted absorption in the higher-order Balmer lines. These inverse P
Cygni profiles have generally been interpreted in terms of material accreting onto the young
stars. The optical observations of unidentified Einstein Observatory X-ray satellite sources
led to the the discovery of many TTSs with weak Hα and IR excess emission (Strom et