7 Comparison with other observations

7.1 NIR data

High resolution spectroscopy of near infrared hydrogen lines in T Tauri stars is not commonly found in the literature. The only publication based on data with spectral resolution and signal-to-noise ratio comparable to those of the data presented in this work is Najita et al. (1996). These authors present Br ${\rm\gamma }$line profiles for five T Tauri stars (DG Tau, AS 353A, V1331 Cyg, AS 205A and S CrA), one Class I source (SVS 13) and one low luminosity embedded object (WL 16). Only two objects (DG Tau and V1331 Cyg) are common to both their and our sample. The Br ${\rm\gamma }$ line profiles shown by Najita et al. for those two objects are very similar to the ones reported here. The Br ${\rm\gamma }$ line profiles from the remaining objects presented in Najita et al. have shapes similar to the typical type I or type IV R profiles presented in this work.

Hamann et al. (1988) present high resolution, but relatively low signal-to-noise, Br ${\rm\gamma }$ spectra of DG Tau, GW Ori, HL Tau, SU Aur and T Tau, all of which are in the sample studied here. Despite the low signal-to-noise ratio, it can be seen that the line profiles displayed in their Fig. 1 have the same characteristics of the ones presented here. Even SU Aur, seen with a slightly redshifted absorption feature, is somewhat similar to the one shown in our Fig. 2.

Comparison with published line profiles other than the ones mentioned above is very difficult due to the much lower spectral resolutions and signal-to-noise ratios. Examples of those are the data shown in Giovanardi et al. (1991) and Evans et al. (1987) where the line profiles are defined by only a few spectral points.

Photometric standard stars or simultaneous photometry is not available for the data set presented here, hence reliable line fluxes could not be determined. Comparison with line fluxes available in the literature is, therefore, not possible. Variability in T Tauri stars, in particular in the NIR magnitudes, implies that any lines fluxes determined using non-simultaneous photometry in conjunction with the measured equivalent widths, would yield unreliable results.

7.2 Balmer lines

The results obtained in Sect. 4 for the NIR lines are in complete contrast to what is found for H ${\rm\alpha}$ by Edwards et al. (1994), Fernandez et al. (1995), Reipurth et al. (1996) and Alencar & Basri (2000). Reipurth et al. (1996), who introduce the classification scheme used in this work show that, for a sample of 43 TTS, 54% of the H ${\rm\alpha}$ line profiles have blueshifted absorptions whereas redshifted absorption features are found in 21% of the lines. Only 5% of the profiles are actually classified as type IV R (IPC). The remaining 25% of the line profiles are classified as type I. As a result the Af of H ${\rm\alpha}$ is predominantly less than unity, contrary to what is observed for the NIR lines. The lack of blueshifted absorptions in the NIR lines and the much higher frequency of IPC line profiles are the most prominent differences between H ${\rm\alpha}$ and the NIR lines.

Edwards et al. (1994) also present high resolution observations of higher members of the Balmer series (from H ${\rm\beta }$ to H ${\rm\delta}$) for a sample of 15 TTS. These authors show that redshifted absorptions are seen frequently in residual line profiles of the higher Balmer lines (henceforth HBLs). Of the 8 TTS shown in Edwards et al. (1994) to display IPC structure at least in one of the Balmer lines presented there, 5 also display an IPC profile at Pa ${\rm\beta }$ and/or at Br ${\rm\gamma }$. GM Aur displays IPC structure in Pa ${\rm\beta }$ but not in any of the Balmer lines shown in Edwards et al. (1994). Furthermore, for the Balmer lines, both redshifted and blueshifted absorptions can coexist on the same line profile (e.g. DF Tau in Edwards et al. 1994 and Alencar & Basri 2000). That is not seen in any of the NIR line profiles presented here. The Af of the HBLs falls mostly between 1 and 2.5. Similar values for this parameter are found for Pa ${\rm\beta }$ and Br ${\rm\gamma }$. The vast majority of the HBLs with redshifted absorption features have them located between about 200 and $300\ {\rm km\ s}^{-1}$. Only two of the objects in the Edwards' sample have the centre of the absorption below $200\ {\rm km\ s}^{-1}$. When compared to the HBLs, the NIR IPC profiles tend to have the centre of the absorption feature located at lower velocities, often below $200\ {\rm km\ s}^{-1}$ (see Tables 6 and 7). Furthermore, the NIR lines analysed here tend to have line wings less extended than those of the HBLs by about $200\ {\rm km\ s}^{-1}$. Visual inspection of the Balmer profiles in Edwards et al. (1994) reveal that most line wings extend up to $500\ {\rm km\ s}^{-1}$, especially in the blue. From Fig. 6, we see that wings in NIR lines extend typically to about $300\ {\rm km\ s}^{-1}$. Also, IPC HBLs tend to have line emission beyond the redshifted absorption feature. In NIR lines, emission seems to stop blueward of the redshifted absorption feature.

In summary, while the NIR lines show very different properties from those of H ${\rm\alpha}$, in some aspects they are similar to higher Balmer lines. They are not as wide as the latter, nor seem to be so much influenced by outflowing material. Like the higher Balmer lines, the NIR lines are more prone to IPC structure than H ${\rm\alpha}$. The different velocities at which the redshifted absorption feature occurs in IPC higher Balmer lines and in IPC NIR lines should provide constraints on models that hope to explain the formation of hydrogen lines in TTS.