Results

The veiling measurements were performed near Pa Beta and near Br Gamma as discussed above. Hereafter, the veiling measured at Pa Beta will be referred to by r_J and that measured at Br Gamma by r_K.

In 25% of the T Tauri stars observed at J and in 30% of the T Tauri stars observed at K photospheric lines are not identified. These stars are: CW Tau, DG Tau, DL Tau, DR Tau, GW Ori, HL Tau, RW Aur, RY Tau, SU Aur, V1331 Cyg, YY Ori and XZ Tau at J and CW Tau, DI Cep, DR Tau, GW Ori, HL Tau, RW Aur, RY Tau, SU Aur, V1331 Cyg and XZ Tau at K. There are two possible explanations for this. On the one hand the photospheric lines can be masked under a noisy T Tauri star spectrum. On the other hand a T Tauri star's spectrum might be significantly veiled so that the photospheric lines cannot be identified. For the above stars lower limits on the J and K veiling were obtained by veiling the spectrum of an appropriate template star up to a level such that the depth of the photospheric absorption features became comparable to the fluctuations on the T Tauri star's spectrum. The results are presented in Table 4.1. The veiling results obtained for the stars for which the photospheric lines were clearly identified are also presented in Table 4.1.

Table 4.1: Results from veiling measurements. Column 1 shows the name of the star, Column 2 its spectral type, Column 3 lists the veiling measurement at Pa Beta, Column 4 the spectral type of the template star used, Column 5 has the veiling measurement at Br Gamma and Column 6 the spectral type of the template star used. C means that no photospheric are identified at optical or at these IR spectra; '-' means that the star was not observed at those wavelengths; a) No appropriate template star for a veiling measurement .
  

With the exception of DI Cep, whenever the photospheric spectrum was identified at one wavelength range (say J) it was also identified at the other (say K). The K band spectrum of DI Cep does not show any clear photospheric feature despite photospheric features being clearly present at J. Having a spectral type G8V, as determined from optical spectroscopy, the Si 4590.17 cm^-1 should be clearly identified in the K band spectrum. A lower limit of is obtained for DI Cep. The J band spectrum is compatible with the G8V classification. For BM And it was only possible to obtain a lower limit on r_J. For this star, one cannot say that photospheric features are not present since the blue wing of Mn 7749.92 cm^-1 is clearly seen at the red edge of the spectrum (see Figure 3.12).

From the data presented in Table 4.1 one arrives to the conclusion that, within the quoted uncertainties, of the 37 stars for which the J veiling was measured, 12 stars have values compatible with . Of the remaining stars can be excluded at the level in 12 cases and at the in 11 cases. The results for the K veiling imply that, of the 23 stars with measured, 5 are compatible with , in 6 can be excluded at the level and in 11 stars can be excluded at the level.

In Figure 4.5 r_J is plotted versus r_K for the stars that had the veiling determined at both wavelength regions.

Despite the large uncertainties present there is some correlation between r_J and r_K. Over-plotted in Figure 4.5 (dotted line) is the line corresponding to r_J = r_K. One sees that there is a direct correlation between r_J and r_K, with r_K growing faster than r_J does.

The histograms of the r_J and r_K measurements for the stars for which the veiling could be determined are shown in Figure 4.6. The method used to compute these histograms, as well as the ones presented in the remaining of this work is described in Appendix A.

The mean values for the distributions are < r_J > = 0.56 and < r_K > = 1.31 and the respective standard deviations are and . Although most stars have relatively low J and K veiling, the distributions clearly have an extended tail towards higher values of veiling.

The histograms of Figure 4.6 do not take into account the stars for which only lower limits where obtained. The lower limits tend to be relatively high, implying that the true distributions are actually enhanced for higher values of veiling when compared to those shown in Figure 4.6. This is shown in Figure 4.7, where the dashed bins, representing the distribution of the lower limits on the veiling, are plotted over the distributions of the veiling measurements (solid lines).

The lowest values of the upper limits occur for early type stars, like GW Ori, RW Aur and RY Tau, for which the photospheric lines in the template stars are themselves not very strong, making the veiling measurements difficult (note that the strong Magnesium feature is, for the T Tauri \ stars, out of the observed wavelength region due to their radial velocities) or for stars with low signal-to-noise ratio, like XZ Tau. The un-identification of photospheric features in the spectra of some T Tauri tends therefore to reflect high veiling states for the near infrared wavelength region.

Furthermore, magnetic fields tend to broaden photospheric lines and thence increase their EW, which is proportional to the magnetic sensitivity , where is the wavelength of the line and is the efective Landé factor [Guenther et al. 1998]. As discussed by Guenther et al. (1998), such an increase in the EW might result in an underestimation of the measured veiling for T Tauri stars if the Landé factor of the photospheric lines used for the veiling determination is not zero. The Landé factors for the photospheric lines used for the veiling determination in this work are [Ruedi et al. 1995]: at J, 0.67 for Ti 7791.23 cm^-1, 1.08 for Ti 7781.77 cm^-1 and 1.50 for Fe 7761.99 cm^-1; at K 1.25 for Ti 4589.50 cm^-1. Therefore, the results presented above for the veiling underestimate the true veiling if there are magnetic fields, as expected.