1 Introduction

Hydrogen lines in emission are the hallmark of Classical T Tauri stars (CTTS). These lines tend to reflect the dynamics of the region where they are formed. Clear examples are normal or inverse P Cygni profiles found in the Balmer lines of many CTTS. Traditionally, the strong hydrogen emission was interpreted in terms of mass loss (Lago 1979; De Campli 1981; Hartmann et al. 1982; Natta et al. 1988; Hartmann & Kenyon 1990). The last decade saw a shift in interpretation towards mass accretion (Calvet & Hartmann 1992; Hartmann et al. 1994; Edwards et al. 1994; Muzerolle et al. 1998a; Muzerolle et al. 1998b), following the magnetospheric accretion model proposed by Camenzind (1990), Königl (1991) and Shu et al. (1994).

Considerable effort has been made in trying to model line profiles, especially those of H ${\rm\alpha}$, both in wind scenarios (Hartmann et al. 1990; Calvet et al. 1992; Grinin & Mitskevich 1991; Mitskevich et al. 1993; Pedrosa 1996) and accretion scenarios (Bertout 1977; Bertout 1979; Bastian 1982; Calvet & Hartmann 1992; Hartmann et al. 1994; Muzerolle et al. 1998a). This has been accompanied by a similar effort on the observational side, aiming at constraining the existing models (e.g. Edwards et al. 1994; Fernandez et al. 1995; Reipurth et al. 1996; Muzerolle et al. 1998b; Alencar & Basri 2000).

A wind scenario for the formation of the hydrogen emission lines observed in T Tauri stars was much discussed over past years. The predominance of blueshifted absorptions in the H ${\rm\alpha}$ line profiles, the success in modeling some of these observations with wind models, together with the unequivocal presence of winds in CTTS has been the main driving force behind this interpretation. Despite the recent shift towards an interpretation based on the magnetospheric accretion model, the debate is far from settled, as it is clearly seen in Alencar & Basri (2000). Also, for a given star, both winds and accretion flows may contribute to hydrogen line emission, with different lines being formed in distinct regions and/or affected differently by those regions.

The successes and failures of the models in explaining hydrogen line profiles in CTTS have been based on the study of lines from the Balmer series. The exception is the work by Najita et al. (1996) where the Br ${\rm\gamma }$ emission line is studied for a very small number of objects.

The strongest hydrogen lines in the near infrared part of the electromagnetic spectrum are lines from the Paschen and Brackett series. These lines arise in higher energy levels when compared to the lower Balmer lines (e.g. H ${\rm\alpha}$ and H ${\rm\beta }$) and they are generally optically thinner than the latter. One expects them to form at different depths into a cloud of hydrogen. Near infrared hydrogen lines should form in the denser parts of the circumstellar envelope of CTTS and, in particular, they should trace infalling material in a magnetospheric accretion scenario. The development of near infrared high spectral resolution high sensitivity spectroscopy allows the study of profiles of near infrared hydrogen lines. Such a study imposes strong constraints on models and contributes to the understanding of the origin of hydrogen emission in CTTS.

The structure of the paper is as follows: observations and data reduction are described in Sect. 2, the line spectra are presented in Sect. 3, classification of the line profiles is done in Sect. 4, the parameters that characterize the line profiles and their statistics are discussed in Sects. 5 and 6. The results are compared with other observations in Sect. 7, and with models in Sect. 8. The results are discussed in Sect. 9 and conclusions are presented in Sect. 10.