Search for a command to run...
We present the rst results from a program of near-infrared spectroscopy aimed at studying the familiar rest-frame optical emission lines from the H II regions of Lyman break galaxies at z ^3. By targeting redshifts that bring the lines of interest into gaps between the strong OH sky emission, we have been successful in detecting Balmer and [O III] emission lines in all ve galaxies observed so far with CGS4 on UKIRT. The typical line uxes are a few times 10~17 ergs s~1 cm~2, approximately 1 order of magnitude lower than the limits reached with wide-eld narrowband imaging surveys. For a Salpeter initial mass function and a km s~1 Mpc~1, cosmology, the Hb luminosities uncorrected for H 0 \ 70 q 0 \ 0.1 dust extinction imply star formation rates of 20270 yr~1 ; these values are greater than those that M _ may have been deduced from the ultraviolet continuum luminosities at 1500 by factors of between A D0.7 and D7. Uncertainties in the shape of the reddening curve and in the intrinsic UV continuum slope do not yet allow us to assess accurately the level of dust extinction ; however, on the basis of the present limited sample, it appears that an extinction of 12 mag at 1500 may be typical of Lyman A break galaxies. This value is consistent with recent estimates of dust obscuration in star-forming galaxies at z 1 and does not require a substantial revision of the broad picture of star formation over the Hubble time proposed by Madau and coworkers in 1996. In four out of ve cases the velocity dispersion of the emission line gas is p ^70 km s~1, while in the fth the line widths are nearly three times larger. Virial masses are suggested, but both velocities and masses could be higher, M vir B (15) ] 1010 M _ because our observations are only sensitive to the brightest cores of these systems where the line widths may not sample the full gravitational potential. The relative redshifts of interstellar absorption, nebular emission, and Lya emission lines di er by several hundred km s~1 and suggest that large-scale outows may be a common characteristic of Lyman break galaxies. The forthcoming availability of highresolution infrared spectrographs on large telescopes will soon allow all of these questions to be addressed in much greater detail.