REMEASUREMENT OF THE H-I-GUNN-PETERSON EFFECT TOWARD QSO-PKS-1937-101WITH KECK OBSERVATIONS

We present the first measurement of the H I Gunn-Peterson effect using the Keck 10 m telescope, observing the high-redshift QSO PKS 1937-101 (z = 3.787). The high-resolution echelle (HIRES) spectra, with FWHM s imilar to 15 km s(-1) and a signal-to-noise ratio (S/N) similar to 50 per spectral resolution element, allows us to resolve many weak lines down to N-HI = 10(12) cm(-2), thus reducing the line-blanketing proble m compared with previous data. Based on intensity-distribution analysi s, we find that a maximum likelihood best fit yields a Gunn-Peterson t ype of opacity tau(GP) = 0.113 +/- 0.020 in addition to a power-law Ly alpha absorption-line population with beta of 1.7 down to N-HI = 10(1 2) cm(-2). There remains systematic uncertainty in this result because of problems extrapolating the spectral continuum from the red side of the Ly alpha emission line. This is consistent with the previous stud y of the same QSO in low S/N data using weighted intensity function an alysis (Fang & Crotts 1995). It indicates that this previous method su cceeds in measuring the Ly alpha forest continuum level at low S/N, wh ich is essential in extending the technique to possible fainter QSOs w ith minimum emission-line contamination for reliable continuum extrapo lation. We further discuss problems of severe line blanketing, even in Keck spectra for QSOs at z greater than or equal to 4.5, and show the effectiveness of the weighted intensity function method in measuring continuum levels in extremely crowded Ly alpha absorption spectra for redshifts as high as z > 5.