=====================Observing shedule 

The observing set-up was fixed early in the first run to insure that the spectra were obtained in a uniform way. Our observing procedure was to first mark the fiducial position of the slit with respect to a visible-light acquisition camera using a bright optically- visible star in zero-order where the grating acts as a mirror producing an image of the star. The slit position was checked every two hours or after large telescope slews. For optically-visible stars, sources were directly placed on the slit using the acquisition camera. For optically-invisible sources that were infrared-bright (m_K < 10^m) we merely adjusted the position of the source in the slit in order to peak-up the source counts. The preferred acquisition method used for fainter stars (10^m < m_K < 12^m) was to move the grating to zero-order and verify directly that the star was centered on the slit. If available, a guide star was acquired for exposures of 10 seconds or more. After checking the orientation of the instrument with respect to the telescope, spectra were obtained at five positions along the slit using guided offsets of the telescope of 10.0 arcsec and 7.0 arcsec per step for the 2.1m and 4m respectively. The slit was oriented east-west in order to minimize the effect of tracking errors on the signal level. In most cases five independent source spectra were obtained in each of three grating/filter combinations: H/2283 (1.55-1.72 µm); K/3162 (2.12-2.25 µm); and K/3343 (2.23-2.36 µm). A complete set of observations was obtained before the next source was acquired.

Flat-field calibration frames were obtained by observing a dome target both with the dome lights on and off. The lights-off frames provided background measurement including dark and bias. Calibration spectra were obtained using arc lamps configured for use with existing optical spectrographs. These included the helium-neon-argon and thorium-argon lamps. Images were obtained both with the lamp lights on and off. A set of bright A0 standard stars was observed in order to correct for telluric absorption features. These standards were observed throughout the night at a variety of airmass values.