The Antarctic plateau provides superb sites for infrared astronomy, a resul
t of the combination of low temperatures, low levels of precipitable water
vapor, high altitude, and atmospheric stability. We have undertaken measure
ments of the sky background from 1 to 5 pm at the South Pole, using a singl
e channel InSb spectrometer, the Infrared Photometer Spectrometer (IRPS), d
uring the winter (dark) period of 1995. The IRPS records the DC level of th
e sky flux through a 4 degrees beam and a variety of broadband and narrowba
nd (1%) filters. It can be scanned in elevation from horizon to horizon thr
ough the zenith. We find a 20-100 times reduction in the background of ther
mal emission compared to that from mid-latitude sites such as Siding Spring
and Mauna Kea, with typical background levels of 80-200 mu Jy arcsec(-2) a
t 2.43 mu m, 100-300 mJy arcsec(-2) at 3.6 mu m and similar to 0.5 Jy arcse
c(-2) at 4.8 mu m. Airglow emission contributes significantly to the sky fl
ux shortward of similar to 2.4 mu m, which is why the K-dark (2.27-2.45 mu
m) band emission does not drop to the 10-20 mu Jy arcsec(-2) levels origina
lly predicted. The darkest window for IR observations from the South Pole i
s from 2.35 to 2.45 mu m, where the fluxes from the atmosphere may drop to
as low as similar to 50 mu Jy arcsec(-2) at times. Airglow dominates the em
ission at J (1.25 mu m) and H (1.65 mu m), but the flux levels of 300-600 m
u Jy arcsec(-2) and 800-2000 mu Jy arcsec(-2), respectively, are also one-t
hird to one-half those at temperate sites. We find no evidence for any sign
ificant contribution from auroral emission to the K-dark band. During twili
ght, when the Sun is <10 degrees below the horizon, scattered sunlight cont
ributes to the sky background with a Rayleigh-type spectrum. Scattered moon
light is also evident in the sky emission at the J band when the Moon is up
.