We investigate the 21 cm signature that may arise from the intergalact
ic medium (IGM) prior to the epoch of full reionization (z > 5). In sc
enarios in which the IGM is reionized by discrete sources of photoioni
zing radiation, the neutral gas that has not yet been engulfed by an H
II region may easily be preheated to temperatures well above that of
the cosmic background radiation (CBR), rendering the IGM invisible in
absorption against the CBR. We identify three possible preheating mech
anisms: (1) photoelectric heating by soft X-rays from QSOs, (2) photoe
lectric heating by soft X-rays from early galactic halos, and (3) reso
nant scattering of the continuum UV radiation from an early generation
of stars. We find that bright quasars with only a small fraction of t
he observed comoving density at z similar to 4 will suffice to preheat
the entire universe at z greater than or similar to 6. We also show t
hat, in a cold dark matter dominated cosmology, the thermal bremsstrah
lung radiation associated with collapsing galactic mass halos (10(10)-
10(11) M.) may warm the IGM to similar to 100 K by z similar to 7. Alt
ernatively, the equivalent of similar to 10% of the star formation rat
e density in the local universe, whether in isolated pregalactic stars
, dwarf, or normal galaxies, would be capable of heating the entire IG
M to a temperature above that of the CBR by Ly alpha scattering in a s
mall fraction of the Hubble time at z similar to 6. In the presence of
a sufficiently strong ambient flux of Ly alpha photons, the hyperfine
transition in the warmed H I will be excited. A beam differencing exp
eriment would detect a patchwork of emission, both in frequency and in
angle across the sky. This patchwork could serve as a valuable tool f
or understanding the epoch, nature, and sources of the reionization of
the universe, and their implications for cosmology. We demonstrate th
at isolated QSOs will produce detectable signals at meter wavelengths
within their ''spheres of influence'' over which they warm the IGM. As
a result of the redshifted 21 cm radiation emitted by warm H I bubble
s, the spectrum of the radio extragalactic background will display fre
quency structure with velocity widths up to 10,000 km s(-1). Broad bea
m observations would reveal corresponding angular fluctuations in the
sky intensity with delta T/T less than or similar to 10(-3) on scales
theta similar to 1 degrees. This scale is set either by the ''thermali
zation distance'' from a QSO within which Ly alpha pumping determines
the spin temperature of the IGM or by the quasar lifetime. Radio measu
rements near 235 and 150 MHz, as will be possible in the near future u
sing the Giant Metrewave Radio Telescope, may provide the first detect
ion of a neutral IGM at 5 less than or similar to z less than or simil
ar to 10. A next generation facility like the Square Kilometer Array I
nterferometer could effectively open much of the universe to a direct
study of the reheating epoch and possibly probe the transition from a
neutral universe to one that is fully ionized.