We report on the X-ray spectral behavior of XTE J1550-564 during its 1998-1
999 outburst. XTE J1550-564 is an exceptionally bright X-ray nova and is al
so the third Galactic black hole candidate known to exhibit quasi-periodic
X-ray oscillations above 50 Hz. Our study is based on 209 pointed observati
ons using the PCA and HEXTE instruments on board the Rossi X-Ray Timing Exp
lorer (RXTE) spanning 250 days and covering the entire double-peaked erupti
on that occurred from 1998 September until 1999 May. The spectra are fitted
to a model including multicolor blackbody disk and power-law components. T
he spectra from the first half of the outburst are dominated by the power-l
aw component, whereas the spectra from the second half are dominated by the
disk component. The source is observed in the very high and high/soft outb
urst states of black hole X-ray novae. During the very high state, when the
power-law component dominated the spectrum, the inner disk radius is obser
ved to vary by more than an order of magnitude; the radius decreased by a f
actor of 16 in one day during a 6.8 crab Rare. If the larger of these obser
ved radii is taken to be the last stable orbit, then the smaller observed r
adius would imply that the inner edge of the disk is inside the event horiz
on! However, we conclude that the apparent variations of the inner disk rad
ius observed during periods of increased power-law emission are probably ca
used by the failure of the multicolor disk/power-law model; the actual phys
ical radius of the inner disk may remain fairly constant. This interpretati
on is supported by the fact that the observed inner disk radius remains app
roximately constant over 120 days in the high state, when the power-law com
ponent is weak, even though the disk flux and total flux vary by an order o
f magnitude. The mass of the black hole inferred by equating the approximat
ely constant inner disk radius observed in the high/soft state with the las
t stable orbit for a Schwarzschild black hole is M-BH = 7.4 M.(D/6 kpc)(cos
i)(-1/2).