It has been proposed recently that the small size and long lifetime of
ultracompact (UC) H II regions could be due to pressure confinement i
f the thermal pressure of the ambient gas is higher than previous esti
mates. We point out that confinement by thermal pressure alone implies
emission measures in excess of observed values. We show that turbulen
t pressure, inferred from observed nonthermal velocities, is sufficien
t to confine UC H II regions and explain their longevity. We predict a
n anticorrelation between the size of UC H Ir regions and the velocity
dispersion of the ambient neutral gas and show that it is consistent
with existing observations.