It has often been hypothesized that the broad line-emitting clouds in quasars and Seyfert galaxies form from thermally unstable condensations in a hot (T ∼ 108) cloud-confining medium. We raise here a number of difficulties with this type of cloud formation. The inherently slow growth of thermal instabilities requires implausibly large, finite initial perturbation amplitudes, particularly if thermal condensations are required to form into dense clouds in times less than the characteristic flow time in quasar (and especially in Seyfert) line-emitting regions. Thermal instabilities require rather unusual (entropy) perturbations for which there is no obvious physical cause. Our studies of the detailed hydrodynamical evolution of thermally unstable regions in quiescent and expanding hot quasar atmospheres indicate that while the radius of the cloud condensation increases during the later, nonlinear stages of the thermal instability, the sonic radius in the gas locally flowing toward the condensation increases even faster, prolonging or indefinitely postponing the time at which the cloud attains (subsonic) pressure equilibrium with the ambient gas. The dynamic pressure in the flow toward the unstable condensation is much larger than the ambient pressure. To attain densities of broad-line clouds, the ratio of the energy density of tangled magnetic fields to the thermal energy density in the initial unperturbed hot gas must be implausibly small, ≲10-6. In general, combined gravitational, pressure, and radiative forces may dislocate the high density cores of thermally unstable regions from the centers of the locally converging inflow; this would significantly restrict the size of the clouds that can form. Thermally unstable protoclouds can also be disrupted or greatly agitated by the strong (anisotropic) radiation field of the quasar before the quasar cloud condensations can contribute to the broad-line profile. For otherwise optimal initial conditions, the density of thermally unstable condensations may exceed values appropriate for the broad-line region ( ∼ 109-1010 cm-3). We conclude that emission-line clouds are unlikely to arise from condensations in a hotter medium. It is more likely that they appear first in the broad-line region having densities comparable to or greater than those of typical line-emitting clouds.