The model of the stratosphere proposed by Strong and Plass assumed that the spectral lines had the Lorentz shape. Actually the center of the line shifts to a lower frequency by an amount proportional to the pressure, and the shape of the line becomes asymmetrical at large distances from the center. The effect of the line shift on atmospheric heat-transfer is shown to be small and can be neglected except for very accurate calculations. On the other hand, for the strong spectral lines that play the dominant role in determining atmospheric heat-transfer, the difference between the results calculated using the Lorentz and asymmetrical line-shapes is large. A discussion is given of the effect of the asymmetry on the actual bands of interest in the atmosphere. The difficulties of making exact theoretical calculations of atmospheric heat-transfer are also discussed.