Why does Prism leave some interpolated X values blank?

When you ask Prism to interpolate X values from Y values you enter, it will leave the result blank in three cases:

• The input value for Y falls out of the vertical range of the standard curve. This happens when you fit a sigmoidal log(dose) vs response curve. If you enter an known  Y value that is larger than the best-fit value of Top, or smaller than the best-fit value of Bottom, there simply is no corresponding X value, and Prism will simply leave that cell of the Results sheet blank. Note that "top" and "bottom" apply to the best-fit values of those fitted parameters, not the maximum and minimum values of the input data. It simply is impossible to interpolate a X value if the Y value is too high or too low to be on the curve. In the example below, it is not possible to interpolate when Y=1700 because it is above the top plateau of the curve.
  
• The interpolated X value falls way outside the output range of X for the fitted curve. Prism will try to extrapolate a bit beyond the range of X values for which it calculates and draws the curve. But it won't extrapolate far from the ends of the curve. You can extend this range. From the Nonlinear regression Parameters dialog, choose the Range tab and enter a lower value for the minimum X and/or a higher value for the maximum X.
• You entered the unknowns in rows above the standard curve values. The unknowns must be entered in rows below the standards.

Prism can also report the 95% confidence interval for the interpolated values. It only does this when the interpolated X value is within the range of the standard curve. If the interpolated value is extrapolated beyond that range, Prism leaves the confidence interval blank.