Residual Plots
When a linear model is applied to a given dataset and the residuals are plotted in the vertical axis, we get the residual plot.
If the linear model is adequate to represent the dataset, then the residual plot should have scattered random points.
But if any pattern can be found on the residual plots, like another line, or any indication of a known function, then the linear model is not adequate. Maybe a non-linear model is suitable to represent the data.
A very quick view of the residual plot indicates a clear pattern of some well-defined curve, like a parabola.
Thus, the choice is:
d. non-random pattern, supporting a non-linear model