For real landscapes, the local slope is related to the area of the drainage basin. Such relations do not always indicate a causal link between drainage area and local slope. We provide a general and quantitative assessment of this statistical effect. Slope-area relations in which slopes decay slower than roughly the cube root of drainage area are insensitive to processes that have shaped the landscape. To illustrate this point, surfaces unrelated to erosion are compared to real landscapes. The fluctuations of slope for constant drainage area provide a means of distinguishing between causal and random effects.