Another way of displaying variation in feature values is to create a gradient
of symbol sizes corresponding to the variable. This works mostly for point
layers but is also sometimes applied to line layers under certain circumstances.
A classic point feature example is city points where the largest cities are sym‑
bolized with big squares and small towns are small squares and other towns
are in between. A line feature such as roads can be displayed with different
thicknesses based on traffic counts. This is different from the dot map method
because the size of the symbol is changed, whereas in the dot map method
the dots stay the same size and only the number of dots changes. One caution
to be aware of when creating a graduated‑symbol map is to avoid the use of
circles as your symbol if you have more than three or so categories since small
changes in circle size may be difficult for the human eye to detect.1 The alter‑
native is to use squares, triangles, or some other shape instead.