The study reports systematic distance errors in reading raised-line maps by
touch, and how they can be reduced. We show that T-shaped road junctions p
roduce the typical error due to overestimating the length of the bisecting
road compared to the bisected road. The error was not reduced when the targ
et location was marked initially by a symbol. However, it was eliminated by
instructions to use both hands concurrently to scan the route relative to
an external square frame surrounding the map layout and to the body midline
. Road junctions, which resemble 'Muller-Lyer' configurations, produced a s
ignificant overestimation of the length of a road that ended in diverging s
ide roads relative to underestimating a road section with converging side r
oads, Spatial anchor cues from landmarks along the route, either alone or i
n conjunction with the spatial frame instructions, eliminated the illusion
equally. We discuss theoretical implications, and the use of map frames and
landmarks as spatial guides to improve map use by touch.