Clinical neglect patients overattend to stimuli on their right, whereas the general population overattend to the left (pseudoneglect). Both phenomena are affected by viewing distance, whereby the attentional biases are attenuated as the stimulus moves from near to far space. Both are also affected by stimulus length and reduce in strength, or even reverse (the crossover effect), as length decreases. To gain an insight into the cognitive/neural mechanisms that underlie the effects of viewing distance and stimulus length, in two experiments we examined the interaction between the variables. In Experiment 1 we asked university students (n = 20) to perform a horizontal landmark bisection task with lines presented at varying lengths (1.2°, 6.3°, and 18.4° of viewing angle) and distances (450 and 1,350 mm). A crossover effect and pseudoneglect were observed for the short and the long lines, respectively. An effect of viewing distance was only observed for long lines. Experiment 2 was the same, except that the lines were rotated to form vertical lines. No crossover effect was observed for the short lines, but an upward bias was observed for the long lines. Once again, an effect of viewing distance was only apparent for the long lines. These results demonstrate that the crossover effect is not a general property of short lines and is specific to the horizontal dimension. Models of crossover therefore need to incorporate processes related to left–right asymmetries. The results also demonstrate that viewing distance only affects long lines, and that this happens irrespective of orientation. A model of viewing distance is discussed that incorporates a right hemisphere mechanism specialized for an interaction between the ventral and dorsal streams.