The data from Experiment 2 support Alvarez and Cavanagh’s
(2005) hemifield-independence hypothesis more
strongly than do the data from Experiment 1. In Experiment
2, the mean error when the targets were in the same
hemifield was 33.6 deg, compared with 30.1 deg when
the targets were in different hemifields [t(19) 5 4.02, p ,
.01]. Whether this 3-deg difference in average error is the
level expected from hemifield independence is uncertain,
however. Alvarez and Cavanagh asserted that the hemifields
were fully independent, meaning that performance
in tracking two targets should be no worse than if only one
target had been tracked. The level of performance when
tracking only one target is unknown for this experiment,
since it was not tested.