When normal emmetropic subjects view dichoptically an image sized differently for the two eyes saccades become rapidly disjunctive. This allows a bifoveal fixation immediately after the saccade. The disjunctivity can be retained for a few minutes under subsequent monocular viewing. This indicates oculomotor learning (e.g. van der Steen and Bruno, 1995; Kapoula et al., 1995). Prior studies used relatively important image size inequality (8% - 10%). The disparity created by such inequality corresponds to that created by a difference of power spherical lenses of about 4 - 5 dioptres. The purpose of the present study was to test weather feeble image size inequality (2%, approx. 1 dioptre) causes also disjunctivity of the amplitude of the saccades.

Six normal subjects participated in the study. All subjects wore an afocal magnifier of 2% in front of their preferred eye; consequently this eye viewed the image 2% larger. Subjects were seated at 1 m in front of a screen where a random-dot pattern was projected. They made saccades at 8° and 16° along the horizontal and vertical principal meridians and along tertiary positions at ±16° in the upper and lower field. Subjects performed this task for a training period of 16 min; the point aimed by each saccade was disparate, i.e. 2% more eccentric because of the magnifier. To test for learning before and after training saccades were recorded under monocular viewing (no disparity). Binocular eye movements were recorded with search coils (Robinson, 1963; Collewijn et al., 1975). A bidimensional calibration (Helmholtz co-ordinates) was used for data analysis.

During the training compensatory disjunctivity of the amplitude of the saccades occurred for the principal horizontal and vertical meridians; the group mean difference in the disjunctivity of saccades between the before and the during training condition was 0.14 ±0.04°, n = 6 subjects for horizontal saccades, and 0.15 ±0.07°, n = 6 for vertical saccades. Such increased disjunctivity persisted even after training suggesting learning: the group mean difference in the disjunctivity of saccades between the before and the after training condition was 0.12 ±0.02°, n = 6 for horizontal saccades, and 0.15 ±0.01°, n = 6 for vertical saccades. In contrast, for horizontal saccades to tertiary positions made in the upper and lower field no consistent changes in the disjunctivity occurred, neither during training nor after the training condition.

In conclusion this study shows that even a 2% image size inequality, frequently present in subjects wearing spectacles, stimulates oculomotor learning leading to persistent disjunctivity of saccades. The limitation of such mechanism for tertiary positions is an intriguing new finding. This limitation could be sensory, e.g. difficulty to evaluate horizontal and vertical disparity at the same time, and/or motor, e.g. difficulty to produce convergent or divergent movements when the eyes are at elevation or depression. Further research is needed.