Cyclovergence is a simultaneously occurring cyclorotation of the two eyes i
n opposite directions. Cyclovergence can be elicited visually by opposite c
yclorotation of the two eyes' images. It also can occur in conjunction with
horizontal vergence and vertical version in a stereotyped manner as descri
bed by the extended Listing's law (or L2). We manipulated L2-related and vi
sually evoked cyclovergence independently, using stereoscopic images of thr
ee-dimensional (3D) scenes. During pursuit in the midsagittal plane, cyclov
ergence followed L2. The amount of L2-related cyclovergence during pursuit
varied between subjects. Each pursuit trial was repeated three times. Two o
f the three trials had additional image rotation to visually evoke cyclover
gence. We could separate the L2-related and visual components of cycloverge
nce by subtraction of the cyclovergence response in matched trials that dif
fered only in the image rotation that was applied during pursuit. This indi
cates that visual and L2-related contributions to cyclovergence add linearl
y, suggesting the presence of two independent systems. Visually evoked cycl
overgence gains were characteristic for a given subject, little affected by
visual stimulus parameters, and usually low (0.1-0.5) when a static target
was fixated. Gain and phase lag of the visually evoked cyclovergence durin
g vertical pursuit was comparable with that during fixation of a static tar
get. The binocular orientations are in better agreement to orientations pre
dicted by L2 then would be predicted by nulling of the cyclodisparities. On
the basis of our results, we suggest that visually driven and L2-related c
yclovergence are independent of each other and superimpose linearly.