Previous work has demonstrated that people can collaborate during search by adopting a strategy of spatially dividing the search labor. But does this strategy hinge on the information available in the task? And can searchers coordinate their behavior in a fine-grained way, at the micro-level? I had groups of 2-4 people engage in 5 collaborative search tasks. In Experiment 1 subjects searched for an oval dot among circular dots. Experiment 2 added color, with red, blue, green, and black dots segregated into irregular regions. Experiment 3 used multiple potential targets (1, 4, or 8 photo-realistic objects) in 14-item displays. I quantified division-of-labor by correlating targets’ properties (e.g., location, color, identity) with individuals’ responses within the collaborating group. Consistent with previous work, subjects in Experiment 1 divided the search labor spatially, splitting the display in halves (2-person condition) or quadrants (4-person condition). Subjects in Experiment 2 divided the search labor by feature, with each of 4 members searching a different color. In Experiment 3 subjects divided the labor by targets rather space or feature, with each searcher taking responsibility for a different potential target from the preview set. Experiment 4 looked for tacit coordination during collaborative searches. Experiment 5 used a shared-gaze methodology to explore coordination at a micro-level; remotely located pairs viewed matching displays with each other’s eyegaze cursor superimposed, so that each could monitor where the other was looking in real time. Stimuli were again colored dots, with only two colors and with one region 1/3 the size of the other. I found that when one person finished searching the smaller region, she often assisted her partner in searching the larger region. This assistance was targeted, such that she tended to look where her partner had not yet searched. I conclude that subjects use spatial division-of-labor collaborative strategies when the task does not allow for more meaningful divisions of labor. When the task allows, simple spatial division-of-labor strategies are replaced by collaborative strategies based on feature and target information. These analyses constitute the first demonstration of behavioral coordination during visual search at the micro-level.