The effects of irrelevant variations in the color and depth of elements on participants' ability to detect and discriminate elements defined by a difference in orientation were compared. Consistent with previous research, it was found that there was no effect or small effects if the targets were single elements in visual search tasks and that there were large effects for targets defined by several elements defining an area in visual segmentation tasks. It is suggested that the reason for the discrepancy between the 2 paradigms lies in the need for grouping processes in segmentation experiments. This notion was examined in 3 additional experiments that manipulated grouping processes through task demands and stimulus design. The data provide tentative support for this notion.