This paper presents a comparison of experimental results with predictions obtained from Tullis' (1984) model of search times for tabular displays. Three levels of information density for displays with and without highlighting were used in a series of experiments. The highlighting of information was done by adding graphic boundaries (lines). Two levels of highlighting were used. A question-answer type of visual search was performed for two different tasks. The search time results are discussed and a method for utilizing Tullis' model for highlighted displays is suggested.