Taken together, the Coherence Principle of Multimedia Learning Theory and the Integrated Systems Hypothesis propose that co-occurring and semantically congruent verbal and visual information should be integrated into one mental representation that enhances memory. The purpose of this paper was to examine how learning pseudowords with matching versus mismatching gestures affects subsequent identification and integration of these newly learned pseudowords into read sentential contexts. Additionally, the pseudowords were manipulated to occur in either semantically congruent or semantically incongruent read sentential contexts, based on the pseudowords' learned definition. To investigate the research question, two experiments utilizing self-paced reading paradigms were employed. Results of Experiment 1 indicated partial support for the Integrated Systems Hypothesis. In Experiment 2, results indicated that pseudowords learned with matching gestures were identified more quickly and accurately after being read in semantically congruent sentences compared to semantically incongruent sentences, as was expected based on the Integrated Systems Hypothesis as well as the Coherence Principle of Multimedia Learning Theory. Additional results and implications are reported.