This paper starts from the assumption that the level of emotionality in hostage-negotiation differs between phases of escalation and de-escalation. Three hypotheses were derived, stating that (1) the overall-level of emotionality should be higher for a hostage-taker as compared to a negotiator; (2) for both the level of emotionality is supposed to be higher in escalating phases than in de-escalating phases; (3) the change from de-escalating to escalating phases should be more pronounced for an hostage-taker than for a negotiator. Seventeen phone calls from an authentic hostage taking, classified as escalating (10) and deescalating (7), were analyzed to test these hypotheses. The Gottschalk-Gleser hostility scale and Rogan and Hammer's message affect were used as indicators of emotionality. Significant main effects and interactions were identified in this case study by both instruments, supporting our hypotheses. Results are discussed against the background of the present research conditions.