In simultaneous interpreting, the source-language speaker can use gestures, facial expressions, proxemics, prosody and graphic material, whereas the interpreter, who is only perceived by her/his listeners through her/his voice in the headphones, has a more limited use of nonverbal elements such as prosody and suprasegmental elements, including intonation, stress, pauses and delivery speed. Ideally, according to Kirchhoff, the interpreter should sound like a speaker expressing him/herself spontaneously and naturally, but such output is difficult to find in real settings. In fact, as noted by Shlesinger and Williams, the interpreter’s situation with overlapping source-speech perception and target-speech production seems to generate unnatural, typical patterns. The paper discusses the possible reasons for this fact and describes the resulting non-verbal phenomena in the target speech on the basis of an authentic English-German corpus which was part of the author’s PhD project at the FASK in Germersheim.