Fritz Lang's M (1931), contains both dialogue sequences and silent sequences with music or sound effects.  Lang edited the sound as if he were editing the visuals. 

We are introduced to the murder in shadow when he speaks to a young girl, Elsie.  We hear the conversation he makes with her, but we see only his shadow, which is ironically shown on a reward poster for his capture. 

Lang then set up a parallel action sequence by intercutting shots of the murderer with the young girl's mother. The culmination of the scene relies wholly on sound for its continuity. The mother calls out for her child. Each time she calls for Elsie, we see a different visual: out of the window of home, down the stairs, out into the yard where the laundry dries, to the empty dinner table where Elsie would sit, and finally far away to the child's ball rolling out of a treed area and  to a balloon stuck in a telephone line.  

With each shot, cries became more distant. For the last two shots, the mother's cries are no more  than faint echo.  

In this sequence, the primary continuity comes from the soundtrack. The mother's cries unify all the various shots, and the sense of distance implied by tone of the call suggests that Elsie is now lost to her mother. 

Later in the film, Lang elaborates on this use of sound to provide the unifying idea for a sequence. In one scene, the minister complains to the chief of police that they must find the killer of Elsie. The conversation reveals the scope of the investigation. As they speak, we see visual details of the search for the killer. The visuals show a variety of activities, including the discovery of a candy wrapper at the scene of the crime and the subsequent investigation of candy shops. Geographically, the police investigation moves all around the town and takes place over an extended period of time. These time and place shifts are all coordinated through the conversation between the minister and the chief of police. 

In terms of screen time, the conversation is five minutes long, but it communicates and investigation that takes place over many days and in many places. We sense the police department's commitment but also its frustration at the lack of results. 

What follows is the famous scene of parallel action where Lang intercut two meetings. The police and the crimal underworld meet separately, and the leaders of both organizations discuss their frustrations about the child murderer and devise strategies for capturing him.  

Rather than simply relying on visuals parallel action, Lang cut on dialogue at one point, starting a sentence in police camp and ending it in the criminal meeting. The crosscutting is all driven by dialogue. There are common visual elements: the meeting setting, the smoking room, the seating, the prominence of one leader in each group. Despite these viusals cues, it is the dialogue that is used to set up the parallel action and to give the audience a sense of progress. Unlike Griffith's train chase, there is no visual dynamic to carry us toward a resolution, nor is there a metric montage. The pace and character of the dialogue establish and carry us through this scene. 

Lang used sound as if it were another visual element, editing it freely. Notably is how Lang used the design of sound to overcome space and time issues. Through his use of dialogue over the visuals, time collapses and the audience moves all about the city with greater ease than if he had straight-cut the visuals 
 

FilmSound.org    www.filmsound.org       Oxford University: "...an excellent collection of resources and links.."