Background: The content of medical records is a potential source of miscommunication between clinicians. Doctors' written entries have been criticised for their illegibility and ambiguity, but no studies have examined doctors' drawings that are commonly used for recording auscultation findings. Objective: To compare doctors' drawings of auscultation findings, based on identical clinical information. Methods: Doctors at the Royal London Hospital and a group of London based general practitioners (GPs) documented a respiratory examination with a drawing of the auscultation findings of bilateral mid and lower zone wheeze and right lower zone crackles. The graphical properties of each drawing were examined and the use of written captions and labels recorded. Drawings were classified into styles according to the use of symbols (defined as discrete characters or icons) and shading (cross-hatching, speckling or darkening) to depict the auscultation findings. The study was conducted between September and November 2011. Results: Sixty-nine hospital doctors and 13 GPs participated. Ten drawing styles were identified from 78 completed drawings. Ten distinct symbols and a range of shading techniques were used. The most frequent style (21% of drawings) combined 'X' symbols representing crackles with musical notes for wheeze. There was inconsistency of representation across the drawings. Forty-seven (60%) drawings used an 'X' symbol exclusively to represent crackles, but six (8%) used 'X' only to represent wheeze, and 10 (13%) used 'X' to represent both findings. 91% of participants included captions or labels with their drawing. Conclusions: There was wide variation in doctors' drawings of identical auscultation findings, and inconsistency in the meaning of symbols both between and within drawings. Doctors risk incorrectly interpreting each other's drawings when they are not effectively labelled. We recommend doctors consider using a written description instead, or draw different findings with distinct symbols or shading, labelling all findings clearly.