The dominance of Mustard's operation for transposition of the great arteries has been challenged by the recent revival of Senning's repair because it promises better long-term results in terms of venous obstruction and atrial haemodynamics. These hypotheses were tested by recording jugular venous flow waveforms transcutaneously in 24 postoperative patients with simple complete transposition using a bidirectional Doppler blood velocimeter. Eight patients had undergone Mustard's operation and 16 the Senning alternative; all had previously had a postoperative cardiac catheterisation. Both groups of patients had similar left ventricular, pulmonary arterial, and systemic venous atrial pressures. No child showed any evidence at catheterisation of either mitral regurgitation or of superior vena caval pathway obstruction. These two findings were endorsed by the transcutaneous Doppler recordings. Jugular venous flow in normal children exhibits two maxima, one of atrial filling during ventricular systole, the other of ventricular filling occurs once the tricuspid valve has opened. Both operative procedures diminished the size of the former phase, but the Mustard did so more. After Mustard's operation forward flow during the atrial filling phase was absent in approximately half the cardiac cycles recorded, and severely diminished in the rest. By contrast, there was approximately a 90 per cent appearance of atrial filling waves after Senning's operation which also provided significantly better atrial function than Mustard's procedure in terms of peak velocity of blood entering the atrium and total atrial filling. It is therefore concluded that both procedures compromise atrial volume and compliance but Senning's repair to a much lesser extent.