The extent of analgesia provided by transversus abdominis plane blocks depends upon the site of injection and pattern of spread within the plane. There are currently a number of ultrasound-guided approaches in use, including an anterior oblique-subcostal approach, a mid-axillary approach and a more recently proposed posterior approach. We wished to determine whether the site of injection of local anaesthetic into the transversus abdominis plane affects the spread of the local anaesthetic within that plane, by studying the spread of a local anaesthetic and contrast solution in four groups of volunteers. The first group underwent the classical landmark-based transversus abdominis plane block whereby two different volumes of injectate were studied: 0.3 ml.kg(-1) vs 0.6 ml.kg(-1). The second group underwent transversus abdominis plane block using the anterior subcostal approach. The third group underwent transversus abdominis plane block using the mid-axillary approach. The fourth group underwent transversus abdominis plane block using the posterior approach, in which local anaesthetic was deposited close to the antero-lateral border of the quadratus lumborum. All volunteers subsequently underwent magnetic resonance imaging at 1, 2 and 4 h following each block to determine the spread of local anaesthetic over time. The studies demonstrated that the anterior subcostal and mid-axillary ultrasound approaches resulted in a predominantly anterior spread of the contrast solution within the transversus abdominis plane and relatively little posterior spread. There was no spread to the paravertebral space with the anterior subcostal approach. The mid-axillary transversus abdominis plane block gave faint contrast enhancement in the paravertebral space at T12-L2. In contrast, the posterior approaches, using both landmark and ultrasound identifications, resulted in predominantly posterior spread of contrast around the quadratus lumborum to the paravertebral space from T5 to L1 vertebral levels. We concluded that the pattern of spread of local anaesthetic differs depending on the site of injection into the transversus abdominis plane. This may have important implications for the extent of analgesia produced with each approach.