This paper points out five anomalies of the Icelandic upper crust that are not compatible with currently accepted construction models. These anomalies are (1) assumption of a single time-averaged plate boundary, (2) irregular depth distribution to seismic layer 3, (3) the assumption of a uniform vertical compositional zonation, (4) the ''excess width'' of Iceland, and (5) abnormal variations in the duration and stratigraphic thickness of chron 5 time lavas in western, northern, and eastern Iceland. A model involving frequent shifts of the volcanic zones in Iceland is outlined that suggests a way to account for the above structural anomalies. The model also suggests that detailed studies of basalt lava facies types may be used effectively to trace small-scale shifts of the Tertiary plate boundary.