Magnetic anomalies over Iceland, measured by Serson et al. (1968), are similar in shape and amplitude to those found over mid-oceanic ridges in general and over Reykjanes Ridge in particular. However, the geology of Iceland does not favour the simple model of sea floor spreading as formulated by Vine and Matthews. The Brunhes period volcanism can neither in place nor in time be related to an opening process of the Central Graben, which actually is a downthrown block and not an opening rift. Furthermore, the structure of Iceland is not symmetric with respect to the Central Graben. The geology of the Central Graben of Iceland does support a model proposed by Thorleifur Einarsson in 1967. In this model elongate ridges of pillow lavas are thought to have piled up on top of parallel volcanic fissures. The actual spreading is negligible. The fissures have been opening at random over a width of about 120 km, and no definite time scale can be set up for the associated magnetic anomalies. This conflict between Icelandic geology and the current views on sea floor spreading, can be evaded by supposing that the mere circumstance that Iceland is an island obscures a spreading process underneath. One might also postulate that Iceland nevertheless should stand as an example of a mid-oceanic ridge which implies that our ideas on sea floor spreading should be thoroughly revised.