The northernmost onshore surface faulting on the San Andreas fault in 1906 was reported at Shelter Cove, a small embayment in the coastline east of Point Delgada. Between 1979 and 1985, several authors questioned both this location for the San Andreas fault and the validity of 1906 surface faulting. Alternatively, they proposed an offshore fault and a landslide origin for the 1906 ground fractures.
F. E. Matthes (1874-1948) described and mapped the surface deformation at Shelter Cove in May and June 1906, documenting evidence of surface faulting for about 3 km between Dead Mans Gulch and Kaluna Cliff. Detailed sketch maps in his notebook indicate a component of right-lateral strike slip and a consistent fault dip to the southwest. Geomorphic and geologic evidence confirm a right-oblique sense of slip and a southwest dip of about 45°; the projected trace of the fault meets the sea at a point about 1.5 km northwest of Kaluna Cliff. Aligned scarps, saddles, and deflected stream courses along this fault imply repeated, nearly colinear, fault displacements during late Quaternary time. Published observations of shear strain and fission-track ages show that a narrow (<1 km) zone containing the mapped fault trace is slipping at depth in a right-lateral sense and is also a profound tectonic boundary separating rocks with different thermal histories. Collectively, the evidence supports both the reported faulting in 1906 and the interpretation that this fault is a northern extension of the San Andreas fault.