In 1990 a deep multichannel seismic reflection line was shot along a flow line across the Labrador Sea. The results from the central portion of this line between magnetic anomalies 25 and across the extinct central ridge are described here. Spreading rates in this part of the Labrador Sea are very low, from 10 mm/yr to less than 3 mm/yr, so that the line provides a unique opportunity to examine the relationship between very slow spreading and crustal structure. A clear division is observed between two types of crust. The oldest crust, between chrons 21 and 25, which formed at a mean half spreading rate of 10 mm/yr, exhibits smoothly undulating basement with only minor normal faulting. This region shares many of the reflection characteristics of North Atlantic crust formed at moderate to low spreading rates. In contrast, the younger central region, between chrons 21 and 13, which formed at a mean half spreading rate of 3 mm/yr, displays evidence of intense normal faulting of the crust, giving a total extension of about 70%. Inward facing normal faults on both sides of the extinct ridge, with large offsets, many of which extend to lower crust or Moho depths, dominate the seismic section. The axial region is characterized by a deep, fault-bounded, median valley.

These results suggest that mechanical extension plays a more important role in seafloor spreading at low spreading rates than previously documented. Integration of the reflection data with previous refraction measurements and with gravity modeling of the region shows variations in crustal thickness which can be correlated with spreading rates. The region formed at a mean spreading rate of 10 mm/yr, whre about 15% extension is observed, exhibits slightly thinner than normal crust (4.8 km or less versus a normal thickness of about 7 km). At a lower spreading rate of 3 mm/yr across the axial region where extension is about 70%, an average crustal thickness of 3 km is obtained. Thus lower spreading rates are associated with regions of thinner crust and greater amounts of extension. While many studies suggest that thin crust at slow spreading rates may result from a reduced magma supply, this study suggests that extension is at least equally important and may be responsible for most of the variations in crustal thickness. The increased cooling of young oceanic lithosphere formed at these very low spreading rates (approximately 3 mm/yr) may have amplified brittle failure in response to plate separation. However, the timing of extension is still uncertain and some of it may be related to postextinction tectonics and not to the spreading process. The role of extension as a control on crustal thickness need to be considered further in studies of crustal generation and magmatism at slow spreading ridges. Extension will decrease the importance of magmatism in generating thin oceanic crust and will favor models of magnetic processes which produce thicker crust. ¿ American Geophysical Union 1995