Rows of small vents, cones, domes and collapse pits are common in volcanic terranes. These features are thought to form above feeder dikes. Although dike orientation can be influenced by the orientation of pre-existing faults and other structures, in general the strongest control on dike orientation is the country rock stress field, with intrusion perpendicular to the least principal stress (Nakamura, 1977; Delaney et al., 1986). The surface vents are therefore generally aligned parallel to the local maximum horizontal stress direction (SHmax). The rift-parallel volcanic vent and fissure alignments in the vicinity of the axial volcanoes suggest that during most of the Pleistocene, the minimum horizontal stress direction (Shmin) was oriented approximately E-W. For example, small cones and domes on the northern flank of Suswa that pre-date caldera collapse are generally aligned parallel to N7°W (Fig. 6). A large number of pyroclastic and lava cone alignments are present in the region of Silali, with an average orientation of N10°E (Fig. 4b). Fault kinematic data from the central rift in the area of the Kinangop Plateau (east of Menengai, Fig. 1) support the volcano-tectonic interpretation of E-W extension during the Late Pliocene to Early Pleistocene (Strecker et al., 1990).

Borehole breakout data are available for several hydrocarbon exploration wells in eastern Kenya and their interpretation suggest that the present-day Shmin is aligned NW-SE (Bosworth et al., 1992). Vents in the large volcanic shields east of the rift valley at the Huri Hills, Mt. Marsabit, and Nyambeni Hills (Fig. 1) are predominantly aligned NE-SW, also indicating NW-SE Shmin. Uppermost basalts at Mt. Marsabit and the Huri shield have been dated ~500 ka and younger (Brotzu et al., 1984; Key, 1987; Charsley, 1987), demonstrating that this NW-SE Shmin is a very young stress field orientation. In the region of the eastern border fault of the central Kenya rift, faulting dated at less than ~500 ka yields a similar NW-SE extension direction (Strecker et al., 1990).

At Suswa, small cones in the Ol Doinyo Nyukie phonolites (circa 100 ka; see above), which post-date caldera collapse, are aligned N60°E (Fig. 6). East of Silali caldera, in the Black Hills, very young trachyte lavas (10±2 ka and 4±2 ka; Ar-Ar; Dunkley et al., 1993) display two fissure and cone alignment trends, one N-S to N20°E and the other N50°E (Fig. 4b). The presence of N50°E to N60°E volcanic alignments in the youngest units of the rift axial volcanoes supports the interpretation that a new stress field has manifest itself within the rift valley axis. This places the WNW-ESE to NW-SE trending long axes of the northern axial calderas sub-parallel to Shmin of this new stress field. As argued by Dunkley et al. (1993), large magma chambers probably existed beneath some or all of the volcano summits prior to caldera collapse. Subjected to non-hydrostatic stress, these chambers would experience variation in shear stress around their walls, and might behave analogously to a tunnel or well-bore under similar circumstances. As discussed below, pieces of the chamber wall are envisioned to spall off to produce a large-scale "borehole breakout" and elongate the chamber in a direction parallel to Shmin. This interpretation would indicate that the NW-SE Shmin was in effect along the rift axis during the past 100 ka, the period of collapse of the northern volcanoes.