Many of these studies also found evidence for additional discontinuities at other depths in the upper mantle. Both Revenaugh and Jordan [1991a] and Shearer [1991] consistently detected an apparent 520-km reflector in long-period waveforms. However, such a discontinuity was not seen in short-period refraction data examined by Jones et al. [1992], or in the short-period discontinuity reflections obtained by Benz and Vidale [1993], suggesting that the 520-km discontinuity may be characterized by a rather gradual change mostly in density, rather than a sharp velocity jump. Revenaugh and Jordan [1991b] found evidence in their ScS reverberation data for discontinuities at average depths of 60 km, 210--270 km, and 300 km. These features were not seen on all of the profiles and often varied greatly in depth, perhaps explaining why they did not appear on the global stacks of Shearer [1991].
At or near subduction zones, Vidale and Benz [1992] found evidence for a P impedance change near 210 km in their short-period stacks of deep earthquakes, while Zhang and Lay [1993] saw apparent sS precursors resulting from reflections near 80, 210 and 330 km for some of their events. Across Eurasia, Goldstein et al. [1992] showed that there must be at least some reflectors between 140 and 200 km depth to explain secondary arrivals on refraction data from Soviet nuclear tests. Shearer [1993] attempted a global survey of upper mantle reflectors in long-period SS precursor data, and found marginal evidence for regional reflectors at a variety of depths in the mantle above the transition zone. Possible explanations for observations of secondary reflectors have included phase changes, differences in anisotropy, and hydration reaction boundaries [ Revenaugh and Jordan, 1991b].