In most earthquakes the Earth's crust cracks like porcelain, Stress builds up until a fracture forms at a line depth of a few kilometers and the crust slips to relieve the stress. Some earthquakes, however, take place hun dreds of kilometers down in the Earth's mantle, where high pressure makes rock so ductile that it flows instead of cracking, even under stress severe enough to deform it like putty. How can there be earthquakes at such depths?

    That such deep events do occur has been accepted only since 1927. when the seismologist Kiyoo Wadati convincingly demonstrated their existence. Instead of comparing the arrival times of seismic waves at different locations, as earlier researchers had done, Wadati relied on a time difference between the arrival of primary(P) waves and the slower secondary(S) waves. Because P and S waves travel at different but fairly constant speeds, the interval between their arrivals increases in proportion to the distance from the earthquake focus, or initial rupture point.

    For most earthquakes, wadati discovered, the interval was quite short near the epicenter; the point on the surface where shaking is strongest. For a few events, however, the delay was long even at the epicenter. Wadati saw a similar pattern when he analyzed data on the intensity of shaking. Most earthquakes had a small area of intense shaking, which weakened rapidly with increasing distance from the epicenter. but others were characterized by a lower peak intensity, felt over a broader area. Both the P-S intervals and the intensity patterns suggested two kinds of earthquakes: the more common shallow events, in which the focus lay just under the epicenter, and deep events, with a focus several hundred kilometers down.

    The question remained: how can such quakes occur, given that mantle rock at a depth of more than 50 kilometers is too ductile to store enough stress to fracture? Wadati's work suggested that deep events occur in areas (now called Wadati-Benioff zones) where one crustal plate is forced under another and descends into the mantle. The descending rock is substantially cooler than the surrounding mantle and(60) hence is less ductile and much more liable to fracture.

    The term "episodic memory" was introduced by Tulving to refer to what he considered a uniquely human capacity-the ability to recollect specific past events, to travel back into the past in one's own mind-as distinct from the capacity simply to use information acquired through past experiences. Subsequently, Clayton et al. developed criteria to test for episodic memory in animals. According to these criteria, episodic memories are not of individual bits of information; they involvemultiple components of a single event "bound" together. Clayton sought to examine evidence of scrub jays' accurate memory of "what," "where," and "when" information and their binding of this information. In the wild, these birds store food for retrieval later during periods of food scarcity. Clayton's experiment required jays to remember the type, location, and freshness of stored food based on a unique learning event. Crickets were stored in one location and peanuts in another. Jays prefer crickets, but crickets degrade more quickly. Clayton's birds switched their preference from crickets to peanuts once the food had been stored for a certain length of time, showing that they retain information about the what, the where, and the when. Such experiments cannot, however, reveal whether the birds were reexperiencing the past when retrieving the information. Clayton acknowledged this by using the term "episodic-like" memory.