SCIENTIFIC CORRESPONDENCE
able of spawning such an SL field. Light from seismic waves Reports of EQLs for large nocturnal earthquakes are not ubiquitous, but SIR - Sonoluminescence (SL) - the purities can appreciably alter the basic neither are they extremely rare. Selected 6 production of light by the action of aqueous spectrum , so that yellow or red characteristics relevant to the SL hypo sound waves in liquid - has been may predominate. thesis that appear in the EQL literature observed and studied in the laboratory SL has been generated in the labora are: (1) distinct blue to bluish-white for more than 50 years. But I believe it tory with ultrasonic pressure amplitudes EQLs reported from coastal Japan10 and 4 2 has been observed in nature for centur of 1-2 bar (0.1-0.2 MPa) , correspond Hawaii ; (2) extensive EQLs from an 11 ies as earthquake lights (EQLs). EQLs ing roughly to an energy density in the onshore alluvial setting ; (3) numerous 3 are largely a coseismic occurrence, ambient fluid of 10-20 erg cm- . accounts of EQLs si¥hted offshore from 1 1 second only to pre-seismic abnormal Whether SL is a viable mechanism for California , Mexico and other coastal animal behaviour for difficulty of reli EQLs hinges on the question of whether zones; and (4) well-defined, blue and able documentation and lack of a verifi it is reasonable that P waves of sub yellow spherical lights in tsunami 12 able explanatory mechanism. audible frequency can supply this energy wavecrests • (Luminescent organisms Many explanations for the generation density in water. may be another source of luminescence 13 of EQLs exist (see refs 1-3). Most prop The density of the kinetic energy, e in tsunami wavecrests , but I do not osed mechanisms require assumptions (per unit volume), induced in the trans believe that they would create well such as the presence of special minerals, mission medium by one cycle of an defined spheres of light.) gases or organisms, or unverified physic advancing seismic wavefront is a stan Natural SL is perhaps not the only al conditions in the fault zone. All have dard result in seismology7 and is given mechanism that produces EQLs, but it is difficulty explaining the persistent re by able to explain a wide range of the ports of EQLs at distances of up to existing reports. The hypothesis predicts e = 2n2p 2 hundreds of kilometres from the earth (A~-r0) that (1) EQLs should not be confined to quake, and at sea or in association with where p is the density of the medium the immediate fault rupture zone; (2) large bodies of both fresh and salt water and A 0 and to are the displacement bodies of water must be present, or wetlands. amplitude and period respectively of the although it is possible that saturated soil I propose that at least some EQLs are seismic wave. Conservatively estimated can sustain SL; (3) EQLs are essentially not the product of high strain accumula values for A 0 of 1-10 em and to of a coseismic phenomenon, which in the tion or shear rupture dynamics in fault 0.1-1.0 s may be obtained from the absence of strong foreshocks or after zones, but rather result from molecular strong ground motion recordings of shocks should not be observed before or reactions in water that has been strongly earthquakes. This yields P-wave energy more than several minutes after the shaken by the compressional (P) waves densities in water of roughly 500-2,000 earthquake; and (4) the EQL spectrum produced by the earthquake. A seismic erg cm-3 at 10-1 Hz and pressure differ should contain a prominent hydroxyl P wave is simply an earthquake entials of 1.3-2.7 bar. Thus, seismic P peak at 310 nm (and for sea water6 a generated sound wave in a solid or waves are capable of supplying pressure sodium peak at 589 nm). Hence the liquid; hence if a P wave induces light changes and energy densities that exceed SL-EQL hypothesis can be tested by emission from liquid, it is a situation the laboratory values that induce SL. means of spectrographic analysis, entirely analogous to SL as generated in Within the water volume irradiated by P although obtaining an EQL spectrum the chemist's laboratory. waves, EQLs would arise as the inte will not be a trivial undertaking. SL is a remarkable consequence of grated light flux from many SL cavita ARCH C. JOHNSTON acoustic cavitation in liquids irradiated tion bursts, all loosely synchronized by Center for Earthquake Research, 4 5 by sound waves • • For it to occur, a the P-wave dilational half-cycles. Memphis State University, cavity or bubble must be created in the In the laboratory, SL produces an Memphis, Tennessee 38152, USA liquid continuum, then rapidly compress illuminance of - 10-8 lumen cm-2 (ref. 1. Derr. J. Bull. seis. Soc. Am. 63, 2177-2187 (1973). ed. The process adiabatically heats the 8), which is visible to the dark-adapted 2. Lockner, D. A., Johnston, M. J. S. & Byerlee, J. D. trapped gas or vapour sufficiently to eye. Thus, to reproduce an illuminance Nature 302, 28-33 (1983). dissociate molecules. On recombination equivalent to moonlight of -10-4 lumen 3. Brady, B. T. & Rowell , G. A. Nature 321, 488-492 2 9 (1986). or return to the ground state, photons cm- , as reported for EQLs in Japan , 4 . Barber. B. P. & Putterman. S. J. Nature 352, 318-320 are emitted. Once a cavity or bubble is -104 ultrasonic SL bursts are required in (1991). 5 . Suslick. K. S. Science 247, 1439-1445 (1990). formed, two types of SL are possible: the laboratory. For the much larger 6 . Suslick. K. S. & Flint. E. B. J. Phys. Chern. 95, 1484 'stable', in which the bubble resonates P-wave cavitation events (with bubble (1990). 7. Kasahara. K. Earthquake Mechanics (Cambridge Un i· and incrementally grows, usually in a radius in the centimetre rather than versity Press. 1981). standing wave field; and 'transient', in micrometre range), the same illuminance 8. Walton, A. J. & Reynolds, G. T. Adv. Phys. 33, 595-660 which the bubble expands and implodes could arise from a single event. To (1984). 9. Musya. K. Bull. Earthquake Res. lnst. Tokyo Univ. 9, all within one cycle of a standing or illuminate a landscape to moonlight 214-215 (1931). travelling sound wave. A travelling P brightness from at least several 10. Terada , T. Bull. Earthquake Res. lnst. Tokyo Univ. 9, 2 4 225-255 (1931). wave should be an efficient stimulus of kilometres distance, 10 -10 individual 11. Fuller, M. L. U.S. Geol. Surv. Bull. 494, 120 (1912). transient cavitation, although long trains SL P-wave bursts would be sufficient. A 12. Musya, K. Bull. Earthquake Res. lnst. Tokyo Univ. 10, 666-673 (1932). or strong-motion P waves may induce P-wave with a dilational half-cycle 13. Terada . T. Bull. Earthquake Res. lnst. Tokyo Univ. 1 , stable cavitation SL as well. wavelength of - 1 km is certainly cap- 25-35 (1934). The observed SL spectrum in water has a peak at 310 nm (in the ultraviolet), Correction arising from the return to the ground IN the letter from M. Allard eta/. (Nature 353, 610: 1991) "Tests for rodent polyphyly", state of the excited hydroxyl radical the lettering on the figure, part a. was incorrect. The correct figure is shown below. o (OH); there is also a poorly understood a continuum throughout the visible wave CLA PPY CLA >----< MDO CLA >----< MDO band. A pure water SL spectrum will appear blue to bluish-white, but the MDO>----< CHI CHI PPY PPY CHI presence of dissolved salts or other im- I 17(11) II 10(5) Ill 6(3) NATURE · VOL 354 · 5 DECEMBER 1991 361 © 1991 Nature Publishing Group