9. M. G. Leakey, C. S. Feibel, I. McDougall, A. C. Walker, Nature 376, 565 (1995). 10. T. D. White, G. Suwa, B. Asfaw, ibid. 371, 306 (1994); ibid. 375, 88 (1995). 11. A. T. Chamberlain and B. A. Wood, J. Hum. Evol. 16, 119 (1987); R. R. Skelton and H. M. McHenry, ibid. 23, 309 (1992); D. S. Strait, F. E. Grine, M. A. Moniz, ibid. 32, 17 (1997); R. R. Skelton and H. M. McHenry, ibid. 34, 109 (1997). 12. F. E. Grine, Ed., Evolutionary History of “Robust” Aus- tralopithecines (de Gruyter, New York, 1988); A. C. Walker, R. E. Leakey, J. M. Harris, F. H. Brown, Nature 322, 517 (1986); H. M. McHenry, in Contemporary Issues in Human Evolution, W. E. Meikle, F. C. Howell, N. G. Jablonski, Eds. (California Academy of Sciences, San Francisco, 1996), pp. 77–92. 13. W. H. Kimbel, T. D. White, D. C. Johanson, Am. J. Phys. Anthropol. 64, 337 (1984); G. Suwa et al., Nature 389, 489 (1997); P. V. Tobias, Olduvai Gorge, Volume 4: The Skulls, Endocasts and Teeth of Homo habilis (Cambridge Univ. Press, Cambridge, 1991). 14. W. H. Kimbel, D. C. Johanson, Y. Rak, Nature 368, 449 (1994). 15. A. C. Walker and R. E. Leakey, Eds., The Nariokotome Homo erectus Skeleton (Harvard Univ. Press, Cam- bridge, MA, 1993). 16. Femur and humerus length were virtually complete in A.L. 288-1 [D. C. Johanson et al., Am. J. Phys. An- thropol. 57, 403 (1982)]. In BOU-VP-12/1, the femur is preserved from the intersection of the medial terminus of the neck with the (missing) femoral head (proximally) to a point on the medial supracondylar line just superior to the gastrocnemius impression (distally). This distance was measured in a sex- and species-balanced sample of Pan, Gorilla, and Homo (N = 60) and used to regress (least squares) femoral length [correlation coefficient (r 2 ) = 0.952; 95% confidence interval of estimate =0.28]. This re- gression computes the BOU-VP-12/1 femur at 348 mm. On anatomical grounds, we believe it to have actually been slightly shorter (about 335 mm). Much of the shaft of the BOU-12/1 humerus is preserved, including the point of confluence between the diaph- ysis and the medial epicondylar apophysis and the distalmost extent of the deltopectoral crest. This distance was used to regress humeral length with the same sample (length estimate = 226 mm; r 2 = 0.876; 95% confidence interval of estimate = 0.40). On anatomical grounds, we estimate the humerus to have been slightly longer (about 236 mm). Radial length was estimated for A.L. 288-1 with multiple linear regressions from the same sample (breadth distal articular surface; maximum diameter radial head; length radial neck; r 2 = 0.929; 95% confi- dence interval of estimate =0.29) and for BOU-VP- 12/1 (radial head to nutrient foramen; maximum diam- eter radial head; length radial neck; r 2 = 0.937; 95% confidence interval of estimate =0.27). These regres- sions estimate a length of 203 mm for A.L. 288-1 and 231 mm for BOU-VP-12/1. On anatomical grounds, the BOU-VP-12/1 estimate appears correct. However, we believe that the A.L. 288-1 radius is underestimated on the basis of a lack of sufficient “anatomical space” with which to accommodate all of the preserved pieces of the bone. A regression limited to a sample of common chimpanzees and bonobos (N = 36) estimates a length of 215 mm (r 2 = 0.529; 95% confidence interval of estimate =0.36). This result appears more probable. Only exceptionally pronounced errors in any of the above predictions would alter the conclusions made in the legend of Fig. 3, nor are these conclusions altered by regressions based only on single hominoid species. 17. The Middle Awash paleoanthropological project is mul- tinational (13 countries), interdisciplinary research co- directed by B.A., Y. Beyene, J. D. Clark, T.W., and G. WoldeGabriel. The research reported here was support- ed by the NSF. We thank N. Tahiro and A. Abdo for their assistance in naming the new species. We thank Y. Haile-Selassie for discovery of the BOU-VP-12/130 ho- lotype and H. Gilbert and D. DeGusta for field and illustrations work. R. Holloway kindly allowed us to cite his BOU-VP-12/130 cranial capacity estimate. L. Gudz made the palate and postcranial drawings. D. Brill made the photographs. P. Reno provided comparative primate data. The Japan Ministry of Education, Science, Sports and Culture provided support to G.S. We thank the Ethiopian Ministry of Information and Culture, the Cen- tre for Research and Conservation of the Cultural Her- itage, and the National Museum of Ethiopia. We thank the Afar Regional Government and the Afar people of the Middle Awash for permission and support. We thank the many individuals who contributed to the camp, transport, survey, excavation, and laboratory work that stands behind the results presented. 25 February 1999; accepted 30 March 1999 Electron Solvation in Finite Systems: Femtosecond Dynamics of Iodide( Water) n Anion Clusters L. Lehr,*† M. T. Zanni,* C. Frischkorn, R. Weinkauf,† D. M. Neumark‡ Electron solvation dynamics in photoexcited anion clusters of I - (D 2 O) n=4–6 and I - (H 2 O) 4–6 were probed by using femtosecond photoelectron spectros- copy (FPES). An ultrafast pump pulse excited the anion to the cluster analog of the charge-transfer-to-solvent state seen for I - in aqueous solution. Evolution of this state was monitored by time-resolved photoelectron spectroscopy using an ultrafast probe pulse. The excited n = 4 clusters showed simple population decay, but in the n = 5 and 6 clusters the solvent molecules rearranged to stabilize and localize the excess electron, showing characteristics associated with electron solvation dynamics in bulk water. Comparison of the FPES of I - (D 2 O) n with I - (H 2 O) n indicates more rapid solvation in the H 2 O clusters. A free electron can be trapped by solvent reori- entation in polar solvents such as ammonia (1) or water (2). These “solvated” electrons play an important role in condensed phase chemistry, including radiation chemistry, electron transfer, and charge-induced reactivity. A microscopic understanding of the electron-solvent and sol- vent-solvent interactions that govern electron solvation is therefore a fundamental and chal- lenging problem. These considerations have motivated femtosecond time-resolved studies that have demonstrated rich and complex dy- namics after electronic excitation of electrons in water (3, 4 ). To gain a complementary perspective on this problem, we studied solvated electron dynamics in finite clusters and compared these results with our understanding of bulk solvation phenomena. We used two-photon anion femtosecond (10 -15 s) photoelectron spectroscopy (FPES) (5, 6 ) to study electron solvation dynamics in the mass-selected an- ion clusters I - (D 2 O) n and I - (H 2 O) n in order to address the following questions: (i) What is a minimum solvent cluster size needed to solvate an electron? (ii) What is a typical time scale for solvent reorientation in a cluster? (iii) What type of solvent motion is involved in electron solvation dynamics? Aqueous solutions of I - exhibit broad electronic bands in the ultraviolet (UV) cor- responding to electron ejection from I - into the solvent (7 ), known as “charge-transfer- to-solvent” (CTTS) states. Excitation of these states is an elegant means of generating sol- vated electrons, as was first demonstrated by Jortner (8). The dynamics of these states have been investigated by Eisenthal (9), Gaudeul (10), Bradforth (11), and their co-workers, all of whom excited the CTTS states to inject an electron into the water and then followed the subsequent electron solvation dynamics by femtosecond absorption spectroscopy. These experimental studies along with simulations by Sheu and Rossky (12) and Staib and Bor- gis (13) show that excitation of the lowest energy CTTS band results in the generation of fully solvated electrons on a 200-fs time scale (11). Once generated, these electrons thermalize with the solvent molecules, and some are then lost through geminate recom- bination with the neutral halogen atom over a time scale of tens of picoseconds. The issue of how the CTTS bands mani- fest themselves in finite clusters was first addressed in experiments by Johnson and co-workers (14 ), in which a diffuse absorp- tion band was seen just above the detachment Department of Chemistry, University of California, Berkeley, CA 94720, USA and Chemical Sciences Di- vision, Lawrence Berkeley National Laboratory, Berke- ley, CA 94720, USA. *These authors contributed equally to this work. †Permanent address: Institut fu ¨r Physikalische und Theo- retische Chemie, Technischen Universita ¨t Mu ¨nchen, D-85748 Garching, Germany. ‡To whom correspondence should be addressed. E- mail: dan@radon.cchem.berkeley.edu R EPORTS www.sciencemag.org SCIENCE VOL 284 23 APRIL 1999 635