such structure is depicted in the upper panel of Fig. 4b. It compares very well qualitatively with the model result shown in the lower panel of Fig. 4b. When domain edges remain fluxional following heterogeneous evaporation, the networks described above for Fig. 4b are not stable, long-lived structures. Nanoparticles continue to move in this case, strongly biased by the interfacial tension of cell boundaries. Cells break up as diffusion concentrates nanoparticle density at the nodes of the network, leaving distinct, worm-like domains. An example of such a pattern generated by our simulations (lower panel of Fig. 4d) compares well with worm-like morphologies observed in experi- ments (upper panel of Fig. 4d). These structures are themselves transitory, because their anisotropy costs significant interfacial free energy. Domains thus eventually become disks, which diffuse and coalesce, as described in the case of homogeneous evaporation. This mechanism of network disintegration strongly resembles that observed in viscoelastic phase separation of a dynamically asymmetric mixture 2 . Our results suggest four basic regimes of drying-mediated nano- particle assembly. They are distinguished by the spatial uniformity of solvent dynamics, and by the fluctuations of nanoparticle domain boundaries following evaporation. When solvent disap- pears homogeneously from the surface, disk-like or ribbon-like domains reminiscent of spinodal decomposition form at early times. We have shown that if these aggregates remain fluxional, they continue to evolve in a self-similar fashion, principally by diffusion and coalescence. If instead domain boundaries are frozen following evaporation, dynamical constraints arrest this growth at an early stage. When evaporation is inhomogeneous owing to infrequent nucleation events, network structures are formed at early times as vapour nuclei meet. These cellular patterns are only stable if interfaces are frozen following evaporation. Otherwise, networks fragment to form distinct domains that asymptotically evolve as in homogeneous coarsening. Many aspects of these self-assembly mechanisms have previously been rationalized on the basis of seemingly distinct physical pic- tures. Our mesoscopic model unifies these pictures and provides a quantitative measure of their importance under different con- ditions. As such, it may serve as a basic guide for designing self- assembled structures with desired nanometre-scale features. Con- sidering the simplicity of its energetics and dynamical rules, the degree of local order and anisotropy out of equilibrium is remark- able. We have omitted many physical details, most notably hydro- dynamic convection, substrate roughness, non-local interactions, and film thickness. These effects can be added to provide micro- scopic realism, but it is interesting that they are not needed to account for the variety of patterns that have been observed in experiments. A Received 9 May; accepted 25 September 2003; doi:10.1038/nature02087. 1. Bray, A. J. Theory of phase-ordering kinetics. Adv. Phys. 43, 357–459 (1994). 2. Tanaka, H. Viscoelastic phase separation. J. Phys. Condens. Matter 12, R207–R264 (2000). 3. Ge, G. & Brus, L. E. Evidence for spinodal phase in two-dimensional nanocrystal self-assembly. J. Phys. Chem. B 104, 9573–9575 (2000). 4. Tang, J., Ge, G. & Brus, L. E. Gas-liquid-solid phase transition model for two-dimensional nanocrystal self-assembly on graphite. J. Phys. Chem. B 106, 5653–5658 (2002). 5. Puntes, V. F., Krishnan, K. M. & Alivisatos, A. P. Colloidal nanocrystal shape and size control: The case of cobalt. Science 291, 2115–2117 (2001). 6. Gelbart, W. M., Sear, R. P., Heath, J. R. & Chaney, S. Array formation in nano-colloids: Theory and experiment in 2D. Farad. Disc. 112, 299–307 (1999). 7. Whitesides, G. M. & Grzybowski, B. Self-assembly at all scales. Science 295, 2418–2421 (2002). 8. Murray, C. B., Kagan, C. R. & Bawendi, M. G. Self-organization of CdSe nanocrystallites into 3-dimensional quantum-dot superlattices. Science 270, 1335–1338 (1995). 9. Freeman, R. G. et al. Self-assembled metal colloid monolayers—an approach to SERS substrates. Science 267, 1629–1632 (1995). 10. Andres, R. P. et al. Self-assembly of a two-dimensional superlattice of molecularly linked metal clusters. Science 273, 1690–1693 (1996). 11. Harfenist, S. A., Wang, Z. L., Alvarez, M. M., Vezmar, I. & Whetten, R. L. Highly oriented molecular Ag nanocrystal arrays. J. Phys. Chem. 100, 13904–13910 (1996). 12. Sear, R. P.,Chung, S. W., Markovich, G., Gelbart, W. M. & Heath, J. R. Spontaneous patterning of quantum dots at the air-water interface. Phys. Rev. E 59, R6255–R6258 (1999). 13. Fried, T., Shemer, G. & Markovich, G. Ordered two-dimensional arrays of ferrite nanoparticles. Adv. Mater. 13, 1158–1161 (2001). 14. Redl, F. X., Cho, K. S., Murray, C. B. & O’Brien, S. Three-dimensional binary superlattices of magnetic nanocrystals and semiconductor quantum dots. Nature 423, 968–971 (2003). 15. Elbaum, M. & Lipson, S. G. How does a thin wetted film dry up? Phys. Rev. Lett. 72, 3562–3565 (1994). 16. Chandler, D. Introduction to Modern Statistical Mechanics (Oxford Univ. Press, New York, 1987). 17. Ge, G. & Brus, L. E. Fast surface diffusion of large disk-shaped nanocrystal aggregates. Nano Lett. 1, 219–222 (2001). 18. Lo, A. & Skoodje, R. T. Kinetic and Monte Carlo models of thin film coarsening: Cross over from diffusion-coalescence to Ostwald growth modes. J. Chem. Phys. 112, 1966–1974 (2000). 19. Maillard,M., Motte, L., Ngo, A. T. & Pileni, M. P. Rings and hexagons made of nanocrystals: A Marangoni effect. J. Phys. Chem. B 104, 11871–11877 (2000). 20. Witten, T. A. & Sander, L. M. Diffusion-limited aggregation, a kinetic critical phenomenon. Phys. Rev. Lett. 47, 1400–1403 (1981). 21. Stowell, C. & Korgel, B. A. Self-assembled honeycomb networks of gold nanocrystals. Nano Lett. 1, 595–600 (2001). Supplementary Information accompanies the paper on www.nature.com/nature. Acknowledgements This work was supported by the United States–Israel Binational Science Foundation. L.E.B. is supported by the Columbia MRSEC. P.L.G. was an MIT Science Fellow throughout most of this work. D.R.R. is a Sloan Fellow and Camille Dreyfus Teacher-Scholar. Competing interests statement The authors declare that they have no competing financial interests. Correspondence and requests for materials should be addressed to E.R. (rabani@tau.ac.il) or D.R.R. (reichman@chemistry.harvard.edu). .............................................................. Proxy evidence for an El Nin ˜ o-like response to volcanic forcing J. Brad Adams 1 , Michael E. Mann 1 & Caspar M. Ammann 2 1 Department of Environmental Sciences, University of Virginia, Clark Hall, Charlottesville, Virginia 22903, USA 2 Climate Global Dynamics Division, National Center for Atmospheric Research, 1850 Table Mesa Drive, Boulder, Colorado 80307-3000, USA ............................................................................................................................................................................. Past studies have suggested a statistical connection between explo- sive volcanic eruptions and subsequent El Nin ˜o climate events 1,2 . This connection, however, has remained controversial 3–5 . Here we present support for a response of the El Nin ˜o/Southern Oscillation (ENSO) phenomenon 6,7 to forcing from explosive volcanism by using two different palaeoclimate reconstructions of El Nin ˜o activity 8,9 and two independent, proxy-based chro- nologies of explosive volcanic activity 5 from AD 1649 to the present. We demonstrate a significant, multi-year, El Nin ˜o-like response to explosive tropical volcanic forcing over the past several centuries. The results imply roughly a doubling of the probability of an El Nin ˜o event occurring in the winter following a volcanic eruption. Our empirical findings shed light on how the tropical Pacific ocean–atmosphere system may respond to exogenous (both natural and anthropogenic) radiative forcing. Coupled ocean–atmosphere experiments have explored the pos- sible response of ENSO to enhanced greenhouse gas concen- trations 10–16 . Results indicate El Nin ˜o-like 11–14 , neutral 15 and even La Nin ˜a-like 16 responses of average conditions (even a ‘neutral’ response represents a La Nin ˜a-like anomaly in the face of large-scale greenhouse warming). Simulations employing the Cane–Zebiak model of tropical Pacific coupled ocean–atmosphere dynamics, which exhibits a stronger dynamical feedback than most global models, produces negative (positive) eastern tropical Pacific sea surface temperature (SST) anomalies in response to a positive (negative) surface radiative forcing 17 . This imposes a La Nin ˜a-like cooling of the mean state in the presence of positive greenhouse warming 10 . It is intriguing in this context to reconsider the con- letters to nature NATURE | VOL 426 | 20 NOVEMBER 2003 | www.nature.com/nature 274 © 2003 Nature Publishing Group