brightness of the ultraviolet Lyman-Ȋ emission (based on data from comets that have been observed at both wavelengths). Orbital calculations 1,6 suggest that this comet was brighter than the brightest comet discovered by the professional surveys for at least six months. Why was it not seen at visible wavelengths? Does this imply that there are many other comparable comets just waiting to be discovered? And, if so, should we care? We care about the efficiency of discover- ing comets for many reasons. The most obvi- ous scientific reason is that we cannot find out where comets come from and how big they are unless we can reduce the biases in our observations. These biases are due to selective discovery in certain parts of the sky and to varying orbital and physical charac- teristics of the comets. For our peace of mind, it is important to know whether comets represent 10% of the potential large impacts on Earth (as is commonly thought) or a much larger fraction. Many people argue that comets dominate at the largest sizes and that the observations set only a lower limit to this fraction. It has long been recognized that selection effects influence the discovery of comets, but the magnitude of these effects is not well understood. A few years ago, Brandt et al. 7,8 proposed a systematic search for small (100-metre) comets using a technique similar to SWAN. But instead of studying emission from hydrogen they suggested using emission from the OH radical (the other fragment produced when cometary water molecules are broken apart by sun- light) as a characteristic marker of cometary activity. Their proposal suffered from inade- quate satellite time, and no comets were found. The comet found by Mäkinen et al. is a prototypical example of something missed by other people because of the strong selec- tion effects in those other searches. A system- atic search for comets in the inner Solar System, using the known characteristics of comets and searching the entire sky with high frequency, would be extremely useful. Although observations of emission by the OH radical would have less of a problem with background noise, a systematic search by SWAN using longer integration times would almost certainly turn up additional comets. It would add tremendously to our knowl- edge of the physical and orbital properties of comets and would provide an insight into the real statistical frequency of cometary impacts on Earth. ■ Michael F. A’Hearn is in the Department of Astronomy, University of Maryland, College Park, Maryland 20742-2421, USA. e-mail: ma@astro.umd.edu 1. Mäkinen, J. T. T. et al. Nature 405, 321–322 (2000). 2. Gehrels, T. (ed.) Hazards due to Comets and Asteroids (Univ. Arizona Press, Tucson, 1994). 3. Grieve, R. A. F. & Shoemaker, E. M. in Hazards due to Comets news and views NATURE | VOL 405 | 18 MAY 2000 | www.nature.com 287 B oth benign and malignant tumours grow in an uncontrolled way. But it is only cells of malignant tumours that invade surrounding tissues and travel to distant organs (metastasize). Conventional wisdom used to hold that invasion and metastasis are late events — often ‘too late’ — in the clinical course of a patient’s cancer. However, we now know that invasion can be both early and clinically ‘silent’. An under- standing of the molecular basis for this aggressiveness could lead to therapies that block the transition of a tumour from benign to malignant, and keep local disease in check. Taguchi and colleagues 1 , writing on page 354 of this issue, have now identified proteins called RAGE and amphoterin as a receptor–ligand pair in a molecular check- point that regulates not only the invasiveness but also the growth and movement of tumour cells — the trio of characteristics required for malignancy. The threat of tumour invasiveness is exemplified by the fact that brain cancer does not need to metastasize to kill a patient. The growth of a brain tumour mass in the Cancer Checkpoint for invasion Lance A. Liotta and Timothy Clair and Asteroids (ed. Gehrels, T.) 417–462 (Univ. Arizona Press, Tucson, 1994). 4. Bailey, M. E. et al. in Hazards due to Comets and Asteroids (ed. Gehrels, T.) 479–533 (Univ. Arizona Press, Tucson, 1994). 5. Marsden, B. G. & Williams, G. V. Catalogue of Cometary Orbits 1999 13th edn (Minor Planet Center, Cambridge, MA, 1999). 6. Marsden, B. G. Minor Planet Electronic Circular 1999-X07 Pseudopod Tumour cell Extracellular matrix Detach Attach Pseudopod protrusion Motility Invasion Proteolysis Growth RAGE Amphoterin Actin mobilization p42/p44 JNK p38 Intercellular adhesion molecules Cell–matrix adhesion molecules Actin filaments Figure 1 Spatial and temporal regulation of cellular invasion of the extracellular matrix. Invasion can be normal (for example, in the case of the outgrowth of neuronal protrusions during brain development) or malignant (for example, in the case of tumours). Top, invasion can be viewed as cellular motility coupled to regulated adhesion and detachment (from the extracellular matrix) and proteolysis (of extracellular matrix molecules). The advance of pseudopods (extensions) of the cell — driven by the formation of actin polymers, a component of the cytoskeleton — may require the action of cell-surface protein-degrading enzymes, as well as other enzymes, receptors and activators. Extracellular matrix degradation must be balanced by antiproteolysis (that is, processes that inhibit proteolysing enzymes) to allow for adhesive traction. Bottom, signal-transduction pathways allow the individual cell to move between phases of pseudopod protrusion, extracellular matrix degradation, antiproteolysis, adhesion and detachment. These pathways split at the level of the mitogen-activating protein kinases JNK, p38 and p42/p44. Blocking the interaction between amphoterin and RAGE suppresses these pathways, as shown by Taguchi et al. 1 . (Minor Planet Center, Cambridge, MA, 1999). 7. Brandt, J. C. et al. Earth, Moon and Planets 72, 243–249 (1996). 8. Brandt, J. C. et al. in Completing the Inventory of the Solar System (eds Rettig, T. W. & Hahn, J. M.) 289–297 (Astron. Soc. Pacific, San Francisco, 1996). 9. http://www.geo.fmi.fi/PLASMA/SWAN © 2000 Macmillan Magazines Ltd