SCIENTIFIC CORRESPONDENCE and the function of the immune system. there are insufficient neutralizing anti­ They may also be dependent on antigen bodies? Therefore, it is not clear to us PRIMARY EDGE WAVEFRONT stimulation and T-cell activation, which what an average level of antibodies really may be more frequent in homosexuals in means. whom Kaposi's is more common. Thus, in our view, the titres cannot be (5) The presence or absence of anti-tat used to draw any meaningful conclusions antibodies or tat protein in the sera of regarding the role of tat or the anti-tat HIV-1-infected individuals obviously antibodies in vivo in the pathogenesis of does not exclude the role of tat in the Kaposi's. pathogenesis of Kaposi's because it could BARBARA ENSOLI be released and taken up by target cells in ROBERT C. GALLO close proximity or by cell-cell contact. Laboratory of Tumor Cell Biology, ( 6) Finally, if high levels of anti-tat National Cancer Institute, antibodies were found, Reiss et al. could National Institutes of Health, argue that high levels should be protective Bethesda, Maryiand 20892, USA by inactivating tat as might occur, for 1. Salahuddin. S.Z .. Nakamura. S. Biberfeld. P.. Kaplan. example, in polio. Therefore, tat would M.H .. Markham. P.D .. Larsson. L. & Gallo, R.C. Science 242. 430-433 (1988). have no role in Kaposi's. But antibodies 2. Vogel. J .. Hinrichs. S.H .. Reynolds. R K.. Luciw. P.A. &Jay. are not always neutralizing, and the mean­ G. Nature 335. 606-611 (1988). 3. Ensoli. B .. Barillari. G . Salahuddin. S.Z . Gallo. R.C. & ing of high levels can be just the opposite Wong-Staal. F. Nature 345. 84-86 (1990). -high levels of antibodies to HTLV-1 are 4. Goudsmit. J. eta/. Proc natn. Acad Sci U.S.A. 85, 4478- not protective (unless present before 4482 (1988). 5 Oevash, Y., Reagan, K., Wood, D., Turner, J., Parrington, infection). High levels, in fact, suggest the M. & Kang. C. Y. Nature 345. 581 (1990). In a non imaging concentrator designed by the presence of more virus. Conversely, what 6. Viscid1. R. P.. Mayur. K .. Lederman. H. M. & Frankel. A. D. edge- method (shown in a) all rays Science 246. 1606 (1989). do low antibodies mean? Would they indi­ 7. Ensoli. B .. Barillair. G .. Salahuddin. S.Z . Galio. R.C. & entering the device at the maximum collection cate that not much tat is present; or that Wong-Staal. F. Nature 345. 84 (1990). angle are directed after one reflection at most to the rim of the exit aperture. The edge-ray principle is slightly modified for the concen• Sunlight brighter than the Sun trator described here, to ensure that light is reflected by total internal reflection near the SIR-We have concentrated terrestrial could attain high concentrations of sun­ top. Hence, the exit angle of the light is sunlight by a factor of 84,000 to intensities light, but would be large and unwieldy. reduced slightly, from 90° to 86°. Panel b of 72 W mm-', establishing a new world We therefore use a two-stage system shows the two-stage approach to concentrat• record for solar flux concentration that incorporating a parabolic and a ing sunlight. exceeds our previous result by a factor of nonimaging concentrator. The overall 1.5 (refs 1 ,2). This concentration exceeds concentration is the product of the con­ In a series of trials over several cold, the intensity of light at the surface of the centrations of the parabolic mirror and of crisp days in January and February we Sun itself, which is 63 W mm-': we have the nonimaging concentrator: measured an average power per area at thus produced the highest concentration the exit aperture of the sapphire concen­ of sunlight in the Solar System. C = n'cos'rj>!sin'fJ (2) trator of 72 W mm-'. This intensity is We use a device designed according to 84,000 times greater than the direct inci­ the principles of nonimaging '. A which falls short of the theoretical limit in dent intensity, 0.86 mW mm·'. nonimaging concentrator is essentially a equation (1) only by the factor of cos'rj>, Potential uses for such high levels of 'funnel' for light, and can concentrate where rp is the rim angle of the mirror. solar flux are only now being explored. In light to intensities four times greater than Using such a technique, we previously the past it has been difficult to make effici­ those obtained with conventional imaging achieved a concentration of sunlight of ent solar pumped lasers because of the devices. (Aberrations in image-forming 56,000 times the intensity at the surface of limited materials available and the lower devices limit their ability to concentrate the Earth'. Our modified apparatus levels of concentration of sunlight that light.) A laser crystal or a solar furnace reaches still. higher concentrations. We could be achieved. We believe our does not need to receive a perfect image of use a 40.6-cm-diameter silver-coated tele­ approach should increase the attained the Sun, rather the maximum power per scope mirror with rim angle 11 .SO (focal efficiencies, with the aim of making a tun­ area. By dispensing with image-forming ratio 2.5). Sunlight is concentrated to a able solar-pumped laser. Other applica­ requirements, we can attain concentra­ 1-cm spot 1m away from the mirror. The tions include the destruction of hazardous tions that approach the theoretical limit light is further compressed to a 1-mm spot waste and high-temperature processing of of: by a sapphire nonimaging concentrator, specialized materials. (1) chosen for its low absorption and high DAVE COOKE index of , n = 1. 76. Substituting PHIL GLECKMAN where n is the index of refraction at the these values into equation (2) gives the HELMUT KREBS target surface and f) is the semiangle sub­ theoretical limit to the concentration of JoE O'GALLAGHER tended by the Sun. The limit can be our new two-stage system, 137,000, which DAN SAGlE derived from both phase space conserva­ is a 31% increase over the theoretical ROLAND WINSTON tion' and thermodynamic arguments'. upper limit of 104,000 in our earlier Department of Physics and Our concentrator is designed according device. The Enrico Fermi Institute, to the 'edge-ray' method\ in which all Another modification is that our con­ The University of Chicago, light rays entering the device at the maxi­ centrator now works by total internal Chicago, Illinois 60637, USA mum collection angle are directed after reflection. In the previous experiment the one reflection at most to the rim of the exit walls of the concentrator were silvered, 1. Gleckman. P. App/. Opt. 21. 4385-4391 (1988). 2. Gleckman. P.. O'Gallagher. J. & Winston. R. Nature 339, aperture (see figure). The other rays are and allowed losses due to absorption. 198-200 (1989). therefore reflected within the aperture Total internal reflection is, however, 3. Welford. W.T. & Winston. R. High Collection Nonimaging Optics (Academic. New York. 1989). itself. nearly loss-free, and increases the effi­ 4. Winston. R. J. opt. Soc. Am. 60.245 (1970). A nonimaging concentrator alone, ciency of the concentrator. 5. Rabl. A. Sol. Energy18. 93 (1976). 802 NATURE ·VOL 346 · 30 AUGUST 1990 © 1990 Nature Publishing Group