SCIENTIFIC CORRESPONDENCE process was advanced to the point that it ton corresponds almost exactly to that Sharov'sl plate 4. Impressions of the flight had become detached. In the photographs exhibited by complete, undisturbed exam­ membrane, the posterior boundary of (Fig. la of ref. 2), neither wing membrane ples of , from the Upper which is evident in Sharov'sl plate 5, Fig. clearly indicates where it was attached to Solnhofen limestonelO. However, la, show that the right cheiropatagium the leg, although illustrationsl.2 of the the evidence for soft tissues, especially the remained intact and did not undergo any right wing show an imprint identical in extent and structure of the wing mem­ post-mortem transport. The wings occupy shape and position to that of the left wing. branes, is much clearer in Sordes. The wing almost identical positions, as Sharov This observation may be suspect because, membranes of Solnhofen are correctly showedl, but the outline of the 3 as reconstructed here, the left wing may preserved as impressions , ,9,10, but even right cheiropatagium is more difficult to have originally attached near the knee. the best examples can be interpreted in a trace in photographs because of the I can accept that a membrane may have variety of ways3,4.11. The decomposition of patchy preservation of the black, mineral­ spanned from the lateral digits to either Sordes was halted at an earlier stage than ized tissues in this area. side of the tail (it would not impede in Solnhofen pterosaurs: extensive tracts Peters' reconstruction also fails to terrestrial locomotion), but I question of black, mineralized soft tissues8 pick out explain the unusual morphology and 2 how a membrane spanning the legs and impressions left by the wing membranes. position of the fifth toe ,S,12 and is attached medially can also be attached to Outlines of the wings can be traced with contradicted by details of the internal the lateral digits. The image of pterosaurs ease and show, quite unambiguously, the composition of the membranes. The with broad bat-like wing membranes is existence of a uropatagium and attach­ uropatagium, preserved in three other traditional, but at least two Solnhofen ment of the cheiropatagium to the leg as individuals as well as the holotype, con­ 3 l specimens ,4.7 were preserved with very far as the ankle ,2,8. tains relatively short, sinuous, loosely narrow wings stretched only between the Peters' narrow-winged reconstruction packed fibres, clearly distinguishing it elbow and wing finger. Thus, either Sordes of Sordes is based on a highly unreliable from the middle and outer parts of the is different from these specimens, or pre­ technique, interpretation of photographs, cheiropatagium, which is characterized l vious interpretations ,2 are in error. and unfounded suppositions of post­ by much longer, straighter, closely pack­ 2 David Peters mortem disturbance. Central to his argu­ ed fibres • 12812 Wood Valley Court, St Louis, ment is the claim that the uropatagium is David M. Unwin Missouri 63131-2051, USA part of the right cheiropatagium which, Department of Geology, somehow, drifted into a symmetrical posi­ University of Bristol, UNWIN AND BAKHURINA REPLY - Recon­ tion between the hindlimbs, the internal Bristol BS8 1RJ, UK structions of the Upper Jurassic fibres fortuitously lining up parallel to the Natasha N. Bakhurina* Sordes pilosus with broad wings attaching lower leg. However, Peters' interpretation Palaeontological Institute, to the hindlimbs and a medial membrane, (his figure b) omits parts of the central Russian Academy of Sciences, the uropatagium, between the legsl), have region of the right cheiropatagium, pre­ Moscow 117647, Russia 4 , been challenged by Peters, and others on served lateral to the right knee and clearly *Present address: Department of Geology, University of the grounds that preservation of the visible together with the uropatagium in Bristol. remains is too poor to permit such infer­ ences. However, comparison of Sordes with other exceptionally preserved Early AGEing and Alzheimer's pterosaurs shows that the holotype is one of the best-preserved individuals in exis­ SIR - Mattson et ai. suggese that in enhances the aggregation of 1: or amy­ tence2,s. As originally interred, the skele­ Alzheimer's disease "glycation is a late loid-~9 into neurofibrillary tangles and ton was complete and fully articulated, event" and that it "results from free senile plaques. Therefore, a contribution with the neck flexed backward as in some radical generation". Despite their cita­ of glycoxidation modifications early, as 9 Solnhofen pterosaurs • Each forelimb is tion of some of our papers to support well as later, in the pathogenesis of 2 4 partially folded and, although parts of the their view - , their opinion is directly Alzheimer's disease should be seriously left forelimb were lost during collection, it contrary to the findings that the protein considered. is evident that, like the hindlimbs, they are PHF-1:/A68, regarded as the precursor Mark A. Smith* S almost perfectly symmetrical with respect to neurofibrillary tangles , is modified Lawrence M. Sayre* to each other, contradicting claims of by advanced glycation end products Michael P. Vitekt post-mortem disturbance. (AGE)2.6. Vincent M. Monnier* Interestingly, the position of the skele- We have also demonstrated that AGE George Perry* epitopes localize to diffuse amyloid-~ *Institute of Pathology,

1. Sharov. A.G. Akad. nauk SSSR pa/eont. Inst. Tr. 130, senile plaques' (lesions that represent Case Western Reserve University, 104-113 (1971). one of the earliest pathological changes Cleveland, Ohio 44106, USA 2. Unwin, D. M. & Bakhurina, N. N. Nature 371, 62-64 7 ), (1994). in Alzheimer's disease as well as tThe Picower Institute, 3. Welinhofer, P. Ann. Naturhist. Mus. Wien 88A, 149-162 neurofibrillary tangles and neuritic senile Manhasset, (1987). plaques. Moreover, the proteins found in New York 11030, USA 4. Padian, K. & Rayner. J. M. V. Am. J. Sci. 293A, 91-166 (1993). these lesions are continuously exposed to 1. Mattson, M. P., Carney, J. W. & Butterfield, D. A. Nature 5. Gekker, R. F. Trudy paleont. Inst. 15, 7-85 (1948) (in glucose in vivo and therefore constantly 373, 481 (1995). Russian). 2. Yan, S. D. et al. Proc. natn. Acad. Sci. U.S.A. 91, 6. Doludenko, M. p" Sakulina, G,V. & Ponomarenko, A.G. subject to Schiff base formation follow­ 7787-7791 (1994). The Geology and Fauna and Flora of a ed by Amadori rearrangements and Unique Locality, Au/Ie (Karatau Range, Southern 3. Smith, M. A. et al. Proc. natn. Acad. Sci. U.S.A. 91, Kazakhstan) (Inst. Geol., Russian Acad. Sci., Moscow, AGE modification. Of relevance is that 5710-5714 (1994). 1990) (in Russian). intracellular AGE modification, contrary 4. Vitek, M. P. et a/. Proc. natn. Acad. Sci. U.S.A. 91, 7. Welinhofer, P. The Illustrated Encyclopedia of Pterosaurs to popular dogma, often occurs very 4766-4770 (1994). 1-192 (Salamander. London, 1991). 5. Trojanowski, J. Q. & Lee, V. M. Am. J. Path. 144, 8. Bakhurina, N. N. & Unwin, D. M. Hist. Bioi. (in the rapidly8. 449-453 (1994). press). Rather than representing a tombstone 6. Ledesma, M. D. et a/. J. bioi. Chern. 269, 9. Wellnhofer, P Abh. bayer. Akad. Wiss. (N.F.) 141,1-133 21614-21619 (1994). (1970). or late event, the process resulting in gly­ 7. Giaccone, G. et al. Neurosci. Lett. 97, 232-238 (1989). 10. Welinhofer, P. Palaeontographica149, 1-30 (1975). cation-related modification appears to 8. Giardino, I., Edelstein, D. & Brownlee, M. J. elin. Invest. 11. Pennycuick, C. J. Bioi. Rev. 63, 209-231 (1988). 94, 110-117 (1994). 12. Bakhurina, N. N. & Unwin, D. M. J. vert. Pafeont. 12 represent a very early event that pro­ 9. Hunt, J. V. & Wolff, S. P. Free Radie. Res. Commun. 9 (Suppl.), 18A (1992). motes the formation of free radicals and 12-13,115-123 (1991). 316 NATURE· VOL 374 . 23 MARCH 1995