SCIENTIFIC CORRESPONDENCE weight, did not flower and died soon after some of these herbicides (such as the ALS manently bent backwards and the feathers the emergence of the broomrape flower­ inhibitors), but is expected to evolve more are free to rotate in the nose-up sense in 2 ing stalk. slowly to the others • their sockets. Nose-down rotation is pre­ The doubled yields afforded by control Daniel M. Joel vented by the feathers' attachment in their of the parasites will more than offset Yeshaiahu Kleifeld sockets. It is also restricted by the nearest­ the added cost of both the transgenic Dalia Losner-Goshen neighbouring feather lying behind it, and seed and the small amount of herbicide, Geza Herzlinger partly overlapping the upper side of the even in underdeveloped countries. Department of Weed Research, trailing vane, so that the feathers are This approach should only be used with Newe-Ya'ar Research Center, pressed together on the down-stroke. As a crops that do not interbreed with related Haifa 31900, Israel result of feather curvature, and less of weeds in the same locality. The use of Jonathan Gressel vane asymmetry, the flight feathers of such transgenics represents a necessary Department of Plant Genetics, modern birds rotate in the nose-up sense stopgap measure until other means are Weizmann Institute of Science, on the upstroke when this is aerodynami­ found, as resistance can rapidly evolve to Rehovot 76100, Israel cally non-functional, letting air through the wing. This rotation occurs about an axis through the feather base, and because Feather asymmetry in Archaeopteryx of the shaft curvature, most of the feather vane is behind this axis to provide the SIR -After comparing flight feather anteriormost primary feathers of modern required nose-up torque. asymmetry of Archaeopteryx with the birds often separate so that the outer part Archaeopteryx flight feathers are as 3 3 asymmetry of several extant birds with of each one acts as an aerofoil on its own . strongly curved as those of modern birds . various flight styles, 'flapping', 'gliding' The vane asymmetry then comes into Its feather curvature alone could there­ and 'flightless', Speakman and Thomson1 operation and effects a proper orientation fore provide the nose-up moment claim that Archaeopteryx was not capable of the feather to the incident air stream. required for individual feathers to rotate of sustained flapping flight. In my view, On random changes of angles of attack, about the axis through the feather base their analysis is flawed and does not sup­ the aerodynamic centre of pressure of a to let air through the wing on the port their conclusion. flat plate, or feather, does move fore and upstroke. Also, the vane asymmetry in First, they report asymmetry values for aft in such a way that it maintains dynamic Archaeopteryx is pronounced enough to Archaeopteryx feathers that are far too low, stability in pitch, and hence in angle of prevent a counteracting nose-down and they make a misleading comparison attack, provided that the span-wise torsion moment to be set up about the local feath­ with extant birds. I have examined high­ axis lies within the interval 27-35% of the er shaft on the upstroke. Feather curva­ 2 4 quality photographs in two fold-out plates , chord length behind the leading edge • ture and vane asymmetry in Archaeopteryx showing the left and right wing of the The feather shaft acts as a local torsion are therefore fully consistent with an Berlin Archaeopteryx specimen at 2.8 and axis in any chord-wise profile. When the active, flapping, flight mode. 4.1 times their natural size. The first three feather shaft lies ahead of the 27% chord R. Ake Norberg feathers are staggered in length, so the point, there is a nose-down pitch moment Department of Zoology, anterior margin is entirely free and clearly that also tends to match the angle of University of Gothenburg, visible all along the outer half of feathers 1 attack to the incident air-stream. Medicinaregatan 18, and 2 in both wings. Also, feather 3 of the The crucial feature of Archaeopteryx S-413 90 Gothenburg, Sweden right wing has its anterior margin free in feather asymmetry is that the shaft of the the outer part that extends beyond feather anteriormost primary feathers is 24--34% SPEAKMAN AND THOMSON REPLY - Nor­ 2. The photographs show the ventral side of the chord behind the leading edge, berg raises several interesting but erro­ of the wing, so, because of the way feathers completely ahead of the critical rear limit neous comments. are arranged in a bird wing, the rear mar­ 35%, as required for self-stability in pitch. (1) He has measured vane asymmetry gin of these feathers is also free and unob­ Consequently, vane asymmetry in for one Archaeopteryx. His measurements scured. At about 25% of the feather length Archaeopteryx primary feathers is pro­ exceed those made by us 1 on two speci­ from the tip, the feather shaft of the four nounced enough to confer automatic mens. Unfortunately, many of his mea­ anteriormost primary feathers (1--4) is pitch control on separated feather tips, sures were made on regions of the feathers located 24--34% of the feather chord and therefore does not indicate a lack of which are overlapped. Since measuring behind the leading edge; thus, the trailing­ powered flight. asymmetry depends on defining both vane is 3.11-1.91 times as wide as the lead­ The asymmetry of the first three pri­ feather margins, his measurements ing-vane. This is much more asymmetrical maries is what matters most. Because of depend critically on a subjectively inferred than 41% of the chord behind the leading their staggered length in Archaeopteryx, position of the hidden margin. Different edge, or the ratio 1.46, reported by Speak­ they all form part of the wing's leading edge observers have inferred asymmetry (ap) for man and Thomson for primary feathers 4, and therefore they are the feathers most the overlapped feathers of Archaeopteryx 5 and 6 of the Berlin specimen. likely to split apart in flight, acting as aero­ ranging from complete symmetrj (ap = My asymmetry values are near the foils- actually, leading-edge slats - on 1.0) to extreme asymmetry6 (ap is approxi­ lower limit of, but partly inside, the range their own, with a need for the automatic mately 4--5 from Fig. 1 in ref. 6). Our mea­ for 'flapping' and 'gliding' birds (Fig. 2 in pitch control that the vane asymmetry gives. surements are more accurate as they were ref. 1). Further, Speakman and Thom­ Third, the function of vane asymmetry made on sections of the feathers which do son's values for extant 'flapping' and 'glid­ is linked to the function of feather curva­ not overlap. The measures he made which ing' birds are from primary feathers 1 or 2, ture3. Flight feathers in bird wings are per- were not overlapped refer to the diminu­ which are more asymmetrical than feath­ tive first and second primaries; these small ers further back (Fig. 1 in ref. 1), but they Scientific Correspondence feathers would not be important in flight. nevertheless compare them with primaries In our measurements of extant birds we Scientific Correspondence is intended to 4, 5 and 6 in Archaeopteryx. I therefore provide a forum in which readers may ra ise ignored the first primary when it was conclude that the feather asymmetry of points of a scientific character. Priority will diminutive (less than half the length of the Archaeopteryx is within the range of mod­ be given to letters of fewer than 500 words second primary); thus, comparing the ern birds using flapping flight. and five references. diminutive primaries of Archaeopteryx to Second, on the wing's downstroke, the our sample of extant birds is invalid. NATURE · VOL 374 · 16 MARCH 1995 221