REPORT No. 411 THEORY OF WING SECTIONS OF ARBITRARY SHAPE By Tanopoxs Tunoponsen SUMMARY This paper presents a solution of the problem of the theoretical flow of a frictionless incompreseible uid past airfoils of arbitrary forms. The velocity of the 2- dimensional jlow is explicitly expressed for any point at the surface, and for any orientation, by an exact expression containing a number of parameters which are funetions of the form only and which may be evaluated by convenient graphical methods. The method ie partiou- larly simple and convenient for bodies of streamline forms. The results have been applied to typical airfoils and compared with experimental data, INTRODUCTION ‘Tho theory of airfoils is of vital importance in aero- nouties, It is true that the limit of perfection as, regards efficiency has almost been reached. ‘This attainment is 0 result of persistent and extensive testing by a large number of institutions rather than of the fact that the important design factors are known. Without the knowledge of the theory of the air flow around airfoils it is well-nigh impossible to judge or interpret the results of experimental work intelligently or to mako other than random improvements aé the exponso of much useless testing. A scioneo can develop on a purely experimental basis only for a certain time, ‘Theory is a process of sys- tematic arrangement and simplification of known facts. As long as the facts are few and obvious no theory is, necessary, but when they become mazty and less simple theory is needed. Although the experimenting itself may require little effort, it is, however, often excoed- ingly difficult to analyzo the results of even simple experiments. . Thero exists, therefore, always a ten- doney to produce more test results than ean be digested by theory or applied by industry. A large number of investigations aro carried on with littlo regard for the theory and much testing of airfoils is done with insuff- cient knowledge of the ultimate possibilities. This state of affuirs is due largely to the very common beliof that the theory of the actual airfoil necessarily would, be approximate, clumsy, and awkward, and therefore useless for nearly all purposes. ‘The various types of airfoils exhibit quite different properties, and it is ono of the objects of aerodynamical science to detect and define in preciso manner the fne- tors contributing to the perfection of the airfoil. Above all, we must work toward the end of obtaining ‘2 thorough understanding of the ideal case, which is, tho ultimate limit of performance. We may then attempt to specify and define the nature of the devia- tions from the ideal case. ‘No method has been availeble for the determination of the potential flow around an arbitrary thick wing section. The exclusive object of the following report is to present a method by which the flow velocity at, any point along the surface of a thick airfoil may be determined with any desired accuracy. ‘Tho velocity of the potential flow around the thick airfoil has been expressed by an exact formule, no approximation hav- ing beon made in tho anelysis. ‘The evaluation for specific cases, howover, requires a graphical determi- nation of some auxiliary parameters. Since the airfoil is perfectly arbitrary, it is, of course, obvious that graphical methods are to some extent unavoidable. Curiously enough, tho theory of actual airfoils as presented in this report has beon brought into a much ‘simpler form than bas hitherto beon the caso with the theory of thin airfoils. In the theory of thin airfoils certain approximations have restricted its application, to smell cambers only. This undesirable feature has been avoided, and the results obtained in this report have 2 complete applicability to airfoils of any camber and thickness. ‘The author hes pointed out in an earlier report that another difficulty exists in the theory of thin airfoils. Te consists in tho fact that in potential flow the velocity at the leading edge is infinite at all angles except one. ‘This particular angle at which tho theory actually applies has been defined es the ideal angle of attack. In the present work wo shall not go any further into this theory, since it is included in the following theory. 25 « special case of rather limited practical importanco. ‘THEORY OF THICK AIRFOTS In tho theory of functions there is a theorem by Riemann which shows thet it is always possible to transform the potential field around any closed con- tour into the potential field around a circle. ‘Tho direot transformation of an airfoil into a circle may, Tiaiind Ga Po, Dak Ho. i Pua Oe Ronen 229230 for analytical purposes, conveniently be performed in two steps. The first step is to transform the airfoil into curve which ordinarily does not differ greatly from a cirelo by the transformation geered @ where f is a complex quantity defining the points in the plano describing the flow around the airfoil and 2” is another complex quantity defining the points in the plano describing tho flow around the almost circular ‘curve. Tho constant a is of dimension length and is, merely @ geometrical scale factor. In the following theory, attention is directed to the fact thet the shape of the curve resulting from transformation (1) is arbi- trary, since the sirfoil shapo is arbitrary. At a later point we shall transform this curve into a circle. ‘The 2/ and tho ¢ planes are shown superposed in Figure 1. It will be noticed that at great distances REPORT NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS This relation may further be conveniently expressed in hyperbolic functions $=2a cosh ¥ cos 0+ 2ia sinh ¥ sin 0 Since }=2+éy, the coordinates of tho airfoil (2, y) are given by 2=2a cosh ¥ cos @ y=2asinh y sin @ a Wo obtain a relation betweon 0 and the coordinates of the airfoil as follows: cosh Yaa eG sinh Yrs aan and since cosh? y—sinh? y=1 (eda) - (Geka) ~ or developed saat oop yl) -@)-@) jimilarly we obtain a relation betwoen y and the (a) where 2 “Povee shoring tho tustormaton om a none eoro Baan alt from the origin 2/—>f; that is, both flows are similar at infinity. In particular, the “angle of attack,” defined as the direction of flow at infinitity with respect to some fixed reference line in the body, is identical in both flows. Near the origin the two flows aro entirely different; one value of 2" is, however, uniquely essociated with a given value of ¢ by tho relation (1). ‘We shall, at a later point, determine the flow in the 2 plane. At present we shall determine the appear- ance of the airfoil when the almost circular curve B is given, or whet amounts to the same thing, we shill determine the curve B when the airfoil is given. In Figuro 1, Cis «circle of unit radius. Since the matter of dimensions is rather important, wo shall avoid confusion in the following by adhering to this length as unity. ‘The curve B is uniquely given by the rela- tion ¢/=aet*# where y is a known or unknown real function of the angle @ where 9 varies from zero to 2 and i is the imaginary unit. Since tho airfoil surface corresponds to the surfaco of the curve, the former is, given from relation (I) as pea ctea ere o aa(et-+e-) cos 044 alee) sin coordinates of the airfoil by using the equation (waceare) + (caatary) + or developed 2 sink? y= pt ye+( Since y is generally small for wing sections it may be more conveniently expressed for purposes of calcu- lation as a series in terms of 5-4» as follows: We heve ’ sinh Y-+cosh y Tro ry Ltsinh y+} sink? y+ - wes(sdny sar sinh y—} sink* y+ - =, 4-14)... ata i(eda) + ave [for ¥< log, 2] ‘Wo are now in a position to reproduce the conformal representation of an aitfoil in the 2’ plane, since for ach point of tho airfoil (z, y) both @ and y have been determined.THEORY OF WING SECTIONS OF ARBITRARY SHAPE ‘The curves constant are ellipses in the £ plane nt (caemany) * (caakary ‘The foci arelocated at (42a, 0). ‘The radius of curva- ture at the end of tho major axis is px Gash YY or ‘This relation is useful for the determination of y near the nose and the tail. ‘Phe leading edge, corresponding to 0=0, is located at 2a cosh av2a(1+ 220+ ap—20+}o ‘Thus we see that the length 4a corresponds to the distance between ‘the point midway between the nose and the center of curvature of the leading edge to the point midway between the tail and the center of curvature of the trailing edge? ‘To establish the magnitude of the velocity at any point (x 1) on the airfoil, wo start in customery manner with the velocity around a circle in 2-dimensional flow. Contrary to usual practice we will, however, make the radius of tho circle equal to ae where ¥ is a small constant quantity. ‘This quantity is shown Jater in this report (equation (0)), to represent the avorago value of y taken around tho cirelo C. ‘Tho potential function of the flow past this cirelo is, (reference 1, p. 83) and the velocity * dt we fF) where T is the circulation, ‘This expression must vanish at the reer stegnation point‘ (Kutte condition) whose coordinate is 2=—aeheH, where a is tho angle of attack and ¢r is shown to be the anglo of zero lift, 17h cose ans etl iy. so mata of convenes only ce tho sao tat the le appears a bey lpi ws possible they feakog tha ant alr cary mnsly cnr ax prety aes et {Bacto taaeralon Bt wil testa tat tevin of te impart {stegsl appearing the append shen mast ely sampling. Tn ache trameraeton iad i calymaterotravealane peal heady evasion teste. * gray equals ttt mags howl vector boat nee, ‘tls worthy ofmeationto nts that ths theory tin a thsepertmay actully ba apple to aooth tae of aizary bape eel emptied. Te tan Mpg ution” has bees ated Ue to tply bolle wth sharp (ot ary aarp) tall eden whos cclallon sr maybe conaldere aed By She ‘ta conln cr some alla asumptiane 114990033 —16 231 Wo obtain r= — 22 v(1— = wtevnw(Seteeteney =4xVaeto sin(arhen) (va) ‘This flow around the circle may now be transformed into the low around any other body. In the particular case in which the flow at infinity is not altered the circulation will not be altered and the force experienced by # body at the origin will remain at the fixed value L=p Vr. We will now transform this circle, defined as made into our curve B defined by the relation z—ast*#, For this purpose wo employ the general transformation 2/=2en ++) % which leaves the flow at infinity unaltered, the constants being determined by the boundary conditions. By definition anette), “*") Consequently or LiL Wet i09) = 3 (det iB ¥-YetH0—9)—7 (Ae tiB.) 3, (008 np—i sin ng) where z has-been expressed in polar form z=r(cos +i sin 9) and by De Moivre’s theorem 1 4-4 (cos ny—i sin ny) ‘Bquating the real and imeginary parts we obtain the ‘two Fourier expansions: Y-verrh[ SF cos net Be inne] (0) snd oon 2] Be cos net sinne] ——@) ‘Tho values of the cocfficionts A% Bs, as well as the quantity Yo, may be determined from (a) as follows: 1 (y c08 nede © Bea! (*ysin ney @ and vont [ae ©232 ‘The quantity @—p is necessary in the following analysis. Let us eliminate the coeficient 4* and 2 in (b) by means of (c) and (@). ‘Wo obtain Oem} cos nest [yin node ~sin need fy cos node ‘The subscript c is added to indicate that the angles so distinguished are kept constant while the integrations are performed. ‘The expression may be simplified @-e)en 2? 4 (oi np cos ng.—cos ng sin ny.)dy HU, [vain mle eade But ( — gy «008 On-+1) Zain n (ped =} cot 22) — ‘Therefore, ‘Tho latter integral is identically zero. (See Wilson, E. B. Advanced Calculus, p. 368, Follow method of exercise 10.) ‘Then & CoE vot de ID For purposes of calculation this integral is expressed in convenient form in the appendix. ‘We shall now resume the task of determining the velocity at any point of the surface of the airfoil. ‘The velocity at the surface of the cirde is $2 (600 equation (VI) and footnote). ‘For corresponding points on the curve B in the e' plane and on the airfoil in the ¢ plano the velocities aro respectively $8. $2, dw ‘The quantities ¢ and 2/ are related by the expression geet REPORT NATIONAL ADVISORY COMMITTEE FOR ABRONAUTICS Hence ae PU VoL atta) bot fede 9) 3 oar) =D (alte) cos o-+ia (o-+6-¥) sin O] = 5 (ea sinh 9 cos 6+ 2a cosh ¥ sin 6] Using the relations (II), is v A 2a sinh y= and 20 cosh y= we obtain a 1 . =F coto+iz tan). ax) Tt now remains to find the ratio 2%. Trom the relation gag @Uetba wwe obtain duende nee E hati Bo ] $F -o(E+ht¢-v0+i0-w)) ne LOH O-o)+ oe But macht from which, Lod dog 2) dog a+yetie)= Stn) ‘Therefore #7 8 y+i@-e) tie) 2 durin ‘This expression may be written de’ a apy 98 den ETO «Ge But wo have or mid p= i d(e—0) +i do and‘THEORY OF WING SECTIONS OF ARBITRARY SHAPE 233 Henco +] Replacing o by 9++« (er, the angle of zero lit, is the | value of 9—0 at tho tail’), we have i | [QE |-2 V tin (e+ 040)+5in (eter) where For a point on the airfoil we have, then, or $2 and trom (1), finaly | pplsin (a+ 040) +5in (aes) (+e) eat ; i? (siakey Fain® 6) CY) oa where and ¥’ indicate 4 and gp» respectively. | where the various symbols have the following signif- ieance: ‘Equations (LX) and (X) give now vis the velocity at any point (¢, y) of the airfoil. Vis the uniform velocity of flow at infinity. a y is the ordinate of the airfoil as measured from de" a aawe (cot Os tema) 5 the z-axis, whore to fix the system of coordi- . nates (24, 0) is the point midway between = nose and center of curvature of tho nos, = (y cot o+ie tan 0) 2 Lat! { ee » (wet otis tan oe and (—2a, 0) is the point midway between ; : : the tail and center of curvature of the tail Because wo aro interested more in the magnitude than | gig thie angle of atteck as measured from the in the direotion of tho velocity wo will write for the | pecs t indented igure numerical value of this expression 60% WV, «, and ¢ are all functions of =. Bauation (KI), expressing the value of the velocity at any point of an airfoil of anyshapo, iseurprisingly simple ‘when the complex nature of the problem is considered. : Ithesthe distinctadvantage of being exact; no approxi. ‘Phe quantity xy cot'e+(Z2) tan’ is readily soon | mations have besn made in tho preceding analysis, We shall note some of the properties of this impor- ag _\@reok OFF tar) TV [gee on to bo oqual to (by relation (11)) tant relation. Because y is generally small, the term. Voce aq kqpis of influence chiefly near the leading edgo, Ce whore sin @ is smell. Tt is noticed, however, that if or also gicg70 for 6=0, equation (XII) yields in all cases sinh y-+sin* 9 This means that the velocity at the nose be- comes infinite for sinh y=0 (thin airfoils). ‘This fact has beon pointed out in en eszlier report. (Reference 2) ‘The quantity 5! or sinh ¥ is thus of con- : 5 siderable significance in the theory of thick airfoils. ‘Tho numerical value of the velocity at the surface of oneeoe See are ing tho elrcle is obtained by equations (VI) and (VID) «8 | gst Ay and v= eet-c'Ao, ‘Wheto Ad iow onal woah follo : Substituting the general point 2—aete+ (te), where | 1 equation (KIT) of (L+e) [sin Oat) tein (wer) a is tho angle of attack as measured from the axis of Le coordinates, in equation (VI) 17 Vink y Fein) CY) wo got aw, Fi de. _ yy —ertlete) —95 Vain (at egdertet oleh Ltd) [a tateteaotetel di [= 7 ¥ me —ViL—008 2+ 9) +2 sin(arte) sin (a+) tt Vera) OY) +4 (Gin 2 (+9) +2 sin (wes) 008 (at g))] ~IS [GB/ = 2 aint (arten) +8 sin (wer) sin (ay) ete © +4 sin? (a+) Velen (S) Jar {88 aa Visin (et plein (ete SE eee Rees eam234 REPORT NATIONAL ADVISORY y near the tail may be expressed as. Vet y AOFE "(ANE + + or ¥ ai aya hy Le Sew W"AOt ve ry at 49=0. The velocity is in this case zero, indieat- ing the presence of the rear stagnation point. If, on the other hand, ¥ is zero, that is, if the tail is per- fectly sharp, ‘The quantity = is infinite if y> is different from zero voy a er for a8=0 and the velocity at the tail is aye? or V" es or enc beye a ® (or the Clark Y, v;* is about 0.88 V? near the tail.) ‘Woe obtain the front stagnation point by letting 0 in equation (XI). Hence et eba= (ate) 0= Garten ter) In a previous report (reference 2) er z has been defined as the ideal angle of attack. It is seen that, for this angle of attack, @ is zero or the steg- nation point occurs directly at the nose. ‘Equation (XII) may also be applied to strut forms, and for such symmetrical shapes takes even a simpler form. ote V* PRACTICAL APPLICATION OF RESULTS ‘Wo will now apply Formula (XII) to the typical case of the Clark Y airfoil and calculate the velocities at points of tho airfoil surface. ‘The detailed method. of procedure is as follows. 1. Tho axis of coordinates is drawn through the points (2a, 0) and (—2a, 0) located respectively at the point midway between the nose and the center of curvature of the nose and the point midway between the tail and the conter of curvature of the tail. (See fig. 6.) The radius of curvature at the leading edgo is 1.75 por cont chord. 2. The points (z, y) of the upper and lower surfeces of the airfoil are determined with respect to this axis. COMMETTAB FOR ABRONAUTICS 8. Sin%, sin 6, and @ are determined by the relation saves Pe) viweni-(B)-(E) 4. is given by the relation yyy v-(sehas ~acztes) te Ce 5. ¥ is plotted as a function of @ wre on pve EL cot mula shown in the appendix: 6. Determine er de by for- eo= — 410.628 y/, 4+ 1.085(41—¥-1) + 0.445(Y4~¥-2) +0.281(Ys~¥-s) +0.104(Y4—¥-)] where ¥, is the slope of the ¥ curve at p= Yi the value of ¥ at e=ytZ, ya at em ert, ote. Yas the value of ¥at p= eZ, ete. 7. From the ¢ versue 6 curve and from tho y versus @ curves ¢’ and y/ are determined. 8. Determine F by the relation Zar in 6 Vlei 9. (0-+2) is determined in radians and degrees. 10. Sing+atd-+tsin(atér) is now caloulated where a is the angle of attack as measured from the axis of coordinates. + sino |a+y) | UL BaF ilein Oate)+ sin (eten)] 2, Bat -(Z) Gressur) ‘The entiro calculation, properly arranged, can be quite acourately obtained in a very short time. COMPARISON WITH EXPERIMENTAL RESULTS In order to compare the theory with experimental results, the geometric anglo of attack ae as measured in tho wind tunnel must be corrected for a number of items, such as finite span and effect of wall interference. ‘We may, however, obtain approximately the apparent, or effective angle of attack ay (in radians as measured from the‘ angle of zero lift) by teking the quotient of the sree of the pressure-distribution curve and 5.5, since itis known that this value of the lift: cocffi-Experinental THEORY OF WING SECTIONS OF ARBITRARY SHAPE 235 ‘sts © 20 40 0 720 Fer contchord 18 Theoretic angie Experimental i “O-- inf RN Experimental a |_| “f= K z 5 4 | a o| [ © @ O20 30 — 0 95 00 0 Percentebord C0 Recontcrsrs = FlounE 2.—Preesure-dltribation carves slong saxtsof ClarkY; Pagatust per cent chord (@e-PW. Qane iW. @arni i, env w236 REPORT NATIONAL ADVISORY COMMITTEE FOR AERONAUTICS - \ ~ _ “1 | P | 24 1 «Wl 2 @) 3 +248, Re ceit onied ) -2 Bu os 8 Te 78 cott chard 1 “1 ell P Giant @° i © ® 1 @ oie Percent chord © Recent onerd Yioons Stent prea alsetbtlon ag pass of Cate Y errs. Dall. Oat W. Wen we cient is very nearly realized in most cases. ‘This has been done in Table IIT, and the angle of attack a, which should be substituted in tho Equation (XII), is given in the last column. ‘Tho pressure distribution curves, Figures 2a, b, ¢, d, and 3a, b, o, d, were obtained, by application of Equation (XI) to the Clark ¥ airfoil. Numerical results are shown in Tables I, TI, and IIT. ‘The experimental values are from original data sheots for N. A.C. A. Technical Report No. 353, and aro not entirely consistent due to difficulties axperienced in these experiments. After the theoretical pressure distribution curves have been obtained, the moments, about any required axis may be found. Table IV cr ar a rT ‘Powns 4~Moment agit onl flick gives some of theso results and Figure 4 shows the ‘comparison with experimental date taken from N. A. C.A. Technical Report No. 312. Lanermy Manonrar Arnonaunican Lanonatony, ‘Narionar Apvisony Cosamrren ror Amronavrics, Lanatey Frau, Va., October 16, 1981.APPENDIX EVAMUATION OF THE FORMULA = (eden — [* yeot 252) tle Don gf peat GED ay Although the above integrand becomes positively and negatively infnite around p=. it is readily verified that for y finite, throughout the range 0—2z, the integral remains finite, the positive and negative infinite strips exactly canceling each other. » Tho value of the integral for any point y, may be accurately obtained by the following device. Wo know that if y is a continuous function and the range ¢t0 ga not too lange in 2 3 foe # cot 2522 des very nearly 4 log — where #4 is tho avorago value of y in the range y: to g Also near p= 9, we may write Pevet (e-90) Wot (Om ‘Thon for ¢ a small quantity Hy (O29 P58) 4 2 ie (Binco the even powers drop out and the lim cot ¢=1).. 0 Lot us now divide the interval 0-2 into 10 parts, starting with g as a reference point. (See fig. 5.) br ‘Tx, Te on, Home int Tbeebiewtie ‘Then, on BfMycot 2G ae sin ?® FF sin 35 nk + a-¥-s) log —22+ (Ya nfs sing sings sin? + W- vs) log: 30 int = ~ 110.028 97.41.0605 (hv) + 0.445 (a4) 0.231 (Us H-s) + 0.104 (Ye Yd] where ‘sis the slope of the y curve at e= 9. vs value of Yat p= yet Eva at p= 9. Fr Yat om pet 2S te at p= eet Zs ote. ‘To evaluate the above integral it is, strictly speak- ing, necessary to know ¥ as a function of g rather than of 6.1 Wehave =0+e. For all flattened or stream- line bodies, however, ¢ is small; for ordinary airfoils itis, in fact, 60 small that ¥(@) may unconditionally be considered equal to ¥(s). For the sake of mathe- matical accuracy we will, however, indicate how the problemmay be solved also for bodies of more irregular contour by successive approximations. We have YOVO+EV Ot + + + As a first epproximation wo noglect tho second and all following terms of this expression. ‘Tho valuo of thus obteined by graphical integration or otherwise is then used in tho expression for y(e) and o second intogration is performed, ete. "The oqutiea fr «bo moliotr integra eaton Aion ict ater: Tartans besten oto al aioe of +0 a oobi is eta | ona ota to any dated crn by erlary Gai Inna. 237238 ‘REPORT NATIONAL ADVISORY APPLICATION OF FLOW FORMULA TO THE SPECIAL ‘CASE OF AN ELLIPTIC CYLINDER As a matter of interest we will assume the form of the body to be thealipse (5255) +(sebng) = and find 7 for zero angle of attack, i. 0. we have Y=Vemconstant, ¥/=0, e=0, 70, a= 0. Equation (XI) becomes setter maahy eee sinh ¥+(s- ey) o (ay) and P-()-- aaa aap This result checks exactly with the form given by Dr. A. F. Zahm in N. A. O. A. Technioal Report No. 253, Flow and Drag Formulas for Simple Quadrics, equation 14. _ REFERENCES 4, Glauert, H.: Blessonts of Airfoil and Aln-Sorew ‘Theory. Cambridge, University Press, 1026. 2. Theodarsea, "Theodore: On tho Theory of Wing Sections ‘with Particular Reference to the Lift Distribution. TR. No. 985, N-A.C.A, 1031.» EXPLANATION OF THE TABLES ‘The first part of Table I refers to the upper surface or to positive ordinates of the Clark Y, the second part to the lower surface or to nogative ordinates. Column 1 gives the location in per cant of the chord} 2 gives the ordinates with respect to the z-axis in this same unit; 3 and 4 give z and y in the present system of coordinates; 5, 6, and 7 give sin* sin 0, and 6, respectively (Equation (11D); 8 gives ¥ (by equation (1V2)); 9 gives ¢ (appendix); 10 and 11 give%Y ena 2 as obtained from y against # and « ageinst @ curves; (See figs. 7 and 8). Column 12 gives the quantity comnetema rox smoNAtitCs oe Oe) ee (1.11); +) F Column 14 gives 0+« in degrees. ‘The volocity at any point 2 and angle of attack a is given by v=V [sin (a+ 0-+0)+ sin (a-+er)}-F and the pressure, P. ‘8 wit) Tt must be noted that a is measured from the lino of 14 # { y 2 faa wer -22 100% oows ¢Olak ¥ srst—ehowtag eat of cornaes flow to the z-axis as shown in Figure 6, and if otherwise measured, must be reduced to this basis. ‘Rowe 7th walt del ee tha dra rae and the eomepeadlog ‘iver meg ~10} ¥ eee (meer eae ae ee reed Be mn nose thper gertace Tail Lower suroce tose ‘Moves Ee) They apa # carve are Ca¥. (0) ‘Te eagtat caro far tha Clark Y289 ‘THEORY OF WING SECTIONS OF ARBITRARY SHAPE TABLE I CLARK ¥ ‘UPPER SURFACE ~ [ravs~auonenua~saan| [feaesaen-onezacan £ |. yearenesesagzgeaa| | T5nRa7y on | veeeseesaanaagg| | aueeaseeaseezes a eaeaaaaEEA| | SHRERS = ~ | segaseseraccsaese| [scsusaensieotdel i SEES apzesanaalensile = ~ | gezeeeenzenenazee| | gaueseszaaazeczee i =! : at «| sasewseusnessses u eit gaeeeesee 9, ote Hae : 3 |pcamarencunaraeaa 4 E |: ovpeparerienaaag g | SResaeneuaeneRee saguasausasagag E |easasgaczuessceas naa [eee = eeeeemnser SERSESRATESEIISUS Shrine =e | samamemarentiee silceneaaageszare i) SHEETS i ee subenssQHG820ete ol E Het Htret oj | GoRSSUIEENEIERE 5 8 : z “ aaageagqate eeees| |aaageasnassaeseee |; [--aanconanwenrea 2 8 — my uy £ |-onaaeeasecgansee | © | .ssasseagsseases®) | sarsuaesamaeres® | | -Eeacasesseaseg| 2 lia arr R 83c3, g aRaS i ARSSERBEER one geanessraseg, ‘tho numerical values for Figure 2a in detail as an example. Seo alao Table I. TABLE IV