CHAPTER Applications of Derivati ‘OVERVIEW This chapter shows how to draw conclusions from derivatives. We use serivatives fo find extreme values of functions. to predict and analyze the shapes of _raphs, to find replacements for complicated formulas, to determine how sensitive formulas are error in meaburoment, and to find the 2aros of functions numoricaly. ‘The key to many of these accomplishments is the Mean Value Theorema, a theorem whose corollaries provide the gateway to integral calculus in Chapter 4. Extreme Values of Functions "This section shows how to locate and identify extreme values of continuous func- The Max-Min Theorem ‘A function that is continuous at every point of a closed interval has an absolute maximum and an absolute minimum value on the interval. We always look for these values when we graph 2 function, and we will see the role they play in problem solving (this chapter) and an the development of the integral calculus (Chapters 4 and 5). ‘Theorem’ ‘The Max-Min Theorem for Continuous Functions Iffis continuous at every point of a closed interval J, then f assumes both an absolute maximum value Af and an absolute minimum value m somewhere in £ That is, there are numbers x; and.x in f with f(a) =m, fl) = My ‘and ms f(s) < M for every other xin 1 (Fg. 3.1 on the following page). ‘The proof of Theorem 1 requites a detailed knowledge of the real number system and we will not give it here 189190 Chapter 3: Applications of Derivatives 3.1. Typical arrangements of a continuous function's absolute maxima and minima fon a closed interval [a6 yesiox 32 Figure for Example 1. Nobset ave 23 On an open interval, a continuous function need not have either 2 ‘maximum or 2 minimum value. The function #12) = has neither a largest nor ‘ emalest value on (0 1). Maxima an miinon strep y= fad in van oie tt gg ak dg °F mR 7 Maxim tir poi, Minima terior poi, ‘misma epi Pastis end EXAMPLE 1 On |~/2,/2), fx) = cose takes on a maximum value of 1 (once) and a minimum value of 0 (ewice). The function g(x) = sin twkes on a ‘maximum value of | and a minimum value of —1 (Fig. 3.2) a As Figs. 3.3 and 34 show, the requirements that the interval be closed and the function continuous ar key ingredients of Theorem 1. Without them, the conclusion (ofthe theorem need not hola 3.4 Even a single point of discontinuity can keep a function from having either @ ‘maximum of @ minimum value on a dosed interval. The function th Tex<0 y=]o° x=0 x O SOQ for all xin D. Absolute maximum and minimum values are called absolute extrema (plural of the Latin exiemum), Absolute extrema are aso called global extrema. Functions with the same defining rule can have different extrema, depending fon the domi EXAMPLE2 — (See Fig. 3.6) Function Domain Absolute extrema rule D on D (if any) a) (00,00) No absolute maximum. Absolute rinimuen of Oat x ») 10,21 Absolute maximum of (2)* = 4 at x =2, Absolute minimum of 0 at x ° 0.2) Absolute maximum of 4 at x =2. [No absolute miniraur, a 0.2 [No absolute extrema, a192 Chapter 3: Applications of Derivatives Definition Local Extreme Values ‘A function f has aJocal maximum value at an interior point ¢ of its domain it F68) © fl6) forall xin some open inzeral containing & ‘A function f bas @ local minimum value at an interior point ¢ of its domain if FG) > Fle) forall x in some open interval containing c. ‘We can extend the definitions of local extrema to the endpoints of intervals by defining f to have # local maximum of local minimum value af an endpoint e if the appropriate inequality holds forall x in some half-open interval in its domain, containing ¢. In Fig. 3.5, the function f has local maxima at c and d and local ‘minima at a, e, and b ‘An absolute maximum is also a local maximum. Being the largest value overall, it is also the langest value in its immediate neighborhood. Hence, alist ofall local ‘maxima will automatically include the absolute maximum if there is one. Similarly, 4 list of all local minirna will include the absolute minimum if there is one. Finding Extrema ‘The next theorem explains why we usually need to investigate only a few values to find « function’s extrema, Theorem 2 ‘The First Derivative Theorem for Local Extreme Values AE fas a local maximum or minimum value at an interior point ¢ of its ‘cal masium vale ‘domain, and if.” js defined atc. then fle=o. Proof To show that f"(c) is zero at & local extremum, we show first that /“(c) ‘cannot be positive and second that fe) cannot be negative, The only nuraber that is neither positive not negative is zero, so that is what "(c) must be. ‘To begin, suppose that fas a local maximum value atx = ¢ (Fig. 3.7) so that “f(2) ~ fle) $0 for all values of x near enough to e. Since ier of f's domain, f'(c) is defined by the two-sided limit tim LO-LO Secant slopes = 0 (oem) “This means thatthe righthand and left hand limits both exist a x = ¢ and equal 47 A carve with oc! masimum value. _£(0) When we examine these limits separately, we find that ‘The slope atc simultaneously the limit of ft. onpostive number and nonnegative Fe) = tim £2 Beane ieo0 ayHow to Find the Absolute Extrema of a Continuous Function f on a Closed interval 1, Evaluate fatal ecical points and endpoints 2, Take the largest and smallest of these vals, 3.1. Extreme Values of Functions 193 Similarly, fim LO=SO 29 necmeteo <0 gy mye and fia) = Jee) ‘Together, (1) and @) imply "(c) = 0. ‘This proves the theorem for local maximum values. To prove it for local ‘minimum values, we simply use f(x) > f(c), which reverses the inequalities in (1) and 2, a ‘Theorem 2 says that a function's first derivative is always zero at an interior point where the function has a local extreme value and the derivative is defined. Heace the only places where a function fcan possibly have an extreme value (local or global) are 1. interior points where Js” =0, 2 interior points where fis undefined, 3. endpoints of the domain of f ‘The following definition helps us to summarize. Definition ‘An interior point ofthe domain ofa function f where fis ero of undefined is a ertical point of f Summary ‘The only domain points where a function can assume extreme valves are critical poitus and endpoi ‘Most quests for extreme values call for finding the absolute extrema of a continuous function on a closed interval. Theorem 1 assures us that such values exist; Theorem 2 tells us that they are taken on only at critical points and endpoints. These points are often so few in number that we can simply list them and calculate the corresponding function values to see what the largest and smallest are. EXAMPLE 3 Find the absolute maximum and minimum values of f(x) = x* on [-2, 1). Solution The function is differentiable over its entire domain, so the only evtieal point is where f'(x)=2r =0, namely x =0, We need to check the function's values at x = 0 and at the endpoints x = ~2 and x = I: Critical point values (0) = 0 Endpoint values: G-=4 f= ‘The function has an absolute maximum value of 4 at x minim value of O at + = 0194 Chapter 3: Applications of Derivatives 3.8. The extreme values of g(t) — 6t— fon [-2, 11 @xample 4), Abolte maxi So gs maxim at Nad minions ‘here mn 3.9. The extreme values of hie) = #° on 1-2,3] occur atx O and x = 3 (Exam ple). 3.10 f(x) =x" has no extremum at leven though #3) = (13) is undefined at x EXAMPLE 4 Find the solu extema values of ¢(¢)= 81 ~#* on [=2, 1) Solution The function is differemiable on its entre domain, so the only extical points occur where g’(r) = 0, Solving this equation gives 8 ano p=2 218 4 point not in the given domain. The function’s local exttema therefore oceur at the endpoins, where we find (-2) = 32 (Absolute minimum) g(D)=7. (Absolute maximum) See Fig. 38, a EXAMPLE 5 Find the absolute extrema of h(x) =x" on [~% . 3]. Solution The firs derivative Hw bas no zeros but is undefined at x = 0. The values of ft at this one critical point and atthe endpoin's x = —2 and x he) M2) hn) ‘The absolute maximum value is 9", assumed at x 6, assumed at x = 0 (Fig. 3.9), While a function’s extreme can occur only at critical points and endpoints, not every critical point or endpoint signals the presence of an extreme value. Figures 3.10 and 3.11 illustrate this for interior points, and Exercise 34 asks you for a function that fails to assume an extreme value at an endpoint of its domain, ay agi ap? a 919, 1 absolute minimum is ? has no extremum at x even though g'bx) = 3x7 is zero atExercises 3.1 195) [As we will see in Seetion 3.3, we can determine the behavior of a function fat critical point ¢ by further examining f°, but we must look beyond what f” does atc itself Exercises 3.1 Finding Extrema from Graphs Jn Exereises 1-6, determine from the graph whether te funtion has any absolute extreme values on a, 6], Then expain how your answer is consistent with Theorem yon yen a4 uti, ff li, oe D . sy y=ne Tee Absolute Extrema on Closed Intervals In Exercises 7-22, find he absolute maximum ae minimum values of cach fonction onthe given interval, Then graph the function. Keay ‘he points onthe graph where the absolate extrema occur, and include their enordinates. 9 firyewt1, —Lexe? 1. F() AL Fe 1 Fo) = 1k he) =o 1a, heey = 3088, 1 1S. gs) = VI=P, 16. ats) 17. Fay= 18. f16) 19. gc) 2. g(x) = sex, a. so 2 fuel, 4e0s7 In Beercises 22-26, find the funtion’ absolute maximums and mnie mum values and say where they are assumed, 2 siya, ~1er58 W. sis -1exe8 25, (= 0, —R=Ee fin) < fla). 2. f decreases on Jif fm) < fn), Corollary 3 ‘The First Derivative Test for Increasing and Decreasing Suppose that fis continuous on la, &) and diferentiable on (a, 6). If f° > 0 at each point of (a,b), then finereases on fc, B If f/ <0 at each point of (a, 6), then fdecteases on [a, 6] Proof Let x, andl x3 be wo points in [a b] with x; < x2. The Mean Value Theorem applied to fon [x %2] seys that Fee) - Ff) = FOl—n) © for some between x1 and a2, The sign of the right-hand side of Eg. (6) is the same as the sign of /"(c) because x2 — 5, is positive. Therefore, (xa) > flsy) if J” is positive on (a, B), and f(x) = F(x) iff’ is negative on (a,b). a EXAMPLE 5 The function f(x) =.x* decreases on (~c0, 0), where f(x) = 2x <0. Ie increases on (0. 00). where f(x) = 2x > 0 (Fig. 3.22) a Function Function ccesing fncesing yeo yoo 3.22 The graph for Example 5Exercises 32 203, Exercises 3.2 Finding ¢ in the Mean Value Theorem Find te value oF values of that satisfy tho equation J@)~ fa) boa {in the eonelusion of the Mean Value Theorem for the functions and intervals in Exereses I -Fo ie L fee s2r—l OU) = 2 fay=e, [O11 spomeed [Ba] pod 88 15. Checking and Using Hypotheses - See ee eee is ciautie Salt eae ec ee ta ee A [=L.8) 6 fixbax, (0,0) at ° 9. ‘The function 20. jon=[fe 884 ae con ee ac » sais the hypothe Mean eT on the itv 10,217 max Roots (Zeros) IL, a) Plot the 20s of each polyromialon a line together with 24, the zeros of is fist derivative, ) ya2-4 ) yar eee es li) y=xt 3084 25, by) yore 4216 4) bb) Use Rolle’stheogem to prove that beiaen every two 22105 OE ag ye ebay ay there Hes re oF ne Dag Suppose that /” is continuous on (a, and that f has three ets inthe interval. Show that fas at Teast ne zero In (@. by, Generalize this resale Show thet if J > 0 throughout an interval (a. then f* has at most one Zero in fa, 6]. What iff <0 chroaghout fa, b] instead? +a Show that a cubic polynomial can have at most thee real zeros. Theory and Examples ‘Show that at some instant during @2-h automobile tp the car's speedometer reading ill equal the average speed forthe wip. Temporature change. I wok 4 see fora thermometer rise from —19°C 10 107°C when it was taken from a freezer and placed in boing water. Show tat somewhere slog the way te mercury was rising at exactly 8.5°Chec. Suppose that fis differentiable om [0,1] and ha is derivative is never ero. Spow that (0) f(1) Show that sind — sina| < [> a) for any numbers and & Suppose that fis differentiable om [a,b] and that f¢) < f(a. (Can you then say anything about the values of fon (a 6]? Suppose that and gare diferenable on [a,b] and that fa ‘(a and fb) = g(b). Show that there isa east one point be- toicen a and D wihere the tangents tothe graphs of f and g are parallel Let f be ditferentiable at every value of x aod suppose that FEU) = A that 7° < 0.0m (00,1, that "> O 08 (1,2) 4) Show that f(x) = I forall by Most /'(l) = 07 Explain. Let f(x) = ps! +gx-tr be a quadeatc function defined om closed interval (a, bl. Show that there is exactly one point ¢ in (g, b) at which fsatisies the conclusion of the Mean Value Thoorem, ‘A surprising graph. Graph the function (4) = sins sine +2) inte +, ‘What does the grap do? Why does the Fnetion behave this way? Give reasons for your answers Ie the graphs of two Functions f(x) and g() star atthe same point inthe plane ai dhe funetions have the same ate of change at every point de the graphs have 1 be identical? Give reasons for your answer. a) Ix decreases on every interval in its204 Chapter 3: Applications of Derivatives 1b) If the conchsion in (a) is relly ve, how do you explain the fae that g(1) = 1 is acwally greater than g(—1) = =1? 26, Let be a funeton defined onan interval [a,b]. What conditions ‘ould you place on fto guarantee tht wate boa where min f” and max f° refer to the minimums and maxienum values of /” on a,b]? Give reasons for your answer. E27. CALCULATOR Use the inequalities in Exercise 26 wo estimate FON if f(a) = 1/1 + x e082) forO = x < 0.1 and f(0) = 1. 28. CALCULATOR Use the inequalities in Exercise 26 to estimate AOA) i€ (3) = 1/0124) for 0 = x = 0. and FO) =2. 29. The geometric mean of a and b. The geometric mean of two positive numbers a and b isthe number Vab. Show thatthe value fofeinthe conclusion ofthe Mean Value Theorem for f(x) = 1/ ‘om an interval [a 6) of positive numbers is © = ab, 30, The arithmetic mean of a and b, The arithmetic mean of two rumbers a and b is the number (a +6)/2. Show that de value ‘of cin the conclusion of the Mean Value Theorem for f(2) = 2° om ny interval fu #) is = 4a 1/2 Finding Functions from Derivatives 31. Suppose that. f(—1)=3 and that 7°(3)=0 for all x Must {Fl8) = 3 for al x? Give reasons for your answer 32 Suppose that (0) = 5 and that f*(s} = 2 forall x Mast fla 2x-+5 forall x? Give reasons for your answer 3B, Suppose that f(x) = 2s foe all x: Find (2) if # f0) » su=0 9 fe) ‘34 What canbe said about functions whose derivatives ae constant? Give reasons for your answer. nin f= max f', In Exercises 35-40, find al} possible functions withthe given deriva 364) 36. a) » 9) ara) bye" oO yer » 8 3 a) » 8 ym sinze yoo 5 oy sin2r 4608 b ya © yavi-sect0 In Exercises 41-44 ind the fonction with the given derivative whose _sraph passes through the point P a £0) (0,0) 2 condemn % a rons-aco, Pf 44. (9) =secttane 70.0) Counting Zeros ‘When we solve an equation f(x) = O numerically, me wsully want to ‘now beforchand how many solutions t Took fr ina given interval Wah the help of Corll tw ean comets Bd on ‘Suppose that 1. Fis comtinows on (a, bj and diferentiable on (a,b) 2 “fha) and f(b) have opposite signs, 3 f'>Oon labor f <0 onia bh ‘Then fhas exocly one ero between a and b It cannot have more than one because itis ether increasing on {a,b} oF decreasing on le, bh. Yr it hes at least one, hy the Intemedite Value Theorem 15), Forexample, (2) =27 43+ I has exacly one zero 1) becouse is diferentable on (—1, 1}, f(-1) = —3 and have opposite signs, and f'(x) = 3x +3 > O for all x Fig. 323), fei.3) 3.23 The only real zero of the polynomial y= x! + 3-41 is the one shown here between —1 and 0. Show tha the functions in Exercises 45-82 have exactly one zero inthe given interval 48. fo)axe3e41, 2-1) 4 6. (= PFS 47 (-00.0)47, gy = Vi+VTFI-4, @, 00) ae Tah Cho 9.1 +6¢(Q)-5 cme 50, 1(0)=20 —cos! 0+ V3, (90, 00) SL. (0) = M008 F545, 9/2) 52. r(0)=tand —c —9, (0. »/2) 3.3. The First Derivative Test for Local Extreme Values 205 CAS Exploration 53. Folles origina theorem 8) Consievet a polyoma /() that has 2e05 a = ~2, O., and 2. W)Gush fan its Cavnatve f° eget how in what you see relied to Rolls original theorem? (See the marginal te on Role) ©) Do gts) sinx ond its derive gdlseate the same hooomenor? 4) How would you sae and prove Rol’ orginal theorem in Tight of what we know today? 3.24 A function’ fist derivative tells how the graph rises and falls. The First Derivative Test for Local Extreme Values This section shows bow to test a function's eritical points for the presence of local extreme values The Test [As we see once again in Fig. 3.24, a function f may have local extrema at some exitical points while filing to have Tocal extrema at others. ‘The key is the sign of Fin the poin’s immediate vicinity, As x moves from left righ, the values of f increase where > 0 and doerease where f” < 0 [Atte points whero fas a minimum value, we se that” < 0 on the interval Immediately tote left and f’ > O onthe interval immediately to the right. (the point i an endpoint, there is only the imervalon the appropriate side to consider) This means thatthe curve is fling (values decreasing) on the left ofthe minimum value and rising (values ineressing) on its right. Similarly, at the points where f hs o maximum value, J” > 0 on the interval immediately to the left and J” < 0 on the interval immesisiey othe right. This means thatthe curve i rising (values Increasing) om the left ofthe maximum value and falling (values deressing) on its sia These observations lead to test forthe presence of local extreme values. Ascte max F soto206 Chapter 3: Applications of Derivatives Theorem 5 ‘The First Derivative Test for Local Extreme Values ‘The following test applies to a continuous function f(x). At a left endpoint a: If f’ < 00" > 0) for.x > a, then fhas a local maximum (oinimum) value at, At a critical point ¢: 1. If f° changes from positive to negative at c (f'> 0 for x <¢ and f'<0 for x > 0), then f has a local ‘maximum value at ¢ sone vite + 4 role ap a eee i ‘At aright endpotnt b: @fio-0 () fiedundefined If f’ < 0(/" > 0) forx 0 for x > c), then f has a local ‘minimum value atc. eat reoypea bh _# @fo=0 (©) £0 ondefined 3. If "does not change sign at ¢ ("has the same sign ‘on both sides of c), then fhas no Tocal extreme value ate. Para pee foo ir>o { [reo @ fo~0 (©) Flo onefied EXAMPLE 1 — Find the citical points of Fo ‘Mentify the intervals on which fis increasing and deere Jocal and absolute extreme values. A ay, xa 4) ing. Find the function's Solution The function fis defined forall real numbers nnd is continuous (Fig. 3.25).9 3.25 The areph of (xample 1), cot) 3.25 The graph of gfe) = =? + 12x45, =3 sx 53 (Example 2), 3.3. The First Derivative Test for Local Extreme Values 207 ‘The fist derivative £0) is zero at x= | and undefined at x =O. There are no endpoints in f's domain, 0 the cttical points, x = 0 and x = 1, are the only places where f might have an extreme value of any kind, ‘These eritcal points divide the x-axisinto intervals on which j” is ether postive ‘or negative. The sign pattem of f' reveals the behavior of f both between and at the critical points, We can display the information in a picture like the following, 4 Sinotcty: 4 | 4 bt mers Signotes = 0) oumeene NON Z ‘To make the picture, we marked the critical points on the x-axis, noted the sign of each factor of /* on the intervals between the points, and “multiplied” the signs of the factors to find the sign of /*. We then applied Corollary 3 of the Mean Value ‘Theorem to determine that f decreases (,) on (—00, 0), decreases on (0, 1), and ‘increases (//) on (1, 20). Theorem 5 tells us that f has no extreme at x = 0 (F" does not change sign) and that fhas a local minimum at x = 1 (f? changes from negative to positive) ‘The value of the local minimum is f(1) = 1!°(1 — 4) = —3. This is also an absolute minimum because the Fanetion’s values fll toward it from the left and ise way from it on the tight. Figure 3.25 shows this value in relation to the function's raph, a EXAMPLE 2 Find she imervals on which —P4itets, 35053 ale is increasing and decreasing. Where does the function assume extreme values and ‘hat are these values? Solution The function fis continuous on its demain, [~: derivative (Fig. 3:26). Te first gs) = 3x74 12-3074) 3x +) —2), defined at all points of {—3, 3], is zero at x = —2 and x = 2. These critical divide the domain of g into intervals on which g’ is either positive or negative. We analyze the behavior of g by picturing the sign pattem of g’:208 Chapter 3: Applications of Derivatives Signet-3ie4: 4 | Simeta-a: | - é Signofgtei =-Me+2N8-D: = cmmingn INE NI FFxtrema’ local local local Yocal We conclude that g has local maxima at x = ~3 and x x= —2and x = 3. The corresponding values of g(x) = and local minima at 84 12e +5 are Local maxima: g(-3)=—4, a wi g(-2) ==, Since g is defined on a closed interval, we also know that g(—2) is the absolute ‘minimum and g(2) is the absolute maximum. Figure 3.26 shows these values in Local mi relation o the funetion’s graph 4 Exercises 3.3 Analyzing f Given f 11 hay = 428 12 hee) = 209 — 18 Answer the following questions about th functions whose dorvaives 13, f(@) = 39° — 40? 4 (0) = 0-0 sao civ i Rome 8 15. f(0) =3 + 16 16 hey = (+9 f8) Wht are the cic points of 7? 2 J aviag 1) Ona inal aoe er eenig? fede Nea a ata 6) Atoms fay cs fase mam nd ml 9, 1) a xy)= ues 1. f@) 2x1) 2payeu—n6+ — Bhe=a/ToaT 229) RP GY=O- WH APRON a, fey tian Me es Fl 6 £@)=6-Da+ DES) 3% sey=sa+D 2 gts) = +5) 2 fay 'P(e42) aye) » 2 Wx) = 1904) 28, ke) = E24) Extremes of Given Functions In Exercises 9-28, a » ° Ra 9. ety Extremes on Half-Open Intervals In Brevises 29-36 Find the intervals on which he fusion i incresing and de- . mee 1) Wenify he functions Toa exvome ales in the given domsn, ‘Then identify the funetion’s focal extreme values, if any, saying, and say where they are assumed. Teer mi 1b) Which ofthe exteme ales, i any, ate abso? ‘Which, ifn, f the ereme ales are absolute? 83-6) GRAPHER You moy with o support your ndings wih a gph GRAPHER You may wishosppont your ndings witha gaph- ‘na or computer rapes ing calculator or computer grapher ea 10. si) 324945 20,3.4 Graphing with y’ andy” 209 444, Lex coe PW 6r-9, -bex<0 @ sin, OSP Sn, tO=0and 82, bo) =Ssin tent, 12 288 48, Sketch the graph ofa differentiable Funetion y = f(x) through ne the point (1. 1) if 7°) = 0 and yi meets 9) ') + Ofer < Land (a) Ofer > 1s 3 ©) f'@) > 0fwrZl Me kaa bIe Lae4l, -oo 2, and g(x) > 1 as e+ 2 Zot gare ) gQi=2 y O for x > 2, and g's) > oe as > 2! BB. fox) =—2eosx eos, EES 46, Sketch ke graph of a cominaous Faction» = AC) such thst 2. f(ay=oe'a Pama, Os eH 9) HO) =0,-2 and W(3) > 80 8x > OF Theory and Examples 447. As xmoves from lft igh through the point c= 2, isthe graph ‘Show that the functions in Exercises 41 and 42 have local exteme of f(x) =x? = 3x +2 rising, ori it falling? Give reasons for lesa he given values of 8, and say which kind of local exteine Your enswer. the fnecon has 48 Find the atersls om which the function fs) ‘a # Osis increasing and decreasing. Describe the reasoning be- ° 41. m0) =30088, O<@< 2, 0=0 md O= OEE RSG. ASRSIE WE hind your answer. Graphing with y’and yx In Section 3.1, we saw the role played by te fist derivative in locating function's extreme values. A function ean have extreme values only at the endpoints of its domain and at its erica points. We also saw that eitical points do not necessarily yield extreme values In Section 32, we saw that almost all the information about 2 differeniable function is contained in its derivative. To recover the funetion completely, the only additonal information we need is the value of the function at any one single point. I a function's derivative is 2x and the graph passes through the origin, the function must be 22. I function's derivative is 2x and the graph passes through the point (0, 4, the function must be 2? +4, In Section 3.3, ve extended our ability to recover information from a function's first derivative by showing how to use i to tell exacly what happens at a cxtcal point. We can tell whether there rally is an extreme value there or whether the 219ph jus continues tose oF Fal, Jn the present section, we show how to determine the way the graph of a