RECOGNISING 7 Oxford Cambridge and RSA Examinations SIXTH TERM EXAMINATION PAPERS (STEP) MATHEMATICS I, Il AND Ill 2004 EXAMINATION HINTS AND ANSWERS STEP 9465-75/SUP/2I04OCR (Oxford, Cambridge and RSA Examinations) is a unitary awarding body, established by the University of Cambridge Local Examinations Syndicate and the RSA Examinations Board in January 1998. OCR provides a full range of GCSE, A level, GNVQ, Key Skills and other qualifications for schools and colleges in the United Kingdom, including those previously provided by MEG and OCEAC. It is also responsible for developing new syllabuses to meet national requirements and the needs of students and teachers. This mark scheme is published as an aid to teachers and students, to indicate the requirements of the examination. It shows the basis on which marks were awarded by Examiners. It does not indicate the details of the discussions which took place at an Examiners’ meeting before marking commenced. All Examiners are instructed that alternative correct answers and unexpected approaches in candidates’ scripts must be given marks that fairly reflect the relevant knowledge and skills demonstrated. Mark schemes should be read in conjunction with the published question papers and the Report on the Examination, OCR will not enter into any discussion or correspondence in connection with this mark scheme. © OCR 2004 Any enquiries about publications should be addressed to: OCR Publications PO Box 5050 Annersiey NOTTINGHAM NG15 ODL Telephone: 0870 870 6622 Facsimile: 0870 870 6621 E-mail publications@ocr.org.ukHints and Answers STEP I, 2004 2 Section A: Pure Mathematics i) (ii) ‘The quickest method is to use Pascal’s triangle: (8-+2V8)" = 27 +319)" (2v8) + 818) (2v8)" + (2v8)" = 27 + SAV + 180 + 40V5 = 207 + 94V5 but for a small power such as 3 itis easy to compute (8-+2V8)” = 29+ 12V5 and then multiply the answer by 3+2V5. You might like to consider how to calculate e.g. (3 +25)" (02) x0? xu It ¥90—70V3 = e ~ dy then 99 - 703 = (ce - dv)’, By Pascal's triangle = 99 ~ 70 = c! ~ 3dV + 3c (dv3)* — (dv32)°. ‘Equating rational and irrational coefficients = c? + 6ed? = 99 and 3c?d + 2d° = 70. These equations are hard to solve algebraically, but the question tells you that c and d are positive integers, and since you can see that c* < 99 and 2d < 70 there aren't many values of c and to try. You should quickly find that ¢ = 8 and d= 2: remember to check that these values satisfy both equations, efficiently, using the fact that u!! 198.4 VI08F=4 2 Notice how the discriminant is the difference of two squares: }98? — 2? = (198 — 2) (198 + 2) = 196 x 200 = 4 x 49 x 2 x 100 = (2x 7 x 10)? x 2. 198 + (2.x 7 x 10) v2 Using the quadratic formula <® x = 994 70V2 = r= 3-2/3 using the answer to part (ii).Hints and Answers STEP I, 2004 3 (i) Gi) (ai) Your graph should look like a staircase: if 0 r= —y+2or2 = —y, s0 the factors are x ~ (-y +2) and x — (9). Since 6y2—y? —21y+22? +122 —dzy-+2%y —Bry?+10 has no term in 2°, there must be a linear factor with no x term i.e. of the form py-+q. Checking a few values of y shows that when =2, the given expression = 0 => (y +2) (x + ay +6) (x +.cy +d) is a factorisation. Notice how you shouldn’t assume a factorisation of the form (z+ 6y +8) (x y +d) as it might be of the form (x4 8y +6) (x + 2y +d), but it is reasonable to assume that the coefficient of x in both factors is 1, since that will ensure a term 222, ‘The claimed factorisation is (y +2) (2? + acy? + (a+-c)ay + (b+ d) 2+ (ad-+ bc) y + bd) so equating coefficients implies that, ac=6 ad + be + 2a0 bd +2 (ad + be) b4d=6 ‘The first and sixth equations tell you that a = —2 and c = —3 or vice versa, and the fourth ‘and seventh equations tell you that b = 5 and d = 1 or vice versa. You need to use the second equation, which tells you that ad + be = ~13, to determine that the factorisation is (+2) (2 ~ 241) (@~3y +5).Hints and Answers STEP I, 2004 10 (iti) Tt may be instructive to find these factors using the quadratic formula, as described at the cend of part (i). ‘The derivative of sec z is secz tan, which can be determined by differentiating (cosz)~'. Using the given substitution, and remembering not only to write the limits in terms of t but, also to write dz in = of dt, 5 7 1 sect tant sect tant + at= f° at [ cos? t dt [aaa ite De y sect tVtant t t : jana, “Ua using the identities 1 + tan® ¢ = sect and cos 2t = 2cos?t — 1 Since (« +1) (z-+3) = 2? +42 +3 = (x +2)* —1, this integral in part (ii) looks similar to the integral in part (i). sect tant aAVETT Ker J sect eect iy ect rH sect tant = ata [dt=t40= arse (2 +2) +0 laos 2e=1 fi * Guede Sole 2+ 2— eects f —— plereaer (2 +2)? —9, let z +2 = sect. 1 je~ f Sete _ Ssect tant terete | EE os 1 ne (242 4 wore (242) +e Since 2? +40 —5Hints and Answers STEP I, 2004 6 Using the formula for the kth term of an arithmetic progression: A: value of kth term = 5k +1 B: value of kth term = 5k +2 C: value of kth term = 5k +3 D: value of kth term = 5k +4 E: value of kth term = 5k or equivalent. ‘Therefore, the sum of any term in B and any term in C can be written (Sk-+2) + (5n+3) = 5 (k-+n-+ 1) which is a term in 2. Notice that it is important to use k and n, to ensure that you ‘are not only adding corresponding terms in B and C. Similarly, the square of any term in B can be written (5k +2)? = 25k? +20k-+4 = 5 (SK? + 4k) +4 which is a term in D, and the square of any term in C can be written (5k +3)? = 25k? +30k +9 = 5 (5k? + 6k-+ 1) +4 which is also a term in D, (i) Since (5k)? = 5 (5k), (5k + 1)? = 5 (5K? +2k) +1 and (5k +4)? = 5 (5k? + 8k +3) +1, 2? ust be in B, A or D (including the results from above) ie. x? =5n, Sn +1 or n+ 4. But (5n+1) +5y =5(n-+y) +1, (5n-+4) + 5y=5(n-+y) +4, and 5n+5y =5(n+y). So 2? + a term in E cannot be a term in C. (ii) You know that 2? must be in B, A or D, so 2* = (x2)? must be a term in A or E, as all terms in E square to give a term in E, and all terms in A and D square to give terms in A. ‘Therefore 2y* must be a term in B or E. No pair of terms from A, B or E can add to give a term in D (consider the ”+4”), so the equation 2 + 2y* = 26 081 974 has no solution for integer z and y.Hints and Answers STEP I, 2004 7 (se=n)e-™ ‘The line joining A to the midpoint of BC has equation y ~ 41 Mga (BEBE) og ae cor ant) ‘The line joining B to the midpoint of AC has equation y — gz = | —-—*——= ] (x — Soin r cquation y= gy = (BERR (op "To find where the two lines meet you need to solve simultaneously these two equations. ‘The many subyripte malo i dif fo do tis correctly, tthe aera rogue tot vanced. You Pitmtps go tate) 3 should find that the lines meet at, ( : You then need to veri that these coordinates satisfy the equation of the li +02 ~ 2a inidpoint of AB, which is y ~ (ESS Je- . rome (SEEM) e-my joining C to the Bte , w-% If AH is perpendicular to BC then 2+% pepe PtP pape 1 since the product of the gradients of perpendicular lines is —1. 2 a-d=-(h- D> ata= ne If the line BH intersects the line AC at right angles => p? + 4? = p2 + q3. The instruction "write down” in the question suggests that very little work is needed to do so: see how this answer is stevtoraly densa the prev ‘Therefore, if AH is perpendicular to BC and BH is perpendicular to AC +g Spitd= pita} and pi +a? srt bat= rh gB nen Pot PL PP = CH is perpendicular to AB = - ah =~ (03 - Pi) > ‘This question aims to show some interesting properties of triangles. The line joining a vertex of a triangle to the midpoint of the opposite side (e.g. A to the midpoint of BC) is called a median: ‘you have proved that the three medians of any triangle are concurrent, ie. they alll meet at the ttt p atete) ‘This point is called the centroid ‘The line through a vertex which is perpendicular to the opposite side is called an altitude: the three altitudes of any triangle are concurrent, and they meet at a point called the orthocen- tre. In this question the triangle has been set up so that the orthocentre has simple coordinates (+ patra, a1 + a2 + 49). You should consider how this has been achieved: what geometric fuct must be true if the coordinates of A, B and C satisfy 93 +g? = pi + a2 = 98 4-027Hints and Answers STEP I, 2004 8 @ (ii) ‘The relationship ag = 2, a) = 7 and an — 7ay—1 + 10ay—p = 0 is called a recurrence relationship: given the values of ay and a you can calculate that az = 29, a3 = 133 and so on. Therefore f (x) = 2+ 7x + 202? + 13329 + => f(z) ~ Tel (x) + 102"f(2) = (24 72 +202? 418829 4...) = (Je + 492? + 20828 4 98124 +.) + (202? + 7024 + 20004 + 138025 + because all the other terms cancel. Q-%e etry) tales seta) Soe = Serger Shimer” = = 24 Tat) organ"? — de — STay42"*? +) 10a,2"*? (if this is unclear, expand the first few terms of each of the sigma). = 12+) (p42 — Tyg + 10a,) 2°? = 12+ (027? because of the given relationship a, — 7a,_1 + 10a,—2 = 0 for n > 2. Tad Tas en by partial fractions. + 1-72 +102? ~ (1-32) (1-52) Since (1—2r)"! = 1422 +42? +823 +... +2"2" + and (1 — 5x)! = 1+ 5x +252? + 12503 +... + 5%" + the coefficient of 2" is 2" +5", therefore ay, = 2" +5". Be In this part of the question you haven’t been given the precise relationship between the terms of the sequence by, so you need to determine that first. n= 2 ba = ph + aby = 40= 1p +54 n= 3. by = phy + ghy = 100 = 40p +109 0 D1, = 6 by byt ~ Olga Part (a) suggests consider g(z) ~ xg (x) ~ 622g (x) = (5+ 102 + 4022 + 10029 +...) — (52-4102 + 402° + 10024 + ..) (800? + 602° + 24024 + 60028 + ..) = 5450 ‘This is sufficient, but you should try to write a convincing argument using sigma notation as above. Stor B+: ay 1 6x? ~ (1+2z)(1—32) 1422 "1-3 3 w(a) = (1~ 2b da? ~ 80° + + (-2)" 2" +n) $4(14 90+ 92? +2729 + .9"2" 4.) = bn = (-2)" + 4(3)" a(x) =Hints and Answers STEP 1, 2004 9 @ (i) It is important to realise that it is not sufficient to choose a specific value of zo ef t0 = 4, evaluate 2; = 4.4 and then claim that the sequence is increasing for all n and for all x9 > 3. When asked to prove that a > 6, a good strategy may be to consider ab, and try to see why a — b is positive, In this problem, therefore, consider ng = RES ha San 6 _ (en — 3) (En —2) 5 5 5 which will be positive for any aq > 3 (or any x, <2). Let n =0: you are told that 29 > 3, 80.21 ~ 19 >0 = a1 > 29 = 21 > 3 since 2 >3 na Sa -m>03m>% my > 3 since 21 >3 => 23 ~ a2 > 0-9 2 > 22 and 90 on. ‘Therefore an41 > tn for all n. As in part (i), consider 6 5Yn — 6 — nm Ua which is positive for any ya between 2 and 3. Let n=0: 2< yy <3 —yw>04 m1 > wo. 6)_6 Also 3, =3- (5-5) = 5 -2> since w <3. » ( ry) ® ‘° 33-n>05n<3 ‘Therefore 2 < yo <1 <3,80 92-1 > 0 w2> H Also 3-92 a- (5-5) $0 2< m1 < v2 <3 and 50 on. ‘Therefore, yas > tw for all n, but yy <3 for all n Unsth — Ya 6 Sas osincen <3 uw e Both of these arguments are most easily written using Mathematical Induction: if you are familiar with this method of proof, then you might like to prove these two results in this way.Hints and Answers STEP I, 2004 10 Section B: Mechanics At the maximum height H, the vertical velocity = usin? ‘Therefore O = (usin8)? ~ 29H => H = “5 At the point P, utcos@ =d and utsind — bgt? = $d a d__gf_a \_ gd? sec? 8 = Fe usinax 45 (4) = dtano ~ Eee op ar a see? gdsec?@ _ sin?@ 9d © Btand— 0 ‘an 29 because the projectile doesn’t hit the roof of the tunnel and therefore its maximum height is less = 10sec? @ sin? @ < 18(2tand— 1) => 5tan?0— 18tand+9<0 = (Stand — 3) (tand- 3) <0 a cunscs = arctan } < 0 < arctan310 Hints and Answers STEP I, 2004 ul a i) ‘A good velocity time graph makes this question much easier to understand. 1 v? =u? +.2(8a) (dy) and 0 = v* — 2a(da), = Dad, =u? + Gad = Gad; <2adg because u? > 0 3d < dy A velocity time graph is very helpful here. ‘The partite tine 3 travels a distance of =, You should not calculate ths distance using v? = u?-4-2as because 8 is the displacement of the particle not the distance it travels, but instead should look at the (positive) area of the relevant triangle on the graph. to accelerate from initial velocity u < 0 to rest, in which time it ‘Then the part le travels a distance D where v? = 0-4 2(3a) (D), so that D = =. vou du-u If u > 0 then the total time taken by the journey = =" u> 0 u= lu) . total distance “ot +2 du?@—u? total time aoe ~ t= [uy vtlul |v dut Bata Ba Note the use of |u| to ensure that you are adding two positive quantities: the expression v-+ 1 is actually a subtraction if u is negative. ‘Therefore, average speed = If u <0 then total time taken by the journey te tae? average speed = “te > average specs (ot Te As in question 8, you can consider the difference between these two: do? a2 du fu? (dv? — 02) (dv + ful) ~ (40? 4 02) (40 2(4v— ful) 240+ 2 (4v = ful) (4v + 820+ 807ful_—__Boful (v— ful) ane (av = |ul) (4v + [ul) |) (40 + [ul) e alaa Hints and Answers STEP I, 2004 12 A good diagram is essential here. Let AB be the left ladder. Let the normal reaction at A be R, and the frictional resistance be F, < Rt recall F; = 1R only if the friction is limiting e.g. the ladder AB is about to slip - and you don’t know that this is the case. Let the normal reaction at, C be S, and the frictional resistance be F; < 4S. Let AB = BC =1 ‘The easiest strategy is to take moments about B, and therefore not worry about the nature of the reaction force acting there between the ladders. ‘Taking moments about B of the forces acting on BC L tv3 1 3B 28x 5=aW x Gt ex 2 ‘Taking moments about. B of the forces acting on AB L 1 W3 gt hey 1 2 Rx Sa Wx Gt TW x Fx ‘Therefore, $= 2W + F,V3 and 6R = 31W +6F, v3. Resolving vertically all forces acting on the system => R+ $= 12. Resolving horizontally all forces acting on the system => Fy = 6R-31W 6 = 65 =6R-19W > S-W => RW —6R=6R-19W ow 53 + R= SF ands => 20W V3 =A BC sips as $< Ry and p= Fi = 2008, If the reaction between the two ladders at the hinge B is included, itis best to consider a horizontal and vertical component of the reaction force acting on each ladder. ‘There is no reason to assume that the reaction force will act purely horizontally: not all the forces act symmetrically on the ladders. The horizontal components will be equal in magnitude but in opposite directions, and the vertical components will be similar. Let the reaction at, B acting on AB be X horizontally (to the left) and Y¥ upwards; let the reaction acting on BC be X to the right and Y downwards. It would be instructive now, for example, to take moments about A and C, and compare the merits of the two solutions.12 Hints and Answers STEP I, 2004 13 Section C: Probability and Statistics ‘A tree diagram is very helpful, if only to clarify the language of the question P(inade by A and perfect) P (perfect) (made by A | perfect) = le = - > prog at PHAN 24 a OAS Bog Now, P (made by A | declared to be perfect) = : + pap+ JAC -p) _ Pet DO-v)+T0-)a+ 7-0-4) since those that are declared to be perfect are either those that are perfect. and have been correctly assigned (with probability 3) or those that are imperfect but have been incorrectly assigned (with probability 3). = 15Ap + 5A (1 — p) = 6p + 2A(1 — p) + 6(1— A) +2(1 A) (1-4) 4g+2 p+ igs13 Hints and Answers STEP 1, 2004 “ @ (ii) P (maximum of 3 numbers < 0.8) = P(all three numbers < 0.8) = 0.8° = 0.512 Similarly, F(2) = the cumulative distribution function of X = P(X <2) =a = f(z) = the probability density function of X = F’(z) = 32? = B(x)= [3 ae= A similar argument would work to calculate E(Y), where ¥ is the minimum of the numbers drawn. If Ho is true, then P(X < 0.8) = 0.8%. We require the minimum N such that 0.8% < 0.05 since then Ho will be rejected. BY" 1 aw 210% aw 10" (5) 13 4N=14 Ifa =0.8 then P (Ho is rejected) = 1 Ifa=0.9 then P (Ho is rejected) =4 Hints and Answers STEP I, 2004 15 io (i) (iii) The first pirate will have some gold if he does not take a lead coin first. i n = P (first pirate has some gold ) = =" ‘There are at least two different approaches that can be used to answer this part. The easiest argument, if you are familiar with it, is to see this as a problem about the number of ways of rearranging a selection of objects. You may recall that there are m! ways of rearranging m different objects in a line. Here there are n+ 2 coins which can be rearranged in (n +2)! ways, but the second pirate wants the two lead coins not to be adjacent. It is easier to count the number of ways to rearrange the n+2 coins so that the two lead ones are adjacent: there are (n+ 1)! x 2! ways of doing this, because you treat the two adjacent lead coins as a single item and therefore rearrange the remaining n+ 1 coins, but the two lead coins ean be swapped around 2! ways so that there are twice as many rearrangements. Hence there are (n + 2)! — (n + 1)! x 2! ways of rearranging the n + 2 coins so that the lead coins are not adjacent - (n$2)-M8(n4 1)! (m+) (n + 2-21) = P (the second pirate has some gold) = “TPE = (ntl xn n “tx (nti nye An alternative argument is the following. If the second pirate has some gold coins, then the coins must have been taken from the chest in the order LG or GLG or GGLG or GGGLG ete, until GGG..GGLGL, where G represents taking a gold coin and L represents taking a lead coin, You should consider why the orders LGG, LGGG, LGGGG ete do not need to be listed: a tree diagram will be helpful. ; = (2x0 ny 2 ynet ‘Therefor, P (the second pirate ha ome god) = (25 a i) (ats s*anS ) non-1 2 na? n a2 2 14 (Sax) nel on BGR OT = TEDHED MPO DEO) bo HY (nth) armen “a If all three pirates have some gold, then a gold coin was chosen first, a gold coin was chosen last, and when the remaining n coins were chosen the lead coins were not adjacent i nm-1 n=? + P{Gll three pirates have some gold) = <7 x Bo B=? because the probability that the last coin is gold must equal the probability that the second coin is gold: once the second coin has been drawn it can be renamed as the last coin, ‘The third fraction has been deduced by adapting the answer from part (il) 2 342 mFl “ned” npn 4d ‘This simplifies toOXFORD AND CAMBRIDGE AND RSA EXAMINATIONS STEP MATHEMATICS PAPER 2: 9470: JULY 2004 : HINTS AND ANSWERS QU) Put the terms with radicals on one side and the terms without on the other and square. Repeat this strategy (8) and the equation 24 ~ 6" + 9, of (*) are c= 0,1, 4. de = 0(+) will be obtained. The roots Squaring may introduce spurious roots, so these munbers must be checked to see that they are roots of the original equation, In fact, they are. (i) Application of $ again leads to (*). Checking shows that 2 = 0, = 1 are roots of the second equation but that x =4 is not. (iii) Again application of $ leads to (*). Checking shows that 2 = 1, 2 = 4 are roots of the third equation but that 2 = 0 is not. Q2 Write Q= 1 alr] +2 = [lx|-a/2)? +2 at/A sa < 2V9= 2-0/4 >0 + Q > 0 for all x It is therefore unnecessary to consider x > 0 and x <0 separately and even more unnecessary to tse caleulas methods eit =3 then Q= (lel ~ Die (e:- ¢ The solutions in 2 of the equation Q = 0 are of the form —22, —x,, x1, 42, where 0 < ay L, where {(@) = acscO-+bsec8. Setting J'(0) =0 will then lead to the required result. (ii) Here, for all @ € (0, x/2), itis necessary that y > , where y is such that b = (y—r) cos -+wsin and 2 ig such that a = rsin@ + weos¢. (Other formulations are possible.) Elimination of « leads ob — 2wese 2p Setting y'(d) = 0 plus some further working will then produce the required result. to y = wesc g + bs Q5 Using the integration by parts rule Sil vooszde = -2. is easy to establish the results ff xsinadx = x and ‘¢ Write sin(x +) =sina-cost + sinécosz and the result f(t) = ¢-+ Asint + Beost, where A and B are ns defined in the question, follows immediately, ¢ Hence write t+ Asi £4 Boost = t+ ff (e+ Asin + Beosa)sin(e +4) de (++) so that as J esin(e +t) de =... = neost — 2sint, ff) sinsrsin( +1) dz =... = (w/2)c08t, JG cosxsin (x +t) da =... = (n/2) sint, then, by considering the coefficients of cost and sint on both sides of (***), it follows that 2+ (n/2)B, B= + (n/2)A> A Alternatively, eq A= 2, B=0. tions for A and B can be obtained by putting t= 0 and ¢ = x/2 in (***). Q6 From the data it follows that the component of b in the direction of a is Sa Hence p = 4a and q = b= Sa « Agnin from the data, lal? Gab +9a.a {(ea)/lal’] a= G/2)b ~ 6/2)0 i follows that (¢.a)a = ~2a and bb 5-18 +9 = 16 [ql =4, so that.‘Thus P = 2a, Q = -(9/2)a+ (3/2)b, R= (7/2)a - (1/2}b +e Q7Goorl sketch graphs of y= © and y = 2sin.x, in the same diagram and aver the interval (0,7), will readily show that the equation f(2r) = 0 has exactly one root in the interval (+/2, x] #-f(3n/) = V2—3n/4 has the same sign as 2— 9n2/16 = 2— 45/8 = -29/8 < 0. Hence as J{n/2) =2—m/2>0 and Fn) In/2, 3/4) = sina /$ + 2oyT =H = sin Ba/A = 1) V3 = Bet ~ 82? + 0.85. (+8). ‘The sign of f(5/8) is the same as that of Ax” ~25n?/64 = 17/5~125/32 = [x/2, 5/8} <0, then Jy O(+) a? = 1/2-41/(2V3) 1/625 <0. Hence fy # A good approximation to x = sin9x/16 may also be obtained in a similar way. In fact, it will be found that f(9x/16) > 0 so that Js = [9x/16, 5%/8} = A=0. Thus SG) Integration leads to the general solution t = A—In(1~ 2) and 2(0) = nai- Gi) Obviously, (I= 2)? < (L42)"/? for all x € (0,1}. Hence multiplying this inequality through by (1 =)" leads immediately to the required result “Arguments which go in the wrong direction, eg 12 < (I-2?)'?? =... a~a > 0, ete,, are invalid. It may be possible to salvage them by replacing ' =#! by‘ <=" = y+ B and hence to t = sin” In the case n = 2, the substitution « = siny will lead to a) +B 0 B=0-2=sint. as the general solution, In particular, 2(0) Note that the question does not allow the use of the standard form f (1—22)-1/? d= sin“™(x) + an athitzary constant, without proof (iii) 1G, is the graph of x for 0 a2 = 4. )- ar - TP S38 (2460430 46741 =04... 47 =14(8-VOr. 2g If the acceleration of P down the slope is az ms, then —4g( Hence if P and Q meet at time 7, then 3 — 2(r — 1)? = 6(r — Note that the condition 7 < 1+ /372 ensures that the collision takes place before P returns to A. (ii) A possible solution is first to show that T= 14 JIB +r =2. , vq(2)=2V6 ms-* then the total KE at t= 2 of P and Q = 80), ‘a Hence as vp(2) = 4 ms Further, gain in PE at = 2 since start of motion = 40 j so that energy lost due to friction = M44 = 80-40 = 24 j Alternatively, and more directly, the work done against friction up to the moment of collision = Aj. frictional force opposing motion of P (or Q) x6 j = 4x6 QII (i) At full engine power, the equation of motion of Ais Po!!? — kv = m(dv/dt), ‘The result f 1/(Pu!/? — kv)dv = —(2/k)In(P = ku'/?) + constant, together with use of the condition v(0) = 0, followed by some algebra will lead to v4 = (P?/R2) (1 — e~#/2)? («), where tug is the velocity of A at time t ‘To obtain up, the velocity of B at time t, substitute 2m for m and 2P for P in (*). ‘Thus op = (AP#/K2) (1 — eH/8m)? (i) 94 = Avg = 9 (1 — eH HE)? = 16 (1 — eH)? = 9 (14 eI)? = 16 Se N/m = 1/3 = vq = 6APP/BLE and vp = 16P?/9K2 (iii) The equation of motion of A is now m(dv4/dt) = —kva, where t is now measured from the instant at which the engine of A is switched off. Since the velocity of A at the start of this phase of the motion is 61P2/8LA, then s vv, result vy = (16P2/9K2)e~#/2 will be obtained. Elimination of t will then lead to ku, = 4P?vq (64P2/81K2)e-/™, By a similar argunnent the Q12 ‘This question generates seven separate tasks and so it is especially important to set out responses in an orderly way. # The sketch is unimodal and falls entirely in the first quadrant of the x ~ y plane. In particular, y'(0+) > O.and y is asymptotie to y = 0 as 00 4For f(x), the constant k is deterinined by fj) kre—a2? dr = 1=> ... = k = 2a/ (178) ‘© For the mode, note first that f’(x) = k [1 - 2a] * which is zero when x = \/T/2a. Asa< 1/20 = VTda> Land f'(2) > 0 for any 2 € (0,1), then in this case m= 1 On the other hand, « > 1/2 => y/ © To determine h, set F(h) = 1/2, where F(x) = fj f(y) dy. This leads to k/2a —(k/2aje~e"* = 12>... oh= VT/Qmp7A +7) a> —hn (207? = 1) eM < L Fe“4) Ho hm. € [0,1] in which ense m eo bec 1 te 1 < eV > Ine“? = 1) < -1/2 > — n(2e-/? = 1) > 1/2. © PX > miX < W)P(X < h) = P(X > maX PIX > MIX < fh) = [12- PV Ba) /(1/2) = 1 = (k/a) [b= eV] =. = eV = * = YL = Q1318 W,, pounds is the gain from draw n, then E(Wy.41) = (b-7—n)/(b—n) x14 /(6—n) x(n) which is zero if m= (b~ r)/(r +1) =& say. © Wass increases as n increases for n < and War decreases as m increases for n> € Hence W,, maxin m when n= [f] +1 =e, say, so that optimal stopping n is 1¢ # For r = Land beven, me = 6/2, in which ease P(first me~1 dram are all white) = (b=m-+1)/b = 1/2-+ 1b, ‘Thus expected total reward = (1/2 ~ 1/b) x 0+ (1/2 + 1/0) [(0/2)/{(b/2) + 1)] x ne =... = bf pounds For r= Land b odd, no =/2 + 1/2 s0 that now P(first ne — 1 draws are all white) = 1/2-+ 1/0. Hence expected total reward = (1/2+1/2b) x {(b/2—1/2)/(b-+1/2)] x (b4+1)/2 =... = (b= 1)/4b pounds. QU The introductory result may be explained by means of a diagram, Alternatively, replacing B by BUC in P(AUB) = P(A) + P(B) ~ P(ANB) will lead to the displayed result: almost innmnediately. # P= Plat least one pudding contains no sixpence) = 3{(2/3)" ~ (1/3)")- oP, BL/81 > 1/3, Pa = 7/27 < 1/3 > min(r) = 6. ant that ca © With r = 6, let A be the e 1 pudding contains > 1 sixpences and let B be the ‘event that each pudding contains 2 sixpences. ‘Th = 7/27 = 20/27 P(ANB) = PUB) 10/81, P(BIA) = (10/81)/(20/27) = 1/6, P(A)STEP III, 2004 Hints and Solutions Section A: Pure Mathematics ‘The substitution u = cosh x should suggest itself (because of the factor of e numerator), and the resulting integral can be tackled by splitting the integrand into part fractions: _ er duit re 1 Ly fo Qeos’a-1 J) P= 12 2f, Vu-1 VQut1 cosh, Vicosh v241)) wg ln (v1) (vn) f= a (n (ea) +e (HE) Similarly, substituting u = sinh:r, and then recognising an arctan integral: [tacts [Ocoee Sy lose (08) SP = stn (Vn) ‘To show that inh int @ sinh fo note that a cosh? = sinh? x + 1, so 2cosh?x — 1 = 1+ 2sinh? =, and the integral required is t second minus the first of those calculated earlier, as a —» 00. b as.a—+00, sinha 00, so arctan (V2sinha) —» 5 as a0, cosha ~+ 00, 50 ee “ V2cosha +1 V2cosha+1 (143) and sinha = b(u-3): a fs (2-4) Ty pees [rae 143 eo 7 1 v24+1 » [t= aH an (Hh t): Vcosha=1 oy and in (as) a Substituting u =e, so that cosh x = dr f cosh x — sinha fo 1+2sinh?«STEP II, 2004 Hints and Solutions 3 (i) (ii) it) Inspection of the denominator shows that the vertical asymptotes are at x = 0, x = 4, and the third term in £(2) tends to zero as |z| + 00, so the oblique asymptote is just yarn, 16 (2e-+1)? #2 (e—4) there is a double root at x = —} and hence the asymptote touches rather than crosses the curve at (—},—$), so is a tangent there. = 0 or (2¢ +1)? = 0, henc ‘The oblique asymptote meets the curve wh £(c) = 0 when 2°(x — 4)? — 16(2e + 1)? =0. ‘The left hand side of this equation is a difference of two squares, so factorises to give (x(x — 4) — 4(20 + 1) (x(a — 4) + 4(2x + 1)) = 0; that is, (2? — 122 - 4) (a +2)? =0, which has a double root at On your sketch you should show: the double root at (—2,0) — the curve has a local maximum here and touches the x-axis; the remaining roots (solutions of x? — 12x — 4 = 0) at « = 64 2V10; the curve approaching the oblique asymptote y = «—4 from below as zr —+ oo, approach- ing it from above as x + ~co and touching it at (— f(x) + co as x + 0 from above or below, f(z) + too as x — 4 from below and {(«) + 00 as 2 4 from above; local minima at some x value with 0 0 and at some x value with -2. y > 2-4 — note that this second minimum is not at the point of tangency with the oblique asymptote.STEP III, 2004 Hints and Solutions 4 ‘The sketch should show a curve with increasing gradient: because the gradient is increasing, the curve lies below the chord joining (a,f(a)) and (b,£(#)) and above the tangent to the ath - (ath curve at, Ce. £(S5-) )- The illustration is clearer if (1) > 0 for a 0, a =n ~1 and b = n to get the quoted result. ‘Take the sum from n = 2 to co of each term in the inequality: the left hand sum is directly as quoted; in the middle sum, you need to notice that it telescopes, so that all the terms except the first cancel in pairs; in the right hand sum, each reciprocal square occurs twice, cancelling the factor of 3, except the first. Ct er p22 nt (nyt) ~ 2\n? SF Finally, combine the two previous results to get: ; 1 For the next part, observe that (ay dial dlliia a(t tpt) qa then 2 (s - z rearranging these inequalities gives i the required bounds on S.STEP II, 2004 Hints and Solutions 5 If circle m has centre Oy then OOn and OOn — OOn41 = a + Pag + Or sina sina Substituting and multiplying by sina gives ry —rng1 = sina(rn + Tny1) which simplifies to the required result. 1=sina T+ sina. armor" \ \iysina ) T¥sina. t-sina\? which is almost a geometric series with common ratio G arnt ) 180 ‘This result then implies that ra = ( ) ro, 80 the total area is (l+sine)? 1 _l+sinta 9 dsina 2)" “Asin “7 l-sina\? 2 - (RS) Area T of triangle OAB 5 cosa (1+sin?a) = 7 cosa (2—cos?a) 6 Ss By differentiation, the maximum = occurs where 2 — 3cos* a = 0 (not sina = 0) and equals 7 a 92 22> 2 fis. [16 4V3 3) 3V37 V3 24 25STEP III, 2004 Hints and Solutions 6 If cos(z — a) = cos then x ~ a however, for example, x = 7, @ cos(a — a) = cost = -1 #1 = cos. Qnm +B sox = a+ B+ 2nm so tanz = tan(a + A) B = O has tanz = tanm = tan0 = tan(a +) but a Writing cos — 7sinz = Reos(x — a) requires R = 50 = 5y2 and tana = -7, so cos — a) = coe, where cos = —, so we enn take tan = 1 tana+tang _ -T+1 i-tanatan§ TE7 The frst of these gives «= Jw or x= $r-+w (since arctan j = arctan $)and the second x =w or r = +w. However, the first solution in each case does not satisfy the original equation (both have sin > 0, so cosz —7sinz <1), 50 r= $x+worn+tw. Hence tan = tan(a+ 8) = b proceeding as in (i), cos(x — a) = cos f, where tana = } and R = 55, 80 cos 8 = 2 v5 7 aa u and so tan = }. Hence tana = Hor 2 eu 73 2tanw a Notice that tan2u = 73> so the solutions are x = w and « = 2u, again eliminating the other two possibilities, w + and 2u +7, by checking in the original equation, Fy = wat wn mw, —w) 9412 = = (Un — Ung) (Wn + Waa ~ Wns.) +) wg An 1 (Un Wn-2) (i) Let ty be tin; then tin +2in-2 — Aint =0, 80 Fy — Fy =0 for n > 2, by (+), 3soF,=-3 forn2>1 but Fy = 0} +3 — dust = (i) In this part, let wp be en. (a) o2+1=4 33 (1-2)? = 04 m= Fa 3 forn>1 2 + vy — d0ntn—t Op — tn2 = 0 OF Yn + Un — AUnt = 0, for n> 2, by (+). (b) Since 1,2,1,2,... satisfies vq —vy_2 =0 forn > 2, Fy is constant, by (+) and since vw = 1, v1 = 2 that constant is ~3, so the sequence satisfies (+). for odd nn >2 for even n > 2, s0 Fy is constant, by (+), and since (c) The sequence 1,2,7, 2,.. with period 4, satisfies ty — un-2 = and Un + Un-2—4Un—1 ‘up = 1,01 = 2 that constant is —3, so the sequence satisfies (+).STEP III, 2004 Hints and Solutions 7 Integration by parts gives tor= al, 80 Ty > Ung = 1 Since Je = fe —Irsiy Yo cpp = (Ih ~ ba) + (a= ba) +--+ Un ~ Iss) =~ Igy and a q oa hh h ee i= ff ppet= 9, cond De tan 21 21 Vi th< Le a rus y? then F (x)y® + ele) = 2 (au + 2g(z), which is a linear differential equation for u(z). a , 6(2) = —1 80 the differential equation is = 3-2 Ldu Qu Me ae a In this ease, f(x) 2 ‘The integrating factor is ef 247 = 22 = 4 g cx? +2. +ooru ‘The solution curves which pass through (1, 1), (2, 2) and (4, 4) are y? +(x —1)? = 1, y? = 2x and (x + 2)? — 2y? = 4 respectively. In drawing these curves it should be made clear that all of them pass through the origin, and that this is their only point of intersection; that the first is a circle with centr (1,0), the gecond a parabola and the third an hyperbola with r+ centre (~2,0) and asymptotes y = 7°,STEP III, 2004 Hints and Solutions 8 Section B: Mechanics Let angle AOM be 26, so APM = 8, and let R, F be the normal reaction and frictional forces of the hoop on the mouse. ‘The forces on the hoop are its weight, the force on the hoop from its suspension, and the reaction on the hoop to the forces R and F of the hoop on the mouse. For the hoop to be in equilibrium, the net moment of these forces about the point of suspension must be zero, but the lines of action of the weight of the hoop, and the force on it from its suspension, pass through the point of suspension, so have zero moment about it, Thus equilibrium of hoop requires the net moment of the reactions to Rand F about the point of suspension of the hoop to be zero; that is, PF x PMcos6 — R x PMsin@ = 0, or F = Rtané. For the mouse (of mass m, say) to have constant speed 1, its equations of motion are: 2 resolving radially inward, R~mgcos20 = """ and resolving tangentially, F ~mg sin20 = 0 2 Combining these three equations gives mgsin20cos0 = (a 08 20+ =) sin which reduces to u? = ag using the double angle identities. ‘To maintain a speed u with u? = ag requires R = mg(cos20 + 1) = 2mgcos?@ and F = mg sin20 = 2mgcos?@tan@ which is greater than :R if @ exceeds arctan j. and hence angle AOM exceeds 2arctan 1, $0, initially, the hoop will begin to rotate in the opposite sense to the mouse’s motion round the circle.10 STEP III, 2004 Hints and Solutions 9 (Ja-2) _, (w~6a) _ 4g ba o=a(1-6(4)) Under Hy, X has a Normal distribution with mean j: and standard deviation © p= PUR -n<9=1-2(1-0(¢)) -20(¢ so is independent of n. p<005 + & (22) < 140% Lo505 = 8 < 0.068. a a 2 <005 24> 1,960. coz. 1.96009 For these both to hold, we must have 0.063 > —* > 1OCR (Oxford Cambridge and RSA Examinations) 1Hills Road Cambridge CB1 2EU OCR Information Bureau (General Qualifications) Telephone: 01223 553998 Facsimile: 01223 552627 Email: helpdesk@ocr.org.uk www.ocr.org.uk For staff training purposes and as part of our quality assurance programme your call may be recorded or monitored ‘Oxford Cambridge and RSA Examinations is a Company Limited by Guarantee Registered in England Registered Office; 1 Hills Road, Cambridge, CB1 2EU Rogistered Company Number: 3484466 ‘OCR is an exempt Charity ‘OCR (Oxford Cambridge and RSA Examinations) Head office Telephone: 01223 552552 Facsimile: 01223 852553 @0cR 2004