ABSOLUTE PITCH-BOTH A CURSE AND A BLESSING

DANIELJ.L EVITIN

Absolut e p itch is not a medi cal condition. There is no descrip tion of it in the DSM-IV (the D iagnostic Statistical Manu al of the American Psychi atr ic Association) . Clinic ians

are neith er besieged by patient s complaining that they have it (and want to tid themselves of it) nor by those who lack it (and deman d some med ical int ervent ion in orde r to acquire that which they consider to be their musical , if not neurological , right ) . Absolute pi tch has no obv ious societal or evoluti onary advantage . Yet it has remain ed one of the most pers istent pu zzles in cogn itive psychology for over 100 years, and it has received renewed inter est from cogniti ve neuroscientists and geneticists hoping to better p inpoi nt its etiology. It is wid ely regarded by many as representin g th e pinn acle of musical abilit y, and by others-includin g some who possess it-as a bother. MOZART had it, HAYDNdid not ; RICHARDCARPENTER has it , PAUL MCCARTNEYdoes not .

N otwith standing the above , many of th ose who possess absolut e pit ch do exper ience tang ible di scomfort , disori entat ion , or anxiety when their pit ch pe rcep tion shifts , often varying with wom en 's menstru al cycles , as a result of certain ph arm acologi cal int erventi ons (carbamazepine in particul ar) or simp le aging . For an und erstandin g of this dis com fort and its p roper perspectiv e, it is useful to define what absolut e pitch is, dema rcate what it is not , and explore it s possible origin s.

Absolut e pitch is generally defined as the ability either to zdentijjJ the chroma (pit ch class) of an isolated tone usin g labels such as C, 261 hert z, or Do or to p roduce a specified tone (e.g ., by sing ing or adjust ing the frequency of a vari abl e ton e generato r) and ro do so wit hout reference to an external standard (l -3) . The identificationhas been du bbed "passive absolut e pi tch", while the reproductionhas been dubbed "act ive absolute pitch" {4}.

W hen someo ne with passive absolut e pit ch hears a car horn , he or she might say, "That 's an E-flat !" In contrast , if I press a key on a pian o and ask som eone witho ut absolut e pit ch

124

M USIC M EET S MEDICINE

what I played, th e per son cannot tell me (unless th e per son watch ed my hand). People with passive abso lu te p itch can name a not e they hear, and those with act ive absolute p it ch can even do the reverse. Ask them co pro duc e a middl e C (the cente r key on a piano keyboard), and they will sing or hum or wh istle th e p itch for you . T hose with both kinds of absolut e pit ch have me mory for th e actual pitches in songs, not just the p itches in relat ion co one another. In fan , when most of them hear a song in a different key (and the refore with di ffer­em pitches ), it sound s wrong CO them , a source of anx iety. An analogy is if, due to a strange neuro log ical condi t ion , you went co the produce mark et and suddenly all of rhe bananas were blue and th e parsley was candy -appl e red .

Wh eth er or not th ey possess absolut e pit ch, some people are able co discern th at a fa­mili ar pi ece is played in the correc t key or can sing a fami liar song in the correct key. N ote the parallel here between th e passive and act ive absolut e pit ch describ ed first: recog nizin g the key of a mus ical pi ece is passive and repro ducing a mu sical pi ece in th e correct key is act ive. Because some people di splay the se abil it ies only with respect co musical pieces- a form of key memory we could say- and the y do not exhibit such memor y for indi vidu al cones, it is useful co distin gui sh between p ieceabsolutep itch and tone absolutep itch (4 ).

Some confusion and misconcept ions exist in the lit eratur e about whet her absolut e pitch involves mor e highl y developedp ercept ual mechanism s, whereas the preponderance of eviden ce is that absolut e pit ch ability is an ability of long-termmemOlJIand linguisticcoding(5, 6). Anoth er source of confusion is th e frequent use of the term peifectpi tch. To many people, it impli es a "perfect" mode of perception or a near-per fect abil ity to dete ct out-of- tune cones. I find th e term misleading because, in fan, absolut e p it ch possessors do not p erceive pi tch any better than th ose without thi s ability (6-8). Absolut e pit ch possessors typi cally tun e p itches to within 20- 60 cent s of target frequen cies (where 1 cent equals 1I100th of a sem iton e) (9) . In passive casks, they regul arly make sem itone and octave errors (9-1l) . Clearly, then , th ere is nothi ng "perfect " about absolut e pitch; it is simply th e ability co p lace or prod uce ton es with in nomin al cat egories.

T he ability co recognize and ident ify absolute pi rch presents th e research scient ist wi th two opposing puzz les. First , wh y do some peop le have absolut e pi tch? Since melodi es are defined by relativ e pit ches, why do some people have th e ability to track absolut e pit ches­inform ati on that has no appa rem value? U nderstand ing speech virtually requi res that we ign ore absolut e pit ch inform ati on . If we did not , we would not be able co understand chil­d ren, who speak an octave or two high er than adults.

125

M USIC M EE TS MEDI CIN E

A contra di ctor y puzzle arises when we consider th at the audito ry system , from the cochlea in the ear up to the cort ex of the brain, cont ain s neuron s that respon d only to spe­cific frequencies. Our ears and our brains indeed regi ster absolute p itch infor mat ion at every stage. T he second que stion then becomes not "W hy do some people have absolut e pi tch?" but rather "W hy doesn 't everyone)" Afte r all , as the lat e psychologist DIXON WARD {3) was fond of point ing out, we do not have to run to a pi ctur e of a rainbow to say that a rooster's com b is red, or run to a bottl e of camphor to ident ify the odor of a skunk. Wh y, th en , if someone plays us a not e, do most of us have to run to the piano to figure out what note it is) As Deut sch {12) points out , th e situat ion is aki n to the med ical condition of color anom ia, in whi ch patient s can readily di stinguish colors with normal disc rirninabiliry, but are un able to att ach labels to the colors {13).

Some p rogres s has been made with respect ro the se que sti ons. An emerg ing bod y of research suggests that both tone and pi ece absolut e pi tch involve two separate cognit ive subskills , long-ter m pitch memory and an appropriate form of lingui sric codin g for attachin g labels to st imuli (14 ). "True" tone absolut e pi tch requi res ind ividual int ernal pit ch standa rds for all 12 chroma. Thi s tem plat e can shift wit h age by as m uch as two sem itones {I S, 16}, and shi fts can also be induc ed neurochem ically {l Z} ,

Absolute pit ch should not be confus ed with relat ive pitch, an ability th at nearly every musi cian learn s. Relat ive pitch refers to th e ability to identif y or produ ce musical intervals, wh ile absolut e pit ch refers to the ability to ident ify or produce individu al m usical pi tches . To illustr ate, if we present peopl e with relat ive pitch wi th the rones A and C, they can identif y the musical interv al as a minor third , or 300 cent s. If, in add it ion , we tell th em th at th e nam e of th e first tone was A , their knowledge of int erval and scale relation s will allow them to identi fy the second rone as C. On th e ot her hand (this is why God gave us two hand s), if we had rold thes e people that th e name of the first tone was D, they would have no reason to disbelieve us and wou ld happil y ident ify the second tone as F- the ton e that is a mi nor th ird above D- and not know th at we had fooled them . This is because possessors of rela­tive p itch , by definiti on, do not have an int ernal templat e or reference system for pi tch as absolut e p irch possessors do {3). In contra st , if we played an A for absolute p irch possessors and told them that it was a D , th ey wou ld know thi s was not correct. Most absolut e pi tch possessors actually have difficult y with relative pit ch tasks in that they do not identif y musi cal int ervals dir ectly by th eir sound, but instead use their kn owledge of scale relati ons

126

MUSI C M EETS MEDI CIN E

to dedu ce the name of an int erval from the ir abil ity to identif y it s compo nent tones. N ore that thi s is the oppos ite stra tegy of relativ e pi tch possessors giv en a reference tone, and who deduce the tone names from their abili ty to identif y the interv al they define .

A musici an with good relative pit ch who has intern alized several, but not all , of the p itc hes of the chromatic scale can often label pit ches as accuratelYas one with true absolute pitc h, but not as rapidlY; such peop le are said to havepseudoabsolutepitcb{I , 7 , I S}. Th e labels used in tone absolut e p it ch are mu sical not e nam es; in pie ce absolut e pit ch , they are names of pieces and lyrics of songs . It has also been argued that the use of nonmu sical, informa l names (such as "th at 's the first not e in the song 'Ho te l California ' ") should also be accepted as evidence of a form of imp licit or lat ent ton e absolut e pit ch {5, 6,14, 19}.

Int eresti ngl y, absolute p itch does not appea r ro be correlated with other mu sical ski lls. Comp osers with ton e absolut e pit ch (e.g ., MOZART, SCRIABIN, MESSIAEN, BOULEz) did not necessarily write better or worse m usic th an compos ers with out it (e.g ., WAGNER,TCHAIKO­VSKY, RAVEL, STRAVINSKY{20D. W hile tone absolute p itch is sometim es an advantage (help ­ing horn p layers to imagine tones before pla ying them , sing ers to perf orm atonal mu sic, and the ori st s to follow large-scale tonal struc tu res by ear), it can also be a hindran ce to certain rasks (e.g ., when playing , singin g , or listenin g to mu sic in a key other than that wr it ten). Regardin g relativep itch, people wi th absolut e pi tch can be less skilled tha n oth er m usicians, calculat ing int ervals and chords from note names rather th an hearing them directly {21 , 22}. Their constant aware ness of mu sical pit ch label s can de tracr from th eir enjoyment of mu sic- as more th an one possessor of ton e absolute p itch has comp lained : "I don 't hear melodies; I hear pi tch names passing by."

The relative frequ ency of absolut e p itch in the general po pulat ion has never been es­tablished . An oft-ci ted estim ate of 1 in 10 000 (23) was not based on scientifi c study, and reliable est ima tes are further confounded by th e follow ing two pro blems : (i) test s for abso­lute pitch , as typi cally admin istered , can be comp leted onl y by mu sicians who have some familiarity with tone nam es and (ii) absolute p it ch is not an ali-or -none abili ty, and th us one needs to dec ide nonarbi traril y on the th reshold that qualifies one as a tru e possessor. Thus the distin cti on betwe en possessors and non possessors is not clear-cut, and it is best to th ink of absolute pit ch ability as fall ing along a cont inuum . Absolute p itch possessors are usuall y consider ed to be those who label 70-10 0% of rando ml y selected , middle- range piano ton es {II}. N ote th at even mu sicians not claim ing absolut e pit ch identi fy ton es above

127

M USIC M EET S M EDICI!' E

chance levels (1/12 = 8. 3% ) wi rh rates up to 40% {I0, II} . Thi s latter result is not surpris­ing given that neur ological inform ati on on absolute pitch is available for all levels of the aud ito ry system (24). Even songbirds {25}, canines {26}, and monk eys {n} demonstrate absolute pitch memory.

DEUTSCHwas the first to recogniz e the conti nuous nature of absolute pitch abilities. She investigated two aspect s of music cognit ion , th e invariance of ton al relation s under trans­positi on and the dimens ionality of intern al pitch represent ations {28- 30}. In these stud ies, the parti cipants were asked to judge th e height of octave-comple x, pit ch-ambi guous rones, known as SHEPARDtones {31}. A pair of such tones, with their focal frequ ency a trir one apart , form a sort of audi tory N ecker cube and are ambi guous as ro whether th e second tone is higher or lower than th e first. The part icipa nts' dire ctio nal jud gm ents were found to be dependent on pit ch class, leading DEUTSCHto conclude that , although the participants were not able to label th e tones, they were neverth eless using absolut e pitch indi rectly. DEUTSCH and her colleagues (32) further specul ated that absolute pit ch "is a comple x faculty which may frequently be present in partial form (p. 135 1]" . More recentl y, D EUTSCH and her colleagues (l 2, 33, 34) have p rovided evidence that speakers of tonal lang uages, such as Mand arin , use absolut e pit ch inform atio n all the t ime in daily conversat ion. In addition , SAFFRAN & GRIEPENTROg {35} dem onstrated impli cit absolu te pitch abiliti es in infants as youn g as 8 months.

Althou gh, in the popu lar media{ 36}, there are some wh o claim th at absolute pit ch is "comp letely inb orn " and that young children are "born with the k nowled ge of note names ", this clearly cannot be true ; tone names must be acqui red along with other lingui sti c term s durin g lang uage acqui sition . Th e real mystery is why some children develop absolute pitch and oth ers do not (5, 37, 38 ). It has been establi shed that mu sicians who start musical training early ate more lik ely to acquire tone absolute p itch than thos e who start it late (38 , 40), Pseudo absolute pitch can be acquir ed in later life, but only with considerable moti­vation , time , and effort {l 8 , 4 1-43 }, Late acquirer s of pseudo absolute pitch are generally less spontaneou s and accurate in their identific ation of pit ches; th ey tend not to develop a complet e int ernal chroma templ ate, filling the gaps by means of relativ e pitch, Younger child ren acquire pi ece absolute pitch more easily than older child ren (shown by singin g a song in its regul ar key (44)), Man y in th e field now believe ther e exists a sensit ive period for the acquisit ion of true absolute pit ch and that specific training to associate tone names with

128

M us IC M E ETS MEDIC INE

their sound is required . Indeed , in region s of Jap an , where the Suzu ki meth od is prevalent and this typ e of tra ining is conducted , absolute pit ch rat es can soar as high as 50% .

T he search for an absolut e pit ch gene (22 , 4 5, 46) may be in vain, g iven that , in a learned skill , "nature" and "nur ture" cann ot easily be separated (12 ,47,4 8) and that ab­solute pitch involves several neurall y d istinc t subproc esses (p itch percep t ion, classificati on , labeling , storage in long-t erm memory, retr ieval from mem ory (19 , 49» . Recent evidence from brain im agin g studies has suggested that th e neur al correlates of absolut e p itch may involve the planum ternporale (SOl and areas of the left posterior dorso -lateral front al cortex (lpDL FC), an area associated with labelin g in cond itional associative learning (491.In this latt er study, IpDLF C was shown to be act ive in tasks of both int erval nami ng and absolute p it ch naming, p roviding neuroanaromi cal confirma tion that it is merely labelingabil ity th at disti nguishe s absolute pit ch possessors from non possessors. The reason why some childre n acqui re thi s abi li ty and others do not may be simp ly because they were taught it and made an effort to learn it. This assumption is not incon sistent with th e not ion that there may inde ed be some ge net ic contr ibut ion in the way of a clu ster of ge nes p rovidin g a ge net ic pred ispositi on toward absolute pit ch . But, if thi s is the case, it is unlike ly that t hese genes encode prote in synth esis for absolute pit ch per se, but , rather, they may encode prot eins tha t con tr ibu te to componentabilit ies that are requi red for the development of absolute pit ch, specifically such subskills as auditor y memory, auditory att ent ion, condi tional associat ive learnin g , catego rical pe rceptio n, and perhaps even a p red ispos it ion towards absolute versus relati ve featur es of cert ain perce ptua l st imu li.

N ew stu dies underway in several laboratorie s are bri ngin g convergi ng techniqu es and evidence to a more th orough unde rstand ing of absolut e pit ch, and th ese incl ude new stu dies of infant and chi ld developm ent, functional neuroanatom y, neuropsycholog ical case studi es, ge netics, psychoph ysiology (including evoked response potent ials), and tradi tional behavioral studi es. Und erstandin g both the nature of absolute pitch and wh y it favors some per sons over others can tell us some thing abou t how the hum an brain processes melodies and p it ch and , ult im ately, can lead to broader theori es with respect to cogniti on, perc ep tion, and th e int eracti on betw een th e tw o.

In the meantim e, tho se who have absolute pit ch report that it can help with mu sical dict at ion and the recollecti on of melodies. Those wi thout it often wish they had it for these reasons, and yet most mu sical tasks in perform ance and com pos ition require only a goo d

l 29

MU SIC M EE TS M ED ICI NE

sense of relati ve pitch . Its value to the working mu sician is thus deb at able. For the research comm unity, its existence is more a blessing tha n a curse, for it g ives us the opp ort unity to bene t und erstand neural organi zat ion and the connect ions betw een neurodeveloprn enr, neur oanatomy , and high er cognit ive fun cti on .

References

1. BACHEM A. Various types of absolut e pitch .] Acousr Soc Am . 1937 ;9: 146-5l.

2 . BAGGALEY]. Measurement of absolute p itch. Psychol Mu sic. 1974 ;22 :11-7.

3. WARDWD. Absolut e pitch . In: DE UTSCH, D, edi to r, The psych ology of music. 2nd ed . San D iego (CA): Academi c Press; 1999 . pp . 265 - 98.

4. PARNCUTT R, LEVITIN D]. Absolut e p itch. In: SADIE S, ed ito r, N ew gro ve di ct ionary of music and mus i­cians . New York (NY): Sr . Martin s Press; 200 1. pp. 37- 39.

5. DEUTSCH D. T he pu zzle of absol ut e pitch. Cur Dir Psychol Sci. 200 2; 11:200 -4.

6 . LEVITIN DJ , ROGERS, SE. Ab solut e p it ch : percep t ion, cod ing , and controv ersies. Trends Cogn Sci . 2005;9 (1)26 -33

7 . BACHEM A. T im e fac to rs in rel ativ e an d abso lute pi tch dererrn inari on. ] Acous r Soc A m. 195 4 ;

26 :751-3.

8. BURNSEM , CAMPBELLS1. Frequency and freq uency -rat io resoluti on by possessors of absolut e and relat ive

pir ch: examp les of catego rica l perceptio n ? ] Acc use Soc Am. 1994 ;96 :2704- 19 .

9· RAKOWSKI A, MORAWSKA-BuNGELER M. In search of rhe crir er ia for absol ut e pit ch . Arch Acou sr . 1987;12:75- 87 .

10. LOCKHEAD G R , BYRD, R . Pr acti cally perf ect p itch . ] Acousr Soc Am. 1981 ;70 :38 7-9.

I I. MIYAZAKI K. The ident ificat ion of mu sical pitc h by absolut e pi tch possessors. Per cept Psychoph ys. 1988;44: 501 -12.

12 . D EUTSCH D. T he enigm a of absolute pi tch . Acoust Toda y. 2006 ;2:11- 9 .

13· G ESCHWIND N , FUSfLLO M. Color-nam ing defect s in associat ion with alexia. Arch N euro l. 1996 ;15: 137­

46

14 · LEVITIN DJ. Abso lut e m emory for musical pitc h : evid ence from rhe pro duct ion of learn ed me lod ies. Per­cep t Psychoph ys. 1994 ;56 :4 14- 23.

15· VERNON E. Absolut e pitch : a case study. Br] Psychol. 1977;68 :485- 9.

16. W YNN VT. Absolut e pitc h revi sit ed. Br] Psycho!. 1992;8 3:129 - 31.

17. CHALOUPKAV, MITCHELL S, M UIRHEAD R . Ob servarion of a reversible , medi cati on- induced change in

pitc h percepti on. ] Acousr Soc Am. 1994 ;96 :14 5- 9.

18 . CUDDYL1. Training the abso lut e ident ificat ion of pi tch . Percept Psyc hophys. 1970; 8: 265- 9.

130

M USIC MEE TS M EDICINE

19· LEVITIN D]. In search of rhe mu sical min d. Cerebrum. 2000 ;2:2- 24.

2 0 . SLONIMSKYN. Per fect pitc h: a rue ful aut opsy. N ew York (NY) : Oxf ord U niversity Pr ess; 1988.

2 1. MIYAZAKI K. Percepti on of mu sical int ervals by absolut e pit ch po ssessors. Mu sic Pe rcep t. 1992;9:4 13­26.

2 2. MIYAZAKIK. Abs olute pit ch as an inability: ident i ficat ion of musical int erval s in a ron al conte xt. Mu sic

Pe rcep t. 199 3;11:55- 72.

23· PROFITA) , BIDDERTG . Perfect p itch . Am) Med Ge ne t. 1988;29:763- 71.

24 . MOORE B. A n int roduc tion ro the psycholog y of heari ng . San Di ego (CA): Academi c Press; 1997.

25. HULSE SH, CYNX] , H UMPAL]. Ab solut e and relative p itch dis crim inati on in ser ial pit ch p ercep t ion by

bird s.) Exp Psycho! G en. 1984;113:38- 54.

26. TOOZE, Z], HARRINGTON, FH , FENTRESS, ] C. Indi vidu all y d ist inct vocalizat ions in timber wolves , can is

lupu s. An im Behav . 1990;40:72 3-30 .

27. D 'AMATO MR. A search for ronal patt ern pe rcep t ion in debus m onke ys. Why monk eys can 't hum a tune.

Musi c Percept. 1988;5:453- 80.

28. D EUTSCHD. The rriron e paradox: an influ ence of langu age on musi c percep t ion. Mu sic Percept. 1991;8:335­47.

29. DEUTSCH D. The rrirone parad ox: im pli cation s for the repr esent ation and com m unication of pit ch srr uc­(LIre. In: JONES MR , H OLLERANS, edi tor s. Cogn it ive bases of mu sical com municat ion . W ashin gtOn (DC) :

Am erican Psychol ogical Associati on; 1992. pp . 115-3 8.

30. DE UTSCH D, KUYPER WL, FISHERY. Th e rrir one pa radox: It s pre sence and form of dist ribut ion in a ge n­

eral pop u lat ion . Mu sic Percepr. 1987;5:79- 92 .

3 1. SHEPARD RN . Cir cularir y in judgment s of re lative pi tch. ] Acou sr Soc Am . 1064;36:2346- 53.

32. DEUTSCHD, MOORE FR , DOLSON M . Th e perceived hei ght of ocrave- relar ed compl exes.) Acous Soc Am.

1986;80:1346- 53.

33. DEUTSCH D , HE NTHORN T, DOLSON M . Abso lute pit ch is demo nst rate d in speaker s of to ne lan gua ges.

J Acoust Soc Am. 1999 ;106:2267.

34. G LANZ]. Study lin ks perfect pit ch to ronal langua g e. N ew York Time s, 5 November 1999. p. 1.

35· SAFFRANJR, GRIEPENTROG G]. Absolut e pit ch in infant au ditory learnin g: eviden ce for de velopment al reorg aniz ation . D ev Psychol . 200 1;37:74-85 .

36. BBe. "Frontie rs" , featu re show on Pe rfect P itch. BBC Radi o 4 . Broadcast da te 9 May 2001.

37. LEVITIN D]. A bsolut e pit ch: self-reference and hwnan memory . Inr ] Comput Anr icipar ory Sys. 1999 ;4:255­66

38. WARDD , BURNS EM . Singin g witho ut aud itory feed back .] Res Singin g Appl Vocal Pedagog . 1978;1(2):24­44 .

39· SERGENTD. Experiment al investi gati on of absolut e pir ch.} Res Musi c Educ. 1969 ;17:135--43.

40 . WELLEK A . D as absolute G ehor und seine Typ en. Z An gew Psych ol Charac rerk unde -Beih . 1938;83:1­368.

131

M USK M EETS MED K IN E

4 r . BRADY PT. Fixed- scale m echani sm of absolute pitch. J Aco us e Soc Am . 1970 ;48 :883- 7.

42. CUDDYL1. Pr act ice effects in rhe abs ol uce ju dg menc of piech. J Aco use Soc Am . 1968;43:1069-7 6.

43. M EYER M . I s rhe m emo ry of absolut e p it ch cap ab le of develop me nt by cra in in g. Psychol R ev.

1899 ;6:514 -6..

44 . SERGENT DC, ROCHE S. Perc epcual sh ifts in th e aud it ory inf ormacio n p rocessing of young ch ild ren. Psy­

chol Mu sic. 1973;1:39-4 8 .

45. BAHARLOO S, J OHNSTON PA, SERVICE SK, G ITSCHIERJ, FREIMER NB. Ab soluce pi tch: an approach for

idencifi caeion of ge neric and non g en eri c co mpo ne n ts. Am J Hum Genet . 1998 ;62:224- 31.

46. G REGERSEN P, KOWALSKY E, KOHN N , MARVIN, E. Early ch ildhood musi c education and p red ispo sitio n

co absolu te pi t ch . Am J Me d G enet. 2000;98:280- 2.

47· J EFFRESS LA A bsolu te p itch . J Acou sr Soc Am. 1962;34 :987.

48. LEVITIN DJ , ROGERS SE . A bsolut e p itch percep ti on, cod ing , and cont roversi es . Tren ds Co g n Sci .

200 6;9(r):26 - 33.

49 · Z ATORRE RJ, PERRYD , BECKET C, W ESTBURY C, EVANSAC. Funcci onal an atomy of musi cal proc essing

in list en er s wieh absolu te and rela tive p itch. Proc N arl Acad Sci USA . 1998 ;95:3172- 7.

50. SCHLAUG G, J ANCKE L, HUANG Y, STEINMETZ H. In vivo evi dence of srrucrural bra in asymm ecry in mu­

sicians . Scienc e. 1995 ;267:699- 70 1.

132

MU SIC M EET S M EDICI N E