daltons (1), giving an estimiate of 0.023disintegrations per 5S RNA molecuLleper 2 maonths. The enmullsion, NTB2,detects tritiLim disintegrations with near-ly a 10 percent efficiency (12); there-fore, 1 5 to 30 grains wouLld accumulateover this region if the hybridizationyield was between 3 to 6 percent. SuLchhybrridization efficiency places the trans-fer RNA genes with their reduLndancy of13 times or more (13) within the limitsof resolution of the system. Indeed,certain sites along the chromosonme armsare frequently lightly labeled with frac-tions containing transfer RNA (Fig. 2,region 60 BC) and need only to hecataloged to cytologically mark thelocation of the numnerous transfer RNAgencs.

D. E. WIMBER*D. M. STEFFENSEN

DJeparbtieu t of BotanY,Un'ietrsity of Illinois, Urb(ina 61801

References and Notes

1. F. 1\. Ritossa, K. C. Atwood, S. Spicgelman.Genetics 54, 819 (1966); K. D. Tartof anlL R.1'. Perry, J. Mol. Biol., in press.

2. D. D. Browni and C. S. Weber, J. Mol. Biol.34. 661 (1968).

3. 1.ow yeast medlioim: 1000 ml of H,O, 55 g ofyellow cornimcal, 25 g of brewer's yeast, 8 gof agar.

4. C. G. Mead, J. Biol. Clhent. 239, 550 (1964).5. K. S. Kirby, Bioclhe,ni. J. 64, 405 (1956); F.

M. Ritossa and S. Spiegelmani. Proc. Nat.Acad. Sci. U.S. 53, 737 (1965).

Upon presentation of a brief tonalstimulus a d-c potential shift can herecorded from the cochlea (1). Thispotential, the summating potential(SP), is reported to be an eluLsive phe-nomenon and difficult to quantify (2),even though a large number of experi-ments have been devoted to it (3, 4).The SP, just as the other inner-ear po-tential, the cochlear microphonic (CM),is presumed to originate in the haircells; it is generally stated, or tacitlyassumed, to appear in opposite polarity(referred to an indifferent referenceelectrode) in the two perilymphaticscalae at a given cochlear location. Weshow here that the latter contention isnot generally correct, and that the re-corded SP at any point in the cochleacan be thought of as the algebraic sum6 NOVEMBER 1970

6. J. 1). Nlandell and A. i). Huershey, A nal. Bio-thlile. 1. 66 (1961)): RNA wsas dissolvsd inbuflcred salt (0.1W1 NaC l. 0.0)5MW Na H,PO0,plI 5.5), loaded onto a MAK coltomiin, .iniI- in against a liniear ().1 to 1 .2W1 NaCl gra, li-ent. Fractions taken between 11.42 to 0.52.1and 0.53 to 11.62M NaCiIvere pooled: thcsessere the 4.S atidciS Jfractionis, respectively.

7. N\. L. Pardtue, S. A. Gerbi, R. A. Eckardit,.1. G. Gall, (Chrono0soina 29, 268 (1970).8. N. SLocoka and T. Yamane, Proc. Nat. A cad.

Sci. fJ.S. 48, 1454 (1962).9. U. E. Locning, Bliochein. J. 102, 251 (1967).

Nine perccnt gels (7 by 0.6 cm) werce ulscdand I to 2 ug of RNA were applied per gel.Flecti-oplioresis was carried ooit for 3 hotirs at2.5 ma per gel. The RNA was meastired inser-ial 1 -mm slices; see R. i-i. Gray, R. Her-zog, D. M. Steffensen, Biochimr. BiophlYs.Acta 157, 344 (1968).

10. D I. Lindsley and E. H. Grell. GenieticVariationis of Drosophila melanogaster (Car-negie Instittition of Washington, Washington,D.C., 1967).

11. 1. Gillam, S. Millward, D. Blew, M. vonTigerstrom, E. Wimmer, G. M. Tener, Bio-chemistrt 6, 3043 (1967). The RNA in two 1-mm slices taken from the leading edge of the55 peak and the 5S peak (located from ali-qulots of eltiting solution) was eldited from thepolyacrylamide gel in 0.1M NaCI, 0.O0Methylenediaminetetraacetate, pH 7.2. The[PIIIRNA was further purified on a benzoylateddiethvlaminoethyl-celltilose coltimn. The elUantwas then dialyzedl against water and lyo-philized to dryness: then a solution of 0.9AfNaCI and 0.09M soditim citrate (pH 7.4) wasadded to yield a final concentration of about0.5 /pg/ml.

12. J. E. Cleaver, Th'vntdine Metabolism andfell Kineticts (North-Holland, Amsterdam,1967).

13. F. M. Ritossa, K. C. Atwood, S. Spiegelman,Genetics 54, 663 (1966).

14. Research stuppor-ted by NIH grants GM 8465and GM 11702 and NSF grant GB-7645. Wethank Lilian Cooper and Doris Wimber fortheir technical assistance.

* Permanent addrcss: Biology Department, Uni-versity of Oregon, Euigene 97403.

8 Jtily 1970

of the potential difference between scalavestibuli and tympani (heretofore desig-nated as differential component or DIF)and the common potential of thesescalae (to he identified as average com-ponent or AVE). These componentsvary independently; both are functionsof stimulus parameters and cochlearrecording site.

Data were obtained from 39 guineapigs. The experimental animals wereanesthetized with urethane; the audi-tory bulla was approached ventro-later-ally and was opened widely; small holeswere drilled through the bony cochlcarwall at appropriate locations; and finewire electrodes (Tu or Ag-AgCl) wereintroduced throuLgh these holes into theperilymphatic scalae. Electrodes wereplaced in pairs, one in the scala vestibuli

and one in the scala tympani ot a giVenturn. Either one or two pairs (If elec-trodes were ulsed. The indifferent elec-trode was an Ag-AgCI disk placed onthe neck muscles. The signals fronm thet\vo active electr-odes were indepenid-ently amplified (60-db gain. 3 secondtime-constant) and processed. ThestimuLli consisted of tone bursts 40msec long, with 100 mnsec betweenhursts, and virtually instantaneous riseand fall times. The sound signal wasdelivered to the bony meatus of theanimals in a closed acouistic system; itwas continuously monitored near theeardrum with a calibrated probe-tubemicrophone. The aamplified signals wereprocessed with an averaging computer.The gated sinusoidal signal that pro-vided the speaker voltage was not syn-chronized with the onset of the burst;consequently the cochlear microphonicphase was changing fromii burst toburst with the result that this a-c signalwas simply averaged out. What re-mained was the onset and offset actioInpotential, some onset CM in responseto ringing in the acouLstic systenm, andthe SP. This method of elimiin-ating theCM is much superior to filtering be-cause it avoids wave-form distortion.The number of samples averagedranged between 16 and 1024. Thisnumber depended on the magnitude ofthe signal; at high stimulus levels afew s.amples sufficed, but at low levelslarger numbers were required. Wewould either average the scala vestibuliand scala tympani potentials (SV andST) or utilize electronic means to sub-tract or add these signals before averag-ing. These latter two signals are definedas SV - ST = 2 DIF and SV + ST =2 AVE. It is easy to demonstrate thateither arithmetic or electronic genera-tion of DIF and AVE components fromSV and ST potentials yield virtuallyidentical results. It should be notedthat when obtained by electronic addi-tion or subtraction, the DIF and AVEcomponents are similar to the resultsof classic differential electrode record-ing when either action potential or CMis rejected (5).The only difference between our

method and the usual differential elec-trode recording is that we do not utilizean electronic bal.ancing technique; in-stead, simple addition or subtraction isperformed. Appropriate symmetry be-tween the two electrodes is achievedby their careful placement and by therejection of inadequate prep.arations(6). The DIF component should he in-

641

Cochlear Summating Potentials: CompositionAbstract. The potential dif/erence actross tile cocileair partitioll anid the overall

potetltial level of a given cochlear crosS section were iieasttredl a.s functions ofstiinuiiltts parattleters atnd spatial lOCatiOn. It was con.firmed that the potential dif-ference is negative in the vicinity of greatest excitatioll, atid it was discoveredthat in the sam?1e region the overall potential level is positive.

on

Sep

tem

ber

11, 2

015

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Sep

tem

ber

11, 2

015

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Sep

tem

ber

11, 2

015

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

on

Sep

tem

ber

11, 2

015

ww

w.s

cien

cem

ag.o

rgD

ownl

oade

d fr

om

1 khz, 90 db

J Sy3

.2 mv

Fig. 1. Averaged scala vestibuli (SV) and scala tympani (ST) summating potentialresponses to tone bursts 40 msec long with 100 msec between bursts. Stimulus fre-quency and sound pressure level are indicated as the parameters. Vertical bars providevoltage calibration. Panels a, b, and c show recordings from one electrode pair inthe first turn, whereas panel d depicts responses from the third cochlear turn. Positivepolarity is up.

terpreted as the potential differenceacross the cochlear partition, and thusas an indication of the combined outputof the sources that are located betweenthe two electrodes. The AVE compo-nent has been generally considered toreflect remote activity, and indeed untilnow it was used merely to reject localCM and to emphasize whole-nerve ac-tion potential which is, of course, a "re-mote" potential. We wish to extendthis interpretation. It appears to us thatthe AVE component might reflect bothremote potential and a local responsewhich, however, has the same polarityin the two opposite perilymphaticscalae. We will show that for certainstimulus conditions this latter constitu-ent of the AVE response predominates,and in these cases this response shouldbe interpreted as the potential level ofa cochlear location (SV and ST) inreference to indifferent tissue (7).

In Fig. 1 some sample SP wave-forms are shown. These are SV andST recordings to demonstrate possiblepolarity relationships between the twopotentials in the two scalae. The firsttwo traces are from the basal turn ata low frequency and a relatively highsound pressure level (SPL). The polar-ity is opposite in the two scalae; SVis positive, ST is negative. The secondtwo traces are from the same location

at the same frequency but at a lowsignal level. Here both traces are nega-tive and virtually identical. The thirdpair of traces are again from the basalturn, but here the frequency is highand so is the intensity. The polarity isopposite, but here SV is negative andST is positive; the magnitudes of thetwo responses are unequal. The finaltwo traces were recorded from thethird cochlear turn at a midfrequencyat high stimulus level. Both traces arepositive but not equal in magnitude.These traces simply serve to demon-strate some of the varieties of relation-ships that can be seen for SP polarity.Note that the magnitude of the SPchanges in some cases during the pres-ence of the stimulus. The change isasymptotic, and the response generallylevels off around 40 msec. All readingsof SP magnitudes are obtained by tak-ing the difference between the baselineand the average magnitude of the SPduring the final 5 msec of the response.There are many possible ways to

depict the quantitative behavior of theSP as a function of stimulus param-eters. It is not our purpose to providea complete parametric picture of thedependenwe of the SP on stimulus in-tensity and frequency and upon record-ing location. Such will be given in amore detailed communication. Here we

merely chose to present one examplewhich, however, is sufficient to dem-onstrate our main points of conten-tion. In Fig. 2 the magnitudes of theDIF and AVE SP components aregiven as the function of stimulus fre-quency with intensity as the parameter.These plots are obtained from oneguinea pig that had electrodes placedin scalae vestibuli and tympani of thesecond cochlear turn. The followingtrends are evident. At low frequenciesthe DIF component is positive at allintensities, and the magnitude is di-rectly related to stimulus strength.Around 4000 hz the DIF componentis negative at all levels and again themagnitude increases with increasingsound pressure. At high frequencies theDIF component vanishes. Note thatthe width of the negative band around4000 hz increases with increasing sig-nal level, but significantly the spreadis toward the low frequencies. At firstglance the AVE plots appear to be themirror image of the DIF plots; thishowever is not the case. Around 4000hz there is a positive band in the AVEplot; the width of this band also widensat higher stimulus levels, and thespread is again primarily toward thelow frequencies. The zero-crossingpoints are generally not the same inthe DIF and AVE plots. For fre-quencies below and above the positiveband the AVE response is negative.

If the recording electrodes are placedat another cochlear site, then a similarbut displaced pattern of DIF and AVEresponses is obtained. To explain, withan electrode pair in the first turn themaximum DIF negativity and AVEpositivity occurs around 10,000 hz.When *the electrodes are in the thirdcochlear turn, these characteristic peaksare seen around 1000 hz. In all casesthe potentials at lesser or higher fre-quencies than that of the characteristicregion show similar patterns to whatis demonstrated here for the electrodesite in the second turn.The general trends can be studied

with optimum clarity in spatial plotssuch as the one given in Fig. 3. Herethe results from three animals, havingelectrodes in different cochlear turns,are combined to provide a spatial rep-resentation of the two potential com-ponents. In other words, SP magnitudeis plotted as a function of distancefrom the stapes. Stimulus frequency isthe parameter, and the intensity is con-stant at 60 db SPL. Each curve is basedon three data points. The general trendsseem to be well represented by these

SCIENCE, VOL. 170

200 hz, 100 db

a

8 khz, 90 db

C

b1 g~~~~~-l - V-- IN&

5

642

plots even though accuracy of spatialdetail might be lacking.

Three zones are seen in both plots;they can be best studied by following,for example, the 4000-hz graphs. Inthe DIF plot there is a negative peakbetween 1 1 and 12 mm; a positivepeak occurs at approximately the samelocation in -the AVE plot. Below thischaracteristic region the DIF responsegoes positive and then diminishes to-ward the base of the cochlea. In thissame region the AVE response is nega-tive. The latter is also negative on theapical side of the positive maximum,but in contrast the DIF response as-sumes zero value toward the apex. Onenotes similar trends and polarity zonesfor the other frequencies.The two types of DIF responses

(negative and positive) appear to cor-respond to the previously identified (3)so-called negative and positive summat-ing potentials, SP- and SP+. If wecompare the spatial patterns of Fig. 3with available frequency maps of theguinea pig cochlea (5), then it is notedthat both the prominent negative DIFpeak and the positive AVE peak occurmore apically than the maximum me-chanical vibration of the cochlear par-tition. These peaks are better corre-lated with the distal end of the travel-ing wave envelope and conceivablywith the hypothesized shear waves inthe cochlea (8). The positive DIF re-sponse appears on the proximal slopeof the traveling wave envelope, whilebeyond the spatial extent of the wave,the DIF response vanishes. It is highlypossible that both negative and positiveDIF responses originate in cochleardistortion processes as has been sug-gested by several authorities for SP-and SP+ (9).The properties of the DIF compo-

nent strongly resemble those of the SPwhen recorded from the scala mediaof the cochlea (4); both are apparentlyrepresentative of the local potential dif-ference across the organ of Corti.We suggest that the two likely

sources that might be responsible forthe generation of the negative DIFcomponent are either the nonlinear dis-tortion of the CM-producing process,or one-way bending of the cilia of thehair cells on the distal envelope of thetraveling wave, or both. The positiveDIF component is likely to arise inone-way bending of the hairs in a dif-ferent direction on the proximal slopeof the traveling wave. Von Bekesy (8)had demonstrated that the principalmotion is different on these two slopes,6 NOVEMBER 1970

I ~~~~~~AVE0.1 0 0 1 -

0~ ~ ~~~~~~~~~~~~~~~~7

T2-1 DIF ___

100 90 ~~~~~~~~~~~~50

0.1 0.2 0.5 1 2 5 10 0.1 0.2 0.5 1 2 5 1oFrequency (khz)

Fig. 2. Magnitudes of DIF and AVE summating potential components as the functionsof stimulus frequency with sound pressure level as the parameter. Electrodes werelocated in the second turn. By definition the DIF and AVE components are obtainedfrom the potentials recorded from scalae vestibuli and tympani as 2 DIF = SV - ST;2 AVE = SV + ST.

thus opposite unidirectional bendingaccompanying the to and fro move-ments is a distinct possibility.The AVE component of the SP

response had not been previously iden-tified. This component is clearly thesum of two constituents, the negativeand positive AVE response. The posi-

S.,0.

-0a)4-

QIL

tive component dominates in the vicin-ity of maximum excitation, whereas thenegative component is evident every-where else in the cochlea. It is our con-,tention that the negative component isa conducted remote response, reflectingdistant activity (10). The identificationof the narrow positive potential zone

60 db

Fig. 3. Composite spatial plot (based on data obtained from three animals havingelectrodes in turns one, two, and three, respectively) of DIF and AVE summatingpotential component magnitude. Abscissa is the distance from the stapes in millimeters;the parameter is stimulus frequency in kilohertz. All data are obtained at 60 db SPL.By definition the DIF and AVE components are obtained from the potentials recordedfrom scalae vestibuli and tympani as 2 DIF = SV - ST; 2 AVE = SV + ST.

643

in the region of the maximum activityis probably the most important contri-bution of this study. This positivity is astimulus-related cochlear response thatis confined to a limited spatial extentof the inner ear, and which is an inde-pendent response from the familiar CMand DIF-SP components which appearas voltages across the organ of Corti.The positive AVE-SP component is thepolarization of an entire segment ofthe cochlea. Since the only electricalconnection between the membranouscochlea and the rest of the body isthrough the internal auditory meatus,via the eighth nerve and blood vessels(11), it can be conjectured -that thepositive AVE component can serve as ahyperpolarizing electrical agent actingon the dendrites that originate in theregion of maximum positivity. Thesource of this positive AVE response isnot yet identified with certainty. Thisresponse is possibly due to a voltagedrop caused by longitudinal current flowwithin the cochlea between regions ofstrong and weak excitation.

PETER DALLOSZ. G. SCHOENY

M. A. CHEATHAMAuditory Research Laboratory andDepartment of Electrical Engineering,Northwestern University,Evanston, Illinois 60201

References and Notes

1. G. von Bek6sy, J. Acoust. Soc. Amer. 22,-576 (1950); H. Davis, C. Fernfindez, D. R.McAuliffe, Proc. Nat. Acad. Sci. U.S. 36,580 (1950).

2. H. Davis, D. H. Deatherage, D. H. El-dredge, C. A. Smith, Amer. J. Physiol. 196,251 (1958).

3. H. Davis, C. Fernalndez, R. Goldstein, ColdSpring Harbor Symp. Quant. Biol. 17, 143(1952); J. R. Johnstone and B. M. Johnstone,J. Acoust. Soc. Amer. 40, 1405 (1967); R.Kupperman, Acta Oto-Laryngol. 62, 465(1966); G. A. Misrahy, E. W. Shinaberger, K.M. Hildreth, Wright Air Development CenterTechnical Report 45-467 (Wright-PattersonAir Force Base, Ohio, 1957), p. 1; P. E.Stopp and I. C. Whitfield, J. Physiol. 175,45P (1964); R. Butler, T. Konishi, C.Fernandez, Amer. J. Physiol. 199, 688 (1960).

4. T. Konishi and T. Yasuno, J. Acoust. Soc.Amer. 35, 1448 (1963); V. Honrubia and P.Ward, ibid. 45, 1443 (1969).

5. I. Tasaki, H. Davis, J. P. Legouix, ibid. 24,502 (1952).

6. P. Dalos, ibid. 45, 999 (1969).7. Under ideal recording conditions the DIF

and AVE responses can be registered inde-pendently from one another. If the electrodesare well placed and if the electrical paths fromthem to the various sources of potential aresymmetrical, then any change that might occurin the potential difference across the coch-lear partition is manifested by the DIF re-sponse only. Conversely, under such condi-tions any overall shift in voltage level by acochlear cross section would modify the AVEcomponent alone. In- practice the electrodesare never perfectly balanced, and consequent-ly the recorded components are not completelyindependent. The errors resulting from a givenelectrode imbalance can be estimated, andthus one can guard against unwarranted useof the data. The electrode imbalance can beobtained, at least- as a first approximation,

644

from the comparison of the cochlear micro-phonic outputs of SV and ST electrodes athigh frequencies (6). It is important to pointout, however, that even when perfect a-cbalance (that is, that for CM) exists betweentwo electrodes, they are not, in general,"balanced" for SP. For example, in Fig. icthe responses are shown from scalae tympaniand vestibuli from the first turn at a highfrequency. In this situation the CM's fromthe two electrodes are virtually identical inmagnitude but opposite in phase; there isbalance. Note, however, that the SP magni-tudes are not the same; the SP from the STis considerably greater than that from the SV.This indicates the presence of a "common-mode" d-c potential, the very item of inter-est in this communication that is labeled asthe AVE component.

8. G. von B&k6sy, J. Acoust. Soc. Amer. 25, 770(1953); S. M. Khanna, R. E. Sears, J. Tonn-dorf, ibid. 43, 1077 (1968).

9. I. C. Whitfield and H. Ross, ibid. 38, 126(1965); M. Engebretson and D. H. Eldredge,ibid. 44, 548 (1968); H. Davis, Ann. Otol.Rhin. Laryngol. 77, 644 (1968).

10. At first glance this contention might appear

A recurring question concerns therelative heritability of general IQ. Oneposition is that IQ can be altered bycertain environmental factors, such asprograms of enrichment (1). The op-posite position suggests that the geneticcomponent of IQ is so significant thatit is unrealistic to hope that enrichmentprograms will be able to change IQvery much, particularly for relativelynormal children (2).The issue is exemplified by two sets

of apparently contradictory findings:(i) the IQ's of severely deprived young-sters can be raised by certain programsof stimulation and compensatory edu-cation (1, 3); but (ii) the correlationbetween the IQ scores of pairs of in-dividuals increases dramatically andconsistently with the degree of theirgenetic, as opposed to their environ-mental, relationship (2, 4).

Part of this controversy may be re-solved by distinguishing between thegeneral level of IQ as reflected in anIQ score assessed at a single age andthe sequential pattern of IQ change thatmight occur over age. It is likely thatmany normal children display substan-tial changes in IQ during childhood andthat these shifts over age constitutemeaningful and reliable trends (5).

to be contradictory to the findings of I.Tasaki and C. Fernandez [V. Neurophysiol.15, 497 (1952)] that the electrical spread inthe cochlea from turn to turn is negligible.Actually there is no contradiction. Tasaki andFermindez obtained their attenuation figure of6 db/mm from differential electrode record-ings; in other words, this attenuation constantwould apply for our DIF component. Whennot using differential electrodes, one observesmuch less attenuation of electrical potentials,probably of the order of 1 to 2 db/mm (6).This latter figure is applicable for the AVEcomponent. Thus, it is quite reasonable to as-sume that the negative AVE component is aresponse that results from current spread fromactive remote regions. That this is so can behighlighted by the prominence of this re-sponse in regions and at frequencies where nolocal traveling wave activity exists, that is,in the higher cochlear turns at high frequen-cies.

11. G. von B&k6sy, J. Acoust. Soc. Amer. 23,18 (1951).

12. Supported by NIH grants.

4 August 1970

Since much of the current emphasis oncompensatory education programs isconcerned with changing IQ, it wouldbe valuable to determine if such pat-*terns possess as much heritability as thegeneral level of IQ.Most of the evidence for the genetic

basis of intelligence has been deter-mined with single-age (or at least sep-arate-age) correlations between genet-ically related individuals. These datademonstrate the heritability of the gen-eral level of IQ, but they do not demon-strate the heritability of IQ changesover age. Even if these methods suggestthat the genetic oontribution to the gen-eral level of IQ is substantial, meaning-ful changes in IQ over age are still pos-sible (as produced by enrichment pro-grams, for example).

In this paper I report an examina-tion of the relative heritability of thegeneral level of IQ as opposed to thepattern of IQ change over age. Thisissue was addressed by a comparison ofthe similarity of patterns of IQ changeover age among siblings and amongparent-child combinations with the sim-ilarity of patterns for unrelated subjectsmatched for year of birth and parentaleducatioal level. These comparisonswere made first when both the. genral

SCIENCE, VOL. 170

Intelligence Quotient Pattern over Age: Comparisonsamong Siblings and Parent-Child Pairs

Abstract. Comparisons between sibling and parent-child pairs with unrelatedcontrol pairs matched for year of birth and parental education were made todetermine the relative heritability of the general level of intelligence quotient asopposed to that of the sequential pattern of IQ change over age (3 to 12 years).There was greater similarity among related siblings relative to matched controlsfor general level than for pattern of IQ over age. Relationships between the IQ'sof children and that of their parents as children were not consistent across age.