VOL. 67, No. S MAY 1967

Psycho log ica l Bu l l e t in

SENSORY ADAPTATION AND BEHAVIORAL COMPENSATIONWITH SPATIALLY TRANSFORMED VISION AND HEARING

R. H. DAY

Monash University

G. SINGER

University of Sydney

An analysis of spatial transformations of perceived space is made in terms ofangular and parallel modifications of the median, horizontal, and frontalplanes of O, and the perceptual and behavioral outcomes of such transforma-tions examined. It is argued that there are 2 independent outcomes: behavioralcompensation and sensory spatial adaptation with aftereffect. The 1st can beregarded as a special case of motor learning similar to that studied in earlyinvestigations with frontal plane transformation (mirror tracing), and the2nd is essentially similar to spatial adaptation which may occur with appro-priate nontransformed stimulation. Both effects can occur simultaneously inthe same direction, but the experimental data presented show that they can bestudied independently. The effects observed by Ivo Kohler are treated asspecial instances of sensory adaptation which occur with transformationsdependent upon sense-organ position and movement. The felt-position hy-pothesis and the reafference theory proposed by Held are shown to be rein-terpretable in terms of motor learning and transfer of learning. Variousmethodological issues in the investigation of motor learning and sensoryadaptation are examined.

If before entering the eye light is spatiallytransformed by means of an optical system(prism, lens, mirror), visually guided behav-ior (reading, walking, card sorting) which isinitially disrupted exhibits a progressive re-turn to pretransformation normality. In addi-tion, under certain spatial transformation con-ditions, visual appearances (shape, size, di-rection) are also reported to undergo changeand to exhibit a tendency to return to pre-transformation appearances. Analogous changesalso occur when hearing is spatially trans-formed by means of "pseudophonic" systems.Since Czermak (1855), Helmholtz (1866),and Wundt (1898) drew attention to theseeffects a variety of explanations have beenproposed in terms of perception (Kohler,1964; J. G. Taylor, 1962), motor learning(Smith Si Smith, 1962), reafference (Held &Freedman, 1963), and "felt position" (Harris,1965). In addition to clarifying the nature ofspatial transformations, the purpose of thispaper is to show that there are two inde-pendent outcomes, sensory spatial adapta-tions and behavioral compensation, which

under some conditions may occur together inthe same direction. The first is a purely sen-sory effect identical to that which occurswithout transformation (Gibson, 1933, 1937;Kohler & Wallach, 1944), and the second isa special case of motor learning identical tothat investigated in the early mirror-tracingstudies (Starch, 1910). The recent interpre-tation of the second effect in terms of feltposition merely deals with learning from aphenomenological viewpoint. It is arguedthat failure to recognize the simultaneous oc-currence of these independent outcomes oftransformed vision and hearing has resultedin inappropriate procedures and a tendencyto treat both as specific characteristics oftransformed conditions rather than as in-stances of effects which occur under a varietyof conditions without transformation.

NATURE OF SPATIAL TRANSFORMATIONS

Despite considerable recent research therehas so far been no attempt to describe or toclassify the classes, directions, extents, andcomplexities of transformations. This omis-

307

308 R. H. DAY AND G. SINGER

sion not only makes comparisons of studiesdifficult, but also obscures relationships be-tween early experiments concerned with mo-tor learning and more recent investigationswhose data have been differently interpreted.

Perceived space can be usefully describedin terms of three hypothetical planes: themedian (sagittal), horizontal, and frontal (orcoronal) planes, which for present purposescan be conceived as intersecting at rightangles at the center of the head. When theobserver is vertical, the median (straightahead) and frontal (right-left) planes arevertical and the horizontal plane is parallelwith the ground plane. Spatial transforma-tions which may vary in class, direction, ex-tent, and complexity can be described interms of these three planes and those parallelto them, the median, horizontal, and fronto-parallel planes.

Classes of Transformation

There are two classes of spatial transfor-mation, which will be termed angular andparallel. Angular transformation occurs whenthe optical or pseudophonic system rotates theplane or planes about their transverse orlongitudinal axes. A wedge prism with basevertical rotates the median plane right orleft about its transverse (i.e., vertical) axisso that objects normally straight ahead nowappear laterally displaced. A double Dove-prism system rotates both the median andhorizontal planes together about their longi-tudinal (i.e., horizontal) axes so that normallyvertical and horizontal contours appear slanted(see Figure 1). A pseudophone with micro-phones lying anterior to one ear and posteriorto the other serves to rotate the median planeabout its transverse axis, and microphoneplacement above one ear and below the otherresults in transformation of the same planeabout its longitudinal axis. A mirror beforethe eyes rotates the frontal plane through180° and the degree of rotation can be variedby slanting the mirror.

Parallel transformations occur when one ormore of the three planes are displaced in alinear fashion without rotation about eitheraxis.1 Thus Weinstein, Sersen, and Weinstein

1 Magnifying and minifying lenses serve to increaseand decrease respectively the magnitude of retinal

(1964) used a mirror system arranged sothat objects at a certain distance appeared tobe farther away. Although parallel transfor-mations of auditory space have not been re-ported so far, they are entirely feasible. Twomicrophones the same distance apart as theears but mounted above the head so that bothare to the same side of each ear would pre-sumably achieve a parallel transformation ofthe median plane. Placing both an equal dis-tance above or behind each ear would pro-duce a parallel transformation of the auditoryhorizontal and frontal planes, respectively.

Direction and Extent of Transformation

Angular transformation can vary between0° and 360° in either direction about bothaxes, and parallel transformations can varyto the limits of visual or auditory acuity. Themirror systems used by Kohler (1964) trans-formed either the median or horizontal planethrough 180° so that objects normally rightappeared left and those originally above ap-peared below. Ewert (1930), Snyder andPronko (1952), and Stratton (1896, 1897)introduced double-convex lens systems so thatboth median and horizontal planes were si-multaneously transformed through 180° abouttheir longitudinal axes. A mirror directly be-fore the eyes transforms the frontal planethrough 180° longitudinally so that objects inback appear in front (but objects in front areobscured by the mirror). Stratton (1899)used two mirrors to transform angularly thefrontal plane through 90°. In most recentstudies, the transformations have been rela-tively small (10-20°) and most usually ofthe median plane. The systems used byYoung (1928) and Willey, Inglis, and Pearce(1937) transformed the auditory medianplane through 180° about its transverse axis,and, more recently, smaller angular transfor-mations of the same plane have been induced(Freedman & Zacks, 1964; Held, 19SS).

Complexity of Transformations

The simplest transformation is small inextent involving a single plane. Complexity

projections thus changing apparent size and dis-tance. In this manner (as when viewing through theeyepiece or mouth of a telescope) parallel transfor-mations of fronto-parallel planes are achieved.

SENSORY ADAPTATION AND BEHAVIORAL COMPENSATION 309

of transformation can vary as a function ofclass combinations (angular and parallel),number of planes simultaneously transformed,and extents. Probably the most complextransformation investigated so far is that usedby Snyder and Pronko (1952) involving 180°angular changes in the median, horizontal, andfrontal planes. This was achieved by meansof a double-convex lens system and a mirror.

Area, oj Spatially Transformed Field

In vision, the angle subtended by the trans-formed field may vary from the total visiblefield to a small part of it. Prisms worn closeto the eyes in spectacle frames transform thetotal field, whereas a prism set in a screenand viewed from a distance alters only thatpart of the field bounded by the prism. Onviewing into a large mirror the total field istransformed with respect to the frontal plane,but a small mirror held in the hand transformsonly that area bounded by the mirror.

In hearing, as in vision, spatial transforma-tions may vary in class, direction, extent, andcomplexity by suitable positioning of micro-phones or other devices. Because of the tem-poral basis of auditory localization, however,the total field is always transformed.

REPRESENTATIVE STUDIES

Helmholtz (1866), using prisms worn be-fore the eyes, angularly transformed the me-dian plane and observed progressive reductionin reaching errors and the occurrence of er-rors in the opposite direction when the prismswere removed. Wundt (1898), with similartransformation, commented on changes in theapparent direction of objects during and afterthe transformation period. Wundt observedthat whereas objects initially appeared angu-larly displaced, this effect progressively di-minished. In probably the best known of allspatial transformation studies, Stratton (1896,1897), rotated the median and horizontalplanes together through 180° about theirlongitudinal axes and noted progressive im-provement in initially disrupted locomotor andmanipulative behavior. The purely perceptualfeatures of this long-term transformation haveremained controversial although later experi-ments (Snyder & Pronko, 1952) have sug-

gested that there were no marked perceptualchanges. Later, Gibson (1933) used wedgeprisms to give a double median plane trans-formation about the transverse axis so thatstraight edges were curved, and he observedthat this perceived curvature progressively di-minished. On removal of the prisms, straightedges appeared curved in the opposite direc-tion. In later experiments, Gibson dispensedwith prisms and established that an inspectedfield of curved lines results in the same effectsas those of prismatic curvature. Subsequentexperiments (Gibson & Radner, 1937) showedthat essentially similar effects occur withslanted lines, thus confirming an earlier find-ing by Verhoeff (1902).

Independent of but parallel with thesevisual transformation investigations were aseries of studies of motor learning which used180° transformations of the frontal plane toproduce a novel learning task. Beginning withStarch (1910), the typical "mirror-drawing"or "mirror-tracing" apparatus prevented adirect view of the hand which was viewed ina mirror so that only part of a fronto-parallelplane was rotated through 180°. The observ-er's task was that of drawing, writing, ortracing a pattern while viewing the hand inthe mirror. The task was further developedand used by Siipola (1935) with the mirrorarranged so that the median and horizontalplanes were also transformed. As in the otherstudies reviewed, over a number of trials therewas a gradual improvement in accuracy.

It can also be noted that while in mostexperiments it has been found that errors inreaching, pointing, and marking occur in theopposite direction after the transformationperiod (Held & Hein, 1958; Held & Schlank,1959), they have also been observed to occurin the same direction as the transformation(Weinstein, Sersen, & Weinstein, 1964). Thiseffect is discussed below.

In summary, investigations of the effects ofspatial transformation have variously empha-sized the behavioral or perceptual outcomes.While Gibson (1933, 1937) was wholly con-cerned with visual perceptual phenomena,Starch (1910) and Siipola (1935) were solelyinterested in changes in motor behavior.Stratton (1896, 1897) reported both percep-tual and motor changes as did Kohler (1964).

310 R. H. DAY AND G. SINGER

Visual and Auditory Adaptation

The perceptual outcomes of spatially trans-formed vision can for the most part be re-produced using appropriate stimulus condi-tions without transformation. After Gibson(1933) demonstrated apparent diminution ofprism-induced curvature with subsequent op-posite curvature when the prism was removed,he showed that inspection of curved and thenstraight lines resulted in the same effects. Theprisms induce a double transformation aboutthe longitudinal axis of the median plane sothat the retinal projections of straight lineslying in a fronto-parallel plane are curved. Anormally viewed curved line of course resultsin a similar retinal projection. Wundt (1898),however, observed that with the median planetransformed about its transverse axis bymeans of prisms there was a progressive dim-inution of the apparent deviation of objectsfrom the straight-ahead. If a field of lineslying in the horizontal plane and parallel tothe median plane is viewed through a wedgeprism, they appear slanted to one side. Withprolonged viewing through the prism the linesappear progressively more straight-ahead. Thisobservation made by one of the authors isessentially that made by Wundt (1898). Ithas also been observed that a field of linesslanted away from the median plane andlying in the horizontal plane appear progres-sively more straight-ahead with prolongednormal viewing.

Essentially similar effects to those demon-strated by Gibson (1933) and Wundt (1898)occur with angular transformations of thefrontal plane. Wertheimer (1912) used alarge slanted mirror so that when an observerstood before it the floors and walls of a roomappeared slanted. This apparent slant dimin-ished with prolonged viewing in the mirror.Although Gibson (1952) has disputed thisfinding, he (Gibson & Bergman, 1959) ob-served an essentially similar effect with a nor-mally viewed tilted surface. Following inspec-tion of a textured surface a subsequently pre-sented vertical surface appeared slanted in theopposite direction (Gibson & Bergman, 1959).

Kohler and Wallach (1944) showed thatfollowing fixation of a figure, the apparentsize of a slightly larger figure increased com-

pared with its apparent size prior to fixation.Similarly, the apparent size of a slightlysmaller figure was decreased. It would be ex-pected, therefore, that if the retinal projectionof an object were reduced in size by means ofa minifying optical system, subsequentlyviewed objects would appear larger. Rock(1965), using a convex mirror to minify reti-nal projections, has demonstrated that suchchanges in judged size occur.

In general terms, visual perceptual changesoccur with spatially transformed space. Ifthe same stimulus conditions are reproducedwithout transformation (e.g., actually curvedlines as opposed to prismatically curved reti-nal projections), essentially similar changesoccur.

Although spatial adaptation and aftereffecthave been observed to occur in hearing (Bart-lett & Mark, 1922; Flugel, 1921; Krauskopf,1954; M. M. Taylor, 1962) with a soundsource to one side of the median plane, thesame effects have not been reported with spa-tially transformed hearing. There is no reasonto suppose, however, that similar effects wouldnot occur if the same conditions were pro-duced by a pseudophonic device. The experi-ments with spatially transformed hearing re-ported by Held (1955) and Freedman andZacks (1964) were concerned with behavioralchanges.

Behavioral Compensation

If a wedge prism transforms the medianplane 20° right about its transverse axis, theobserver reaching for an object quickly willat first move too far to the right. After sev-eral attempts, however, accuracy is achieved;and when, without visual guidance, the ob-server points straight-ahead, his direction ofpointing will be to the left of his pretransfor-mation straight-ahead. The leftward compen-sation for rightward angular transformationpersists. Similar effects can be easily demon-strated with a variety of directions, extents,and complexities of transformation. The es-sential feature of these effects is a change inthe direction of responses such as pointing,reaching, marking, walking, or card sortingduring the transformation period such thatcompensation is made for the effects of trans-formation. The posttransformation responses

SENSORY ADAPTATION AND BEHAVIORAL COMPENSATION 311

in the absence of visual or auditory guidancetend to be in the same direction as those de-veloped during the transformation period.

Independent Occurrence oj the Two Effects

Gibson and Radner (1937) showed thatadaptation to slanted lines achieved a maxi-mum when the field of lines was slanted at15-20° and was negligible at 45°, a findingwhich has been confirmed by Logan (1962)and by Morant and Seller (1965) for pris-matically slanted lines. In a recent unpub-lished experiment by the present authors,a physically horizontal line was viewedthrough a double Dove-prism system (Figure

1) so that the median and horizontal planeswere angularly transformed about their hori-zontal axes. In one condition, the prisms werearranged so that the retinal projection of theline was slanted 20°; and in the other condi-tion, at 45°. The mean aftereffect (differencebetween pre- and posttransformation adjust-ments to horizontality with normal viewing)for two groups of 10 observers each was1.57° (p<.0l) for the 20° slant and .16°(p > .05) for 45°. The transformation periodwas 3 minutes. Thus the visual adaptationeffect is negligible at 45° transformation andsignificant for the smaller slant.

FIG. 1. Apparatus used to induce spatial adaptation by visual inspection of an optically slanted (but ob-jectively horizontal) bar and behavioral compensation by observation of the hand moving across theoptically transformed bar.

312 R. H. DAY AND G. SINGER

If, on the other hand, the observer moveshis extended hand across a physically hori-zontal bar while viewing both hand and bar(Figure 1) and sets it so it seems horizontalat IS-second intervals during the 3-minuteperiod, he will adjust it to compensate forthe visual transformation in the opposite di-rection to the slant. This compensation forvisual transformation persists so that it occurswhen the observer makes an adjustment toapparent horizontality without visual guidanceafter the transformation period. In an exten-sion of the experiment, two groups of 12 ob-servers adjusted the bar so that it seemedhorizontal at IS-second intervals during the3-minute transformation period, observing thebar and their hand moving across it. Theprisms were arranged so that the retinal pro-jection of the horizontal bar was slanted 20°for one group and 45° for the other. Theaftereffect (difference between pre- and post-transformation adjustments without visualguidance) was 6.86° for the 20° transforma-tion and 8.23° for the 45° transformation.

This latter finding is similar in principle tothat reported by Harris (1965) in which theobserver moved his viewed hand across aprismatically curved (but physically straight)line.2 It is clear from the data reported here,however, that the visual and behavioral effectsare independent. Whereas visual adaptationwith aftereffect is negligible at 45° of pris-matically induced slant, the behavioral effectis substantial. At 20° slant, the visual effectis larger than at 45°, while the behavioraleffect is larger at 45° than at 20°. Sincethere is no visual effect at 45°, it seems un-likely that this phenomenon is due to changesin the registration of eye movements as sug-gested by Harris (1965) for prismatically in-duced curvature. Such a hypothesis wouldpredict that the visual effect would be greaterat 45° than at 20°. There is, of course, noessential difference between visual adaptationderiving from the inspection of prismaticallycurved and prismatically slanted lines.

2 Learning to move the hand across an objec-tively straight but prismatically curved line is, ofcourse, essentially similar to a mirror-tracing task.Whereas mirror tracing usually involves a 180°transformation of the frontal plane, Harris's taskinvolved a double median plane transformationabout its longitudinal axis.

Simultaneous Occurrence of the Two Effects

Under a variety of spatial transformationconditions, sensory adaptation and behav-ioral compensations can occur simultaneouslyin the same direction. Consider a wedge prismwhich transforms the median plane through20° to the right set into a sheet of plate glasswith the prism surface flush with the glassand the whole forming the top of an open-sided box. If a rod is placed a few inches be-low the underside of the glass, that sectiondirectly beneath the prism is displaced to theright so that the rod appears to be discon-tinuous. A cover with a line the same widthas the rod and coincident with it can be placedover the glass and prism. With fixed head po-sition, the observer fixates the rod at the cen-ter of the prism for 2 minutes, after which thecover is quickly replaced and the line fixatedat the point coincident with the previous fix-ation point. Since the central section of therod was prismatically displaced right, the samesection of the line appears displaced left. Thisis a typical visual spatial aftereffect. Now, ifwhile viewing the rod the observer quicklyreaches beneath the glass and touches thesection beneath the prism 10 times during a2-minute period, he will at first reach too farright. After several responses, accuracy willbe achieved; and when the cover is replacedand the observer is required to touch itsunderside directly beneath a point coincidentwith the center of the prism, his response willbe too far left. That is,' compensation fortransformed vision during the transformationperiod persists and is revealed in posttransfor-mation responses without visual guidance. Thepoint to note is that the visual spatial after-effect and behavioral aftereffect both deriv-ing from median plane transformation are inthe same (leftward) direction.

The independence of these two effects canbe demonstrated by requiring the observerduring the second part of the experiment toreach out quickly and touch the rod, allowingabout 3 seconds for the response and closinghis eyes for about 20 seconds between eachresponse. In this case, behavioral compensa-tion will be marked but the visual effect neg-ligible since there is insufficient time for it todevelop fully and what does develop will dis-

SENSORY ADAPTATION AND BEHAVIORAL COMPENSATION 313

sipate during the periods when the eyes areclosed (Hammer, 1949; Oyama, 19S3). Withspatial transformation of the total visual fieldby means of prisms worn before the eyes fora long period, both visual and behavioraladaptation with subsequent aftereffects onremoval would be expected to occur.

Whether or not visual adaptation occurs inaddition to behavioral compensation dependson the type and extent of spatial transforma-tion. For example, for the reasons alreadyoutlined, visual adaptation with aftereffectwould not be expected with the median planeangularly transformed 45° about its longi-tudinal axis. Since with 180° transformationsof the median or horizontal planes (or withboth, as studied by Stratton, 1896, 1897),the straight-ahead remains straight ahead andvertical and horizontal edges remain so;adaptation with aftereffect would not be ex-pected.

Behavioral Compensation and KinestketicAftereffect

In addition to vision and hearing, sensoryspatial adaptation with consequent aftereffectoccurs following prolonged kinesthetic stimu-lation (Day & Singer, 1964; Gibson, 1933;Kb'hler & Dinnerstein, 1947). Under certainconditions a kinesthetic aftereffect and be-havioral compensation may occur together inthe same direction. For this reason it is nec-essary to distinguish between them.

If a blindfolded observer moves his ex-tended hand from side to side across a slantededge, a subsequently presented horizontal edgeis judged to be slanted in the opposite direc-tion. If, however, the observer is required toadjust the edge so that it feels horizontal, hewill set it in the same direction as the originalslant. Movement across an objectively hori-zontal edge does not result in an aftereffectif the poststimulation task is that of adjust-ing the edge to apparent horizontality (Day& Singer, 1964). With the apparatus shownin Figure 1, the actual slant of the bar andoptical slant (i.e., median and horizontal planetransformation) can be varied independently.In a recent series of experiments8 in whichbar slant and optical slant were independently

8 These experiments were conducted by Mrs.Margaret Austin, University of Sydney, 1966.

varied in both directions, the observer movedhis hand from side to side across the bar whileviewing it, after which a nonvisual posttrans-formation adjustment to apparent horizontal-ity was made. The results from these experi-ments show that the kinesthetic aftereffectand behavioral compensation may summateor cancel as a function of the extent and di-rection of bar and optical slants. If, for ex-ample, the bar was slanted left and opticalslant was right, the two effects summed. Ifboth bar and optical slant were in the samedirection, the net effect was zero since thekinesthetic aftereffect and behavioral compen-sation occur in opposite directions, as Harris(196S) has implied. If the bar was objectivelyhorizontal but optically slanted, a pure meas-ure of behavioral compensation was obtainedsince, as pointed out above, the kinestheticaftereffect does not occur under this condi-tion.

It seems likely that in some studies theoccurrence of a kinesthetic aftereffect has af-fected the results which have been interpretedin terms of "negative behavioral compensa-tion." Weinstein, Sersen, and Weinstein(1964) attempted to repeat an experiment byHeld and Schlank (19S9) and obtained post-transformation responses opposite in directionto those originally obtained. A mirror systemwas arranged to give a parallel transforma-tion of the frontal plane so that when theresting hand was viewed it appeared at agreater distance from the body than whennormally viewed. During the transformationperiod, the hand was moved rapidly (68movements per minute) toward and awayfrom the body. Before and after this 6-min-ute period the observer was required to marka point without viewing his hand. The rangeof movement of the hand and the position ofmarking strongly suggest that a kinestheticaftereffect occurred in a direction opposite tobehavioral compensation.

NATURE OF SENSORY ADAPTATION ANDBEHAVIORAL COMPENSATION

Sensory Adaptation

The essential effects of optical and pseudo-phonic transformation are changes in the spa-tial properties of stimulation similar to thosewhich can be achieved by appropriate pat-

314 R. H. DAY AND G. SINGER

terns without transformation. The processesunderlying the effects of protracted stimula-tion produced by either means, however, arefar from clear. Although a number of theorieshave been proposed to explain spatial after-effects (Gibson, 1937; Kohler & Wallach,1944; Osgood & Heyer, 1952), none is en-tirely satisfactory. It is sufficient to note forpresent purposes, however, that spatial after-effects in vision occur with retinal stabiliza-tion of the projected image (Ganz, 1964),thus ruling out ocular movement as a deter-minant. Howard and Templeton (1964)demonstrated that effects from slanted linescannot be attributed to ocular torsion asproposed by Ogle (19SO).

Interocular transfer. It is well established(Gibson, 1933; Kohler & Wallach, 1944)that following appropriate stimulation ofone eye with the other occluded, a spatialaftereffect occurs when the occluded eye isused for testing. The occurrence of theeffect in the previously nonstimulated eye isgenerally referred to as interocular transfer.It would be expected that if spatial adapta-tion with aftereffect consequent upon trans-formed vision represents the same effects aswith appropriate nontransformed vision, inter-ocular transfer would also occur with trans-formed stimulation. That this is so has beendemonstrated by Ebenholtz (1966), Hajosand Ritter (196S), and Pick, Hay, and Wil-loughby (1966). The question of whether ornot such demonstrations of interocular trans-fer provide evidence for the central origin ofthe processes involved remains unsettled(Day, 1958).

Behavioral Compensation

It is argued here that changes in behaviorconcomitant with transformed visual andauditory input are a special case of motorlearning essentially similar to and exhibitingthe same characteristics as the learning proc-ess observable with the pursuit-rotor, finger-maze or mirror-tracing apparatus. To this ex-tent it seems unnecessary to attribute specialproperties to behavior deriving from trans-formed sensory input as has been common inrecent contributions (Harris, 1965; Held &Freedman, 1963). That is, it seems unneces-

sary and unparsimonious to regard changesin behavior deriving from transformation asessentially different from changes occurringin a variety of alternative motor-learningsituations.

If an observer wearing wedge prisms whichtransform the median plane transversely 20°right regards an object which is objectivelystraight-ahead, it appears 20° to the right. Inreaching quickly for the object, the observermay at first miss it but after several attemptshe will reach accurately. If reaching is con-tinued, proprioceptive information for a par-ticular direction of limb movement (previ-ously associated with the nontransformed me-dian plane) and visual information becomeassociated. In brief, an appropriate responsepattern to the spatial properties of stimula-tion is learned. When the prisms are removedand the observer reaches straight ahead with-out visual guidance, the response deviates leftof the actual straight-ahead point. Since theoriginal proprioceptive information forstraight-ahead has become associated with 20°right, the new nontransformed straight-ahead(without visual guidance of the limb) mustof necessity be to the left. In this sense, be-havioral compensation is representative of thelearning process which occurs in a wide vari-ety of situations and is subjected to extensiveenquiry. An appropriate response pattern de-velops in relation to a certain pattern of stim-ulation. Presented with a target moving rap-idly in a circular path, as with the pursuit-rotor, the individual, after initial inaccuracies,exhibits an appropriate response pattern whichimproves with practice. If the frontal plane istransformed through 180° by means of amirror placed before the observer and he isrequired to trace a star pattern while observ-ing his hand in the mirror, a typical learningcurve can be plotted on the basis of eithertime taken or number of errors per trial. Ifthe median plane is less drastically trans-formed through 20° and the observer is re-quired to reach for an object while observinghis hand, a typical learning curve results(Hamilton, 1964). The main difference be-tween early studies of this process (Siipola,1935; Starch, 1910) and more recent inves-tigations (Held & Freedman, 1963) is thatthe early experiments did not make use of

SENSORY ADAPTATION AND BEHAVIORAL COMPENSATION 315

pre- and posttransformation tests to measurethe transfer of learning to nontransformedconditions, and the term "sensorimotor adap-tation" has been coined in referring to thechange. Further, in recent investigationsthere has been a failure to distinguish clearlybetween the purely sensory (adaptive) phe-nomena and learning or to recognize that thefrequently used median plane transformationsachieved by wedge prisms are identical inprinciple to the frontal plane transformationsderiving from mirrors. The classical mirror-tracing procedure is a standard procedure fordemonstrating the course of motor learningand bilateral transfer of learning. Althoughthe learning is more rapid (since the trans-formation is smaller and less complex), me-dian plane transformations or any other spa-tial transformation would serve as well. Al-though the evidence strongly suggests thatchanges in behavior with transformation rep-resent a motor-learning process, the term "be-havioral compensation" can be convenientlyused in referring to learning under these con-ditions.

Rate oj behavioral compensation. The evi-dence suggests that the rate of motor learn-ing with transformed input varies as a func-tion of both the extent and complexity oftransformation and the nature of the re-sponses. In those studies involving 180°transformations of median and horizontalplanes about longitudinal axes (Ewert, 1930;Snyder & Pronko, 1952; Stratton, 1896, 1897)and involving complex response patternssuch as perambulation and card sorting, thetime to achieve pretransformation efficiencywas protracted, taking up to 2 or 3 days.Learning with simpler and less extensivetransformations involving less complex re-sponses such as reaching or marking is rela-tively rapid, occupying a few seconds (Hamil-ton, 1964). So far, there has been no syste-matic study of the effects of extent, direction,and complexity of transformation on motorlearning, although Siipola (193S) investigatedthe effects of 180° transformation of thefrontal and horizontal planes singly and to-gether. She found no difference between thethree conditions. The accumulated data froma number of studies indicate, however, thatthe extent of transformation may be a criti-

cal variable affecting rate of learning to a setcriterion.

Intermanual transfer oj behavioral compen-sation. Kalil and Freedman (1966) foundthat after viewing one hand with transformedvision behavioral compensation occurred inthe other hand in a nonvisual posttransforma-tion test. This finding was not confirmed byHarris (1963) or Mikaelian (1963). Hamil-ton (1964) noted intermanual transfer withunrestricted head and body movement butfound that it did not occur with restrictedmovement. Intermanual or, as it was calledearlier, bilateral transfer, is well establishedwith 180° transformation of the frontal planein the typical mirror-tracing experiment(Ewert, 1926; Siipola, 1935). In fact, it hasalso been found to occur between hand andfoot (Bray, 1928). Since the recent investiga-tions of intermanual or bilateral transfer havebeen restricted to relatively small spatialtransformations usually of the median planewhereas the earlier studies employed 180°changes, it seems that transformation extentmay be a determinant of such transfer. Afurther point of difference between early andrecent studies concerns the procedure formeasuring transfer. The older procedure(Ewert, 1926; Siipola, 1935) involved a seriesof trials viewing one hand with transformedvision followed by trials with the other handunder the same conditions. More recent stud-ies (Harris, 1963; Mikaelian, 1963) have in-vestigated transfer from transformed visionof one hand to a nonvisual test with the other.A double transfer is involved in this proce-dure—transfer between a transformed visualcondition and a nonvisual condition andtransfer between one hand and the other. Itis probable that the earlier transfer studiesconstituted a more sensitive test of transfer ofmotor learning between limbs and weresounder in terms of experimental design. The"double transfer" procedure must lead to aconfounding of transfer from visual to non-visual conditions with transfer from one limbto the other.

Intermodal transfer. Harris (1963) demon-strated that following behavioral compensa-tion for median plane transformation, thecompensation occurred in pointing to the di-rection of a sound source. Since visual and

316 R. H. DAY AND G. SINGER

auditory directions are normally in accord(e.g., the telephone is both seen and heardto the left or right), it follows that if newmotor responses are learned with transformedvision they would also be manifest during anonvisual posttransformation test with anauditory stimulus. So far, the opposite trans-fer from hearing to vision has not been dem-onstrated although it would clearly be of in-terest to do so.

Further Evidence for Behavioral Compensa-tion as a Special Case of Motor Learning

In a recent experiment (Day & Singer,1966) using the apparatus shown in Figure1, the number of trials during the transfor-mation period and during the posttransforma-tion phase were systematically varied. The barwhich was objectively horizontal was opti-cally transformed through 20° and the ob-server's task was that of adjusting it to theapparent horizontal. Nonvisual posttransfor-mation adjustments were compared with thosemade similarly before transformation. Threegroups made a single adjustment to the hori-zontal while viewing the angularly trans-formed bar, and three further groups madeseven adjustments. While two groups (1 and7 trials) made IS posttransformation adjust-ments at 1-minute intervals, two furthergroups (1 and 7 trials) made a single ad-justment after 7 minutes, and the remaining

£3-

""Gp.4

0 1 2 3 4 6 6 7 8 9 10 11 12 13 14INTERVAL AFTER EXPOSURE(Mia)

FIG. 2. Dissipation of behavioral compensationunder two conditions of transformation (1 and 7trials) and three conditions of posttransformation(15 trials at 1-minute intervals, 1 trial after 7minutes and 1 trial after 14 minutes).

two groups (1 and 7 trials) made a singleadjustment after IS minutes. The data fromthis experiment are plotted for the six condi-tions in Figure 2. It is clear from the graphthat the frequency of transformation trialsdetermines the degree of learning and thatdissipation of the learned response during theposttransformation period is accelerated as afunction of the number of posttransformationtrials without visual guidance. If the transfor-mation period is regarded as a training pe-riod during which new relationships betweenstimulus and response are learned, then theposttransformation phase can be treated as anextinction period. During this latter period itwould be expected that in the absence of trans-formed stimulation the recently acquired re-sponses would give way to previously learnedand more firmly established responses.

METHODOLOGICAL ISSUES

By far the most common procedure in in-vestigating sensory spatial adaptation and be-havioral compensation is that involving testsbefore and after a period of spatial transfor-mation. The difference between pre- and post-transformation measures serves as an index ofeither visual or auditory adaptation or of be-havioral compensation. The use of this pro-cedure raises two questions—that concerningthe appropriateness of the tests to reveal oneor the other effects, and that concerningtransfer of learning from the transformationperiod to the posttransformation test.

Appropriateness of Pre- and Posttransforma-tion Tests

It has been pointed out that both sensoryspatial adaptation and behavioral compensa-tion may occur simultaneously. Whether oneor both effects are revealed depends on thenature of the pre- and posttransformationtests. If after a period of wearing wedgeprisms to transform the median plane throughits transverse axis the observer is requiredto point straight ahead without viewing hishand and arm, then behavioral compensationwill be exhibited. If, on the other hand, theobserver is required to report when a field oflines in the horizontal plane appears straightahead, then a perceptual change will be re-vealed.

SENSORY ADAPTATION AND BEHAVIORAL COMPENSATION 317

In a recent experiment, Mikaelian and Held(1964) used a dual prism system to trans-form spatially the median plane through itslongitudinal and transverse axes and the hori-zontal plane through its transverse axis. Theoutcome of this complex transformation wasthat a point normally straight ahead was dis-placed left and upward and a normally verti-cal line was rotated into a slanted position.There were two tests before and after trans-formation: adjustment of the line to apparentverticality and rotation of the body by legmovement until a point of light appearedstraight ahead. The first is a perceptualtest and the second is a behavioral test.Using the conditions of Mikaelian and Held(1964), it would be possible to devisea behavioral test in relation to the slantedline (e.g., pointing to the top of the line) anda perceptual test for the straight-ahead (e.g.,adjusting a point to appear straight ahead).As a matter of fact, McLaughlin and Rifkin(1965) have shown that following transverseangular transformation of the median plane, aspatial aftereffect occurs in adjusting apointer to the straight-ahead. It is clear thatappropriate behavioral and perceptual testscould be devised for the complex transforma-tion used by Mikaelian and Held (1964)such that behavioral compensation and sen-sory spatial adaptation could be shown tooccur for the two aspects of transformationseparately investigated.

In a large number of recent investigations,perceptual tests have preceded and followeda transformation period during which theobserver was required to engage in variousactivities. Thus, Ebenholtz (1966), usingangular transformation of median and hori-zontal planes, required observers to walkthrough long corridors and to undertake vari-ous manipulative tasks (jigsaw puzzles, dartthrowing). The pre- and posttransformationtests involved judgments of verticality. It ishighly likely that if the observers had merelyinspected a field of vertical lines for the sameperiod of time the same results would havebeen obtained since spatial adaptation is de-pendent on stimulation rather than motor ac-tivity during the transformation period. Thesame argument applies in the case of an ex-periment by Freedman and Stampfer (1964)

in which judgments of the auditory straight-ahead were made before and after a period ofpseudophonically transformed hearing duringwhich the observer walked about. Had the ob-server been seated for the same period thesame results probably would have occurred.

Transfer from Transformation to Test Phase

The experimental situation in which an ob-server learns to make certain responses withspatially transformed vision or hearing fol-lowed by responses with normal sensory in-put (but usually with his responding limboccluded) is similar in many respects tomany transfer-of-learning experiments. Theobserver learns to make appropriate responseswith one form of sensory input and later re-sponds with another. The posttransformationresponses (reaching, pointing, marking) tendto be in the direction of those developed dur-ing the transformation phase. That is, re-sponses learned under one condition transferto another. In this sense, it would be expectedthat the magnitude of behavioral compensa-tion as revealed by the difference between pre-and posttransformation tests would be a func-tion of those variables known to affect trans-fer of learning. The relationship betweentransformation and posttransformation phasesin terms of task similarity, relative task diffi-culty, degree of learning in the transforma-tion phase, and time interval (Ellis, 1965)would be expected to affect compensation.The data from Day and Singer (1966) whichindicated that one trial during the transfor-mation phase resulted in less compensationthan seven trials are in agreement with atransfer-of-learning interpretation. Furtherevidence in support of this view was providedby Freedman, Hall, and Rekosh (1965) whoshowed that similar activities during trans-formation and posttransformation phases re-sulted in greater transfer than dissimilar ac-tivities.

THEORETICAL ISSUES

In view of the distinction which has beenmade between sensory adaptation to and be-havioral compensation for spatially trans-formed vision and the identification of thelatter as motor learning, the role of active andpassive responding (Held & Freedman, 1963)

318 R. H. DAY AND G. SINGER

and the felt-position hypothesis (Harris,1965) can be considered.

Active and Passive Responding duringTransformation

Held and his associates (Held & Freedman,1963; Mikaelian & Held, 1964) proposed thatthe self-induced movement is a fundamentaldeterminant of what they called sensorimotoradaptation to spatially transformed input.Recent experiments have shown, however, thatfor some conditions of transformation, exter-nally induced or "passive" movement duringtransformation is as effective as self-inducedor "active" movement (Pick & Hay, 196S;Singer & Day, 1966b; Templeton, Howard,& Lowman, 1966). In a recent study, (Singer& Day, 1966a) it was shown that when thepassive, resting limb was viewed through awedge prism, behavioral compensation oc-curred if the observer was required to judgethe position of his limb in terms of a movingscale located between prism and eye. If, onthe other hand, the observer merely viewedhis limb without judging position, compensa-tion was negligible.

If, as argued here, changes in behavior withspatial transformation represent learning, itfollows that the observer, in order to compen-sate for the altered spatial input, must begiven an opportunity to learn. Although somelearning would be expected to occur as aresult of merely observing the resting limb orthrough observing another individual learning(Siipola, 1935), maximum learning would bemore likely to occur when the observer ac-tually responds. It does not seem at all sur-prising that when the observer's responses arerestricted, as when he is moved about in awheelchair (Held & Bossom, 1961), or whenthe movements are induced by an externalforce, he learns less. It has been pointed outthat the occurrence of responses learned dur-ing the transformation phase in the post-transformation phase can reasonably be re-garded as an instance of transfer of learning.If little learning takes place during transfor-mation as a consequence of "passive" move-ment (i.e., limited opportunity to learn) andthe responses in the two phases are markedlydifferent, it is to be expected that little or notransfer would be manifested.

There is a further point which arises inconnection with Held's theory. Proprioceptivefeedback, which is central to his argument, isby no means dependent on self-induced move-ment as implied in the theory. Stimulation ofreceptors mediating the sense of position andmovement occurs with passive as well as ac-tive movement (Mountcastle & Powell, 1959).Only the muscle spindles rely on self-inducedmovement for their stimulation. Thus, follow-ing passive movement during the transforma-tion phase, transfer would be expected tooccur to a passive posttransformation task.Less transfer of learning would be likely tooccur from a task requiring passive move-ment to one demanding active movement, asis common with tasks dissimilar in the typeof responses involved.

Felt-Position Hypothesis

Harris (1965) has proposed that changesoccurring with spatially transformed inputcan be largely attributed to changes in theposition sense. He called the perception ofposition "felt position" but made clear thatthe hypothesis applies when position informa-tion is unconscious.

In discussing his own experiments, Harriscontended that since the observer was notallowed to make any active movement withhis viewed arm during transformation, a sim-ple motor-learning interpretation is inade-quate to explain the changes which occurred.But it is well established that even inspectionof a pictorial representation of a motor taskaffects the learning of that task positively(Gagne & Baker, 1950). In fact, Siipola(1935) showed that subjects who observedothers performing the mirror-tracing taskbenefitted from such perceptual pretraining.In other words, it is not absolutely necessaryto perform the task to learn it; merely ob-serving the arm or hand with spatially trans-formed vision is sufficient to result in somelearning. The argument advanced by Harristhat the elimination of active movement whileviewing the hand through a prism attests tothe inadequacy of a motor-learning explana-tion ignores the role of perceptual pretrainingin motor learning.

The second point, although more abstract,is more fundamental in evaluating Harris's

SENSORY ADAPTATION AND BEHAVIORAL COMPENSATION 319

explanation in terms of the felt position oflimbs or body parts. The study of perceptionhas long been as equally concerned with phe-nomena as it has been with responses. De-scriptive phenomenology and behavioral psy-chophysics represent the extremes of a con-tinuum of viewpoint with shades of approachbetween. The study of learning, on the otherhand, has been almost exclusively concernedwith the objective observation, recording, andmeasurement of responses. Subjective dataare seldom sought. But, and this is the coreof the argument, responses are accompaniedby sensory stimulation. The proprioceptiveend-organs function in the maintenance andcontrol of posture and movement. If a new setof responses is learned, a new set of proprio-ceptive perceptual phenomena must also oc-cur. If after many months of reaching to theright to pick up the telephone, the positionof the telephone is altered to the left, the indi-vidual tends to continue reaching right. Grad-ually, he learns to reach left and the right-reaching response disappears. At the sametime there is a change in the proprioceptive"sensation" accompanying the movements.Reaching left is accompanied by differentfelt-positions and felt-movements than reach-ing right. If with median plane transformationthe observer learns to reach right to pick upan object which is actually straight ahead,the learning which occurs has its sensorycomponent. To argue that changes in behav-ior with spatially transformed input are dueto changes in proprioception is merely tooffer an explanation in terms of the proprio-ceptive accompaniments of learning. If, asHarris (1965) has indicated, this interpreta-tion applies when position information is notavailable to consciousness, then there wouldseem to be little difference between the felt-position hypothesis and one couched in termsof motor learning.

INTERACTION BETWEEN SENSORY AND MOTORPROCESSES: DIRECTION-CONTINGENT EFFECTS

The data treated so far strongly suggestthat under conditions of visual and auditorytransformation, independent sensory effectsand learning occur. While sensory spatialadaptation is essentially similar to that oc-curring with appropriate nontransformed

stimulation, the learning effects are the samein principle to those observable in a varietyof alternative situations involving perceptual-motor coordination. There is, however, athird group of phenomena which, althoughapparently related to the sensory effects, has amotor component. These have been termed"gaze-contingent" effects by Pick and Hay(1966) and were first observed and intensivelystudied by Kohler (1964).

The refraction of light passing through aprism is a function of the angle at which lightenters the prism. If when viewing through aprism the eyes are turned right, the transfor-mation is different from that when the eyesare turned left. The light falling on the eyeenters the prism at a different angle for eachdirection of regard. Similar variations in thecharacteristics of transformation occur whenthe eyes remain in a fixed position and thehead is turned from right to left. Kohler(1964) has shown that with prolonged wear-ing of prisms and other devices, adaptationto these gaze-contingent transformations takesplace but that on removal, the consequentaftereffect varies with direction of regard.That is, if the eyes are turned one way, theaftereffect is different from that when theeyes are turned another. It is also claimedthat if spectacles whose eyepieces consist ofhalf blue and half yellow glass are worn for asufficiently long period, chromatic aftereffectswhich are gaze-contingent also occur. Thislatter observation was not confirmed by Har-rington (1965), one of whose observers worethe bicolor lenses for 146 days with nomarked changes.

Spatial aftereffects of the type studied byVerhoeff (1902) and Gibson (1933) can beinduced either by transformation or by suita-ble nontransformed stimulation. The questionimmediately arises concerning whether gaze-contingent effects might also occur with non-transformed stimulation. Both Carlson (1964)and Hein and Sekuler (1959) reported thatadaptation to curved and tilted lines can varyas a function of eye position and head posi-tion with nontransformed vision. Carlson(1964) proposed that perceptual adaptationis basically an immediate localized conse-quence of stimulation but can become gen-eralized through a process of conditioning by

320 R. H. DAY AND G. SINGER

temporal contiguity. Thus the data reportedby Kohler (1964) with transformed visionand by Carlson (1964) and Hein and Sekuler(19S9) with normal viewing indicate thatsensory spatial adaptation with aftereffectunder conditions of varying direction of gazemay become associated with visual direction.

Another question concerns the possible oc-currence of direction-contingent effects insenses other than vision. Since sensory spatialadaptation and aftereffect occur in the audi-tory modality, it is conceivable that direction-contingent auditory effects might also occur.A pseudophone with a fixed sound sourcewould presumably result in variations in spa-tial and other properties with movements ofthe head. In view of the possibility of sucheffects occurring with sensory stimulationother than vision, the term "direction-con-tingent" is preferable.

CONCLUSIONS

Over SO years ago a 180° rotation of thefrontal plane about its longitudinal axis wasused by Starch to demonstrate the course ofwhat he called "trial-and-error learning." Thistechnique has since been used in a large num-ber of studies to demonstrate numerous fea-tures of the learning process, including that ofbilateral transfer. More recently, other classes,directions, and extents of visual and auditorytransformation have been used to study simi-lar changes in behavior. For the most part,recent investigations have tended to regardthese behavioral changes as unique. An analy-sis of spatial transformation in terms of angu-lar and parallel shifts in the median, hori-zontal, and slanted planes shows that there isno difference in principle between frontalplane transformations induced by mirrors andmedian and horizontal plane changes inducedby prisms, lenses, and other optical systems.In addition to motor learning under theseconditions, however, there also occur sensoryadaptive effects of a type essentially similarto those which take place with appropriatenon transformed stimulation. With spatialtransformation, both effects may occur simul-taneously and in the same direction. The basicargument proposed here is that these effectsare independent but since they occur togetherunder conditions of optical and pseudophonic

transformation of space they have been con-fused. Perceptual tests have frequently beenused to index motor learning, and behavioraltests have been used to demonstrate percep-tual changes. To further complicate the pic-ture, Ivo Kohler and his colleagues haveshown and others (Pick & Hay, 1966) haveconfirmed that varying visual transformationwith eye movement results in gaze-contingentadaptation with aftereffect. Such direction-contingent effects may also occur with appro-priate nontransformed stimulation. It is alsoconceivable that such effects may take placein hearing as well as vision.

Since behavioral compensation represents aspecial case of motor learning, it is to beexpected that amount of practice, task diffi-culty, and other factors determine its causeand dissipation. Experimental data have beenreported to show that this is so and thatthe method of studying the learning process isessentially a transfer-of-learning paradigm.

REFERENCES

BARTLETT, F. C., & MARK, H. A note on local fatiguein the auditory system. British Journal of Psy-chology, 1922, 13, 215-218.

BRAY, C. W. Transfer of learning. Journal of Ex-perimental Psychology, 1928, 11, 443-467.

CAKLSON, V. R. Eye movement and adaptation tocurvature. Scandinavian Journal of Psychology,1964, 5, 262-270.

CZERMAK, J. Physiologische Studien, III. Sitzungs-berichte Akademie der Wissenschaften in Wien,1855, 17, 563-600.

DAY, R. H. On interocular transfer and the centralorigin of visual aftereffects. American Journal ofPsychology, 1958, 71, 784-789.

DAY, R. H., & SINGER, G. Spatial aftereffects withinand between kinesthesis and vision. Journal ofExperimental Psychology, 1964, 68, 337-343.

DAY, R. H., & SINGER, G. The basis of behavioralchanges with spatially transformed vision. Paperread at International Congress of Psychology,Moscow, August 1966.

EBENHOLTZ, S. M. Adaptation to a rotated visualfield as a function of degree of optical tilt and ex-posure time. Journal of Experimental Psychology,1966, 72, 629-634.

ELLIS, H. The transfer of learning. New York: Mac-millan, 1965.

EWERT, P. H. Bilateral transfer in mirror drawing.Journal of Genetic Psychology, 1926, 33, 235-249.

EWERT, P. H. A study of the effect of inverted ret-inal stimulation upon spatially coordinated behav-ior. Genetic Psychology Monographs, 1930, 7, Nos.3-4.

SENSORY ADAPTATION AND BEHAVIORAL COMPENSATION 321

FLUOEL, J. C. On local fatigue in the auditory sys-tem. British Journal of Psychology, 1921, 11, 105-134.

FREEDMAN, S. J., HALL, S. B., & REKOSH, J. H.Effects on hand-eye coordination of two differentarm motions during compensation for displacedvision. Perceptual and Motor Skills, 1965, 20,1054-1056.

FREEDMAN, S. J., & STAMPFER, K. Changes in audi-tory localization with displaced ears. Paper readat Psychonomic Society, Niagara Falls, Ontario,October 1964.

FREEDMAN, S. J., & ZACKS, J. I. Effect of active andpassive movement upon auditory function duringprolonged atypical stimulation. Perceptual andMotor Skills, 1964, 18, 361-366.

GAGNE, R. H., & BAKER, K. E. Stimulus prediffer-entiation as a factor in transfer of training. Jour-nal of Experimental Psychology, 1950, 40, 439-451.

GANZ, L. Lateral inhibition and the location of vis-ual contours: An analysis of flgural aftereffects,Vision Research, 1964, 4, 465-481.

GIBSON, J. J. Adaptation, aftereffect and contrast inthe perception of curved lines. Journal of Experi-mental Psychology, 1933, 16, 1-31.

GIBSON, J. J. Adaptation with negative aftereffect.Psychological Review, 1937, 44, 222-243.

GIBSON, J. J., & BERGMAN, R. The negative after-effect of the perception of a surface slanted in thethird dimension. American Journal of Psychology,1959, 3, 364-374.

GIBSON, J. J., & RADNER, M. Adaptation, aftereffect,and contrast in the perception of tilted lines. I.Quantitative studies. Journal of Experimental Psy-chology, 1937, 20, 453-467.

HAJOS, A., & RITTER, M. Experiments to the problemof interocular transfer. Ada Psychologica, 1965,24, 81-90.

HAMILTON, C. R. Intermanual transfer of adapta-tion to prisms. American Journal of Psychology,1964, 77, 457-462.

HAMMER, E. R. Temporal factors in flgural after-effects. American Journal of Psychology, 1949, 62,337-354.

HARRINGTON, T. L. Adaptation of humans to coloredsplit-field glasses. Psychonomic Science, 1965, 3,71-72.

HARRIS, C. S. Adaptation to displaced vision: Visual,motor, or proprioceptive change? Science, 1963,140, 812-813.

HARRIS, C. S. Perceptual adaptation to inverted,reversed, and displaced vision. Psychological Re-view, 1965, 72, 419-444.

HEIN, A. V., & SEKULER, R. A technique for pro-ducing displacement aftereffects contingent uponeye and head position. American Psychologist,1959, 14, 437. (Abstract)

HELD, R. Shifts in binaural localization after pro-longed exposure to atypical combinations of stim-uli. American Journal of Psychology, 1955, 68,526-548.

HELD, R., & BOSSOM, J. Neonatal deprivation andadult rearrangement: Complementary techniques

for analyzing plastic sensori-motor coordinations,Journal of Comparative and Physiological Psy-chology, 1961, 61, 33-37.

HELD, R., & FREEDMAN, S. J. Plasticity in humansensorimotor control. Science, 1963, 142, 455-462.

HELD, R., & HEIN, A. Adaptation of disarrangedhand-eye coordination contingent upon re-afferentstimulation. Perceptual and Motor Skills, 1958, 8,87-90.

HELD, R., & SCHXANK, M. Adaptation to disarrangedeye-hand coordination in the distance-dimension.American Journal of Psychology, 1959, 72, 603-605.

HELMHOLTZ, H. VON Treatise on physiological optics.(Orig. publ. 1866) Vol. 3. (Trans. & Ed. by J. P.Southall) New York: Dover, 1962.

HOWARD, I. P., & TEMPLETON, W. P. Visually-in-duced eye torsion and tilt adaptation. Vision Re-search, 1964, 4, 433-437.

KALIL, R. E., & FREEDMAN, S. J. Intermanual trans-fer of compensation for displaced vision. Percep-tual and Motor Skills, 1966, 22, 123-126.

KOHLER, I. The formation and transformation ofthe perceptual world. (Trans, by H. Fiss) Psycho-logical Issues, 1964, 3(4).

KOHLER, W., & DINNERSTEIN, D. Figural aftereffectsin kinesthesis. In A. Michotte, Miscellanea psycho-logica. Louvain: Catholic University of Louvain,1947. Pp. 196-220.

KOHLER, W., & WALLACE, H. Figural aftereffects:An investigation of visual processes. Proceedingsof the American Philosophical Society, 1944, 88,264-357.

KRAUSKOPF, J. Figural aftereffects in auditory space.American Journal of Psychology, 1954, 67, 278-287.

LOGAN, J. A. An examination of the relationship be-tween visual illusions and flgural aftereffects. Un-published doctoral dissertation, University of Syd-ney, 1962.

MCLAUGHLIN, S. C., & RIPKIN, K. I. Change instraight-ahead during adaptation to prism. Psycho-nomic Science, 1965, 2, 107-108.

MIKAELIAN, H. Failure of bilateral transfer in modi-fied eye-hand coordination. Paper read at EasternPsychological Association, New York, April 1963.

MIKAELIAN, H., & HELD, R. Two types of adapta-tion to an optically-rotated visual field. AmericanJournal of Psychology, 1964, 77, 257-263.

MOUNTCASTLE, V. G., & POWELL, T. P. S. Centralneural mechanisms subserving position sense andkinesthesis. Bulletin of the Johns Hopkins Hospi-tal, 1959, 105, 173-200.

MORANT, R. B., & BELLER, H. K. Adaptation toprismatically rotated visual fields. Science, 1965,148, 530-531.

OGLE, K. N. Researches in binocular vision. Phila-delphia: Saunders, 1950.

OSGOOD, C. E., & HEYER, A. W. A new interpreta-tion of figural aftereffects. Psychological Review,1952, 59, 98-118.

OYAMA, T. Experimental studies of figural after-effects: I. Temporal factors. Japanese Journal ofPsychology, 1953, 23, 239-254.

322 R. H. DAY AND G. SINGER

PICK, H. L., & HAY, J. C. A passive test of theHeld reafference hypothesis. Perceptual and Mo-tor Skills, 1965, 20, 1070-1072.

PICK, H. L., & HAY, J. C. Gaze-contingent adapta-tion to prismatic distortion. American Journal ofPsychology, 1966, 79, 443-450.

PICK, H. L., JR., HAY, J. C., & WILLOUGHBY, R. H.Interocular transfer of adaptation to prismatic dis-tortion. Perceptual and Motor Skills, 1966, 23(1),131-135.

ROCK, I. Adaptation to a minified image. Psycho-nomic Science, 1965, 2, 105-106.

SUPOLA, E. M. Studies in mirror drawing. Psycho-logical Monographs, 1935, 46(6, Whole No. 210),66-77.

SINGER, G., & DAY, R. H. The effects of spatialjudgments on the perceptual aftereffect resultingfrom transformed vision. Australian Journal ofPsychology, 1966, 18, 63-70. (a)

SINGER, G., & DAY, R. H. Spatial adaptation andaftereffect with optically transformed vision: Ef-fects of active and passive responding and the re-lationship between test and exposure responses.Journal of Experimental Psychology, 1966, 71,725-731. (b)

SMITH, K. U., & SMITH, W. K. Perception and mo-tion. Philadelphia: Saunders, 1962.

SNYDER, F. W., & PRONKO, N. H. Vision with spatialinversion. Wichita: University of Wichita Press,1952.

STARCH, D. A demonstration of the trial and errormethod of learning. Psychological Bulletin, 1910,7, 20-23.

STRATTON, G. M. Some preliminary experiments onvision without inversion of the retinal image. Psy-chological Review, 1896, 3, 611-617.

STRATTON, G. M. Vision without inversion of theretinal image. Psychological Review, 1897, 4, 341-360, 463-481.

STRATTON, G. M. The spatial harmony of touch andsight. Mind, 1899, 8, 492-505.

TAYLOR, J. G. The behavioral basis of perception.New Haven: Yale University Press, 1962.

TAYLOR, M. M. The distance paradox of the figuralaftereffect in auditory localization. Canadian Jour-nal of Psychology, 1962, 16, 278, 282.

TEMPLETON, W. B., HOWARD, I. P., & LOWMAN, A.E. Passively generated adaptation to prismaticdistortion. Perceptual and Motor Skills, 1966, 22,140-142.

VERHOEFF, F. H. A theory of binocular perspective,and some remarks upon torsion of the eyes, thetheory of the vicarious fovea, and the relation ofconvergence to the perception of relief and forms.Annals oj Ophthalmology, 1902, 11, 201-229.

WEINSTEIN, S., SERSEN, E., & WEINSTEIN, D. S. Anattempt to replicate a study of disarranged eye-hand coordination. Perceptual and Motor Skills,1964, 18, 629-632.

WERTHEIMER, M. Experimentelle Studien iiber dasSehen von Bewegung. Zeitschrift fur Psychologie,1912, 61, 121-165.

WILLEY, C. F., INGLIS, E., & PEARCE, C. H. Reversalof auditory localization. Journal of ExperimentalPsychology, 1937, 20, 114-130.

WUNDT, W. Zur Theorie der raumlichen Gesicht-wahrnehmungen. Philosophiche Studien, 1898, 14,11.

YOUNG, P. T. Auditory localization with acousticaltransposition of the ears. Journal of ExperimentalPsychology, 1928, 11, 399-429.

(Received May 19, 1966)