0042~6989191 $3.00 + 0.00 Copyright Q 1991 Pergamon Press plc

Yision Res. Vol. 31, No. 4. pp. 717-734, 1991 Printed in Great Britain. All rights reserved

PROPERTIES OF THE FEEDBACK LOOPS CONTROLLING EYE GROWTH AND REFRACTIVE IN THE CHICKEN

STATE

FRANK SCHAEFFEL’*and HOWARD C. HOWLAND’

fiir Experimentelle Ophthaimolo~e (Leiter: E. Zrenner), Ob dem Himmelre~ch 9, D-7400 Tiibingen, F.R.G. and 2Section of Neurobjolo~ and Behavior, Cornell University, Ithaca, NY 14853, USA, lFo~hungsstelle

(Received 8 November 1989; in revised form 23 April 1990)

Abstract-Recent experiments in chickens provide evidence that axial eye growth and refractive state are guided by mechanisms sensitive to refractive error. To determine whether or not the sign of refractive error is derived from longitudinal chromatic aberration we raised chicks with spectacle lenses in monochromatic light. The eyes showed an appropriate growth response to correct for the defocus imposed by the lenses no different than in previous experiments in white light. Thus, in normally accommodating chicks chromatic cues are not necessary for emmetropi~tion to occur. We examined the linearity of feedback loops controlling axial eye growth: positive spectacle lenses were found to inhibit axial growth very efficiently making the eyes shorter than normal whereas negative lenses had little effect on axial elongation: feedback loops for regulation of axial growth are highly nonlinear and act most efficiently on the myopic side. We found that, subsequent to a period of binocular deprivation of form vision, the refractive errors acquired are highly correlated in both eyes. Since both eyes grew without visual feedback we conclude that the gains in the feedback loops that control axial growth must be similar in both eyes. We suggest that the gains are genetically determined and are typical for each individual. Chicks made near-sighted in both eyes by “deprivation of form vision” were corrected by appropriate negative lenses. Three out of five chicks recovered from myopia despite the correction. Also two chicks that were made near-sighted in one eye recovered with no regard to the correcting lens. Three chicks remained more myopic than the correcting lens required and finally started to recover while the lens was still in place. Two out of three chicks that were made far-sighted showed recovery despite appropriate correction by positive lenses. We conclude that there must be a nonvisual mechanism highly sensitive to abnormal eye shape. During expt (4) we found unexpectedly that the development of form deprivation myopia is inhibited if no part of the retina in an animal is exposed to normal visual experience. The result indicates that some communication between both eyes exists, although form deprivation myopia itself has been shown to develop independently in both eyes. Refraction

Eye growth

Emmetropization

Chicken

Eyes of vertebrates can attain a high optical precision. Particularly, the focal Iength of the optics is normally well matched to the axial length of an eye such that the focal plane is coincident with the photoreceptor plane. On the other hand deviations from the optimal condition are frequently encountered in humans and increasing interest has been focused on the question of how such a precise matching is attained in emmetropic eyes. A number of recent experimental studies in animals have shown that axial eye growth is influenced by visual experience and, in the chicken, evidence is accumulating that axial *To whom correspondence

should be addressed.

growth is indeed controlled by mechanisms sensitive to refractive error. Although the idea of a “feedback regulation of axial eye growth by mechanisms sensitive to refractive error” is a plausible one, it has not yet been shown that such mechanisms exist in other animal models than the chicken, and recent experiments in cats treated with lenses (Ni & Smith, 1989) failed to show that compensatory changes in axial growth occur in both directions as should be expected from the signs of the lenses used. Similarly, in a study where contact lenses were used to defocus the eyes of cynomolgus monkeys, no consistent correlation between refractions after treatment and sign and power of the lenses were seen (Crewther, Nathan, Kiely, Brennan & Crewther, 1988). However, there is a basic difference in our lens experiments

717

718

FRANK SCHAEFFEL and HOWARD C. HOWLAND

in the chicken: monkeys and cats are frontaleyed animals with the accommodation coupled in both eyes. Since the treatments always occurred monocularly (to allow for an untreated control eye) accommodative tonus was guided by the well seeing, untreated eye. This produces the same amount of defocus in the treated eye for all fixation distances and may therefore be the reason for form deprivation myopia (Ni & Smith, 1989) with no regard to the sign of refractive error. Also in the cat experiments the lenses were often very powerful compared to the accommodative r nge of the animals (Ni & Smith, 1989; Smith, II arwerth, Crawford & von Noorden, 1987). Two major lines of evidence suggest that the refractive error in the chick eye is sensed and used to trigger a compensatory change in axial eye growth: first, if chicks are raised with ophthalmic lenses of different sign in front of their eyes, axial growth and refractive state are changed in the appropriate direction to compensate for the imposed defocus. The result can be obtained in normal birds with accommodation intact (Schaeffel, Glasser & Howland, 1988) and in birds unable to accommodate due to lesions in the Edinger Westphal nuclei (Schaeffel, Troilo, Wallman & Howland, 1990). Second, chicks recover from myopia or hyperopia induced by visual deprivation, mainly by an appropriate adjustment of axial growth (Troilo, 1989). Recovery occurs in the absence of accommodation and even, to some extent, after the optic nerve has been cut (Troilo, 1989). A number of questions arise from these results. (1) Apparently both the magnitude and sign of the refractive error can be sensed in local retinal areas. While the amount of defocus could be extracted from the high spatial frequency content in the retinal image, it is not clear how the sign could be determined. One possibility would be that the difference in the position of the focal planes for light of different wavelengths provides the information. Because blue light is focused closer to the principal plane than red light, red and blue color fringes around single light spots in the retina are interchanged if a defocus changes its sign. To test whether the sign of defocus is derived from chromatic cues we raised chicks with lenses in monochromatic light (589 nm). (2) In a recent mathematical model designed to explain refractive development in the chicken in response to the different treatments (Schaeffel & Howland, 1988a) it has been assumed that the

feedback loops for regulation of axial growth are linear. We tested whether this is true: in case of linearity. positive and negative lenses of the same absolute power should produce the same change in refractive state and axial growth, with only the signs different. We raised chicks with positive lenses in front of one eye and no treatment in the other and compared the results to earlier ones in which two lenses of different signs have been used for the same animal. (3) One interesting feature of refractive development in all animal models tested so far is that removal of higher spatial frequencies in the retinal image (“deprivation of form vision”) results in excessive axial growth and a high refractive errors of variable degrees (“form deprivation myopia”, FDM; monkey: e.g. Wiesel & Raviola, 1977; cat: Gollender, Thorn & Erickson, 1979; tree shrew: McKanna & Casagrande, 1978; pigeon: Bagnoli, Porciatti & Francesconi, 1985; chick: Wallman, Turkel & Trachtman, 1978; Hodos & Kuenzel, 1984). FDM also develops with the eye occluded and the optic nerve cut (Troilo & Wallman, 1987; Wildsoet & Pettigrew, 1988) or after accommodation has been abolished by lesions of the appropriate nuclei (Troilo, 1989). In addition, FDM can be induced regionally by partial visual deprivation (Wallman, Gottlieb, Rajaram & Fugate-Wentzek, 1987) which suggests that it is generated by a direct local influence of the retina on the sclera. In terms of a model of feedback regulation, form deprivation (FD) can be considered as an open-loop condition because the retinal image quality is no longer dependent on how the eye grows. Both the larger scatter in refractions and axial lengths are consistent with the idea of an open-loop system. Here, we asked the following question: if the same feedback mechanism is responsible for FDM (open-loop, with a hyperopic bias) and the changes in eye growth after lens treatment (closed-loop) then refractive errors following FD should be correctible by appropriate lenses and the corrected eyes should not return to their normal refraction. If recovery still occurs either additional mechanisms must be involved or FDM represents a condition that cannot be explained by the feedback systems running open-loop. (4) A striking feature of FDM is the large variability in refractions and axial lengths in a population (Wallman & Adams, 1987). In a recent model (Schaeffel & Howland, 1988a) the variability has been explained by different gains

Eye growth and refractive state in chicken

in the feedback loops of the different individuals while the eyes grew without visual feedback. We here tested the hypothesis that the gains are geneti~lly dete~ined and are ~ns~nts that are typical for each individual; if both eyes are symmet~cally occluded and are both growing without visual feedback, then refractive errors should be highly correlated sub~uent to a period of occlusion. METHODS

We used infrared photoretinoscopy as described in earlier papers (Schaeffel, Farkas & Howland, 1987; Schaeffel et al., 1988). The t~hnique has been further improved by computer programs that allowed one to display 20 subsequent video frame windows on the video monitor at once. The frames were grabbed over a period of about 1.5 set while the infrared LEDs at increasing eccentricities were flashed. Because every refraction occurred at least twice within the time frame of 1.5 set (for a photoretinoscope with five eccentricities), variations in accommodative tonus could be easily detected. We used two criteria to assure that we were measu~ng the resting refractive state. Firstly, only those refractions were used where no accommodative changes in refractive state had occurred during the 1.5 set and, secondly, only the most hyperopic readings were included. In the absence of an interesting fixation target the chickens quickly reiaxed their accommodation and a stable resting refractive state was observed. Single frame windows could be zoomed until they filled the whole video monitor. Intensity profiles could be placed in any desired position in the horizontal and vertical meridian and were used to verify the position and height of the photoretinoscopi~ crescents. For a precise photorefraction the critical parameter is the height of the light crescent seen in the pupil. The crescent can be well estimated if the light fractions are small but less precisely, if the pupil is almost filled with light. In Figs 5 and 8, we therefore plotted the averages of the refractions obtained from five different eccentricities along with their standard deviations to permit an evaluation of the precision with which the light fractions could be determined. If the standard deviations are large the pupil was flooded with light for the first eccentricity which did not permit one to make a dependable estimate of the light fraction. VR3114-H

~e~ure~ent

719

of cornea! radius of curvature

We used infrared photokeratometry as described earlier (Schaeffel & Howland, 1987). The technique is very precise with the standard deviations of subsequent measurements not exceeding 0.04mm (equivalent to about I D in the chick). A -scan ultrasound We used a “Echorule”, Phacosystems Inc. (3M) with a transducer probe 9D122. Because the probe was designed for the human eye, it comes focused such that its best spatial resolution was at some 2 cm from the transducer tip. We adjusted the probe to improve resolution for the chick eye by fixing a rubber tube (length 8 mm, dia. 3 mm) in front of it. By moving the reflecting surfaces in the chicken eye away from the transducer tip into the range of highest sensitivity, the precision of measurement could be si~ificantly increased, particularly for anterior chamber depth and thickness of the lens. The tube was filled with water and its tip covered with parafilm “‘M” (American Can Co.). Care was taken to avoid air bubbles being trapped in the tube because they resulted in a strong increase of measu~ment noise. Once the tube was well aligned in the axis of the transducer we always obtained very clear echoes from cornea, both surfaces of the lens and the vitreo-retinal interface. A gain of 30% of the maximal attainable was s&cient and the background noise was virtually zero. We found that the alert chicks tolerated ultrasound measurements without complaint after one or two drops of I-paracaine (proparacaine hydrochloride 0.5%, benzalkonium chloride 0.01%) were instilled on the cornea for local anesthesia. The tip of the rubber tube was moistened with water and the cornea was touched for about 1 set for measurement. After some practice’the alignment of the transducer by eye was precise enough to result in standard deviations of 0.06mm from 4-6 repeated measurements. In all the chicken eyes contributing to this study (n = 52) axial length was also measured by a video technique (Schaeffel et al., 1988). The correlation between axial length measured by ultrasound the day before dissection (x) and axial length measured in a video image (y) was very high: y = -0.64-l-l.Ix,R=O.P56(n=52 eyes). We also recorded posterior nodal distances in all eyes as described in previous papers (Schaeffel et al., 1988; Schaeffel & Howland,

720

FRANKSCH~~L

and HOWARD

1988b). Posterior nodal distance was found to be always proportional to axial length, and the same conclusions in this paper could be drawn from posterior nodal distance as from axial length. However the measurement of image magnification from trans-scleral images was difficult for some eyes from this study which were strongly defocused, and axial length is a parameter that is more frequently used. We therefore relied on the precision of the ultrasound and present only axial lengths. We carefully examined the optical quality of the eye subsequent to the ultrasound measurements to ensure that no cornea1 damage had occurred. The technique allowed one to track daily changes in axial growth (see Fig. 1C) without affecting the normal development of the optical quality of the eye. Experimental treatments Lenses were attached to small leather hoods as described earlier (Schaeffel et al., 1988), and removed for a few seconds every l-2 h for cleaning. They were worn for 12 h a day. Hoods with lenses were removed during the night where the chicks remained in the dark. In expt 1 individuals were treated with lenses of different sign in both eyes. The procedure seemed justified because it had been found earlier that accommodation and refractive development are largely independent in both eyes (Schaeffel, Howland & Farkas, 1986). We also wanted to keep the experiment comparable to the earlier

C. HOWLAND

experiment in white light. In those cases where monocular occluders or two different lenses were used in both eyes the sides were chosen randomly because there is some evidence that both eyes of the chick are not perfectly interchangable (Rogers, 1989). Occluders were handmade from a translucent plastic foil similar to the ones described by Wallman et al. (1978) and were attached to the feathers around the eye under light ether anesthesia at day 6 or 7 by instant glue. They were removed at day 14-16 by cutting the feathers that kept them in place. No inflammations or injuries were ever observed and the eyes had clear optics after removal of the occluders. Restriction of the field of view to frontal vision was achieved as described by Wallman et al. (1987). Those chicks which became near-sighted by about 8 D and far-sighted by about 8 D were chosen for subsequent correction by the appropriate negative and positive lenses, respectively. We chose refractive errors of about 8 D to assure significant defocus in the eyes on one hand but suitable correction on the other. Stronger lenses would have introduced additional optical distortions. Lenses were worn for at least a week until the course of refractive development was obvious (see Figs 4-6). A summary of all experiments including the numbers of chickens used is shown in Table 1. Animals were kept under a 12/12 hr light/dark cycle. White light was provided by fluorescent tubes and by an additional 60 W

Table 1. Summary of experiments and numbers of chickens used Experiment

1

Number of chicks

Question

Treatment

Role of chromatic a~rration?

12 Chicks raised with lenses in monochromatic hght. 6 Postive lens treatment monocularly; fellow eye untreated. IO (A) Complete binocular occlusion to attain hyperopia. 5 (B) Binocular occlusion with (in addition data from a previous frontal vision to attain experiment with 5 chicks used) myopia. 3” (A) Binocular deprivation subsequent correction of hyperopic refractive error with positive lens. 3” (8) Binocular deprivation with frontal vision, subsequent correction of myopic refractive error with negative lens. 5 (C) Monocular deprivation subsequent correction with negative lens. 2 (D) Monocular deprivation recovery with no lens correction.

Inhibition of axial eye growth by myopic defocus? Is the gain in visual feedback loops a systemic constant?

Can recovery from FDM be prevented by spectacle correction of refractive error?

“Same chicks as those used in expt 3.

721

Eye growth and refractive state in chicken

light bulb. The irradiance on the cage floor was 0.3 W/m2. Monochromatic light was provided by two large low-pressure sodium lamps. Their output was carefully examined with a spectroscope. Only very weak emission lines were found aside of the main double sodium lines at 589 nm. Also, a human observer could not see any colors on a Kodak color test table when viewing the table under sodium lamp illumination. The irradiance at the cage floor was comparable to one in white light (about 0.3 W/m*). The reason for choosing the sodium lines as light sources for this experiment were two-fold: (1) it is known from electroretinographic studies (Wortel, Rugenbrink & Nuboer, 1987, and personal observation) that the light sensitivity in the chicken is maximal between 510 and 610nm; and (2) the sodium lines provide a relatively cheap and easy way to attain high energies of narrow-band monochromatic light. Animals

As in our earlier studies, most of the chickens originated from Cornell K-strain. Twelve chicks were raised in monochromatic light with a + 4 D lens in front of one eye and a - 4 D lens in the other. Six chickens were raised with positive lenses in one eye and no lens in the other (1 chick: +4 D; 5 chicks: +8 D). Ten chicks were deprived of sharp vision by translucent occluders in both eyes over the entire visual field. Althou~ this procedure may seen questionable we only continued the experiment after we had observed that the chicks appeared not to be unhappy (e.g. they were able to orient in the cage and locate their food and water supplies). We also tracked the weight curves of all animals involved in this study and found that the double occluded chicks became as heavy as all the others. Five chicks attained binocular occlusion with a frontal notch cut into the occluder to allow for frontal vision. In Fig. 3A, data from an additional five chicks from an older experiment are included, which had been similarly treated. At the time of the previous experiment an ultrasound was not available. The ocular dimensions of these five chicks could therefore not be included in Fig. 3B. Five chicks were occluded monocularly. In addition, data from two chicks o~~nating from a black eggstrain (Bavaria, F.R.G.) and three chicks from an egg-strain from Tiibingen, F.R.G., were used. These chicks were also monocularly occluded and the resulting myopia was subsequently car-

rected by a - 8 D lens. AI1 chickens had access to water and food ad libitum. Food was distributed on paper that covered the cage floor which was changed twice a day. RESULTS

Refractive development in chicks reared with lenses in monochromatic light, and comparison to white light

Refractive development of chickens treated with positive lenses in one eye and negative lenses in the other is shown in Fig. 1. Although the time scales in the graphs (A, original experiment in white light, from Schaeffel et al., 1990, and B, monochromatic light of 589nm) are different, it can be seen that the time course of refractive development is very similar. In the original experiment in white light, eyes treated with negative lenses had a final refraction (Day 28) of - 1.7 ) 0.7 D (n = 18) as opposed to eyes treated with positive lenses which had a refraction of +2.9 2 1.2 D (n = 18, d.f. 34, t = 14.04, P < 0.001, t-test). The comparable refractions from the experiment in monochromatic light (day 14) were very similar: negative lenses: -0.90 + 1.01 D (n = 12), and positive lenses: + 2.63 + 1.20 D (n = 12, d.f. 22, t = 7.79, P < 0.001). While there was no difference between the hyperopic refractions in white and monochromatic light in eyes which had been treated with positive lenses (d.f. 28, f = 0.6, NS), the eyes with the negative lenses were slightly different in both experiments and were more myopic in white light (d.f. 28, t = 2.38, P c 0.05). However, this difference could be att~buted to the fact that in the original experiment some chicks also wore stronger lenses (-8 D). In Fig. 1C the average differences in axial lengths in both eyes is shown for the experiment in mon~hromatic light. It can be seen that almost immediately after the beginning of the lens treatment (arrows) the eyes start to grow differently, the eye with the negative lens becoming longer than the eye with the positive lens (at day 17, average difference: 0.363 & 0.177, d.f. 11, t = 7.096, P < 0.0005, one sample t-test). As has been found in earlier studies (Schaeffel & Howland, 1988b) the difference in axial length could mostly account for the difference in refractive state [predicted (y) vs measured refraction (x): y = 0.93 + 1.49x, r = 0.9141. Additional consideration of anterior chamber depth and cornea1 radius of curvature

722

FRANK~CHAEFFEL

and

HOWARD C. HOWLAND

did slightly improve the prediction as can be seen from the slope of the regression line which was closer to 1: y =0.72+ 1.10x, r =0.853. Also, the average absolute prediction error was reduced (1.77 + 2.2 D, as opposed to the prediction from axial length only: 2.45 + 2.69 D). No consistent changes in either thickness of the lens (-0.038 + 0.083 mm, d.f. 11, r = - 1.602,

NS, one sample t-test) or anterior chamber depth (0.049 + 0.11 mm, d.f. 11, t = 1.492, NS) were found in response to the lens treatment. There was a small difference in cornea1 radius of curvature between eyes treated with positive vs negative lenses (-0.058 + 0.069 mm, t= -2.775, P < 0.01, one sample t-test). The average difference in radius of curvature is White

(A)

Hyperopic

Positivelenses h=181 51 ._ 4 _ ..*-**.*..* Negative lenses ln=lW

v)

P J L

,P z

0

2

4

6

8

10

12

14

light

(0.3

r 5

T

16

1

W/m2

--4

18

20

22

24

26

28

30

Myopic

Monochrom. light (589nm,

(B)

5

0.25

W/m* r

r>

4

3

I/

[

2

4 -

3

-

2

.[. .......... I -

“i;; ;;;;;;;;;;;- ........................

0

..% . . -1

......... i

0

2

4

6

8

10

5

1

0

- -1

- -2

12

14

16

Age (days)

Fig. 1. Development of refractive state during lens wearing. (A) Chicks raised in white light with positive and negative lenses (ranges: +4 D, and -4 D and -8 D, respectively) develop quickly hyperopia and myopia. Error bars denote standard deviation from 18 chicks in which one eye was treated with a positive lens and the fellow eye with a negative lens. Arrows denote beginning of lens treatment. Replotted from Schaeffel et al. (1990). (B) The same lens experiment as performed in monochromatic light with 12 chicks (low pressure sodium lamps, 589 nm). Again the eyes treated with positive lenses develop hyperopia and the fellow eyes treated with negative lenses become weakly myopic. Note the different time scale compared to (A). (C) Ultrasound measurements of axial length in the chicks described in (B). It can be seen that immediately after application of the lenses the axial growth was different in both eyes: the eyes with the positive lenses became shorter than the fellow eyes treated with a negative lens. Error bars are calculated from the standard deviations from 4-6 ultrasound measurements. The results of the experiment in monochromatic light show little difference compared to those in white light (see text).

123

Eye growth and refractive state in chicken

negative lenses (Fig. 1A) shows that negative lenses are not very efficient in making the eyes myopic although they still have a significant effect [ - 0.43 f 1.07 D (untreated eyes at day 17) vs - 1.7 & 0.7 D (eyes wearing negative lenses at day 28), d.f. 22, t = 2.71, P < 0.02, two-tailed t-test]. By contrast, positive lenses produce strong shifts towards hyperopia which is particularly obvious if one considers the decline of hyperopia which normally occurs in the first weeks of life (Wallman & Adams, 1987; Schaeffel & Howland, 1988a). The hyperopia again is mostly axial in origin: Fig. 2B shows the average difference in axial lengths of both eyes. The eyes with the positive lens became significantly shorter [on average by 0.423 + 0.251 mm (at day 17), d.f. 11, t = 4.136, P < 0.005, one sample t-test]. There were also slight differences in cornea1 radius of curvature, the eyes wearing positive lenses having flatter corneae (+0.07 f 0.08 mm, d.f. 5, t = -2.378, P ~0.05). In addition, the anterior chamber was reduced in depth after treatment with a positive lens (-0.228 +0.18mm,d.f. 5, t =3.041, PcO.025).

equivalent to about 1.5 D in a chicken eye of the age group considered. Does myopic defocus inhibit axial growth?

In six chickens axial eye growth was compared in both eyes one of which was myopically defocused by a positive lens ( + 4 D, 1 chick, and +8 D, 5 chicks) while the other eye was untreated. The subsequent refractive development is shown in Fig. 2A. The treated eyes developed hyperopia which reached its maximum after about five days of lens wearing (+4.00 + 1.56 D, n = 6). The untreated control eye remained close to emmetropia (+0.45 f 0.61 D). The single chick with the +4 D lens was slightly less hyperopic than the average of the other five ones treated with +8 D lenses (3.2 + 0.7 D vs 4.6 f 1.68 D) but the significance of this observation cannot be established. In all chicks the difference in the refractions of both eyes was highly significant (n = 6, d.f. 22, t = 5.19, P < 0.001). On the other hand, a comparison of the refractions from untreated control eyes (Fig. 2A) to the eyes treated with (A)

White

Hyperopic -

Positive lenses [+ 4Dl11+6D(51]

*****..*.**

Untreated,

contralateral

light (0.3

W/m2

1

(n= 61

Myopic

-z

(B)

0.6

1

5 5

0.6

_$

0.4

s5

0.2

C

0

-I I

T T I I

Eye wearing

-I

,

I,

,

0

2

4

6

,

.,

0 Age

10

.,

., 12

. 14

. 16

.

. 16

(days)

Fig. 2. (A) In three chicks, the left eye was treated with a positive lens (+4 and +8 D) and the fellow eye remained untreated. In three other chicks the sides of treatment were reversed. It can be seen that the eye wearing a positive lens quickly became hyperopic whereas the untreated fellow remained functionally emmetropic. Error bars denote standard deviations from six chicks, arrows denote beginning of lens treatment. (B) From ultrasound measurements it can be seen that the eye with the positive lens remains shorter than the untreated fellow eye: myopic defocus imposed by positive lenses can inhibit axial eye growth.

FRANK SCHAEFFELand HOWARD C. HOWLAND

724

We saw no changes in thickness of the lens (-0.025+0.098mm, d.f.5, t =0.623, NS). The predictions of the interocular differences in refraction from axial length alone showed some overestimation of refractive error as can be seen from the slope being quite different from 1 [measured (x) vs predicted (y): y = 4.46 + 1.93x, n = 6, r = 0.931. The prediction became better, when cornea1 radius of curvature and anterior chamber depth where considered as well according to the schematic eye model described by Schaeffel and Howland (1988b): y = 0.45 + 1.68x, n = 6, r = 0.94. Correlation of refractions and axial growth in both eyes of individuals: are the gains systemic constants?

After symmetrical occlusion of both eyes, refractive errors acquired were highly correlated ‘47

in the left and the right eyes of 20 chickens (Fig. 3A): y = -0.602 +0.905x, r = 0.906, d.f. = 18, P < 0.001, t-test. While there was already a high correlation of axial lengths in both eyes of the different individuals before the treatment started (y = 0.884 + 0.892x, r = 0.894, n = 15) the correlation remained high while both eyes became severely ametropic during the period of deprivation of form vision (y = 0.601 +0.948x, r = 0.938; Fig. 4B). For the purpose of comparison, the average axial length of five untreated control eyes of the same experimental group is shown in Fig. 3B by the broken lines (9.45 f 0.12 mm). The result indicates that the changes in axial lengths during deprivation are correlated in both eyes. It is therefore consistent with the idea of a gain in the mechanism controlling axial growth that is a systemic constant in an animal.

(A)

-141-14

I -10

I -6

1 -2

Refraction

8.5

I 2

(left

length

(left

f 10

1 14

eye)

9.5

9.0

Axial

6

10.0

10.5

eye, mm)

Fig. 3. (A) Twenty chickens were raised with symmetrical occluson in both eyes: 10 chicks had a binocular complete occlusion for a week (producing mainly hyperopia, triangles), and 10 chicks were binocularly occluded with frontal vision allowed (resulting mainly in myopia, circles). After termination of the occlusion, refractions between the left and right eyes were compared. They were found to be highly correlated. (B) A comparison of the axial lengths before and after treatment in 15 chicks showed the following: as could be expected, before treatment axial lengths are correlated in both eyes of the different individuals. More interestingly, the deviation of axial growth from the normal condition (dashed lines) which occurs during deprivation, is also correlated in both eyes, suggesting that the gains in the mechanisms responsible for the development of ametropia, are similar. Note that some data points are hidden because they are plotted on top of each other.

Eye growth and refractive state

Spectacle correction of refractive errors acquired by deprivation

Formdeprivation myopia (FDM) is characterized by a high variability in refractive errors and axial lengths (Wallman et al., 1978; Fig. 7 in the present study). Only about 25% of our from deprived chickens had refractive errors close to 8 D. They were chosen for spectacle correction. Because refractive errors were correI4

1

(A)

125

in chicken

lated in both eyes after symmetrical deprivation, the contralateral eyes whose refractive errors remained uncorrected were available for comparison. On the other hand, because two ametropic eyes may influence each other during recovery we repeated this experiment with chicks which were monocularly deprived. (A) Correction of hyperopic eyes. Refractive development and axial growth in six eyes (three chickens, one eye corrected and the other not)

Sotid

symbds:

corrected

Open

symbols:

not

eyes I+801

corrected

T

T

12 10

Ii i

4

2

Ii

.-.,o_.

0

.

:::::::

,..,..,.:,.,.,.

.._., -0

. . . . :....~~:::::::.~.~.~;;;~....

*

f3

.

.,’

1

-2

‘1.0 1 (9) z10.5

T

-

E 5

10.0

-

P

2 6 P a

9.59.0 1

6.5-l

0 f 5 8

.r 8 C E0) 0

12

14

(C)

Chick 1

16

18

20

22

Chick 2

24

Chick 3

1

I 12

Solid

symbols:

predicted

Own

symbols

:measured

14

16

18

20

22

24

12

‘14

16

18

20

22

24

14

16

18

20

22

24

Age (days)

Fig. 4. (A) Three chicks were hyperopic by about 8 D in one eye subsequent to binocular complete occlusion. These eyes were corrected by a +8 D lens. The contralateral eyes remained uncorrected and were allowed to recover. Surprisingly, in two of the chicks the corrected eyes also recovered (0, A). In the third chicken the corrected eye remained hyperopic with the correction (m). The same symbols denote both eyes in the same individuals. Error bars denote standard deviations of the refractions from five eccentricities. (B) Development of axial lengths in the eyes from (A). The same symbols as in (A) denote the same eyes. Error bars denote standard deviations from 4-6 ultrasound measurements. The differences in axial length alone could not account for the differences in refractive state as shown in (A). (C) Predicting the intraocular differences in refractive state from measured ocular parameters. The correction with the positive lenses results in consistent flattening of the cornea (as shown in Fig. SC). Only if axial length, cornea1 refractive power, and anterior chamber depth were considered could the refraction be successfully predicted.

726

FRANK SCHAEFFEL and HOWARD C. HOWLAND

are shown in Fig. 4A and B, respectively (the same symbols denote the same chick). A striking observation is that two of the eyes corrected by a +8 D lens still recovered from hyperopia. After nine days of correction, the refractive state was only 2-3 D more hyperopic than contralateral fellow eyes. In a third chicken both eyes started out being about 8 D hyper-

epic. While the uncorrected eye recovered from hyperopia within 7 days, the eye corrected by a + 8 D lens remained hyperopic and, towards the end of the observation period, the refractive error even seemed to increase. Axial growth is shown in Fig. 4B. It can be seen, that interocular differences are larger than in normal, untreated chicks. In those, axial lengths rarely differ by

Left

eye

Hyperopia corrected with + 6D lens Right Hyper@a

eye not

corrected

11.111-11. Binocular deprivation of form vision

Binow~w deprivation of form vision on-axis; normal

2

frcntol

vision

cornea flatter

0.6,(c) After

correction

of

hyperopic eyes

myo~la:

t 2

4

6

f3

10

becomes in corrected

12

14

16

18

I

20

’

I

k

22

24

-=8

Ape (days1

Fig. 5. Effects of spectacle correction (A) Three chicks were made hyperopic by about variability

on the recovery from form deprivation ametropia-average data. by binocular occlusion. Subsequently, the eyes that were hyperopic 8D were corrected. Both uncorrected and corrected eyes showed recovery although the was high because one chick did remain hyperopic with the correction (see Fig 4A). (B) Three

chicks were made 8 D myopic by binocular occlusion with frontal vision still allowed. Subsequently one eye was corrected and the other not. Recovery was not affected by the lens correction. (C) Morpholo~cal changes in the eyes of hyperopic and myopic chicks corrected with lenses. Subsequent to the correction, hyperopic eyes became transiently shorter and myopic eyes became transiently longer than their uncorrected fellow eyes (see arrows). In addition, the cornea1 radius of curvature decreased very consistently in the eyes corrected with positive lenses (straight dashed line, note small standard deviation from the three chicks. To present differences in cornea1 radius of curvature (mm) in the same graph, we used the same ordinate). No effect of the lens correction on cornea1 radius of curvature could be seenin the negative lenses (data not shown). The result shows that, despite general recovery, the growth of the eyes had been a&&cd by the presence of the correcting lenses.

727

Eye growth and refractive state in chicken

myopic in both eyes and were subsequently corrected by -8 D lenses. The refractive development is shown in Fig. 5B. In all chicks there was recovery from myopia despite the correction. In one chick the corrected eye even recovered faster than the uncorrected eye, in another one recovery was apparently slowed down by the lens correction (individual data not shown). The different behavior is responsible for the large standard deviations shown in Fig. SB. In contrast to the corrected hyperopic chicks described above, cornea1 radius of curvature was unaffected by the lens correction (data not shown). A summary of the morphological changes in eye growth subsequent to the lens correction is given in Fig. SC. Immediately after removal of the occluders there was no consistent difference in axial lengths. However, the eyes started growing differently as soon as the lenses were in place: following correction of hyperopia the corrected eyes were transiently shorter in all chicks (average difference after 4 days of correction: 0.39 + 0.12 mm, d.f. 2,

more than 0.1 mm (Schaeffel et al., 1988). However, the differences in axial lengths shown in Fig, 4B were not sufficient to predict the large differences in refractive state as shown in Fig. 4A: we saw a consistent flattening of the corneae in all three eyes corrected by positive lenses (denoted in Fig. 5C). Here, a precise prediction of refractive state was possible only if axial length, cornea1 radius of curvature and anterior chamber depth were taken into account (Schaeffel & Howland, 1988b). As described earlier for other treatments (Schaeffel & Howland, 1988b) the crystalline lens was not affected by any manipulations of visual experience. A comparison of predicted and measured interocular differences in refractive state is shown for the three chicks in Fig. 4C. Figure 5A shows the average data from the three corrected hyperopic chicks. Due to the one chicken in which recovery was, in fact, prevented by the lens correction the average refractions have a large standard deviation.

(B) Correction ofmyopic eyes in binocularly treated chicks. Three chicks were about 8 D 4

:

No correction

..

..a..................

. . . . . . . . . . . . . . ..*.......*...

FDM

-16_1.... 0

..,,,,.,,

. .._..._.. 5 10 ..*...*...,.*...**.....*

15

-24-l 20

Lens

FDM

Age (days)

. . 0

. L .,,,

I,,

5 10 . . . . . ..~.......~...~...

,,_.. 15

,.,.. 20

/ ..,. 25

Fig. 6. Correction of monocular “form deprivation myopia” with spectacle lenses. (A) Development and recovery from myopia in two control chicks, not corrected with lenses. Solid symbols: treated eyes; open symbols: untreated control eyes. (B) and (C) Spectacle correction of FDM. Note that the eyes cross the point of best focus, thereby reversing the sign of refractive error. (D) Incomplete recovery in three chicks. These chicks remain on the myopic side of the “correction” for almost a week and Anally started to recover while the lens was still in place.

j0

FRANK SCHAEFFELand HOWARD C. HOWLAND

728

t = 5.29, P < 0.025, one sample t-test). Following correction of myopia all corrected eyes remained longer (individual data: + 0,44, + 0.12, -t 0.12 mm) although the difference did not achieve statistical significance (average difference after 3 days of correction: 0.23+0.18mm, d.f.2, t =2.21, P ~0.1, NS). The temporary change in axial growth was not sufficient to keep the eyes at the refractive state adapted for the lenses. Figure 5C also shows the flattening of the corneae in the corrected hyperopic eyes which occurred in a very consistent fashion (after 9 days of lens correction: cornea1 radius + 0.26 _+0.02 mm, d.f. 2, t = 29.10, P < 0.0005). (C) Correction of myopic eyes in monocularly treated chicks. Figure 6A shows the time course

of recovery from form deprivation myopia for two control chickens which were not corrected by lenses. It can be seen that the changes in refractive state during development and recovery from myopia are about equally fast. Recovery is entirely unaffected by a “correcting” lens in front of the eye in two cases (B and C). Here, the eye clearly “overshoots” the point of optimal correction, thereby reversing the sign of refractive error. A more complex pattern is observed in three other chicks (case D). These chicks remained more myopic than the lens correction required for about a week. They finally started to recover but recovery seemed to be independent from the presence or absence of the lens because it started without obvious reason while the lens was still in place. Unexpected occlusion

hyperopia

in binocular

complete

Binocular complete occlusion for 9 days was originally performed in 10 chicks with the intention to attain similar degrees of myopia in both eyes for subsequent lens correction. Unexpectedly, some chicks were rather hyperopic after removal of the occluders (Fig. 7). They were significantly more hyperopic than untreated controls (d.f. 23, t = 3.31, P < 0.01, t-test) and highly significantly more hyperopic than chicks with similar binocular occlusion but intact frontal vision (d-f. 28, t = 5.31, P < 0.001). Axial lengths were not significantly shorter than in normal chicks (d.f. 21, t = 1.38, NS) but much shorter than in binocularly occluded chicks with frontal vision (d.f. 26, t = 5.94, P < 0.001). As has been described by Wallman et al. (1987) binocular occlusion of all

but the frontal visual field resulted in significant amounts of myopia (P < 0.01 to P < 0.001) in the occluded eyes with significantly longer axial lengths than normal chicks. Our results from the different kinds of occlusion are summarized in Fig. 7. The striking observation is that development of “form deprivation myopia” is inhibited if no part of the retina is exposed to normal visual experience.

DISCUSSION

We have examined some properties of the mechanisms that influence eye growth and refractive state in the chicken. Longitudinal chromatic aberration not necessary for emmetropization: evidence ,for an accommo dative feedback loop?

We have found that color vision is not necessary for a regulation of axial growth so as to minimize refractive errors. We believe that the results of the experiment in monochromatic light are consistent with the idea of an accommodative feedback loop for regulation of refractive state. It can not be easily imagined how a local growth regulating mechanism in the retina (Wallman et al., 1987) could determine the sign of a defocus imposed by spectacle lenses if chromatic cues are not involved. One would expect that a more complex processing or a kind of memory, recording the visual experiences would be necessary to determine whether a given defocus is myopic or hyperopic. Because such a complex process can not be expected in the retina we conclude that some central mechanisms must be responsible for the changes in eye growth. While the result is suggestive of an involvement of accommodation in the control of axial eye growth, it is clear that a lens experiment conducted in monochromatic light with chicks that are unable to accommodate (Schaeffel et al., 1990) is necessary to finally answer this question. Linearity qf the feedback refractive state

loops controlling

We found that positive lenses which mimic myopia can inhibit axial eye growth and make the eye shorter than the contralateral untreated eye. Although we have found earlier that positive lenses produce hyperopia, we still could not show that the eyes were actually shorter than normal eyes because we always treated the

Eye growth and refractive state in chicken

contralateral eye with a negative lens. The lack of an untreated control eye in this experiment has been pointed out by Holden, Hodos, Hayes and Fit&e (1988). The previous experiments could not separate possible effects of deprivation (including axial elongation) from the effects of pure optical defocus by the lenses. The result of the current experiment now clearly demonstrates that, during treatment with positive lenses, no effects of form depriviation are superimposed because the lens treated eye is actually shorter than an untreated one. Because positive lenses seem to have a stronger effect on axial growth than negative

lenses we conclude that the feedback loops for regulation of axial growth are highly nonlinear. This is in agreement with an earlier observation (see Fig. 4 in Schaeffel & Howland, 1988a) showing that negative lenses have a much weaker effect on refractive development than positive lenses. This situation in chicks contrasts with the interpretation that, in humans, it is near work that effectively produces myopia. In chicks it seems to be the degree of myopic defocus (induced by the positive lens) that inhibils axial growth as opposed to the degree of hyperopic defocus (here induced by a negative lens) that triggers increased axial elongation and myopia.

Blnoculor

10

6

6

b’peropiC

4

2

6.6 Myopic

(SD: 0.09)

9 9.29.49.69.61010.210.4

Range (mm)

b4onocular

- 0.46

occlusion

0 -2-4-6-E-10

Range ( D1

729

occlusion

9.45

GD:O.12)

6 J4

Fig. 7. Effects of different kinds of occlusion on refractive state. In our strain, binocular complete occlusion resulted in significant amounts of hyperopia and short eyes (n = 20). Binocular occlusion, but with frontal vision still allowed, produced typical form deprivation myopia (FDM) with long eyes (n = IO). Monocular occlusion produces FDM in the occluded eye only (a = 5). The control eyes were very similar to each other in refraction and axial lengths.

730

FRANK SCHAEFFEL and HOWARD C. HOWLAND

Is the gain in the feedback loops a variable that is genetically determined for each individual?

FDM provides a very strong visual signal for the eye to grow out of the normal range. A recent interpretation of FDM (Schaeffel & Howland, 1988a) was that the feedback mechanisms regulating axial growth are set on a hyperopic bias and are running open loop with a variable gain. Because FDM has been shown to develop independently in both eyes (Yinon, Xose & Shapiro, 1980; Yinon, 1984) refractive development should be correlated in both eyes of an individual if the gain is a systemic, genetically determined constant. Figure 3 has shown that this assumption has some justification. An even more powerful experiment would be to breed those chicks that acquire similar refractive errors subsequent to a period of deprivation. If the gain is a constant that is genetically determined one should, after some generations, be able to obtain a population of chicks that develop the same amount of FDM after a defined period of deprivation. The interpretation of a variable gain also allows us to speculate why some people acquire “school myopia” (Curtin, 1985) and others not. If the gains in the feedback mechanisms for regulation of eye growth are heterogeneous in the population, one should expect that eyes in different individuals can adjust with different speed to a new working distance. On the other hand people may never get myopic in spite of near work because the gains are too low once the eyes have grown to their final sizes. Can recovery from FDM be prevented appropriate spectacle correction?

by

The results of these experiments (Figs 4-6) are very heterogeneous. In one hyperopic chick the refractive error was preserved during spectacle correction whereas the others recovered. In the myopic chicks all eyes showed recovery in spite of the correction. In both hyperopic and myopic lens-corrected chicks there was a transient change in axial growth due to the presence of the lens. Also, there was a consistent effect of the positive lens correction on cornea1 curvature, but no effect of the negative lenses. This observation, along with the finding of a flatter cornea in chicks treated with positive lenses (expt 2) might indicate that some emmetropization may also occur by a change in refractive power in the anterior segment of the eye. The assumption is in agreement with a conclusion

of Wildsoet and Pettigrew (1988). Two of the monocularly treated chicks recovered from myopia despite the correcting lens in front of their eyes (Fig. 6B, C). Here, the recovery seemed to be inde~ndent from the current refractive error in the eye. Also, in the three other chicks (Fig. 6D) there was no correlation between magnitude and time course of the recovery and the presence of the lens. The heterogeneous results from these experiments can be interpreted in the following way: (1) there exists a nonvisual mechanism that can recognize abnormal eye shapes induced by FDM that allows for recovery; (2) mechanisms sensitive to refractive error are still active but could be detected only in expts A and B because they are only superimposed on the much stronger effect of a “shape-sensitive” mechanism. Chicks have been found to recover from FDM even in monochromatic light (Gottlieb et al. and Wildsoet & Pettigrew, personal communications). Similar to the result of our lens experiment in monochromatic light it could be concluded from these recovery experiments that chromatic cues are not necessary for regulation of eye growth. However the presence of a nonvisual mechanism that can detect abnormal eye shape makes the inte~retation of these recovery experiments more difficult. ‘Form-deprivation-hyperopia”

We found that complete binocular occlusion unexpectedly produced hyperopia (Fig. 7). By contrast, visual exposure to only a small fraction of the visual field resulted in myopia as described by Wall~an et al. (1987). Here, our observations also nicely agreed with the finding of local myopia in the chick, as described by the above authors. Figure 8 shows infrared photoretinoscopy in a chick that was wearing binocular occluders for a week which permitted only frontal vision. After removal of the occluders the eyes were slighly hyperopic or emmetropic in the frontal visual field (middle trace). However, both eyes were -8 D myopic both on axis and in the rear visual field (marginal traces). In our experiment, local myopia was consistently observed in all chicks that became myopic. The finding of ‘~fo~-deprivation-hy~ropia” subsequent to binocular complete occlusion indicates some communication between both eyes (recall that chicks monocularly deprived with one total occluder became myopic). In fact, in those chicks, a powerful signal must be present

Fig. 8. Local myopia in the chicken after treatment with bmocular occluders that allowed only frontal vision. Cienem%y, for myopic eyes the pboto~n~p~ re&x is seen in the bottom of the pupil and for hyperopic eyes in the top. The refractions took place from five different eccentricities of increasing magnitude which results in decreasing light fractions in the subsequent frames. It can be seen that, after 1 week of occlusion, the chick is hyperopic or about emmetropic in the frontal visual field but quite myopic in the deprived regions of the visual field. Note also the symmetrical refraction in both eyes.

733

Eye growth and refractive state in chicken

which can antagonize the mechanism that would normally produce FDM. Currently we have no info~at~on via which pathway the signal might reach the contralateral eye. We realize that Yinon et al. (1980) obtained myopia in both eyes of chicks subsequent to a period of 3-3.5 months of binocular lid suture. There are two differences to our experiment however: firstly we occluded the eyes for only 7 days because myopia develops very rapidly and sufficient amounts of myopia were produced withm this period of time. It could be that the continuous action of the mechanism producing FDM finally overcomes the inhibitory signal presence of which we assumed above. Secondly, lid suture may involve additional nonvisual stimulation of eye growth in comparison to occluders. We also note that the difference between strains of chicks make a general systems analysis difhcult. However, it is our observation (after now having been working with five different strains of chicks) that the basic conclusions are not affected by the choice of the strain and eventual differences may only concern the quantitative aspects of a result. Also, the high statistical significance (P < O.OOf, see Fig. 7) of the occurrence of “form deprivation hyperopia” argues against the assumption that the effect was random or specific for Cornell K-strain. Conchsion The results of the present study support the previous finding that axial eye growth in the chick is controlled via feed-back regulation by mechanisms sensitive to refractive error. However, gains of these mechanisms are variable in different individuals and the effects of their action on axial eye growth may be difficult to see if more dominant mechanisms act simultaneously. “Form deprivation myopia” (FDM) develops in spite of a resulting abnormal eye shape. Our experiments show, however, that an additional mechanism that is sensitive to abnormal eye shape must be present because eyes grow back to normal shape in spite of an appropriately focused retinal image, As a simple model, there seems to be three mechanisms of increasing power: one sensitive to refractive error (closed-loop), a second mechanism sensitive to abnormal eye shape (ciosed-loop) and finally, a mechanism that is responsible for the development of FDM. The last one apparently has the strongest impact but may be identical to the first one, but with the foop of feedback regulation open.

Acknowledgeraentts-This work has been supported by NIH grant EY-02994 to HCH and, in part, by grant Zr 1/3-I from the German Research council. We thank Josh W&man for criticaf discussions, and Greg Bock for help with camputer programming, and two anonymous referees for helpful comments.

REFERENCES Bagnoli, P., Porciatti, V. & Francesconi, W. (1985). Retinal and tectaf responses to alternating gratings are unaf%cted by monocular deprivation in pigeons. Brnin Research, 338, 341-345. Crewther, S. G., Nathan, J., Kiely, P. M., Brennan, N. A. Jr Crewther, D. (1988). The effect of defocussing contact lenses on refraction in cynomolgus monkeys. Ciinical Vision Sciences, 3, 221-228. Curtin, Il. J. (1985). The myopias: Basic science and clinical management. Philadelphia: Harper & Row. Gollender, M., Thorn, F. & Erickson, P. (1979). Development of axial dimensions following eye lid suture in the cat. Vision Resetarch, 19, 221-223. Hodos, W. & Kuenzel, W. J. (1984). Retinal image degradation produces ocular enlargement in chicks. Inuestigutive Ophthalmology and Visual Science, 25, 652-659.

Holden, A. L., Hodos, W., Hayes, B. P. & Fitzke, F. W. (1988). Myopia: Induced, normal, and clinical. Eye, 2 &q@.), S242-S256. McKanns, J. A. Br Casagrande, V. A. (1978). Reduced lens development in lid-suture myopia. E~~r~e~~u~ Eye Research, 26, 7 I S-723.

Ni, J. Br Smith, E. L. III (1989) Effect of chronic optical defocus on the kitten’s refractive status. Vision Research, 29, 929-939.

Rogers, L. J. (1989). Left and right in the chick visual system: Asymmetrical thalamic projections to the hyperstriata in the young chicken. In E&r, J., Menzel, R., PiIiiger, H. J. & Todt, D. @is.), Neural mechunisms of behavior @. 197). Stuttgart: Thieme. Schaeffel, F. & Hawland, H. C. (1987)+ Cornea1 accommodation in chick and pigeon. Journal of Comparative Physiology A, 160, 375-384.

Schaeffel, F. & Howland, H. C. (1988a). A mathematical model for emmetropization in the chicken. Journal of the Optical Society of America A, 5, 2080-2086.

Schaeffel, F. & Howland, H. C. (1988b). Visual optics in normal and ametropic chickens. Clinica! Vision Scierrces. 3, 83-93. Schaeffel, F,, Howland, H. C. & Farkas, L. (1986). Natural accommodation in the growing chicken, Vision Research, 26, 1977-1993. Schaeffel, F., Farkas, L. & Howland, H. C. (1987). Infrared photoretinoscope. Applied Optics, 26, 1505-I 509. Schaeffel, F., Glasser, A. & Howland, H. C. (1988). Accommodation, refractive error, and eye growth in chickens. Vision Research, 28, 539457. SchaeRel, F., Troilo, D., Wallman, J. & Howland, H. C. (1990). Growing eyes that lack a~ommodation grow to compensate for imposed defocus. Visual Neuroscience, #+ 177-183.

Smith, E. L. III., Harwerth, R. S., Crawford, M. L. J. & von Noorden, G. K. (1987) Observation on the effects of form deprivation on the refractive status of the monkey. investigatiue 12361245.

Uphthalmology

and

Usual

Science,

28,

734

FRANKSCHAE~FEL and HOWARDC. HOWLAND

Troilo, 0. (1989). The visual control of eye growth in chicks. Ph.D. Dissertation, City College of City University of New York. Troilo, D. & Wailman, J. (198’7).Visual deprivation causes myopia in chicks with optic nerve section. Current Eye Research,

6, 993-999.

Wallman, J. & Adams, J. (1987). Developmental aspects of experimental myopia in chicks: Susceptibility, recovery and relation to emmetropization. Y&ion Research, 27, 1139-l 163. Wallman, J., Turkel, J. & Trachtman, J. (1978). Extreme myopia produced by modest changes in early visual experience. Science, 201, 1249-125 I. Wallman, J., Gottlieb, M. D., Rajaram, V. & FugateWentzek, L. (1987). Local retinal regions control local eye growth and myopia. Science, 237, 73-77. Wiesel. T. N. & Raviola, E. (1977). Myopia and eye

enlargement after neonatal lid fusion in monkeys. Nurure, 266, 6668. Wildsoet, C. F. & Pettigrew, J. D. (1988). ~x~rirnen~l myopia and anomalous eye growth patterns unaffected by optic nerve section in chickens: Evidence for local control of eye growth. Clinical Vision Sciences, 3, 99-107. Wortel, J. F., Rugenbrink, H. & Nuboer, J. F. W. (1987). The photopic spectral sensitivity of the dorsal and ventral retinae of the chicken. ~o~r~u~ of CompurQriL~e Physiorogy A, 160, 151-154. Yinon, U. (1984). Myopia induction in animals following alteration of the visual input during development: A review. Current Eye Research, 3, 677-690. Yinon, U., Rose, L. & Shapiro, A. (1980). Myopia in the eyes of developing chicks following monocular and binocular lid closure. Vision Research, 20, 137-141.