Resilience in Cognitive Development

Jerome Kagan

Each scientific discipline, during successive eras in its growth, is loyal to one member of a set of opposed assumptions that typically form the axioms of the discipline. Holton (1973) has called these polarized premises themata . Debate over whether matter is particulate or wavelike, whether the universe is steady-state or expanding, or whether growth is continuous or discontinuous are among the themes that scientists have debated in the past and will continue to discuss in the future because, as Bohr wisely noted, the propositions are likely to be complementary rather than incompatible.

This paper considers the evidence bearing on a pair of opposed themata that have given direction to the empirical study of human psychological development. One proposition holds that the experiences of infancy produce a set of dispositions that have a continuous influence throughout life, implying that some of the effects of early experience are not malleable to change. The opposed position is

that the infant is resilient and the effects of early experience—which can be dramatic—are reversible under proper environmental conditions. The debate centers on the degree of modifiability of psychological structures established early in life. There is unequivocal support for the view that the experiences of the young infant have a powerful contemporary effect on his behavior, temperament, and knowledge. This hypothesis is unchallenged. But it is less clear how stable these early structures are, especially if the environment should change in a serious way. Stated in the interrogative, how resilient—or responsive to change— are the cognitive structures and behavioral dispositions shaped during the first three years of life.

Although the infant is influenced by his environment from the moment he is born, the equally popular assumption that the effects of those early experiences can extend long into the future does not, at the moment, have unequivocal support. Hence we should ask why many psychologists, parents, and educators have been reluctant to examine the validity of that second statement. In addition to the obvious influence of psychoanalytic theory and the persuasive animal data, there are additional reasons, both phenomenological and philosophical. Each person feels a compelling sense of continuity and connectedness when he reflects on the experiences of his early childhood. This sense of the past's contribution to the present derives from man's need to regard his life as coherent and his past decisions as part of a rationally causal chain. A second, more speculative, basis for believing in the extended power of early experience could be a derivative of one of the central maxims of Western Protestantism—preparation for the future. Application of that maxim to child rearing would lead parents to award validity to the idea that if children are treated optimally during the early years, the healthy attitudes, talents, and behaviors established during that first era should provide protection against possible traumas during adolescence and adulthood. Proper early familial treatment, like early vaccination, might inoculate the child against vulnerability to future distress. Finally, faith in the permanent influence of early experience is in accord with the commitment to political egalitarianism that is so strong in Western Protestant democracies. If society treats children properly during the opening years there is at least the hope that the distress, incompetence, and hopeless-

ness that prevent full political participation by all adult citizens could be eliminated and a truly egalitarian society established.

I recently interviewed a 14 1/2-year-old girl who spent most of the first 30 months of her life in a crib in a small bedroom with no toys and a sister one year older than herself. The mother, who felt unable to care for her fourth child, restricted her to the bedroom and instructed her 8-year-old daughter to care for the child. When she was removed to a foster home at 2 1/2 years of age, she was severely malnourished, retarded in weight and height, and so retarded psychologically that she was untestable. She has remained with the same foster family for the last 12 years. At present her full scale IQ is 88; she performs normatively on a wide battery of cognitive tests and her interpersonal behavior is not seriously different from that of an average rural Ohio adolescent.

Koluchová (1972) has recently reported a similar developmental history for twin Czechoslovakian boys who were placed in total isolation by their stepmother and father from 18 months to 7 years. Most of the time the boys were in a small unheated closet, but they were often locked up for long periods in the cellar of the house. The boys were never allowed outside the house and were inhumanly treated. When the children were removed from the home at age 7 they were physically ill, psychologically untestable, and displayed extreme surprise and fear to common events like automobiles and toys. This extreme behavior, which resembled that of Harlow's surrogate-reared monkeys, gradually abated in the hospital environment. The boys were sent to a children's home for six months and then to a foster home. When the boys were tested at 11 years of age—only a few years after their emergence from the isolated environment—their full scale Wechsler IQ scores were 95 and 93, and the physician noted that they appeared above average for their age.

The Guatemalan Study

I recently observed infants and children living in an isolated, subsistence farming village called San Marcos in the highlands of northwest Guatemala. As a result of parental treatment, frequent illness, lack of experiential variety, and mild malnutrition the one-year-olds were quiet, nonsmiling, minimally alert, motorically flaccid and temperamentally passive. This profile of characteristics is

in sharp contrast to the modal profile of middle class American infants who are highly vocal, smiling, alert, and active. Experimental and observational procedures designed to assess level of cognitive development among these Indian infants revealed that, relative to the Americans, they were 3 to 12 months behind the latter depending on the cognitive system studied.

The Guatemalan infants were markedly less attentive than the Americans to visual and auditory events, and this difference was greater at one year that it was at 5 months (Kagan and Klein 1973). The Guatemalan infants were retarded relative to the Americans in their tendency to reach for an attractive object that they watched being hidden, and not one of a group of 12 infants revealed facial surprise following a sequence in which he watched an object being hidden under a cloth but saw no object when that cloth was removed. These observations suggest a retardation of about 4 months in the display of behavioral signs diagnostic of what Piaget has called object permanence.

A third source of data came from observations of stranger anxiety. Each of sixteen infants between 8 and 20 months was observed following the first exposure to a strange male. The first age at which obvious apprehension or crying occurred was 13 months, suggesting an approximate lag of 5 months between the Guatemalan and American infants. Information on nonmorphemic babbling and the onset of meaningful speech also supported a diagnosis of slower growth, for there was no marked increase in frequency of babbling or vocalization between 8 and 16 months among twelve infants observed at home. Comparable observations in American homes reveal a significant increase in babbling and the appearance of morphemic vocalization for some children. Furthermore, meaningful speech usually appears first at 2 1/2 years, about 12 to 18 months later than the time of average display of initial words in American children. These data, together with the extremely depressed and withdrawn appearance of the Guateman infants, suggest that for a small set of universal competences displayed by all children during the first two years of life, the Indians were significantly late in attaining these abilities. Since more than 90 percent of the infants were homogeneously passive, nonalert, and quiet, it is unlikely that the recovery of intellectual functioning to be reported was the result of the selective mortality of a small group of severely retarded infants.

The home environments. It is believed that the restricted experiences of the Guatemalan infants were responsible for their slower rate of growth. During most of the first year, the infant is tightly clothed and restricted to the inside of a windowless hut about 75 feet square, constructed of bamboo walls and a roof of thatched grass. The dirt floor contains an open fire, some wood, a straw mat, and a few clay receptacles. Ears of corn, cups, and pots hang from the walls and roof. The light level inside the hut at noon is low and approximates the level outside at dusk. The infant spends approximately a third of his time on his mother in a sling, a third sitting or lying on a straw mat, and the final third sleeping in his hammock. The infant has no conventional toys with which to play, and adults are minimally interactive with him. Time-sampled observations of infants in the home revealed that play or vocalization directed at the baby by others (parents, relatives, or older children) occurred less than 10 percent of the time, in contrast to 25 to 40 percent of the time in American homes. As a result, the babies were generally very quiet.

By 13 to 16 months, however, when the baby becomes mobile and is allowed to leave the hut, he encounters the greater variety inherent in the outside world. He engages an environment that includes domestic animals, other children, trees, rain, clouds, and makes the accomodations those experiences require. The 8- to 10-year-old is assigned tasks and responsibilities, such as helping the father in the field, caring for infants, cooking, cleaning, and carrying water. During the postinfancy years the child becomes increasingly alert and active and it is relevant, therefore, to ask if these older Guatemalan children, who were slow in attaining the universal competences of infancy, are different from less isolated Guatemalan or urban Western children with respect to some of the universal cognitive competences of preadolescence.

The Competence of Older Children

Tests designed to assess cognitive processes believed to be part of the natural competence of growing children were administered to samples of Guatemalan and American children. We tried to create tests that were culturally fair, recognizing that this goal is, in the extreme, unattainable. We assumed, along with many psychologists, that perceptual analysis, recall and recognition memory, and inference are among the universal cognitive abilities of children

(even though they do not exhaust that set), and our tests were designed to evaluate those processes.

Recall memory for familiar objects. The ability to organize experience for commitment to long-term memory and to retrieve that information on demand is a basic cognitive skill. It is generally believed that the form of the organization contains diagnostic information regarding cognitive maturity for, among Western samples, both number of independent units of information as well as the conceptual clustering of that information increase with age.

A twelve-object recall task was administered to the Indian children of San Marcos and to children from a Ladino village seventeen kilometers from Guatemala City whose infant experience was not as restricted. The eighty subjects from the Ladino village were 5 and 7 years old, equally balanced for age and sex. The fifty-five subjects from San Marcos were between 5 and 12 years of age (26 boys and 29 girls).

The twelve miniature objects to be recalled were common to village life and belonged to three conceptual categories: animals (pig, dog, horse, cow); kitchen utensils (knife, spoon, fork, glass); and clothing (pants, dress, underpants, hat). Each child was first required to name the objects and if the child was unable to he was given the name. The child was told that after the objects had been randomly arranged on a board he would have 10 seconds to inspect them, after which they would be covered with a cloth, and he would be required to say all the objects he could remember.

Table 1 contains the average number of objects recalled and the number of pairs of conceptually similar words recalled—an index of clustering—for the first two trials. The maximum clustering score for a single trial was 9 points. All the children showed a level of clustering beyond chance expectation (which is between 1.5 and 2.0 pairs for recall scores of 7 to 8 words), recall scores increased with age for children in both village (F ranged from 11.2 to 27.7, p < .05), and there was no significant difference in performance between the two samples. Indeed, the San Marcos children performed slightly better than the Ladino youngsters.

No 5- or 6-year-old in either village and only twelve of the forty 7-year-olds in the Ladino village were attending school. School for the others consisted of little more than semiorganized games. Moreover, none of the Indian children from San Marcos had ever left

the village, and the 5- and 6-year-olds typically spent most of the day within a 500-yard radius of their homes. Hence, school attendance and contact with books and a written language do not seem to be prerequisites for clustering in young children.

The recall and cluster scores obtained in Guatemala were remarkably comparable to those reported for middle class American children. Appel et al. (1971) presented twelve pictures to Minneapolis children in Grade 1 (approximately age 7), and fifteen pictures to children in Grade 5 (approximately age 11) in a single trial recall task similar to the one described here. The recall scores were 66 percent for the 7-year-olds and 80 percent for the 11-year-olds. These values are almost identical to those obtained in both Guatemalan villages. The cluster indexes were also comparable. The American 7-year-olds had a cluster ratio of 0.25; the Indian 5- and 6-year-olds had a ratio of 0.39.^[1]

Recognition memory. The robust performance on recall was also found on a recognition memory task for thirty-two photos of faces, balanced for sex, child versus adult, and Indian versus Caucasian, administered to thirty-five American and thirty-eight San Marcos children 8 to 11 years of age. Each child initially inspected thirty-two chromatic photographs of faces, one at a time, in a self-paced procedure. Each child's recognition memory was tested by showing him thirty-two pairs of photographs (each pair was of the same sex,

[1] The cluster index is the ratio of the number of pairs recalled to the product of the number of categories in the list times one less than the number of words in each category.

age, and ethnicity), one of which was old and the other new. The child had to state which photograph he had seen during the inspection phase. Although the American 8- and 9-year-olds performed slightly better than the Guatemalans (82 versus 70 percent) there was no significant cultural difference among the 10- and 11-year-olds (91 versus 87 percent). Moreover, there was no cultural difference at any age for the highest performance attained by a single child. These data are in accord with those of Kagan et al. (1973) and Scott (1973).

The favored interpretation of the poorer performance of the younger children in both recognition memory studies is that some of them did not completely understand the task, and others did not activate the proper problem-solving strategies during the registration and retrieval phases of the task.

It appears that recall and recognition memory are basic cognitive functions that seem to mature in a regular way in a natural environment. The cognitive retardation observed during the first year does not have any serious predictive validity for these two important aspects of cognitive functioning for children 10 to 11 years old.

Perceptual analysis. The Guatemalan children were also capable of solving difficult Embedded Figures Test items. The test consisted of twelve color drawings of familiar objects in which a triangle had been embedded as part of the object. The child had to locate the hidden triangle. The test was administered to the rural Indian children of San Marcos, as well as to less isolated Indians living close to Guatemala City (labeled Indian₁ in figure 1), Ladino villages, and two groups from Guatemala City.

The Guatemala City middle class children had the highest scores and, except for San Marcos, the rural children the poorest. The surprisingly competent performance of the San Marcos children results, we believe, from the more friendly conditions of testing. This suggestion is affirmed by an independent study in which a special attempt was made to maximize rapport and comprehension of instructions with a group of rural isolated children before administering a large battery of tests. Although all test performances were not facilitated by this rapport-raising procedure, performance on the Embedded Figures Test was improved considerably. It is important to note that no 5- or 6-year-old was completely incapable of solving some of these problems. The village differences in mean

Fig. 1.
Mean number correct on the Embedded Figures Test

score reflect the difficulty that the rural children had with three or four of the harder items. This was the first time that many rural children had even seen a two-dimensional drawing and most of the 5-, 6-, and 7-year-olds in San Marcos had no opportunity to play with books, paper, pictures, or crayons. Nonetheless, these children solved seven or eight of the test items. As with recall and recognition memory, the performance of the San Marcos child was comparable to that of his age peer in a modern urban setting.

Perceptual inference. The competence of the San Marcos chil-

dren on the Embedded Figures Test is affirmed by their performance on a test administered only in San Marcos and Cambridge and called Perceptual Inference. The children (60 American and 55 Guatemalan, 5 to 12 years old) were shown a schematic drawing of an object and asked to guess what that object might be if the drawing were completed. The child was given a total of four clues for each of thirteen items, where each clue added more information. The child had to guess an object from an incomplete illustration, to make an inference from minimal information (see figures 2 and 3).

Fig. 2.
Sample item from the Perceptual Inference Test

There was no significant cultural difference for the children 7 to 12 years old, although the American 5- and 6-year-olds did perform significantly better than the Indian children. In San Marcos, performance improved from 62 percent correct on one of the first two clues for the 5- and 6-year-olds to 77 percent correct for the 9- to 12-year-olds. The comparable changes for the American children were from 77 to 84 percent (see figure 4).

Fig. 3.
Sample item from the Perceptual Inference Test

Familiarity with the test objects was critical for success. All the San Marcos children had seen hats, fish, and corn and these items were rarely missed. By contrast, many American children failed these items. No San Marcos child not attending school, and therefore unfamiliar with books, correctly guessed the book item; whereas most of those in school guessed it correctly. As with memory and perceptual analysis, the retardation seen during infancy did not predict comparable retardation in the ability of the 11-year-old to make difficult perceptual inferences.

Conceptual inference. The San Marcos child also performed well on questions requiring conceptual inference. In this test, the child was told verbally three characteristics of an object and required to guess the object. Some of the examples included: what has wings, eats chickens, and lives in a tree; what moves trees, cannot be seen, and makes one cold; what is made of wood, is used to carry things, and allows one to make journeys. There was improved performance with age; the 5- and 6-year-olds obtained an average of 9 out of 14 correct, and the 11- and 12-year-olds obtained 12 out of 14 correct. The San Marcos child was capable of making inferences from both visual and verbal information.

This corpus of data implies that slower attainment of selected

Fig. 4.
Number correct on the Perceptual Test

universal competences of infancy does not have any important implications for the competences of the preadolescents with respect to perceptual analysis, perceptual inference, recall, and recognition memory. Indeed, the Guatemalan 10-year-olds performed at a level comparable to those of American middle class children.

The suggestion that basic cognitive competences emerge at dif-

ferent times and that the child retains the capacity for actualization of the basic human competences until a late age is not substantially different from the earlier conclusions of Dennis and Najarian (1957). Although the forty-nine infants 2 to 12 months of age living in poorly staffed Lebanese institutions were seriously retarded on the Cattell developmental scale (mean developmental quotient of 68 compared with a quotient of 102 for a comparison group), the 4 1/2- to 6-year-olds who had resided in the same institution all their lives performed at a level comparable to American norms on a memory test (Knox Cubes) as well as on Porteus mazes and the Goodenough Draw-a-Man-Test.

Of more direct relevance is Dennis's (1973) recent follow-up study of sixteen children who were adopted out of the same Lebanese institution between 12 and 24 months of age—the period durin which the San Marcos infant leaves the unstimulating environment of the dark hut—with an average developmental quotient of 50 on the Cattell Infant Scale. Even though the assessment of later intellectual ability was based on the culturally biased Stanford-Binet IQ test, the average IQ, obtained when the children were between 4 and 12 years old, was 101, and thirteen of the sixteen children had IQ scores of 90 or above (Dennis 1973).

This finding is in substantial agreement with a recent follow-up study of sixty-five 4 1/2-year-old children who had spent their first 2 to 4 years in an institution in which an exclusive relation between an infant and one caretaker was actively discouraged. Of the original group of sixty-five, fifteen were now living with their natural mothers, twenty-four had been adopted, and twenty-six were still living in the institution. There was no difference among the three groups at 4 1/2 years of age on Wechsler IQ scores (the IQ means ranged from 100 to 115). Although the institutionalized children had been retarded in language development when they were 2 years old, they were not retarded with respect to British norms at 4 1/2 years. "No evidence of cognitive retardation, verbal or otherwise, was found in a group of 4-year-old children institutionalized since early infancy. . . . As far as reversibility of the ill effects of institutionalization is concerned, cognitive retardation was reversed even within the institution between ages 2 and 4" (Tizard and Rees 1974:97, 98; see also Rheingold and Bayley 1958).

Animal Investigations

More dramatic support for the notion that psychological development is resilient comes from recent experimental studies with animals. Several years ago Harlow's group demonstrated that although monkeys reared in isolation for the first six months displayed abnormal and often bizarre social behaviors they could, if the experimenter were patient, solve the complex learning problems normally administered to feral-born monkeys. The prolonged isolation did not destroy their cognitive competence (Harlow, Schiltz, and Harlow 1969). More recently, Suomi and Harlow (1972) have shown that even the stereotyped and bizarre social behavior shown by 6-month isolates can be altered by placing them with female monkeys three months younger than themselves over a 26-week therapeutic period. "By the end of the therapy period the behavioral levels were virtually indistinguishable from those of the socially competent therapist monkeys" (Suomi and Harlow 1972:491; see Gomber and Mitchell 1974 for a similar result).

Even imprinting toward a nonnatural object in a laboratory context seems to be reversible. In a laboratory context Hess attempted to imprint ducklings to human beings. For twenty continuous hours newly hatched ducklings were exposed to adults and, before long, followed the adults. The ducks were then given a female mallard that had hatched a clutch of several ducklings several hours before. After only an hour and a half of exposure to the female, the human imprinted ducklings followed the female on her first exodus from the nest. The laboratory imprinting had been reversed (Hess 1972). This phenomenon is analogous to changes in the object of primary attachment among primates. Rhesus monkeys were raised from birth with cloth surrogates, their mothers, or a peer monkey for three to ten months. Then all the monkeys were separated from these objects of primary attachment and gradually exposed to spayed, adult female dogs. Initially most of the monkeys were fearful, but this behavior disappeared quickly and after seven hours all monkeys approached the dogs and eventually clung to them. Soon the monkeys displayed the classic signs of attachment—clinging and following. The initial attachment had been changed (Mason and Kenney 1974).

These dramatic alterations in molar behavior are in accord with replicated reports of recovery of visual function in monkeys and

cats deprived of patterned light soon after birth (Wilson and Riesen 1966, Baxter 1966, Chow and Stewart 1972). Kittens deprived of light for one year recovered basic visual functions after only ten days in the experimenter's home (Baxter 1966); kittens who had one or both eyes sutured for nearly two years were able to learn pattern discriminations with the deprived eye only after moderate training (Chow and Stewart 1972). Even complex cognitive functions can recover following removal of frontal cortex in young monkeys. In a recent dramatic study, rhesus monkeys were given bilateral, orbital prefrontal lesions during the first, fourth, or eighth week of life and compared with age-matched, unoperated controls at 1 to 1 1/2 years and gain at 2 years. The monkeys were tested on spatial delayed response, visual pattern discrimination, spatial delayed alternation, and object discrimination reversal. The task most sensitive to the lesion, spatial delayed alternation, requires the animal to alternate his responses to the right and left food wells on succeeding trials with a 5-second delay between trials. The operated monkeys were seriously impaired when tested at 1 and 1 1/2 years of age, regardless of when the operation had been performed (1, 4, or 8 weeks). But when they were 2 years old, they had recovered that competence and were not significantly different from the unoperated controls. The investigators suggested that this result was because of maturation of other cortical regions during the second year of life (Miller, Goldman, and Rosvold 1973).

If the extreme behavioral and perceptual sequellae of isolation and brain ablation in monkeys and cats can be altered by such brief periods of rehabilitative experience or time for recovery, it is not difficult to believe that the rural Guatemalan infant is capable of as dramatic a recovery over a period of nine years. These data do not indicate the impotence of early environments, but rather the potency of the environment in which the organism is functioning. There is no question that early experience seriously affects kittens, monkeys, and children. If the first environment does not permit the full actualization of psychological competences, the organism will function below his ability as long as he remains in that context. But if he is transferred to an environment that presents greater variety and requires more accomodations, he seems more capable of exploiting that experience and repairing the damage wrought by the first environment than some theorists have implied.

Conclusions

The total corpus of information implies that the young animal retains an enormous capacity for change in early patterns of behavior and cognitive competence, especially if the initial environment is seriously altered. The data offer no firm support for the popular belief that certain events during the first year can produce irreversible consequences in either human or infrahuman infants. If one limits the discussion to universal cognitive competences, in contrast to culturally specific skills, it appears that a slower rate of mastery of the universal abilities during the first two years places no serious constraints on the eventual attainment of many of the competences of preadolescence. For most of this century developmental psychology has been friendly toward the pole of the irreversibility-reversibility theme that posited irreversible effects of early experience. The extreme form of that position is as unlikely as the opposite pole that assumes complete capacity for resilience of all dispositions at any age. The purpose of this paper has been to persuade the receptive reader to move just a little closer toward the latter view. The first messages written on the tabula rasa may not necessarily be the most difficult to erase.

References

Appel, L. F., R. G. Cooper, N. McCarrell, J. S. Knight, S. R. Yussen, and J. H. Flavell. 1971. The Developmental Acquisition of the Distinction between Perceiving and Memory. Unpublished Manuscript.

Baxter, B. L. 1966. Effect of Visual Deprivation during Postnatal Maturation on the Electroencephalogram of the Cat. Experimental Neurology 14:224-237.

Chow, K. L., and D. L. Stewart. 1972. Reversal of Structural and Functional Effects of Longterm Visual Deprivation in Cats. Experimental Neurology 34:409-433.

Dennis, W. Children of the Creche. Century-Crofts. 1973.

Dennis, W., and P. Najarian. 1957. Infant Development under Environmental Handicap. Psychological Monographs 71, no. 436.

Gomber, J., and G. Mitchell. 1974. Preliminary Report on Adult Male Isolation-reared Rhesus Monkeys Caged with Infants. Developmental Psychology 10:298.

Harlow, H. F., K. A. Schiltz, M. K. Harlow. 1969. The Effects of Social Isolation on the Learning Performance of Rhesus Monkeys, Proceedings of the Second International Congress of Primatology (C. R. Carpenter, ed.), Vol. I. Karger.

Hess, E. H. 1972. Imprinting in a Natural Laboratory. Scientific American 227:24-31.

Holton, G. 1973. Thematic Origins of Scientific Thought. Harvard University Press.

Kagan, J., R. E. Klein, M. M. Haith, and F. J. Morrison. 1973. Memory and Meaning in Two Cultures. Child Development 44:221-223.

Kagan, J., and R. E. Klein. 1973. Cross-Cultural Perspectives on Early Development. American Psychologist 28:947-961.

Koluchová, J. 1972. Severe Deprivation in Twins: A Case Study. Journal of Child Psychology and Psychiatry 13:107-111.

Mason, W. A., and M. D. Kenney. 1974. Redirection of Filial Attachments in Rhesus Monkeys: Dogs as Mother Surrogates. Science 183:1209-1211.

Miller, E. A., P. S. Goldman, and H. E. Rosvold. 1973. Delayed Recovery of Function Following Orbital Prefrontal Lesions in Infant Monkeys. Science 182:304-306.

Rheingold, H. L. and N. Bayley. 1959. The Later Effects of an Experimental Modification of Mothering. Child Development 30:363-372.

Scott, M. S. 1973. The Absence of Interference Effects in Preschool Children's Picture Recognition. Journal of Genetic Psychology 122:121-126.

Suomi, S. J., and H. F. Harlow. 1972. Social Rehabilitation of Isolatereared Monkeys. Developmental Psychology 6:487-496.

Tizard, B., and J. Rees. 1974. A Comparison of the Effects of Adoption, Restoration to the Natural Mother, and Continued Institutionalization on the Cognitive Development of 4-year-old Children. Child Development 45:92-99.

Wilson, P. D., and A. H. Riesen. 1966. Visual Developments in Rhesus Monkeys Neonatally Deprived of Patterned Light. Journal of Comparative and Physiological Psychology 61:87-95.