33

Patterns of Behavior

Richard E. Wimer and John L. Fuller

This chapter is divided into three major sections. The first surveys a wide range of studies on mouse behavior. It is also concerned with apparatus and testing procedure and with behavioral differences which have been established between genetic stocks. The second section deals with studies of environmental factors and their contribution to behavioral phenotypes. The final section deals with techniques for the genetic analysis of behavior.

SURVEY

Maintenance behavior

Eating and hoarding. That eating is cyclic is well known to all mouse breeders, for the sound of gnawing on pellets becomes audible in the mouse quarters late in the day. More precise records of daily variations in feeding activity can be obtained by the cumulative recorders used in operant conditioning studies ( Anliker and Mayer, 1956). Normal mice show a strong 24-hour periodicity, the maximum rate occurring during the night. Genetically obese ( ob/ ob) mice and mice made obese by hypothalamic lesion or gold thioglucose injection eat at a uniform rate with only short irregular respites.

Hoarding of food is not seen in the laboratory unless conditions are arranged to favor it. Smith and Ross ( 1953a, 1953b) found that C3H mice would carry food pellets to their cage from a supply connected temporarily by an alley. The effects of food deprivation were complex, and satiated animals sometimes hoarded more than deprived ones. The relevance of hoarding to eating behavior was questioned, since they found that wet and dry cotton packs were retrieved more readily than food pellets. A comparison of three strains on the hoarding test ranked BALB/c, C3H, and C57BL/10 in order from highest to lowest ( Smith and Powell, 1955).

Drinking. Drinking, like eating, is cyclic and occurs mostly during the night. Mice given water ad libitum typically consume 4 to 6 ml each 24 hours. A number of strains manifest polydipsia. Parous MA/J and Ma/MyJ females usually consume 10 to 50 ml per day; virgin females and males show a less extreme increase in water intake ( Hummel, 1960). The polydipsia is probably caused by a reduction in antidiuretic hormone brought on by cystic degeneration in the posterior pituitary. An apparently similar condition, not inherited in a simple Mendelian fashion, has been reported in STR/N mice ( Silverstein, 1961).

The intake of water by mice is considerably above minimum requirements. The effects of restricting time of access to water were studied by Hudson (1964, personal communication). Access was permitted for either one or two periods per day, with a total duration of 2, 4, 6, 8, or 1 minutes. The results, shown in Figure 33-1, indicate a rather rapid loss of weight ( 10 to 17 per cent of body weight) which leveled off at about 5 days. Partial recovery of the loss occurred in all except the 1 x 2-minute and 2 x 1-minute groups. Weight recovery in general paralleled ingestion curves. More water was drunk when access times were scheduled 12 hours rather than 24 hours apart. It is apparent that water was being physiologically conserved, since weight increased while intake was far below the normal level with 24-hour access. The results suggest that 3 days of adaptation to scheduled drinking should precede experiments using water deprivation as a means of controlling drive.

Preference. A well-stocked grocery store is an excellent cafeteria for the determination of food preferences in mice. Chitty and Southern ( 1954) reported that the first supplies to be attacked were the cereal foods such as rolled oats, oatmeal, whole and ground rice, macaroni, and vermicelli. In a less preferred class were substances containing fats and proteins such as meats, butter, wax. etc. Finally in the lowest class were sugar, chocolate, and a variety of dried fruits and preserves. Mice were said to avoid legumes even when other food was scarce. The earliest systematic experiment on food preferences was reported in the same source. Southern and his associates found rolled oats mixed with 20 per cent olive oil to be particularly attractive to mice.

Hoshishima et al. ( 1962) studied differences in preference for NaCl, saccharin, acetic acid, and phenylthiocarbamide (PTC) in strains NA-II, aa, SM, 0-61, C57, and C3H, Mice had access to distilled water and a test solution. Concentration of test solution was increased until threshold, defined as the lowest concentration of test solution necessary to produce significant differences between average 24-hour intake of the test solution and distilled water, was reached. Inspection of their data suggests comparatively little strain variation in threshold for NaCl and acetic acid, but large variation for saccharin and PTC. Other preference studies have employed sodium sucaryl ( Smith and Ross, 1960), sucrose octaacetate ( Warren, 1963), propylthiouracil ( Jacobs, 1962, and sucrose ( Rodgers and McClearn, 1962 a).

The largest body of experimental research on preference concerns alcohol ( Rodgers and McClearn, 1962b). McClearn and Rodgers ( 1959) measured intake of 10 per cent alcohol and water using strains C57BL/Crgl, A/2Crgl, DBA/2NCrgl, BALB/cCrgl, C3H/2Crgl, and AKR/Crgl. Strain C57BL/Crgl consumed significantly greater proportions of alcohol than the other strains, which did not differ. In a later study, Rodgers and McClearn ( 1962a) reported preference among water and six concentration of alcohol. Strain C57BL showed highest preference for a 12.5 per cent solution, and C3H preferred 10 per cent. Strains BALB/c and A avoided all concentrations of alcohol. Mirone ( 1958) investigated the effects of parental or earlier direct contact with alcohol on voluntary consumption of alcohol.

Biochemical differences related to alcohol metabolism have been found between C57BL/Crgl and DBA/2Crgl mice. The high-preference strain (C57BL) metabolizes ethanol more rapidly than the low-preference strain ( Schlesinger, 1964). The results are consistent with the hypothesis that high-preference animals drink more alcohol because less acetaldehyde, a toxic intermediary in alcohol metabolism, accumulates in their blood. The content of alcohol dehydrogenase in the liver of C57BL/6 mice increased significantly after forced consumption of 10 per cent ethanol, but other liver enzymes remained unchanged. It is possible that the behavioral phenotype can be explained by a combination of biochemistry and reinforcement theory.

Social behavior

Laboratory mice live in groups sharing small cages, except when isolated for experimental purposes. Wild mice may also attain high population densities in grain ricks or other unusually favorable conditions ( Elton, 1942). Though they commonly live in close proximity to one another, mice are not regarded as highly social mammals because group organization is minimal. Nonetheless, social behavior as shown in fighting, mating, and care of young has been investigated for many purposes, ranging from theoretically oriented investigations of the causes of fighting ( Scott and Fredericson, 1951) to a search for drugs to control aggressiveness ( Scriabine and Blake, 1962). Although some strain comparisons have been made, the genetics of social behavior has not been widely studied.

Fighting. Mouse breeders are well aware of the propensity for fighting shown by many strains, particularly by males. Severe wounds are inflicted in combat, and battles to the death are not uncommon. Several investigators have described the fighting of mice (reviewed in Scott and Fredericson, 1951). According to van Abeelen ( 1963a), the elements of male-male interaction include fixing (staring at rival), dancing, boxing, kicking, nosing, wrestling, biting, chasing or fleeing, and submissive posture. Tail rattling and fur fluffing are often seen when mice are paired and may represent emotional responses ( Scott, 1947).

Laboratory mice housed in cages typically develop a social organization based upon exclusive dominance of one male ( Uhrich, 1938). The despot ordinarily retains his position for some months. Linear dominance is occasionally observed in which an α-male dominates a β-male, who in turn dominates others below him in the social order. Equal or unsettled dominance relationships are seen in newly assembled groups. Only when mature animals were matched against juveniles was dominance not correlated with weight in Ulrich's colony of heterogeneous albino mice. A strong "home-cage" effect was seen in the usual advantage of residents over an introduced stranger, regardless of his previous high social status in another situation. Success in fighting was not the basis of success in mating.

Methods of studying fighting quantitatively can be divided according to whether subjects fight to a decision or are separated when fighting commences. The tests may also be separated into noncompetitive ones in which animals fight in bare cages and competitive ones in which the animals have restricted access to food or to shelter from electric shock ( Scott and Fredericson, 1951).

Fighting tests are usually conducted un special pens with a removable center partition, thus minimizing handling. Male mice will fight readily in a bare arena, but somewhat more consistent results are reported when hungry mice compete for a loose pellet of food ( Fredericson, 1950; Fredericson et al., 1955), Females (C57BL/10) which seldom fight spontaneously will fight regularly in a competitive situation ( Fredericson, 1952). Success in competition to escape noxious stimuli has been used to evaluate aggressiveness of different strains ( Doner et al., 1952) and effects of hormone treatments ( Bevan et al., 1960). A somewhat different means of inducing fighting is to give foot shocks to mice confined within a narrow space. Most pairs attack each other promptly ( Tedeschi et al., 1959).

Fighting until decision results in training for dominance or submission. Some investigators purposely pretrain for one or the other. Fighters are readily trained by the "dangler" method in which other mice, held by forceps so they cannot defend themselves, are presented to the potential fighter ( Scott, 1946; Bauer, 1956). Mice may be trained for submission by exposure to such a trained fighter ( Kahn, 1951). The round-robin method of matching members of a set of subjects in all possible pairings involves training of a less directed type ( Beeman, 1947), but most subjects become habitually dominant or submissive during the course of testing.

Nondecision procedures usually involve isolation after weaning ( King, 1957a; Tollman and King, 1956), though shorter isolation periods have been employed in some drug assays ( Scriabine and Blake, 1962). In such methods a pair of mice is taken as the unit, and aggressiveness is measured by the latent period until the beginning of a fight or the accumulated attacking time over a fixed period ( Catlett, 1961). The advantage of the latency measure is that subjects can be separated before fighting is inhibited by injury or reinforced by victory. A great variety of other measures have been used ranging from simple occurrence or nonoccurrence of fighting in a pair to rating scales ( Lagerspetz, 1964) and elaborate counts of specific acts ( Banks, 1962).

Some methods of measuring social dominance do not involve actual fighting. In the tube-dominance test, mice deprived of food are trained to go down a narrow tube to a chamber where food is provided ( Lindzey et al., 1961). On the dominance trial, two animals are started simultaneously at opposite ends, and the animal which retreats is scored as submissive ( Figure 33-2).

The amount of fighting between males varies according to previous experience ( Scott, 1944). New males (C57BL/10) were accepted peacefully by mated males living in a large multiple-escape pen. After a period of isolation, during which residents were trained to fight, the pairs fought vigorously when they were again brought together.

Castration reduces male aggressive behavior as it does in many species. Beeman ( 1947) found that aggressive behavior in male C57BL/10 and BALB/c mice was eliminated if 25 or more days elapsed between extirpation of the testes and the initial encounter. She reported that implants of testosterone propionate promptly restored aggressive behavior to the level characteristic of the strain.

Bevan and his collaborators reported somewhat different findings. Replacement androgen therapy only partially restored aggressive behavior of castrated SWR mice, and doses of 600 μg per day of testosterone propionate even depressed it ( Bevan et al., 1958). When pretraining, castration, and testosterone therapy were varied independently, the effects of the hormone on fighting success appeared to be mediated chiefly through an effect on body weight. Effects of pretraining were more striking than those of androgen ( Bevan et al., 1960). An additional complication in the androgen-aggression relationship is that testosterone induces spontaneous aggression in prepuberally castrated males but not in females ( Tollman and King, 1956). Some of the discrepancies between these studies are certainly byproducts of differences in techniques of measurement and in the handling of the subjects. It appears likely, however, that some genetic variation exists in the association between androgen and fighting and that more is involved than the sensitization of an "aggression center" by hormones.

Reduction of aggression is the purpose of much drug treatment of behavioral disorders in man. A comparison of various tranquilizers showed that meprobamate was very effective in reducing fighting in mice, though it had no effect on spontaneous motor activity ( Tedeschi et al., 1959). Janssen et al. ( 1960), who used spontaneous fighting rather than foot shock-induced fighting as an index found striking differences in the antifighting potency of drugs relative to their motor effects. It appears that tests of fighting behavior in mice have merit for evaluation of potential tranquilizing agents ( Scriabine and Blake, 1962). One may conjecture that an ideal drug would reduce spontaneous fighting but would leave competitive fighting intact, since success in obtaining food must have survival value.

Studies on the genetics of fighting in mice have been restricted, with one exception, to strain comparisons. Ginsburg and Allee ( 1942) found the descending order of fighting ability in three strains to be C57BL, C3H, and BALB/c. Several researchers have used BALB/c as a relatively nonaggressive strain, but little effort seems to have been made to survey the large number of available strains. Early experience in victory or defeat can produce great variation in fighting within a fixed genotype ( Scott, 1946; Kahn, 1951). Furthermore, the type of test used is important. Pairs of BALB/c males usually share food pellets, but BALB/c males paired with C57BL/10 males become aggressive ( Fredericson and Birnbaum, 1954). Lindzey et al. ( 1961) compared social dominance in pairings between three inbred strains using the tube-dominance apparatus. The outcomes in terms of percentage of animals winning a majority of contests for each pairing were A/alb (100) and DBA/8 (0); A/alb (86) and C3H (14); C3H (86) and DBA/8 (14).

Lagerspetz ( 1964) obtained good separation of aggressive and nonaggressive lines within two generations of selection from a heterogeneous population of albinos. Increased divergence was obtained in later generations, though the lines appeared to be fairly stable by the fourth generation. Only males could be tested because fighting was infrequent among females. The results suggested that a few genes make major contributions to the variation in fighting behavior observed in the original stock. Losing of fights reduced the aggressiveness score of both lines; winning enhanced fighting of initially aggressive animals but not of nonaggressive ones. Animals aroused by an immediately preceding fight would cross an electrified grid to attack an opponent or would chose an opportunity to fight rather than remain peaceful.

Mating. Courtship follows a pattern similar to that of other laboratory rodents but has species-specific characteristics. The basic sequence consists of elements described as sniffing, following, mounting, mounting-with-intromission, and post-copulatory grooming ( van Abeelen, 1963a; Grant and Mackintosh, 1963; Lipkow, 1960; McGill 1962). The sexually aroused male often crawls in front of the female or even under her ("rooting," McGill, 1962). There may be one to more than 100 intromissions before ejaculation, which is often marked by the male rolling over on his side, often carrying the female with him. At ejaculation the male accessory glands produce a secretion which hardens to form the vaginal plug. Only one ejaculation usually occurs per day.

Inbred strains differ quantitatively in many aspects of mating behavior. Males of C57BL/6 typically gain intromission rapidly and ejaculate in 20 minutes. DBA/2 males are slow to achieve intromission, but ejaculate quickly thereafter. Slowest to achieve ejaculation are BALB/c males (average latency of 1 hour), largely because of a long period of courtship ( McGill, 1962). Even more striking genetic effects are found in recovery time, the interval between successful copulations ( McGill and Blight, 1963a). Recovery time in C57BL/6 males averages 4 days but only 1 hour in DBA/2 males. The F₁ hybrids resemble DBA/2 parents, and backcrosses to C57BL/6 yield intermediate latencies. The physiological and psychological bases of various components of mating behavior are complex ( McGill and Blight, 1963b). Levine ( 1958) found that albino males (strain ST) sired 76 out of 88 litters from ST females when competing against pigmented males (CBA). Success in mating was not correlated with social dominance in a test of aggression. Although ST males had a slight advantage in initial encounters, all CBA males achieved dominance in continued competition ( Levine, 1963). Solution of this apparent contradiction must be based upon direct observation of mating behavior in a competitive situation.

Caretaking. The parturient female generally constructs a hollow nest of an available material. She spends much time in the nest huddled over the young. The young are licked at frequent intervals, most often in the perineal region. Retrieving of young removed from the nest, amount of huddling, and nest building have been used as the basis of a quantitative rating scale for maternal behavior ( Leblond and Nelson, 1937; Leblond, 1940).

Young mice elicit retrieving much more effectively than older ones. The percentage of successful retrievals with 1-day-old young was 83; 5-day-old, 78; 10-day-old, 54; and 15-day-old, 11. Leblond ( 1940) found that the expression of maternal behavior in mice was little influenced by hormones of parturition and lactation. Maternal behavior could be evoked in intact and hypophysectomized males and virgin females by leaving young in the cage, a process called "sensitization." Males were found which retrieved dislocated young mice as readily as did females.

The extensive investigations of Beniest-Noirot ( 1958) have confirmed the independence of so-called maternal behavior from hormonal control. Postparturient and virgin females and males performed equally well on her maternal behavior tests, which included retrieving, nest building, nursing care, and assumption of a nursing position. Similarly, adult mice of both sexes with and without breeding histories ate placentas, gnawed umbilical cords, and cleaned young when confronted with newborn mice. These activities are not maternal in a strict sense but are responses made to appropriate stimuli. The only probable hormonal effect noted by Beniest-Noirot was increased defense of the nest by postparturient females. Caretaking responses, though not dependent upon prior experience with infant mice, were more frequent in mice which had spent 3 days with their dam's second litter.

Whisker-eating, mutilation, and cannibalism of young are disturbances of caretaking behavior which vary in incidence among inbred strains. Their genetics has been reported only in a preliminary fashion (Hauschka, 1952a, 1952b). Cannibalism is often a serious problem to the animal breeder and experimentalist, but no thoroughly satisfactory explanation has been advanced for its occurrence in certain lines and individuals. Disturbing the mother may be an inciting factor. Beniest-Noirot ( 1958) found more cannibalism in mothers selected for strong defense of their nests against intruders. Some experimenters working with neonatal mice routinely remove the mother and anesthetize her lightly with ether, while the young are being handled, to reduce the incidence of cannibalism.

Emotionality

Willingham ( 1956) measured defecation, urination, activity, freezing, emergence into open and enclosed areas, and squeaking in several different situations. The matrix of intercorrelations was factor-analyzed, and six factors were extracted. The first, called "elimination," had significant loadings on all the defecation and urination measures. The second factor was named "freezing." Willingham concluded that there are a variety of nearly independent types of emotional behavior and that the concept of general emotionality is an oversimplification.

Thompson ( 1953) scored 15 mouse strains on incidence of defecation in the open field ( Table 33-1). Several other investigators have used elimination as a measure of timidity ( Lindzey, 1951; Fuller et al., 1956).

An important challenge to the validity of elimination as an index of emotionality has recently come from Bruell ( 1963), who observed defecation in 25 different genotypes. Males defecated more than females, and hybrid males defecated more than their inbred sires. However, hybrid females defecated less than their inbred dams. Bruell's interpretation was that the defecation of male mice in a strange environment is a territory-marking response. Thus, increased defecation in hybrids was interpreted as restoration of adaptive responses depressed by inbreeding.

Antalfi ( 1963) has shown that freezing (William's second factor) in the presence of stress may be less strong in laboratory mice than in wild house mice. Freezing also has possible adaptive significance.

Learning

Conditioned responses. The earliest attempt to establish conditioned responses in mice came, most appropriately, from Pavlov ( 1923):

The latest experiments (which are not yet finished) show that the conditioned reflexes, i.e., the highest nervous activity, are inherited. At present some experiments on white mice have been completed. Conditioned reflexes to electric bells are formed, so that animals are trained to run to their feeding place on the ringing of the bell. The following results have been obtained:

The first generation of white mice required 300 lessons. Three hundred times it was necessary to combine the feeding of the mice with the ringing of the bell in order to accustom them to run to the feeding place on hearing the bell ring. The second generation required for the same result, only 100 lessons. The third generation learned to do it after 30 lessons. The fourth generation required only 10 lessons. The last generation which I saw before leaving Petrograd learned the lesson after 5 repetitions. The sixth generation will be tested after my return. I think it very probable that after some time a new generation of mice will run to the feeding place on hearing the bell with no previous lesson.

We do not know what Pavlov found on his return to Petrograd.

Denenberg ( 1958; 1959a, 1959b; 1960) studied the formation of conditioned emotional responses in C57BL/10J and C57BL/10Sc mice. In one version of his procedure, the buzzer was sounded for 3 seconds and then followed by a short pause, after which the shock was presented for 1 second ( Denenberg, 1958). Typical responses to shock were described as yelping, walking, running, climbing, and rearing. The occurrences of any of these responses to the buzzer was considered evidence of conditioning.

Operant conditioning. Studies on operant conditioning in the mouse have used sensory reinforcers — lights, primarily — and lever contact rather than pressing. Indeed, one of the earliest reports of a successful utilization of light as a reinforcer used mice ( Kish, 1955). Various studies have used males of strains C57BL/10J ( Kish, 1955), C57BL/6J ( Barnes and Baron, 1961; Kish and Baron, 1962; Baron and Kish, 1962), and DBA/1J ( Barnes et al., 1959), with age range of 8 to 17 weeks. Pretraining in darkness is important ( Kish and Baron, 1962) and the light source should be weak ( Barnes et al., 1959). Barnes and Baron ( 1961) studied the effects of luminous line drawings as reinforcers, concluding that their effectiveness increased as they became more complex.

Barnes and Kish ( 1957) reported that termination of high intensity sound acts as a negative reinforcer, i.e., termination can be used to strengthen a response. Use of sound as a positive reinforcer for lever contact has also been studied. The definitive study is that of Barnes and Kish ( 1961) who reported the effects of 10 frequencies at five different intensity levels. Reinforcing effects were quite weak. In a later study comparing directly the effectiveness of light and sound, Baron and Kish ( 1962) reported that low intensity sound may have aversive properties. Clearly, light and luminous drawing are better choices as reinforcers.

Escape and avoidance. Essman and Jarvik ( 1961a) devised an apparatus consisting of a large container of water with an escape ramp fastened on one corner. The animal was placed in the water on the side opposite the escape ramp and time to reach and climb the escape ladder were recorded. Learning is extremely rapid ( Figure 33-3). The technique has been used for strain comparisons by Winston ( 1963) and for comparison of mutants by Denenberg et al. ( 1963).

The essence of avoidance conditioning is that the organism is trained to make a response which postpones the impending occurrence of noxious stimulation. A number of methods have been devised for the study of avoidance learning in mice. One technique employs a box with a grid floor partitioned in the middle by a hurdle. A warning stimulus — light or buzzer — begins a few seconds before shock is delivered to the floor grids on the side on which the animal is located. The animal is able to avoid shock by jumping the hurdle. The warning stimulus goes on again a few seconds or minutes later, and the animal must quickly jump back if it is again to avoid the shock. Excellent descriptions of the use of such procedures with mice are given by Caldwell ( 1962) and by Royce and Covington ( 1960). Caldwell's technique worked moderately well with the female CF#1 mice he used.

Royce and Covington ( 1960) compared nine inbred strains maintained at The Jackson Laboratory. Nearly one-third of the animals could not be conditioned in 700 trials — half or more of strains C57BR/cd, AKR, C57BL/6, and A. Table 33-2 presents mean number of trials to a criterion of five successive avoidances for the strains in which conditioning was easiest. Subsequent work ( Carran et al., 1964) showed that strain differences in conditioning rate are dependent in part upon the shock voltage used. Whether the basis of the voltage dependence is associated with strain differences in shock sensitivity was not made clear. Two strains differing in conditionability were shown to diverge in measured skin resistance during training. King and Mavromatis ( 1956) and Stanley and Monkman ( 1956) also studied avoidance conditioning of mice.

A different kind of device has also been widely used. As described by Denenberg et al. ( 1958), it consists of a narrow rectangular box with partitions. The floor is an electrifiable grid. Near one end is the starting chamber in which the animal is restrained by a guillotine door. A safe chamber, its grid floor covered with wood, is located at the other end. The door of the starting chamber is opened and a 100-watt bulb is turned on. Five seconds later, the floor is electrified. The animal may escape or avoid the electric shock by opening the door of the safe chamber and entering. Use of the device was reported in several articles ( Denenberg and Bell, 1959; Denenberg et al., 1963; Bell and Denenberg, 1963).

Ekström and Sandberg ( 1962) and Jarvik and Essman ( 1960) have also reported avoidance conditioning techniques. Ekström and Sandberg trained mice to avoid shock by jumping to a wire net hung on the wall. A rather large number of training trials was required for only moderate success.

The Jarvik and Essman device produced extremely rapid learning ( Figure 33-4). Subjects were placed on platform A and the shelf B was then pulled out. A step down from the platform to the shelf resulted in shock to the animal. Learning was measured by placing the animal back on the platform and measuring the amount of time it delayed before stepping down again. Comparison of shocked and nonshocked animals 24 hours after a single training trial resulted in 10 per cent of the previously shocked and 90 per cent of the unshocked animals leaving the platform in 10 seconds or less. This and similar devices have been used in three other studies ( Essman and Alpern, 1964; Essman and Jarvik, 1961).

Maze learning. Early work on maze learning in mice was done by Yerkes ( 1907) and Vicari ( 1923, 1924, 1929). More recently, Lindzey and Winston ( 1962) compared the leaning performance of several strains in a six-unit T maze. Twenty animals each of strains A/alb, C3H/Bi, DBA/8, and C57BL/1 were given one trial per day for 14 days in a study in which all animals received prior handling. Strain C3H/Bi was poorer than the other three strains. A second experiment indicated differential strain responsiveness to handling. Winston ( 1963) used a similar maze and procedures with strains A/alb, C3H/Bi, and DBA/8, and also showed strain C3H/Bi to be poorer. Enclosed mazes have also been used by Hall ( 1959) and Smith and Bevan ( 1957).

Morgan ( 1963) compared the effectiveness of food and a dark hiding place as reinforcers for mice of strains SW and C57BL. Animals were given 20 trials on the first day and 10 trials per day thereafter. All groups attained a level of 9 out of 10 correct by the end of 60 trials. Groups running to a dark hiding place learned faster. No strain differences or interaction with reinforcer were found. A diagram of the apparatus is presented in Figure 33-5.

McClearn ( 1958, 1965) compared the performance of strains C3H/NT, C57BL, and BALB/c on an elevated maze. In decreasing order of performance, the strains were BALB/c, C57BL, and C3H/NT. Lindzey and Winston ( 1962) and Winston ( 1963) also found C3H mice to be poor learners, possibly because they are blind or nearly so ( Chapter 32).

Meier and Foshee ( 1963) compared performances of strains AKR, BALB/c, C3H/He, C57BL/6, DBA/2, and noninbred CF#1 in a three-choice-point water maze in which escape from water was used as a reinforcer. The strains divided into two groups on the basis of time to reach an escape ramp; slower strains were BALB/c, AKR, and C3H/He. No error score analyses were included. This and a second experiment ( Meier, 1964) demonstrate that level of performance for some strains depends in part upon the age at which the tests are given.

Discrimination learning. Much of the literature on discrimination learning is reviewed in Chapter 32, where information on sensory capacity of mice is given. However, two techniques of special interest are better described here.

What appears to be an extremely promising apparatus for the study of discrimination learning was devised by Keeler ( 1927) and used but once ( Figure 33-6). The device was made of two cylinders of sheet metal. The floors of the central cylinder and those of the external chambers were independently wired for shock. There were six doorways from the central cylinder. A light was turned on in one of the external chambers and the floor of that compartment was isolated from the circuit, all others being electrified. The mouse was placed in the center cylinder, the center floor was electrified, and the mouse had to escape to the lighted compartment. Learning curves of normal and rodless ( Chapter 32) mice tested in the device are presented in Figure 33-7. It is evident that learning was extremely rapid.

A discrimination device with which we have had personal experience is the water maze devised by Waller et al. ( 1960; Figure 33-8). Animals were trained to swim either the white or black side. The entire bottomless maze could be inverted in the water to reverse the orientation of the correct alternative. C57BL/6J males were given five trials per day for 12 consecutive days at one of three water temperatures. Waller et al. found that animals swam faster at lower temperatures, but that there was no difference in error performance among the three water temperatures used.

We have successfully used the device for black-white discrimination learning with six strains ( Wimer and Weller, 1965). All but one strain performed significantly above chance by the fourth day. A total time investment of about 40 minutes per animal was required to achieve nearly errorless performance in some strains. In decreasing order of performance, the strains were C57BL/6J, DBA/2J, RF/J, AKR/J, and A/HeJ. Except for strains AKR/J and A/HeJ, the difference in performance between adjacent strains was not significant. However, differences between all strains separated by one other were significant. Our experience suggests that water temperature may affect accuracy of performance in some strains. DBA/2J and C57BL/6J discriminate horizontal from vertical striations with approximately equal success and at a level of performance only very slightly inferior to that for black-white discrimination (Wimer, unpublished data).

Exploratory activity

Exploratory activity — activity in an unfamiliar environment — is a comparatively well-studied mouse behavior. Thompson ( 1953) measured activity levels of 15 inbred strains in an enclosed arena ( Table 33-3). In a later study he measured arena and Y-maze activity in large samples of five strains ( Thompson, 1956). Thompson's two studies showed that there is considerable variability between strains and that the ordering of the strains is invariant in the two testing devices, with the exceptions of strains C57BL/6J and AKR/J. McClearn ( 1959) attempted to establish the generality of Thompson's results with strains C57BR/cd, C57BL/10, LP, AKR, BALB/c and A/J by observing their behavior in other situations in which one might reasonably expect the same ordering to be maintained. McClearn's findings indicated that Thompson's characterizations of strains has wide generality. The basis for these strain differences is not known. Thiessen ( 1961) attempted to relate body weight to activity, but his results were inconclusive.

McClearn's ( 1960) study of the effects of varying level of illumination suggests that there are limits to the conditions under which ordering of strains on activity is maintained. He measured activity levels of strains C57BL/Crgl and A/Crgl in an open field with barriers, in bright white light and in dim red light. Note in Figure 33-9 that strain A showed higher activity when tested under dim red light, whereas strain C57BL showed less. Although variation in level of illumination did not reverse the ordering of the strains, there was a significant interaction of strain and illumination. Difference in eye pigmentation between the strains might be partially responsible for the observed phenomenon. However, McClearn ( 1959) found that hybrids of strain C57BL/10 and A/J were intermediate in activity. Thus, more than eye pigmentation was involved, for the hybrids are dark-eyed like the C57BL/10 parents.

There are two other studies of exploratory activity using mice ( Wimer and Sterns, 1964; Wimer and Fuller, 1965).

ENVIRONMENT AND PHENOTYPE

The purpose of this section is to present information on the effects of maintenance conditions and early environment on behavioral phenotypes.

Maintenance conditions

Light cycle. Cycles of approximately 24 hours (circadian rhythyms) in laboratory mice have been so well demonstrated that artificially controlled light-dark cycles in the animal room must be regarded as essential for quantitative behavioral and physiological studies.

Strain I mice were much more susceptible to audiogenic seizures during the night ( Halberg et al., 1955), and the peak hours of susceptibility could be shifted by artificially reversing the lighting cycle in the animal quarters ( Halberg et al., 1958). Halberg et al. ( 1959) have shown also that variation in seizure susceptibility is merely one of many biological processes changing rapidly during the day (e.g., such diverse phenomena as blood cortisone, number of mitoses in adrenal cortex, and total body activity). The importance of controlling time-of-day effects in psychological studies is clearly indicated. Higher paroxysmal activity in electrocorticograms (ECoG) has been found during the evening hours and may relate to increased susceptibility to audiogenic seizure ( Harner, 1961). Since ECoG's were recorded under pentobarbital anesthesia, it is equally possible that differences were related primarily to a cycle in drug susceptibility. The relationship of the diurnal cycle to drug sensitivity is complex. Under constant illumination, variation in pentobarbital sleeping time with time of day was absent; under cyclic lighting, grouped subjects slept longer when tested in the light period but not in the dark ( Davis, 1962) ( Table 33-4). Periodicity in activity in waltzing mice was maintained in constant darkness, and even occurred in mice born and reared in the dark, though here the cycles were not correlated with the natural day-night cycle ( Wolf, 1930).

Temperature. Temperature control is also widely regarded as essential in the animal room. Mice, however, have adapted quite successfully to ambient temperatures as low as -3°C, provided that ample nesting material was supplied (Barnett, 1956, 1959). Improved nest building was a major means of adaptation in both A and C57BL mice. Increased heat production of mice reared in the cold enabled them to withstand low-temperature stress much better than animals reared in heated laboratories, showing that physiological as well as behavioral adaptation had occurred. Improved body insulation played a minor role in adaptation, which must therefore have been based on increased heat production.

Population density. The relation between crowding and the endocrine system has been widely studied. Increased adrenal size and decreased size of male sex structures in crowded animals led Christian ( 1955) to postulate such endocrine effects as a mechanism for regulation of population density. Some studies have failed to confirm all of Christian's findings and have thus cast doubt on this hypothesis of density control. Southwick and Bland ( 1959) found no difference in the adrenal weights of CFW male mice reared in isolation or in groups of 2, 4, 8, and 16 per cage. However, they did not find adrenal enlargement in wounded mice (presumably socially subordinate animals subjected to chronic stress). The males in groups of four or eight actually sired more young than isolates when females were introduced to their cages, a result contrary to Christian's hypothesis.

The effects of population density upon adrenal activity seem to be largely dependent upon the type of social organization in the group. A number of investigators have used eosinopenia and other measures as indicators of the stressfulness of various types of social grooming imposed by the experimenter ( Southwick, 1959; Vandenbergh, 1960; Thiessen et al., 1962; Thiessen and Nealy, 1962; Bronson and Eleftheriou, 1963; Bronson, 1963). Enhanced adrenal function parallels social tension (grouping of strange males, subjection to defeat by fighters, as examples) rather than physical crowding alone. These findings are pertinent to research in ecological pathology, which is concerned with the effects of housing conditions upon cancer and other diseases.

C3H mice kept in isolation in a long-term experiment on nutrition showed a high incidence of convulsion (12 in 20) contrasted with subjects housed in one large cage (3 in 20). The mean lifespan of the isolates was 434 days; of the group-reared, 500 days ( J.T. King et al., 1955). Group rearing also favored survival of C57Bl/10 mice weaned early, apparently as a result of better heat conservation ( J.A. King and Cannon, 1955). Prior social experience raises the lethal threshold for amphetamine and counteracts the enhanced susceptibility found when mice reared in isolation are aggregated just prior to toxicity testing ( Mast and Heimstra, 1962).

It must appear from these latter studies that the relationship between the behavioral and physiological levels of integration are complex. The laboratory mouse is one of the best subjects for investigation of these interactions and has the further advantage of permitting an evaluation of genetic effects.

Early environment Environment during both prenatal and postnatal developmental periods has a marked effect on behavioral phenotypes.

Prenatal. Lieberman ( 1963) provided an excellent demonstration of the effects of the prenatal environment. Pregnant females of strain C57Bl/6 were given (1) saline injection, (2) epinephrine injection, (3) norepinephrine injection, (4) hydrocortisone injection, or (5) were crowded in a cage with 10 aggressive males. All treatments were given during the second trimester of pregnancy. Offspring were tested at 35 days of age in an open field test. In comparison with the saline controls, offspring of crowded and epinephrine-injected dams showed increased activity and decreased defecation. Offspring of hydrocortisone- and norepinephrine-injected dams showed decreased activity and increased defecation. Some, but not all, differences attained acceptable standards of statistical significance.

Weir and De Fries ( 1964) subjected pregnant females of strains BALB/cJ and C57BL/6J to a daily regimen of swimming in a water tank, exposure to intermittent loud tones in a tilt box, and placement in a brightly illuminated open field. Later test of the activity of offspring in an open field showed complex interactions of the treatment effect with genotype.

Postnatal. It is generally believed that certain kinds of experience may have the most profound and long-lasting effects if encountered early in postnatal life. That such experience may affect phenotype was dramatically demonstrated by Denenberg et al. ( 1964), who reared C57BL/10 young mice under a variety of conditions. For example, some mice (controls) were reared with other mice both before and after weaning, while other animals were transferred to a lactating female rat and reared with two male and two female offspring of her litter. Mice reared with rats from a very early age tended to be less active in an open field and less aggressive than mice reared with other mice. Preference for own species was very strongly affected by the treatment: Mice reared with rats preferred them to other mice.

A recent study by Ressler ( 1962) provides an excellent example in which strain of both parent and offspring of C57BL/10 and BALB/c mice contributed to an environmental effect. All offspring were foster-reared by parents of the same or different strain. Ressler found that BALB/c parents handled young of both strains more than C57BL/10 parents did. In addition, young BALB/c mice received more handling from both strains of parents than did C57BL/10 young.

Since variations of handling in infancy by humans have been shown to influence a variety of behavioral characteristics in adulthood, the superficially slight difference in the early environment of the young mice might be important. Ressler ( 1963) reported on differences in operant conditioning. Pressure on a wire mesh door was used as the response. For the first 15 minutes of testing, each press merely actuated a counter, while in the second half it briefly illuminated the dark box in which the animal was kept. The strain of the young mice had a significant effect on the number of presses during the dark phase, while the strain of foster parents had a significant effect on the visual exploration score (number of presses during the phase in which the light was operative minus presses during the dark phase). Young of both strains engaged in more visual exploration if reared by BALB/c parents. Results are presented in Table 33-5.

King and his associates have produced differences in aggressiveness within strains by varying size of social group. King and Gurney ( 1954) treated C57BL/10 males in one of three ways. Some were housed individually from 20 to 100 days of age. Others were reared with sires and male sibs until 45 days and then isolated until over 100 days. A third group was reared with female sibs until 45 days and then isolated. Latency of fighting was measured after bringing together individuals reared similarly. Animals isolated at 20 days were found to be much less aggressive. King ( 1957b) used strains C57BL/10J and BALB/cJ and a variety of treatments incompletely replicated in both strains. King stated that his results showed C57BL/10J males isolated at 20 days to be slower to fight, whereas BALB/cJ males were unaffected by this treatment. Similar early treatment produced no major effects on sexual behavior ( King, 1956).

The effect of early human handling on emotionality and fighting behavior in C57BL/10 mice has been studied by Levine ( 1959). Handling reduced the freezing component of emotionality and made animals more aggressive. Hall and Whiteman ( 1951) and Lindzey et al. ( 1960) also studied the effects of early experience on adult emotionality. Effects on other behavior have been reported by Baron et al. ( 1962), Bell and Denenberg ( 1963), Denenberg ( 1958, 1959, 1960), Denenberg and Bell ( 1960), and Stanley and Monkman ( 1956).

GENETIC ANALYSIS OF BEHAVIOR

The diversity of mouse stocks affords a corresponding diversity of possible analytic techniques for exploiting the genetic material, which varies from lines segregating at a single locus to those differing at many loci. The purpose of this section is to present examples of various techniques which have been applied to the genetic analysis of mouse behavior.

Stocks segregating at a single locus

The establishment of a behavioral difference between genotypes of a strain segregating at only a single locus is highly informative, for all possible pathways between gene and phenotype must converge on one biochemical process or on one regulator controlling a few related processes. A small number of studies on behavioral differences have been reported.

Studies of obesity ( ob) illustrate the possibilities for physiological and biochemical analyses after establishment of a genetic effect. Reared under typical laboratory conditions, ob/ ob mice attain a weight twice that of littermate non-obese sibs.

They characteristically eat heartily and are inactive and infertile. Both excessive weight and infertility may be eliminated by restriction of food intake ( Runner and Gates, 1954; Lane and Dickie, 1954). Increased food consumption could be due either (1) to a defect in central mechanism regulating caloric intake or (2) to differences in the hedonic attributes of food ingestion per se. A series of experiments were performed to identify the underlying mechanisms.

Fuller and Jacoby ( 1955) decreased the caloric value of the regular laboratory food by diluting it with cellulose or increased it by adding fat. Normal mice responded rapidly to decreased caloric value by increasing consumption, but obese mice increased intake by a smaller amount. Both groups ate less of the fat-enriched diet, but the adjustment was again much greater in the nonobese mice. When bitter substances — quinine or caramelized sugar — were added to food, both obese mice and controls sharply reduced their consumption. However, controls slowly increased their ingestion over a few days, whereas obese mice continued to eat reduced amounts. Results, then, favored a defect in central mechanisms producing adaptive regulation of food intake.

The behavior of obese mice appears very similar to that produced in rats by surgical intervention in the hypothalamus. However, the hypothalamus of obese mice appears to be normal histologically ( Maren, 1952). This does not, of course, eliminate the possibility either of subtle structural differences or of a biochemical disorder ( Chapter 19). Injection of gold thioglucose may result in obesity, and because of this similarity of effect to hypothalamic lesions it has been assumed that the hypothalamus is the site of action. Hollifield et al. ( 1955) compared ob/ ob mice with those made obese by injection of gold thioglucose. Gold thioglucose produced obese mice which did not increase activity during food deprivation nor show decreased activity following refeeding. Hereditarily obese mice, however, responded in a manner similar to controls. We may conclude (1) that the mechanism of hereditary obesity differs from that produced by gold thioglucose and (2) that hereditarily obese mice do show some of the behavioral effects of food deprivation normally associated with drive.

Known genes may share control of a phenotype with other genes which are not separately identifiable. An example is the dilute gene ( d), which has attracted attention since Coleman ( 1960) demonstrated that homozygous d/ d mice have a phenylketonuria-like condition. Since phenylketonurics often have convulsions, it was considered possible that the high incidence of audiogenic seizures in DBA/2 mice ( d/ d) might be caused in part by their deficiency in phenylalanine hydroxylase activity.

Direct evidence on this point is conflicting. No correlation between dilute phenotype and seizure susceptibility was found in (DBA/2J x C57BL/6J)F₁ hybrids ( Fuller et al., 1950), but Huff and Huff ( 1962) found an excess of seizures in dilute animals from another source. Huff and Fuller ( 1962) failed to find an effect on susceptibility of d-locus substitution in a stock of DBA/1J mice segregating for + and d. They suspected that a mutation might have abolished the enzyme-inhibiting effects of the d/ d genotype and thus interfered with its facilitation of seizures. Comparison between this and earlier studies was rendered difficult by an intervening change in diet of the mice. An interaction between the dilute locus and diet was found by Coleman and Schleisinger ( 1965). Mice from homozygous dilute strains (DBA/2J, BDP/J, and P/J) showed progressive increase in seizure susceptibility during 7 weeks on a low-pyridoxine (vitamin B₆) diet; nondilute C57BL/6J mice showed no such increase. Heterozygotes (D2B6F₁) showed a small increase.

The experiments using gene substitutions at the ob and d loci have concentrated on the analysis of the behavioral effects in physiological terms. Other studies of mutants have been directed toward finding an association between a gene and behavior.

Ashman (1957, personal communication) used strains SEC/1Gn-se and SEC/2Gn- d to study the effects of the alleles normal-ear (+) and short-ear ( se), and dense coat color (+) and dilute coat color ( d). Constitution of the stocks made it possible to compare +/+ se/+ with +/+ se/ se, and d/+ se/ se with d/ d se/ se. Ashman found no differences in activity in a tilt cage and open field, in conditionability, or in susceptibility to audiogenic seizure.

Les ( 1958) compared mutant and wild-type mice in stocks congenic at other loci. The mutants were furless ( fs), short-ear ( se), dilute ( d), yellow ( A^y), albino ( c), black-and-tan ( a^t), misty ( m), and hairless ( hr). He obtained measures of locomotor activity, transporting shavings, nibbling through a cardboard partition, defecation and urination in an open field, and adult body weight. More fortunate than Ashman, he established several behavioral differences. For example, animals homozygous for furless ( fs/ fs) and hairless ( hr/ hr) were less active than their normal sibs ( fs/+ and hr/+), and yellow mice ( A^y) were less active than their normal sibs (+/+). Differences in nibbling and adult body weight were also found.

Though not based upon stocks segregating at a single locus, results obtained by Winston and Lindzey ( 1964) suggested a behavioral effect associated with the albino locus. Time to swim to an escape ramp was measured for pigmented (+/+) strains C3H/Bi, DBA/8, and JK and albino ( c/ c) strains A and BALB/c. The albino strains had longer time-scores. Of greater interest here were the performances of segregating stocks: An intercross of F₁ hybrids (A x DBA/8) x (A x DBA/8) resulted in F₂ offspring with coat color phenotypes albino ( c/ c) and pigmented ( c/+ and +/+). Albinos had higher swimming scores. Backcross (A x DBA/8) x A also resulted in both albino ( c/ c) and pigmented offspring ( c/+), and again albino mice were slower (had higher scores). The authors suggested that the recessive gene for albinism is also a recessive gene for slow water-escape behavior. Since the mating systems used did not minimize the likelihood of contribution of other genes in the A stock linked with the c allele and other genes in the DBA/8 stock linked with the + allele, no definite conclusions can be reached ( Meier et al., 1965).

Stocks differing at many loci

Since most behavioral variation appears to be determined by many pairs of genes with small individual effect acting together ( Broadhurst and Jinks, 1963), it may often not be possible to detect effects of a single-gene substitution. Several investigators using genetically more complex materials with discernable behavioral variation have applied the techniques of quantitative genetics to behavior. (For details of methods and assumptions see Chapter 9.)

Scaling. For a simple biometric analysis, the phenotypes of quantitative behavior genetics should have certain measurement characteristics. First, the effects of heredity and environment should act independently of each other, for interactions always make conceptualizations more difficult, and they frequently strain the statistical assumptions made in performing analyses. Specifically, variances in P₁, P₂, and F₁, all presumably environmental in origin, should be homogeneous for stocks reared under similar controlled conditions. Second, and for the same reasons, the genetic effects should be additive. Once the phenotypic values for P₁, P₂, and F₁ have been measured, the observed values for the F₂ and the two backcrosses should equal their expected values on some additive genetic model. Change of scale may eliminate interactions of genetic and environmental factors and produce genetic additivity. Such transformations should be considered whenever needed, for momentary awkwardness of working on an unfamiliar scale may be repaid by the long-term reduction in analytic complexity.

Bruell ( 1962) analyzed patterns of behavior according to Mather's ( 1949) two scaling criteria paraphrased above. Bruell's description of rationale and procedures for testing appropriateness of scale merit careful study. He measured the exploratory activity of strains A, C57BL/10, F₁'s, F₂'s, and backcrosses. The raw score was found adequate for the additivity criterion, and the variances of P₁, P₂, and F₁ were homogeneous. Figure 33-10 shows the discrepancies between observed and predicted values for the F₂ and the two backcrosses. The apparent overdominance of the F₁ was not significant. In an analysis of activity-wheel data, Bruell ( 1962) compared raw scores, square root scores, and logarithmic scores for additivity, finding only the log scale to be adequate. He was unable to achieve homogeneity of variance among P₁, P₂, and F₁. Bruell also analyzed time to climb down a pole in a similar manner.

Physiological studies have sometimes indicated than inbred strains may be more variable than their F₁ hybrids. Mordkoff and Fuller ( 1959) studied variability in activity in stocks C57BL/6J, DBA/2J, B6D2F₁, and DX (a 4-way cross of strains C57BL/6J, DBA/2J, BALB/cJ, and C3HeB/FeJ; Green, 1964). A reduced variability in hybrids was or was not shown, depending upon the measure used. Ranked from highest to lowest in variability between animals within a group, the order was C57BL/6J, DX, B6D2F₁, and DBA/2J. However, the coefficient of variation (s/x) being used, they ranked C57BL/6J, DBA/2J, B6D2F₁, and DX. Reduced variability in hybrids has been hypothesized to be due to better developmental homeostasis — higher stability resulting from more biochemical and physiological versatility in the developing young heterozygous animal. Morkoff and Fuller raised the question of whether developmental homeostasis would have the same meaning for behavior, suggesting that behavioral variability might produce better homeostasis. There is clearly need for careful distinction between homeostatic mechanisms and their consequences. A later study by Schlesinger and Mordkoff ( 1963) measured both locomotor activity and oxygen consumption in strains C57BL/6J, DBA/2J, and the F₁. They interpreted their results as showing less variability among F₁ animals on both measures.

Observations on the inheritance of seizure susceptibility in mice illustrate how polygenic systems, environmental effects, and choice of scale may combine to affect the results of an analysis. Fuller et al. ( 1950) selected strain DBA/2J (extremely susceptible to convulsions) and strain C57BL/6J (highly resistant) as parental stocks. Also tested were the F₁, F₂, and backcrosses. The measure was risk of convulsion during a 2-minute period of bell ringing. Particularly striking was the difference in seizure incidence between the original (32 per cent) and replicated (74 per cent) F₁ groups.

Fuller et al. ( 1950) assumed that a threshold model would account for their results. Two things are important in understanding the model: (1) the general nature of audiogenic seizures and (2) the measure sued. Seizures result in some manner from massive discharges of central nervous system neurons produced by the cumulative effects of impulses entering auditory nerves. Differences in seizure susceptibility presumably reflect a continuum of differences in the rate of accumulation and dissipation of the effects of stimulation. Some mice convulse quickly during a test, others slowly; some severely, others lightly; some show great excitement without convulsing, others are overtly unresponsive. Because of problems of reliability and scaling, the phenotypic measure generally used is dichotomous. That is, animals either do or do not convulse.

Fuller et al. ( 1950) proposed that the genotype of DBA/2 led to a phenotype almost always seizure susceptible and that of C57BL/6 to one almost always resistant. The F₁ genotype produced an intermediate phenotype leading to a seizure only with an environmental push. Figure 33-11 illustrates what might have happened if the two F₁ groups were reared or tested under somewhat different environmental conditions. It is important to remember that a phenotype characterized as being in one of two classes may be polygenically determined and that results of a genetic analysis are determined in part by the scale used.

Crosses, intercrosses, and backcrosses. As we have seen, the simplest quantitative model for any behavioral phenotype is that it is an additive function of heredity and environment. By use of this model and some basic statistical concepts of multiple-gene action, it is possible (1) to estimate the relative determinations of phenotype by genotype and (2) to make generalizations concerning the average contributions of genes from one parental stock as contrasted with those issuing from others.

McClearn ( 1961) crossed strains C57BL and A, intercrossed the resulting F₁ to produce an F₂, and backcrossed the F₁ to both parental stocks. The phenotype measured was number of squares traversed in an open field in a 3-minute period. Means of square root transforms, required to obtain homogeneity of variance among parental and hybrid stocks, are presented in Table 33-6. Deviation of the F₁ mean from the midpoint value of 7.45 indicated dominance of C57BL genes over those from strain A for this phenotype. Note that the ranking of groups corresponded generally to the percentage of C57BL genes present. The relative contribution of the genetic factor — heritability in the broad sense — was estimated by measuring the variability in the parental and F₁ groups (in all of which variability between animals within stocks was presumably environmental in origin) and subtracting that value from the observed variability of the F₂'s (in which both genetic and environmental sources of difference between individuals were operating). The ratio of residual genetic variability of the F₂'s to their total observed variability was 0.69, indicating a rather high degree of genetic determination. A variety of related analyses were performed by McClearn and Rodgers ( 1961) in their studies of alcohol preference, by McClearn ( 1959) in his study of exploratory behavior, and by Fuller and Thompson ( 1960, p. 268) for activity. Heritabilities computed from crosses between inbred lines should not be used to estimate heritabilities in natural populations of unknown genetic diversity. Their greatest value is to guide experimenters to model systems which are promising for the analysis of genetic processes by a variety of methods.

Another quantitative technique is the diallel cross (Chapters 2, 9). In this technique, k strains are crossed in all possible combinations to produce k (k - 1) hybrids, and phenotypic measures are taken on members of all groups. From the resulting k x k matrix of all phenotypic values classified by genotype, it is possible to assess (1) the general contribution (general combining ability) of genes from a specific parental strain when combined in hybrids with genes from all other strains, and (2) detect deviations (specific combining ability) from this general contribution occurring in specific crosses. Fuller ( 1964) applied this technique to the study of genetic factors in alcohol preference. The phenotypic values of parental stocks and hybrids are shown in Figure 33-12. The parental strains differed significantly among themselves, and the analysis to evaluate the differences in general and specific combining ability as measured in the hybrids showed both factors to be significant. Genes from the two extreme strains, C57Bl/6J and DBA/2J, had consistent effects of raising or lowering alcohol preference scores, whereas genes from strains A/J and C3HeB/FeJ had negligible general combining ability. For example, hybrids of A/J resulting from a cross with low-preference DBA/2J mice were like DBA/2J, hybrids from a cross with high-preference C57BL/6J mice were like C57BL/6J, and those from a cross with C3HeB/FeJ were almost identical with A/J. The dangers in generalizing from one specific cross are evident.

Bruell has related degree of heterosis or dominance observed in the F₁ to closeness of relation of parental strains. His stocks consisted of 31 F₁ hybrid groups produced from 13 parent strains, some of which were unrelated, moderately related, or related as sublines of a single strain. Bruell ( 1964a, 1964b) measured exploration of a strange environment and running in activity wheels. Hybrids typically explored more and were more active in activity wheels. Furthermore, the degree of superiority of the hybrids over the parental strains increased as relatedness of the parental strains decreased ( Figure 33-13). Bruell ( 1965) also measured latency of descent from a pole and emergence from a dark tunnel into an open field. Mode of inheritance for these activities tended to be intermediate. Barnett and Scott ( 1964) suggested that mode of inheritance in the F₁ hybrid depends upon the adaptive significance of the behavior, heterotic inheritance being observed most often in behavior having homeostatic function. Bruell's ( 1964a) position is similar.

Selection

Selection is an important but neglected technique for behavior geneticists. Parent-offspring correlations obtained during very early phases of selection provide an estimate of heritability, and speed of response to selection provides a clue to the number of loci involved ( Chapter 9). Examination of correlated responses to selection — changes in brain size or cell density, size of adrenals, etc. — could provide important insights into the mechanisms underlying a behavior pattern. Once phenotypically extreme stocks have been produced, they may be of value for a variety of purposes.

The power of selection as a tool for understanding the mechanisms underlying a behavior pattern is amply demonstrated in a study concerned with audiogenic seizures — a response to intense auditory stimulation characterized by running, convulsion, and death in highly susceptible mice. Frings and Frings ( 1953) selected separate lines for (1) high incidence of clonic-tonic seizures but low death rate, (2) high clonic seizure incidence, (3) low seizure susceptibility, and (4) susceptibility over a restricted age range. Results of selection indicated separate genetic control for age of susceptibility, severity of seizure, and resistance to death during seizure.

Two other behavioral studies have used the mouse: Dawson ( 1932) selected for running speed in a runway, and Lagerspetz ( 1964) developed strains differing in aggressiveness.

Multivariate phenotype

It is likely that genetically diverse stocks such as different inbred strains of mice will be behaviorally diverse for a variety of activities. Vanderpool and Davis ( 1962) employed a time-sampling technique to study the frequencies and durations of seven classes of orienting behavior in strains C57BL/6J, BALB/cJ, and DBA/2J. A variety of strain differences and similarities emerged.

Survey procedures have also been applied to identify behavioral differences associated with specific genetic loci. Thus, van Abeelen ( 1963a, 1963b, 1963c) applied ethological techniques, using an inventory of behavioral elements, including staring at the observer, hair fluffing, tail rattling, sniffing, reconnoitering, eating, digging, grooming, shaking the fur, wrestling, submissive posture, mounting, and many others. He reported, for example, that stocks carrying pink-eyed dilution ( p, I) stared at the observer less and groomed their fur more. Stocks carrying jerker ( je, XII) displayed clearly disturbed exploratory and aggressive behavior. Frequency of eating was not affected, but the eating posture was different.

Survey techniques have considerable potential value. First, they provide a profile of differences on a variety of activities. Second, they may identify specific activities and patterns of activities with higher heritabilities than those typically studied. A variety of multivariate statistical techniques useful in analysis are now available and practical ( Cooley and Lohnes, 1962; Kendall, 1961).

SUMMARY

Decades of labor by geneticists have made available an extremely wide variety of genetically controlled mouse stocks. No such diversity exists in any other animal species except the fruit flies. There is a large body of detailed and behaviorally relevant genetic, anatomical, physiological, and biochemical knowledge concerning these mouse stocks. Further, as shown in early sections of this chapter, experience with apparatus and techniques for behavioral studies is substantial.

Though the presentation was expressly intended for behavior geneticists, the house mouse should be of interest to many different kinds of behaviorists. General experimental psychologists who seek behavioral laws applying across organisms should find value in both the genetic uniformity within inbred strains and the genetic variability between them. Uniformity within strains has the special advantage that observed differences among individuals are presumably environmental in origin, so that one has a genetically "noiseless" system in which to evaluate the consequences of experiential treatments. Genetic variability between strains offers the possibility of repeating treatments across genotypes as a reasonable first step to estimating their generality. Physiological psychologists will find a wide range of anatomical characteristics among standard inbred strains. Some effects produced by mutant genes — for example, retinal degeneration ( rd) — can probably not be copied by any experimental treatment ( Chapter 32).

Interest in the behavior of mice is not restricted to behaviorists. Biological problems in such fields as reproductive physiology and possibly resistance to cancer have been found to have psychological aspects. The mouse has been exploited rather extensively by neurochemists, though little advantage has been taken of the genetic diversity within the species. If this is remedied, psychochemistry may come to depend upon the cooperation of behaviorists, chemists, and laboratory mice. It should be noted that psychotropic drug assay and screening procedures make use of large numbers of mice, generally of unrecorded ancestry.

A final comment is required to place the genetic and environmental determinants of behavioral phenotype in proper perspective. Our writing has, as a matter of convenience and clarity of exposition and as a reflection of our own interests and those of the anticipated readers, placed great and oversimplified emphasis on genetic determinants of behavior. Let us set the record straight: Just as it is impossible to conceive of an organism without a genotype, so it is also impossible to conceive of one which does not develop and is not maintained and tested in some environment. Some of the genetically associated differences in behavior reported may originate in differences in the pre- and postnatal environments which the different stocks afford (Ressler, 1962, 1963). Both direction and degree of other genetically associated differences may depend upon more general experimenter-controlled conditions of maintenance ( Cooper and Zubeck, 1958; Lieberman, 1963) and test ( McClearn, 1960).

We firmly believe that conceptions of behavior not jointly determined by heredity and environment are absurd ( Moltz, 1965). More-traditional psychologists have been criticized for failure to pay adequate attention to genetic factors in behavior. Behavior geneticists potentially run the opposite risk. Intelligently united, both approaches will increase our understanding of behavior.

¹The writing of this chapter was supported in part by Public Health Service Research Grant MH 01775 from National Institute of Mental Health.

Anliker, J., and J. Mayer. 1956. An operant conditioning technique for studying feeding-fasting patterns in normal and obese mice. J. Appl. Physiol. 8: 667-670.
See also PubMed.

Antalfi, S. 1963. Biological determination of the intensity of the alarm reaction in house mice. J. Comp. Physiol. Psychol. 56: 889-891.
See also PubMed.

Banks, E.M. 1962. A time and motion study of prefighting behavior in mice. J. Genet. Psychol. 101: 165-183.
See also PubMed.

Barnes, G.W., and A. Baron. 1961. The effect of sensory reinforcement on extinction behavior. J. Comp. Physiol. Psychol. 54: 461-465.
See also PubMed.

Barnes, G.W., and G.B. Kish. 1957. Reinforcing properties of the termination of intense auditory stimulation. J. Comp. Physiol. Psychol. 50: 40-43.
See also PubMed.

Barnes, G.W., and G.B. Kish. 1961. Reinforcing properties of the onset of auditory stimulation. J. Exp. Psychol. 62: 164-170.
See also PubMed.

Barnes, G.W., G.B. Kish, and W.O. Wood. 1959. The effect of light intensity when onset or termination of illumination is used as reinforcing stimulus. Psychol. Re. 9: 53-60.
See also PubMed.

Barnett, S.A. 1956. Endothermy and ectothermy in mice at - 3°C. J. Exp. Biol. 33: 124-133.

Barnett, S.A. 1959. The skin and hair of mice living at a low environmental temperature. Quart. J. Exp. Physiol. 44: 35-42.
See also PubMed.

Barnett, S.A., and S.G. Scott. 1964. Behavioral "vigor" in inbred and hybrid mice. Anim. Behav. 12: 325-337.

Baron, A., J.J. Antonitis, and S.F. Schell. 1962. Effects of early restriction and facilitation of climbing on later climbing behavior in mice. J. Comp. Physiol. Psychol. 55: 808-812.
See also PubMed.

Baron, A., and G.B. Kish. 1962. Low-intensity auditory and visual stimuli as reinforcers for the mouse. J. Comp. Physiol. Psychol. 55: 1011-1013.
See also PubMed.

Bauer, F.J. 1956. Genetic and experimental factors affecting social reaction in male mice. J. Comp. Physiol. Psychol. 49: 359-364.
See also MGI.

Beeman, E.A. 1947. The effect of male hormone on aggressive behavior in mice. Physiol. Zool. 20: 373-405.

Bell, R.W., and V.H. Denenberg. 1963. The interrelationships of shock and critical periods in infancy as they affect adult learning and activity. Anim. Behav. 11: 21-27.

Beniest-Noirot, E. 1958. Analyse du comportement dit maternal chez la souris. Cent. Nat. Rech. Sci. Monogr. Franc. Psychol. No. 1.

Beven, J.M., W. Bevan, and B.F. Williams. 1958. Spontaneous aggressiveness in young castrate C3H male mice treated with three dose levels of testosterone. Physiol. Zool. 31: 284-288.

Bevan, W., W.F. Daves, and G.W. Levy. 1960. The relation of castration, androgen therapy and pretest fighting experience to competitive aggression in male C57BL/10 mice. Anim. Behav. 8: 6-12.

Broadhurst, P.L., and J.L. Jinks. 1963. The inheritance of mammalian behavior re-examined. J. Hered. 54: 170-176.
See also PubMed.

Bronson, F.H. 1963. Density, subordination and social timidity in peromyscus and C57BL/10J mice. Anim. Behav. 11: 475-479.

Bronson, F.H., and B.E. Eleftheriou. 1963. Adrenal responses to crowding in Peromyscus and C57BL/10J mice. Physiol. Zool. 36: 161-166.

Bruell, J.H. 1962. Dominance and segregation in the inheritance of quantitative behavior in mice, p. 48-67. In E.L. Bliss [ed.] Roots of Behavior. Hoeber-Harper, New York.

Bruell, J.H. 1963. Emotional defecation in mice, a territory marking response. Amer. Psychol. 17: 445. (Abstr.)

Bruell, J.H. 1964a. Inheritance of behavioral and psychological characters of mice and the problem of heterosis. Amer. Zool. 4: 125-138.
See also MGI.

Bruell, J.H. 1964b. Heterotic inheritance of wheel-running in mice. J. Comp. Physiol. Psychol. 58: 159-163.
See also PubMed.

Bruell, J.H. 1965. Mode of inheritance of response time in mice. J. Comp. Physiol. Psychol. 60: 147-148.
See also PubMed.

Caldwell, D.F. 1962. Effects of adrenal demedulation on retention of a conditioned avoidance response in the mouse. J. Comp. Physiol. Psychol. 55: 1079-1081.
See also PubMed.

Carran, A.B., L.T. Yendall, and J. R. Royce. 1964. Voltage level and skin resistance in avoidance conditioning of inbred strains of mice. J. Comp. Physiol. Psychol. 58: 427-430.
See also PubMed.

Catlett, R.H. 1961. An evaluation of methods for measuring fighting behavior with special reference to Mus musculus. Anim. Behav. 9: 8-10.

Chitty, D., and H.N. Southern [ed.] 1954. Control of Rats and Mice. Vol. 3. House Mice. Oxford, London. 240 p.

Christian, J.J. 1955. Effect of population size on the adrenal glands and reproductive organs of male mice in populations of fixed size. Amer. J. Physiol. 182: 292-301.
See also PubMed.

Coleman, D.L. 1960. Phenylalanine hydroxylase activity in dilute and nondilute strains of mice. Arch. Biochem. Biophys. 91: 300-306.
See also PubMed.

Coleman, D.L., and K. Schlesinger. 1965. The effects of pyridoxine deficiency on audiogenic seizure susceptibility in inbred mice. Proc. Soc. Exp. Biol. Med. 119: 264-266.
See also PubMed.

Cooley, W.W., and P.R. Lohnes. 1962. Multivariate Procedures for the Behavioral Sciences. Wiley, New York. 211 p.

Cooper, R.M., and J.P. Zubeck. 1958. Effects of enriched and restricted early environments on the learning ability of bright and dull rats. Can. J. Psychol. 12: 159-164.
See also PubMed.

Davis, W.M. 1962. Day-night periodicity in pentobarbital response of mice and the influence of socio-psychological conditions. Experientia 18: 235-237.
See also PubMed.

Dawson, W.M. 1932. Inheritance of wildness and tameness in mice. Genetics 17: 296-326.
See also PubMed.

Denenberg, V.H. 1958. Effects of age and early experience upon conditioning in the C57BL/10 mouse. J. Psychol. 46: 211-226.

Denenberg, V.H. 1959a. Interactive effects of infantile and adult shock levels upon learning. Psychol. Rep. 5: 357-364.

Denenberg, V.H. 1959b. Learning differences in two separated lines of mice. Science 130: 451-452.
See also PubMed.

Denenberg, V.H. 1960. A test of the critical period hypothesis and a further study of the relationship between age and conditioning in the C57BL/10 mouse. J. Genet. Psychol. 97: 379-384.

Denenberg, V.H., and R.W. Bell. 1959. Relationships between social reactions and avoidance conditioning. Physiol. Zool. 32: 51-56.

Denenberg, V.H., and R.W. Bell. 1960. Critical periods for the effects of infantile experience on adult learning. Science 131: 227-228.
See also PubMed.

Denenberg, V.H., G.A. Hudgens, and M.X. Zarrow. 1964. Mice reared with rats: modification of behavior by early experience with another species. Science 143: 380-381.
See also PubMed.

Denenberg, V.H., J.A. King, and D.W. Ehrenfeld. 1958. Effects of testosterone propionate on the avoidance learning of male C57BL/6 mice. Physiol. Zool. 31: 244-247.

Denenberg, V.H., S. Ross, and M. Blumenfield. 1963. Behavioral differences between mutant and non-mutant mice. J. Comp. Physiol. Psychol. 56: 290-293.
See also PubMed.

Doner, R.D., C. Inman, and R.T. Davis. 1952. Competitive behavior between inbred strains of mice. Proc. S.D. Acad. Sci. 31: 172-176.

Ekström, N.A., and F. Sandberg. 1962. A method for quantitative determination of the inhibitory effect on the conditioned avoidance response in mice. Drug Res. 12: 1208-1209.

Elton, C. 1942. Voles, Mice, and Lemmings. Oxford, London. 496 p.

Essman, W.B., and H. Alpern. 1964. Single trial conditioning: methodology and results with mice. Psychol. Rep. 14: 731-740.

Essman, W.B., and M.E. Jarvik. 1961a. A water-escape test for mice. Psychol. Rep. 8: 58.

Essman, W.B., and M.E. Jarvik. 1961b. Impairment of retention for a conditioned response by ether anesthesia in mice. Psychopharmacologia 2: 172-176.
See also PubMed.

Fredericson, E. 1950. The effects of food deprivation upon competitive and spontaneous combat in C57 black mice. J. Psychol. 29: 89-100.
See also PubMed.

Fredericson, E. 1952. Aggressiveness in female mice. J. Comp. Physiol. Psychol. 45: 254-257.
See also PubMed.

Fredericson, E., and E.A. Birnbaum, 1954. Competitive fighting between mice with different hereditary backgrounds. J. Genet. Psychol. 85: 271-280.
See also PubMed.

Fredericson, E., C.D. Fink, and J.R. Parker. 1955. Elicitation and inhibition of competitive fighting in food deprived mice. J. Genet. Psychol. 86: 131-141.
See also PubMed.

Frings, H., and M. Frings. 1953. The production of stocks of albino mice with predictable susceptibilities to audiogenic seizures. Behavior 5: 305-319.
See also MGI.

Fuller, J.L. 1964. Measurements of alcohol preference in genetic experiments. J. Comp. Physiol. Psychol. 57: 85-88.
See also MGI.

Fuller, J.L., R.M. Chambers, and R.P. Fuller. 1956. Effects of cortisone and of adrenalectomy on activity and emotional behavior of mice. Psychosomat. Med. 18: 234-242.
See also PubMed.

Fuller, J.L., C. Easler, and M.E. Smith. 1950. Inheritence of audiogenic seizure susceptibility in the mouse. Genetics 35: 622-632.
See also PubMed.

Fuller, J.L., and G.A. Jacoby, Jr. 1955. Central and sensory control of food intake in genetically obese mice. Amer. J. Physiol. 183: 279-283.
See also PubMed.

Fuller, J.L., and W.R. Thompson. 1960. Behavior Genetics. Wiley, New York. 396 p.

Ginsburg, B., and W.C. Allee. 1942. Some effects of conditioning on social dominance and subordination in inbred strains of mice. Physiol. Zool. 15: 485-506.

Grant, E.C., and J.H. Mackintosh. 1963. A comparison of the social postures of some common laboratory rodents. Behavior 21: 246-259.

Green, E.L. 1964. Fitness of populations of irradiated mice: plan of experiments. Genetics 50: 417-421.
See also PubMed.

Halberg, F., J.J. Bittner, R.J. Gully, P.G. Albrecht, and E.L. Brackney. 1955. 24-hour periodicity and audiogenic convulsions in mice of various ages. Proc. Soc. Exp. Biol. Med. 88: 169-173.
See also PubMed.

Halberg, F., E. Jacobsen, G. Wadsworth, and J.J. Bittner. 1958. Audiogenic abnormality spectra, twenty-four hour periodicity, and lighting. Science 128: 657-658.
See also PubMed.

Halberg, R., R.E. Peterson, and R.H. Silber. 1959. Phase relations of 24-hour periodicities in blood corticosterone, mitoses in cortical adrenal parenchyma, and total body activity. Endocrinology 64: 222-230.

Hall, C.S., and P.H. Whiteman. 1951. The effects of infantile stimulation upon later emotional stability in the mouse. J. Comp. Physiol. Psychol. 44: 61-66.
See also PubMed.

Hall, M.J. 1959. The effects of early infantile experiences on maze learning in the mouse. Proc. W. Va. Acad. Sci. 31: 79-82.
See also PubMed.

Harner, R.N. 1961. Electrocorticography and frequency analysis in mice; circadian periodicity in electrocerebral activity. Electroenceph. Clin. Neurophysiol. 13: 752-761.

Hauschka, T.S. 1952a. Whisker-eating mice. J. Hered. 43: 77-80.
See also MGI.

Hauschka, T.S. 1952b. Mutilation patterns and hereditary cannibalism in mice. J. Hered. 43: 117-123.
See also MGI.

Hollifield, G., W. Parson, and K.R. Crispell. 1955. Studies of food drive and satiety in mice with gold thioglucose-induced obesity and the hereditary obesity-diabetes syndrome. Metabolism 4: 537-541.
See also PubMed.

Hoshishima, K., S. Yokoyama, and K. Seto. 1962. Taste sensitivity in various strains of mice. Amer. J. Physiol. 202: 1200-1204.
See also PubMed.

Huff, S.D., and J. L. Fuller. 1964. Audiogenic seizures, the dilute locus, and phenylalanine hydroxylase in DBA/1 mice. Science 144: 304-305.
See also PubMed.

Huff, S.D., and R.L. Huff. 1962. Dilute locus and audiogenic seizures in mice. Science 136: 318-319.
See also PubMed.

Hummel, K.P. 1960. Pituitary lesions in mice of the Marsh strains. Anat. Rec. 137: 366. (Abstr.)

Jacobs, B.B. 1962. Propylthiouracil taste inheritance in mice. J. Hered. 53: 183-186.

Janssen, P.A.J., A.H. Jageneau, and C.J.E. Niemegeers. 1960. Effects of various drugs on isolation-induced fighting behavior of male mice. J. Pharmacol. Exp. Therap. 129: 471-475.
See also PubMed.

Jarvik, M.E., and W.B. Essman. 1960. A simple one-trial learning situation for mice. Psychol. Rep. 6: 290.

Kahn, M.W. 1951. The effects of severe defeat at various age levels on the aggressive behavior of mice. J. Genet. Psychol. 79: 117-130.
See also PubMed.

Keeler, C.E. 1927. Rodless retina, an ophthalmic mutation in the house mouse, Mus musculus. J. Exp. Zool. 46: 355-407.

Keeler, C.E. 1928. Blind mice. J. Exp. Zool. 51: 495-508.

Kendall, M.G. 1961. Course in Multivariate Analysis, 2nd ed. Hafner, New York. 185 p.

King, J.A. 1956. Sexual behavior of C57BL/10 mice and its relation to early social experience. J. Genet. Psychol. 88: 223-229.
See also PubMed.

King, J.A. 1957a. Intra- and interspecific conflict of Mus and Peromyscus. Ecology 38: 355-357.

King, J.A. 1957b. Relationships between early social experience and adult aggressive behavior in inbred mice. J. Genet. Psychol. 90: 151-166.
See also PubMed.

King, J.A., and H. Cannon. 1955. Effects of social relationships upon mortality in C57BL/10 mice. Physiol. Zool. 28: 233-239.

King, J.A., and N.L. Gurney. 1954. Effect of early social experience on adult aggressive behavior in C57BL/10 mice. J. Comp. Physiol. Psychol. 47: 326-330.
See also PubMed.

King, J.A., and A. Mavromatis. 1956. The effect of a conflict situation on learning ability in two strains of inbred mice. J. Comp. Physiol. Psychol. 49: 465-468.
See also PubMed.

King, J.T., Y.C.P. Lee, and M.B. Visscher. 1955. Single versus multiple cage occupancy and convulsion frequency in C3H mice. Proc. Soc. Exp. Biol. Med. 88: 661-663.
See also PubMed.

Kish, G.B. 1955. Learning when the onset of illumination is used for reinforcing stimulus. J. Comp. Physiol. Psychol. 48: 261-264.
See also PubMed.

Kish, G.B., and A. Baron. 1962. Satiation of sensory reinforcement. J. Comp. Physiol. Psychol. 55: 1007-1010.
See also PubMed.

Lagerspetz, K. 1964. Studies on the aggressive behaviour of mice. Ann. Acad. Sci. Fenn. B 131: 1-13.

Lane, P.W., and M.M. Dickie. 1954. Fertile, obese male mice. J. Hered. 45: 56-58.

Leblond, C.P. 1940. Nervous and hormonal factors in the maternal behavior of the mouse. J. Genet. Psychol. 57: 327-344.

Leblond, C.P., and W.O. Nelson. 1937. Maternal behavior in hypophysectomized male and female mice. Amer. J. Physiol. 120: 167-172.

Les, E. 1958. A study of the effects of single locus heterozygosity on traits which may have survival value in eight stocks of laboratory mice. Ph.D. Diss. Ohio State Univ.

Levine, L. 1958. Studies on sexual selection in mice. I. Reproductive competition between albino and black-agouti males. Amer. Natur. 92: 21-26.

Levine, L. 1963. Factors affecting mating competition in mice, p. 249-250. In S.J. Geerts [ed.] Proc. XI Int. Congr. Genet. (The Hague, 1963) Vol. 1. Pergamon Press, New York (Abstr.)

Levine, S. 1959. Emotionality and aggressive behavior in the mouse as a function of infantile experience. J. Genet. Psychol. 94: 77-83.
See also PubMed.

Lieberman, M.W. 1963. Early developmental stress and later behavior. Science 141: 824-825.
See also PubMed.

Lindzey, G. 1951. Emotionality and audiogenic seizure susceptibility in five inbred strains of mice. J. Comp. Physiol. Psychol. 44: 389-393.
See also PubMed.

Lindzey, G., D.T. Lykken, and H.D. Winston. 1960. Infantile trauma, genetic factors, and adult temperament. J. Abnorm. Soc. Psychol. 61: 7-14.
See also PubMed.

Lindzey, G., and H. Winston. 1962. Maze learning and effects of pretraining in inbred strains of mice. J. Comp. Physiol. Psychol. 55: 748-752.
See also PubMed.

Lindzey, G., H. Winston, and M. Manosevitz. 1961. Social dominance in inbred mouse strains. Nature 191: 474-476.
See also PubMed.

Lipkow, J. 1960. Die Begattung bei der weissen Maus. Z. Tierpsychol. 17: 182-187.

Maren, T. 1952. A summary of preliminary observations on genetically and chemically obese mice, p. 12-13. Conference on the Obese Mouse. Roscoe B. Jackson Memorial Laboratory, Bar Harbor, Maine.

Martin, G.M., D. Manring, and E.P. Benditt. 1960. Experimental turpentine abscesses in aggressive and non-aggressive inbred strains of mice. Fed. Proc. 19: 20. (Abstr.)

Martin, P.G., and H.G. Andrewartha. 1962. Success in fighting of two varieties of mice. Amer. Natur. 96: 375-376.

Mast, T.M., and N.W. Heimstra. 1962. Prior social experience and amphetamine toxicity in mice. Psychol. Rep. 11: 809-812.

Mather, K. 1949. Biometrical Genetics. Dover, New York. 158 p.

McClearn, G.E. 1958. Performance differences among mouse strains in a learning situation. Amer. Psychol. 13: 405. (Abstr.)

McClearn, G.E. 1959. The genetics of mouse behavior in novel situations. J. Comp. Physiol. Psychol. 52: 62-67.
See also PubMed.

McClearn, G.E. 1960. Strain differences in activity of mice: influence of illumination. J. Comp. Physiol. Psychol. 53: 142-143.

McClearn, G.E. 1961. Genotype and mouse activity. J. Comp. Physiol. Psychol. 54: 674-676.

McClearn, G.E. 1965. Genotype and mouse behavior, p. 795-805. In S.J. Geerts [ed.] Proc. XI Int. Congr. Genet. (The Hague, 1963) Vol 3. Pergamon Press, New York.

McClearn, G.E., and D.A. Rodgers. 1959. Differences in alcohol preference among inbred strains of mice. Quart. J. Stud. Alc. 20: 691-695.

McClearn, G.E., and D.A. Rodgers. 1961. Genetic factors in alcohol preference of laboratory mice. J. Comp. Physiol. Psychol. 54: 116-119.

McGill, T.E. 1962. Sexual behavior in three inbred strains of mice. Behaviour 19: 341-350.

McGill, T.E., and W.C. Blight. 1963a. Effects of genotype on the recovery of sex drive in the male mouse. J. Comp. Physiol. Psychol. 56: 887-888.
See also PubMed.

McGill, T.E., and W.C. Blight. 1963b. The sexual behaviour of hybrid male mice compared with the sexual behavior of males of the inbred parent strains. Anim. Behav. 11: 480-483.

Meier, G.W. 1964. Differences in maze performances as a function of age and strain of house mice. J. Comp. Physiol. Psychol. 58: 418-422.
See also PubMed.

Meier, G.W., and D.P. Foshee. 1963. Genetics, age, and the variability of learning performances. J. Genet. Psychol. 102: 267-275.
See also PubMed.

Meier, G.W., and D.P. Foshee; H.D. Winston and G. Lindzey. 1965. Albinism and water escape performance in mice. Science 147: 307-308.
See also PubMed.

Mirone, L. 1958. The effect of ethyl alcohol on growth and voluntary consumption of alcohol by successive generations of mice. Quart. J. Stud. Alc. 19: 388-393.
See also PubMed.

Moltz, H. 1965. Contemporary instinct theory and the fixed action pattern. Psychol. Rev. 72: 27-47.
See also PubMed.

Mordkoff, A.M., and J.L. Fuller. 1959. Variability in activity within inbred and crossbred mice. J. Hered. 50: 6-8.

Morgan, H.S. 1963. The effects of food and darkness on learning in two inbred strains of mice. M. A. Thesis. Cornell Univ.

Pavlov, I.P. 1923. New researches on conditioned reflexes. Science 58: 359-361.

Ressler, R.H. 1962. Parental handling in two strains of mice reared by foster parents. Science 137: 129-130.
See also PubMed.

Ressler, R.H. 1963. Genotype-correlated parental influences in two strains of mice. J. Comp. Physiol. Psychol. 56: 882-886.
See also PubMed.

Rodgers, D.A., and G.E. McClearn. 1962a. Mouse strain differences in preference for various concentrations of alcohol. Quart. J. Stud. Alc. 23: 26-33.
See also PubMed.

Rodgers, D.A., and G.E. McClearn. 1962b. Alcohol preference of mice, p. 68-95. In E.L. Bliss [ed.] Roots of Behavior. Hoeber-Harper, New York.

Ross, S., and W. Smith. 1953. The hoarding behavior of the mouse. II. The role of deprivation, satiation, and stress. J. Genet. Psychol. 82: 299-307.
See also PubMed.

Royce, J.R., and M. Covington. 1960. Genetic differences in the avoidance conditioning of mice. J. Comp. Physiol. Psychol. 53: 197-200.

Runner, M.N., and A. Gates. 1954. Sterile, obese mothers. J. Hered. 45: 51-55.

Schlesinger, K. 1964. Genetic and biochemical determinants of alcohol preference and alcohol metabolism in mice. Ph.D. Diss. Univ. California, Berkeley. 98 p.

Schlesinger, K., and A.M. Mordkoff. 1963. Locomotor activity and oxygen consumption: variability in two inbred strains of mice and their F₁ hybrids. J. Hered. 54: 177.
See also PubMed.

Scott, J.P. 1944. An experimental test of the theory that social behavior determines social organization. Science 99: 42-43.
See also PubMed.

Scott, J.P. 1946. Incomplete adjustment caused by frustration of untrained fighting mice. J. Comp. Physiol. Psychol. 39: 379-390.

Scott, J.P. 1947. "Emotional" behavior of fighting mice caused by conflict between weak stimulatory and weak inhibitory training. J. Comp. Physiol. Psychol. 40: 275-282.

Scott, J.P., and E. Fredericson. 1951. The causes of fighting in mice and rats. Physiol. Zool. 24: 273-309.

Scriabine, A., and M. Blake. 1962. Evaluation of centrally acting drugs in mice with fighting behavior induced by isolation. Psychopharmacologia 3: 224-226.
See also PubMed.

Silverstein, E. 1961. Effect of hybridization on the primary polydipsic trait of an inbred strain of mice. Nature 191: 523.

Sirlin, J.L. 1956. Vacillans, a neurological mutant in the house mouse linked with brown. J. Genet. 54: 42-48.
See also MGI.

Smith, M.P., and S. Ross. 1960. Acceptance of sodium sucaryl by C57 black mice. J. Genet. Psychol. 96: 101-104.
See also PubMed.

Smith, R.P., and W. Bevan. 1957. Maze performance, emotionality, audiogenic seizure susceptibility in rats and mice treated with 2-dimethylaminoethanol. Proc. Soc. Exp. Biol. Med. 96: 382-385.
See also PubMed.

Smith, W., and E.K. Powell. 1955. The role of emotionality in hoarding. Behavior 8: 57-62.

Smith, W., and S. Ross. 1953a. The hoarding behavior of the mouse. I. The role of previous feeding experience. J. Genet. Psychol. 82: 279-297.
See also PubMed.

Smith, W., and S. Ross. 1953b. The hoarding behavior of the mouse. III. The storing of "non-relevant" material. J. Genet. Psychol. 82: 309-316.
See also PubMed.

Southwick, C.H. 1959. Eosinophil responses of C57BR mice to behavioral disturbance. Ecology 40: 156-157.
See also PubMed.

Southwick, C.H., and V.P. Bland. 1959. Effect of population density on adrenal glands and reproductive organs of CFW mice. Amer. J. Physiol. 197: 111-114.
See also PubMed.

Stanley, W.C., and J.A. Monkman. 1956. A test for specific and general behavioral effects of infantile stimulation with shock in the mouse. J. Abnorm. Soc. Psychol. 53: 19-22.
See also PubMed.

Tedeschi, R.E., D.H. Tedeschi, A. Mucha, L. Cook, P.A. Mattis, and E.J. Fellows. 1959. Effects of various centrally acting drugs on fighting behavior of mice. J. Pharmacol. Exp. Therap. 125: 28-34.
See also PubMed.

Thiessen, D.D. 1961. Mouse exploratory behavior and body weight. Psychol. Rec. 11: 299-304.
See also MGI.

Thiessen, D.D., and V.G. Nealey. 1962. Adrenocortical activity, stress response and behavioral reactivity of five inbred mouse strains. Endocrinology 71: 267-270.
See also PubMed.

Thiessen, D.D., J.F. Zolamn, and D.A. Rodgers. 1962. Relation between adrenal weight, brain cholinesterase activity, and hole-in-wall behavior of mice under different living conditions. J. Comp. Physiol. Psychol. 55: 186-190.
See also PubMed.

Thompson, W.R. 1953. The inheritance of behavior: behavioral differences in fifteen mouse strains. Can. J. Psychol. 7: 145-155.
See also MGI.

Thompson, W.R. 1956. The inheritance of behavior; activity differences in five inbred mouse strains. J. Hered. 47: 147-148.

Tollman, J., and J.A. King. 1956. The effects of testosterone propionate on aggression in male and female C57BL/10 mice. Brit. J. Anim. Behav. 4: 147-149.

Uhrich, J. 1938. The social hierarchy in albino mice. J. Comp. Psychol. 25: 373-413.

Van Abeelen, J.H.F. 1963a. Mouse mutants studied by means of ethological methods. I. Ethogram. Genetica 34: 79-94.

Van Abeelen, J.H.F. 1963b. Mouse mutants studied by means of ethological methods. II. Mutants and methods. Genetica 34: 95-101.

Van Abeelen, J.H.F. 1963c. Mouse mutants studied by means of ethological methods. III. Results with yellow, pink-eyed dilution, brown and jerker. Genetica 34: 270-286.

Vandenbergh, J.G. 1960. Eosinophil response to aggressive behavior in CFW albino mice. Anim. Behav. 8: 13-18.

Vanderpool, D.L., and R.T. Davis. 1962. Differences in spontaneous behavior among inbred strains of mice. Psychol. Rep. 10: 123-130.

Vicari, E. 1923. Hybridization and behavior. Eugen. Race State 2: 75-77.

Vicari, E. 1924. Non-inheritance of the effects of training. Science 59: 303.
See also PubMed.

Vicari, E.M. 1929. Mode of inheritance of reaction time and degrees of learning in mice. J. Exp. Zool. 54: 31-88.

Waller, M.B., P.F. Waller, and L.A. Brewster. 1960. A water maze for use in studies of drive and learning. Psychol. Rep. 7: 99-102.

Warren, R.P. 1963. Preference aversion in mice to bitter substance. Science 140: 808-809.
See also PubMed.

Weir, M.W., and J.C. De Fries. 1964. Prenatal maternal influence on behavior in mice: evidence of a genetic basis. J. Comp. Physiol. Psychol. 58: 412-417.
See also PubMed.

Willingham, W.W. 1956. The organization of emotional behavior in mice. J. Comp. Physiol. Psychol. 49: 345-348.
See also PubMed.

Wimer, R.E., and J.L. Fuller. 1965. The effects of d-amphetamine sulphate on three exploratory behaviors. Can. J. Psychol. 19: 94-103.
See also PubMed.

Wimer, R.E., and H. Sterns. 1964. Controlled visual input and exploratory activity in C57BL/6J mice. Percept. Mot. Skills 18: 299-307.

Wimer, R., and S. Weller. 1965. Evaluation of a visual discrimination task for the analysis of the genetics of a mouse behavior. Percept. Mot. Skills 20: 203-208.

Winston, H.D. 1963. Influence of genotype and infantile trauma on adult learning in the mouse. J. Comp. Physiol. Psychol. 56: 630-635.
See also PubMed.

Winston, H.D., and G. Lindzey. 1964. Albinism and water escape performance in the mouse. Science 144: 189-191.
See also PubMed.

Wolf, E. 1930. Die Aktivität der japanischen Tanzmaus und ihre rhythmische Verteilung. Z. Vergl. Physiol. 11: 321-344.

Yerkes, R.M. 1907. The Dancing Mouse. Macmillan, New York. 290 p.