Maintenance and Decline of the Suppression
of Infanticide in Mother Rats
L. C. PETERS,1 T. C. SIST AND M. B. KRISTAL2
Department of Psychology, State University of New York
at Buffalo, Buffalo, NY 14260
Received 16 November 1990
Maternal behavior Infanticide Filicide Rats Delivery Lactation Cannibalism
THE behavior of maternal rats includes pup-directed caretaking behaviors such as retrieving of young to the nest, cleaning the young, and a specific hovering posture that warms the pups and allows them to nurse (25,26). In addition to caring for their young actively, it is essential that mother rats not harm them. Although virgin rats generally kill freshly delivered young that are still covered with fetal fluids and membranes, mothers given such foster young very seldom kill (16, 21, 27). This indicates that infanticide is somehow suppressed by the time rats become mothers. Like caretaking behavior, the suppression of infanticidal responses first appears within a day or so before delivery in primigravid rats. However, some rats that show a suppression of their infanticidal tendencies when tested before delivery do not actively care for foster pups. Other near-term pregnant rats do actively care for older, cleaned foster pups, but will kill any freshly delivered young they are given (21).
It is not clear how long the suppression of infanticidal tendencies lasts. In laboratory tests of Long-Evans rats, it generally seems that only hours-old neonates, uncleaned of fetal fluids and membranes, are killed by virgin females (21). Is the suppression of infanticidal tendencies maintained only for a day or so after birth? Alternatively, does it ever decline? Various patterns have been seen in other rodent species. A male gerbil kills any young it is given until some time during its mate's first pregnancy, and thenceforth never kills again, whereas female gerbils refrain from killing foster young only when they have young of their own (5). In at least some environments, female hamsters kill a proportion of their own litters, even after investing several days of care in them (3). In the present experiments, we a) studied the duration of the suppression of infanticide in mother rats; b) investigated the mechanisms by which it is suppressed; and c) explored the degree to which that suppression is independent of caretaking behavior.
The details of the method described here briefly can be found in Peters and Kristal (21).
Subjects were 170 adult female Long-Evans hooded rats kept on a 14-10 h light-dark schedule, with lights on at 0600 (EST). No rats other than those indicated in Experiment 2 were tested more than once. Prior to testing, all subjects were screened for estrous cyclicity and for spontaneous attraction to placenta.
Tests were conducted in clear, wire-topped plastic cages, 21 x 42 x 20 cm. Minimal bedding was used so that stimulus pups could not be buried or dried out by it.
Long-Evans pups were used for tests of infanticide and maternal responsiveness. Those used in infanticide tests were obtained 3 h before use by cesarean section (hereafter referred to as "cesarean pups" or neonates) of donor rats on Day 21 of gestation. Fetal fluids and membranes were allowed to remain on the pups and were reconstituted with saline before testing. Placentas were removed, however, so that placentophages and nonplacentophages would not react differently to the stimulus pups (12,13). In an effort to exclude moribund cesarean pups, we monitored the amplitude of their vocalizations on an oscilloscope attached to a Holgate Mk IV ultrasonic receiver. Only those pups were used that had responded with a minimum deflection in response to 10 s of prodding with a cotton swab.
Before rats were tested for infanticide, there was a 5-h period of habituation to the test cage beginning at 1200 h. At 1700 h, the 60-mm infanticide test was begun by placing a single cesarean pup (to keep the number of pups exposed to harm to a minimum) into the end of the test chamber opposite the corner where the subject slept or lay. Sixty mm after the stimulus pups were presented, the experimenter returned and determined whether each subject was a pup killer or a nonkiller.
EXPERIMENT 1: THE SUPPRESSION OF INFANTICIDE DURING LACTATION AND AFTER WEANING
The first experiment was designed to determine how long after delivery mother rats continue not to kill neonates. We tested a group of rats in the second week of lactation [when suppression toward strange conspecifics is most intense (6)], a group that was near the end of lactation, and a group whose young had been weaned two weeks earlier. The behavior of these groups was compared to that of large groups of virgins and newly delivered mothers that also figured in a previous study (21), although, in many cases, those rats were tested at the same time and in the same place as the rats first discussed here.
Group L9 consisted of 17 rats tested on Days 8, 9, 10 or 11 of lactation (Day 8, n = l; Day 9, n = l0; Day 10, n = 4; Day 11, n = 2). Group L20 was 10 rats tested on Days 19, 20 or 21 of lactation (Day 19, n = 4; Day 20, n = 4; Day 21, n = 2). Group NLP was 10 nonlactating primiparae tested 14 days after they had been separated from their 25- or 26-day-old litters and had resumed normal estrous cyclicity as determined by vaginal smears. Group Virgin was 52 virgins, and Group LI was 20 mothers tested on the first full day of lactation. All lactating subjects were primiparous.
Subjects were tested for infanticide in the manner described above. During the 5-h period of habituation to their test cages, the lactating rats had no contact with pups. Ten minutes after the end of the infanticide test, most of the subjects in Groups L9, L20 and NLP were tested for 60 min for caretaking responses (retrieving, licking, crouching) toward 3 foster pups, 3-10 days old, placed in the end of the test cage opposite the sleeping corner. In this way, it was possible to see if a rat was maternally responsive even if it had killed a cesarean pup.
RESULTS AND DISCUSSION
Proportions of groups were compared using the standard error of the difference between two proportions (4) because of the combination of large n's and low expected cell frequencies.
Compared to virgins, significantly smaller proportions of primiparous mothers killed cesarean pups not only on the first day after delivery (Group Virgin: 50 of 52 killed; Group L1: 2 of 20 killed; s p1-p2 = 0.12; z = 7.3, p < 0.001), but also in the second week of lactation (Group L9: 4 of 17 killed; s p1-p2 = 0.12; z = 6.3, p < 0.001), near the end of lactation (Group L20: 7 of 10 killed, s p1-p2 = 0 .09; z = 2.78, p = 0.003), and even two weeks after weaning (Group NLP: 8 of 10 killed; s p1-p2 = 0.08; z = 1.90, p = 0.029). In Fig. 1, it can also be seen that, although the suppression of infanticidal tendencies was maintained in some rats through lactation and beyond, larger proportions killed as the interval after delivery increased.FIG. 1. Incidence of infanticide (toward newly cesarean-delivered pups) and maternal behavior (toward normally delivered 3-day-old pups) among groups of lactating primiparous rats. Values for pregnant rats can be obtained from Peters and Kristal (21). (*Significantly different from virgins, p < 0.05.)
Maternal responsiveness to older pups followed a similar pat tern, as shown in Fig. 1. All 12 Group L9 rats given older pups responded maternally to them. Nine of the 10 Group L20 rats given older (3-10 days old) pups responded maternally, even though their own pups were much older (19-21 days old). None of the 8 Group NLP rats given older pups responded maternally to them.
It appears that the suppression of infanticide is, in fact, maintained, as caretaking is maintained, well beyond the first day or so when the hormonal conditions are present that cause rapid induction of caretaking (30). Infanticide is first suppressed prepartum, presumably due to hormonal or other physiological changes of pregnancy, since the females in which this was seen had neither undergone delivery nor had received previous exposure to young (21). That the suppression of infanticide is maintained well into the lactation period indicates one of two things:
(a) If suppression of infanticide is maintained by the same mechanism
that causes the onset of the suppression of infanticide prepartum, then
the mechanism must be different from the short-lived physiological basis
for the onset of caretaking behavior. The suppression would be based on
a physiological state that is present both shortly prepartum and through
much of lactation, e.g., the secretion of oxytocin (14,15) or prolactin;
or (b) if the suppression of infanticide is maintained through lactation
by a mechanism different from that responsible for its onset prepartum,
then it might be maintained either by physiological conditions specific
to the lactation period or else by nonphysiologic variables, such as experience
with young. Experience with young seems to be the principal basis for the
maintenance of caretaking behavior through lactation (26). In the following
experiments, we examined the effect of previous contact with young on rats'
infanticidal tendencies in an attempt to understand the factors that maintain
the suppression of infanticide through at least the first two weeks of
lactation and why infanticidal tendencies reappear as the time for weaning
Twenty-six pup-naive virgin rats were tested. Group NO- TOUCH rats (n = 10) were exposed to cesarean pups presented in protective wire cubes. Group TOUCH rats (n = 11) were ex posed to unprotected cesarean pups. A Control Group (n = 5) consisted of subjects exposed to empty cubes.
Group TOUCH rats were individually housed and tested in 10-gallon glass tanks. Group NO-TOUCH and Control rats were housed during the 7-day exposure period in galvanized wire animal cages, 24.5 x 18 x 18 cm. All rats had lab chow and water available ad lib. Small cages were used to decrease the likelihood that the subjects could avoid the protected pups, since Terkel and Rosenblatt (29) demonstrated that virgins become maternal more rapidly when they are housed with pups in small cages.
The cesarean pups were protected from adults by two layers of screen. The inner layer consisted of a pouch made from brass mesh (2 mm square) that was large enough to allow the enclosed pup some movement. The outer layer was a 7-cm cube made from strong galvanized mesh (1 cm square) that opened on one side so the inner sleeve could be removed. Herrenkohl and Lisk (8) have shown that virgins can be sensitized, although more slowly, with only the exteroceptive cues from pups behind a wire screen.
At 1200 h on the first day of testing, the subjects exposed to unprotected pups (Group TOUCH) were moved into the glass tanks. Minimal bedding was provided, making it difficult to bury pups or for the amniotic fluid on the pups to be dried off. One 60-min infanticide test was conducted at 1700 h on each of six consecutive days.
The day before exposure to neonates began, NO-TOUCH and Control subjects were moved to their small test cages and presented with empty wire cubes. The following day, at about 1700 h, the cubes were removed and those of the NO-TOUCH Group were replaced with cubes containing cesarean pups prepared in the usual way, except that the membranes covering them had not been moistened (1 pup/cube/subject). The Control subjects, which were not housed in the same room as the experimental group, had their cubes returned empty daily. This procedure was repeated with each member of Group NO-TOUCH receiving a fresh cesarean pup daily until it had been exposed to pups for nearly 7 full days. At 1200 h on the day after the seventh presentation, both the experimental and control groups were moved from their metal cages to plastic test cages, and at 1700 h were given the infanticide test. At the end of the test, subjects were given a 60-mm test for maternal responsiveness to 3- to 10-day-old pups. Afterward, vaginal smears were obtained to determine the subjects' ovarian cycle stages.
RESULTS AND DISCUSSION
All 11 subjects in Group TOUCH killed pups on every one of the six test days. There seems to have been a trend for virgins to approach and kill cesarean pups with shorter latencies as they acquired more experience with them, although the particular data that would demonstrate this conclusively were not systematically recorded. Apparently, if repeated exposure to neonates has any suppressing effect on infanticide, it is overridden by some effect of killing experience that promotes further killing. Therefore, the data obtained from the NO-TOUCH Group were important in order to ascertain if the development of suppression of infanticidal tendencies was possible while avoiding any effects of killing experience.
Even after prolonged exposure to protected pups, Group NO- TOUCH did not differ from the Control Group; all 15 rats killed cesarean pups. Moreover, none of the rats was maternal to older pups.
The amount or type of pup contact in this experiment may not have been
sufficient to produce a suppression of infanticidal tendencies in experimental
subjects. In particular, the exteroceptive stimuli to which Group NO-TOUCH
was exposed may not have been adequate. In the subsequent experiment, attempts
were made to intensify the contact that virgins had with young. In this
way, we hoped to make a fair test of whether experience with newborn pups
can be sufficient to cause the infanticidal tendencies of virgins to be
suppressed to an extent comparable to that of mothers in the second week
Thirty-five sensitized (maternal) virgin rats were used. Group NOPUP consisted of 10 subjects tested for infanticide without older pups present after a 5-h habituation period in which no pups were present. Group PUP consisted of 10 subjects tested with 3- to 10-day old stimulus pups present during both the 5-h habituation period and the infanticide test. The older pups were allowed to remain with the subjects because the hormonal condition of sensitized virgins has been shown to vary with the presence or absence of pups (11). Group CUBE consisted of 10 subjects exposed to empty cubes. The Control Group consisted of 5 subjects exposed to empty cubes. Group CUBE and the Control Group were also tested with the older stimulus pups still present during test-cage habituation and the infanticide test.
Subjects had been induced to behave maternally (sensitized) by presenting each with a stimulus litter of three or four pups 3 to 10 days old. After each subject spent 24 h with a stimulus litter, the litter was returned to its mother and replaced with a fresh one. Sensitization was judged to have occurred when a subject retrieved, licked, and hovered over its pups within 15 min after they were removed from the nest and scattered about the cage. After the rat had met these criteria on 2 consecutive days, it was considered eligible for inclusion in the experiment. Assignments to groups were made such that subjects in the pairs NOPUP/PUP and CUBE/Control each had taken about the same number of days to sensitize.
Wire cubes (each containing either a fresh cesarean pup or no pup) were hooked tightly under the wire tops of the plastic cages as in Experiment 2. They were removed daily at around 1700 h. At about the same time, each subject's group of three older stimulus pups was removed and replaced with another that had been fed. At 1130 h on the fourth through seventh days of exposure to cesarean pups, each subject's older pups were re moved from the nest and scattered at the opposite end of the cage.
At 1200 h on the seventh day of exposure to cubes, all subjects were given the infanticide test in fresh plastic cages. Three fresh 3- to 10-day-old pups, without cubes, were also placed into each PUP, CUBE and Control test cage. At the end of the infanticide test, noninfanticides were observed for 5 mm to discover whether they were maternal to their (unprotected) cesarean pups. The attractiveness of cesarean pups relative to older pups for each noninfanticidal subject was then judged by scattering all four pups at the end of the cage opposite the nest and observing the latency to respond maternally to each pup. At the completion of testing, vaginal smears were obtained from all subjects.
Contact with older pups that was more than sufficient to induce maternal behavior in virgins did not significantly reduce the proportion of subjects killing neonates. Despite all the experience they received with older pups, all 10 rats in Group NOPUP and 7 of 10 in Group PUP killed neonates (p = 0.105, Fisher exact probability).
Infanticide was not inconsistent with maternal responsiveness, as at least 6 of 7 of the Group PUP infanticides continued to treat the older pups maternally during the infanticide test, whereas 2 of 3 noninfanticides were maternal toward both kinds of pup (the remaining rat did not respond maternally to either). No attempts were made to determine if Group NOPUP rats were maternal during or after the test with cesarean pups, although all of them were maternal on the morning before the test.
PERCENTAGES OF INFANTICIDES AMONG GROUPS OF VIRGIN FEMALE RATS EXPOSED
|*Older pups remained
with subject during the test for infanticide of neonates.
†Significantly different from Control, p<0.05.
After 7 days of exposure to pups protected by cubes, 7 of 10 Group CUBE subjects killed cesarean pups. That infanticide was suppressed in three subjects does not seem to be due to the 7 days of contact with protected neonates, however, since only 2 of 5 Control rats killed (Table 1). The proportions of infanticides in the CUBE (70%) and Control (40%) Groups did not differ statistically (p = 0.29,. Fisher exact probability).
All rats in both the CUBE. and Control Groups continued to respond maternally toward older pups, but there was surprisingly little maternal responsiveness toward cesarean pups. All 9 noninfanticidal sensitized virgins (3 from Group PUP, 3 from Group CUBE, and 3 Controls) had been continuously housed with older pups during their infanticide tests. By the end of those tests, most of the cesarean pups had been retrieved and all had been at least partly cleaned. Only in 2 of 9 cages, however, were neonates in a nest with the three older pups. The differences in the treatment of older pups and neonates were best seen in the retrieval test in which maternal responses were noted during the 15 mm after the older pups and cesarean pups (by then clean and dry) were scattered about the end of the cage opposite the nest. Eight of the 9 noninfanticidal virgins retrieved all three older pups to their nests, then licked and crouched over them, but only one retrieved its cesarean pup to the nest (p = 0.0017, Fisher exact probability). Two other noninfanticides retrieved cesarean pups to sites other than their nests, licked them briefly, and then returned without them to crouch over the older pups. None of the three that retrieved its cesarean pup did so before all older pups. had been retrieved and licked. It is interesting that, although neonates seemed to be inferior stimuli for maternal care in these sensitized virgins, the virgins nevertheless did not kill them.
Of maternal virgins tested for infanticide while still in the presence of older pups (Group PUP, Group CUBE and Controls), only 16 of 25 (64%) killed cesarean pups. This proportion was clearly lower than that of virgins in Experiment 1 that had not been given periods of prolonged exposure to pups (50 of 52, or 96%) (s p1-p2 = 0.08; z= 3.77, p < 0.0002).
The proportion of infanticides among maternal virgins with pups present, Groups PUP, CUBE and Control (64%), was significantly higher than that found among Group L9 lactating mothers (24%) in Experiment 1 (s p1-p2 = 0.16: z = 2.58; p = 0.005).
Although only 16 of 25 (64%) maternal virgins that had been allowed to remain with their pups killed cesarean pups, all 10 Group NOPUP maternal virgins killed cesarean pups after they had been separated from their pups for 5 h (s p1-p2 = 0.16; z = 2.20; p = 0.014).
There was no indication that the rats that killed tended to be in a
different stage of the estrous cycle than those that did not kill.
We found in Experiment 1 that most mother rats still did not kill foster cesarean pups in the second week of lactation. Experiment 2 demonstrated that infanticide was not suppressed if killers were repeatedly exposed to unprotected, freshly delivered pups. There also seems to have been no inhibition of infanticide brought. about by prolonged exposure of virgins to neonates housed in wire cubes. This was so when this was the only experimental treatment (Experiment 2) and also when it was combined with the induction of maternal behavior and continued exposure to older pups (Experiment 3). In Experiment 3, we found that more intense exposure to older pups did result in a significant reduction in the proportion of virgin infanticides (from 100% to 64%), provided that older pups were present during both the habituation and test periods.
It is still necessary to conclude that something besides pup contact contributes to suppression of infanticide in mother rats. The noninfanticidal virgins differed from noninfanticidal lactating mothers. All infanticide tests given to mothers followed a 5-h period in which the mothers had no contact with pups. Despite the separation, most mothers did not kill foster neonates, whether they were in the first day of lactation (L1, when 10% killed) or in the second week of lactation (L9, when 24% killed). The suppression of infanticide in maternal virgins, however, in no case endured a 5-h separation from the pups that produced it. Virgin and mother noninfanticides differ in that only the latter are capable of withholding killing responses after a 5-h separation from pups.
Some conclusions can be drawn about why mother rats continue to be noninfanticides through much of their lactation period. First, contact with pups does appear to play a role. The proportion of maternal virgins that did not kill stimulus pups in the third experiment (9 of 25, or 36%) might have been even greater if the subjects had been sensitized with pups that were younger, and hence more like the neonates used in the infanticide test. It is unclear whether it is necessary for virgins to be maternally responsive for pup contact to turn them into noninfanticides. In any case, it seems likely that the pup contact that mothers receive while being maternal is involved in maintaining the suppression of infanticide -- infanticide that is first suppressed prepartum by a physiological change, presumably hormonal, associated with pregnancy (21).
Second, because mothers are affected little or not at all by a 5-h separation, whereas infanticide-suppressed virgins are, other factors may be involved besides pup contact. Mother rats differ from virgins in both a) the physiological modifications they have undergone as a consequence of lactation, and b) their history of having experienced parturition, including exposure to neonates during a period when infanticide was suppressed. Either or both of these differences could be responsible for the behavioral differences of mothers and sensitized virgins.
If physiological aspects of lactation are involved in suppressing infanticide, they may be the same physiological factors that appear to be responsible for initiating the suppression of infanticide prepartum. At least two hormones, prolactin and oxytocin, that are necessary for lactation and that are secreted in response to suckling and other pup stimuli, are also secreted in the hours immediately before parturition (7,17). It may be relevant that prolactin, at least, is also present at greater concentrations in virgins that have been sensitized than in those that have not been exposed to pups long enough to become sensitized (11,28). However, the increase in prolactin levels that is seen in sensitized virgins after reunions with their pups is only a fraction of the increase seen in lactating mothers. This opens up the possibility that sensitized virgins are not as resolutely noninfanticidal as are mothers with pups of the same age because the latter have a more highly developed neuroendocrine response to young (14,15).
It is also possible that mothers acquire a habit of suppressing infanticide during the onset phase of the suppression. Unlike the sensitized virgins in Experiment 3, mothers in the second week of lactation (L9) in Experiment 1 had the experience of not killing neonates before they underwent infanticide tests. In the period shortly after their own parturitions, they may have learned to respond to neonates in ways other than killing, and these habits may have influenced their responses to the cesarean pups with which they were subsequently tested. A similar long-term strengthening of maternal caretaking by pup contact specifically during the first few hours after delivery has been demonstrated by Bridges (1,2). Perhaps the nonkilling by two of ten rats in Group NLP (the nonlactating primiparae of Experiment 1) was an expression of a long-term change in behavior not based in hormonal conditions or recent exposure to young.
2 Requests for reprints should be addressed
to Dr. M. B. Kristal, Department of Psychology, SUNY at Buffalo, Buffalo,