NEST DEFENSE AND SURVIVAL OF OFFSPRING IN HIGHLY AGGRESSIVE WILD CANADIAN FEMALE HOUSE MICE

Nest defense behavior was examined in wild female house mice (Mus dome sticus) that were derived from a stock initially trapped in Alberta, C anada. The first objective was to determine whether behavior toward pu ps prior to mating was related to the intensity of postpartum aggressi on in a variety of social situations. Therefore, prior to the experime nts we screened virgin females for their behavior toward a newborn pup [60% of the females exhibited infanticide and 40% were noninfanticida l: 7% were parental (retrieved and hovered over the pup) and 33% ignor ed the pup]. Infanticidal and noninfanticidal females were then mated with males and used in four experiments. In Experiment 1 the females w ere housed individually prior to delivery, while in Experiment 2 the f emales were allowed to remain with their mates; in both situations all females successfully reared litters of similar sizes. Male and female intruders (that had all exhibited infanticide when previously tested with a pup) were placed separately into a test cage containing a lacta ting female during the first four days after delivery. Regardless of t he presence of the stud male, previously infanticidal females were mor e aggressive (exhibited more attacks per min) toward both male and fem ale intruders than were previously noninfanticidal females; infanticid al females also exhibited more of both forms of attack (offensive and defensive) and also attacked with greater intensity than did noninfant icidal females. The number of attacks toward intruders of both sexes i ncreased for both infanticidal and noninfanticidal females between Day 1-4 postpartum, but very high rates of attack were observed on all da ys by the lactating females, including the day of delivery. In Experim ents 3 and 4 only the most aggressive (previously infanticidal) female s were tested. In Experiment 3, two unrelated, unfamiliar females were mated separately and then were housed together just prior to delivery , which was planned to occur 3-4 days apart. In 5 of the 15 cages, all pups disappeared on the day of delivery of the second female to deliv er her litter. In the remaining 10 cages, it appeared that none of the pups produced by the 20 females were killed. Thus, in this experiment , 66% of pups survived to Day 4 postpartum. In Experiment 4, two previ ously infanticidal female siblings, which had been housed together sin ce birth, were placed together with a stud male. In all 9 cages only o ne female became pregnant and delivered pups, but only 3 litters survi ved to Day 4 (no litters were observed being attacked during intruder tests). In contrast to kin-selection theory and our expectation that a sibling would contribute to communal rearing and nest defense, housin g siblings together thus resulted in the lowest overall reproductive s uccess (3/18; 17% of the females produced litters that survived to Day 4 postpartum) of any of the social conditions examined. We discuss th e implications for social structure and population dynamics of the ext remely high aggressiveness and infanticidal tendency, regardless of ki nship, of females from this stock of wild mice.