SHORT COMMUNICATIONS 1253

rover' and the E2W3, 'sitter' stocks. This result indicates that the third pair of chromosomes as well as the second pair influences 'digging' behaviour.

The numbers of vials sown for each stock is shown in Table L Since each vial was sown with I0 eggs, stock 'survivorship' estimates (a composite of egg hatchability and larval survivorship) could be calculated. In all cases, stocks sharing the W3 chromosomes showed significantly higher 'survivorship' (P < 0.008) than stocks with the E3 chromosomes.

Godoy-Herrera suggested that inter-strain differences in digging patterns provide evidence for the genetic control of digging behaviour. Differences in the digging behaviour of tile four stocks tested in this study were con- sistent with his interpretation. The tendency for 'rover' larvae to perform more digging and crawling behaviour than 'sitter' larvae reflects some joint genetic influences in crawling and digging behaviours that are to some extent caused by differences in the second pair of chromo- somes. However, in contrast to crawling behaviour where major differences in this behaviour could be attributed to the second pair of chromosomes, the genetic control of digging behaviour is also affected by the third pair of chromosomes.

In a series of natural population studies, Sokolowski (1980, in press) showed that both 'rover' and 'sitter' forager types were found within a single pear. If natural popu- lations of 'rovers' also tend to dig more than 'sitters,' then greater dispersal (both horizontal and downward) would be expected in larvae of the 'rover' morph. Because the adult female D. melanogaster lays her eggs on or close to the surface of the medium, the amount the larvae dig into the medium measures the microdispersal. The ability of a larva to utilize and compete for a distant food resource may depend on the amount of digging and crawling behaviour it exhibits.

MARLA B. SOKOLOV~rSKI Department of Biology, York University, Downsview, Ontario, Canada M3J IP3.

References

Barker, J. S. F. 1971. Ecological differences and com- petitive interaction between Drosophila melano- gaster and Drosophila simulans in small laboratory populations. Oecologia, 8, 139-156.

Godoy-Herrera, R. 1977. Inter- and intra-populational variation in digging in Drosophila melanogaster larvae. Behav. Genet., 7, 433-439.

Godoy-Herrera, R. 1978. Selection for digging behaviour in Drosophila melanogaster larvae. Behav. Genet., 8, 475-479.

Hirsch, J. & Ksander, G. 1969. Studies in experimental behaviour genetics: V. Negative geotaxis and further chromosome analyses in Drosophila melanogaster. J. eomp. physiol. Psychol., 67, 118-122.

Parsons, P. A. 1975. The comparative evolutionary biology of the sibling species Drosophila melano- gaster and Drosophila simulans. Q. Rev. Biol., 50, 151-169.

Sewell, D., Burnet, B. & Connolly, K. 1975. Genetic analysis of larval feeding behaviour in Drosophila melanogaster. Genet. Res. Cam., 24, 163-173.

Sokal, R. R. & Roll, R. 1969. Biometry. San Francisco: W. Freeman.

Sokolowski, M. B. 1980. Foraging strategies of Droso- phila melanogaster. Beh~v. Genet., 10, 291-302.

Sokolowski, M. B. In press. 'Rover ' and 'sitter" larval foraging patterns in a natural population of Drosophila melanogaster. Drosoph. Inf. Serv.,

(Received 6 November 1981; revised 1 June 1982; MS. number: AS-161)

Mate Takeover and Possible Infanticide by a Female Northern Jaeana (Jacana spinosa)

Birds in the pan-tropical family Jacanidae breed on floating vegetation in marshes. Female jacanas employ a suite of tactics that make their reproductive strategy unique among female vertebrates: attainment of larger body size than maIes, defence of limited breeding habitat, simultaneous reproduction with up to four mates, and lower investment in offspring care than males (Jenni 1974; Jenni & Betts 1978; Ridley 1978). Observations that I made during a study of northern jacanas (Jacana spinosa) in Costa Riea reveal that the reproductive strategy of female jacanas sometimes involves aggressive takeovers of mates from neighbouring territorial females and suggest that females may destroy the current offspring of mates acquired by takeovers, as detailed below.

At 0605 hours on 25 June 1981, a bigamous territorial female, here designated F1, landed within the territory of a bigamous neighbouring female (F2), who flew to the intruder and attacked her. The females beat each other with their wings and pecked with their bills. After two minutes of fighting, F2 retreated under water and swam away. F2 had fought with a third female in an adjacent territory just before this fight and perhaps had been weakened enough to enable F1 to defeat her.

One of F2's mates, male M2a, stood near FI and F2 as they fought. He attacked F1 when his mate disappeared but failed to evict her. F1 only occasionally retaliated against M2a's attacks by attempting to peck him. In contrast to M2a's behaviour, F1 invited copulation from him and made r~est-construction movements.

M2a had a nest with a full clutch of eggs. He increased the intensity of his attack as F1 wandered nearer his nest, but this had no effect. He then performed a distraction display: he crouched and drooped his wings (as if brood- ing), and slowly rocked the wings. This behaviour is easily distinguished from other displays of jacanas. M2a gave two more distraction displays to F1 during the two hours of observation; both occurred while F1 was near his nest. F1 appeared to ignore the display in all three instances.

The defeated female had one other mate (M2b), whose young offspring were concealed in emergent vegetation after the fight between F1 and F2. FI encountered M2b during her inspection of F2's territory. M2b attacked F1 when she approached the vicinity of his offspring. His vocalizations apparently summoned his defeated mate, who reappeared and fought F1 again, only to retreat under water once aga:n. M2b continued the attack without his mate but F1 retaliated by chasing him from his territory. He returned shortly thereafter, however, and was still present when observations were terminated at 0800.

F1 was at M2a's nest when observations began the following day at 0900, yet M2a was no longer hostile toward her. A check of the nest at 1100 revealed it to be empty. The pair courted on this day and F1 began laying

1254 A N I M A L B E t I A V I O U R , 3 0 , 4

. eggs for this male eight days after she evicted his mate. F2, M2b, and their brood moved into an unoccupied area adjacent to their former territory. �9 These observations document one way that female

jaeanas acquire additional mates: aggressive takeovers of males from neighbouring females. This behaviour may account for several other instances of females gaining additional mates and territory during my study, but I did not witness how these gains occurred. Non-territorial female jacanas are also known to evict female residents and acquire their mates (Jenni & Collier 1972). A mate takeover by a polyandrous female has also been reported in the spotted sandpiper , Act#is maeularia (Oring & Maxson 1978)..

The fact that M2a's clutch disappeared within one day after F l ' s takeover suggests that F1 removed his eggs

f r o m the nest. The distraction displays that M2a per- formed while F1 was at his nest imply that he responded to F1 as if she were a potential destroyer of his eggs. Jenni (personal communication) saw female jacanas de- stroy eggs of new mates after mate takeovers. This provides evidence of a phenomenon previously unknown among birds: infanticide whereby the aggressor destroys :the current offspring of its newly-acquired mate and there- by initiates breeding with that mate. Trail et al. (1981) review other types of intraspecific infanticide that occur in birds. See Hrdy (1979) for a general review of infanticide. The present findings are particularly noteworthy because infanticide following a mate takeover has been known only as a male reproductive tactic (Hrdy 1979). The alternative tactic to this type of infanticide, to wait for the mate to complete raising its eutxent offspring, entails 'the cost of delayed reproduction. In the present example the delay entailed by that tactic would have been at least 40 days, a substantial port ion of tl~e breeding season.

The observations reported here indicate that poly- androus female jacanas employ some of the same repro- .ductive tactics as polygynous males in several mammalian species, namely, takeovers of mates and destruction of current offspring of mates acquired by takeovers (Hrdy 1979). Why these tactics a r e unreported in polygynous avian species is unclear. The observations also suggest 'that distraction displays can occur intraspecifically in species in which asymmetries in fighting ability between the sexes render direct attack of intruders of the opposite sex ineffective.

I thank the Costa Rican Ministry of Agriculture and Ranching and the personnel of the Dr Rafael Rodriguez National Wildlife Refuge fer their assistance. J. Altmann, S. Attmann, P. Bergstrom, S. Hrdy, M. Pereira, J. Silk, and G. Woolfenden kindly commented on the manu- script. Research was supported by the Organization for Tropical Studies (RIAS Fund), American Museum of Natural History (Chapman Fund), American Ornitho- logists' Union (Van Tyne Fund), N I M H (Behavioral Biology Training Grant MH-15181), NAS (Bache Fund), University of Chicago (Hinds Fund), and NSF (Grant DEB-8113258). I thank D. Jenni for sharing his unpub- lished observations with me.

MARTIN L. STEPHENS Allee Laboratory, of Animal Behavior, University of Chicago, 940 E. 57th St. Chicago, IL, 60637, U.S.A.

References Hrdy, S. B. 1979. Infanticide among animals: a review,

classification, and examination of the implications

for the reproductive strategies of females. Ethol. Sociobiol., 1, 13-40.

Jemai, D. A. 1974. Evolution of polyandry in birds. Am. Zool., 14, 129-144.

Jenni, D. A. & Betts, B. J. 197& Sex differences in nest construction, incubation, and parental behaviour i n the polyandrous American jacana (dacana spinosa). An#n. Behav., 26, 207-218.

Jenni, D. A. & Collier, G. 1972. Polyandry in the American jacana, daeana spinosa. Auk, 89, 743- 765.

Oring, L. W. & Maxson, S. 1978. Instances of simultane- ous polyandry by a Spotted Sandpiper Aetitis maeularia. Ibis, 120, 349-353.

Ridley, M. 1978. Paternal care. Anita. Behav., 26, 904-932. Trail, P. W., Strahl, S. D. & Brown, J. L. 1981. Infanti-

cide in relation to individual and flock histories in a communally breedfiag bird, the Mexican jay (Aphelocoma ultramarina). Am. Nat.. 118, 72-82.

(Received 7 April 1982; revised 10 Jane 1982; MS. number: As-176)

Male Newts Prefer Large Females as Mates Although females are usually considered to be more choosy than males concerning whom they accept as mates (Halliday 1978), males have a potential for mate selection and the extent to which they are choosy is expected to vary with their parental investment (see Halliday, in press). Males have been shown to mate preferentially with larger, more fecund females in a number of species and in several of these, the males make a substantial invest- ment in their progeny through nutrients passed to their mates during copulation, e.g. the mormon cricket, Anabrus simplex (Gwynne 1981) and the checkered white butterfly, Pieris protodiee (Rutowski 1982).

In the Nor th American red-spotted newt, Notoph- thalmus viridescens, the male provides the female with nothing but genes. However, because the male does not continue to produce mature spermatozoa during the breeding season (Adams 1940), each mating depletes his sperm supply. This being the case, the male newt might be expected to maximize his rate of fertilization with every female he inseminates. In many urodele species, female fecundity is correlated positively with body size (Kaplan & Salthe 1979). Thus males should prefer larger females as mates, the size of the female being a reliable indicator of her fecundity. ! here report the results of a simple experiment designed to test this hypothesis.

The newts used in this study were collected in January 1982 from a single pond in Wilson County, Tennessee, and were maintained in the laboratory as described in Verretl (1982). The experiment reported here was conducted in March 1982. A male newt was placed in an observation aquarium measuring 6 1 • with a water temperature of about 20 C, and allowed 10 rain to settle. Two females were then placed in the aquarium, one at least 0.5 g heavier and 5 mm longer than the other. The newts were observed for a period of 20 min, during which time I noted the latency to the first courtship attempt by the male (amplexus or hula display) and the lmmber of attempts directed to each female. Forty such trials were conducted, all involving different animals. See Verrell (1982) for a full account of the sexual behaviour of this species.

A total of 117 courtship attempts were made over the 40 encounters. Significantly more attempts were directed to the larger female (65 G) than to the smaller one (35 ~ ) (X 2 = 10.5, P < 0.001). The mean latency to approach