BIOLOGY OF RODENTS 



205 



be an adaptation to reduce intraspecific competition under 

 conditions where a variety of foods is readily available. 



The mice were smaller in body size than mainland 

 forms (9 g vs. 15+ g); this was hypothesized by Berry 

 and Jackson (1979) to be an adaptation to the constant 

 high temperatures in a situation where predators are gen- 

 erally absent. Both populations were considered to be 

 western (rather than Asiatic) in origin, supporting their 

 hypothesized introduction after World War II. 



The high density Polynesian rat population on Japtan 

 islet interfered with specific trapping efforts there, and no 

 comparable studies were possible. We did occasionally 

 observe different pelage forms there, suggestive of animals 

 having escaped from laboratory colonies maintained during 

 the test program on this islet and interbreeding with the 

 local population. 



Two breeding peaks (January through February and 

 July through August) were observed. Average litter size 

 was 4.0. 



Habitat Selection 



The Polynesian rat is a ground-dwelling rat, although it 

 has extensive arboreal highways, well odor-marked, 

 through the low vegetation. Frequently palm fronds were 

 used, which we confirmed with direct field observations 

 under red-light conditions and later in a simulated environ- 

 ment in our home laboratory using fronds brought back 

 from Enewetak. Nests were rarely found but were con- 

 sidered to be under or among surface debris (e.g., piles of 

 coconuts) or in shallow burrows. We never observed their 

 feeding in the crowns of coconut trees, even where islet 

 distribution of rats was allopatric. This species was trapped 

 on the trunks of coconut trees and was found feeding on 

 freshly damaged small nuts on Japtan. 



This rat was found on the more densely vegetated 

 islets but was absent from some of the smaller and/or 

 more remote islets, especially on the western side of the 

 atoll. We conclude, on the basis of conversations with 

 Enewetak elders, that before the atomic tests, these rats 

 were present only on those islets that had been regularly 

 used in the past for coconut harvesting. 



On islets such as Enjcbi and Runit as well as Medrcn 

 and Enewetak, which initially were infested by Polynesian 

 rats (based on early observations and records), clearing and 

 construction activities and test device detonation (in the 

 case of Enjebi and perhaps Runit) eliminated them. These 

 islets are now occupied by only the roof rat. Whether 

 interspecific competition was a factor in these local extinc- 

 tions of the Polynesian rat is not known. Accidental 

 transport of rats by man (initially by the Micronesians in 

 their canoes, later by AEC and military personnel in supply 

 craft) is considered the primary mode of spreading rats 

 around the atoll. Both tidal flow patterns and abundance of 

 predatory fish reduce the likelihood of direct water 

 transport. 



The roof rat infested coconut crowns when these trees 

 were present. Although they readily climbed available 



vegetation, they traveled easily on the ground surface. 

 They used surface debris or shallow burrows for nest sites; 

 they also dug around bunker foundations. This species 

 prospered on the more disturbed islets, those having less 

 cover, but was absent from more remote islets having 

 minimal human activity during the atomic test program. 



The Polynesian rat and the roof rat readily invade 

 structures, and almost all storage and inhabited buildings 

 were infested. The facility at Lojwa during the cleanup 

 operations had a particularly difficult time with invading 

 Polynesian rats. The initial population density was high, 

 and vegetation removal forced the concentration of surviv- 

 ing rats in a perimeter strip. As soon as buildings were 

 erected, rats took up residence. 



Popuiation Density and Home Range 



Live-trapping, mark-and-release studies were under- 

 taken for roof rats on Runit and Polynesian rats on Japtan. 

 Roof rat population density in this open, grass-sedge-shrub 

 environment was about 20 animals 11,000 m^^ or about 

 one animal 550 m~^ (Jackson, 1967). This density may 

 have been exceeded in subsequent years on Japtan with 

 diminished human disturbance; frequently, we caught two 

 rats in the same snap trap. 



In these environments, the home range (as measured 

 by standard diameter) for roof rats was 67 m for females 

 and 100 m for males. For Polynesian rats, it was 50 m 

 (both sexes). 



Food Habits 



Rats, even though at the apex of the terrestrial food 

 pyramid, are largely vegetarians in this environment. In a 

 study of food habits. Fall et al. (1971) found arthropxsd 

 (insects, centipedes) remains somewhat more frequently in 

 roof rat than Polynesian rat stomachs (33% vs. 10%). 

 Additional studies at Enewetak and on the Northern 

 Marshall Islands Radiological Survey reaffirmed these pat- 

 terns; however, the volume of animal matter was small 

 (<2%) (Temme, 1982). Seasonal fruits and seeds, as well 

 as vegetative plant structures, were recognized in the 

 stomach contents. 



Roof rats occasionally were seen at night foraging on 

 the beach and even out onto the exposed reef flat. Fish 

 (trapped in tidal pools) may well have been caught or 

 scavenged, though we did not observe this. 



In a simulated predation situation with roof rats con- 

 fined to beach enclosures, we did, with starlight scopes 

 and under direct moonlight, observe rats enter ghost crab 

 {Oci>pode sp.) burrows, pull the crab out, dismember, and 

 eventually eat the crab. Often the eye stalks were the ini- 

 tial target of the rat. 



When we captured rats in the vicinity of ground-nesting 

 tern colonies, we occasionally found bird remains in the 

 stomachs. We were unable to determine if this resulted 

 from predation or scavenging. In one instance, we found' 

 opened eggs in a f>ortion of a sooty tern colony and 



