12 
PROCEEDINGS OF THE CALIFORNIA ACADEMY OF SCIENCES 
Series 4, Volume 65, Supplement I 
Grant and Banack, 1995). The Samoan flying fox is nocturnal on Fiji but is well known to be more 
diurnal in the Samoa Islands, and perhaps diumality is facilitated by the absence of birds of prey. 
Hunting and deaths as a result of cyclones impacting food availability and habitat structure are 
the major recognized mortality factors and have serious implications (see details under “Manage¬ 
ment Practices and Concerns” below). Monitoring of a limited number of roosting sites of 
P. samoensis in American Samoa before and after two cyclones showed continued occupancy. 
However, mortality of immature P. samoensis as a result of the cyclones appeared to be high (Pier¬ 
son et al., 1996a). Craig et al. (1994) modeled population growth projections for a likely initial size 
of 400 Samoan flying foxes with no further hunting or hurricanes, and calculated that it would take 
from 13^10 years to reach a desirable size of 1,500 bats. If hunting continued at a rate considered 
typical, the population would likely decrease steadily or at best not achieve the desirable size for 
more than a century, even in the absence of catastrophic storms (but see section on “Management 
Practices and Concerns below). 
Population Trend: According to Flannery (1995:289), “This beautiful flying-fox is still read¬ 
ily observed in many parts of Fiji, but it is now on the verge of extinction in parts of Samoa.” A 
petition to the U.S. Fish and Wildlife Service to list the species as endangered prompted a series of 
field surveys beginning in the mid-1980’s (Wilson and Engbring, 1992; Craig et al., 1994). Surveys 
were improved each year to result in an index based on standardized counts of daytime-active bats 
from a series of viewing stations (mean number of bats seen/size of viewing area/20-minute peri¬ 
od for five replicate counts), primarily on Tutuila (Craig et al., 1994; Morrell and Craig, 1995). 
This index, which typically ranged from about six to eight from 1986 to 1990, dropped drastically 
to about 1.5 in 1991 and 1992 (Craig et al., 1994). This decline was attributable to damage from a 
severe hurricane in 1990. An approximation of population size for Tutuilla was calculated by mul¬ 
tiplying the index by the total available suitable area and making somewhat arbitrary corrections 
for inactive (undetectable) bats, species misidentifications, and bats moving among survey areas. 
This index suggested that only about 200-400 P. samoensis remained on Tutuila after Hurricane 
Ofa (Craig et al., 1994), in comparison with a pre-Ofa estimate of less than 700 (Pierson et al., 
1996a). Craig et al. (1994) noted that out of the four islands of Tutuila, Ta’u, Olosega, and Ofu, 
nearly 70% of the remaining population in 1992 occurred on Tutuila, and about 25% on Ta’u. A 
population size of 300-500 was suggested for Tutuila in the early 1990s (Banack and Grant, 2003). 
Additional surveys were conducted in 1995-1996 using somewhat different techniques, and in 
1996 the number of P. samoensis on Tutuila was estimated at 854 (Brooke, 1997). However, arriv¬ 
ing at population size or trend estimates for this species has been problematic due to the ad hoc 
nature and numerous changes in survey methods over time (Utzurrum et al., 2003), as well as 
recent recognition that on American Samoa many of these bats are active at night (Brooke, 2001; 
Banack and Grant, 2003). As a result, counts based on diurnal activity alone could be underesti¬ 
mates (Brooke, 2001). Population growth rate estimates based on a simple model have been made 
to assess the likely importance of mortality factors (Craig et al., 1994; see “Mortality” above). 
Population Genetics: Analysis of mitochondrial D-loop and microsatellite nuclear DNA of 
Samoan flying foxes was conducted based on samples from American Samoa (Olosega, Ofu, and 
Tutuila), Fiji, and Samoa (Russell et al., 2016). Mitochondrial DNA from 19 bats in American 
Samoa showed high haplotype diversity and nucleotide diversity; observed heterozygosity at six 
microsatellite loci did not deviate from Hardy-Weinberg expectations (Russell et al., 2016). Genet¬ 
ic structuring among those islands sampled across American Samoa, Samoa, and Fiji was apparent 
from the mitochondrial DNA analysis but not from micro satellite nuclear DNA analysis, which 
revealed detectable genetic differences only among archipelagos rather than individual islands 
(Russell et al., 2016). 
