Conservation of endangered species - Ecological Applications: Toward a Sustainable World - page 118

Fig. 5.4 Fender's blue
butterfl y population
dynamics recorded for
populations with a
mean population size of
(a) more than 25 and
(b) less than 25
butterfl ies. FR, Fern
Ridge; WC, Willow
Creek. (After Schultz &
Hammond, 2003.)

(a)

WC Bailey
WC Main
WC North

Baskett Butte
Butterfly Meadows
Fir Butte
Oak Ridge

10,000

1000

100

10

1

1992 1994 1996 1998 2000 2002

Year

(b)

FR Eston Lane
FR Spires Lane
Gopher Valley

McTimmand's valley
Mt Crook

1000

100

10

1

1992 1994 1996 1998 2000 2002

Year

5.3.2 A primate in
Kenya - how good
are the data?

In their study of the Tana River crested mangabey ( Cercocebus galeritus galeritus ) in
Kenya, Kinnaird and O'Brien (1991) used a similar approach to estimate the popula-
tion size needed to provide a 95% probability of persistence for 100 years. This
endangered mammal is confi ned to the fl oodplain forest of a single river where it
declined in numbers from 1200 to 700 between 1973 and 1988 despite the creation
of a reserve. Its naturally patchy habitat has become progressively more fragmented
because of agricultural expansion. The key parameters were estimated from just a
few years of data to be an average population growth rate of r = 0.11 with variance

V 0.20. This corresponds to a minimum viable population of 8000. Given the

amount of habitat actually available, Kinnaird and O'Brien concluded that the
mangabeys could not attain such a population size. In fact, they think it unlikely
that this naturally rare and restricted species has ever been so abundant. The data
were probably defi cient; for example, environmental variation in r may be smaller
than estimated if the mangabeys are able to undergo dietary shifts in response to
habitat change.
While these simple census-based approaches have the advantage of requiring a
minimum of information, the results must be treated with caution. Lotts et al. (2004)
analyzed a series of published data sets and concluded that ideally at least two
decades of census data are needed, because predictions based on short time series
produce unreliable estimates of variance in growth rate - as with the mangabeys.
They noted, however, that the approach can be useful to rank populations according
to extinction risk, the approach taken with the butterfl ies in Section 5.3.1. It is cer-
tainly not without some value.

=