which occurs only in mesic grassland along the western
edge of Marin County, California, and is isolated by over
100 km from the nearest adjacent subspecies to the north
(Lidicker 1998). In the last 3.5 decades, it has been found
only within Point Reyes National Seashore. Highly vulner-
able examples would include the Vancouver Island marmot,
which occurs only in a narrow band of subalpine habitat
in the southern half of Vancouver Island (Johnson 1989),
an area heavily impacted by timber harvesting and recre-
ational facilities. Another is the saltmarsh harvest mouse
(Reithrodontomys raviventris), which is restricted to the
saltmarshes of San Francisco Bay (California), especially to
areas dominated by pickleweed (Salicornia) and that are
adjacent to upland areas suitable as refuges during very
high tides (Shellhammer 1998).
Ephemeral patterns
In this category, I include species whose role in the com-
munity is temporally strongly variable. Hibernators, for ex-
ample, largely withdraw from active participation for sig-
nificant parts of the annual cycle. Among rodents, these
include marmots, ground squirrels, dormice, and zapodids.
Seasonally migratory species similarly come and go from
communities. This is quite common among rodents. In cold
climates, house mice tend to move into buildings in the win-
ter and crop fields in warmer periods. The root vole (Mi-
crotus oeconomus) in Finland moves between upland mead-
ows and marsh borders (Tast 1966), Norwegian lemmings
(Lemmus lemmus) breed in alpine tundra but winter in sub-
alpine scrub (Kalela 1961; Kalela et al. 1971), various Afri-
can rodents move in and out of a Zambian flood plain on
a seasonal basis (Sheppe 1972), and where California voles
live in grasslands adjacent to saltmarshes, they will mostly
move into the marshes during the dry summers and then
migrate back into the uplands when winter rains commence
(Harding 2000). Many other examples are known for such
interhabitat migrations. From a conservation perspective,
it is essential that protected areas include all the required
habitat types, not only the one associated with the breeding
season.
An especially important and widespread ephemeral dem-
ographic pattern is that of multiannual cycles in numbers.
These are best known among arvicoline rodents, where the
phenomenon has been intensively studied (Taitt and Krebs
1985; Lidicker 1988, 1991; Stenseth 1999). Cyclic behav-
ior raises several conservation concerns. First of all, popu-
lations must be large enough to survive periods of low den-
sity. Related to this, the effective population size (Ne) will
often be closer to the low densities experienced by a popu-
lation than to the more conspicuous peaks. Low Nerisks
the genetic consequences of small population sizes (Allen-
dorf 1986; Ralls et al. 1988; Lacy 1997). These dangers in-
clude inbreeding, genetic drift, skewed sex ratios, and pos-
sibly the loss of adaptive polymorphisms. The second area
of concern is that of the habitat mosaic available to such
cyclic populations. Often when cyclic species crash in num-
bers they survive only in especially favorable patches or ref-
ugia. As numbers grow, they progressively occupy the less
favorable matrix around these refugia, and at peak num-
bers will inhabit all habitat types in which survival is pos-
sible, even for the short term (Wolff 1980a; Lidicker 1985).
This scenario has several critical implications. The refugia
must be large enough and numerous enough to assure sur-
vival through the low-density episodes. They also must
be close enough together that recolonization can readily oc-
cur into patches that do become extinct. It may also be the
case that secondary habitats are essential for such species
to persist. This is because without dispersal sinks (Lidicker
1975) refugia may sustain high densities for long inter-
vals, such that significant vegetation damage is perpetrated,
and /or predators become abundant, and /or parasites may
invade and spread easily. Evans (1942) has documented a
case where bank voles (Clethrionomys glareolus) were able
to survive population crashes only because some individu-
als persisted in secondary habitat.
The final class of ephemeral patterns is that of irrup-
tive species. Such populations occur at low densities most
of the time, but occasionally irrupt to very high numbers.
Irruptions are usually attributed to an unusual combina-
tion of favorable conditions that permit a large increase in
numbers. Irruptive species generally have high potential re-
productive rates, so they can opportunistically take advan-
tage of favorable circumstances. In South Australian wheat-
growing areas, house mice irrupt to “plague” proportions
when rare summer rains permit the construction of bur-
rows needed for successful reproduction at the same time
that food is abundant (Newsome 1969). Rodents in tem-
perate forests increase greatly during so-called mast years
(Pucek et al. 1993; Ostfeld et al. 1996). In Japan (Tanaka
1957) and southern South America, rodents have been ob-
served to reach plague proportions following the massive
flowering of bamboo, a phenomenon which seems to oc-
cur only rarely. Jaksic and Lima (2003) summarize data for
rodent outbreaks (ratadas) in South America that include
twenty-eight associated with bamboo flowering and twenty-
seven with above-average rainfall.
Persistence of irruptive species will depend on conditions
being suitable for long-term survival at low numbers. How-
ever, it may also be critical for these species to occasionally
experience irruptions. A brief episode of high numbers will
allow recolonization of empty or new habitat patches, and
most importantly, a thorough mixing of local gene pools.
Such periodic genetic revolutions will serve to reverse any456 Chapter Thirty-Eight