278
of relative tree species dominance over time in conjunction with changes in climate and fire regime.
Plant macrofossil analysis provided crucial specieslevel identification for the vegetation reconstruction. Forests dominated by Abies lasiocarpa and
Picea engelmannii occupied the site from 8000 to
2600 years ago, but a stable Pinus contorta forest
has been present during the past 2600 years, perhaps reflecting a shift to drier climate conditions
(Fall, 1997).
19.4.2 Plant Macrofossils in Middens
Members of the genus Neotoma, commonly known
as packrats, share the habit of accumulating plant
material in middens within caves and rock crevices
or within stick houses (Wells, 1976). This habit has
been extensively exploited by researchers in North
America to reconstruct vegetation over the past
40,000 years (Betancourt et aI., 1990). The contribution of fossil packrat middens has been especially important in arid regions of westem North
America, which may have relatively few other
sources of fossil plant records (Van Devender and
Spaulding, 1979). When deposited in a dry, protected environment and encased in crystalline
urine, middens contain extremely well preserved
plant fragments that are suitable for radiocarbon
dating (Betancourt et al., 1986). Each midden or
distinguishable layer in a midden is deposited over
one to several years and therefore represents a point
in time. Packrats collect plant material within foraging ranges that are usually less than 1 ha in size,
resulting in a sample of vegetation whose spatial
extent is very local in nature (Betancourt et al.,
1986). Because middens that are preserved over
time are located in caves and rocky shelters, they
generally contain vegetation samples that are restricted to rocky habitats (Spaulding, 1990). No information about patch or age class structure of the
vegetation is provided.
The contents of a midden result from the habitat preferences, den-building behavior, and diet of
packrats (Vaughan, 1990). Some packrat species
are dietary generalists and widely sample the plant
species present in an area, while other species are
dietary specialists (Vaughan, 1990). For this reason, packrat middens may provide a biased representation of the vegetation in an area, and it is desirable to determine which animal species is the
agent of deposition whenever possible. It may be
difficult to distinguish modem from ancient midden deposits in the field (Finley, 1990). Some deposits appear to have been produced within a very
Methods for Determining Historical Range of Variability
short time, and mUltiple radiocarbon dates from
such deposits give a uniform age (Wells, 1976).
However, in other cases, layers in a midden may
have been deposited at different times, requiring
that care be taken in collecting samples to avoid
mixing material of different ages (Wells, 1976).
In addition to plant macrofossils, middens also
contain pollen, which is derived from several
sources and has been studied in conjunction with
macrofossils (King and Van Devender, 1977;
O'Rourke, 1991). Pollen present in packrat fecal
pellets reflects the animal's dietary habits (King
and Van Devender, 1977). In nonfecal portions of
the midden, pollen may have been airborne, settling
on the midden during its construction, or may have
adhered to plants brought to the midden by packrats
(King and Van Devender, 1977; O'Rourke, 1991).
Therefore, nonfecal pollen is potentially contributed by both local and regional sources (Anderson and Van Devender, 1995). Midden pollen
is analyzed using methods developed for pollen
from sediments (King and Van Devender, 1977).
The pollen analysis can be used to confirm the
macrofossil data and, conversely, identification of
macrofossils can narrow taxonomic assignments to
pollen found in middens (King and VanDevender,
1977). However, such pollen records have been
found to be highly variable between middens of
similar age and location, and the spatial extent represented by the sample is unknown (i.e., can vary
from local to regional; King and Van Devender,
1977).
EXAMPLES
1. A number of chapters in Betancourt et aI.
(1990) describe studies of postglacial vegetation
history from packrat middens in various parts of
the western United States. Other animals besides
packrats produce middens that can be analyzed to
reconstruct plant assemblages. They include porcupines (Erethizon dorsa tum) in North America
(Van Devender et al., 1984; Betancourt et aI., 1986)
and various rodents in South America (Markgraf et
aI., 1997). Chapters in Betancourt et aI. (1990) describe middens from hyraxes (Procaviidae) and
dassie rats (Petromuriidae) in Africa (Scott, 1990),
hyrax middens in the Middle East (Fall et al., 1990),
and stick-nest rats (Leporillus) in Australia (Nelson et aI., 1990).
2. Spaulding (1990) compared separate pollen
analysis and packrat midden studies conducted in
localities in western North America as a means of
testing pollen-based vegetation reconstructions.
Comparisons enabled exploration of the strengths
of relative tree species dominance over time in conjunction with changes in climate and fire regime.
Plant macrofossil analysis provided crucial specieslevel identification for the vegetation reconstruction. Forests dominated by Abies lasiocarpa and
Picea engelmannii occupied the site from 8000 to
2600 years ago, but a stable Pinus contorta forest
has been present during the past 2600 years, perhaps reflecting a shift to drier climate conditions
(Fall, 1997).
19.4.2 Plant Macrofossils in Middens
Members of the genus Neotoma, commonly known
as packrats, share the habit of accumulating plant
material in middens within caves and rock crevices
or within stick houses (Wells, 1976). This habit has
been extensively exploited by researchers in North
America to reconstruct vegetation over the past
40,000 years (Betancourt et aI., 1990). The contribution of fossil packrat middens has been especially important in arid regions of westem North
America, which may have relatively few other
sources of fossil plant records (Van Devender and
Spaulding, 1979). When deposited in a dry, protected environment and encased in crystalline
urine, middens contain extremely well preserved
plant fragments that are suitable for radiocarbon
dating (Betancourt et al., 1986). Each midden or
distinguishable layer in a midden is deposited over
one to several years and therefore represents a point
in time. Packrats collect plant material within foraging ranges that are usually less than 1 ha in size,
resulting in a sample of vegetation whose spatial
extent is very local in nature (Betancourt et al.,
1986). Because middens that are preserved over
time are located in caves and rocky shelters, they
generally contain vegetation samples that are restricted to rocky habitats (Spaulding, 1990). No information about patch or age class structure of the
vegetation is provided.
The contents of a midden result from the habitat preferences, den-building behavior, and diet of
packrats (Vaughan, 1990). Some packrat species
are dietary generalists and widely sample the plant
species present in an area, while other species are
dietary specialists (Vaughan, 1990). For this reason, packrat middens may provide a biased representation of the vegetation in an area, and it is desirable to determine which animal species is the
agent of deposition whenever possible. It may be
difficult to distinguish modem from ancient midden deposits in the field (Finley, 1990). Some deposits appear to have been produced within a very
Methods for Determining Historical Range of Variability
short time, and mUltiple radiocarbon dates from
such deposits give a uniform age (Wells, 1976).
However, in other cases, layers in a midden may
have been deposited at different times, requiring
that care be taken in collecting samples to avoid
mixing material of different ages (Wells, 1976).
In addition to plant macrofossils, middens also
contain pollen, which is derived from several
sources and has been studied in conjunction with
macrofossils (King and Van Devender, 1977;
O'Rourke, 1991). Pollen present in packrat fecal
pellets reflects the animal's dietary habits (King
and Van Devender, 1977). In nonfecal portions of
the midden, pollen may have been airborne, settling
on the midden during its construction, or may have
adhered to plants brought to the midden by packrats
(King and Van Devender, 1977; O'Rourke, 1991).
Therefore, nonfecal pollen is potentially contributed by both local and regional sources (Anderson and Van Devender, 1995). Midden pollen
is analyzed using methods developed for pollen
from sediments (King and Van Devender, 1977).
The pollen analysis can be used to confirm the
macrofossil data and, conversely, identification of
macrofossils can narrow taxonomic assignments to
pollen found in middens (King and VanDevender,
1977). However, such pollen records have been
found to be highly variable between middens of
similar age and location, and the spatial extent represented by the sample is unknown (i.e., can vary
from local to regional; King and Van Devender,
1977).
EXAMPLES
1. A number of chapters in Betancourt et aI.
(1990) describe studies of postglacial vegetation
history from packrat middens in various parts of
the western United States. Other animals besides
packrats produce middens that can be analyzed to
reconstruct plant assemblages. They include porcupines (Erethizon dorsa tum) in North America
(Van Devender et al., 1984; Betancourt et aI., 1986)
and various rodents in South America (Markgraf et
aI., 1997). Chapters in Betancourt et aI. (1990) describe middens from hyraxes (Procaviidae) and
dassie rats (Petromuriidae) in Africa (Scott, 1990),
hyrax middens in the Middle East (Fall et al., 1990),
and stick-nest rats (Leporillus) in Australia (Nelson et aI., 1990).
2. Spaulding (1990) compared separate pollen
analysis and packrat midden studies conducted in
localities in western North America as a means of
testing pollen-based vegetation reconstructions.
Comparisons enabled exploration of the strengths