Download The effects of climate change on Merriam`s kangaroo rat, Dipodomys

Journal of Arid Environments (2001) 49: 581–591
doi:10.1006/jare.2001.0806, available online at http://www.idealibrary.com on
The effects of climate change on Merriam’s kangaroo
rat, Dipodomys merriami
Terri L. Koontz, Ursula L. Shepherd* & Diane Marshall
Department of Biology, University Honors Program,
University of New Mexico Albuquerque, NM 87131, U.S.A.
(Received 15 June 2000; accepted 20 February 2001, published electronically 24 September 2001)
We examined the relationship between body size of Dipodomys merriami and
specific climate variables for the years 1989 through 1996 at the Sevilleta
LTER Station. Earlier work demonstrated a 2–33C increase in both summer
and winter temperatures at Sevilleta over this time period. In that study,
Neotoma albigula’s adult mean body mass decreased as temperature increased.
Dipodomys merriami’s adult mean body mass differed significantly among
years for both sexes but there was no trend across years. There was also no
relationship to climate variables. Abundance showed a positive relationship
with one axis of a PCA of the correlated climate variables.
2001 Academic Press
Keywords: global warming; Dipodomys merriami; kangaroo rat; body size;
Bergmann’s Rule; Neotoma albigula; woodrat; Sevilleta
Introduction
Analysis of surface air temperature shows an increase in global temperature of about
0)3–0)63C over the past 130 years (Gates, 1993; Trenberth, 1993). However, this
warming trend has not been uniform worldwide (Houghton et al., 1996). In the
south-western United States, both summer and winter temperatures have increased at
the Sevilleta National Wildlife Refuge LTER by 2)5–33C for the years 1989 through
1996 (Smith et al., 1998). These figures are important since temperature increases like
these may influence animal communities by altering species’ ranges and/or abundances
(Tracy, 1992). Brown & Kurzius (1987) suggested that heteromyid rodent species
respond to their local environment by acting individualistically. This may include
responses to changes in temperature or precipitation in local environments. In addition,
with a shift in species distributions, whole ecosystems may change via the invasion or
disappearance of species.
Smith et al. (1998) demonstrated that Neotoma albigula responded to both summer
and winter temperature increases at the Sevilleta with a decrease in adult MBM (mean
body mass). They found no relationship between adult MBM and precipitation. Smith
et al. (1998) utilized data from three study areas (Rio Salado Larrea, Rio Salado
grassland and Two-22) at the Sevilleta.
* Corresponding author. Fax: 505-277-4271. E-mail: [email protected]
0140}1963/01/110581#11 $35.00/0
2001 Academic Press
582
T. L. KOONTZ ET AL.
In an effort to understand how community structure might change as a result of
changes in climate, we examined the relationship between adult MBM and abundance
of another rodent species with these changes in temperature and with precipitation. In
addition, we examined the relationship between abundance differences with temperature and precipitation across years. The questions we addressed were: (1) have
changes occurred in adult MBM in D. merriami from 1989 to 1996 at the Sevilleta; if so
(2) were these MBM changes correlated with temperature or precipitation variables for
the year preceding their capture; (3) was D. merriami’s response to climate variables of
the same magnitude and direction as that of N. albigula; and finally (4) did the
abundance of D. merriami have any relationship to climate variables?
Dipodomys merriami was chosen because it provided a large sample size and because it
is a dominant species (i.e. highly abundant) in the area. More importantly, D. merriami
was chosen because it is a member of a keystone guild, along with D. spectablis and
D. ordii, giving it an influential role in the dynamics of plant and animal communities
(Brown & Heske 1990, Heske et al., 1993, 1994). Knowing this, it would be beneficial to
discover D. merriami’s response to a dramatic increase in temperature for a localized area.
Along with its large sample size, dominance, and its ecological impacts to communities, D. merriami has many traits similar to those of Neotoma albigula: it is a nocturnal,
burrowing rodent, and is adapted to a desert environment. However, it is in a
different family, Heteromyidae, and therefore has a different physiology and
phylogenetic history. Baumgardner & Kennedy (1993) reported that D. merriami
exhibits an increase in body size, as measured by various morphological characteristics,
with decreasing latitude, suggesting an unusual relationship between body size and
temperature for this species. Bergmann’s Rule states that, within a species, body size
decreases as the species gets closer to the equator (Van Voorhies, 1996). Dipodomys
merriami and some other Dipodomys species violate Bergmann’s Rule (Baumgardner
& Kennedy 1993). We hypothesized that D. merriami, being a highly adapted desert
rodent species, would be a likely candidate to respond differently to increasing
temperatures.
Materials and methods
Study area
The Sevilleta National Wildlife Refuge LTER (National Science Foundation Longterm Ecological Research) site is located along a transition zone between several biomes:
Great Plains Grassland, Great Basin Shrub Steppe, and Chihuahuan Desert. The
Sevilleta is in Socorro County in central New Mexico at approximately 34315 to 34345
degrees latitude, and 107310 to 106350 degrees longitude.
Sites
Data were obtained from five sites at the Sevilleta: Five Points Grassland, Five Points
Larrea, Rio Salado Grassland, Rio Salado Larrea, and Two-22 (a juniper site). These
sites were chosen because they have been trapped consistently from 1989 to 1996. The
two Five Points study sites are located on the east side of the Sevilleta and are part of the
Great Plains Grassland and the Chihuahuan Desert biomes. Five Points Grassland
contains mainly grasses, with black grama (Bouteloua eriopoda) the dominant species and
a very small amount of creosote bush (Larrea tridentata). Five Points Larrea has the same
species, but the landscape is dominated by creosote bush. Both Five Points sites contain
two species of Opuntia cacti (Opuntia phaeacantha and Opuntia clavata). The two Rio
Salado study sites are located on the west side of the Sevilleta and are a part of the
EFFECT OF CLIMATE CHANGE ON MERRIAM’S KANGAROO RAT
583
Chihuahuan Desert biome. While the Rio Salado sites and Five Points sites are both
made up of two predominant habitats, creosote and grassland, Rio Salado Grassland
and Rio Salado Larrea are much more diverse. Rio Salado Grassland is characterized by
burrograss (Sceleropogon brevifolius) and black grama. The Rio Salado Larrea site
contains both creosote and mesquite (Prosopsis glandulosa), and the dominant grass is
galleta (Hilaria jamesii). Site Two-22 is also located on the west side of the Sevilleta at
the base of the Ladrone Mountains in the Great Basin Shrub Steppe. The area is rocky,
and the dominant species at this site is juniper ( Juniperus monosperma). Bush muhly
(Muhlenbergia porteri), black grama, and blue grama (Bouteloua gracilis) are the dominant grasses at this site. Also, the succulents found here are plains prickly pear (Opuntia
polyacantha), tree cholla (Opuntia imbricata), banana yucca (Yucca baccata), and Spanish bayonet (Yucca glauca).
Species
Dipodomys merriami adult female mean body mass ranges from 36)5 to 43)8 g and adult
male mean body mass from 39)5 to 46)5 g (Reynolds, 1960). Dipodomys merriami
inhabits areas of creosote bush with a variety of soil types. It feeds on seeds of various
types, including mesquite, creosote, and grama grass (Hoffmeister, 1986).
Dipodomys merriami females and males are reproductively active during all months in
the Chihuahuan Desert (Brown & Zeng, 1989). However, there are peaks of pregnancy
in the months of May and September (Reynolds, 1960). Newborns reach sexual
maturity within the first year (Reynolds, 1960). Longevity of D. merriami in the
Chihuahuan Desert averages 43 months, about 3)5 years (Brown & Zeng, 1989).
Mammal trapping
Mammal trapping webs were installed at each of these sites in 1989. Each site contains
five trapping webs. The webs are trapping areas where metal traps are set on 12 transects
radiating from a center point with 12 traps on each transect. Four traps are set
perpendicular to one another at the center stake, giving a total of 148 traps in all. At each
site, two of the five webs are designated as removal webs. The specimens are harvested
and become part of the permanent collection of the Museum of South-western Biology
at the University of New Mexico. The remaining three webs are mark-recapture webs
where animals are identified and measured, then released at the site (Parmenter et al.,
1989; M. Friggens, pers. comm.). Trapping crews record adult body mass for all
mammals at all the sites twice a year—late spring, and late summer. Data from these
webs become part of the archived data set available at the Sevilleta. Adult body masses of
D. merriami were taken from this data set for Five Points, Rio Salado, and Two-22 for
the first trapping period in late spring.
Data
Archives of both the meteorological and mammal trapping data began in 1989 at the
Sevilleta. We used animal data for the years 1989–1996, utilizing only the first trapping
period in late spring of each year. Species, year collected, body mass, sex, and habitat for
each animal measured were retrieved from archived records. Total abundance counts
were recorded for sites and years.
To answer the questions pertaining to MBM, we divided individuals into adults,
sub-adults or juveniles based on the archived records. It was often difficult to tell
whether an individual was a sub-adult or an adult. To alleviate this problem we excluded
584
T. L. KOONTZ ET AL.
any individual with a body mass less than two standard deviations below the mean mass
for each year and by sex. For females, all animals identified as pregnant were removed
from the data set. Fifteen percent (282) of the total population (1856) was removed
from the data set either because they were pregnant or identified as juveniles or
sub-adults.
The temperature data were separated into the following categories: average maximum
hot, average hot, average minimum cold, and average cold. Precipitation data were
divided into summer precipitation (May–October) and winter precipitation (November–April). Meteorological data came from Smith et al. (1998) for Rio Salado and
Two-22 sites. For the Five Points site, we used records from the meteorological station
just north of the site. All data were collected for the years 1989–1995. We used the same
methods as Smith et al. (1998) to obtain the climate variables. For December 1990, we
used the monthly averages for the temperature variables because of missing data in the
daily averages.
Statistical analysis
A one-tailed Wilcoxon’s test was used to test whether the Five Points climate variables
exhibited trends similar to those found at Rio Salado and Two-22 (SAS, PROC
NPAR1WAY WILCOXON, SAS Institute Inc., 1996). A non-parametric analysis of
variance, Kruskal-Wallis test, based on Wilcoxon scores was used to identify trends. We
then compared adult MBM among sex, location, and year by means of ANOVA (SAS,
PROC GLM, SAS Institute Inc., 1996). Adult MBM was the dependent variable and
sex, location within the Sevilleta, year of data collection, and their interactions were the
independent variables. All variables were treated as fixed. Significance tests were
constructed using type III sum of squares due to imbalance in the data set. Tukey tests
were used to determine which years or locations differed significantly for adult
MBM. Adult MBM of females and males were matched with the climate variables from
the previous year. For example, MBM of adult females in May 1990 was assigned to
climate variables from 1989. Only 7 years of rodent data could be matched with climate
variables because climate variables were not available for 1988 to be matched with
rodent data for 1989. The same process was followed to examine the response of
abundance to climate variables. Using Pearson correlation coefficients, we tested
whether climate variables were correlated (SAS, PROC CORR, SAS Institute Inc.,
1996). Since the results showed that the climate variables were strongly correlated, we
ran a principal component analysis to create independent variables (SAS, PROC
PRINCOMP, SAS Institute Inc., 1996). With the new independent variables, we ran
a simple regression with female and male adult MBM and abundance (SAS, PROC
REG, SAS Institute Inc., 1996). Linearity between adult MBM and abundance with
single climate variables at the Sevilleta was examined using simple regressions (SAS,
PROC REG, SAS Institute Inc., 1996). Finally, we ran a Spearman correlation with
adult MBM and abundance (SAS, PROC CORR, SAS Institute Inc., 1996).
Results
More than 1500 adult individuals of D. merriami were trapped during the study period
excluding pregnant females (1574 individuals, 631 females, 943 males) (See
Appendix 1). A total of 74 pregnant females were removed from the data set. More
adults were caught at the Rio Salado site (n"690) than at the Five Points site (n"579)
or the Two-22 site (n"305). In addition, the largest number of adults were caught in
1993 (n"254) and in 1995 (n"276), while the fewest were caught in 1996 (n"86).
We found that adult MBM for females at the Sevilleta ranged from 36)4 to 44)7 g over
EFFECT OF CLIMATE CHANGE ON MERRIAM’S KANGAROO RAT
585
Figure 1. Dipodomys merriami adult mean body mass (MBM) changes and abundance counts
for the years 1989–1996 at the Sevilleta for Five Points (
), Rio Salado (
), and Two-22
(—). (a) adult MBM for females; (b) adult MBM for males; (c) abundance counts.
the study period with a mean of 39)3$6)0 g, while adult MBM for males ranged from
41)4 to 45)8 g with a mean of 43)5$4)6 g (Fig. 1, Appendix 1). Abundance counts
varied from a high of 159 individuals at Five Points in 1995 to a low of 23 individuals at
Five Points in 1996 (Fig. 1, Appendix 1).
Smith et al. (1998) demonstrated that temperature has increased by 2–33C at the
Two-22 and Rio Salado sites at the Sevilleta between 1989 and 1996. The KruskalWallis test based on Wilcoxon scores showed that the Five Points site exhibited the same
trend over this time period (Fig. 2). However, Five Points differed significantly
from Rio Salado for average hot, average minimum cold, and average cold temperatures
across the years. Five Points also differed significantly from Two-22 for average
maximum hot and average minimum cold temperatures across years. Precipitation
586
T. L. KOONTZ ET AL.
Figure 2. Temperature and precipitation changes for the years 1989–1996 at the Sevilleta for
Five Points (
), Rio Salado (
), and Two-22 (
). (a) Average maximum temperature"average of maximum temperatures for the months of July, August, and September:
standard deviation ranged from 3)0 to 4)8; (b) average hot temperature"average of mean
temperatures for the months of July, August, and September: standard deviation ranged from 2)6
to 3)8; (c) average minimum temperature"average of minimum temperatures for the months
of December, January, and February: standard deviation ranged from 3)3 to 5)1; (d) average cold
temperature"average of mean temperatures for the months of December, January, and
February: standard deviation ranged from 3)3 to 5)3; (e) sum for summer precipitation"sum
amount of precipitation from May to October for the year; (f ) sum for winter precipitation"sum
amount of precipitation from November to April for the year. (e, f ) Five points (
); Rio
Salado (
), Two-22 (
).
values showed the same trend among all sites across the years, and did not differ
significantly from site to site across years.
While adult MBM differed significantly among years, there was no directional
trend (Fig. 1, Appendix 1). The Tukey Test showed several comparisons significant at
EFFECT OF CLIMATE CHANGE ON MERRIAM’S KANGAROO RAT
587
Table 1. Results of correlation PCA of six climate variables at three sites for years
1989–1996
Prin
Prin
Prin
Prin
Prin
Prin
1
2
3
4
5
6
Eigenvalue
Difference
Proportion
Cumulative
3)09
1)42
0)92
0)48
0)06
0)03
1)67
0)51
0)44
0)43
0)03
0)00
0)52
0)24
0)15
0)08
0)01
0)005
0)52
0)75
0)91
0)99
0)10
1)00
Eigenvectors
MaxHot
AveHot
MinCold
AveCold
Sppt
Wppt
Prin 1
Prin 2
Prin 3
Prin 4
0)45
0)52
0)28
0)51
!0)23
!0)38
!0)27
!0)03
0)67
0)36
0)51
0)29
0)51
0)38
!0)32
!0)11
0)56
0)41
0)05
0)15
0)16
0)001
!0)61
0)76
su"21.
the 0)05 level. Adult females at Rio Salado and Two-22 were significantly larger in 1992
than females in almost all other years: 1990, 1993, 1994, and 1995. At the Five Points
site, 1992 females were significantly larger than females in 1995. Females at all sites were
significantly smaller in 1993 than in 1991. Dipodomys merriami males were significantly
smaller at all sites in 1990 than in 1996.
We found that adult MBM differed by sex (df."1, type III sum of
squares"827, F value"33, p value"0)0001) and year (df."7, type III sum of
squares"640, F value"3)65, p value"0)0007). There were also two significant
interaction effects on adult MBM: location and year (df."23, type III sum of
squares"2112, F value"3)66, p value"0)0001) and sex, location and year (df."21,
type III sum of squares"2)1, p value"0)0025).
Since each climate variable correlated with at least one other climate variable in all
locations we constructed a principal component analysis (PCA) to create new independent variables. Next we did simple regressions of adult MBM and of abundance against
these new variables. We found no relationship between these new variables with either
female or male adult MBM. Axis Three of the PCA was significant in explaining
differences in abundance (R2"0)59, p value"0)0001) (Table 1). We performed
regressions against the climate variables for females and males at each location. There
were no relationships between adult MBM of D. merriami females or males with
temperature or precipitation at any location. However, simple regressions showed
a positive relationship between abundance counts and summer precipitation at two of
the three sites, Five Points (R2"0)70, p value"0)02) and Rio Salado (R2"0)87,
p value"0)002). There was no correlation between adult MBM and abundance.
Discussion
We found significant changes in adult MBM for both females and males. Although adult
MBM of D. merriami varied from location to location and from year to year, these
588
T. L. KOONTZ ET AL.
changes were not related to the climate variables measured. This is a very different
result than that reported by Smith et al. (1998) for N. albigula. However, abundance of
D. merriami did show a relationship to climate variables.
These results regarding adult MBM could be due to the conservative approach we
used to obtain the adult population for this study. Removal of individuals smaller than
two standard deviations below mean adult body mass of each sex assured that all
remaining individuals were adults. However, this protocol would make it somewhat
more difficult to detect a decrease in adult MBM across years. We believe this
method was necessary because of the difficulties in discriminating between subadults and adults in a species that exhibits great flexibility in its breeding time. Also,
although some adults may have inadvertently been excluded, our sample size was
large enough that any significant decrease in mean mass of the population should still
have been evident. This approach did not impact our ability to see increases in adult
MBM.
It would have been useful to be able to determine whether mortality of D. merriami
across years differed with changes in temperature or whether individual adult body
masses changed in response to changes in climate variables. However, mark-recapture
data available from the Sevilleta did not allow us to recognize individuals from year to
year.
In addition, the response of other species to changing climate variables in any year
may have affected adult MBM in D. merriami for that or the following year. For
example, if N. albigula populations were high in 1990, that might affect adult MBM
of D. merriami in 1990 or in 1991. Some other species could directly or indirectly impact
food resources utilized by D. merriami.
All these issues are of interest. However, our primary goals were to evaluate the
response of local populations of D. merriami to the increase in temperature seen at the
Sevilleta for the years 1989 through 1996, and to compare these responses with those
previously reported for N. albigula during the same period. Since the Sevilleta has
experienced the increase in temperature reported by Smith et al. (1998), we were
provided an unusual opportunity to observe how D. merriami may react to long-term
climate change. The large sample size available provided for a robust evaluation of
D. merriami’s response to the temperature and precipitation variables we considered.
Given that this species is highly adapted to a desert environment and given our large
sample size, we concluded that D. merriami’s adult MBM was not vulnerable to the
amount of temperature change that occurred during this time period at the Sevilleta.
There are a number of mechanisms that may explain why D. merriami adult MBM did
not respond to temperature changes that greatly affected N. albigula. Although both
D. merriami and N. albigula use burrows, Kenagy (1973) showed that D. merriami nests
are deeper than 1 m while Brown (1968) has noted that N. albigula dens contain tunnels
only 10–30 cm below ground. Since D. merriami burrows are deeper, they are likely to
provide more protection from abiotic factors than the shallower N. albigula burrows.
Hence, woodrats are more susceptible to changes in ambient temperatures. Also, while
these two species live in the same environment, summer temperatures at the Sevilleta are
known to be only a few degrees below N. albigula’s lethal zone (Brown, 1968; Smith
et al., 1998). Dipodomys merriami on the other hand is active during the summer at
temperatures as high as 303C suggesting that they are not as adversely affected
(Kenagy, 1973).
Dipodomys merriami’s adult MBM did not respond to climate changes. This could be
because the species’ physiology allows it to deal with water stress in a variety of ways.
Kangaroo rats are especially well adapted to desert environments due to the aid of
external cheek pouches that allow them to reduce the loss of water from salivary actions
(Vander Wall, 1993). A woodrat’s diet consists mainly of succulents whereas kangaroo
rats are granivores and feed predominantly on dry seeds (Hoffmeister, 1986). Even
though D. merriami’s diet consists mainly of dry food, water balance is maintained, in
EFFECT OF CLIMATE CHANGE ON MERRIAM’S KANGAROO RAT
589
part, by the ability of the kidneys to concentrate urine and by the metabolic water from
seeds (MacMillen & Hinds, 1983; Hulbert & MacMillen, 1988).
Our findings are consistent with those of Zeng & Brown (1987) who showed that
D. merriami in south-eastern Arizona, also a part of the Chihuahuan Desert, varies and
adapts its life history traits (especially its reproductive strategies) in response to environmental change. Brown & Zeng (1989) also showed that D. merriami varies its body mass
throughout the year. This ability to vary its reproductive efforts coupled with the
overall longevity of the species makes D. merriami more able to withstand the level of
increased temperature that was experienced at the Sevilleta. Baumgardner & Kennedy
(1993) reported an inverse relationship between latitude and body size as measured by
the skeletal structure of D. merriami, while Baumgardner (1989) reported that
D. merriami showed no correlation between body size and either annual temperature or
precipitation. These results agree with our findings and lead us to believe that indeed the
degree of change reported in these abiotic variables was not sufficient to affect
body size or mass in D. merriami. It is still possible that larger changes in temperature
and precipitation could change body size and/or mass in D. merriami.
This does not mean that the species might not have experienced other consequences
of climate change (i.e. differing mortality, changes in abundance, or individual
responses such as differential reproductive activity). Although our data set did not
allow us to investigate the questions of differing mortality or individual responses,
we did test whether changes in climate variables affected abundance.
Dipodomys merriami abundance showed a relationship to the new independent climate
variable depicted by Axis Three of the PCA (Table 1). The loadings on this axis
demonstrate that when summer precipitation and summer maximum temperatures were
high, abundance increased. Both high winter precipitation and high summer average
temperatures also had a positive impact on abundance in years when summer precipitation and maximum summer temperatures were high. The last two variables, minimum
cold and average cold temperatures, were not as important in predicting abundance
because they were weighted less on the axis. Interestingly, both loaded onto the axis with
the opposite sign from all the other variables. We’re confident about these results; since,
they were further supported by the simple regression that demonstrated that at two sites
increases in summer precipitation alone was a causative factor in increasing abundance.
This increase in abundance could have resulted from several mechanisms. It is
possible that in times when both precipitation and temperatures were high there was
more food due to increases in seed output by D. merriami’s preferred foods. It is also
possible that competitors were not able to withstand the temperature changes and
suffered population declines. Therefore, more food or other resources became
available to D. merriami. When temperatures were high and rainfall was low, this pattern
reversed (see 1996) and populations crashed at all three sites. Since the data set did not
include more years, it is difficult to evaluate whether this decline would hold up over
time. However, it makes intuitive sense that the absence of rainfall would affect food
availability for this species.
Dipodomys merriami MBM clearly responded differently to changes in temperature and precipitation variables than N. albigula. Asking the additional question about
how D. merriami abundance responded to these variables added new insight. The
increase in abundance as temperatures rose suggests that D. merriami can withstand
changes of this magnitude at least over the short term. Other researchers (Zeng
& Brown, 1998; Brown & Zeng, 1989) have reported that D. merriami has a very plastic
reproductive strategy, and these results support those earlier findings. All these factors
suggest that D. merriami is likely to be among the more successful species in responding
to climate shifts.
Global warming is now well documented (Gates, 1993; Trenberth, 1993; Houghton
et al., 1996; Smith et al., 1998). Our results suggest that in the event of severe warming in
the south-western U.S., some species will alter their morphology to adapt to changes
590
T. L. KOONTZ ET AL.
while others will not. If some species are unable to alter their morphology or life history
traits sufficiently to survive these severe climatic changes, the differing responses could alter species composition in the southwest and may have widespread
ecological consequences (e.g. changes in vegetation structure or changes in the seed
bank). Dipodomys merriami and other kangaroo rats may become even more dominant
species in arid systems because of their ability to withstand temperature changes with no
visible consequences, at least regarding adult MBM and population abundance.
The United States Department of Fish and Wildlife and the Sevilleta National Wildlife Refuge
LTER made this study possible by allowing us to work on the Sevilleta. We would like to thank
D. Moore, M. Friggens, and R. Parmenter for access to the archived data at the Sevilleta. Thanks
to Joe Massotti and Ed Bedrick for assistance with statistical analysis. A special thanks to David
Gray for helping whenever help was needed. Also, we would like to thank Jeremy Littell for
listening to concerns and giving helpful advice. Finally, this research was funded by the National
Science Foundation (NSF) through Research Experience for Undergraduates (REU) grant
C BIR-9424121. This is publication 181 for the Sevilleta LTER.
References
Baumgardner, G.D. (1989). Geographic patterns of size and correlation with selected abiotic
variables In: Morphologic variation in kangaroo rats (genus Dipodomys), pp. 46–92. Ph.D.
dissertation, Memphis State University. 469 pp.
Baumgardner, G.D. & Kennedy, M.L. (1993). Morphometric variation in kangaroo rats (genus
Dipodomys) and its relationship to selected abiotic variables. Journal of Mammalogy, 74: 69–85.
Brown, J.H. (1968). Adaptation to environmental temperature in two species of woodrats,
Neotoma cinerea and Neotoma albigula. Miscellaneous Publications of the Museum of Zoology,
University of Michigan, 135: 1–48.
Brown, J.H. & Heske, E.J. (1990). Control of a desert-grassland transition by a keystone rodent
guild. Science, 250: 1705–1707.
Brown, J.H. & Kurzius, M.A. (1987). Composition of desert rodent faunas: combinations of
coexisting species. Annales Zoology. Fennici, 24: 227–237.
Brown, J.H. & Zeng, Z. (1989). Comparative population ecology of eleven species of rodents in
the Chihuahuan Desert. Ecology, 70: 1507–1525.
Gates, D.M. (1993). Climate Change and its Biological Consequences. Sunderland, Mass.: Sinauer
Associates. 280 pp.
Heske, E.J., Brown, J.H. & Guo, Q. (1993). Effects of kangaroo rat exclusion on vegetation
structure and plant species diversity in the Chihuahuan Desert. Oecologia, 95: 520–524.
Heske, E.J., Brown, J.H. & Mistry, S. (1994). Long-term experimental study of Chihuahuan
Desert rodent community: 13 years of competition. Ecology, 75: 438–445.
Hoffmeister, D.F. (1986). Mammals of Arizona. Tucson: University of Arizona Press. 602 pp.
Houghton, J.T., Meira Filho, L.G., Callander, B.A., Harris, N., Kattenberg, A. & Maskell, K.
(1996). Climate Change 1995: the Science of Climate Change. New York, NY: Cambridge
University Press. 572 pp.
Hulbert, A.J. & MacMillen, R.E. (1988). The influence of ambient temperature, seed composition
and body size on water balance and seed selection in coexisting heteromyid rodents. Oecologia,
75: 521–526.
Kenagy, G.J. (1973). Daily and seasonal patterns of activity and energetics in a heteromyid rodent
community. Ecology, 54: 1201–1219.
MacMillen, R.E. & Hinds, D.S. (1983). Water regulatory efficiency in heteromyid rodents:
a model and its application. Ecology, 64: 152–164.
Parmenter, R.R., MacMahon, J.A. & Anderson, D.R. (1989). Animal density estimation using
a trapping web design: field validation experiments. Ecology, 70: 169–179.
Reynolds, H.G. (1960). Life history notes on Merriam’s kangaroo rat in southern Arizona.
Journal of Mammalogy, 41: 48–57.
Smith, F.A., Browning, H. & Shepherd, U.L. (1998). The influence of climate change on the
body mass of woodrats Neotoma in an arid region of New Mexico, USA. Ecography, 21:
140–148.
EFFECT OF CLIMATE CHANGE ON MERRIAM’S KANGAROO RAT
591
Tracy, C. R. (1992). Ecological responses of animals to climate. In: Peters, R.L. & Lovejoy, T.E.
(Eds), Global Warming and Biological Diversity, pp. 171–179. New Haven: Yale University
Press. 386 pp.
Trenberth, K.E. (1993). Northern hemisphere climate change: physical processes and observed
changes. In: Mooney, H.A., Fuentes, E.R. & Kronberg, B.I. (Eds), Earth System Responses to
Global Change: Contrasts Between North and South America, pp. 35–59. San Diego: Academic
Press. 365 pp.
Vander Wall, S.B. (1993). Salivary water loss to seeds in yellow pine chipmunks and Merriam’s
kangaroo rats. Ecology, 74: 1307–1312.
Van Voorhies, W.A. (1996). Bergmann size clines: a simple explanation for their occurrence in
ectotherms. Evolution, 50: 1259–1264.
Zeng, Z. & Brown, J.H. (1987). Population ecology of a desert rodent: Dipodomys merriami in the
Chihuahuan Desert. Ecology, 68: 1328–1340.
Appendix. Summary of Dipodomys merriami female and male data for the
Sevilleta from 1989 to 1996
Females
Males
Year
Location
Abundance
n
MBM (g)
S.D.
n
1989
1990
1991
1992
1993
1994
1995
1996
1989
1990
1991
1992
1993
1994
1995
1996
1989
1990
1991
1992
1993
1994
1995
1996
5points
5points
5points
5points
5points
5points
5points
5points
RioSalado
RioSalado
RioSalado
RioSalado
RioSalado
RioSalado
RioSalado
RioSalado
Two-22
Two-22
Two-22
Two-22
Two-22
Two-22
Two-22
Two-22
104
59
70
115
80
52
159
23
104
77
109
123
119
124
132
39
54
35
47
51
77
37
42
30
46
16
27
36
24
16
63
7
35
17
25
35
44
54
43
11
11
12
17
12
36
15
16
13
40
40
41)9
39)5
40)6
42)1
38)3
39)1
40)2
37)3
39)5
44)7
38)8
37)3
38)1
44)6
40
36)5
39)6
42)3
34)7
38)4
41)5
36)4
3)6
2)3
4)2
5)7
5)3
8)9
4)3
2)3
4)4
6)8
5
8)6
6
6)7
5)1
5)1
2)6
5
3)8
2)7
7)3
7)4
4)2
5)9
47
40
33
60
50
30
68
16
53
46
53
56
68
57
67
26
22
18
24
26
32
19
19
13
n"sample size, MBM"adult mean body mass, S.D."standard deviation.
MBM (g) S.D.
42)5
43)2
45)2
43)3
43)8
44)2
42)3
43)8
42)6
42
43)9
44)5
42)5
44)1
44
45)8
45
41)4
42)9
44)8
44)5
43)9
44)2
44
3)9
4
4)2
4)3
3)2
3)6
5)3
2)3
4)7
5)7
5)4
4)2
4)7
5)5
4)6
4)1
3)8
3)9
5)2
3)6
6)2
3
4)6
2)4