Prevalence of malaria parasites (Plasmodium floridense and Plasmodium azurophilum) infecting a Puerto Rican lizard (Anolis gundlachi): A nine-year study

Abstract

The prevalence of malaria parasites was studied in the lizard Anolis gundlachi over a 9-yr period at a site in the wet evergreen forest of eastern Puerto Rico. Three forms of the parasite infected the lizards; these were Plasmodium floridense, Plasmodium azurophilum in erythrocytes, and P. azurophilum in white blood cells. Overall prevalence of infection for 8 samples during the study period was significantly higher for males than females (32% of 3,296 males and 22% of 1,439 females). During the study, the site experienced substantial climatic and physical disturbance including rising temperature, droughts, and hurricanes that severely damaged the forest. Parasite prevalence in the first sample, 8 mo after the massive hurricane Hugo, was slightly, though significantly, lower than for subsequent samples. However, overall prevalence was stable during the 9-yr period. The results show malaria prevalence is more constant at the site than found for 2 studies in temperate forests, and that the Puerto Rico system may be an example of the stable, endemic malaria described by standard models for human malaria epidemiology.

Full text

511
J. Parasitol., 86(3), 2000, p. 511–515
䉷
American Society of Parasitologists 2000
PREVALENCE OF MALARIA PARASITES (PLASMODIUM FLORIDENSE AND
PLASMODIUM AZUROPHILUM) INFECTING A PUERTO RICAN LIZARD
(ANOLIS GUNDLACHI): A NINE-YEAR STUDY
Jos. J. Schall, Anja R. Pearson, and Susan L. Perkins
Department of Biology, University of Vermont, Burlington, Vermont 05405
ABSTRACT
: The prevalence of malaria parasites was studied in the lizard Anolis gundlachi over a 9-yr period at a site in the wet
evergreen forest of eastern Puerto Rico. Three forms of the parasite infected the lizards; these were Plasmodium floridense,
Plasmodium azurophilum in erythrocytes, and P. azurophilum in white blood cells. Overall prevalence of infection for 8 samples
during the study period was significantly higher for males than females (32% of 3,296 males and 22% of 1,439 females). During
the study, the site experienced substantial climatic and physical disturbance including rising temperature, droughts, and hurricanes
that severely damaged the forest. Parasite prevalence in the first sample, 8 mo after the massive hurricane Hugo, was slightly,
though significantly, lower than for subsequent samples. However, overall prevalence was stable during the 9-yr period. The
results show malaria prevalence is more constant at the site than found for 2 studies in temperate forests, and that the Puerto
Rico system may be an example of the stable, endemic malaria described by standard models for human malaria epidemiology.
Among the first mathematical models in ecology were those
of Ronald Ross (1911) who sought to explain the observed
spatial and temporal variation in malaria prevalence among hu-
man populations in malarious regions. Indeed, Ross might well
be regarded as a founder of mathematical ecology (Anderson
and May, 1991). These models became classics in epidemiology
when developed by Lotka (1923) and in modern form by Mac-
donald (1957). Macdonald recognized that at some sites malaria
is absent or rare, at other sites prevalence is fairly low but prone
to sudden epidemic outbreaks (epidemic malaria), and at others
malaria is stable at high prevalence (endemic malaria). The
model shows that the average number of blood meals taken by
the vector(s) will determine if malaria is absent (few meals),
epidemic (more meals), or endemic (many meals) (Aron and
May, 1982). The outlook proposed by Ross, Lotka, and Mac-
donald has long been important in malaria epidemiology (An-
derson and May, 1991) and continues to influence even sophis-
ticated landscape analyses of malaria distribution (Craig et al.,
1999).
Under conditions allowing endemic malaria, environmental
perturbations can lead to changes in prevalence of the parasite
in both human and vector hosts, but the system would rebound
to its original, stable state (Aron and May, 1982). However,
stability models in ecology typically assume only relatively mi-
nor variation over time in the model’s terms (Yodzis, 1989).
Greater intensity of disturbance can lead to unpredictable dy-
namics that fall outside the local stability predicted by the mod-
els. For example, endemic malaria might be expected in the wet
tropics, but even these habitats can experience significant cli-
matic disturbance, both acute due to tropical storms and secular
if the regional climate is altered (Lugo and Scatena, 1996). This
is the issue confronted here. Will prevalence of malaria para-
sites remain stable in a wet tropical system that experiences
such climatic disturbances or long-term changes?
We have examined the prevalence of 2 species of malaria
parasite, Plasmodium floridense and Plasmodium azurophilum,
in their primary vertebrate host, the anole lizard Anolis gundla-
chi in the wet evergreen forest of eastern Puerto Rico. Eight
samples were taken from June 1990 through March 1999. Al-
Received 16 July 1999; revised 9 September 1999; accepted 9 Sep-
tember 1999.
though the tropical wet forest is generally regarded as a stable
environment, the site in Puerto Rico has suffered regular dis-
turbance by direct hurricane hits and periodic drought condi-
tions and perhaps a long-term increase in temperature (below).
This variation in environmental conditions has had a substantial
impact on the forest (Reagan and Waide, 1996; Zimmerman et
al., 1996) as well as the anoles (Reagan, 1991, 1996; Schall
and Pearson, 2000). A preliminary study revealed that preva-
lence of P. floridense and P. azurophilum remained high in the
anole over a 6-mo period (
艐
30%; Schall and Vogt, 1993) sug-
gesting an endemic, stable condition for the parasite–host sys-
tem. We asked if this is possible in the severely disturbed trop-
ical environment in eastern Puerto Rico.
STUDY SITE AND METHODS
The study was conducted at the El Verde Field Station inthe Luquillo
Experimental Forest of eastern Puerto Rico (18
⬚
19
⬘
N, 65
⬚
45
⬘
W). The
habitat is described by Waide and Reagan (1996). In brief, the site is a
wet evergreen forest at approximately 400 m elevation, in rough terrain
of steep slopes. The overall biodiversity is lower than mainland tropical
sites; for example, only 8 species account for 75% of tree density. Eight
species of Anolis occur at the site, but only A. gundlachi is commonly
infected (prevalence was
⬍
1% in the other species; Schall and Vogt,
1993), so only A. gundlachi anoles were sampled in our long-term
study. The parasite species that infect A. gundlachi are widespread
throughout the eastern Caribbean (Staats and Schall, 1996).
Lizards were collected along trails cutting through a 36-ha portion of
the field station. Samples were taken during 8 periods: (1) May–July
1990, (2) January 1991, (3) July and August 1996, (4) February 1997,
(5) July 1997, (6) January 1998, (7) May 1998, and (8) March 1999.
The animals were captured by hand or with a slip noose on the end of
a pole. They were kept in mesh sacks until evening when a drop of
blood was extracted from a toe clip to make a thin smear (thick smears
are useless because the erythrocytes of lizards are nucleated). The
smears were stained in Giemsa at pH 7.0 for 50 min. Each lizard was
measured (snout-to-vent length in mm) and gender determined.
Smears were scanned under oil at 100
⫻
for 6 min during which
approximately 10,000 erythrocytes were examined (Schall and Brom-
wich, 1994). For smears positive for infection, the species of parasite
was recorded for 6 samples. Plasmodium azurophilum infects both
erythrocytes and 2 classes of white cells (Telford, 1975; Schall, 1992).
Ayala and Hertz (1981) suggested these could be 2 species of parasite
using different cell classes, and ongoing gene-sequencing studies sup-
port this view (S. Perkins, unpubl. obs.). Therefore, P. azurophilum
infections were characterized by cell class infected and each infection
was scored as P. floridense and P. azurophilum in erythrocytes (RBC)
or P. azurophilum in white blood cells (WBC).
There are 2 potential sources of error emerging from this protocol.

512 THE JOURNAL OF PARASITOLOGY, VOL. 86, NO. 3, JUNE 2000
F
IGURE
1. Top panel: Mean daily low temperature by year at El
Verde, Puerto Rico site from 1980 to 1998 showing general warming
trend during the period of the study from 1990 to the present. Bottom
panel: Mean total rainfall per month for the period 1975–1998, reveal-
ing a seasonal rainy and drier periods.
F
IGURE
2. Total rainfall during the month (30 days) prior to the start
of each sample at the El Verde, Puerto Rico site, and for the 30-day-
period 4 mo prior to the sample. Summer (S) and winter (W) samples
are indicated. Although the wet and dry seasons seen in Figure 1 were
weakly present, there was substantial variation in rainfall during the
study period.
First, infections with very low parasitemia could be missed during the
6-min scan. This underestimation of percentage of hosts infected has
long been known in malaria studies (Macdonald, 1926). This seems not
to be a problem for most studies of lizard malaria because, in contrast
to human malaria, lizard malaria parasites generally produce higher par-
asitemia in their vertebrate hosts that is readily detected during micro-
scopic scanning (Bromwich and Schall, 1986). To test this conclusion,
Perkins et al. (1998) used the polymerase chain reaction (PCR) to detect
infections of another lizard malaria parasite, Plasmodium mexicanum,
in fence lizards (Sceloporus occidentalis) in California. They were able
to detect infections too weak to find during lengthy microscopic ex-
amination of blood smears but found that such infections were rare.
That is, the highly sensitive PCR method was only marginally more
effective in determining parasite prevalence. As the purpose of our
study was to determine changes in prevalence over time, an underesti-
mation of prevalence would be important only if the presence of very
low parasitemia infections differed among sample periods. Second, the
method will underestimate mixed infections when an infection reveals
high parasitemia for only 1 or 2 of the parasite classes being scored.
But again, this would be damaging to our study only if parasitemia
varied in different ways over time for the different parasites.
The lifespan of A. gundlachi anoles is most likely only 1–2yr.Anolis
stratulus, another of the El Verde anoles, has a population turnover of
about 1.4 yr (Reagan, 1996), and A. gundlachi, a larger lizard, probably
lives slightly longer. The rapid turnover of the lizards means that point
estimates of prevalence would reflect any changes in the transmission
biology of the parasite.
Weather data were taken from records maintained by the field station
staff. We used rainfall recorded for the previous 1–12 mo and mean
daily lowest temperature for 1–12 mo prior to each sample period. The
daily low temperature was presumed relevant if the unknown vector(s)
take blood meals at night. Records of hurricanes were extracted from
the U.S. Weather Service web site.
RESULTS
Weather
As expected for a tropical forest, temperature varies only
slightly during the year at the El Verde site. From 1975 to 1998,
mean monthly low temperature differed only 3 C from January
to July. However, there has been a slight warming trend since
the study began in 1990 (Fig. 1). Although rainfall is generally
high year-round at the site, there is a relatively dry period dur-
ing January–April, and a shorter summer dry season in June
(Fig. 1). Figure 2 shows the total rainfall for the month just
prior to the beginning of each of our samples and for the fourth
month prior to our samples. There was a weak seasonal trend
in rainfall, but more striking is the variation in rainfall over the
9-yr period. Two of our samples came during droughts, July
1997 and March 1999. The study began in the summer of 1990
following the 17–19 September 1989 strike by Hurricane Hugo.
Subsequent severe hurricanes included Luis and Marilyn in
September 1995, Bertha in July 1996 during our third sample,
and Georges in September 1998 (Fig. 3).
Prevalence
We sampled 4,735 A. gundlachi (1,439 females and 3,296
males). Prevalence of the parasites (all taxa combined) was
higher for males (
␹
2
⫽
53.8, 1 df, P
⬍
0.0001), so data are
partitioned by gender for subsequent analyses. Prevalence of
malaria for male lizards did not differ by season (winter vs.
summer) (
␹
2
⫽
3.68, P
⬎
0.05). An equivalent analysis for
females is not possible because 2 of the winter samples had
very small sample sizes, so we have only 2 useful winter sam-
ples for females.

SCHALL ET AL.—PREVALENCE OF LIZARD MALARIA PARASITES 513
F
IGURE
3. Total prevalence of malaria parasites for male (top panel)
and female (bottom panel) Anolis gundlachi at the El Verde, Puerto
Rico field site for 8 sample periods. Data combine 3 forms of malaria
parasite, Plasmodium floridense and P. azurophilum in erythrocytes,
and P. azurophilum in white blood cells. Summer samples (S) are con-
nected by solid lines, and winter samples (W) by broken lines. Per-
centage of lizards infected is given with 95% confidence interval for
each sample. Sample sizes, and timing and name of major hurricanes
are also given. Results show stable prevalence of infection over a 9-yr-
period.
F
IGURE
4. Percentage of infected Anolis gundlachi lizards with each
of 3 forms of malaria parasite (Plasmodium floridense and P. azuro-
philum in erythrocytes [RBC], and P. azurophilum in white blood cells
[WBC]). Percentage of infections for the 3 parasite forms will sum to
⬎
100 because many infections contained more than 1 form of the par-
asites. Results show stable prevalence of each of the parasite forms over
a 9-yr-period.
For both males and females, the prevalence varied among
samples (Fig. 3; Male
␹
2
⫽
57.7, 7 df, P
⬍
0.001; Female
␹
2
⫽
18.9, 5 df, P
⬍
0.01). Post hoc cell contribution tests showed
that only the first 2 samples for males and the first sample for
females (the size of the second sample of females was too small
to add to the analysis) contributed to the variation among sam-
ples. Thus, only the samples immediately after Hurricane Hugo
differed, with a lower proportion of the lizards infected.
Separating the 3 parasites reveals that P. azurophilum in RBC
remained the dominant parasite over the 9-yr period (60–80%
of infections), whereas P. azurophilum in WBC and P. flori-
dense remained at about 10–30% of infections (Fig. 4). For
each parasite, data were cast into a contingency table, and re-
sults show no significant difference among samples for P. azur-
ophilum in WBC (
␹
2
⫽
10.1, 5 df, P
⬎
0.05). The other 2
parasites did vary over time (P. azurophilum in RBC
␹
2
⫽
15.6,
5 df, P
⫽
0.009; P. floridense
␹
2
⫽
26.8, 5 df, P
⫽
0.0002).
Post hoc tests show that it is the first sample that leads to this
result; subsequent to that sample, P. floridense became more
common and P. azurophilum less common.
The little difference in malaria prevalence among samples
suggests that neither rainfall nor temperature would be corre-
lated with prevalence of the parasites. This proved to be correct.
No correlation existed for cumulative rainfall from 1 to 12 mo
before the sample and prevalence of the parasites in either male
or female lizards (P
⬎
0.05) or for total rainfall in a single
month from 1 to 12 mo prior to the sample (P
⬎
0.05). Like-
wise, mean low temperature by month or for cumulative months
for 1–12 mo before the sample was not correlated with preva-
lence (P
⬎
0.05).
DISCUSSION
The El Verde site, although a wet evergreen tropical forest,
has experienced substantial environmental variation over the
past decade. Our study began 8 mo after Hurricane Hugo dev-
astated the forest, knocking down most of the larger trees and
stripping away the canopy. Reagan (1996) describes the land-
scape after Hugo as resembling ‘‘a forest of telephone poles
and 3–5 m deep piles of leaf and branch debris.’’ During our
first sample, the remaining trees had grown new leaves, but the
canopy still appeared bare and the forest floor was well lit. Two
anole species normally abundant in the canopy (Reagan, 1996),
A. stratulus (21,500/ha) and Anolis evermanni (1,500/ha), but
normally almost absent near the forest floor, were commonly
seen at the base of trees and on fallen branches (Schall and
Vogt, 1993). The density of A. gundlachi reached only 18% of
prehurricane numbers and recovered to only 35% 13 mo later
(Reagan, 1991). By our third sample in July and August 1996,
much of the canopy had recovered. The situation was once
again reversed after Georges that knocked down few trees but
removed almost all leaves at canopy level. Our last sample was
taken 6 mo after that storm. The trees had cast out new leaves,

514 THE JOURNAL OF PARASITOLOGY, VOL. 86, NO. 3, JUNE 2000
but the canopy still appeared open and the forest floor was again
well lit.
In addition to these physical disturbances to the forest, rain-
fall and temperature also differed among our sample periods.
Monthly rainfall during our study varied from almost none (1–
2 cm) to nearly a meter. Rainfall in the months prior to the
samples varied substantially, even when season is held constant
(Fig. 2). As a last indication of environmental change, the mean
low temperature at El Verde has been rising over the past 8 yr
(Fig. 1). These differences in rainfall and temperature had an
effect on the anoles; Schall and Pearson (1999) found that a
measure of body condition (relative body mass) declined during
cooler and drier periods.
Despite these objective and subjective indications that the
environment at El Verde is unstable, prevalence of Plasmodium
infection in both male and female lizards showed little differ-
ence among the samples. The higher prevalence of malaria par-
asites observed in male A. gundlachi is similar to the pattern
seen in many populations of lizards harboring such parasites,
although the cause is unknown (Schall, 1996).
The relative proportions of the 3 kinds of parasite (P. flori-
dense and P. azurophilum in RBC and WBC) remained con-
stant over the 10-yr period. The prevailing difference in prev-
alence of P. floridense and P. azurophilum in RBC suggests
the 2 parasites may exploit different vectors. If so, then the
stability of prevalence of both species over time would be even
more remarkable, because 2 independent parasite–host–vector
systems could be involved.
The stability of malaria prevalence at El Verde contrasts with
findings from the 2 other long-term studies conducted on lizard
malaria, both from temperate environments with mild winters.
In Georgia, Jordan and Friend (1971) followed P. floridense in
Sceloporus undulatus for 13 yr (1958–1970). Prevalence ap-
peared to have followed a cycle over that time falling from 53%
infected, to 12%, and rising again to 47%. Jordan suggested a
secular trend in rainfall drove the changes. In California, Schall
and Marghoob (1995) during a 13-yr-study found an apparent
cycle of about 10 yr duration of P. mexicanum in S. occiden-
talis. Subsequent additional data for a total of 20 yr of obser-
vations support this conclusion (Schall, 1996; J. Schall, unpubl.
obs.). Prevalence at the California site varied among years from
10% to 37% but was not correlated with any environmental
measure. Schall and Marghoob (1995) suggested the prevalence
of P. mexicanum was following a stable limit cycle that would
result if nonlinearities exist in the relationships between the
parasite, vectors, and lizard.
The high and constant prevalence of lizard malaria at El
Verde suggests that the parasite–host system there would match
the stable, endemic condition pictured by the Macdonald (1957)
epidemiological model. This model is general in its application
(Aron and May, 1982), such that even with likely differences
in vector biology and immune response in lizard versus human
malaria, its conclusions would be relevant for the El Verde liz-
ard malaria system. What allows the important terms of the
model (such as number of lifetime blood meals taken by the
vectors and the density of those vectors) to remain within the
boundaries necessary to allow such stability even in the face of
major habitat changes as observed at the Puerto Rico site? The
Macdonald (1957) epidemiology model suggests that the an-
swer lies in the ecology of the vector(s). The identity of the
vector(s) of lizard malaria at El Verde is unknown, but perhaps
the variable environment has selected for vectors that are un-
affected by changing rainfall patterns or damage to the forest
structure. The vectors of P. mexicanum in California are psy-
chodid sandflies that spend most of their lives in rodent burrows
where the environment is constant and emerge only on nights
when temperature and humidity are appropriate (Fialho and
Schall, 1995; Schall and Marghoob, 1995). Environmental con-
ditions above ground are therefore disconnected from the trans-
mission biology of the parasite. The vectors of P. floridense
and P. azurophilum may have behaviors that buffer them from
droughts or hurricanes and that allow them to rebound in den-
sity very soon after acute habitat disturbance. Our sampling
protocol would thus have missed short-term disruptions in
transmission and the resulting drop in infection prevalence.
Some authors argue that large scale climate changes will alter
the distribution and abundance of parasites of both veterinary
(Baylis et al., 1999) and human public health importance (Rog-
ers and Packer, 1993; Linthicum et al., 1999), including changes
in prevalence of human malaria (Martens et al., 1995). El Nin˜o
events have been associated with increase in malaria prevalence
in South America (Nicholls, 1993; Bouma and Dye, 1997), and
a warming trend has been correlated with increase in malaria
in Africa (Loevinsohn, 1994) and Sri Lanka (Patz et al., 1996).
In contrast, the results from El Verde suggest that substantial
climatic disturbances do not always lead to changes in Plas-
modium prevalence. The causes of this difference in response
to environmental changes presents an interesting problem in
ecological parasitology.
ACKNOWLEDGMENTS
We thank the staff of the El Verde Field Station for their
assistance throughout this project. Helping collect lizards were
S. Vogt, M. McKnight, D. Whitaker, A. Smythe, J. Meisler, J.
Wolf, H. McKinny, B. Reardon, A. Wargo, and C. Bliss. In the
laboratory, we benefited from help with slide scanning by M.
Milas, J. Martin, T. Smith, and A. Wargo. The work was funded
by an LTER grant from NSF and a grant from Vermont
EPSCoR to J.J.S. The last sample was funded by grants from
the University of Vermont HELiX program and the President’s
Office. This study was conducted under an approved protocol
of the University of Vermont animal care committee.
LITERATURE CITED
A
NDERSON
, R. M.,
AND
R. M. M
AY
. 1991. Infectious diseases of humans,
dynamics and control. Oxford University Press, Oxford, U.K., 757p.
A
RON
, J. L.,
AND
R. M. M
AY
. 1982. The population dynamics of ma-
laria. In The population dynamics of infectious diseases: Theory
and applications, R. M. Anderson (ed.). Chapman and Hall, Lon-
don, U.K., p. 139–179.
A
YALA
,S.C.,
AND
P. M. H
ERTZ
. 1981. Malaria infection in Anolis lizards
on Martinique, Lesser Antilles. Revista do Instituto de Medicine
Tropicale Sao Paulo 23: 12–17.
B
AYLIS
, M., P. S. M
ELLOR
,
AND
R. M
EISWINKEL
. 1999. Horse sickness
and ENSO in South Africa. Nature 397: 574.
B
OUMA
, M.,
AND
C. D
YE
. 1997. Cycles of malaria associated with El
Nin˜o in Venezuela. Journal of the American Medical Association
278: 1772–1774.
B
ROMWICH
, C. R.,
AND
J. J. S
CHALL
. 1986. Infection dynamics of Plas-
modium mexicanum, a malarial parasite of lizards. Ecology 67:
1227–1235.
C
RAIG
, M. H., R. W. S
NOW
,
AND
D.
LE
S
UEUR
. 1999. A climate-based

SCHALL ET AL.—PREVALENCE OF LIZARD MALARIA PARASITES 515
distribution model of malaria transmission in sub-Saharan Africa.
Parasitology Today 15: 105–111.
F
IALHO
,R.F.,
AND
J. J. S
CHALL
. 1995. Thermal ecology of a malarial
parasite and its insect vector: Consequences for the parasite’s trans-
mission success. Journal of Animal Ecology 64: 553–562.
J
ORDAN
,H.B.,
AND
M. B. F
RIEND
. 1971. The occurrence of Schellackia
and Plasmodium in two Georgia lizards. Journal of Protozoology
22: 241–244.
L
INTHICUM
, K. J., A. A
NYAMA
,C.J.T
UCKER
,P.W.K
ELLEY
,M.F.M
YERS
,
AND
C. J. P
ETERS
. 1999. Climate and satellite indicators to forecast
Rift Valley fever epidemics in Kenya. Science 285: 397–400.
L
OEVINSOHN
, M. 1994. Climatic warming and increased malaria inci-
dence in Rwanda. The Lancet 343: 714–718.
L
OTKA
, A. J. 1923. Contributions to the analysis of malaria epidemi-
ology, Parts I–V. American Journal of Hygiene, Supplement 3: 1–
121.
L
UGO
,A.E.,
AND
F. N. S
CATENA
. 1996. Background and catastrophic
tree mortality in tropical moist, wet, and rain forests. Biotropica
28: 585–599.
M
ACDONALD
, G. 1926. Malaria in the children of Freetown, Sierra Le-
one. Annals of Tropical Medicine and Parasitology 20: 239–262.
. 1957. The epidemiology and control of malaria. Oxford Uni-
versity Press, London, U.K., 201 p.
M
ARTENS
, W. J. M., L. N
IESSEN
,J.R
OTMAS
,T.J
ETTEN
,
AND
A. M
C
-
M
ICHAEL
. 1995. Potential impact of global climate change on ma-
laria risk. Environmental Health Perspectives 103: 458–464.
N
ICHOLLS
, N. 1993. El Nin˜o—Southern oscillation and vector-borne dis-
ease. The Lancet 342: 1284–1285.
P
ATZ
, J., P. E
PSTEIN
,T.B
URKE
,
AND
J. B
ALBUS
. 1996. Global climate
change and emerging infectious diseases. Journal of the American
Medical Association 275: 217–223.
P
ERKINS
, S. L., S. M. O
SGOOD
,
AND
J. J. S
CHALL
. 1998. Use of PCR for
detection of subpatent infections of lizard malaria: Implications for
epizootiology. Molecular Ecology 7: 1587–1590.
R
EAGAN
, D. P. 1991. The response of Anolis lizards to hurricane-induced
habitat changes in a Puerto Rican forest. Biotropica 23: 468–474.
. 1996. Anoline lizards. In The food web of a tropical rain forest,
D. P. Reagan and R. B. Waide (eds.). University of Chicago Press,
Chicago, Illinois, p. 321–345.
,
AND
R. B. W
AIDE
. 1996. The food web of a tropical rain forest.
University of Chicago Press, Chicago, Illinois, 616 p.
R
OGERS
, D.,
AND
M. P
ACKER
. 1993. Vector-borne diseases, models, and
global change. The Lancet 342: 1282–1284.
R
OSS
, R. 1911. The prevention of malaria, 2nd ed. Murray, London,
U.K., 669 p.
S
CHALL
, J. J. 1992. Parasite-mediated competition in Anolis lizards.
Oecologia 92: 64–68.
. 1996. Malarial parasites of lizards: Diversity and ecology. Ad-
vances in Parasitology 37: 255–333.
,
AND
C. R. B
ROMWICH
. 1994. Interspecific interactions tested:
Two species of malaria parasite in a west African lizard. Oecologia
97: 326–332.
,
AND
A. B. M
ARGHOOB
. 1995. Prevalence of a malarial parasite
over time and space: Plasmodium mexicanum in its vertebrate host,
the western fence lizard Sceloporus occidentalis. Journal of Animal
Ecology 64: 177–185.
S
CHALL
, J. J.,
AND
A. R. P
EARSON
. 2000. Body condition of a Puerto
Rican anole, Anolis gundlachi: Effect of a malaria parasite and
weather variation. Journal of Herpetology (in press).
,
AND
S. P. V
OGT
. 1993. Distribution of malaria in Anolis lizards
of the Luquillo Forest, Puerto Rico: Implications for host com-
munity ecology. Biotropica 25: 229–235.
S
TAATS
, C. M.,
AND
J. J. S
CHALL
. 1996. Malarial parasites (Plasmodium)
of Anolis lizards: Biogeography in the Lesser Antilles. Biotropica
28: 388–393.
T
ELFORD
, S. R. 1975. Saurian malaria in the Caribbean: Plasmodium
azurophilum sp. nov., a malaria parasite with schizogony and gam-
etogony in both red and white cells. International Journal for Par-
asitology 7: 299–314.
W
AIDE
, R. B.,
AND
D. P. R
EAGAN
. 1996. The rain forest setting. In The
food web of a tropical rain forest, D. P. Reagan and R. B. Waide
(eds.). University of Chicago Press, Chicago, Illinois, p. 1–16.
Y
ODZIS
, R. 1989. Introduction to theoretical ecology. Harper and Row,
New York, New York, 384 p.
Z
IMMERMAN
, J. K., M. R. W
ILLIG
,W.R.W
ALKER
,
AND
W. L. S
ILVER
.
1996. Introduction: Disturbance and Caribbean ecosystems. Biotro-
pica 28: 414–423.