SUMMARY
The fluctuations of a Puerto Rican dry forest bird population were sampled by netting over a 5-year period that included drought conditions. The absence of normal May and June rains most affected the resident bird populations that breed during this period. Frugivorous species declined most sharply, while gleaning insectivores declined least. Winter resident gleaning insectivores actually increased as resident insectivores declined. It is suggested that insects may provide a scarce, but predictable, resource that attracts winter residents, and thus, in turn, restricts the breeding season of resident insectivores. Such restriction exposes breeding birds to variable early wet season conditions. The relative unpredictability of the fruit, seed and flower resource seems to limit winter residents that use these foods and releases resident members of these guilds from some of the restrictions in breeding. Despite pronounced population declines, no regular member of the Guanica Forest bird community disappeared during the drought.
Few long-term studies describe population dynamics of tropical avian
communities during normal or abnormal periods. Karr et al. (1982) re-
ported comparable measurements of bird density and community structure
in Panamanian forest over a period of nearly a decade and Willis (1974)
has long-term data for selected species on Barro Colorado Island, Panama.
Although Diamond (1971) and Terborgh and Faaborg (1973) reported sup-
posed cases of turnover (change in species composition over time without
change in the equilibrium species number (see Diamond and May 1977)
in island communities (cases which have been questioned [Lynch and
Johnson 1974]), no one has examined the role of actual population fluc-
tuations in the turnover process.
Here ! report on the population dynamics of a West Indian bird com-
munity over a 5-year period that included severe drought conditions. The
effects of these conditions on total populations, populations of foraging
guilds, and winter resident populations are shown. Possible reasons for
differences among species are discussed, along with the way that such
"ecological crunches" (Wiens 1977) may have led to patterns found in the
Guanica Forest bird community.
STUDY AREA AND METHODS
This study was done in seasonally dry scrub in the Guanica Forest of southwestern Puerto
Rico (for detailed habitat description see Terborgh and Faaborg 1973). The vegetation is
situated on a limestone platform and is generally short and sclerophyllous, with many ar-
boreal cacti.
Population size was determined by mist netting, as described earlier (Terborgh and Faa-
borg 1973). Net lines were operated from dawn to dusk, usually for 3 days. This allowed us
to construct regressions of the capture rate of birds over time and to predict a total capture.
On a couple of occasions only 2 days of netting were done; thus 2-day totals are used.
Addition of the third day only accentuates the difference between good and bad periods, as
the third day of netting during drought periods often yielded fewer than 10 birds. In all cases,
the number of unbanded birds netted declined throughout the sample period, suggesting a
resident population; during the drought period drastic drops in capture rates often occurred
after the first day.
One line of 16 nets was operated in January of 1972. In 1973 a second line was erected
and both lines were operated subsequently in January or early February of each year. The
area was visited and netted in June of 1973, but due to possible seasonal differences in
territorial behavior, these data are not presented here. All birds captured were banded,
measured and released (for longevity records see Faaborg and Winters 1979, 1980), and age
and sex were recorded when possible.
20
Faaborg ß POPULATION FLUCTUATIONS DURING DROUGHT 21
20 1970 /,. "/ 1971
10
J F M A M J J A S O N D J F M A M J J A S O N D
1.72 ......... 1973 .
2ø f
- lO
-I, J F M A M J J A 5 O N D J F M A M J J A 5 O N D
2( 1974 / ///";...., 1975
; ,/ ,,, , ; .
lC i
J F M A M J J a 50 N O J F M a M J J a $ O NI O
MONTH
FIc. 1. Rainfall amounts for Ensenada for the period 1970-1975 (dashed line) with the
normal amount (sohd line). Shaded area designates a net deficit of rainfall during the first
six months of a year. Data are from Calvesbert (1970 and supplements).
Rainfall data are from the U.S. Weather Bureau station at Ensenada, a coastal community
4 miles west of the study area. Average rainfall for this area is characterized by a small peak
during May and a much larger peak in September and October. A dry season usually occurs
from December to April.
The winter bird community can be separated into a permanent resident and winter resident
component. The majority of breeding by permanent resident species, plus a few summer
residents, occurs from late April to June (Wetmore 1916, Bond 1943), although a few species
(mostly frugivores and nectarivores) breed at other times. Most non-breeding species arrive
in September and stay until April, a period that encompasses both the wettest and driest
parts of the year. Virtually all of these winter residents are gleaning insectivores (Faaborg
and Terborgh 1980). Permanent residents were divided into the simple foraging guilds of
frugivore (including both fruits and seeds), flycatching insectivore, gleaning insectivore, and
nectarivore based upon known diets and foraging habits (Wetmore 1916).
RESULTS
Drought conditions.---Monthly rainfall totals for 1970-1975 are shown
graphically in Fig. 1 along with the normal monthly amounts. One year
was near normal, three were below normal (one nearly 40 cm below), and
two were above normal in annual rainfall. However, annual rainfall may
not be the best way to understand the effect of climate on bird populations
in a habitat such as this, one in which potential evapotranspiration nearly
always exceeds rainfall. Rather, the length and severity of the dry season
22 THE WILSON BULLETIN ß Vol. 94, No. 1, March 1982
250
200
150
100
50
72
i I i
73 74 75 76
YEAR
FIG. 2. Population fluctuations of the total resident bird population for 16-net, two-day
samples over five years (dotted line) and 32-net, two-day samples over four years (solid line).
and the occurrence of rains before and during the breeding season may
be critical. The yearly totals are greatly affected by heavy autumn rains
occurring after the breeding season and often associated with tropical
storms.
As shown in Fig. 1, the period 1970-1972 was characterized by normal
or above normal precipitation before and during the breeding season. The
fall rains did not occur to any extent in 1972 and the dry season started
over a month early. Below normal rainfall characterizes 10 months of 1973,
with the first 6 months running less than one-half the normal amount (13.54
cm vs 29.90 cm). While the fall of 1973 was about normal, the first 6
months of 1974 were drier than 1973 (12.17 cm). The drought was broken
by heavy rains in the fall of 1974 and again in the fall of 1975, although
the first 6 months of 1975 were very dry (7.85 cm). Thus, after three
relatively normal years, the breeding seasons of 1973 through 1975 were
characterized by drought conditions, while the dry seasons of 1972-73 and
1973-74 were relatively severe.
Population fluctuations of resident birds.-Fig. 2 shows population fluc-
tuations for the total sampled community for the single netline (1972-1976)
and the combined netlines (1973-1976). (Capture data by species and guild
are listed in the Appendix.) Total populations were high in 1972 and
Faaborg ß POPULATION FLUCTUATIONS DURING DROUGHT 23
120
/
16 NETS
I GI I I I I
72 73 74 75 76
YEAR
\F 32 NETS
I I I I
73 74 75 76
FIG. 3. Population fluctuations by guild for the 16-net, two-day sample over five years
(left side) and the 32-net, two-day sample over four years (fight side). F is frugivore, N is
nectarivore, FI is flycatching insectivore and GI is gleaning insectivore.
peaked in 1973. Observations during June of 1973 revealed no active
breeding; by January of 1974 the total population had declined sharply to
less than 40% of its peak. The 1974 sample showed only about 6% juvenile
birds and a high recapture rate of previously banded birds. A slight pop-
ulation increase occurred in 1975 followed by a slight reduction in 1976.
Fig. 3 shows the population fluctuations of resident birds by guild. The
largest decrease occurred with frugivores, which also showed the greatest
increases in 1973 and 1975. Resident gleaning insectivores declined rela-
tively slowly from the 1973 peak, while flycatching insectivores declined
sharply in 1974 but recovered somewhat in 1975 and 1976. Captures of
nectarivores (the Bananaquit and two hummingbirds) decreased in 1973,
probably as a result of a lack of flowers resulting from the early dry season.
This decline continued throughout the remainder of the study. Humming-
birds had virtually disappeared from the Guanica Forest by 1975 and Ba-
nanaquits behaved largely as a gleaning insectivore during the drought.
The greatest single-species declines were also shown by frugivores. The
Common Ground Dove fluctuated from 54 to 2 individuals during the crash
while Puerto Rican Bullfinches dropped from 49 to 16 and the Black-faced
24 THE WILSON BULLETIN' Vol. 94, No. 1, March 1982
70
60
50
40
30
20
10
73 74 75 76
YEAR
FIG. 4. Population fluctuations of winter residents (dotted line) and permanent resident
insectivores (gleaners and flycatchers, solid line) with the total density of wintering insecti-
vores (dashed line).
Grassquit from 29 to 6. Two of the largest frugivores (the Pearly-eyed
Thrasher and Red-legged Thrush) actually increased in numbers during
the severest part of the drought.
The flycatching Puerto Rican Tody declined steadily and the Stolid Fly-
catcher decreased by half in 1974 but recovered in 1975 and 1976. Among
gleaning insectivores, only the Puerto Rican Vireo showed a substantial
drop (7 to 2), while others declined slowly. The Bananaquit and Antillean
Mango demonstrated similar decreases in 1974, but the hummingbird con-
tinued to decline in subsequent years, whereas the more generalized Ba-
nanaquit remained relatively stable in numbers.
Winter resident population fluctuations.--Populations of winter resi-
dents fluctuated in the opposite direction of resident bird populations (Fig.
4). The total of resident insectivores (flycatchers and gleaners) and winter
residents indicates an almost constant population of insect eating birds.
Combining resident flycatchers and gleaners and comparing them with the
winter residents is not unreasonable, as several of the winter residents,
particularly the American Redstart, also flycatch; some overlap in re-
source use must occur. (Adding the partially insectivorous Bananaquit to
Faaborg ß POPULATION FLUCTUATIONS DURING DROUGHT 25
this group would give combined densities of 121, 96, 97 and 87 for
1973-1976, a particularly stable figure for the last 3 years when nectarivory
could not have been a major factor for the Bananaquit.)
Numbers of the five dominant winter resident species (Parula Warbler,
Prairie Warbler, Black-and-white Warbler, American Redstart and Oven-
bird) fluctuated considerably from year to year. During the last 2 years of
the study when resident populations were lowest, five more species were
netted. While the Prothonotary Warbler, Hooded Warbler and Northern
Waterthrush were netted only once, 10 Cape May Warblers were caught
in 1975 and 1976 and five Indigo Buntings were captured in 1976. Neither
of these species had been previously seen in the Guanica Forest.
DISCUSSION
With the observation that most Puerto Rican birds breed during the
May-June rainy season, it is not difficult to see how lack of rainfall during
the first half of the year could drastically affect breeding success and thus
population levels. The occurrence of autumn rainfall, even if heavy, may
be of relatively little consequence if breeding season conditions are poor.
Of course, an extended dry season can have an effect on survivorship, but
good conditions in autumn will not help recruitment to the population.
The data suggest that the insectivorous guilds, particularly gleaning
insectivores, are least affected by drought conditions, while the frugivores,
although able to attain high densities in good conditions, are very sensitive
to drought. This difference could be explained by the resources involved.
Frugivores feed upon what is basically the yield of primary productivity;
with the absence of water there is probably an almost immediate decline
in the amount of fruit and seeds produced. In contrast, insectivores feed
upon the yield of secondary productivity; even in years when few seeds
are produced there are leaves on which insects can feed. Insects may
emerge from pupae or hatch in dry conditions following favorable periods
and they may linger after conditions become unfavorable. While little is
known about island insect densities (Janzen 1973), during the dry season
in Costa Rica even the driest locations maintained moderate densities of
insects (Janzen and Schoener 1968). Thus, although insects may be rela-
tively scarce, they may be somewhat buffered from climatic fluctuations
and birds feeding on insects may show more stable populations than fru-
givores or nectarivores.
This difference between fluctuations in insectivorous and non-insectiv-
orous guilds has been used to explain the almost total dominance of glean-
ing insectivores among winter resident species (Faaborg and Terborgh
1980). A bird population attempting to winter within the confines of an
island must have a predictable resource. While the availability of fruits,
26 THE WILSON BULLETIN' Vol. 94, No. 1, March 1982
seeds and flowers may be too variable, the insect resource is stable enough
to support winter residents.
Variations in the time required by the different resources to respond to
the onset of the wet season could accentuate the differences between guild
characteristics. While many plants may flower and fruit rather rapidly
after rains (often using stored energy), many insects may require a longer
period of time to go through life-cycles and achieve larger sizes and higher
densities. Janzen and Schoener (1968) found reduced proportions of sub-
adult insects in the driest sites sampled during the dry season in Costa
Rica. Fewer young insects during this period must delay insect population
growth during favorable conditions.
The above climatic and resource patterns suggest that it might be best
to breed later in the year as rains are then more predictable. This may
not be possible due to the high densities of winter resident insect gleaners
that arrive in September. Attempting to compete with these winter resi-
dents and breed may be impossible, so residents must use the early wet
period when conditions are more variable but competition is reduced (see
Keast and Morton [1980] for more evidence of this type). Following this
logic, frugivores and neetarivores should not be as constrained and the
data from Wetmore (1916) suggest this is so. The extent to which these
latter species feed insects to their young may put some constraints on
their breeding season. A final complexity to the above situation is added
by the fact that three other insectivorous species (the Black-whiskered
Vireo [Vireo altiloquus], Yellow-billed Cuckoo [Coccyzus americanus] and
Common Nighthawk [Chordeiles minor]) breed in the Guanica Foreat at
the same time as the permanent residents. This complex set of seasonal
movements suggests many interesting questions, but more data on re-
source distributions, philopatry of migrants, and other factors will be need-
ed before any answers can be offered.
Although many resident species declined sharply, no regular member
of the Guanica Forest bird community disappeared during this drought.
So, while the above data show some interesting responses to severe con-
ditions, this "ecological crunch" was not sufficiently severe to cause even
the local extinction of a regular community member. Perhaps it is the
occurrence of droughts and other climatic variations such as this or worse
that (1) has selected for the proper family composition on these islands
(Terborgh 1973), (2) has dictated a level of species saturation in each
habitat (Terborgh and Faaborg 1980) and (3) has resulted in rules for struc-
turing the guilds composing these communities (Faaborg 1982). In view
of the presumed antiquity of most Puerto Rican residents, it would have
been surprising to find local extirpation caused by a 3-year drought. A few
uncommon species that are common in other habitats on the island (Puerto
Faaborg' POPULATION FLUCTUATIONS DURING DROUGHT 27
Rican Emerald, Stripe-headed Tanager) did seemingly disappear, but it
appears that previous climatic and resource factors have selected a set of
species that can coexist through these stress periods when resources are
limited and competition is important (Wiens 1977, MacArthur 1972:21).
During more favorable conditions, each species may increase in density
in response to the types of resources it best uses, probably with less
interaction with other community members. During these periods species
from other habitats may temporarily invade the Guanica Forest and winter
residents may face increased competition from resident insectivores. The
members of this community seem well adapted to such severe conditions,
thus island turnover must in fact be a rare event in the West Indies.
ACKNOWLEDGMENTS
Financial support for visits to the Guanica Forest was provided by the Frank M. Chapman
Fund of the American Museum of Natural History, the Research Council of the Graduate
School, University of Missouri-Columbia, and an NSF pre-doctoral grant. I thank Jose Colon
of the U.S. Weather Bureau, San Juan, Puerto Rico, for providing rainfall data. The Puerto
Rican Department of Natural Resources kindly gave permission and support for the research.
James Bond, Joe Wright, Wayne Arendt, Janice Winters, James Karr and Cameron Kepler
all made comments on versions of the manuscript.
LITERATURE CITED
BONO, J. 1943. Nidification of the passerinc birds of Hispaniola. Wilson Bull. 55:115-125.
---. 1971. Birds of the West Indies, Houghton-Mifflin, Boston, Massachusetts.
CALVESBERT, R. J. 1970. Climate of Puerto Rico and U.S. Virgin Islands. Climatography
of the United States No. 60-52. U.S. Govt. Printing Office, Washington, D.C.
DIAMOND, J. M. 1971. Comparison of faunal equilibrium turnover rates on a tropical island
and a temperate island. Proc. Natl. Acad. Sci. 68:2742-2745.
AND R. n. MAY. 1977. Species turnover rates on islands: dependence on census
interval. Science 197:266-270.
FAABORG, J. R. 1982. Trophic and size structure of West Indian bird communities. Proc.
Natl. Acad. Sci. USA 1979. In Press.
28 THE WILSON BULLETIN ß Vol. 94, No. 1, March 1982
-- AND J. E. WINTERS. 1979. Winter resident returns and longevity and weights of
Puerto Rican birds. Bird-Banding 50:216-223.
-- AND J. E. TEaBOaCH. 1980. Patterns of migration in the West Indies. Pp. 157-163
in Migrant birds in the neotropics: ecology, behavior, distribution, and conservation (A.
Keast and E. S. Morton, eds.). Smithson. Inst. Press, Washington, D.C.
-- AND J. E. WINTERS. 1980. More returns from the Guanica Forest, Puerto Rico. J.
Field Ornith. 51:368.
JANZEN, D. H. 1973. Sweep samples of tropical foliage insects: effects of seasons, vegetation
types, elevation, time of day, and insularity. Ecology 54:687-708.
-- AND T. W. SCHOENER. 1968. Differences in insect abundance and diversity between
wetter and drier sites during a tropical dry season. Ecology 49:96-110.
KAaa, J. R., D. W. SCHEMSKE AND N. BaO
Faaborg' POPULATION FLUCTUATIONS DURING DROUGHT 29
APPENDIX
TWO-DAY CAPTURE TOTALS BY SPECIES AND GUILD FOR THE COMBINED 32 NETS
1973--1976 AND THE 16 NET LINE 1972-1976 (IN PARENTHESES) a
Yearly captures
2 weight
Species (g) 1972 1973 1974 1975 1976
Zenaida Dove (Zenaida aurita)
Common Ground Dove
(Columbina passerina)
Northern Mockingbird
(Mimus polyglottos)
Pearly-eyed Thrasher
(Margarops fuscatus)
Red-legged Thrush
(Mimocichla plumbea)
Blue-hooded Euphonia
(Euphonia musica)
Stripe-headed Tanager
(Spindalis zena)
Puerto Rican Bullfinch
(Loxigilla portoricensis )
Black-faced Grassquit
(Tiaris bicolor)
Yellow-faced Grassquit
(Tiaris olivacea )
Total frugivores
Flycatching insectivores
Puerto Rican Tody (Todus mex-
Stolid Flycatcher
(Myiarchus stolidus)
Caribbean Elaenia
(Elaenia martinica)
Grey Kingbird
(Tyrannus dominicensis)
Total flycatchers
Gleaning insectivores
Puerto Rican Lizard Cuckoo
(Saurothera vieilloti)
153.0 -- -- 1 (1) -- --
35.4 (5) 54 (40) 2 (1) 7 (2) 14 (10)
43.8 -- 1 (1) -- -- --
100.7 (4) 3 (1) 5 9 (1) 10 (5)
74.5 (4) 2 (1) 5 (3) 8 (3) --
15.7 .... 1
29.7 -- 7 (2) 1 -- --
32.1 (27) 49 (36) 16 (12) 15 (7) 8 (2)
9.7 (10) 29 (24) 6 (6) 11 (11) 15 (14)
7.0 .... 1 (1)
-- (50) 145 (105) 36 (23) 50 (24) 49 (32)
5.4 (3) 12 (6) 5 (1) 2 (1) 2 (1)
22.9 (14) 18 (13) 10 (5) 19 (12) 21 (10)
22.0 -- -- -- 1 (1) --
47.6 -- -- -- 1 (1) --
-- (17) 30 (19) 15 (6) 23 (15) 23 (11)
77.1 -- 2 (1) -- -- 1 (1)
30 THE WILSON BULLETIN ß Vol. 94, No. 1, March 1982
APPENDIX
CONTINUED a
2 weight
Species (g)
Yearly captures
1972 1973 1974 1975 1976
Puerto Rican Vireo
(Vireo latimeri)
Adelaide's Warbler
(Dendroica adelaidae)
Black-cowled Oriole
(Icterus dominicensis)
Troupial (Icterus icterus)
Total gleaning insectivores
Nectarivores
Puerto Rican Emerald
(Chlorostilbon maugaeus)
Antillean Mango
(Anthracothorax dominicus
Bananaquit (Coereba fiaveola
Total nectarivores
Winter residents
Black-and-white Warbler
(Mniotilta varia)
Prothonotary Warbler
(Protonotaria citrea)
Parula Warbler (Parula americana)
Cape May Warbler (Dendroica
tigrina )
Prairie Warbler (D. discolor)
Ovenbird (Seiurus aurocapillus)
Northern Waterthrush
(S. noveboracensis)
Hooded Warbler (Wilsonia citrina)
American Redstart
(Setophaga ruticilla)
Indigo Bunting
(Passerina cyanea)
Total winter residents
Total captures
11.2 -- 7 (3) 2 2 (1) --
6.7 (3) 7 (5) 9 (4) 3 (2) 3
37.2 -- 1 1 2 --
72.2 (3) 4 (2) 2 (2) 8 (6) 4 (4)
-- (6) 21 (11) 14 (6) 15 (9) 8 (7)
2.8 (1) ....
5.4 (8) 11 (8) 6 (5) 2 (2) 2 (2)
9.4 (21) 55 (13) 34 (8) 36 (21) 21 (12)
-- (30) 66 (21) 40 (13) 38 (23) 23 (14)
9.7 (5) 6 (2) 7 (2) 2 4
13.0 .... 1
7.4 (2) -- 5 4 (1) 8 (1)
10.0 -- -- -- 7 3
6.9 (3) 2 (2) 11 (4) 7 (5) 6 (4)
18.7 -- 2 3 1 1
15.0 .... I (1)
11.0 -- -- -- 1 --
7.2 (1) 5 7 1 (1) 6
13.2 .... 5 (5)
-- (11) 15 (4) 33 (6) 23 (7) 35 (11)
-- (114) 277 (160) 138 (54) 149 (78) 138 (75)
Nomenclature is from Bond (1971) and weights are from Faaborg and Winters (1979) who list statistical data.