EFFECTS O F F O R E S T F R A G M E N T A T I O N ON A M A Z O N I A N DNDERSTORY B I R D C O M M U N I T I E S .
Richard 0 . B i e r r e g a a r d , J r . (*) Thomas Ε . Lovejoy (**)
SUMMARY
Data faom an intensive miit-netting monk.-necaptune pfiogiam in the centnal Amazon
demonstrate significant changes In the undehStor.y avian community -in iò olated patches of
/ and 10 ha of teAAa fifune forest. Following isolation, capture rates increase signifi
cantii/ ah birds fleeing the. Celled forest entered the. newly fom\ed forest fragments,.
Movement to and from the. reserve tò restricted, ai> witnessed by an increase in recapture
percentages fallowing isolation. Species of birds that axe. obligate, anmy ant followers
disappeared at the. time. the. surrounding habitat WOO removed from 1- and 10-ha areas. The
complex mixed-species insectivorous flocks typical of Amazonian forests deteriorated
yiithin 2 years of iòolation of 1 - and 10-ha forest fragments. Several species of mid-story
insectivores changed their foraging behavior after iòolation of mail forest reserves.
INTRODUCTION
Th e Biological Dynamics of Forest Fragments project (Minimum Critical Size of Ecosystems) is a multidisciplinary attempt to understand the effects of fragmentation of Amazonian tropical rainforest. Through systematic surveys of a broad range of organisms in different sized tracts of forest before and after isolation by deforestation of sur-rounding areas for cattle p a s t u r e s , w e are studying the ecological changes associated l^iith the fragmentation of tropical f o r e s t s . I ns ights ga i ned may 1 ead to methods of managing
reserves that maximize species carrying capacity in areas scheduled for deforestation. T h e principal question addressed by the project is not whether or not there w i1 1 be ecosystem decay associated w i t h forest fragmentation (which w e take as g i v e n ) , but rather, to what extent does the decay depend on interspecific r e l a t i o n s h i p s , on specific autecolo jical requirements or on microhabitat requirements. T h e importance of stochastic e v e n t s , including sampling effects at the moment of isolation and the vagaries attendant on small populations (Soule, 1 9 8 0 ; Franklin, 1 9 8 0 ; Lovejoy & O r e n , 1 3 8 1) will also be assessed.
There is general agreement that in order to maintain viable populations of species
(*) World Wildlife Fund 1250 24 th S t . N W , W a s h i n g t o n , DC 20037 - E U A . (**) Smithsonian I n s t i t u t i o n , W a s h i n g t o n , DC 2 0 5 6 0 .
with large territorial requirements (principally large p r e d a t o r s ) , Amazonian forest re serves should be on the order of hundreds of thousands of hectares (Terborgh, 1 9 7 5 ) . However, there is little doubt that much of the remaining f o r e s t s w i l l be fragmented into areas much smaller.
This study is the first to quantitatively sample the avian community in small forest tracts before and after fragmentation from surrounding vi rgin forest. Although the data are from very small reserves (l and 10 h a ) , it is hoped that our analyses wi 11 provide insights into processes which might otherwise be obscured or might be much longer in becoming apparent in studies of larger tracts. Future analyses of post-isolation data from 100- and 1000-ha reserves will show the extent to w h i c h the phenomena documented in smaller reserves may apply to larger a r e a s .
A program of periodic sampling in the project's study-tracts before and after their isolation provides the quantitative basis for an assessment of the effects of isolation on this part of the ecosystem. Data from mist-net captures of understory birds from isolated forest reserves w e r e compared to pre-isolation data frcm the same reserves to investigate changes related to isolation of the former virgin forest tracts. Birds of the forest understory probably are particularly susceptible to the effects of fragment ation of virgin forest because physiological and behavioral contraints limit or preclude their movement through areas w h e r e no forest cover is a v a i l a b l e . The 1 arge number of uncter story birds w h o s e ranges are delimited by rivers (Meyer de S c h a u e n s e e , 1966; S i c k , 19&7) reflects this tendency.
Comparisons w e r e made of isolated forest reserves of d i f f e r e n t s i z e s (l and 10 ha) to test the effect of size on the resultant avian community c o m p o s i t i o n . As well as com parisons o f p r e - and post-isolation data from the isolated patches t h e m s e l v e s , post-iso lation data were compared to n e a r - b y , non-isolated forest plots studied simultaneously to control for the effects of temporal variation in avian community s t r u c t u r e .
METHODS
The study areas are situated on nutrient p o o r , clayey or sandy yellow latosols (Chauvel, 1982) 70 km north of M a n a u s , Brazil (Fig. 1 ) . T h e average annual rainfall is approximately 2200 mm and is st-ongly s e a s o n a l , with the dry season extending from June through Octover and the peak oJ the rainy season in February and March (Anon., 1 9 7 8 ) .
The forest canopy is about. 30-40 m h i g h , with a few emergent trees towering up to 55 m above the forest f l o o r . T h e canopy is even and c l o s e d , permitting little sunlight to penetrate to the rather open understory, w h i c h is characterized by many stemless palms (Guillaumet & K a h n , 1 9 8 2 ) . The woody plant community is diverse as many as 2 5 0 ; woody species have been recorded in a single 10 ha plot (Rankin et a l . , 1 9 8 8 ) . The p r e dominant woody plant families are B u r s e r a c e a e , S a p o t a c e a e , L e c y t h i d a c e a e , and Legumino-sae.
The isolation of the reserves w a s the result of deforestation for cattle pastures
by three large firms investing in the Manaus free trade z o n e . The fi rst step i η the process w a s the clearing of the understory so that chain sawyers had access to canopy t r e e s . Understory clearing began in May or J u n e , w i t h the cha i η saws arr i ving about a month later. Depending on the area being c l e a r e d , felling continued into August and sometimes into O c t o b e r . The felled area w a s burned as late as possible in the dry s e a s o n , usually in the end of O c t o b e r .
Six isolated reserves w e r e studied, varying in size and the time at which they w e r e isolated from the surrounding f o r e s t . One 1-ha and one 10-ha reserve w e r e isolated
in I 9 8 O , w h i l e two 1-ha and two 10-ha reserves w e r e isolated in 1983- Fourteen sites in undisturbed forest provided control data.
Distance to continuous forest varied from reserve to reserve. T h e 1980 10-ha iso-« late at its closest point is slightly less than 100 rq from a nanrow penn i nsula of stream, forest, itself less than 100 m w i d e . On all other sides it is m o r e than 500 m to the nearest continuous forest. S i m i l a r l y , the 1980 I-ha isolate is roughly 80m from continu ous forest on its west side and hundreds of m e t e r s to virgin forest on the other s i d e s . One of the 1983 10-ha isolated, reserve 12 0 7 , is isolated on the south by a 500 ha clearing and on the remaining three sides by only 100 m from virgin f o r e s t . All other reserves i sq
lated in 1983 are isolated by at least 200 m.
All reserves w e r e sampled with transects of tethered mist-nets (EBBA type A T X , 12 m χ 2 m , 36 mm m e s h ) once every three or four w e e k s , wi th o c c a s i o n a1 breaks in the rou
tine of up to six m o n t h s . Post-isol ation reserves w e r e sampled w i t h high p r i o r i t y , so that few long gaps exist in post-isolation d a t a . One-ha reserves w e r e sampled w i t h an 8 net line and 10-ha reserves w i t h a line of 16 n e t s .
Nets w e r e opened from dawn (about 0600 h) until 1400 h. All bi rds c a p t u r e d , except h u m m i n g b i r d s , were banded with uniquely numbered aluminum bands and released at the point of c a p t u r e . Hummingbirds w e r e marked by cutting the tips of tail feathers in unique c o m b i n a t i o n s . In a d d i t i o n , all birds seen or heard by the banders iη the reserve on e a c h banding day were recorded.
T h e percentage of captures on a given day of birds already banded w e r e calculated for all non-isolated reserve n e t - l i n e s . An asymptotic regression (Snedecor & C o c h r a n , 1 9 6 7 ; Patterson, I960) w a s fit to these " p r e - i s o l a t i o n " data to obtain £he best fit curve describing the relation between the number of days a given reserve had been sampled and the expected proportion of recaptures on that sample d a y . Recapture percentages w e r e analyzed from the 1980 isolates and compared to values predicted by pre-isolation d a t a .
Capture rates from both 1980 and 1983 isolated reserves w e r e compared before and after isolation to measure the effect of isolation on popu 1 at ion s i ze and/or avian act iv_i_ ty in the recently isolated reserves.
The relative abundances of a group of 25 of the more abundant species in the 1980 isolates (1 and 10 ha) w e r e compared to pre-isolation values from the same reserves to de_ termine what ecological guilds were most susceptible to the effects of isolation. The species selected w e r e the most abundant representatives of the following g u i l d s ; o b l i gate army ant followers (Willis, 19 6 7 ) , mixed-spec i es fl ock spec ies , frugivores, t ree-fal I
s p e c i a l i s t s , hummingbirds and a general category of abundant insectivorous species occu^ pying small territories throughout the study a r e a .
Capture rates w e r e a l s o calculated for each species from non-isolated reserves and in each size class of isolated reserves (both 1980 and 1983 isolates) beginning approxj_ mately 6 months after isolation (l J a n . of the year following i s o l a t i o n ) . For the species
in the guilds mentioned a b o v e , these rates w e r e compared to the capture rates from undis^ turbed forest s a m p l e s .
RESULTS
ι
From October 1 9 7 9 through 5 January 1 9 8 5 , 1 9, ^ 9 ^ captures of 1 0 , 9 1 8 individuals o f 1 ^ 1 species w e r e recorded in non-isolated reserves (see A p p . 1 ) . In isolated reserves an additional 2 , 6 7 3 captures w e r e recorded in isolated forest fragments o f 1 and 10 h a . Four species not encountered in the non-isolated reserves were captured iη the isolates
(see A p p . 1 ) .
The log-transformed rank abundance distribution from undisturbed forest (Fig. 2 b ) w a s linear, with only the most common s p e c i e s , the Whiteplumed A n t b i r d , Pithys a l b i -frons, (7.66¾ of individuals banded, see Fig. 2 a , A p p . 1 ) deviating slightly from the linear trend seen among the other 1*»0 s p e c i e s . This linear distribution indicates the
« 2
§ s
ja -a
ο —
Φ £> -ο
ο
Φ U C CO•σ
c
3 η <οω
>
(Λ ΌCD
Φ w Φο
8
7 ·
6
5
4
3
21
Ο F i g .2a
• • • • • • « • • • ^10 15 2 0 2 5 3 0 3 5 4 0 4 5 1 4 3
S p e c i e s r a n k by abundance
Fig- 2b
( a b u n d ) = 1 . 5 6 - . 0 5 * rank
2 0
4 0 6 0 8 0 100 1 2 0 140
160
S p e c i e s rank bg abundance
F i g . 2 . T h e rank abundance distribution of b i r d s netted in the understory of undisturbed terra firme forest north of M a n a u s , B r a z i l . 2 a ) The relative abundance of each s p e c i e s , a s represented by individuals b a n d e d , plotted in decreasing order for the 143 species encountered in the netting s a m p l e . 2b) A semi-log representation of the rank abundance distribution presented in F i g . 2 a .
absence of a single dominant species or group of species and a long tail of very rare species. Following Karr ( 1 9 7 1 ) , if w e define as rare any spec i es represented by less than of 2¾ o f individuals banded, 1 2 7 s p e c i e s , or 9 0 % of the community, a r e rare. T h e large number of rare species is typical of tropical forests (Karr, 1 9 7 1; Lovejoy,1975; P e a r s o n , 1 9 7 7 ; W o n g , 1 9 8 6 ) .
In the non-isolated control sites the 25 most abundant species represented 71.7¾ of all captures and 67.3¾ of all birds banded. T h e 2 5 t h most abundant spec ies represented only 1.47¾ of all birds banded. Of the remaining 116 s p e c i e s , 18 w e r e recorded from a single capture.
T h e understory avian community in the Manaus area is predominantly insectivorous; 79.2 of the individuals banded (67.96¾ by biomass) and 83.33¾ of al 1 captures a r e strictly or predominantly insectivorous (Table l ) . Obligate army ant fol1owers and regular mixed species flock followers comprise 37.43¾ of the individuals captured (32.13¾ by biomass) and 41.06¾ of captures in non-isolated reserves (Table 1 ) .
Frugivores accounted for only 14.7¾ of the i nd i vidua 1 s b a n d e d , and 12.7¾ of captures in undisturbed forest. H o w e v e r , because of the large size of several species in the g u i l d , biomass the group represented 23.4¾ of the total biomass (Table l ) .
The most common frugivore, the White-crowned Manakin (Pipra p i p r a ) , w a s the fourth most commonly banded species (4.49¾ of all individuals , 2.1¾ by biomass) . The Ruddy Q u a i l -dove (Geotrygon m o n t a n a ) , although not numerically dominant in the sample (the 26th most abundant species, representing 1.47¾ of b a n d i n g s ) , w a s by biomass the most important species in our sample from undisturbed forest, si ightly ahead of the Wh i te-pl umed Antbird
(6.53¾ v s .
6.mriÃpp.
D .Capture rates from undisturbed, virgin forest net 1ines decl ine gradually over time according to the relation y = .227 - .06 χ where y is the capture rate/net-hour and χ is
the number of days since the first sample in the course of a monthly sampling regime (Bierregaard & Lovejoy in p r e s s ) . W h i l e capture rate declined slowly, the percent r e capture on a given sample day approaches an asymptote of 4 2 . 7 % according to t h e equation y = .427 - ·429 ( · 3 6 4 )κ w h e r e y is the recapture percentage and χ is the days since first
sample as above (Bierregaard & Lovejoy in p r e s s ) .
Capture rates and recapture percentages in i sol ated reserves d i ffered marked 1 y from those in undisturbed forest. As deforestation proceeded around the I98O isolates, the most dramatic effect witnessed in the understory avian community w a s a dramatic increase
in avian activity, as reflected quantitatively by increased capture rates (see Fig. 3)i and in the amount of avian vocalizations heard by the banding crews working in the reserves. T h e mean capture rate during the 7 capture days before isolation w a s not sig_ nificantly different between the 1 - and 10-ha reserves isolated in 1 9 8 0 , but was sίgϊηϊfJ_
cantly higher on the 7 capture days fol lowing isolation in both reserves (Student's t-test for paired o b s e r v a t i o n s , 1- h a , p< . 0 5 ; 10-ha ρ < . 0 1 , Table 2 ) . T h e increase in activity in the smaller, 1- h a reserve was greater than in the 10-ha ( p< . 0 5 , Table 2 ) .
After six m o n t h s , capture rates declined precipitously to below pre-isolation rates in the 1- h a area and roughly to pre-isòlation levels in the 1 0- h a reserve.
Replication of this pattern w a s seen in all but o n e o f the 1983 isolates (Fig. 3 ) . The one exception to the general tendency w a s reserve 1 2 0 7 , w h i c h w a s surrounded on three sides by continuous forest only 100 m a w a y . A s in 1 9 8 0 , the I ha reserve isolated in I983 showed greater increases in capture rates than the 1 0 ha reserves.
L o C \ CO <D L_ -4-J C L O . 7 0 . 6 0 . 5 0 . 4 0 . 3 0 . 2 0 . 1 0 0.0 .40 . 3 0 . 2 0 10 0-0 • - • - • ) 1-ha i s o l a t e s ι—π J N o n - i s o l a t e s \ ι
Urn
<Va
/cP
f - r i L A • • - 4 0 0 - 2 0 0 0 2 0 0 4 0 0 6 0 0 8 0 0Days since isolation
} 1 0 - h a i s o l a t e s
• - • N o n - i s o l a t e s
. 3 .
•I:Vr .-.
- 2 0 0 0 2 0 0 4 0 0 6 0 0 8 0 0 1 0 0 0
Days since isolation
F i g . 3 . Capture rate (birds/net-hour) as a function of time in three 1-ha reserves(above) and in two 10-ha reserves (below) before and after isolation (indicated by the vertical l i n e ) . Expected values are shown as calculated from net-lines in c o n tinuous forest. Points are three-point running m e a n s .
Recapture rates also changed after the isolation of a small (1- or 10-ha) reserve! (Fig. 4 ) . Coincidental with the increased capture rates, the percentage of previously banded birds captured on a given day decreased for about three months after isolation began (Fig. 4 ) . Following this decrease, recapture rates climbed rapidly to somewhere between 60 and 90%,considerably above the asymptote o f 4 3 % in non-isolated reserves.
CD
L_
Z3
-é->
CL
CS
u CD c CDu
CDCl
8 0
7 0
6 0
5 0
4 0
3 0
2 0
10
0.0
9 0
8 0
7 0
6 0
5 0
4 0
3 0
2 0
10
0.0
o-o -»
• - • 11-ha isol.
• - • Non-isol.
/ • la
• \ i M \
hi
:1/7
/ • la
• \ i M \
- 4 0 0 - 2 0 0 0 2 0 0 4 0 0 6 0 0 8 0 0
Days since isolation
O - o
ΛS-S MO-haisol
A_
AJ
• - • Non-isol.
• ^ ¾
3•'JesptikAi
mm
[•
•
•
" °l
f 'rf
• ^ ¾
3•'JesptikAi
mm
- 1 0 0 0 - 5 0 0 0 5 0 0 1 0 0 0 1 5 0 0
Days since isolation
P i g . 4 . Recapture rates as a function of time in 1- and 10-ha isolated reserves. T h e vertical lines marks the time of isolation for the two sets of reserves. T h e expected v a l u e s , calculated from net-lines in undisturbed forest, approaches an asymptote of slightly over 4 0 % . Points are three-point running m e a n s .
Subsequent to isolation of the r e s e r v e s , resprouting from fel led trees and an e x u berant flush of second growth species connected the reserves to the adjacent extensive forest. T w o years after isolation, capture rates in thetwo 1980 isol ates increased s o m e -what and recapture rates dropped sharply. T h e s e changes iη avian act ivi ty w e r e witnessed when the second growth attained a height sufficient to have a heavily shaded understory (approx. 3-4 m ) · Individuals of understory species that had d i s a p p e a r e d , such as the Ferrug_i_ nous-backed Antbird (Myrmeciza f e r r u g i n e a ) , or had decreased in frequency in the n e t s , Fuch as the White-crowned M a n a k i n , began entering the reserves. I η the case of the Ferrug_i_ nous-backed A n t b i r d , the recolonization of the reserve w a s by the last individual of the species caught in the reserve over 18 months previously.
Different ecological guilds responded differently to the isolation of the 1980 re serves (Fig. 5 ) . In a d d i t i o n , species within guilds responded differently to isolation. The fastest and most dramatic change w a s thé:disappearance of the obi igate army ant fol lowing
s p e c i e s . This group virtually disappeared within two capture days of isolation. The disappearance of the Formicarid members of the g u i l d , the White-plumed Antbird and the Rufous-throated Antbi rd (Gymnopithys rufigula) w a s vi rtual ly complete and p e r m a n e n t . Since isolation, only 2 individuals of the Rufous-throated Antbird have been netted in isolated reserves, and these only in the 10-ha reserve isolated in three sides by only 100 m. In c o n t r a s t , other species that follow army ant s w a r m s , the Whi te-chinned Woodcreeper (Den-drocíncla m e r u l a ) and 3 much rarer, large D e n d r o c o l a p t i d s , the Plain-brown(Dendrocincla f u l i g i n o s a ) , the Barred (Dendrocolaptes c e r t h i a ) and the Blackbanded (D. picumnus) w o o d
c r e e p e r s , have been netted more than 25 t imes (17 for the mpre general ist Plainbrown W o o d -creeper) i n b o t h l O a n d 1-ha r e s e r v e s , in most cases over army ant s w a r m s . Ant swarms have been observed to invade the reserves through as much as 250 m of pasture and low second g r o w t h .
More slowly, but no less d e f i n i t i v e l y , the mixed species flocks gradually lost species in the 1980 isolates (Fig. 5 ) . T w o s p e c i e s , the Wedge-billed Woodcreeper (Gly-phorhynchus spirurus) and the Chestnut-rumped Woodcreeper (Xi(Gly-phorhynchus parda lotus) became relatively more abundant in both 1980 isolates over the course of t i m e . Capture rates for the Chestnut-rumped Woodcreeper w e r e , in fact higher in 1-ha than in 10-ha isolates and higher in both than in undisturbed forest (Fig. C) . Capture rates for the Wedge-billed Woodcreeper w e r e higher in 10-ha isolates than in undisturbed forest w h i l e in 1-ha iso
lates the capture rate w a s virtually identical to that recorded in non-isolated s a m p l e s . In the 1980 isolates, tree-fal 1 and edge special ists increased in relative abundance, w h i l e understory frugivores declined initially but seemed to recover later on (Fig. 5 ) . The largest Piprid, (ca. 30 g . ) , the Thrush-like Manakin (Schiffornis t u r d i n u s ) , more
insectivorous than its c o n f a m i 1 i a l s , persisted only in the 10 ha reserve (Fig. 5 ) . For most common frugivores, capture rates in 1-ha reserves w e r e much lower than in 10-ha reserves or the non-usolated samples (Fig. 7 ) . The T h r u s h - l i k e M a n a k i n and the White-fronted Manakin (Pipra s e r e n a ) disappeared completely from 1-ha isolates, w h i l e capture rates for the White-crowned Manakin and the White-throated t h r u s h , T u r d u s albi_ c o l l i s , w e r e similar in undisturbed forest and in 10-ha isolates but much lower in 1-ha t r a c t s .
iijyl
i f c t S a m p l e p e r i o d : 1 / 1 1 / 7 9 - 1 / 6 / 8 0 n = 1 0 0 ( 1 - h a ) 1 2 4 ( 1 0 - h a ) _ | _ 1 - h a r e s e r v e J _ 1 0 - h a r e s e r v e 10 5 +\MÁ
S a m p l e p e r i o d : 1 / 1 - 3 1 / 1 2 / 8 1 n = 9 8 ( 1 - h a ) 161 ( 1 0 - h a ) 15 S a m p l e p e r i o d : 1/1 - 3 1 / 1 2 / 8 2 n = 2 7 ( 1 - h a ) 6 2 ( 1 0 - h a )i
S a m p l e p e r i o d : 1/1 - 3 1 / 1 2 / 8 3 n = 3 0 ( 1 - h a ) 5 3 ( 1 0 - h a ) o L.ε
D . C T J O— O)
c
σ> Ό X O L U Cl Cl tf) Q . J 3 U C O υ o c o . c E oA n t s Flocks Fruit Misc. Edge Nect
SPECIES / FEEDING STRATEGIES
5 . Relative abundances of a representative group of 25 species before and the sub sequent periods following isolat ion of the 1980 isolates. T h e ecological guilds represented a r e : Ants - Obligatory army a n t f o l l o w e r s ; Flock - M i x e d s p e c i e s , understory insectivorous f l o c k s ; Fruit - Understory f r u g i v o r e s ; M i s c - Common insectivores and one frugivore/insectivore; Edge - Tree-fall and forest edge s p e c i a l i s t s ; Nect - N e c t i v o r e s . Species a b b r e v i a t i o n s : p a - P i t h y s a l b i f r o n s ; gr - Gyranopithys r u f i g u l a ; dm - Dendrocincla m e r u l a ; ta - Thamnomanes aredesia c u s ; tc - Thamnomanes c a e s i u s ; m l - Myrmotherula l o n g i p e n n i s ; mg - M y r m o t h e r u l a guttural i s ; g s - G l y p h o r h y n c h u s s p i r u r u s ; d s - D e c o n y c h u r a stictolaema; xp - X i phorhynchus p a r d a l o t u s ; a i A u t o m o L u s i n f u s c a t u s ; nib M y i o b i u s b a r b a t u s ; eg -Corapipo g u t t u r a l i s ; pp - P i p r a p i p r a ; ps - P i p r a s e r e n a ; st - S c h i f f o m i s t u r d i n u s ; h p - Hylophylax p o e e i l i n o t a ; ta - T u r d u s a l b i c o l l i s ; pc - P l a t y r i n c h u s c o r o n a t u s ; ca - Cyphorhinus a r a d a ; ga - G a l b u l a albigularis; h e - H y p o c n e m i s c a n t a t o r ; m c - Microbates c o l l a r i s ; p r - Percnostola rufifrons;tf - T h a l u r a n i á f u r c a t a .
CONTINUOUS FOREST
(Captures/net-hour) χ 1 0 ~
3F i g . 6. Capture rates for Mixed species flock species comparing rates in isolated r e serves to those in continuous f o r e s t s . F o r each s p e c i e s , the average capture rates in both 1- and 10-ha isolates a r e plotted against the average capture rates in continuous f o r e s t . T h e open diamond represents the capture rate in 1-ha iso lates and the dark diamond that from 10-ha r e s e r v e s . Species abbreviations as in Fig. 6, with t h e inclusion o f : ro R h y n c h o c y c l u s o l i v a c e u s ; x m Xenops m i -n u t u s ; m m - M y r m o t h e r u l a m e -n e t r i e s i i ; m a - M y r m o t h e r u l a a x i l l a r i s ; h o - H y l o p h i lus o c b r a c e i c e p s ; ms - M y r m o t h e r u l a g u t t u r a l i s . Species whose points lie on or near t h e diagonal line show equal capture rates in isolated and non-isolated r e s e r v e s .
0 10 2 0 3 0 4 0 50 60
CONTINUOUS FOREST
(Captures/net-hour) χ 1 0 ~
3F i g . 7. Capture rates for frugivorous species comparing rates in isolated reserves ta those in continuous f o r e s t s . F o r each s p e c i e s , the a v e r a g e capture rates in both) 1 - and 10-ha isolates are plotted against the a v e r a g e capture rates in contintj o u s f o r e s t . S y m b o l s as in F i g . 7 , abbreviations as in F i g . 6 w i t h the inclusion, o f : gm - G e o t r y g o n m o n t a n a ; m m - M i o n e c t e s m a c c o n n e l l i .
Five of the 6 most commonly captured edge and tree-fall specialists had higher capture rates in 1-ha isolates than in 10-ha isolates, and 2 of t h e s e , the Yellow-bi1 led J a c a m a r , Galbula a l b i r o s t r i s , and the Ferruginous-backed Antbird had greatly elevated capture rates in 1-ha reserves when compared to both 10-ha isolates and undisturbed forest (Fig. 8 ) .
H u m m i n g b i r d s , particularly the Long-tailed Hermit (Phaethornis superei 1iosus) in creased in relative abundance in the 1980 isolated 1-ha reserve (Fig. 5 ) . Capture rates from all 1- and 10-ha isolates for the Fork-tailed Woodnymph (Thalurania furcata) w e r e indistinguishable from those in undisturbed forest (Fig. 9 ) . The two other c o m m o n , and congeneric understory hummingbirds reacted oppos i tel y to forest fragmentat i o n . The Long-tailed Hermit w a s captured m o r e frequently in 1-ha isolates than in 10-ha fragments or in undisturbed f o r e s t , w h i l e the Straight-billed Hermit (Phaethornis b o u r c i e r i ) showed a very slight decrease in capture rates when compared to 10-ha isolates or undisturbed
forest. Capture rates for both species in 10-ha isolates w e r e very similar to those in non-isolated forest samples (Fig. 9 ) .
Comparing capture rates from undisturbed forest w i t h those after 6 months of iso lation of 1- and 10-ha r e s e r v e s , w e see that the trends witnessed in the 1980 isolates seem to be general (Figs. 6 - 9 ) . T h e s e data show that species are not s imply becoming more or less important in relative a b u n d a n c e , but a l s o are showing dramatic changes in
their levels of activity in the r e s e r v e s , as measured by capture r a t e s .
Soon after isolation of 1- and 10-ha reserves capture rates of several species of the m i d d l e story of the forest increased dramatically,' suggesting an increase in abundance or activity levels. Snethlage's Tody-tyrant (Hemitriccus m i n o r ) , the Olivaceous Flatbill (Rhynchocyclus o l i v a c e u s ) , the 01ivaceous Woodcreeper (S ϊ t_ tasomus griseicapi 1 lus) , and the Black-throated Trogon (Trogon rufus), in particular, w e r e relatively common in virgin f o r e s t , but w e r e rarely captured in mist-nets because they normally foraged above the 2 m that the nets samp] e. Snethl age 1 s T o d y - t y r a n t , for e x a m p l e , is heard calling on roughly 50¾ of all capture d a y s , yet it is extremely rare in the nets in undisturbed forest (Table 3 ) . The capture rate for isolated 10-ha reserves w a s more than 10 times that in undisturbed forest and nearly 30 times the undisturbed forest rate in 1-ha reserves (Table 3 ) .
Capture rates for all these mid-story species w e r e higher in isolated reserves than in the undisturbed forest and in general tended to be higher in the 1-ha isolates than in the 10-ha t r a c t s , despite the higher capture rates for all species combined in the 10-ha isolates ( 3 5 6.87 v s . 2 3 3 .^ 1 captures/10000 net-hours) (see A p p . 1 ) .
This pattern w a s not pervasive throughout the mid-story insectivores, however . Aq least two common flycatchers of the mid-stories of undisturbed f o r e s t , the Forest Elaenia (Myiopagis gaimardi) and the Yellow-margined Flycatcher (Tolmomyias a s s i m i l i s ) did not show similar changes in foraging h e i g h t .
5 10 15 20 25 3 0 3 5 4 0 4 5
CONTINUOUS FOREST
(Captures/net-hour) χ 1 0 ~
3F i g . 8 . Capture rates f o r tree-fall and forest adge s p e c i a l i s t s , comparing rates in i s o lated reserves to those in continuous f o r e s t s . For each s p e c i e s , the average capture rates in both 1- and 10-ha isolates are plotted against the average capture rates in continuous f o r e s t . Symbols as in F i g . 7 , abbreviations as in F i g . 6, w i t h the inclusion o f ; h n H y l o p h y l a x n a e v i a ; mf M y r m e c i z a f e r r u g i -n e a .
CONTINUOUS FOREST
(Captures/net-hour) χ 1 0
- 3F i g . 9. Capture rates for h u m m i n g b i r d s , comparing rates in isolated reserves to those in continuous f o r e s t s . For each s p e c i e s , the average capture rates in both 1-and 10-ha isolates are plotted against the average capture rates in continuous forest. Symbols as in F i g . 7 , abbreviations as in F i g . 6, w i t h the inclusion o f : pb -Phaethornis b o u r c i e r i ; ps - -Phaethornis supereLliosus.
DISCUSSION
The most immediate and dramatic effect of forest fragmentat ion was the near doubl ing of capture rates in 1 - and 10- h a reserves (Fig. 3 ) . The increased c a p t u r e rates m a y in part reflect greater foraging activity of the birds in a reserve w h o s e area is| smaller than their normal h o m e range. The drop in recapture percentages coincidental with the increase in capture rates (Fig. 4) showed that an influx of displaced individu als fleeing the felled forest is certainly an important factor. M o r e detailed analyses of the trends in capture and recapture rates are presented by Bierregaard & Lovejoy (in p r e s s ) .
The influx of birds witnessed in the 1980 isolates a f t e r the surrounding capoeira (second growth) reached a height sufficient to have a shaded understory suggests the presence in undisturbed forest of a floating population of birds looking for suitable territory. The recolonization of the reserves by birds that had disappeared from the iso
lates emphasizes the importance suitable habitat connecting a potential colonizing source and the forest tract in question. Recolonization was first w i t n e s s e s in the 1-ha reserve, as expected, given Its greater proximi ty to undi sturbed forest (the recolonization source) than the 10-ha reserve.
Clearly, the avian understory community undergoes drastic changes in its species composition following fragmentation into 1 - and 10-ha reserves surrounded by pasture and low second g r o w t h . Other studies by W i l l i s (1979) in the State of São P a u l o , B r a s i1 , Leek (1979) in E c u a d o r , and Terborgh ( 1 9 7 4 ) , W i l l i s ( 1 9 7 4 ) , W i l l i s & Eisenmann (1979) and Karr ( 1 9 8 2 ) on B a r r o Colorado Island in Panama have a l s o reported the dísappearanee of various species of birds following fragmentation o r Insularization of once continuous tropical forest into areas of up to 1 , 5 0 0 h a .
W i l l i s ( 1 9 7 9 ) has shown that avian communities in remnant w o o d l o t s of 2 1 , 2 5 0 , and 1,A00 ha in.the A t l a n t i c forest regions of the State of Sio Paulo are depauperate when compared to a hypothetical list o f the avifauna that should have been present at the moment of isolation of the w o o d l o t s . What cannot be determined from his data is exactly how many species have been lost from each size wood lot. It is c e r t a i n , for instance, that the 1 0- h a remnant did not have the full complement of 230 species when it was isolated, but rather some s u b - s a m p l e , whose composi t ion would have depended on both a random sampling effect and on the micro-habitat the remnant prior to isolation. Willis' (1979)calculations of percentage species lost in ei.ch woodlot size are consequently difficult to interpret. Analyses of species loss "rom the 1,500- h a Barro Colorado Island (Willis & Eisen_ m a n n , 1 9 7 9 ; K a r r , 1 9 8 2 ) , are based on species lists compiled from the hi 1 1 top that became Barro Colorado Island when the Chagres River was dammed in 1912. H o w e v e r , s i nee systematic surveys were not conducted prior to insularization, Karr ( 1 9 8 2 ) suggested the most a p -propriate method to analyze changes in the ecosystem was to compare the avifauna w i t h that extant on the adjacent mainland. This approach is usually the only one available to biogeographers concerned w i t h questions of species loss from islands (Brown, 1 9 7 7 ; Soule et a l . , 1 9 7 9 ; Terborqh & W i n t e r , I98O) and suffers the same lack of quantitative p r e - i s o
lati on data as W i l l i s1 ( 1 9 7 9 ) analyses mentioned a b o v e .
The most drastic result of forest fragmentation studied here w a s the reduction of continuous forest to patches smaller than the home range of a given individual. The d i s a £ pearance of the obligate army ant following species from our isolated study plots w a s ρ red ictable on the has ΐ s of knowl edge of thei r home range requi rements. Wi 1 1 i s & On i k i (1979) states that individual White-plumed A n t b i r d s use an area 3 - 5 km across at Reserva D u c k e , w h i l e Harper (1988) and Bierregaard (unpubl. data) have shown that individual birds of this species use an area of between 1 to 2 km in d i a m e t e r , w i t h little evidence of movements more than 2 kms through the forest. Regardless of this small d i f f e r e n c e , it is not surprising that individuals of the obi igate army ant fol lowing species did not persist in a 10- h a forest fragment (ca. 350 m a c r o s s ) . Army ant swarms themselves use an area larger than 10 ha (Harper, u m p u b l . d a t a ; F r a n k s , 1 9 8 2 ) .
T h e disappearance of the mixed species flock from the 10- h a 1980 isolate w a s not predictable purely on the basis of known home range information. T h e s e flocks typically
use an area of about 5 ha in western Amazonia (Munn & T e r b o r g h , 1979) and about 1 2 ha in the Manaus area (Powell, u n p u b l . d a t a ) and use the same area over a period of years (Munn & T e r b o r g h , 1979; P o w e l l , u n p u b l . d a t a ) .
It is tempting to suggest a relation between the disappearance of the f 1 ock and the greatly increased levels of avian activity immediately following isolation. W e r e indi viduals of flock species so busy defending their home range from invaders that they suffered higher mortality? T h e r e w a s an obvious increase in vocalizations of flock species in the reserves after isolation. Were resources depleted to such low levels that even the residents could not find enough food? Our data are currently insufficient to answer these q u e s t i o n s . H o w e v e r , observations in a 10- h a reserve (1207) isolated without any noticeable post-isol at ion crowding (see Fig. 3) show that the mixed species flock there "shed" species in roughly the same order as from the 1980 isolate w h e r e crowding w a s pronounced (D. S t o t z , p e r s . c o m m . ) .
Capture rate data from all isolates show that only certain flock species disap peared from the reserves. H o w e v e r , one of the species that disappears from the reserves, the Cinereous Antshrike (Thamnomanes caesius) is one of the more important nuclear species (sensu W i n t e r b o t t o m , 19^9) of the flocks ( P o w e l l , p e r s . c o m m . ) . After the loss of the Cinereous A n t s h r i k e , the species that do persist in the isolates no longer traveled as coheçent f l o c k s . It is particularly interesting that the Dusky-throated Antshrike (Tham nomanes ardesiacus) shows higher capture rates in both 1 - and 10- h a isolates than its extremely similar c o n g e n e r , the Cinereus A n t s h r i k e . This may ref1ect the extent to which the Cinereus Antshrike depends on traveling within a cohesive flock. As much as 50¾ of its prey may be insects that have been flushed by other species traveling wi th the flock (Powell, 1 9 8 5 ) . As certain species are lost from the flock due to some modification in the habitat occasioned by isolation, the flock may become too smal 1 to susta ί η the C inereous A n t s h r i k e .
Several flocking species seem little affected by iiolat ion. The Long-winged and the White-flanked Antwrens (Myrmotherula longipennis and M . a x i l l a r i s ) are captured more
frequently in 1-ha than in 10-ha isolates, although the differences are very si ight, and at approximately the same rates in isolates as in undisturbed forest (Fig. 6 ) . Powell
(1985) has suggested that the co-occurence of up to 7 species of antwrens in the genus Myrmotherelula may be due to the flocking species foraging along w i t h b i r d s Z to 3 times heavier than they a r e , through an area m u c h larger than w o u l d be necessary to support a pair of a n t w r e n s . T h i s , Powell a r g u e s , may result in a density of antwrens below the carrying capacity of the e c o s y s t e m , which in turn would ameliorate the effect of inter specific competition. Were this the c a s e , it would not be surprising to find pairs p e r sisting in 1- h a isolates. Gradwohl & Greenberg (1980) report territory s i zes of slightly over 2 ha for mixed species flocks in Panama w h o s e n u c l e a r species a r e the Checkered-throated Anteren (Myrmotherula fulvíventris) and the Dot-winged Antwren (Microrhopias q u i x e n s i s ) , birds of similar weight to the antwrens in the Manaus f l o c k s . It would be interesting to study the foraging behavior and success rates of antwren pairs feeding independently from flocks to assess the value of flock membersh i ρ to these s p e c i e s . This
in turn might permit us to predict w h e t h e r or not antwrens in small isolates will persist. The Chestnut-rumped Woodcreeper (Xíphorhynchus p a r d a l o t u s ) i s the largest (ca. 36 g.) of the common understory species to persist in 1-ha isolates and seems to do so by way of a very diverse array of foraging s t a t e g i e s , including foraging with canopy f l o c k s . The elevated capture rates recorded for the species in isolates (Fig. 6) may reflect a high level of activity necessary to procure sufficient food in a small forest f r a g m e n t . Mist-netting captures of the Wedge-billed Woodcreeper with no other mixed-species flock members near-by suggests that the species is not a permanent member of the f1o c k s . Powell's (pers. comm.) observations of the species iη the project reserves suggested that
it used a home range roughly one-half the size of that used by the mixed-species f l o c k s , and that each pair will join a flock when there is one foraging in its territory. Thus there is almost always a pair of Wedge-billed Woodcreepers wi th every flock, but each pa i r of Wedge-billed Woodcreepers spends a substantial amount of time foraging away from flocks. Given this independence from the f l o c k s , it is not surprising that the W e d g e -bills persist in 10-ha isolates.
T h e persistence of hummingbirds in the isolated reserves is interesting in light of W i l l i s ' ( 1 9 7 9 ) observations that this group w a s under-represented Í η the smalI est (20 ha) woodlot he studied in Sao P a u l o . Perhaps because the reserves are separated from c o n tinuous forest by 100 to several hundred m e t e r s , individuals can m o v e to the adjacent forest on a daily or seasonal basis w h e n resources are lacking. As w e l l , there may be plants in the connecting second growth that serve as food s o u r c e s . This will be tested
in the future, as further felling will increase the isolation of these and other re serves from continuous forest to at least 1 k m , and maintenance by fazenda personnel w i1 1 replace the second growth with p a s t u r e . W e can offer no explanation for the intriguing difference in response to isolation shown by the Long-tailed and S t r a i g h t - b i1 led Hermits. Detailed field observations of the species will be necessary to understand this d i f f e r ence.
If all bird species w e r e equally likely to be captured by mist n e t s , capture data from mist nets would reflect actual relative abundances of the species under study. H o w e v e r , species are disproportionately represented in mist net data because of m o r p h o logical or behavioral factors (Bierregaard, u n p u b l . d a t a ) . Bi rds that are either too smal1 (hummingbirds) or too large (tinamous) for the mesh of the net size used will be less likely to be caught than other s p e c i e s . Birds that move through the forest primarily by walking also have a reduced probability of becoming entangled in the n e t s . I t i s a s s u m e d in this study that these biases will be the same in virgin forest and isolated r e s e r v e s , such that differences in the mist net data reflect some real difference in the two habi tats (virgin forest v s . isolated forest r e m n a n t s ) , either in behavior or in real popu lation densities. H o w e v e r , the significance of such differences can only be interpreted based on an understanding of the biology of the birds under s t u d y , an understanding that can only be gained by direct field o b s e r v a t i o n s .
T h e appearance of common mid-story birds (see Table 3) in understory nets in iso lated reserves could easily have been misinterpeted as the result of an increase in the population sizes of these species w e r e it not for our observational data on their a-bundance and foraging behavior in the undisturbed forest. Based on these d a t a , w e know that the increases in post-isolation relative abundance for these is due principally to b e h a v i o r a l , rather than populational (immigration) c h a n g e s .
The behavioral shifts in foraging height that bring these species i nto the nets in isolated reserves suggest that birds need a given volume of forest through which to f o r a g e , rather than just an given a r e a , per s e . They seem to be increasing their verti cal foraging range to compensate for the reduction in horizontal s p a c e . Such species may be more persistent than species of the lowest stratum in smaller reserves because they can adjust their foraging volume vert i ca 11 y and hori zon tal 1 y, wh i 1 e ground forag i ng species can only increase the three dimensional s p a c e , or home range vol u m e , hori zonta 11 y. If this
is insufficient to compensate for lost foraging territory, they may be forced to increase the rate of foraging in the available v o l u m e , perhaps preventing adequate replenishment of food resources.
Three alternative explanations could account for the observed shifts in foraging h e i g h t s ; competitive release, physiological restraints attendant on the changes iη m i c r o climate in newly isolated reserves (Karr, p e r s . c o m m . ) , and changes in the distribution of food resources (Malcolm, p e r s . c o m m . ) . The latter 2 explanations might be linked, as micro-climatological changes could cause changes in insect activity attracting birds to lower levels of the f o r e s t .
T h e competi tive release explanation would suggest that rather than being forced to expand vertical foraging b e h a v i o r , mid-story species may be changing their foraging downwards to use a volume of forest vacated by understory species which did not persist in the reserve for some other reason. Although it is not always clear what species might be competitively excluding the mid-story species which moved down to net lev^l s after iso lation (particularly the Black-throated T r o g o n ) , the speed with which they do suggests that such an interspecific interaction is not the causative f a c t o r . The mid-story birds
a r e appearing in the nets long before all but the o b l i g a t e army ant fol lowers have b e g u n to disappear from the r e s e r v e s . For the Olivaceous W o o d c r e e p e r , w h i c h shows the most pronounced increase In activity at net level, several other simi lar sized Dendròcolaptids, such as the Wedgebilled and Spotthroated (Deconychura stictolaema) W o o d c r e e p e r s , c o n -tinue to be caugh*- in the nets well after the Olivaceous a p p e a r s in the s a m p l e s . In f a c t , the competitive effect may be just the r e v e r s e , with increased resource u s e by invading (descending) mid-story species mitigating against the p e r s i s t e n c e of the understory species. The Spot-throated Woodcreeper d o e s , inded, disappear completely from 1-ha tso lates.
The extent to which midstory species are fleeing unfavorable m i c r o c l i m a t i c c o n -ditions in the isolated reserves will be difficult if not impossible to determine and must be considered a viable alternative explanation to the observed p h e n o m e n o n . H o w e v e r , in this regard, the lack of such vertical displacement in the Forest Elaenia and the Yellow' margined Flycatcher ( S t o t z , p e r s . c o m m . ) , two of the most common mid-story insectivores in the undisturbed forest is particularly intriguing* T h e elucidation of the ecologicaJ
or physiological differences leading to this assyrnetry will require more detailed aut-* ecological studies of the species involved. Perhaps these species tend to compensate for
restricted home ranges by shifting their foraging up into the canopy rather than down into the u n d e r s t o r y , or perhaps their h o m e ranges a r e so small that they need not expand their foraging volume at a l l .
ACKNOLWLEDGMENTS
A . M c G r a t h , L. van d e r T o k , D. G r a h a m , F. J o y c e , P. O y e n s , D.Walczak, Β. W i n n , E . P e r k i n s , C. Strang, T. T e s c h e , G. B r a n d ã o , Μ. M a g a l h ã e s , Μ. B e r t e , P. P o l s h e k , D. S t o t z , C. Q u i n t e l a , R. D o w n e r , J. B a t e s , and N . Belfiore most energetically and capably ran the net lines and assisted in data management since the beginning of the project. T h e current manuscript has benefitted greatly from discussions with and suggestions m a d e by B . Zim m e r m a n , L. H a r p e r , J. B a t e s , B . K l e i n , J. K a r r , and especially D. S t o t z , J. Malcolm, M. W o n g , and an anonomous reviewer.
This study was supported by the World Wildlife F u n d - U S , the Instituto Nacional de Pesquisas da Amazônia ( I N P A ) , the Instituto Brasileiro de Desenvolvimento Floresta' ( I B D F ) , a grant from the National Park S e r v i c e , C o o p e r a t i v e Agreement C X- 0 0 0 1 - 9 - 0 0 ^ 1 , the Man and the Biosphere program of U S - A I D , and represents publication number 3 1 in the Minimum Critical Size of Ecosystems Project (Dinâmica Biológica de Fragmentos F l o r e s
RESUMO
Análises de mais de 19.000 capturas em cinco anos de um programa dz marcação e Ae
captura [captura zm rede ζ anithamznto) foram feitas para comparar a estrutura da comu
nidade dz avzs entre uma, floresta virgem de terra firme ζ as reservas florestais imola
das de 1 a Π ha localizadas em
3
fazendas do Distrito Agropecuário da SUERAMA. Pela pri
meira vez, dados sobre a perda de espécies em fragmentos florestais, ou "ilhas dz
habi-tat", podem ser comparados com recenseamentos quantitativos feitos nos próprios locais
zm estudo, antes da insularização. A comunidade dz aves em sua condição original e mui
to diversificada; 141 espécies de aves foram anilhadas em redes dz nylon do sub-bosquz.
Nenhuma espécie representou mais dz %% de todos os indlviduos anilhados ou 10% de todas
as capturas. A comunidade e principalmente insetZvora; mais de que %0% dos indivZduos
anilhados são principalmente insetZvoros, dos quais,
371
são seguidores obrigatórios dz
formigas de correição ou membros de bandos muito evoluídos de espécies mistas. Um peque
no Viprideo está. entre as 4 espécies mais anilhadas, mas a maior parte das atividades
dos frugZvoros e beija-flores ocorre no dossel. Análises , destas comunidades zm
reser-vas isoladas de 1 a 10 ha mostram que os seguidores obrigatórios de formigas são os mais
sensíveis a fragmentação da floresta. Bandos de espécies mistas deterioram-se também,
perdendo espécies antes de dois anos de isolamento. Espécies cujo habitat preferencial
são clareiras formadas por queda dz árvores aumentam em importância relativa e em nZvel
de atividade nas amostras pós-isolamento. Com o isolamento, os pássaros residentes
en-frentam um influxo de pássaros oriundos da floresta derrubada. Esta tendência ζ refle
tida no aumento do número de capturas e o decréscimo da proporção de pássaros ja anilha
dos nas amostras de cada dia. diversas espécies de insetZvoros dos estratos intermedia
rios aumentam o alcance vertical de seus hábitos alimentarei diante da restrição
hori-zontal de seus territórios nas reservas isoladas de 1 e 10 ha. Isto sugzrz quz territo_
rios devem ser considerados em termos de volume do habitat em vez de área.
T a b l e 1. Percent of 10,918 individuals employing various feeding strategies, number of species, percent of individuals corrected by b i o m a s s , and number of c a p t u r e s .
Diet N o . of species % Individuals % Biomass % Captures
Nectar feeders 8 4.85 0.95 3.05
Fruits, seeds 36 14.6 22.96 12.7
Fruits, vertebrates 3 0.05 0.43 0. 0 3
Insects (over army ants) 3 14.43 15.93 18. 8 6
Insects 64 41.3*4 34.07 41.98 Insects, fruits 4 0.04 0.03 0.02 Insects, vertebrates 1 0.38 1.74 0.27 Mixed-species flocks 21 23.0 1 6. 1 9 22.2 F i s h , insects 2 0.14 0.22 0.09 Vertebrates 6 1.16 7.48 0.8 Total
1
*3
100.0 100.0 100.0Table 2 . Comparisons of mean capture rates (number of birds/net-hour) in a 1-ha and a 10-ha forest reserve before and after isolation. Means are shown for each c a s e from 7 capture d a y s . The results of Student-T statistic for paired m e a n s are
indicated. N S = no difference between the means was detected. * = null hyphothe sis of equal means rejected at Ρ á 0 . 0 5 . *** = null hyphothesis rejected at
p = . σ ο ι . Ρre-isolation Post-isolation 1-ha 10-ha 0. 2 6
NS
0.19 0.45*
0.31 APPENDIXSpecies of birds captured in continuous forest and isolated reserves of 1 a n d l O h a . Diet categories a r e : f - f r u i t , i - insects, η - nectar, ρ - fish, msf - forage for insects in mixed-species f l o c k s , ν - vertebrates. In the " b a n d " columm is listed the number of individuals of each species banded in continuous forest. One individual flycatcher and Emberiziid w e r e not identified to species and have been excluded from this a p p e n d i x , bringing the actual total of birds banded in continuous forest to 10,918 of 14 3 s p e c i e s . Columns labelled " p r n h " , "prnh 1" and "prnh 1 0 " indicate capture rates (captures per 1000 net-hours) in continuous forest, 1-ha isolates and 10-ha isolates, respectively. The column labelled "biom" indicates for each species the total biomass represented by the individuals banded for that s p e c i e s . Weights w e r e estimated for species indicated with an *. For all other s p e c i e s , weights a r e based on our o w n capture d a t a .
Referências bibliográficas
Projeto RadamBrasil - 1978. Folha SA 2 0 M a n a u s . Rio d e J a n e i r o . Ministério de Minas e E n e r g i a : Departamento Nacional de P r o d u ç ã o M i n e r a l , p . 2 6 1 .
Bierregaard J r . , R. o . & L o v e j o y , Τ . E . - I n p r e s s . B i r d s in A m a z o n i a n Forest fragments: E f f e c t s of insularization. Proc. X I X Int. Ornith. C o n g .
Brown, J. H . - 1 9 7 1 . Mammals on m o u n t a i n t o p s : n o n equilibrium insular b i o g e o g r a p h y . Am. Hat., 1 0 5 : 4 6 7 - 4 7 8 .
C h a u v e l , A . - 1 9 8 3 . Os latossolos a m a r e l o s , á l i c o s , argilosos dentro d o s ecossistemas das bacias experimentais do INPA e d a região v i z i n h a . A c t a A m a z ô n i c a , S u p l . , 1 2 ( 4 ) : 4 7 - 6 0 .
F r a n k l i n , I. R. - 1 9 8 0 . Evolutionary change in small p o p u l a t i o n s . I n : S o u l e , Μ. E . & W i l c o x , B . A . e d s . Conservation B i o l o g y : A n Evolutionary-ecological Perspective.' Sinauer A s s o c . Inc. Sunderland, M A . p . 135-149.
F r a n k s , N . - 1 9 8 2 . Ecology and population regulation in the army a n t Eciton b u r c h e l l i . I n : L e i g h , E . G . J r . ; Rand, A. S.; W i n d s o r , D . M . e d s . T h e ecology of a tropical forest: Seasonal rhythms and long-term c h a n g e s . Smithsonian Institution. W a s h i n g t o n , D C . p . 3 8 9 - 3 9 5 .
J r a d w o h l , J. & G r e e n b e r , R. - 1 9 8 0 . T h e formation of antwren flocks on Barro C o l o r a d o I s l a n d , P a n a m a . A u k , 9 7 : 3 8 5 - 3 9 5 .
G u i l l a u m e t , J. & K a h n , F. - 1 9 8 2 . Estrutura e dinamismo da f l o r e s t a . A c t a A m a z ô n i c a , S u p l . , 12(4):61-77.
H a r p e r , L . H . - 1 9 8 8 . T h e persistence of ant-following birds in small A m a z o n i a n forest f r a g m e n t s . A c t a A m a z ô n i c a , 1 8 ( 3 - 4 ) :
K a r r , J. - 1971. Structure of avian communities in selected Panama and Illinois h a b i t a t s . E c o l . M o n o g r . , 4 1 : 2 0 7 - 2 3 1 .
- 1 9 8 2 . Avian extinction on Barro Colorado Island. Panama: a r e a s s e s s m e n t . A m . H a t . , 119:220-239.
L e e k , C. F. - 1 9 7 9 . Avian extinctions in an isolated tropical wet-forest p r e s e r v e . Ecua' d o r . A u k , 9 6 : 3 4 3 - 3 5 2 .
Lovejoy, Τ . E . - 1975. Bird diversity and abundance in A m a z o n forest c o m m u n i t i e s . L i v i n g Bird, 13:127-191.
L o v e j o y , Τ. E . & O r e n , D. C. - 1 9 8 1 . The m i n i m u m critical size of ecosystems. In: B u r g e s s , R. L . & S h a r p e , D. M . e d s . Forest Island Dynamics in Man Dominated L a n d s c a p e s . N e w Y o r k , Springer-Verlag. p. 7-12.
Meyer d e Schauensee, R. - 1 9 6 6 . T h e species of b i r d s of South A m e r i c a and their d i s t r i b u t i o n . N a r b e t h , P a . L i v i n g s t o n . 577 p.
M u n n , C. A . & T e r b o r g h , J. W. - 1979. Multi-species territoriality in neotropical foraging flocks. C o n d o r , 8 1 : 3 3 8 - 3 4 7 .
P a t t e r s o n , H. D. - 1 9 6 0 . A further n o t e o n a simple method for fitting an exponential c u r v e . B i o m e t r i k a , 4 7 : 1 7 7 - 1 8 0 .
P e a r s o n , D. - 1 9 7 7 . A pantropical comparison of bird community structure on six lowland rain forest s i t e s . C o n d o r , 7 9 : 2 3 2 - 2 4 4 .
P o w e l l , G. V. N . - 1 9 8 5 . Sociobiology and adaptive significance of interspecific foraging flocks in the n e o t r o p i c s . O r n i t h . M o n o g r . , 3 6 : 7 1 3 - 7 3 2 .
Rankin, J. Μ . ; P r a n c e , G. T . ; S i l v a , Μ . F.; R o d r i g u e s , W . Α . ; U e h l i n g , Μ . Ε .
-R e s u l t a d o s preliminares d e um levantamento florestal de 31 ha d e terra firme na A m a zônia C e n t r a l . D e s c r i ç ã o geral da v e g e t a ç ã o e d a d o s t a x o n õ m i c o s . (em p r e D a r a ç ã o ) . Sick, Η . - 1 9 6 7 . R i o s e enchentes na Amazônia como obstáculo para a avifauna. A t a s S i m p .
B i o t a A m a z ô n i c a . , 5:495-,520.
S n e d e c o r , G. W . & C o c h r a n , W . G. - 1 9 6 7 . Statistical M e t h o d s . A m e s , Iowa, Iowa State University P r e s s . p . 4 6 7 - 4 7 1 .
S o u l e , Μ . Ε . - 1 9 8 0 . Thresholds for s u r v i v a l : m a i n t a i n i n g fitness and evolutionary p o tential. I n : S o u l e , Μ . E . & W i l c o x , B . A. e d s . Conservation Biology a n E v o l u t i o n ary-ecological P e r s p e c t i v e . Sunderland, M A . Sinauer A s s o c . , I n c . p . 1 5 1 - 1 6 9 .
S o u l e , Μ . E . ; W i l c o x , Β . Α . ; H o l t b y , C. - 1979. Benign n e g l e c t : a model of faunal c o l lapse in the game reserves of East A f r i c a . B i o l . C o n s e r v . , 1 5 : 2 5 9 - 2 7 2 .
T e r b o r g h , J. W. - 1 9 7 4 . Preservation of natural diversity: the problem of extinction prone s p e c i e s . B i o s c i e n c e , 2 4 : 7 1 5 - 7 2 2 .
- 1 9 7 5 . Faunal equilibria and the desing of w i l d l i f e p r e s e r v e s . I n : G o l l e y , F. B. & M e d i n a , E . e d s . Tropical E c o l o g y . N e w Y o r k , Springer-verlag. p . 3 6 9 - 3 8 0 .
T e r b o r g h , J. W . and W i n t e r , B . - 1 9 8 0 . Some causes of extinction. I n : S o u l e , Μ. E . & W i l c o x , B . A. e d s . C o n s e r v a t i o n b i o l o g y : a n evolutionary-ecological p e r s p e c t i v e . Sunderland, M a s s . S i n a u e r . p . 1 1 9 - 1 3 3 .
W i l l i s , E . 0. - 1 9 6 7 . The behavior of Bicolored A n t b i r d s . U n i v . C a l i f . P u b l . Z o o l . , 7 9 : 1 - 1 3 2 .
- 1 9 7 4 . Populations and local extinctions of birds on Barro Colorado Island, Pana m a . E c o l . M o n o g r . , 4 4 : 1 5 3 - 1 6 9 .
_ 1 9 7 9 . The composition of a v i a n communities in remanescent woodlots in southern B r a s i l . P a p é i s A v u l s o s , Z o o l o g i a , São P a u l o , 3 3 : 1 - 2 5 .
W i l l i s , Ε . 0. & E i s e n m a n n , Ε . - 1 9 7 9 . A revised list of b i r d s on Barro Colorado Island, P a n a m a . S m i t h s o n . C o n t r i b . Z o o l . , 2 9 1 : 1 - 3 1 .
»
W i l l i s , E . 0 . & O n i k i , Y . - 1 9 7 8 . Birds and army a n t s . A n n . R e v . E c o l . S y s t . , 9:243-2 6 3 .
Winterbottom, J. M . - 1 9 4 9 . Mixed bird parties in the t r o p i c s , w i t h special reference to northern Rhodesia. A u k , 6 6 : 2 5 8 - 2 6 3 .
W o n g , M. A . - 1986. Trophic organization of understory birds i n a M a l a y s i a n dipterocarp f o r e s t . A u k , 103:100-116.
