Documents

11 pages
9 views

Influences of the season and the habitat structure on the foraging ecology of two coua species in the western dry forest of Madagascar

of 11
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Share
Description
Influences of the season and the habitat structure on the foraging ecology of two coua species in the western dry forest of Madagascar
Transcript
  C. R. Biologies 329 (2006) 691–701http://france.elsevier.com/direct/CRASS3/  Animal biology and pathology / Biologie et pathologie animales Influences of the season and the habitat structure on the foragingecology of two coua species in the western dry forest of Madagascar Philippe Chouteau  Laboratoire d’écologie, École normale supérieure, 46, rue d’Ulm, 75230 Paris cedex 05, France Received 13 July 2005; accepted after revision 16 June 2006Available online 7 August 2006Presented by Pierre Buser Abstract I studied the foraging ecology of Coquerel’s Coua ( Coua coquereli ) and Giant Coua ( Coua gigas ), which occur in the dry forestin west Madagascar. This kind of forest is characterised by an alternating of a dry and a rainy season. The foraging behaviour wasdescribed in several dimensions: i.e. height and proportion of perching, rate of capture (estimating the food availability), foragingtechniques, substrates used, type and size of the captured prey. Their foraging behaviour was different according to the seasonand to the proximity of water. Habitat structure was important to take in consideration to study their foraging behaviour. Theytended to use the same pattern of techniques and substrates, but differed by the proportions they used these variables and alsoby the possibility to climb into the dense understorey vegetation. Seasonal variation has probably an important effect on the preyavailability and the nature of prey captured. The diet of both species is also discussed. I suggest that change in habitat structureand resources levels could be important to take in consideration for the conservation of these forest birds. To cite this article: P. Chouteau, C. R. Biologies 329 (2006). © 2006 Académie des sciences. Published by Elsevier SAS. All rights reserved. RésuméInfluences de la saison et de la structure de l’habitat sur la stratégie alimentaire de deux espices de couas dans l’Ouestde la forêt sèche de Madagascar. J’ai étudié les stratégies alimentaires utilisées par deux espèces de couas terrestres de la forêtsèche de Madagascar, le coua de Coquerel ( Coua coquereli ) et le coua géant ( Coua gigas ). Cette forêt dite sèche est caractériséepar une alternance d’une longue saison sèche et d’une courte saison des pluies, durant laquelle la végétation présente un maximumde développement et où les proies animales de ces oiseaux sont potentiellement plus abondantes et plus diversifiées qu’en saisonsèche. J’ai analysé les stratégies de recherche et de prise de nourriture de ces oiseaux avec les variables suivantes : hauteur etproportion de proies capturées en hauteur dans la végétation (lorsque ces oiseaux, surtout terrestres, grimpent dans la végétation),techniques et substrats utilisés, taux de capture (permettant d’avoir une estimation de la disponibilité des proies capturées), typeet taille des proies capturées. Les stratégies sont étudiées et comparées pour chacune des espèces, durant les deux saisons et dansdeux parcelles qui diffèrent de par leur degré d’éloignement d’une rivière (3 km). Ces deux parcelles ont été utilisées pour la coupesélective du bois, ce qui élimine a priori l’effet de la dégradation sur les résultats obtenus. Une analyse préliminaire montre quel’éloignement de la rivière conduit à une modification de la structure de la végétation : la hauteur de la canopée diminue avecl’éloignement, la végétation du sous-bois tend à augmenter dans la station la plus éloignée. La variation saisonnière conduit àun recouvrement de la canopée plus grand dans la station proche de la rivière, et la structure de la litière est aussi modifiée. Lesstratégies utilisées par ces deux espèces de couas varient suivant la saison et suivant l’éloignement de la rivière. Les deux espèces E-mail address: philippe.chouteau@numericable.fr(P. Chouteau).1631-0691/$ – see front matter © 2006 Académie des sciences. Published by Elsevier SAS. All rights reserved.doi:10.1016/j.crvi.2006.06.005  692 P. Chouteau / C. R. Biologies 329 (2006) 691–701 diffèrent par la possibilité de grimper dans la végétation du sous-bois, par la taille et le type de proies capturées. Les substrats etles techniques sont les mêmes, mais diffèrent par les proportions utilisées. Toutes les variables considérées peuvent aussi différersuivant la saison et l’éloignement de la rivière. Ces deux espèces de couas sont similaires, mais différent par la taille. Aucunecompétition interspécifique n’a été remarquée, probablement parce qu’elles utilisent des micro-habitats différents pour se nourrir.Ces oiseaux incorporent une grande proportion de matière végétale dans leurs régimes alimentaires. La structure de l’habitat seraitdonc un paramètre essentiel à prendre en considération dans toute étude sur les stratégies de recherche et de prise de nourriture deces oiseaux. En conclusion, je suggère que les changements dans la structure de l’habitat et du niveau des ressources disponiblesseraient un paramètre important à prendre en considération dans une optique de conservation de ces oiseaux forestiers. Pour citer cet article:P. Chouteau, C. R. Biologies 329 (2006). © 2006 Académie des sciences. Published by Elsevier SAS. All rights reserved. Keywords: Coua coquereli ; Coua gigas ; Dry forest; Foraging behaviour; Madagascar; Seasonality  Mots-clés: Coua coquereli ; Coua gigas ; Forêt sèche; Comportement nourricier; Madagascar; Saisonnalité 1. Introduction The pattern of resource use by the animal speciesis one of the central topics in the study of biologicalcommunities. Some studies have confirmed a close rela-tionship between resource availability and bird-foragingecology[1,2]. Thespatio-temporalvariationinfooddis-tribution can affect the abundance and the foraging be-haviour of birds[3–5], particularly the food resourcesfor insectivorous birds, which change drastically withthe season[6].In addition, the behaviour of birds is also known to vary seasonally: birds adjust importantbiological activities (breeding, moulting) according toseasonal fluctuations of local conditions.Tropical habitats can fluctuate seasonally, particu-larly rainfall, affecting the food availability for birdspecies[7,8].Patterns of leaf growth and loss deter- mine also the associated production. The massive leaf production during the wet season supports an increaseddensity of arthropods upon which many birds feed[9].In a tropical dry forest, prey can vary in quantity andquality, especially the litter arthropods[10,11]or the in-sects linked to new leaves availability[12].Such changes may influence foraging tactics: foodavailability constrains the type of foraging techniquesthat can be used efficiently by a bird. However, thebird does not modify its foraging techniques as easilyas other foraging characteristics, e.g., substrate choice[13]. Rosenberg[14]suggested that the foraging tactics used by birds considerably affects the accessibility of prey on different substrates. In addition, Hejl and Verner[4]indicated that some species of birds living in thesame habitat undergo similar changes in foraging be-haviour and diet as temporal changes affect resourcesabundance and distribution.Most studies on seasonal foraging ecology have beendone in northern temperate zones, but the effects of sea-sonality in tropical forests, especially in Madagascar,have been little studied. I present here the results of atwo-year study aimed at investigating the seasonal vari-ability of two terrestrial coua species in a western dryforest in Madagascar. I hypothesized that, tracking foodavailability, the two species will seasonally change theirforaging strategies. In addition, in two forest plots dif-fering by their proximity to water, I evaluated whethervegetation structure influences their foraging behaviour. 2. Material and methods 2.1. Study site The study site was in the 10000-ha forestry con-cession of the ‘Centre de formation professionnelleforestière’ de Morondava (CFPF), in the Kirindy forest,near the western coast of Madagascar. The concessionwas 60 km northeast of Morondava (44 ◦ 39  E, 20 ◦ 03  S)in the deciduous forest, in the region called Menabe[15]. Rainfall varied widely from year to year, andranged from 300 to 1400 mm per year (80% falling inJanuary–February) with an annual average of 800 mm[16]. Temperature variability may be very large: meandaily maximum are around 36 ◦ C and the minimumaround 19 ◦ C[17].Foraging data were recorded during the rainy season (December to February) and the be-ginning of the dry season (March to June) in 1997 and1998.The Kirindy River crosses the forest from east towest (Fig. 1). Proximity of the water is important to thedistribution and structure of the vegetation[18,19].The study took place in plot CS-6, logged in 1980[20]and plot N-5, logged in 1990[21]. In this study, plot CS-6, near the Kirindy River, will be referred to asgallery forest and plot N-5, 3 km from the river, will bereferred to as dry forest. Both forests will be referred to  P. Chouteau / C. R. Biologies 329 (2006) 691–701 693Fig. 1. Map of the Kirindy forest (CFPF forest concession). The letters and numbers indicated the different forest plots. The field station was locatednear the ‘piste Conoco’. Field studies were performed in the plots CS-6 (gallery forest) and N-5 (‘dry forest’). as habitat. The study area lay on around 100 ha in eachone of the two plots(Fig. 1). Seven hundred and sixty-one trees (dbh > 40 cm)were logged in CS-6 for a yield of 1.3 m 3 of timber / ha[20].1289 trees were logged in N-5 for a yield of 1.07 m 3 of timber / ha[21].The fact the two plots were logged eliminated a possible interaction of the degrada-tion due to logging on the habitat structure. The fact thatten years separated the two logging operations betweenthe two plots was probably not important on the modi-fication of the habitat structure, because the tropical dryforest is known to have a long regeneration time[22].A possible modification of the vegetation structure be-tween the two plots could be only due to the seasonalvariation. 2.2. Vegetation structure I measured vegetation structure on 50 points ran-domly chosen into each plot. These points were alongseveral transects made in the two plots and during thetwoseasons(see[23]formoredetailsaboutthemethod-ology used). I estimated the following variables: canopyheight; canopy cover (estimated at the nearest 10%); es-timated mean distance visible at 1.5-m height, measuredin two different directions randomly chosen and numberof lianas and stems (dead and live) in 4 m 2 around thepoint into the understorey vegetation. Canopy height,number of lianas and number of stems were only usedto compare the two plots (CS-6 and N-5) but not used tocompare the seasonal variation, because these variableswere supposed not to change between seasons. Visibil-ity was used as an index of the understorey vegetationdensity: vegetation density = 1 / visibility. 2.3. Study species Two terrestrial couas were encountered in Kirindy:Coquerel’s coua ( Coua coquereli ) and the Giant Coua( Coua gigas ). This last species is around 50% biggerthan the first one (Chouteau, unpublished data). Therewas no morphological variation associated with sex.Some previous works analysed some biological aspectsof these species (see[24],for a review of the ecology of  these species).Densities of Coquerel’s Coua and Giant Coua in eachhabitat were previously measured during the rainy sea-son[23]:Coquerel’s Coua was more abundant in the logged gallery forest (24.2 individuals / km 2 ) than in thelogged dry forest (17.3 individuals / km 2 ).Giant Coua was slightly more abundant in the loggedgallery forest (3.7 individuals / km 2 ) than in the loggeddry forest (2.4 individuals / km 2 ). 2.4. Data collection Observations of foraging events were opportunistic,but I attempted to observe at various times of day, al-though couas were difficult to locate. I obtained sev-eral foraging sequences during at least 1 min, but no  694 P. Chouteau / C. R. Biologies 329 (2006) 691–701 more than 5 min, and I let a period of 30 min betweentwo successive observations. Although some investiga-tors recommend taking only the first foraging event foranalysis,Iretainedallofthemtoinsurerecordingincon-spicuous foraging events and to reduce biases towardsmost common foraging techniques and substrates used.I ringed five Coquerel’s Coua and four Giant Couain plot CS-6, 11 Coquerel’s Coua and two Giant Couain plot N-5[25].In addition, data were also obtained in both plots on some unringed birds. Furthermore, I madeseveral observations, separated in time and space, mostprobably concerning different individuals. However, thesame individuals were potentially recorded during thetwo seasons in a same plot.In the logged forest, I obtained 54 sequences duringthe rainy season and 52 sequences during the dry seasonfor Coquerel’s Coua (respectively 66 and 42 for GiantCoua).In the dry forest, I obtained 56 sequences during thedry season and 52 sequences during the rainy seasonfor Coquerel’s Coua (respectively 40 and 46 for GiantCoua).During these observations, I tried to identify the preycaptured. Because couas were particularly difficult tohandle and were easily stressed, I did not use emeticadministration to obtain samples. 2.5. Foraging behaviour  Fiveforagingvariableswereconsidered:meanheightreached by climbing in the vegetation above the groundwhen the birds foraged, foraging techniques used, sub-strates used, prey size, and capture rate.Terrestrial couas foraged mainly on or near theground(inthefirst30-cmlevelabovetheground).How-ever, some preys were captured higher in the vegetation.To capture these prey, the birds had to leave the groundand climb in the vegetation. I was interested to knowif the two species differed in their ability to climb andto capture prey high in the vegetation and what was theproportion of prey captured by this way. In this case, al-though I considered that couas were typically terrestrialbirds, which foraged on or near the ground, I hypothe-sized that, if there were a possibility to climb and searchfor prey in understorey vegetation, they did it efficiently.The heights of the places where the birds foraged wererecorded at the nearest 10 cm, by comparison with theheight of the observer. Giant Coua can climb until 10 mhigh from the ground (personal observation).I recorded the kind of capture techniques the birdsused as follows, modified from Remsen and Robinson[26]. ‘Glean’: prey captured on the substrate, withoutextension of the legs or the neck, and without a manipu-lation of the substrates. Preys captured by gleaning wereusually spotted nearby ( < 0 . 3 m) and the attack did notinvolve a flight component. ‘Lunge’: prey captured byrunning, on the litter (running prey) or in the air (flyingprey). ‘Leap’: prey captured by a jump (without usingthe wings), above the ground. ‘Sally’: prey captured bya powerful jump, with use of the wings. Prey capturedby sallying was always higher than prey captured byleaping. ‘Probe’: prey captured after a manipulation of the substrate (i.e. by using the bill to gap the dead leaveson the litter or to chase the prey into the dead curled-leaves or behind a part of bark). ‘Other’ included preycaptured by other techniques, as with extension of theneck and/or the legs. For the statistical analysis, sallyand leap were pooled together.Substrates used were defined as: ‘ground’ (on andinto leave litter), ‘leaves’, ‘trunk’, ‘stem’, ‘other’ (in-cluding dead trunk and air for flying prey). I assumedthat these substrates would harbour different prey types,and some differences in their use could help to segre-gate the foraging strategies used by these two couas.However, because of the difficulty to distinguish eas-ily between a stem or a dead leaf during the dry season,‘leaves’ and ‘stems’ were pooled together in this analy-sis as ‘foliage’.Prey size was estimated from bill length. Three preysize classes were defined as: small ( < 0 . 5 cm); medium(0.5–1.5 cm) and long ( > 1 . 5 cm). The biggest animalprey was easily recognized. Seeds were identified be-cause they were usually grouped on the ground undera particular tree. I was able to identify the remainingseeds after the bird ate them. But because all seeds werenot identified, particularly the smallest ones, I presumedthat the number of eaten seeds was underestimated.It was impossible to estimate the prey size if the birdswallowed the prey quickly, or if the individual was notwell visible. The proportion of prey size not measuredwas 22.7% for Coquerel’s Coua and 19.1% for GiantCoua for both plots and during the two seasons.The attack rate was used as an indirect measure of the prey availability. I usually could not always deter-mine whether an attack was successful, so this measurereferred only to the rate at which prey was attacked,not captured. To calculate it, I retained more than 20periods for each species and into each habitat. Theseperiods were at least equal to 20 min to reduce thebias introduced by inactive behaviour (preening, bask-ing, singing, and resting). These periods were chosenspatially and temporally separated on different individ-uals to assure the independence of the data. To assessthe attack rate, I used the total number of events of cap-  P. Chouteau / C. R. Biologies 329 (2006) 691–701 695 ture, divided by the total duration for each period. I usedeach period as an independent data point to compare theattack rate. 2.6. Data analysis For techniques and substrates variables, expressed asproportions, I used a multivariate analyse of Variance(MANOVA) to compare the interseasonal variation (thetwo seasons compared into each habitat) and to com-pare the intraseasonal variation (one season comparedbetweenthetwohabitats),withseasonorhabitatusedasindependent variable and techniques or substrates usedas dependant variables. I did not include the ‘other’ cat-egories in the MANOVA to avoid non-independence of proportions[27]. Statistical analyses have been donewith Statistica[28].The proportion of climbing was analysed by chi-square tests (procedure FREQ,[29]). Height reached byclimbing and rate of capture were calculated by a Man–Whitney test (procedure NPAR1WAY,[29]). 3. Results 3.1. Vegetation structure In the gallery forest (plot CS-6), no variation wasrecorded for the canopy cover and for the density of understorey vegetation (Table 1). However, litter wasdeeperduringthedryseason,butlittercoverwasgreaterduring the rainy season (Table 1).In the dry forest (plot N-5), I recorded a significantdifference in the canopy cover and in the density of un-derstorey vegetation (greater during the rainy season).I recorded no difference in the litter depth, but littercover was greater during the rainy season (Table 1).Between the two plots, I recorded a significant differ-ence for the canopy height (greater in the gallery forest).During the rainy season, no difference was recordedfor the canopy cover, but this characteristic habitat wasmodified during the dry season, with a greater canopycover in the gallery forest (Table 1). Density of the un- derstorey vegetation was more important in the galleryforest than in the dry forest during the rainy season, butno difference was recorded during the dry season (Ta-ble 1). I did not record any difference for the number of lianas between the two pots, but there were more stemsin the gallery forest. Litter structure (depth and cover)differed between the two plots and during the two sea-sons (Table 1). 3.2. Seasonal foraging ecology of Coquerel’s Coua In the gallery forest (plot CS-6), no difference wasrecorded in the percent of prey captured by climbingand in the height reached by climbing. In addition, theproportion of prey size and the attack rate did not varybetween seasons(Table 2andFig. 2). A difference was recorded between the two seasonsfor the techniques used ( F  4 , 101 = 30 . 79, P < 0 . 001).In the logged forest, Coquerel’s Coua used more oftenglean during the dry season and more often leap andsally during the rainy season (Table 2). I recorded a sig-nificant difference in the substrates used by Coquerel’sCoua between seasons ( F  3 , 102 = 13 . 31, P < 0 . 001)with ground more often used during the dry season butthe foliage more often used during the rainy season (Ta-ble 2).In plot N-5, no difference was recorded in the per-cent of prey captured by climbing and in the heightreached by climbing (Table 2). The proportion of prey size varied between seasons with more small prey cap-tured during the rainy season (Fig. 2). The attack ratevaried between seasons with more prey captured duringthe dry season(Table 2). A difference was recorded be- Table 1Variations of the habitat structure, according to the plots and the seasonsHabitat variables Gallery forest (plot CS-6) Dry forest (plot N-5) N-5 vs. CS-6Rainy season Dry season Result Rainy season Dry season Result Rainy season Dry seasonCanopy height (m) 14 . 10 ± 2 . 42 14 . 25 ± 3 . 04 – 10 . 85 ± 3 . 55 10 . 63 ± 2 . 21 – t  = 5 . 40 *** –Canopy cover (%) 70 . 60 ± 14 . 6 73 . 17 ± 22 . 49 t  = 0 . 70 ns 74 . 50 ± 13 . 07 59 . 49 ± 20 . 16 t  = 5 . 62 *** t  = 1 . 54 ns t  = 4 . 20 *** Visibility at 1.5 m (m) 6 . 39 ± 2 . 70 7 . 50 ± 2 . 20 t  = 1 . 87 ns 5 . 03 ± 2 . 66 6 . 22 ± 1 . 80 t  = 4 . 66 *** t  = 3 . 96 *** t  = 0 . 57 nsNumber of lianas ( / 4 m 2 ) 1 . 38 ± 1 . 18 – – 1 . 88 ± 1 . 55 – – t  = 1 . 87 ns –Number of stems ( / 4 m 2 ) 0 . 58 ± 0 . 76 – – 2 . 53 ± 2 . 55 – – t  = 5 . 16 *** –Litter depth (mm) 14 . 90 ± 5 . 39 32 . 50 ± 9 . 89 t  = 10 . 31 *** 18 . 59 ± 5 . 71 19 . 21 ± 8 . 27 t  = 0 . 37 ns t  = 2 . 96 ** t  = 6 . 49 *** Litter cover (%) 92 . 60 ± 12 . 26 55 . 83 ± 24 . 38 t  = 6 . 15 *** 72 . 36 ± 38 . 74 22 . 95 ± 11 . 61 t  = 5 . 81 *** t  = 3 . 46 *** t  = 4 . 78 *** Variations are calculated by a t  -test (see text). ** P < 0 . 01; *** P < 0 . 001; ns: not significant ( P > 0 . 05).
Related Documents
View more...
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks