ORNITOLOGIA NEOTROPICAL ______Volume 12 2001 No. 3 ______

ORNITOLOGIA NEOTROPICAL 12: 189–204, 2001 © The Neotropical Ornithological Society

USE OF SPACE BY A PAIR OF SALVIN’S (MITU SALVINI) IN NORTHWESTERN COLOMBIAN AMAZON

Juan Luis Parra1, Marta Agudelo, Yaily Molina & Gustavo Londoño

Departamento de Biología, Universidad de Los Andes, Bogotá, Colombia.

Resumen. – Uso y preferencia de hábitats por un grupo familiar de Paujiles Camaranos (Mitu sal- vini) en Amazonia noroccidental Colombiana. – Entre Enero y Julio de 1999 estudiamos el área de rango de hogar, el uso y preferencia de hábitats, y sus relaciones con la disponibilidad de frutos en un grupo familiar de Paujiles Camaranos (Mitu salvini) en el Parque Nacional Natural Tinigua, Amazonia noro- ccidental Colombiana. El grupo de paujiles era seguido continuamente a lo largo del día, mientras registrá- bamos sus actividades, el tiempo invertido en cada hábitat, y la trayectoria recorrida. El uso de los hábitats fue evaluado mediante el tiempo invertido y el área utilizada de cada hábitat. La disponibilidad de frutos fue estimada por medio de trampas y conteos visuales. La disponibilidad de los hábitats se estimó como la representación proporcional de cada hábitat en el área de estudio y en el rango de hogar. El área del rango de hogar durante el período de estudio fue de 72 ha e incluyó los bosques ripario, inundable, de ladera y maduro. Se presenciaron variaciones marcadas en el uso relativo de los hábitats que estuvieron frecuente- mente relacionadas con cambios en la distribución y abundancia de los recursos alimenticios, principal- mente frutos en el bosque maduro e inundable, y posiblemente invertebrados en el bosque de ladera y ripario. El tiempo total invertido y el área utilizada en el bosque maduro e inundable estuvieron significati- vamente correlacionados con cambios en la disponibilidad de frutos. Nuestros resultados sugieren que la búsqueda de alimento, específicamente frutos e invertebrados, es uno de los factores principales en deter- minar los movimientos de los paujiles. Sin embargo, durante la etapa reproductiva, actividades como defensa del nido y territorio, o proteger y ayudar a alimentar a la hembra, ejercen también influencias con- siderables. Se comenta acerca de la importancia de cada hábitat para los paujiles y se discuten implicaciones para su conservación. Abstract: – We studied home-range size, habitat-use, habitat preferences, and their relationships with fruit availability by a family of Salvin’s Curassows (Mitu salvini) between January and July 1999 at Tinigua National Park in the northwestern Colombian Amazon. We followed the group closely throughout the day, registering its activities, the time spent in each habitat, and the trajectory traveled. Habitat-use was evalu- ated in terms of area used and time spent in each habitat. Fruit availability was estimated in each habitat by ______1Current address: Cra 30 # 147-33, Apto 201, Int.6, Bogotá, Colombia, 2165873. E-mail: [email protected]

189 PARRA ET AL. fruit traps and visual counts. Habitat availability was estimated as the proportional representation of each habitat in the study site and the home range. Home-range size was 72 ha during the study period and it included riparian, flooded, open-degraded, and mature forests. There were marked temporal shifts in the relative use of habitats, often related to changes in the distribution and abundance of food resources, mainly fruits in flooded and mature forests, and possibly invertebrates in open-degraded and riparian for- ests. Area used and total time spent in flooded and mature forests were correlated with fruit availability on a weekly basis. Flooded forest, where nesting places and some of the most important food items can be found with a relatively high abundance, was the most preferred habitat overall. Our results suggest that the search for food, mainly fruits and invertebrates, plays a determinant role on the ’s space use. Nonetheless, during the breeding period, activities such as territory and nest defense, or protecting and helping feeding the female may become of high relevance. The importance of each habitat and conserva- tion implications for the curassows are discussed. Accepted 2 February 2001. Key words: Mitu salvini, Salvin’s Curassow, Paujil Camarano, home-range size, habitat-use, habitat selection, temporal variation.

INTRODUCTION movements are frequently reported for fru- givorous and nectarivorous and tend to Knowing the main factors determining the be related with changes in the distribution of movements of an organism is very useful, food (Loiselle & Blake 1992). Large terrestrial since it allows us to make predictions about herbivorous birds such as cracids, trumpeters, changes in the behavior of individuals (e.g., tinamids, and phasianids, are very suitable Pyke 1983) and changes in the structure of study subjects in this area, since they can be populations (Turchin 1998). Movements can followed closely, and detailed and continuous be associated with external and internal fac- observations can be registered. However, tors (Bell 1980, Greenwood & Swingland information on these ’s movements is 1983), whose relative importance can vary in scarce and the few studies reveal no general time, space and between individuals (Cody pattern. Groups of White-winged Trumpeters 1981). Despite the difficulty in identifying the (Psophia leucoptera) defend large territories (72 actual features used by from just the ha) all the year round. While the size of these correlates of its distribution (Cody 1985), territories varies in relation to changes in some general factors have been suggested. overall fruit abundance, movements between Body weight and the distribution and abun- habitats inside the territory appear to be inde- dance of prey appear to be the main factors pendent of relative food availability (Sherman associated with home-range and territory size & Eason 1998). Home-ranges recorded for (Schoener 1968, Mace et al. 1983). The distri- individuals of the Boucard Tinamou (Cryp- bution and abundance of resources (e.g., turellus boucardii) are also large (20 ha in aver- food, mates, nesting, perch and roosting age) and show strong temporal variations sites), competition, and predator density are associated with breeding condition (Lancaster suggested as some of the main factors deter- 1964). During the breeding season, home- mining the use and selection of habitats in range sizes increase, possibly as an adaptation birds (Cody 1985). to encounter more mates. Wild Turkeys The high seasonality in fruit, flower, and (Meleagris gallopavo) also show considerable leaf production in the tropics (Van Schaik et seasonal variation in home-range size and al. 1993), makes the study of space use in her- habitat-use that appears to be mostly driven bivorous birds very interesting. Small scale by breeding activities, such as the search for

190 SPACE USE BY A GROUP OF SALVIN’S CURASSOWS

FIG. 1. Annual distribution of rainfall at the Center for Ecological Investigations La Macarena (CIEM) (mean ± 2 SD, n = 4 years, from Kimura et al. 1994). Black dots represent the amounts of rainfall regis- tered during our study period (Jan.–Jul., 1999). suitable nest habitat (Badyaev et al. 1996). vini), and Black ( alector) curassows show Nevertheless, food distribution and abun- considerable temporal variation in home- dance were also suggested as important fac- range size and habitat-use, presumably related tors influencing space use. Curassows are to changes in fruit availability (Bernal & Mejía considered one of the most important terres- 1995, Santamaría & Franco 2000). Although trial herbivorous birds in tropical lowland most of these studies document temporal forests, representing a high percentage of the variations in space use, few have simulta- avian biomass (Terborgh 1986) and probably neously measure the possible factors driving executing an important role in the forest it. Understanding these factors is essential for regeneration process through seed predation a better comprehension of population and dispersal dynamics (Strahl et al. 1997b). dynamics, and to improve conservation Little ecological knowledge is available for efforts. most of the species. However, movements In this paper, we present results on the between habitats have been documented. The pattern of space use by a pair of Salvin’s Yellow Knobbed Curassow (Crax daubentonii) Curassows and their two chicks during the moves from riparian to seasonally dry forests first seven months of 1999 in northwestern with the onset of rains (Strahl et al. 1997a). Amazonia. Our objectives were to 1) estimate Apparently, this behavior is associated with the curassows’ home-range size, 2) determine the species breeding biology. The Cozumel which habitats were used by the group, 3) Curassow (Crax rubra griscomi) concentrates find if the use of habitats reflected their avail- near water bodies during the dry season ability, and 4) if changes in habitat-use were (Martinez-Morales 1999). Salvin’s (Mitu sal- related to changes in fruit availability. We also

191 PARRA ET AL. contrast our results with those obtained in canopy, and some Rubiaceae (Psychotria spp.) previous studies on the same pair of curas- and “platanillos” (Heliconia spp.) in the under- sows (Yumoto 1999, Santamaría & Franco story; 3) Open-degraded forest: Localized on 1994, 1996, 2000). steep slopes and erosion fronts. This forest is notably heterogeneous, but can be character- STUDY AREA ized by the high diversity of vines, lianas, and bamboos, and by the presence of some dis- The study took place at the CIEM (Center for tinctive plants like Carludovica palmata and Ecological Investigations La Macarena), a some Melastomataceae; 4) Mature forests: biological field station located in the Amazon Localized on high flatlands and hill ridges. forests between the eastern Andes and the This forest contains the most diverse flora. Sierra de La Macarena. The station lies on the The canopy is continuous with emergent western margin of the Rio Duda, about 13 km trees reaching 30 m. The understory is rela- from its entrance into the Rio Guayabero (c. tively clear. Distribution of the habitats in the 2°40’N, 74°10’W; 420 m.a.s.l.). Annual mean study area can be seen in Figure 2. temperature and rainfall are 25°C and 2600 Previous studies on fruit phenology in the mm respectively (Kimura et al. 1994). There is study site (Stevenson et al. 1998) showed fruit- a marked seasonality in rainfall, with a dry ing peaks at the beginning of the rainy season season from about December to March, and (Mar.–May) and during the dry season (Dec.– the rest of the year rainy. Precipitation mea- Feb.). Fruit production declines between sured during our study period did not differ these two peaks reaching to a minimum at the from average. We defined three climatic sea- end of the rainy season (Sep.–Nov.). sons on the basis of the precipitation cycle registered: 1) a dry season (months with less METHODS than 100 mm total rainfall), 2) beginning of rains (transition months with more than 100 Study subject. The Salvin’s Curassow inhabits mm rainfall but without a constant precipita- lowland tropical forests east of the Andes tion pattern), and 3) middle of rains (rainy (Delacour & Amadon 1973). Since 1990, at months with a high and similar amount of CIEM, a curassow couple was habituated to rainfall) (Fig. 1). human presence, allowing detailed observa- Four types of habitats have been tions of several ecological aspects for long described at the study site (Hirabuki 1990): 1) and continuous periods of time (Santamaría Riparian forest: located on recently formed & Franco 2000). Salvin’s Curassows have a sand beaches, involving from the earliest complex diet including fruits (35–45%), seeds stands of succession, dominated by grami- (15–25%), flowers (1–3%), leaves (10–15%), noids (Gynerium sp.) and bushes (Tessaria inte- invertebrates (5–15%), soil (3–4%) and others grifolia), to older stands, dominated by young (4–6%) (Agudelo 2000). Consumption of Cecropia trees and a variety of “platanillos” these food categories shows considerable (mostly Heliconia spp.); 2) Flooded forest: temporal variations, sometimes coupled with Localized on flatlands that are partially inun- variations in habitat-use (Santamaría & dated seasonally. The seasonally flooded part Franco 2000). Their breeding period starts in is dominated by Laetia corymbulosa and Helico- the beginning of the dry season (Dec.) and nia marginata. The rest of the forest is charac- can last until May if the first reproduction tri- terized by a higher diversity, mostly Guarea als are unsuccessful (Santamaría & Franco guidonia, Ficus spp. and old Cecropia trees in the 1994). We believe curassow behavior is little

192 SPACE USE BY A GROUP OF SALVIN’S CURASSOWS

FIG. 2. Temporal variation of the home-range areas used by the Salvin’s Curassows through the study period. First illustration (Total) represents total home range area. Boundaries of the study site were arbi- trarily defined. affected by observer presence although we days a week, from 18 January to 24 July (187 cannot discard the possible influence on the days) 1999. behavior of potential predators and other interacting species. The group was followed Home-range and habitat-use. Habitat-use was continuously from 06:00 to 18:00 h, seven evaluated on the basis of two measurements:

193 PARRA ET AL. time spent and area used by the group of throughout the year, not just for the adults, curassows in each habitat. We followed the but also for their chicks during the parental- curassows closely (3–5 m), noting the time of care period. At a minor scale, thus being entrance and departure from any habitat. dependent upon the previous decision, the Each habitat was delimited by a forest trail, group of curassows must decide how to use except for riparian forest, delimited by a con- the available habitats within their home-range spicuous depression in the soil level and a (second order of selection sensu Johnson marked change in vegetation structure and 1980). At this scale, decisions need not to be composition. We also recorded information influenced by all the factors taken into about the trajectory traveled by the curassows, account previously. In fact, decisions at this i.e., we noted trail crossings, visits to fruit scale could be based on entirely different rea- patches, streams, and any place, which loca- sons. An approximation to visualize how hab- tion was previously known by us. The study itats are selected within the home-range is to site counts with an elaborated trail system, analyze habitat-use at a smaller time-scale, for each trail marked every 25 m, which allowed example weeks. We obtained the weekly pro- us to determine the location of the birds. We portions of time spent in and area used of used the information gathered to outline the each habitat, and contrasted them with the daily trajectory of the group on detailed maps proportions available in the home-range. of the study site. The area used by the group For the first order of selection, we graphi- was estimated by overlapping the daily trajec- cally contrasted the proportions of area of tories and taking the peripheral trajectories each habitat in the home-range with the pro- (or parts of a trajectory) as the boundaries of portions of area of each habitat in the study the area used. Then, using the Sigma Scan site. Boundaries of the study site were arbi- Image software, we obtained the areas used trarily defined by us, based on the range cov- by the group. Home-range was defined as the ered by the trail system. Assuming curassow area used by the curassows during their forag- movements were at random and that they ing, rest, parental care and vigilance activities showed no preference for any habitat, the (Burt 1943), and was calculated using all daily proportion of time spent in each habitat trajectories registered. should be proportional to the area occupied by that habitat in the study area or home- Habitat preferences. Habitat selection can be range. We contrasted the proportions of time viewed as a hierarchical process, in which ani- spent in each habitat (total time in habitati/ mals must make decisions at various stages or total observation time) with habitat availabil- scales (Johnson 1980). Measurements of habi- ity, estimated as the proportional representa- tat-use, availability, and the criteria on which tion of each habitat in the study site. decisions are taken, can vary at different For the second order of selection, we scales (Sodhi et al. 1999). For instance, one of obtained the mean and 95% confidence inter- the decisions the group of curassows must vals (CI) for the proportions of area of each make is which areas to include in their home- habitat used weekly, and contrasted them with range, and how to distribute their time among the proportions of area of each habitat in the the available habitats in the study site (first home-range. Using time, we also calculated order of selection sensu Johnson 1980). At this the mean and 95% CI of the proportions of scale, curassows must take into account that time invested in each habitat on a weekly their home-range includes areas with suitable basis, and contrasted them with the propor- nesting sites, where food can be harvested tions of habitats in the home-range. To calcu-

194 SPACE USE BY A GROUP OF SALVIN’S CURASSOWS

FIG. 3. Temporal variation (A) in habitat use evaluated from the total time spent per week by Salvin’s Curassows in each habitat, (B) in habitat use evaluated from the total area used per week in each habitat, and (C) fruit abundance in each habitat. late mean and CI for the proportions, we p is the actual proportion), so resultant data performed “arcsine” or “angular transforma- would have a nearly normal distribution (for tions” on the proportions: p’ = arcsine √p details, see Zar 1996). After obtaining the (where p’ is the transformed proportion and mean and CI for these transformed values,

195 PARRA ET AL.

TABLE 1. Spearman's correlation between habitat cally consume fallen fruits or fruits from use (total time spent, total area used, and average plants within their reach, standing on the daily time spent foraging per week per habitat) and ground; they rarely feed while perched on fruit availability. trees or higher than 1.5 m. To estimate the density of fallen fruits, we placed five fruit Time Area Time foraging traps of 1 m2 systematically (one every 11 m) Riparian 0.577*1 0.476 0.310 in each plot. To estimate the density of fruits Flooded 0.780** 0.264 0.648* from plants within the curassow’s reach, we Open-degraded 0 . 0 1 1 0 . 4 7 3 0.099 counted the number of fruits less than 1.5 m Mature 0.617* 0.672* 0.744** high, within each plot. Each week, the fruits 1Correlation significance (2-tailed): * P < 0.05, **P consumed by the curassows were counted < 0.01. and identified with the help of the guide for the fruits of the region (Stevenson et al. 2000). they were changed back into proportions by Dry weight of the fruits was obtained by dry- the formula: p = (sin p’)2 . ing the fruits in an oven until the weight stabi- When estimating more than two CI simul- lized. Fruit abundance was expressed as dry taneously, the probability that any one interval weight of fruits/m2 per week per habitat. This is incorrect increases beyond α and is partially measurement was obtained by taking the dependent upon the number of estimates means for the plots[Σdry weights of fruits (g) being made (Hopkins & Gross 1970, fide Neu per trap (m2)/n, where n = number of traps in et al. 1974). For this reason, it is suggested to the plot] and averaging them for a habitat 2 use t(1-α/2k),ν values in the construction of the [Σmean for a plot (g/m )/n, where n = num- CI for the mean proportions: p ± t(1-α/2k),ν sp , ber of plots in a habitat]. where p = mean proportion, t = Student’s t, To determine if habitat-use was related to α = significance level, k = number of habi- fruit availability, we performed Spearman cor- tats, ν = degrees of freedom, sp = standard relation analyses between weekly fruit avail- deviation. If the proportion available lies ability and: i) total time spent, ii) total area below (above) the lower (upper) limit of the used, and iii) average daily time spent foraging CI for the proportion of use, then the conclu- for a week in each habitat. To prevent the use sion is that the group of curassows is choos- of temporal autocorrelated measurements of ing (avoiding) that habitat relative to the habitat-use in consecutive weeks, we only others (Cherry 1996). For the analyses, we used data from interspersed weeks (i.e., weeks used only “complete days”, defined as those 1, 3, 5 and so on), which were not correlated. with a minimum observation time of 10 h (n We used only “complete days” for the analy- = 156 days). sis.

Fruit counts. We estimated the abundance of RESULTS fruits available for the group of curassows by establishing 0.01 ha plots (50 x 2 m) in each Home-range and habitat-use. The group of curas- habitat, in a number proportional to habitat sows used 72 ha during the seven months of area in the study site. Twelve plots were observation. This area included all habitats placed randomly in mature forest, and five in defined for the study site: riparian (3.5%), each of the rest (we oversampled riparian for- flooded (35.6%), open-degraded (28.8%) and est in order to obtain a variability measure mature (32.1%) forests (Fig. 2). There were within this habitat). Salvin’s Curassows typi- considerable changes, both in the use among

196 SPACE USE BY A GROUP OF SALVIN’S CURASSOWS

FIG. 4. Contrast between proportions of use (evaluated from the area used and time spent in each habitat in the total home range by Salvin’s Curassows) and proportions available in the study site (first order of selection). habitat types and in the use of each habitat home range did not differ from the propor- through time (Fig. 2). The overall pattern of tional representation of this habitat in the habitat-use based on area used per habitat study site. Results are presented in Figure 4 was, in general, very similar to the one (first order) and Table 2 (second order). encountered with time spent per habitat Notice that results for the first order of selec- (Figs 3A and 3B). Most or all measurements tion were only evaluated graphically. of habitat-use were statistically correlated to fruit availability in flooded and mature Fruit counts. Fruit availability varied through forests, while none or just one in open- the study period with a pattern similar to the degraded and riparian forests (Table 1, Figs one encountered in previous studies includ- 3A–3C). ing all fruits (Stevenson et al. 1998). Results must be compared with caution, since our Habitat preferences. At both orders of selection, results only include fruits consumed by the curassows spent a greater proportion of time curassows. There were two peaks in fruit in and used a larger proportion of area of availability including all habitats: one in late flooded forest than expected by its availabil- March–early April (beginning of rains), fol- ity, but spent a lesser proportion of time in lowed by the major peak in June–July (middle and used a smaller proportion of area than of rains). Nevertheless, the peak in number of expected in mature forest. Open-degraded species in fruit was recorded in late March forest was used in proportions according to and April (24–27 spp.). Relative fruit avail- its availability at both orders of selection. The ability between habitats is shown in Figure proportion of area used weekly of riparian 3C. A total of 70 species of fruits were forest was smaller than expected by its avail- recorded by the fruit counts out of a total of ability in the home range. Nonetheless, the 93 species of fruits consumed by the curas- proportion of area used of this forest in the sows during the study period (Agudelo 2000).

197 PARRA ET AL.

TABLE 2. Proportions and 95% confidence intervals of time spent and area used weekly by a pair of Salvin's Curassows and their chicks in relation to the proportions available in their home range.

Habitat Area Availability Time Used 95% confidence Available Used 95% confidence proportion (%) interval proportion (%) proportion (%) interval Mature forest 0.111* (0.066–0.167) 0.321 0.085* (0.051–0.126) Open-degraded 0.262 (0.219–0.309) 0.288 0.263 (0.203–0.328) Flooded forest 0.557* (0.465–0.647) 0.356 0.571* (0.482–0.659) Riparian forest 0.021* (0.015–0.029) 0.035 0.030 (0.020–0.041)

*Selected habitats (preferred or avoided).

DISCUSSION Theoretical and experimental studies have found an inverse correlation between home- Home-range size. The group of curassows used range size and food abundance (Hixon 1980, fairly large areas, as expected by their size, Schoener 1980, Keeley & McPhail 1998). Our weight and mostly herbivorous diet (Schoener results showed that area used by the curas- 1968, Mace et al. 1983). Previous studies on sows was positively related to fruit availability the same group of curassows (1990–1994) in mature forest (Table 1). Nonetheless, if we (Santamaría & Franco 2000), and on a group perform the correlation analysis with the total of Black Curassows (Aug. 1993–Jan. 1994) area used during a week and total fruit avail- (Bernal & Mejía 1995) registered even larger ability, we find no relationship between the areas: 150 and 185 ha, respectively. Differ- two variables. Indeed, the correlation coeffi- ences in home-range sizes found between cient turns out to be negative (rS = –0.214, P these studies and ours seem to reflect our lack = 0.482, n = 13). Figure 5 shows the relation- of observations between August and Decem- ship between total fruit availability and home ber, a transition period between the rainy and range size on a weekly basis. The relationship the dry season, characterized by an overall is not strong or consistent and this might be fruit scarcity and a highly clumped distribu- due to the fact that curassows are not strict tion of the fruits available (Stevenson et al. frugivores (Agudelo 2000, Santamaría & 1994, 1998). Curassows movements during Franco 2000). More data on the abundance of this period appear to be the most extensive, as other food resources, like seeds, leaves and the birds visit new places in search for the few invertebrates are needed to rigorously test this fruit patches available (Santamaría & Franco relationship. 2000), sometimes wandering out of their usual home-range limits (X. Bernal, pers. Habitat-use. The group of curassows showed com.). Unlike White-winged Trumpeters, marked changes in habitat-use over time, whose territory boundaries are inflexible, thus which appear to be related in flooded and not permitting expansions during food-scar- mature forest principally with changes in the city periods, curassow’s home-range bound- distribution and abundance of fruits. Fruit is aries appear to be fairly flexible, allowing the the food-category most consumed by this group to stretch or expand its home-range group of curassows (Agudelo 2000), as has limits according to external conditions and been reported for other curassows and eco- their internal requirements. logically related birds (Crax alector: Érard et al.

198 SPACE USE BY A GROUP OF SALVIN’S CURASSOWS

FIG. 5. Relationship between total area used by Salvin’s Curassows and total fruit availability on a weekly basis.

1991, Jiménez et al. 2001; Mitu mitu: Cuadros pal source of protein, should be of 1989; Psophia leucoptera: Sherman & Eason high importance in the diet of curassows and 1998; P. crepitans and Tinamus major: Érard et their chicks. The group of Salvin’s Curassows al. 1991). consume a variety of invertebrates from mol- Variation in the use of flooded forest was lusks to insects (Santamaría & Franco 2000), highly related to changes in fruit availability in but they show a conspicuous affinity for this habitat. During the beginning of rains crabs (Potamoidea), which can be found when fruit was scarce in this habitat, the readily in streams and small water bodies, group of curassows spent less time and used specially during the rainy season (pers. less area relative to other seasons when fruit observ.). In the dry season (December–Feb- abundance was greater. The fruit most con- ruary), streams and water bodies are dry. sumed by the group of curassows was Guarea During this period, crabs appear to estivate guidonia, which is one of the dominant trees in deep in burrows (Burggren 1988), and are the nonflooded areas of flooded forest not available to curassows. With the onset of (Stevenson et al. 1999). The two peaks of fruit rains, crabs come out of their holes to start production of Guarea guidonia (January and reproductive and feeding activities, becom- May–June) coincide with the peaks of total ing again available for curassows. A relatively time spent in and area used of flooded forest. high number of streams are found in open- On the other hand, changes in the use of degraded forest (Fig. 2) where the curassows open-degraded forest were not associated spent long periods of time searching for crabs with changes in fruit availability. We rather from the beginning of the first rains in mid- think these changes are related to shifts in February until middle of rains, when the invertebrate availability (pers. observ.). Con- streams’ water level becomes high, preventing sumption of invertebrates is very variable on curassows from their use. Curassows used a daily basis, but in average, they are con- this habitat more than expected during Feb- sumed in relatively large quantities among ruary and the middle of rains when fruit seasons (Agudelo 2000) and being the princi- abundance was relatively low (Fig. 3).

199 PARRA ET AL.

Use of mature forest showed marked able to track short term variations in the changes that seem to have been influenced by abundance and distribution of different variations in fruit availability. In the dry and resources and that the group might be show- middle of rains seasons, use of this habitat ing an optimal foraging behavior, assuming was minimal. In contrast, during the begin- that time spent foraging is positively related to ning of rains, when fruit and possibly crab the rate of energy gain (Stephens & Krebs availability were at peak, curassows spent 1986). large amounts of time in this habitat. How- Other factors also appear to have a strong ever, based on our observations, we think that influence on curassow movements. In Janu- the group of curassows used this habitat prin- ary, although most of the time was spent in cipally to obtain fruits which could be easily flooded forest, a relatively small area was used found in known patches that were visited (Figs 3A and 3B). During this month, the repeatedly. female was incubating in flooded forest and Although total time spent in riparian for- the male remained close to the nest, booming, est was related to fruit availability in this habi- resting, and feeding. It is possible that during tat, we think that major changes in the use of this month, use of space was principally this forest were in response to changes in determined by breeding activities, like defend- invertebrate availability. In the middle of rains, ing the territory, and protecting and feeding streams coming from open-degraded forest the female in the case of the male, or incubat- formed a large pool in riparian forest where ing and defending the nest in the case of the the curassows invested large amounts of time female. looking for dead invertebrates and crabs. Other factors such as predator density Nevertheless, we cannot discard the possibil- and competition for other resources could ity of the important influence of fruits in also influence habitat-use at smaller or greater determining the use of this habitat. time scales. Relative use of habitats was also related to relative fruit availability among habitats. In Habitat preferences. Although the group of the dry and middle of rains seasons, when curassows showed a marked preference for flooded forest was the habitat most used flooded forest, this does not mean that other (Figs 3A and 3B), fruit availability in this hab- habitats were less important. Our study has itat was the highest relative to others (Fig. demonstrated that all habitats play an impor- 3C). At the beginning of the rainy period, tant role on the curassows movements, at when the most used habitats were open- least during some periods of the year. The degraded and mature forests (Figs 3A and flooded forest is a key habitat for this group, 3B), relative fruit availability in these habitats not only because it provides suitable nesting was the highest (Fig. 3C). Shifts in habitat-use sites, but also because some of the most by this group of curassows appear to be a important food items, including fruits (i.e., strategy to maintain the consumption of fruit Guarea guidonia, Trichilia pleeana, Costus spiralis, and invertebrates constant throughout the Psychotria psychotriaefolia), leaves (i.e., Crateva year. Fruits and invertebrates might be the tapia, Trichanthera gigantea) and invertebrates principal sources of energy and animal pro- (i.e., earthworms, leaf-litter spiders, coleop- tein for these birds (Jiménez et al. 2001). Rela- teran larvae, millipedes) are found in relatively tionships found between fruit availability and high abundance. average daily time spent foraging within habi- The group of curassows spent a lesser tats support the notion that curassows are proportion of time and area than expected in

200 SPACE USE BY A GROUP OF SALVIN’S CURASSOWS mature forest. Santamaría & Franco (1996) Curassows are capable of surviving within a found that this habitat was one of the most single habitat, and if so, what characterizes used aside from the flooded forest. Although this habitat. use was not compared to availability in the At our study site, the flooded forest Santamaría & Franco study, we believe that seems to be the only habitat providing suit- observed inconsistencies are due to a change able nest sites for the curassows. If flooded in the distribution of habitats. Part of the forests are indeed necessary habitats for the areas of mature forest in the Santamaría’s curassows’ successful breeding, more atten- study did in fact belong to open-degraded tion should be focused on this species, and its forests as reported in other studies conservation status should be revalued, at (e.g., Stevenson 1994). This situation could least regionally. Although the Salvin’s Curas- have led to an overestimation of mature sow is not considered a threatened species, forest use, and could also have been the cause mainly because a large part of available habi- of the minimal use of open-degraded tat is presumed to be still intact (del Hoyo et forests registered by them. However, al. 1994), flooded forests, at least in north- mature forest was the most used habitat dur- western Amazonia, are rapidly being colo- ing some periods (April), and it is also possi- nized (Hilty 1985). ble that we underestimated its use by not Until now, we only dealt with implications taking into account the August–December for the species in concern. We still do not period. We also cannot discard the possibility have the information to speculate on implica- of different patterns of habitat-use between tions for the ecosystem function and stability. years. Although curassows appear to be very poor seed dispersers (de los Rios 1997, Yumoto Conservation implications. Distinctive features of 1999, Santamaría & Franco 2000), their role this group’s use of space, like the use of large as seed predators might be fundamental in areas including diverse habitats, and the sea- maintaining plant community structure and sonal shifts in habitat-use and habitat prefer- diversity. Loss of these birds could have great ences, set considerable complications for its consequences, not only in the composition of conservation. Protecting small isolated forest avian communities, but also in some impor- fragments, one of the most popular “conser- tant ecological processes. vation strategies” nowadays, seems not to be a good alternative for these birds. Although ACKNOWLEDGMENTS some curassows, such as Crax globulosa (Ben- nett 2000) and Crax mitu (Karr et al. 1990), We thank Rosario Castañeda, Iván Jiménez, appear to use a single habitat, the exclusive Yadira Rojas and Juliana Cursio for their field nesting requirements and the complex diet of assistance. IDEA WILD and Fundación Salvin’s Curassows make this possibility OMACHA donated basic equipment for the remote. Small scale movements, such as project. We also thank Los Andes University, those exhibited by this group of curassows, specially Carlos Arturo Mejía for letting us have been poorly documented. Species use the CIEM biological field station. S. with low tolerance of habitat disruption Arango, J. Blake, C. Cadena, C. Howell, I. that undergo seasonal movements are likely Jiménez, M. Martinez-Morales, G. Kayser, to be more susceptible toforest frag- and an anonymous referee made substantial mentation (Stotz et al. 1996). It is of great contributions and comments on earlier drafts concern to examine if groups of Salvin’s of the paper.

201 PARRA ET AL.

REFERENCES Greenwood, P. J., & I. R. Swingland. 1983. Animal movement: Approaches, adaptations, and con- Agudelo, M. S. 2000. Efecto de la variación tempo- straints. Pp. 1–6 in Greenwood, P. J., & I. R. ral, individual y del ciclo reproductivo sobre la Swingland (eds.). The ecology of animal move- dieta de un grupo de Paujiles Camaranos Mitu ment. Clarendon Press, Oxford. salvini en el Parque Nacional Natural Tinigua, Hilty, S. L. 1985. Distributional changes in the Meta. Tesis de Grado, Univ. de Los Andes, Colombian avifauna: A preliminary blue list. Bogotá, Colombia. Ornithol. Monogr. 36: 1000–1012. Badyaev, A. V., W. J. Etges, & T. E. Martin. 1996. Hirabuki, Y. 1990. Vegetation and landform struc- Ecological and behavioral correlates of varia- ture in the study area of La Macarena: A physi- tion in seasonal home-ranges of Wild Turkeys. ognomic investigation. Field Stud. New World J. Wildl. Manage. 60: 154–164. Monkeys La Macarena Colombia. 3: 35–48. Bell, W. J. 1980. Searching behaviour. Chapman Hixon, M. A. 1980. Food production and competi- and Hall Animal Behaviour series, Cambridge, tor density as the determinants of feeding terri- UK. tory size. Am. Nat. 115: 510–530. Bennett, S. E. 2000. The status of the Pirui (Crax del Hoyo, J., A. Elliot, & J. Sargatal. 1994. Hand- globulosa) in Colombia: A brief overview. Bull. book of the birds of the world. Volume 2: New Netherlands Committee for the World Conser- World vultures to Guineafowl. Lynx Editions, vation Union (IUCN)/Birdlife/World Pheas- Barcelona. ant Association (WPA) Cracid Specialist Group Jiménez, I., J. L. Parra, M. Agudelo, G. Londoño, & (CSG)10: 18–22. Y. Molina. 2001. Temporal variation in the diet Bernal, X., & C. A. Mejía. 1995. Uso del espacio of a pair of Black Curassows (Crax alector). In por un grupo de Crax alector. Fundación para la Brooks D. M., J. Carroll, J. C. Eitniear, & F. Investigación y Tecnología-Banco de la Gonzalez-G. (eds.). Biology and conservation República, Bogotá, Colombia. of Neotropical in the new mille- Burggren, W. W., & B. R. McMahon. 1988. Biology nium. Spec. Publ. 3, Center for the Study of of the land crabs. Cambridge Univ. Press, Cam- Tropical Birds (CSTB), San antonio, Texas. bridge, New York. Johnson, D. H. 1980. The comparison of usage Burt, W. H. 1943. Territoriality and home range and availability measurements for evaluating concepts as applied to mammals. J. Mammal. resource preference. Ecology 61: 65–71. 24: 346–352. Karr, J. R., S. K. Robinson, J. G. Blake, & R. O. Cherry, S. 1996. A comparison of confidence inter- Bierregaard., Jr. 1990. Birds of four Neotropi- val methods for habitat use-availability studies. cal forests. Pp. 237–269 in Gentry A. H. (ed.). J. Wildl. Manage. 60: 653–658. Four Neotropical rainforests. Yale Univ. Press, Cody, M. L. 1981. Habitat selection in birds: The New Haven, Connecticutt. roles of vegetation structure, competitors, and Keeley, E. R., & J. D. McPhail. 1998. Food abun- productivity. Bioscience 31: 107–113. dance, intruder pressure, and body size as Cody, M. L. 1985. Habitat selection in birds. Aca- determinants of territory size in juvenile steel- demic Press, Orlando, Florida. head trout (Oncorhynchus mykiss). Behaviour 135: Cuadros, B. T. 1989. La dieta del Paujil (Mitu mitu): 65–82. O las vicisitudes de ser frugívoro. Bol. Lima 66: Kimura, K., A. Nishimura, K. Izawa, & C. A. 87–90. Mejía. 1994. Annual changes of rainfall and Delacour, J., & D. Amadon. 1973. Curassows and temperature in the tropical seasonal forest at La related birds. Am. Mus. Nat. Hist., New York. Macarena field station, Colombia. Field Stud. Érard, C., M. Théry, & D. Sabatier. 1991. Régime New World Monkeys La Macarena Colombia alimentaire de Tinamus major (Tinamidae), Crax 9: 1–3. alector () et Psophia crepitans (Psophiidae) Lancaster, D. A. 1964. Life history of the Boucard en forêt Guyanaise. Gibier Faune Sauvage 8: Tinamou in British Honduras. Part II: Breeding 183–210. biology. Condor 4: 253–276.

202 SPACE USE BY A GROUP OF SALVIN’S CURASSOWS

Loiselle, B. A., & J. G. Blake. 1992. Population Trumpeters territories. Ecology 4: 1147–1159. variation in a tropical bird community. Bio- Sodhi, N. S., C. A. Paszkowski, & S. Keehn. 1999. science 42: 838–845. Scale-dependent habitat selection by American Mace, G. M., P. H. Harvey, & T. H. Clutton-Brock. Redstarts in aspen-dominated forest frag- 1983. Vertebrate home-range size and ener- ments. Wilson Bull. 111: 70–75. getic requirements. Pp 32–53 in Greenwood, P. Stephens, D. W., & J. W. Krebs. 1986. Foraging the- J., & I. R. Swingland (eds.). The ecology of ani- ory. Princeton Univ. Press, Princeton, New Jer- mal movement. Clarendon Press, Oxford. sey. Martinez-Morales, M. A. 1999. Conservation sta- Stevenson, P. R., M. J. Quiñones, & J. A. Ahumada. tus and habitat preferences of the Cozumel 1994. Ecological strategies of woolly monkeys Curassow. Condor 101: 14–20. (Lagothrix lagotricha) at Tinigua National Park, Neu, C. W., C. R. Byers, & J. M. Peek. 1974. A Colombia. Am. J. Primatol. 32: 123–140. technique for analysis of utilization-availability Stevenson, P. R., M. J. Quiñones, & J. A. Ahumada. data. J. Wildl. Manage. 38: 541–545. 1998. Annual variation in fruiting pattern using Pyke, G. H. 1983. Animal movements: An optimal two different methods in a lowland tropical foraging approach. Pp. 7–31 in Swingland I. R., forest, Tinigua National Park, Colombia. Bio- & P. J. Greenwood (eds.). The ecology of ani- tropica 30: 129–134. mal movement. Clarendon Press, Oxford. Stevenson, P. R., M. C. Castellanos, & A. Medina. de los Rios, P. T. 1997. Dispersión y predación de 1999. Elementos arbóreos de los bosques de semillas por Paujiles: Mitu salvini (Aves: un plano inundable en el Parque Nacional Nat- Cracidae) en la Amazonia, Colombia. Tesis de ural Tinigua, Colombia. Caldasia 21: 38–49. grado, Univ. de Los Andes, Bogotá, Colombia. Stevenson, P. R., M. J. Quiñones, & M. C. Castell- Santamaría, M., & A. M. Franco. 1994. Historia anos. 2000. Guía de frutos de los bosques del natural del Paujil Mitu salvini y densidades río Duda La Macarena, Colombia. Asociación poblacionales de los crácidos en el Parque para la Defensa de La Macarena, Netherlands Nacional Natural Tinigua-Amazonia Colombi- Committee for IUCN, Tropical Rain Forest ana. Wildlife Conservation Society (WCS), Programme, Bogotá, Colombia. Santafe de Bogotá, Colombia. Stotz, D. F., J. W. Fitzpatrick, T. A. Parker III, & D. Santamaría, M., & A. M. Franco. 1996. Historia K. Moskovits. 1996. Neotropical birds: Ecol- Natural del Paujil (Mitu salvini) en el Parque ogy and conservation. Univ. of Chicago Press, Nacional Natural Tinigua: Base para un Chicago. manejo sostenible?. Pp. 61–70 in de la Cruz H. Strahl, S., J. L. Silva, & R. Buchholz. 1997a. Vari- O. (ed.). Investigación y manejo de fauna para ación estacional en el uso del hábitat, compor- la construcción de sistemas sostenibles. Fun- tamiento de grupo, y un sistema dación Centro para la Investigación en Siste- aparentemente polígamo en el Paujil copete de mas Sostenibles de Producción Agropecuaria – plumas, Crax daubentoni. Pp. 79 in Strahl S. D., CIPAV, Cali, Colombia. S. Beaujon, D. M. Brooks, A. Begazo, G. Santamaría, M., & A. M. Franco. 2000. Frugivory Sedaghatkish, & F. Olmos (eds.). The Cracidae: of Salvin’s Curassow in a rainforest of the Their biology and conservation. Hancock Colombian amazon. Wilson Bull. 112: 473– House Publishers, Surrey, British Columbia. 481. Strahl S. D., S. Beaujon, D. M. Brooks, A. Begazo, Schoener, T. W. 1968. Sizes of feeding territories G. Sedaghatkish, & F. Olmos. 1997b. The among birds. Ecology 1: 123–141. Cracidae: Their biology and conservation. Schoener, T. W. 1980. Simple models of optimal Hancock House Publishers, Surrey, British feeding-territory size: a reconciliation. Am. Columbia. Nat. 121: 608–629. Terborgh, J. 1986. Keystone plant resources in the Sherman, P. T., & P. K. Eason. 1998. Size determi- tropical forest. Pp. 330–344 in Soulé M. E. nants in territories with inflexible boundaries– (ed.). Conservation biology: The science of manipulation experiments on White-winged scarcity and diversity. Sinauer Associates, Sun-

203 PARRA ET AL.

derland, Massachusetts. consumers. Annu. Rev. Ecol. Syst. 24: 353–377. Turchin, P. 1998. Quantitative analysis of move- Yumoto, T. 1999. Seed dispersal by Salvin’s Curas- ment. Sinauer Associates, Sunderland, Massa- sow, Mitu salvini (Cracidae), in a tropical forest chusetts. of Colombia: Direct measurements of dispersal Van Schaik, C. P., J. W. Terborgh, & S. J. Wright. distance. Biotropica 31: 654–660. 1993. The phenology of tropical forests: Adap- Zar, J. H. 1996. Biostatistical analysis. 3d ed. Pren- tive significance and consequences for primary tice Hall, Upper Saddle River, New Jersey.

204