The Diets of Neotropical Trogons, Motmots, Barbets and Toucans

The Condor 95:178-192 0 The Cooper Ornithological Society 1993 THE DIETS OF NEOTROPICAL TROGONS, MOTMOTS, BARBETS AND TOUCANS J. V. REMSEN, JR., MARY ANN HYDE~ AND ANGELA CHAPMAN Museum of Natural Scienceand Department of Zoology and Physiology, Louisiana State University,Baton Rouge, LA. 70803 Abstract. Although membership in broad diet categoriesis a standardfeature of community analysesof Neotropical birds, the bases for assignmentsto diet categoriesare usually not stated, or they are derived from anecdotal information or bill shape. We used notations of stomachcontents on museum specimenlabels to assessmembership in broad diet categories (“fruit only,” “ arthropods only,” and “fruit and arthropods”) for speciesof four families of birds in the Neotropics usually consideredto have a mixed diet of fruit and animal matter: trogons (Trogonidae), motmots (Momotidae), New World barbets (Capitonidae), and toucans (Ramphastidae). An assessmentof the accuracyof label data by direct comparison to independentmicroscopic analysis of actual stomachcontents of the same specimensshowed that label notations were remarkably accurate.The specimen label data for 246 individuals of 17 speciesof Trogonidae showed that quetzals (Pharomachrus)differ significantly from other trogons (Trogon) in being more fiugivorous. Significant differences in degree of frugivory were found among various Trogonspecies. Within the Trogonidae, degreeof frugivory is strongly correlated with body size, the larger speciesbeing more frugivorous. The more frugivorous quetzals (Pharomachrus)have relatively flatter bills than other trogons, in ac- cordancewith predictions concerningmorphology of frugivores;otherwise, bill morphology correlated poorly with degree of fiugivory. An analysis of label data from 124 individuals of six speciesof motmots showed that one species(Electron platyrhynchum)is highly insectivorous,differing significantlyfrom two others that are more frugivorous(Baryphthengus martii and Momotus momota). An analysis of 135 individuals of 12 species of barbets showed that although “fruit only” predominated among almost all species,arthropods are more frequently recorded in the stomachsof speciesin the generaEubucco and Capito than in Semnornis.The highly frugivorous diet of Semnornisspecies is yet another parameter in which they resemble toucansmore than New World barbets. Data from 326 individuals of 32 speciesof toucans showed that the family is remarkably homogeneousin the predomi- nance of fruit in the stomachs of all species.These data suggestthat the degree to which toucansprey upon bird eggsand nestlings,and animal matter in general,is overemphasized. Although our data suggestthat it is safe to assign toucans to a “fruit only” category in community analyses,such assignmentsmust be taken on a genus-by-genusor species-by- speciesbasis in trogons, motmots, and barbets. Key words: Diet; guild; Neotropicalforest bird communities;Trogonidae; Momotidae; Capitonidae;Ramphastidae. INTRODUCTION Karr 197 1,199O; Terborgh 1980; Faaborg 1985; Quantitative analysesof the diets of most species Bierregaardand Lovejoy 1989; Bierregaard 1990; of Neotropical forest birds are virtually nonex- Blake et al. 1990; Karr et al. 1990; Robinson and istent. Therefore, classifications of these species Terborgh 1990; Blake and Loiselle 199 l), and into foraging or diet guilds is usually based on their accuracy is critical for assessingpotential bill morphology, anecdotal observations, or ex- community convergence (Terborgh and Robin- trapolations from nearestrelatives; in some stud- son 1986). Only two such studies(Remsen 1985, ies, the basis for the classification is not stated. Cardosoda Silva and Oniki 1988) however, have Such guild assignments are a frequent and nec- supportedthe assignmentswith analysis of stom- essaryfeature of studies of bird communities of ach contents and quantitative foraging data from Neotropical forest localities (e.g., Orians 1969; the study site itself. Becauseof their heterogeneousdiet, five families of Neotropical birds pose special problems I Received 30 September 1992. Accepted 12 October 1992. for guild assignments: trogons (Trogonidae), z Presentaddress: 11555 Southfork Ave., Apt. 3058, motmots (Momotidae), barbets (Capitonidae), Baton Rouge, LA 708 16. toucans (Ramphastidae), and tanagers (Thrau- [1781 TROGON, MOTMOT, BARBET AND TOUCAN DIETS 179 pidae). Isler and Isler (1987) compiled existing termined the degreeof accuracy of label data by data on stomach contents of the tanagers that direct comparison of label notations of the pre- show that the family is highly heterogeneouswith parator of the specimen to actual contents de- respectto the broad categoriesof insectivory and termined by an independent observer. Then we frugivory. Some speciesseem exclusively insec- used label notations as a first approximation to tivorous, others exclusively frugivorous, and the question of whether speciesin four families others have a mixed diet. Diets of most species are insectivorous, frugivorous, or mixed in diet. in the other four families, however, have yet to More sophisticated categorization awaits de- be analyzed in a comprehensive way. Although tailed, year-round studies of foraging behavior many species in these four families have been and diet, and quantitative, microscopic analyses included as frugivores in compilations of frugiv- of the stomach contents themselves. orousbirds ranging from world-wide (Snow 198 1) to local (Wheelwright et al. 1984) in scope, the MATERIALS AND METHODS degree to which frugivores feed on animal prey Most label data on stomach contents were com- is usually unknown. Most community studies piled from the bird collection of the Museum of (referencesabove) have assignedall specieswith- Natural Science, Louisiana State University. in each of these four families to a single dietary These specimenswere collected primarily in Bo- guild, usually the “omnivore” category. livia and Peru during the dry season; a few in- What is the best way to make such guild as- dividuals were collected in Panama and Costa signments? Direct observations, such as those Rica. Some stomach contents from specimen la- presented by Stiles and Skutch (1989) provide bels from other museums (see Acknowledg- valuable information on diet but are usually bi- ments) were also included (primarily from the ased in favor of conspicuous feeding behaviors dry seasonin Ecuador, Peru, and Brazil, and from and large and identifiable prey items. Although Costa Rica). A few stomachswith contents enu- biasesin digestion rates and poor taxonomic res- merated by Van Tyne (1929) Wagner (1944) olution of food items limit the usefulness of Wetmore (1968) and Boume (1974) were also stomach contents in describing the diets of birds included. Using adults only, each specimen’s no- (Rosenberg and Cooper 1990 and references tation was categorizedas follows: (1) arthropods therein), analysis of stomach or crop contents only; (2) fruit only, including seeds,fruit skins, potentially provides the best data on diet com- or pulp; (3) mixture of arthropods and fruit; (4) position at the level of resolution necessaryfor unidentifiable vegetable matter; or (5) uniden- guild assignments. Although microscopic anal- tifiable “mush.” Presenceof vertebrate remains ysis of the contentscan generatequantitative data was also noted. on diet composition (e.g., Orejuela 1980; Sherry Because these data were taken by many dif- 1984; Rosenberg 1990; Chapman and Rosen- ferent preparatorswho presumably varied in their berg, 199 l), such analyses require considerable skills in recognizing food items, the data set is time investment and expertise in the identifica- heterogeneous. The specimens were collected tion of diet items. Notations on labels of museum from different regions, habitats, and seasons. specimensconcerning stomach contents typical- Sample sizes were insufficient to determine ly lack taxonomic precision and do not state rel- whether these variables contributed to the het- ative proportions of components of the contents. erogeneity. For a few species with the largest These data are also certainly biased against de- sample sizes, informal comparisons did not re- tection of small fragments of arthropods and veal any substantial effectsof region, season,sex, partly digested.fruit pulp. Although inferior to or breeding condition. Clearly, the next step is microscopic analysis in many ways, label nota- to obtain sample sizes adequate for quantitative tions can be usedto addressquestions concerning comparisons among regions, habitats, and sea- broad diet categories,such as whether a taxon is sons. largely insectivorous or fiugivorous. Analysis of Another limitation is the degree of accuracy label data can yield novel results, such as the and resolution of label notations, which seldom degreeof insectivory in hummingbirds (Remsen contain notations on relative proportions of con- et al. 1986). tents. A single fragment of an arthropod in a To explore the usefulness of specimen label stomach full of fruit hypothetically received the data for making guild assignments, we first de- same category assignment in our scheme as a 180 J. V. REMSEN, JR., M. A. HYDE AND A. CHAPMAN stomach with a single fruit fragment in one oth- conservatively a priori at 0.01 rather than the erwise full of arthropod remains. It is unlikely usual 0.05; however, to avoid overlooking trends that most preparatorswould have detected small that might be biologically significant, we mention arthropod fragments, such

Load more