Roosting behavior.

Demography and Natural History of the Common Fruit Bat, Artibeus jamaicensis, on Barro Colorado Island, Panama CHARLES O. HANDLEY, JR., DON E. WILSON, and ALFRED L. GARDNER EDITORS SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY • NUMBER 511 SERIES PUBLICATIONS OF THE SMITHSONIAN INSTITUTION Emphasis upon publication as a means of "diffusing knowledge" was expressed by the first Secretary of the Smithsonian. In his formal plan for the Institution, Joseph Henry outlined a program that included the following statement: "It is proposed to publish a series of reports, giving an account of the new discoveries in science, and of the changes made from year to year in all branches of knowledge." This theme of basic research has been adhered to through the years by thousands of titles issued in series publications under the Smithsonian imprint, commencing with Smithsonian Contributions to Knowledge in 1848 and continuing with the following active series: Smithsonian Contributions to Anthropology Smithsonian Contributions to Astrophysics Smithsonian Contributions to Botany Smithsonian Contributions to the Earth Sciences Smithsonian Contributions to the Marine Sciences Smithsonian Contributions to Paleobiology Smithsonian Contributions to Zoology Smithsonian Folklife Studies Smithsonian Studies in Air and Space Smithsonian Studies in History and Technology In these series, the Institution publishes small papers and full-scale monographs that report the research and collections of its various museums and bureaux or of professional colleagues in the world of science and scholarship. The publications are distributed by mailing lists to libraries, universities, and similar institutions throughout the world. Papers or monographs submitted for series publication are received by the Smithsonian Institution Press, subject to its own review for format and style, only through departments of the various Smithsonian museums or bureaux, where the manuscripts are given substantive review. Press requirements for manuscript and art preparation are outlined on the inside back cover. Robert McC. Adams Secretary Smithsonian Institution SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY • NUMBER 511 Demography and Natural History of the Common Fruit Bat, Artibeus jamaicensis, on Barro Colorado Island, Panama Charles O. Handley, Jr., Don E. Wilson, and Alfred L. Gardner EDITORS SMITHSONIAN INSTITUTION PRESS Washington, D.C. 1991 ABSTRACT Handley, Charles O., Jr., Don E. Wilson, and Alfred L. Gardner, editors. Demography and Natural History of the Common Fruit Bat, Artibeus jamaicensis, on Barro Colorado Island, Panama. Smithsonian Contributions to Zoology, number 511, 173 pages, 69 figures, 62 tables, 1991.—Bats were marked and monitored on Barro Colorado Island, Panama, to study seasonal and annual variation in distribution, abundance, and natural history from 1975 through 1980. Data gathered advances our knowledge about flocking; abundance; feeding strategies; social behavior; species richness; population structure and stability; age and sex ratios; life expectancy and longevity; nightly, seasonal, and annual movements; synchrony within and between species in reproductive activity; timing of reproductive cycles; survival and dispersal of recruits; intra-and inter-specific relationships; and day and night roost selection. Barro Colorado Island (BCI) harbors large populations of bats that feed on the fruit of canopy trees, especially figs. These trees are abundant, and the individual asynchrony of their fruiting rhythms results in a fairly uniform abundance of fruit. When figs are scarce, a variety of other fruits is available to replace them. This relatively dependable food supply attracts a remarkably rich guild of bats. Although we marked all bats caught, we tried to maximize the number of Artibeus jamaicensis netted, because it is abundant Q-h of the total catch of bats on BCI), easily captured by conventional means (mist nets set at ground level), and responds well to handling and marking. An average Artibeus jamaicensis is a 45 g frugivore that eats roughly its weight in fruit every night. These bats prefer figs and often seek them out even when other types of fruit they might eat are far more abundant. They commute several hundred meters to feeding trees on the average, feeding on fruit from one to four trees each night, and returning to a single fruiting tree an average of four nights in succession. The bats tend to fly farther when fewer fig trees are bearing ripe fruit, and they feed from fewer trees, on the average, when the moon is nearly full. These bats, like their congeners, do not feed in the fruiting tree itself. Instead, they select a fruit and carry it to a feeding roost typically about 100 m away before eating it. We utilized radio telemetry to assess feeding rates from the number of "feeding passes"—transits between fruit tree and feeding roost. Bats are often netted while carrying fruit, revealing their diet. Feces also reveal dietary information. Adult female A. jamaicensis live in harems of three to 30 individuals with a single adult male. On BCI the harem groups roost during the day in hollow trees. There is presumably a large population of surplus males that roost together with nonadults of both sexes in foliage. Females commute an average of 600 m from their day roosts to feeding sites, and harem males travel less than 300 m. Twice a year most females give birth to a single young, once in March or April, and again in July or August; active gestation averages about 19 weeks. Juveniles are first netted when they are about ten weeks old, and females usually first bear young in March or April following their year of birth. OFFICIAL PUBLICATION DATE is handstamped in a limited number of initial copies and is recorded in the Institution's annual report, Smithsonian Year. SERIES COVER DESIGN: The coral Montastrea cavernosa (Linnaeus). Library of Congress Caialoging-in-Publicalion Data Handley, Charles O. (Charles Overtoil), 1924Demography and natural history of the common fruit bat, Artibeus jamaicensis, on Barro Colorado Island, Panama / Charles O. Handley, Jr., Don E. Wilson, and Alfred L. Gardner. p. cm. — (Smithsonian contributions to zoology, no. 511) Includes bibliographical references (p. ). 1. Artibeus jamaicensis—Panama—Barro Colorado Island. I. Wilson, Don E. II. Gardner, Alfred L. III. Tide. IV. Series. QL1.S54 no. 511 [QL737.C57] 591 s-dc20 [599.4] 91-13449 @ The paper used in this publication meets the minimum requirements of the American National Standard for Permanence of Paper for Printed Library Materials Z39.48—1984. Contents Page 1. INTRODUCTION by Charles O. Handley, Jr., Don E. Wilson, and Alfred L. Gardner 1 2. PHYSIOLOGY by Eugene H. Studier and Don E. Wilson 9 3. REPRODUCTION IN A CAPTIVE COLONY by Lucinda Keast Taft and Charles O. Handley, Jr 19 4. REPRODUCTION ON BARRO COLORADO ISLAND by Don E. Wilson, Charles O. Handley, Jr., and Alfred L. Gardner 43 5. SURVIVAL AND RELATIVE ABUNDANCE by Alfred L. Gardner, Charles O. Handley, Jr., and Don E. Wilson 53 6. POPULATION ESTIMATES by Egbert G. Leigh, Jr., and Charles O. Handley, Jr.. . 77 7. MOVEMENTS by Charles O. Handley, Jr., Alfred L. Gardner, and Don E. Wilson 89 8. ROOSTING BEHAVIOR by Douglas W. Morrison and Charles O. Handley, Jr. . . 131 9. FORAGING BEHAVIOR by Charles O. Handley, Jr., and Douglas W. Morrison . . 137 10. FOOD HABITS by Charles O. Handley, Jr., Alfred L. Gardner, and Don E. Wilson 141 11. DIET AND FOOD SUPPLY by Charles O. Handley, Jr., and Egbert G. Leigh, Jr.. 147 12. APPENDIX: METHODS OF CAPTURING AND MARKING TROPICAL BATS 151 LITERATURE CITED 167 FRONTISPIECE.—Artibeus jamaicensis. Pencil sketch by Nancy Moran, Barro Colorado Island, Panama^ October, 1976. Demography and Natural History of the Common Fruit Bat, Artibeus jamaicensis, on Barro Colorado Island, Panama 1. Introduction Charles O. Handley, Jr., Don E. Wilson, and Alfred L. Gardner BCI Bat Project In 1974, Charles Handley was invited to develop a project to monitor bats as part of the Smithsonian Tropical Research Institute's (STRI) long-term environmental monitoring program on Barro Colorado Island (BCI). The STRI monitoring program, launched in 1970 and supported by the Smithsonian Environmental Sciences Program (ESP), sought to monitor a wide array of biotic and physical environmental components of the island continuously over a long period of time. The BCI Bat Project was born under the administrative title: "Biomass and energetics of selected populations in Panama: Bats." We wanted to monitor demographic parameters and natural history of all the bats regularly found on BCI. Based on our earlier experiences, we thought the fauna might total 40 species of bats. The length of the project was designed to continue through a generation of bats, however long that might be. The only clue to possible duration was the report (Wilson and Tyson, 1970) of a seven-year-old Artibeus jamaicensis on BCI. At the outset it was evident that ESP funds were spread over Charles O. Handley, Jr., and Don E. Wilson, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560. Alfred L. Gardner, NERC, US. Fish and Wildlife Service, National Museum of Natural History, Washington, D.C. 20560. Review Chairman: W. Ronald Heyer, Smithsonian Institution. Three anonymous reviewers are gratefully acknowledged. too many projects to be able to support a really meaningful monitoring project for bats. Clyde Jones, then Director of the National Fish and Wildlife Laboratory, U.S. Fish and Wildlife Service, offered to provide both financial and personnel support, and the Bat Project became a joint venture of the Smithsonian Institution and the U.S. Fish and Wildlife Service, with Michael A. Bogan, Alfred L. Gardner, and Don E. Wilson joining Handley as field crew leaders. Handley made several trips to BCI in 1975 and 1976 to become familiar with the island and its bats, as well as to determine what was feasible and how to organize the project. On 2 July 1977 a year-round capture and marking program began. With the help of collaborators and dozens of volunteers, we took turns manning the field survey on BCI until November 1980 when this phase of the Project was completed. Thereafter, Handley continued work on BCI on a periodic basis—the fall of 1981, the fall of 1982, and the 12 months from September 1984 through August 1985—with the support of the Smithsonian's ESP, STRI, and National Museum of Natural History (NMNH). The purpose of the continuing study was to maintain the pool of marked bats, refine the demographic data, and gain further information on the biology of the bats, particularly their responses to food sources. During the reconnoitering phase at the beginning of the project we focused much of our attention on developing a reliable, long-lasting, harmless marking system. We established colonies of bats at the National Zoological Park (NZP) that we used in marking experiments and in establishing SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY standards for describing age and reproductive state. As a result, we discarded forearm banding, tattooing, and heat branding and settled on necklacing with stainless steel ball chain. Bandbearing necklaces were first attached to free-living bats on BCI on 18 October 1976. By the fall of 1980 we had placed 18,953 necklaces on bats. The project evolved rapidly. We soon realized that our resources were not sufficient to monitor simultaneously all species of bats on BCI in an effective manner. We settled on studying the bat that had proved to be our principle catch—Artibeus jamaicensis Leach, the common fruit-eating bat. At first we established netting stations at places that, based on experience elsewhere, "looked good for bats," such as stream valleys, trails through open forest, and gaps along ridges * The use of product brand names in this publication is not intended as an endorsement of the products by the Smithsonian Institution. where underbrush and other vegetation did not interfere with the nets or obstruct fly ways. However, it became apparent that abundance of bats coincided with nearby fruit sources and that what at first appeared to be complicated distributional patterns for the island's bats proved to be nothing more that a direct correlation between foraging activity and the uneven distribution and availability of fruit. We improved our netting success by making systematic surveys along the trails on BCI to locate pellets of fruit pulp dropped by feeding bats and then setting our nets nearby where they were most likely to catch bats. We realized early in the project that squeaking bats often attracted others, and in 1978 we began to use the wooden Audubon Bird Call as a substitute to attract bats into the nets. At a good site, an Audubon Bird Call*, an occasional squeaking bat, and one or two nets could produce enough bats to keep everyone busy for hours. In the beginning we sometimes caught only two or three bats in a night, and we were satisfied with catches of 30 bats. Later, TABLE 1-1.—Bats recorded on Barro Colorado Island, Gatun Lake, Panama. Family EMBALLONURIDAE Rhynchonycteris naso (Wied) Saccopteryx bllineala (Temminck) Saccopteryx leptura (Schreber) Cormura brevirostris (Wagner) Centronycteris maximiliani (Fischer) Family NOCTOJONIDAE Noctilio albiventris Desmarest Noctilio leporinus Linnaeus Family MORMOOPIDAB Pteronotus gymnonotus Wagner Pteronotus parnellii (Gray) Family PHYLLOSTOMIDAE Subfamily PHYLLOSTOMINAE Micronycteris brachyotis (Dobson) Mkronycterts hirsuta (Peters) Micronycteris megalolis (Gray) Micronycteris nicefori Sanbom Micronycteris schmidtorum Sanbom Macrophyllum macropkyllum (Schinz) Tonatia bidens (Spix) Tonatia silvicola D'Orbigny Mimon crenulalum (E. Geoffroy) Pkyllostomus discolor Wagner Pkyllostomus haslalus(PMas) Pkylloderma stenops Peters Trachops cirrhosus (Spix) Chrolopterus auritus (Peters) Vampyrum spectrum (Linnaeus) Subfamily GLOSSOPHAOINAE Glossophaga commissar is i Gardner Glossophaga soricina (Pallas) Lonchophylla robusta Miller Subfamily CAROLLINAE Carollia brevicauda (Schinz) Carollia castanea H. Allen Carollia perspicillata (Linnaeus) Subfamily STURNIRINAE Sturnira luisi Davis Subfamily STENODERMATTNAE Uroderma bilobatum Peters Uroderma magnirostrum Davis Platyrrhinus helleri (Peters) Vampyrodes caraccioli (Thomas) Vampyressa nymphaea Thomas Vampyressa pus ilia (Wagner) Chiroderma villosum Peters Mesophylla macconnelli Thomas Artibeus hartii Thomas Artibeus jamaicensis Leach Artibeus lituratus (Olfers) Artibeus phaeotis (Miller) Artibeus watsoni Thomas Ametrida centwio Gray Centurio senex Gray Subfamily DESMODONTINAE Desmodus rotundus (E. Geoffroy) Family THYROPTERTOAE Thyroptera disci/era (lichtenstein and Peters) Tkyroptera tricolor Spix Family VESPERTIUONIDAE Myotis albescens (E. Geoffroy Myotis nigricans (Schinz) Rhogeessa tumida H. Allen Family MOLOSSIDAE Tadarida laticaudata (E. Geoffroy) Molossus bondae J.A. Allen Molossus coibensis J.A. Allen Molossus molossus (Pallas) NUMBER 511 we were disappointed with less than 100 bats per night, and we logged many nights with catches exceeding 200. Our best catch came on 25 October 1979 at a giant Ficus dugandii with ripe fruit where we netted 282 bats in about four hours. As of 1985, we had found 56 species of bats on BCI (Table 1-1). Bonaccorso (1979) categorized the bats of the island into nine "feeding guilds." The distribution of the 56 species among Bonaccorso's guilds is: canopy frugivores (14), groundstory frugivores (4), scavenging frugivores (2), omnivores (4), sanguinivores (1), gleaning carnivores (12), slow-flying hawking insectivores (14), fast-flying hawking insectivores (4), and piscivores (1). A. jamaicensis, the major subject of this report, is a canopy frugivore. By every measure A. jamaicensis is the most widespread and abundant bat on BCI (Table 1-2). On a yearly basis, it averaged 60% of the total catch of bats, and we caught it almost every night that nets were set. Altogether, in the period 1975-1980 (including bats captured before marking with necklaces began), we recorded 17,820 captures of A. jamaicensis. By the end of 1980, we had learned enough about A. jamaicensis including its populations, reproduction, movements, foraging, and physiology to justify a pause to TABLE 1-2.—Measures of abundance of bats captured on BCI during 1979. Bats were netted on 157 nights, and captures (including both marks and recaptures) totalled 9118 bats. Tabulations are by frequency of capture (number and percentage of nights caught), number caught (total), catch per species night (total of a species caught divided by number of nights it was caught), and catch per netting night (total of a species caught divided by total nights of netting). Species Artibeus jamaicensis Uroderma bilobatum Artibeus lituratus Chiroderma villosum Carollia perspicillata Vampyrodes caraccioli Phyllostomus discolor Artibeus phaeotis Carollia castanea Micronycteris hirsuta Vampyressa pusilla Vampyressa nymphaea Pteronotus parnellii Tonatia silvicola Tonatia bidens Artibeus walsoni Micronycteris megalotis Glossophaga soricina Phyllostomus hastatus Mimon crenulatum Rhogeessa tumida Trachops cirrhosus Cormura brevirostris Platyrrhinus helleri Micronycteris brachyotis Macrophyllum macrophyllum Micronycteris nicefori Saccopteryx bilineata Desmodus rotundus Glossophaga commissarisi Myotis nigricans Micronycteris schmidtorum Phylloderma stenops Centurio senex Vampyrum spectrum Uroderma magnirostrum Lonchophylla robusta Carollia brevicauda Artibeus hartii Mean catch per species night 36.3 9.2 6.0 4.4 4.0 3.9 3.3 2.2 2.0 1.8 1.8 1.8 1.7 1.6 1.6 1.6 1.6 1.6 1.6 1.5 1.5 1.4 1.4 1.3 1.2 1.2 1.1 1.1 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 1.0 Rank 1 2 3 4 5 6 7 8 9 10 10 10 13 14 14 14 14 14 14 20 20 22 22 24 25 25 27 27 29 29 29 29 29 29 29 29 29 29 29 Total 5484 551 717 172 428 325 130 235 154 91 87 51 86 113 77 67 35 28 26 50 3 70 7 31 13 11 21 8 12 8 7 5 5 3 2 2 1 1 1 Mean catch per netting night 34.9 3.5 4.6 1.1 2.7 2.1 0.8 1.5 1.0 0.6 0.6 0.3 0.6 0.7 0.5 0.4 0.2 0.2 0.2 0.3 0.02 0.4 0.05 0.2 0.08 0.07 0.1 0.05 0.08 0.05 0.05 0.03 0.03 0.02 0.01 0.01 0.01 0.01 0.01 Nights caught N 151 60 120 39 107 83 40 108 78 52 49 28 51 71 49 43 22 17 16 33 2 49 5 23 11 9 19 7 12 8 7 5 5 3 2 2 1 1 1 % 96 38 76 25 68 53 25 69 50 33 31 18 32 45 31 27 14 11 10 21 1 31 3 15 7 6 12 4 8 5 4 3 3 2 1 1 1 1 1 10SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY in a T, range of 8.0°-33.2°C. The regression coefficient for this relationship (0.933; SE = 0.202) does not differ significantly from 1.0 and we conclude that in A. jamaicensis, Tb parallels Ta. Individual bats maintain a constant range of Tb 6.6°-8.3°C higher than T, throughout the tested T, range. Previously, Morrison and McNab (1967) reported little daily fluctuation in Tb cycles in captive A. jamaicensis at an estimated T, of 27°C. During daylight, some of their data points (13/128 = 10.1%) were significantly lower than average. They demonstrated that A. jamaicensis exhibited the thermoregulatory pattern of a homeothermic endotherm throughout a T, exposure range of 5°-39°C. McNab (1969) obtained similar results in captive A. jamaicensis held for up to two weeks before testing; no bats had reduced Tb relative to their mean temperatures. We reinvestigated these contradictory results and concluded that although differences in thermoregulatory patterns might have been related to genetic differences in the populations studied, more likely the variation reflected different methodological approaches (Studier and Wilson, 1979). The primary significance of our 1979 study was the demonstration of a transition from a heterothermic pattern of Tb regulation on the day of capture to a homeothermic pattern after three days of captivity. Such a "captivity effect" for A. jamaicensis may help to resolve apparently conflicting thermoregulatory patterns in other mammals, for example, Myotis lucifugus (Stones and Wiebers, 1967; Studier and O'Farrell, 1972) and Peromyscus leucopus (Gaertner et al., 1973; Hill, 1977). Although the captivity effect explains divergent data on thermoregulation, it does not show which data represent the natural thermoregulatory pattern. Studier and O'Farrell (1972) found that the Tb of Myotis lucifugus and M. thysanodes in their natural roosting sites was highly variable and similar to the data for newly caught, laboratory-tested individuals. Apparently, data from bats tested soon after capture better reflect natural thermoregulatory performance, whereas data from captive bats held longer reflect their greatest homeothermic capabilities. We did not attempt to determine which components of the captivity effect are responsible, either singly or in combination, for the changeover in thermoregulatory performance, but several possibilities exist. The increased homeothermic response probably does not result from thermal acclimation. We did not try to hold captive bats at constant T,. Instead, T, for captive bats fluctuated in slightly muted fashion with that of BCI's natural environment (Studier and Wilson, 1970). Furthermore, the captivity effect does not seem to be a general stress response because stress would be greatest during the initial hours of captivity; thus bats would be expected to exhibit the most rigid homeothermy in day zero testing. The captivity effect may result from individual or combined actions of reduced activity while caged or the continual presence of excess food, which the bats can eat at will. In its natural environment, A. jamaicensis is probably a homeothermic endotherm during periods of feeding and flight activity, but loosens Tb control (becoming a nonhomeothermic endotherm) during roosting (nonfeeding and nonflying episodes). The slight reduction in Tb at such times would conserve large amounts of energy (Studier, 1981). Heterothermy in A. jamaicensis conserves 38.7% and 67.4% of the energy that would be required of homeothermic individuals at ambient temperatures of 30°C and 25°C, respectively (Studier and Wilson, 1970). This would amount to a major energy cost reduction during a roosting period. We assume that the slight Tb reduction in A. jamaicensis (Tb of 35.2°C at Ta of 30°C and Tb of 32.5°C at Ta of 25°C) would not reduce responsiveness to environmental stimuli during roosting nor would it preclude initiation of flight. Thermoregulation probably is dependent on nutritional state in captive bats. In the wild, A. jamaicensis feeds on fruit that varies seasonally from scarce to plentiful. Individual bats may undergo a natural period of diel torpor, whereas captive animals with unlimited food may never become torpid as long as their food supply is constant and plentiful. Captive bats routinely weigh more than wild-caught individuals (see Section 3, Reproduction in a Captive Colony). The significant questions regarding possible effect of nutrition on the variability of Tb in A. jamaicensis are: Does the heterothermic pattern of bats tested immediately after capture reflect undernourished individuals? Or, is the homeothermic pattern of bats kept in captivity a response to overnourishment and inactivity? Undernourished or not, poor Tb regulation in newly caught A. jamaicensis is not related to physiological competence, but reflects reduced metabolic heat production and rate of depletion of energy stores. Manakins, which are small frugivorous birds, also exhibit heterothermy on BCI (Bartholomew et al., 1983). McNab (1969) reported a resting metabolic rate of 1.70 cc/g/hr for A. jamaicensis within its thermal neutral zone (TNZ) and a thermal conductance of 0.17 cc/g/hr/°C (below its TNZ). As with other frugivorous bats, this mass-specific standard metabolic rate is slightly higher than would be predicted by Kleiber's (1932) classic relation between metabolism and mass in mammals. The energetic cost of high Tb homeothermy in A. jamaicensis is, therefore, higher than expected for a mammal of its size. McNab (1983) presented extensive arguments concerning energetics, body sizes, and the limits to endothermy that can be expressed by a minimum boundary curve for high Tb homeothermy. This curve estimates the smallest mass at which continuous endothermy can occur for a given metabolic rate. When this minimum boundary curve is drawn by relating the Tb to Ta differential as a function of body mass, data for A. jamaicensis falls almost exactly on the curve (McNab, 1982). A. jamaicensis, therefore, would be predicted to be marginally able to maintain its reported high homeothermic Tb to Ta differential. It is, therefore, not surprising that undernourished (or normally nourished) bats tested immediately after capture maintain a markedly lower Tb and, consequently, lower Tb to Tt NUMBER 51111 differentials. As stated previously, the slight Tb reduction found in A. jamaicensis saves 38.7%-67.4% of the energy required for maintenance of higher Tb. Such energy conservation would seem critically important in a species such as A. jamaicensis that has severely limited reserves of stored energy. BODY FAT.—Total body fat is a direct indication of overall nutritional status in vertebrates. Annual variability in body fat content in numerous Neotropical bats, including A. jamaicensis, has been reported (McNab, 1976; 1982). However, Neotropical frugivores and nectarivores demonstrate less of the seasonal variation and none of the gender-related differences that characterize temperate zone insectivores. All Neotropical bats exhibit low peak fat reserve levels when compared with temperate zone bats (Baker et al., 1968; Ewing et al., 1970; Pagels, 1975; Weber and Findley, 1970). The extremely low fat reserves reported by McNab (1976) for A. jamaicensis emphasize its need to reduce daily energy expenditures. Lack of fat reserves and high intake of dietary carbohydrate suggest that glycogen should be examined as the normal energy reserve of A. jamaicensis. Glycogen levels have not been reported for A. jamaicensis, but comparable data are available from megachiropterans that, although not related to A. jamaicensis, are nutritional equivalents. Daily variation in glycogen and fat levels in liver and flight muscle tissue in Eidolon helvum, a Paleotropical frugivorous bat, have been reported by Okon et al. (1978). Their findings show that liver glycogen levels at sunrise (90.0 mg/g) are extremely high in comparison with levels seen in large domesticated mammals (Watt and Merrill, 1963). Liver glycogen then drops precipitously in E. helvum until sunset when levels (35.0 mg/g) reach a range normal for large mammals. Breast muscle glycogen in this bat remains low (6.0-8.1 mg/g) and constant throughout the roosting period. The extreme drop in liver glycogen suggests that glycogen (as glucose) is the primary energy source of E. helvum throughout its roosting period. Fat concentrations in liver and breast muscle in E. helvum show slight increases at sunset, but the range of all values (5.0-10.3 mg/100 g) is nearly two orders of magnitude less than fat levels in liver and muscle in other mammals, large and small (Kirkham and Allfrey, 1972; Watt and Merrill, 1963). These glycogen levels reflect the high carbohydrate, low lipid composition of the diet of these bats. Van der Westhuyzen's (1978) report on another Paleotropical fruit bat, Rousettus aegyptiacus, provides additional support for the extreme importance of glucose or glycogen and the relative unimportance of fat as an energy source in tropical frugivorous bats. He reported the diurnal cycle of several metabolites including glucose, free fatty acids, lactic acid, and pyruvic acid in captive bats during normal feeding cycles as well as after a "prolonged" fast. The most salient features of this study are the cycles of blood glucose and free fatty acid levels. He found that blood glucose levels follow the expected pattern and fall within normal concentrations for mammals in general. Most of the bats studied died during 31-32 hour fasts; however, bats that survived showed no further change in blood glucose level at 35.3 mg/100 ml. The diurnal pattern of free fatty acid (FFA) plasma levels follows the expected general inverse relation to blood glucose levels. Nighttime FFA's are essentially constant at about 0.5 milliequivalents/liter (mEq/L), which is quite normal for mammals. If food is withheld for three hours after sunset, plasma FFA concentrations rise to 4.0 mEq/L concomitant with the fall in blood glucose level. Although histochemical studies of the muscles of bats (Armstrong et al., 1977; Talesara and Kumar, 1974) demonstrate the relative importance of fats and glucose as energy substrates, muscle enzyme profile studies such as those of Muller and Baldwin (1978) and, especially, those of Yacoe et al. (1982), are particularly germane to the present discussion. Yacoe et al. (1982) determined enzyme activity levels for citrate synthetase, hexokinase, 3-hydroxyacyl-CoA dehydrogenase (HOAD), and phosphorylase in two frugivorous species (one of which was Artibeus lituratus) and eight insectivorous species of bats. The four enzymes measured are indirect indicators (in sequence) of citric acid cycle capacity, blood glucose oxidation capacity, beta-oxidative capacity, and glycogenolytic capacity. As expected, citrate synthetase activity was extremely high, among the highest reported for mammalian skeletal muscle, and there were no interspecific differences. Elevated HOAD activity levels in all species indicated the expected high capacity for fatty acid oxidation. Enzymes participating in glucose storage, mobilization, and cell entry, however, provided the most intriguing picture. Hexokinase activity in the frugivorous species was from two to three times higher than in the insectivorous species. Phosphorylase activity in all species was on the high end of the normal mammalian range, and, although not statistically significant, phosphorylase activity in A. lituratus was higher than in any other species tested. Frugivorous bats, therefore, retain the capability of rapidly metabolizing fats as a fuel source but also have unusually high glycogenolytic ability. The combination of high aerobic and high glycogenolytic activities previously has been thought to be mutually exclusive in mammalian muscle fiber (Burleigh and Schimke, 1969). Such a combination defies easy classification in the slow (I) and fast (IIA) and fast (IIB) categories for mammalian muscle fiber types (Lamb, 1984). Although similar results have been reported for Australian bats (Muller and Baldwin, 1978), extremely high capacities for aerobic glucose oxidation have been reported primarily in flight muscle of insects, which have a normal diet with high glucose density (Beenakkers, 1969; Beenakkers et al., 1975; Heinrich, 1979). Because of its presumed high dietary glucose density and extremely high glucose assimilation efficiency (Morrison, 1980b), A. jamaicensis, along with other frugivores and nectarivores, should have much more glucose available for oxidation than do bats of other dietary preferences. However, there is a significant energy penalty for converting dietary carbohydrate to fat (Martin and Lieb, 1979). 12SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY From the foregoing, it may be presumed that A. jamaicensis produces little fat from glucose and does not take advantage of the reduced weight and high caloric density gained by storing energy as fat. Furthermore, A. jamaicensis must exhibit an extreme facility in the storage, mobilization, and turnover of glycogen. Finally, the overall energy reservoir in A. jamaicensis is severely limited. In their natural environment, these bats maintain no significant positive daily energy balance that would allow for storage of surplus caloric energy. SALIVATION.—Like many frugivores, A. jamaicensis has exceptionally large salivary glands (Phillips et al., 1977). Results of our histological examination of these glands in A. jamaicensis (Studier, Boyd et al., 1983) are in agreement with those of Wimsatt (1956). In spite of the large size of the glands there is no relative increase in number of ducts, nor is there evidence of relative excess production of serous secretion. The functional significance of these structural observations include possible increases in salivary amylase production, increased involvement in regulation of mineral and water balance, and overall increase in production of saliva, all related to the unusually large size of the glands. Salivary components may act chemically to neutralize alkaloids in figs. Mucus may hold together the pellets that are dropped during feeding (Dalquest et al., 1952). We have suggested that the abundant saliva may strongly buffer gastric secretions, thus preventing gastric contents from becoming acidic, while simultaneously coating the gastric epithelium with an extensive buffering barrier (Studier, Boyd et al., 1983). If gastric fluid remains above pH 4, salivary amylase should continue to function. This may be of considerable importance in view of the rapid passage time and consequent brief digestion exhibited by A. jamaicensis (Morrison, 1980b; Studier, Boyd et al., 1983). The gastric glands in the stomach and duodenum of many frugivorous bats contain typical parietal and zymogen cells (Bhide, 1980; Forman, 1972; Rouk and Glass, 1970). Brunner's glands are reduced or absent in frugivorous bats, including species of Artibeus. The function of Brunner's glands among mammals is still debated, but for a long time a connection with protection of the duodenum from damage by highly acidic chyme leaving the stomach has been suspected. If saliva of A. jamaicensis has sufficient buffering capacity to prevent chyme from becoming highly acidic, reduction or absence of Brunner's glands in these frugivores would be compensated. Salivary bicarbonate concentration has been shown to Be directly proportional to the rate of saliva formation (Burgen and Emmelin, 1961). Because the bicarbonate buffer system accounts for the bulk of the buffering power of the saliva (Izutsu, 1981), increased or high relative rates of saliva production may indicate elevated salivary buffering capacity in A. jamaicensis. For further discussion of this possibility and for details of gastric ultrastructure see Phillips et al. (1984). Caloric Balance Total daily energy requirements can be estimated by various methods (Kunz and Nagy, 1988). Maintenance energy (ME) requirements are most simply calculated as a function of body mass (W). For endothermic mammals, which maintain constant high Tb, ME (Kcal/day) = 106 W0-75, where W is in kilograms (National Research Council, 1978). For an A. jamaicensis weighing 45 g, estimated ME equals 10.3 Kcal (43.3 Kj)/day (Table 2-1). Daily energy requirements for an A. jamaicensis weighing 45 g also can be estimated based on time partitioning (Table 2-1). The first estimate is based on Morrison's (1978d) data, in which he partitioned flight time into various activities and flight distances. Metabolic cost of flight was based on the wind tunnel studies of Thomas (1975) on other frugivorous bats where metabolic cost of horizontal flight for a 45 g bat is 0.30 Kj. At an average flight speed of 5 m/sec, the cost of flying 100 m is 0.092 Kj. Appropriate increases in the energy cost of flight for transport of figs to feeding roosts also are included (Morrison, 1978d). These flight costs are added to energy needs for maintaining the basal metabolic rate. Using data from McNab (1969) for wellor overfed A. jamaicensis (1.7 cc/g/h), we estimated a basal energy cost of TABLE 2-1.—Daily energy expenditures for a 45 g Artibeus jamaicensis. All values are Kcal/days; equivalent Kj/day are given in parentheses. See text for further details. Parameters Based on mass* Based on time partitioning Flight costs'" searching commuting between feeding passes feeding passes transmitter e subtotal Basal metabolism (wellor overfed)'1 Total**1 Basal metabolism (normallyor underfed)8 Total1"' Based on time partitioning Flight costsf Basal metabolism (wellor overfed)d Totals Basal metabolism (normallyor underfed)* Totalf4« Kcal/day 10.30 0.20 0.28 0.19 1.04 0.11 1.82 8.77 10.59 3.87 5.69 3.20 8.77 11.97 3.87 7.07 Kj/day (43.30) (0.84) (1.17) (0.79) (4.33) (0.50) (7.63) (36.70) (44.33) (16.20) (23.83) (13.40) (36.70) (50.10) (16.20) (29.60) • National Research Council, 1978. b Morrison, 1978d. c Additional 1% increment as cost of carrying transmitter in flight. d McNab, 1969. e Studier and Wilson, 1979. 'Morrison, 1980b. NUMBER 51113 8.77 Kcal/day. Using metabolic rates of underfed (= normally fed) A. jamaicensis (0.75 cc/g/h; Studier and Wilson, 1979), reduces basal energy cost to 3.87 Kcal/day. In either circumstance, daily energy cost of basal metabolism represents the major fraction (67.9%-82.7%) of total daily energy requirements. An additional time-partitioning energy budget (Morrison, 1980b) is based simply on a minimum estimated flight time of 45 min/day (Table 2-1). Although this second time-partitioning budget markedly increases metabolic energy expenditure for flight (from 1.8 to 3.2 Kcal/day), basal metabolic costs still represent the major fraction (54.8%- 73.1%) of total daily energy budgets. Kunz (1980) proposed a daily energy budget for bats in general in which caloric needs (in Kcal/day) equal 0.92m0-767 (where m is mass in grams). Kunz's (1980) estimate of daily energy needs, however, includes data from lactating and pregnant bats as well as from reproductively inactive individuals. Therefore, he overestimated caloric needs. The estimates for total daily energy costs to a reproductively inactive, endothermic A. jamaicensis, maintaining a high Tb are remarkably consistent with a total range of 10.3-12.0 Kcal/day (43.3-50.1 Kj/day) and probably represent realistic estimates for a bat that remains constantly homeothermic. The minimal daily energy budget of 5.7 Kcal/day (23.8 Kj/day) for heterothermic individuals is surely an underestimate because the bats are not continuously heterothermic. The marked reduction in daily energy demand associated with heterothermy, however, may be invaluable to free-flying individuals that have a marginal energy intake. Fleming (1988) estimated daily energy budgets of 41.9-47.3 Kj/day for Carollia perspicillata, a smaller phyllostomid with a more varied diet and different foraging strategy. The similarity between these figures is striking. A. jamaicensis extracts 55.5 g of juice per 100 g of fresh fruit of F. insipida (Morrison, 1980b). Morrison reported that juice from these ripe figs contained 0.315 Kcal/g; we found 0.415 Kcal/ml (Studier, Boyd et al., 1983). The specific gravity of an artificial fig juice solution (111 mg glucose per ml) is 1.042 g/cc. Most of the energy in fig juice is in dissolved glucose and the energy assimilation efficiency of fig juice by A. jamaicensis is 98.3% (Morrison, 1980b). Using the maximum estimate for a daily energy budget of 12.0 Kcal/day, an A. jamaicensis weighing 45 g would need to assimilate all the energy from 28.9 to 36.5 ml of fig juice. Given the 98.3% assimilation efficiency, this would require 29.4-37.2 ml of ingested fruit pulp juice or 55.1-69.8 g of fresh ripe fruit. At 7 g per fruit, an A. jamaicensis would require 8-10 whole fruits of F. insipida per day to meet its caloric requirement entirely from the ingestion of the fruits of figs. If average-weight ripe fruits weighing 5.6 g (Morrison, 1978a) were ingested, required intake would be 9.8-12.5 figs. These numbers of whole figs correspond nicely to the 7 ± 2 nightly feeding passes observed by Morrison (1978a) when A. jamaicensis was feeding exclusively on fruits off. insipida. In summary, A. jamaicensis has meager, if any, fat reserves; probably exhibits extraordinary glucose assimilation, storage, mobilization, and glycolytic capacities; probably has extreme daily fluctuation in glycogen levels with little reserve capacity; and is marginally able to maintain caloric balance on a normal daily intake of 7 ± 2 Ficus insipida fruits. Ingestion of nine fruits probably would allow A. jamaicensis to maintain a high Tb for a 24-hour period but ingestion of fewer fruits would not allow caloric balance as a high Tb homeotherm. Such bats would conserve energy and thus remain in caloric balance by a drop in regulated Tb, and a corresponding drop in energy needs. Nitrogen Balance Nitrogen excretion is related to metabolic rate. Consequently, nitrogen requirements are appropriately related to metabolic body mass (W°-75) (Brody, 1945; Kleiber, 1975). The daily nitrogen requirement (mg/day) for high Tb homeotherms is 200 W075, where W is in kilograms (National Research Council, 1978). The nitrogen requirement can be converted to a minimum protein requirement by multiplying by 6.25 (Herbst, 1988). Alternatively, dietary protein requirement can be calculated as a function of ingested caloric intake. Minimum protein for maintenance is 10.7 mg protein per Kcal ingested. For a 45 g A. jamaicensis, the daily minimum protein requirement is 122 mg, based on body mass alone. Using maximal estimated daily energy expenditure (12.0 Kcal/day from Table 2-1), calculated daily minimum maintenance protein intake is 128 mg. The protein density of F. insipida fruit juice is 4.7 mg/ml (calculated from Morrison, 1980b). Assuming fig juice to be the only source of dietary nitrogen of A. jamaicensis, constantly homeothermic individuals maintaining high Tb would require 26.0-27.2 ml of fig juice/day (at 100% assimilation) to maintain nitrogen balance. This estimated required intake is less than the calculated daily volume of fig juice needed to maintain caloric balance in high Tb homeothermic individuals (29.4—37.2 ml/day). Digestibility of protein in low-fiber diets in a variety of mammals ranges from 77% to 90% (Maynard and Loosli, 1969). Assuming an assimilation of 85% for A. jamaicensis, daily minimum intake of fig juice would rise to 30.6-32.0 ml/day. This would mean that the daily volume of fig juice necessary to meet protein needs is equal to or less than that needed for caloric balance and that maintenance of nitrogen balance is less of a problem for the bats than maintaining caloric economy. The calculated protein minimum, however, is for the "ideal" protein whose amino acid composition exactly reflects the needs of the subject. Rasweiler (1977) pointed out that animal proteins generally have amino acid compositions that correspond more closely to mammalian requirements and may be more readily digestible than proteins of plant origin, which often are incomplete in terms of essential amino acids. 14SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY The calculated requirement for fig juice volume given above is surely a minimum or low estimate of actual juice needs. As is the case for caloric economy, it seems likely that a high Tb homeotherm such as A. jamaicensis could maintain nitrogen balance only marginally on a diet of F. insipida fruit. Again, the ability of a free-living A. jamaicensis to reduce its regulated Tb acts as a safety valve, not only for caloric balance, but also by reducing dietary nitrogen requirements to a level sufficient to allow nitrogen balance. Although studies of the structure and function of kidneys of Neotropical bats will be discussed in more detail with respect to water and mineral balance, some of the data are pertinent to nitrogen balance. Studier, Wisniewski et al. (1983) examined kidneys in 25 species of Neotropical bats and suggested that renal morphology is primarily a function of dietary protein density. Among the phyllostomids, members of the primarily frugivorous/nectarivorous subfamilies Glossophaginae, Carolliinae, and Stenodermatinae (including A. jamaicensis), most of which have low-protein diets, possess renal medullae that cannot be subdivided readily into inner and outer zones. Members of the phyllostomid subfamilies Phyllostominae and Desmodontinae, as well as all members thus far studied in all other families of New World bats, have renal medullae that are easily subdivided into inner and outer zones. The dietary preferences of this second group of bats are varied, but all species regularly consume some food of animal origin, which is high in protein. Our data on urine composition further support the proposed relation between dietary protein density and renal function in Neotropical bats (Studier and Wilson, 1983). We found urinary ammonia and urea nitrogen levels for A. jamaicensis (495 mg% N), although highly variable, to be markedly lower than levels for the insectivorous species, Myotis nigricans (1887 mg% N). If the minimal nitrogen requirement for a 45 g A. jamaicensis is 19.54 mg/day (= 200 x 0.045075), then total urinary N is 434 mg/kg/day, which approximates urinary nitrogen excretion rates in many mammalian herbivores (Altman and Dittmer, 1961). If the two abnormally high values for urinary ammonia and urea nitrogen we found in A. jamaicensis are disregarded, the recalculated average level is 370 mg% N (n = 17, Studier and Wilson, 1983). Water Balance We now have sufficient data to present a rough water balance account for A. jamaicensis (Table 2-2). Values in Table 2-2 represent estimates for a 45 g, high Tb homeothermic individual. As previously discussed, daily intake of fig juice is 29.4-37.2 ml. Swallowed juice is 90% water (Morrison, 1980b). Daily intake of water in food, therefore, is 27.6-34.9 g. Estimated metabolic water assumes complete aerobic oxidation of all glucose in the consumed juice (111 mg/ml, from Studier, Boyd et al., 1983). The calculated value for metabolic water is probably slightly low because oxidation of the small amounts TABLE 2-2.—Water economy budget for a 45 g, high Tb, homeothermic Artibeus jamaicensis. All values are g/day. EWL • evaporative water loss. See text for further details. Parametersg/day WATER GAINS In food From metabolism Drunk Total gain WATER LOSSES EWL at rest EWL in flight In urine In feces Total loss UNACCOUNTED LOSSES 27.6-34.9 1.9-2.4 0 29.5-37.3 5.2 1.3 9.4-11.8 11.8-14.9 27.7-33.2 1.8-4.1 of dietary fats and proteins will produce a little additional water. There is no indication that A. jamaicensis needs to consume free water when feeding on figs. The studies of Phillips et al. (1984) on gastric ultrastructure in Artibeus also indicated a lack of importance of free water ingestion. Assuming constant temperature and humidity, evaporative water loss (EWL) at rest is calculated as log EWL = log 0.398 + 0.672 log W where EWL is grams of water/animal/day and W is in grams (Studier, 1970). We estimated EWL in flight based on 45 minutes total flying time/day (Morrison, 1980b) and scaling up the EWL (957 mg/h) found by Carpenter (1969) for flying Leptonycteris sanbomi (24.4 g). Estimates of urinary water loss and water lost in fecal matrix are derived from Morrison's (1980b) data showing that urinary water represents 31.8% and fecal water represents 40.1% of swallowed fig juice, respectively. If all ingested nitrogen appeared in the urine, A. jamaicensis would produce urine volumes of 4.0-5.3 ml/day, assuming a minimum daily required nitrogen intake of 19.54 mg/day and urinary ammonia and urea nitrogen levels of 370-495 mg%. These values are considerably lower than estimates based on Morrison's (1980b) data, but still represent remarkably high urine output. Considering the multiple sources and assumptions used to construct the water economy budget for A. jamaicensis (Table 2-2), the values come remarkably close to balancing and are probably accurate. Some useful comparisons can be made with values for other mammals. Water turnover rates for a 45 g A. jamaicensis (655-829 ml/kg/day) are extreme in contrast to values of 40-273 ml/kg/day reported for a wide variety of mammals (Altman and Dittmer, 1961). Similarly, preformed water consumption in A. jamaicensis of 613-775 ml/kg/day is elevated in comparison with values for other mammals (35-211 ml/kg/day; Altman and Dittmer, 1961). These NUMBER 51115 exceptional values relate to the essentially liquid diet of A. jamaicensis, coupled with its poor renal water conservation, and relatively small size. Comparative values for total body turnover and preformed water turnover are 330 ml/kg/day and 150 ml/kg/day, respectively, for an 8.4 g Myotis thysanodes (O'Farrell et al., 1971), and 250 ml/kg/day and 180 ml/kg/day, respectively, for M. lucifugus (O'Farrell et al., 1971). Total body water turnover in Pizonyx (= Myotis) vivesi is estimated at 480 ml/kg/day (Carpenter, 1968). Preformed water turnover for a common vampire, Desmodus rotundus, weighing 34.2 g is 410 ml/kg/day (Wimsatt, 1969), assuming bovine blood to be 78.5% water (McNab, 1973). Preformed water turnover also can be estimated from published data for the nectarivorous bat, Leptonycteris sanborni, at 465-712 ml/kg/day (Howell, 1974), and the nectarivorous bird, Selasphorus flammula, at 692 ml/kg/day (Hainsworth and Wolf, 1972). Urinary output in A. jamaicensis varies from 209 to 262 ml/kg/day in contrast to many other mammals with values that vary from 2.5 to 74 ml/kg/day (Altman and Dittmer, 1961). These rates of urine production in A. jamaicensis (6.5-8.2 microliters/min) are roughly 30 times the maximal rate of urine production (0.23 microlitcrs/min) found by Bassett and Wicbers (1979) for the insectivorous M. lucifugus. Having collected urine samples from many species of bats, we have no difficulty believing that rates of urine production in A. jamaicensis and other frugivorous/nectarivorous bats far exceed urine volumes considered normal in other species. As mentioned earlier, A. jamaicensis and other frugivorous and nectarivorous bats possess kidneys with undivided renal medullae. Such species routinely produce natural urine of low osmotic pressure and low urinary nitrogen levels (Studier and Wilson, 1983) compared with those species with renal medulla divisible into inner and outer zones (Studier, Wisniewski et al., 1983). Mean maximal urine concentration (MMUQ in A. jamaicensis is 972 mOsm/kg (Studier, Boyd et al., 1983). The total medullary (M) thickness to cortical (C) thickness ratio (M/C) for A. jamaicensis (2.4), while typical for frugivores, is markedly lower than M/C for bats of other feeding preferences and does not vary with habitat aridity (Studier, Wisniewski et al., 1983). Geluso (1980) presented a highly predictive equation relating MMUC of insectivorous bats to M/C in which MMUC (mOsm/kg) = 702 + 387 (M/C). Based on this equation, A. jamaicensis would produce MMUC of 1,620 mOsm/kg, a value nearly double the observed MMUC. Similar observations hold true for other frugivorous/nectarivorous species with undivided renal medullae (Carpenter, 1969; Studier and Wilson, 1983). None of these species produce natural urine concentrations approaching those predicted by Geluso's equation, suggesting that his formula does not apply to bats with undivided renal medullae (Studier and Wilson, 1983). We made several attempts to induce MMUC in A. jamaicensis, that included dehydration/starvation, loading with strongly hyperosmotic salt solutions, and feeding dehydrated fruits of figs (Studier, Boyd et al., 1983). Although none of these methods worked well, ingestion of dehydrated figs, the functional equivalent of Geluso's (1975, 1978) "water denied" experiments with insectivorous bats, resulted in the most uniform and highest urine concentrations for this species. It is of particular interest that although dehydration and salt loading caused an expected and predictable increase in osmotic pressure of the blood, these treatments were not associated with the expected increase in osmotic pressure of the urine (Studier, Boyd et al., 1983). Dehydration in Myotis lucifugus, induced by Bassett and Wiebers (1979), however, resulted in expected increases of osmotic pressures in both blood and urine. There seems to be a slow or minimal release of antidiuretic hormone (ADH) or a slow or reduced renal tubular response to ADH in response to rising osmotic pressure in the blood of A. jamaicensis. Urine samples taken from A. jamaicensis at sunset in May were significantly more concentrated and less variable than samples taken in November (Studier, Boyd et al., 1983). The May samples were taken at the end of the dry season and the November sampling occurred toward the end of the wet season (Smythe, 1974). Presumably heat/dehydration stress is greater in the dry season than in the wet. In May, there was a marked rapid decrease in total urine concentration 0.5-1.5 hours after sunset in free-flying A. jamaicensis. This decrease is associated with rapid food passage time in this species, ingestion of adequate hypotonic fluid for rehydration, and rapid assimilation and equilibration with the ingested fig pulp juice. Urine then became progressively more concentrated throughout the remainder of the night. Osmotic pressures of urine in captive, rehydrated (taken two hours after sunset) individuals were identical throughout the night with urine concentrations of free-flying A. jamaicensis (Studier, Boyd et al., 1983). This suggests that rehydration occurs early in the nightly feeding period and is unaffected by subsequent feeding bouts. We know from other studies that A. jamaicensis feeds sporadically throughout the night (Morrison 1978a; 1978b; 1978c). Mineral Balance How herbivorous mammals ingest adequate amounts of dietary sodium (Na+) has generated considerable interest (Blair-West et al., 1968; Cowan and Brink, 1949; Dalke et al., 1965; Herbert and Cowan, 1970; Jordan et al., 1973; Stockstad et al., 1953; Weeks and Kirkpatrick, 1976; 1978). Sodium levels in most plants and plant parts are typically low (Likens and Bormann, 1970; Sauchelli, 1969; Weeks, 1978) and related to soil Na+ levels, which in turn are highly affected by the Na+ levels in rainfall (Blair-West et al., 1968) and the frequency of rainfall. Tropical rain forest may be particularly susceptible to loss of nutrients by leaching due to rapid decomposition of litter and heavy, frequent rains (Jordan and Herrera, 1981). Although some data on characteristics and composition of the soils of BCI are available (Knight, 1975), there is no 16SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY information on soil sodium levels. Fig fruits contain little sodium (Diem, 1962; Heinz International Research Center, 1964; Oates, 1978). Sodium levels in ripe fruits of F. insipida and F. yoponensis have been measured at 0.49 and 0.48 mg/g dry weight, respectively (Nagy and Milton, 1979). We determined sodium levels in figs carried by A. jamaicensis to be similar to those found by Nagy and Milton (1979), but potassium concentrations were lower (Studier, Boyd et al., 1983). Sodium density in dried pulp is about 2.4 times the sodium level of dried seeds and is identical to the sodium level of pulp juices. Thus, the consumption of fig juice rather than the entire fruit greatly increases dietary Na+ density in A. jamaicensis and significantly reduces the weight of figs estimated to be required for maintenance of Na+ balance. On the other hand, the K+ level of fig juice is less than half the concentration of that ion in dried pulp. This implies that K+ is concentrated in specific organelles within the fig pulp and is not likely to be extracted by A. jamaicensis. It is released in the process of homogenization of the dried pulp for laboratory analysis. The lowered dietary K+ level obtained from pulp juices as opposed to whole pulp probably also lowers the Na+ requirements of A. jamaicensis as it does in other mammals (Meyer et al., 1950; Staaland et al., 1980; Weeks and Kirkpatrick, 1978). Minimal sodium requirements have been estimated for few small mammals. Sodium required for growth in laboratory rats and mice is estimated at 10 and 18 mg/animal/day, respectively (National Research Council, 1978). Assuming the average requirement for growth in A jamaicensis is 14 mg/animal/day, bats would require 69 ml (72 gms) of F. insipida fruit juice at 8.8 mEq/L sodium level to maintain sodium balance. Such amounts far exceeds the requirements of fig juice previously calculated for maintenance of caloric and nitrogen balance (29.4-37.2 and 30.6-32.0 ml/day, respectively). Estimated daily Na+ requirements for growth are probably higher than requirements for maintenance. However, it would appear that acquiring adequate sodium for daily requirements from fig juice is more likely to be a nutritional limiting factor than ingestion of sufficient juice to meet caloric and nitrogen needs. Mammals suffering Na+ deficiency characteristically exhibit hypertrophy and hyperplasia of the zona glomerulosa of adrenal glands and greater development of striated and excretory ducts within salivary glands (Blair-West et al., 1968). These features, acting through the renin-angiotensinaldostcrone system, result in renal and salivary Na+ retention. The adrenal zona glomerulosa of A. jamaicensis shows no obvious hypertrophy or hyperplasia when compared with the condition in Neotropical insectivorous bats, although possible subtle differences may yet be detected (Studier, Boyd et al., 1983). We have not found an increase in the number of ducts within salivary glands, another fact that argues against a probable Na+ deficiency in A. jamaicensis. The inordinate size of these glands, however, indicates a proportionate increase in absolute number of ducts per gram of body mass of the bat. Energy and Water Balance during Lactation Because the foregoing discussion has concerned nutritional requirements and balances needed for maintenance, it is useful to estimate nutritional increments needed under stress such as during lactation (see Section 3, Reproduction in a Captive Colony). A 45 g A. jamaicensis should produce milk at a rate of 12.3 gm/day based on Linzel's (1972) measure of milk production as a function of body mass (daily milk production in Kg/day = 0.126075 Kg/Kg). Milk of A. jamaicensis contains 2.3 Kcal/g (Jenness and Studier, 1976). This energy level is comparable to milk energy content for other bats, but is higher than milk energy levels for many large mammals. The high energy content of A. jamaicensis milk is in accord with Ben Shaul's (1962) suggestion that mammals that nurse their young on a scheduled basis produce milk of higher energy content than mammals that nurse continuously or on demand. If one assumes that food energy can be converted to milk energy with no cost, the production of 12.3 g of milk per day at 2.3 Kcal/g imposes a minimal additional caloric requirement for A. jamaicensis of 28.3 Kcal/day. This would raise the total daily caloric requirement to about 40 Kcal/day for a high Tb, homeothermic, lactating female compared to 12.0 Kcal/day for a nonlactating individual. The total milk protein level for A. jamaicensis is 4.7 g% (1.1 g% casein and 3.6 g% whey protein) (Jenness and Studier, 1976). Again assuming no energy costs for milk protein synthesis, a high Tb, homeothermic, lactating female would require an additional incremental daily protein intake of 578 mg for a total daily protein requirement of 700-706 mg compared to 122-128 mg/day for a nonlactating individual. Milk of A. jamaicensis is about 70% water; therefore, production of 12.3 g of milk would require an additional intake of 8.6 ml of water for a total daily water need of 38.1-45.9 ml in lactating females compared to the 29.5-37.3 ml water requirements for nonlactating bats. During lactation, therefore, the additional incremental needs for extra caloric and nitrogen intake are massive in comparison to additional water needs. Whereas estimated milk production may seem somewhat high, energy efficiency for milk production is certainly not 100%. Brody (1945) calculated gross energetic efficiency of milk production to be 28%-34% for humans and 44%-48% for rats. We suggest that during periods of high nutritional demands, such as growth and especially pregnancy and lactation, A. jamaicensis would be expected to lower their regulated Tb to reduce caloric and nitrogen needs and that during such times, these bats may supplement their staple fig diet with food items of higher caloric and protein density. Studies directed at shifts in food habits during pregnancy and lactation could test this hypothesis. Summary Studies on energy balance in A. jamaicensis were focused on this bat's use of the fruits of Ficus insipida, the food that it used NUMBER 51117 most on BCI. Probably these bats always are in slight food stress in the wild and are more likely to behave as facultative heterotherms. Poor body temperature regulation in newly captured bats is not related to physiological competence, but reflects reduced metabolic heat production and rate of depletion of energy stores. A. jamaicensis falls very near the minimum boundary curve for high body temperature homeothermy, suggesting that energy savings from heterothermy may be critical to this bat in the wild. Artibeus jamaicensis has extremely low fat reserves and a high intake of dietary carbohydrate, suggesting that glycogen may be its normal energy reserve. This bat, in its natural environment, maintains no significant positive daily energy balance that would allow for storage of surplus caloric energy. The energy cost of basal metabolism represents the major fraction (55%-83%) of total daily energy requirements. Daily energy budgets of 10-12 Kcal/day combined with the caloric content of daily food intake requires an intake of 9.8-12.5 figs per day, a figure not inconsistent with the 7 ± 2 nightly feeding passes observed in radiotelemetry studies. Nitrogen and protein requirements can be met by the amount of daily food intake necessary to maintain caloric balance, although questions remain about the digestibility of plant proteins. Urinary ammonia and urea nitrogen levels are markedly lower in A. jamaicensis than in bats that use animal protein. Based on our studies of caloric, nitrogen, and water balances, there is no indication that A. jamaicensis needs to consume free water when feeding on figs. Water turnover rates for A. jamaicensis (655-829 ml/kg/day) are extreme in contrast to values reported for a wide variety of other mammals. Mean maximal urine concentration in A. jamaicensis is 972 mOsm/ kg, a strikingly low value indicating poor urine concentrating ability. Published predictive equations for maximal urine concentration do not apply to frugivorous bats. Dehydration in A. jamaicensis results in increased blood osmotic pressure, but without a concomitant increase in urine osmotic pressure. Sodium is potentially limiting in tropical animals restricted to a frugivorous diet, but structural features of the kidneys and salivary glands of A. jamaicensis argue against any probable chronic sodium deficiency in these animals. During periods of high nutritional demands such as growth, pregnancy, and lactation, A. jamaicensis probably lowers its regulated body temperature to reduce caloric and nitrogen needs and may use other, more energyand protein-rich food resources. 3. Reproduction in a Captive Colony Lucinda Keast Taft and Charles O. Handley, Jr. We report on the reproductive biology of Artibeus jamaicensis based upon studies of a colony held captive at the National Zoological Park (NZP). Although information is available on many aspects of the reproduction of bats, most is based on temperate zone Vespertilionidae. The most recent reviews of preand postnatal development among the vespertilionids are those of Orr (1970) and Tuttle and Stevenson (1982). Racey (1988) summarized methodology for reproductive assessment and Wilson (1988) provided information on maintaining bats for captive studies. There is little information, however, on reproduction and ontogeny of other bats. Kleiman and Davis (1979) reviewed the available literature on development and maternal care in phyllostomids. The most detailed information on ontogeny is for Carollia perspidllata (Kleiman and Davis, 1979) and Desmodus rotundus (Schmidt and Manske, 1973). Phyllostomid bats show far greater diversity of social systems, dietary habits, reproductive strategies, and selection of roost sites than do the vespertilionids (Baker et al., 1976,1977, 1979). The selective pressures with which phyllostomids must cope are quite different from those encountered by the better-known vespertilionids of the temperate zone. Comparisons between the two families should be productive. Each year most female A. jamaicensis give birth to a single young during two reproductive episodes. This pattern of bimodal polyestry is common to many Neotropical frugivorous and nectarivorous bats (Wilson, 1979). A. jamaicensis is unusual, however, because development of the implanted blastocyst is delayed for 2.5-3 months during one of the birth episodes (Fleming, 1971; Fleming et al., 1972). Delayed development has been documented only in Panamanian populations of A. jamaicensis, but it is likely to prove widespread. Delayed development also occurs in a phylogenetically and ecologically diverse array of bats, Macrotus (Bradshaw, 1962), Miniopterus (Medway, 1971), Hipposideros (Bernard and Meester, 1982), and Haplonycteris (Heideman, 1988). Lucinda Keast Taft, National Zoological Park, Smithsonian Institution, Washington, D.C. 20560. Charles O. Handley, Jr., National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560. Parturition and neonatal appearance have been described (Bhatnagar, 1978; Jones, 1945,1946), but little is known about early development of the young. In Panama^ where the young of A. jamaicensis usually are bom in tree holes (Morrison, 1975), the study of early development is difficult because many tree hole roosts are inaccessible and the bats show a strong tendency to desert a roost if disturbed. Therefore, observations on captives is still the best means of gaining information on reproduction and development. Novick (1960), who was successful in breeding and long-term maintenance of A. jamaicensis in captivity, believed that freedom from handling was necessary for successful reproduction. Improved artificial diets (Rasweiler and de Bonilla, 1972) have enhanced the successful maintenance of a number of Neotropical frugivorous and nectarivorous phyllostomids (Greenhall, 1976; Rasweiler, 1975,1977; Rasweiler and de Bonilla, 1972; Rasweiler and Ishiyama, 1973). Kleiman and Davis (1979) found that captive Carollia perspidllata successfully bred in spite of the handling necessary for weekly examinations of the adults and young. History of the Research Colony at NZP We established a research colony of A. jamaicensis at the NZP in Washington, D.C, with 24 bats captured in June 1978 near Corozal, Panama^ 35 km SE of Barro Colorado Island (BCI). We studied reproduction and development in this colony from July 1978 through August 1981 and, consequently, we can describe several aspects of the biology of A. jamaicensis in considerable detail. Among the original 24 bats, three males and 12 females were subadults, born early in 1978. Four males and five females were adults. The females were neither lactating nor obviously pregnant at capture. Prior to the first births in captivity (January 1979), two adult males and four subadult females died. One of these was a female that apparently did not adjust to the captive feeding regime. On the 47th day of captivity all bats were wing-banded with plastic bands and three days later another female died, possibly of complications from bandingassociated trauma. The other four bats died because of 19 26SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY TABLE 3-7.—Length of interbirth intervals in the NZP colony of Artibeus jamaicensis. Interval between birth groups i-n Il-IIIt III-IV iv-vt V-VI Number of individual females 7 6 14 13 13 Mean interval in days (range) 134.6(120-153) 162.2(132-184) 131.2(113-158) 190.9 (159-220) 120.2(112-137) 198.29 443.37 372.34 377.74 50.74 Comparison* X Yt X zt X Source Between Within Total SS 40561.185 13388.740 53949.925 ANOVA (one-way) df MS 4 10140.296 48 52 F 36.354 P < 0.001 * Student-Newman-Keuls multiple comparison test Identical letters denote that differences were not significant at the 5% level. t Intervals where delayed embryonic development was presumed to have occurred. tP<0.01. development occurred in the captive females at approximately the same time it occurred in free-living females, then the interbirth intervals between BG-II and BG-III and between BG-IV and BG-V would be expected to be longer than interbirth intervals where development proceeded normally. These data and interbirth intervals for individual females between consecutive birth groups (Table 3-7) suggest that delayed embryonic development did occur in the captives but was of diminished duration. The average interbirth interval for normal prenatal development would be about 122 days and the interbirth interval with delayed embryonic development would be about 213 days according to Fleming's (1971) data for free-living Panamanian A. jamaicensis. Differences between intervals presumed to represent normal gestations were not significantly different from one another in the NZP bats as revealed by a posteriori testing (Table 3-7). The presumed delayed intervals, however, were significantly different from the normal intervals as well as from each other. The overall trend in the colony's reproductive output was toward a shortening of the interbirth intervals and earlier onsets of successive birth groups than expected if the captives had remained synchronized with the wild population. Of the two intervals where delayed development was expected, mean length of the second interval (BG-IV to BG-V) was nearly 30 days longer than that of the first, although BG-V still occurred approximately three months ahead of the Panamanian cycle. Loss of some environmental clue, such as day length, may have impacted the cycle of the captive bats. Juvenile Mortality Forty-three babies died before weaning (Table 3-4). Factors contributing to deaths were as follows (by age groups): NEONATAL PERIOD (1st day; 14 deaths).—Mortality was the result of premature birth, still birth, congenital deformation, and maternal neglect. PREVOLANT PERIOD (2-30 days of age; 17 deaths).—This period was characterized by mothers frequently leaving their infants unattended, or in the company of other young and/or adults. Mothers were in estrus during the early part of this period and were frequently pursued by males. A higher than normal density of males (never less than five adult males) may have contributed to increased levels of stress in the colony. The greatest percentage of infant mortality occurred at this time, possibly due to this stress. NEWLY VOLANT PERIOD (31-60 days of age; 5 deaths).— Young were newly volant, but not yet weaned. Initially, they remained in the roost, exercised their wings and pectoral muscles, and accompanied their mothers on short flights. If they were not strong enough for a return flight to a roost area and could not find a substrate up which they could crawl to a roost, or a place high enough to initiate another flight, they would weaken and die unless they were retrieved by their mothers or discovered by the keepers. WEANING PERIOD (61-110 days; 7 deaths).—Juveniles were becoming more independent, weaning was nearing completion, and the mothers were starting to reject their young by chasing them from the roost. Transition to subadulthood was beginning. Preweaning mortality among infants of wild-caught females was 35.4% (Table 3-8). Much of this mortality was attributable to two females (F4 and F12) who consistently failed to rear offspring (seven babies between them for BG-I to BG-VI) because of congenital defects (e.g., skeletal deformities) or intestinal ncmatode infestations in the young. These females were able to rear young successfully beyond weaning age when they were removed from the colony and placed in a larger flight NUMBER 51127 TABLE 3-8.—Correlation of infant and juvenile mortality and level of maternal experience. Numbers of infant deaths in various age groups are tabulated in an r-by-c contingency table with expected values in parentheses (Choi, 1978). X1 = 20.697, tf= 4, P < 0.001. Origin and reproductive experience of females Wild-caught Captive-born multiparous primiparous Total for colony Number of births 65 10 14 89 Number of deaths during preweaning <, 30 days 15(21.91) 3 (3.37) 12 (4.72) 30 31-110 days 8 (7.30) 1 (1.12) 1 (1.57) 10 Number surviving beyond weaning 110+days 42 (35.79) 6 (5.51) 1 (7.71) 49 Percent preweaning mortality 35.4 40.0 92.9 44.9 Percent postweaning survival 64.6 60.0 7.1 55.1 TABLE 3-9.—Analysis of mortality among captive-bom bats by sex. Numbers of infant deaths in various age groups are tabulated in an r-by-c contingency table with expected values shown in parentheses (Choi, 1978). X1 = 5.860, df= 3, 0.25 > P > 0.10. Sex Males Females Total Total births 45 38 83 Total deaths 23 11 34 <.\ 7 (4.34) 1 (3.66) 8 Age in days 2-30 10 (8.67) 6 (7.33) 16 31-110 6 (5.42) 4 (4.58) 10 Survivors 110+ 22 (26.57) 27 (22.43) 49 cage with fewer bats. Thus, they may have been in poor nutritional condition from excess stress or social exclusion from food, factors that could have affected their offspring's growth in utero and postpartum development Excluding the young born to females F4 and F12 would reduce mortality of juveniles of wild-caught mothers to about 27%. Preweaning mortality among offspring of multiparous (experienced) captive-born mothers was 40%, attributable entirely to the deaths of all young born during BG-VI. We have no explanation for this catastrophe. In contrast, approximately 64% of the young born to primiparous (inexperienced) females did not live more than a day, and nearly 86% did not survive a month. Sixty percent or more of the young of experienced mothers (wild-caught females and multiparous captive-born females) survived beyond weaning, but few (7%) of the young of primiparous females survived that long. Apparently, preweaning mortality and postweaning survivorship are dependent on the mother's level of reproductive experience (Table 3-8). Deaths of young were not evenly distributed between the sexes (Table 3-9). In our colony, males died at a rate greater than expected and females died at a rate lower than expected, but the differences were not significant. Lower conception rates and increased infant mortality among primiparous captive-born females may be attributable to a variety of causes. In natural populations, young A. jamaicensis probably are forced to disperse from their natal roost before birth of the next young, and thus parenting behavior would not be acquired through learning. Despite proximity to experienced females at the NZP, inexperience may have operated to increase mortality of infants born to our captive, primiparous females. Increased mortality due to inbreeding could also have been involved, because the inexperienced females could have mated with their male parent, male siblings, or other related males. However, the subsequent successful reproduction in these females suggests that this was not an important factor. Kleiman (1980) discussed several physiological and behavioral means whereby reproduction may be suppressed in socially subordinate female mammals in captive situations. Abortions, stillbirths, inadequate maternal care, and one apparent case of depressed lactation were the causes of death for several of the first young born to our captive-reared females. In Panama and Mexico, groups of 4-11 adult female A. jamaicensis accompanied by their juvenile offspring and a single adult male, occupied tree hole roosts, whereas foliage roosts (palm fronds and subcanopy trees) were occupied by solitary males or small groups of subadults (Morrison, 1979). Tree hole roosts probably are preferred over foliage roosts by reproducing females, as they provide greater protection from rain, predators, and fluctuations in temperature (see Section 8, Roosting Behavior). While foraging, female A. jamaicensis usually leave their young in the roosts (Fenton, 1969). Trune and Slobodchikoff (1976) found that clustered Antrozous pallidus were less agitated and had less weight loss, lower oxygen consumption, smaller body temperature-ambient 28SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY temperature differentials, lost less heat to the environment, and conserved more metabolic energy while inactive, than bats roosting individually. Any or all of these factors could be important to prenatal and postnatal development. Our captive-born female Artibeus jamaicensis first produced young while they were still in nonclustering "subgroups." The implication is that these females were not in the appropriate social environment (harem) necessary for infant development. For some free-living populations, tree hole roosts may be critical resources for reproduction and the cluster of bats within the roosts may be vital as a source of "helpers" to maintain the appropriate thermal conditions for the young bats. These factors might explain the failure of our primiparous captive females to successfully rear offspring. Neonatal Physiognomy Between 29 January 1979 and 30 April 1981, at least 78 full-term bats were born at NZP. Unless otherwise noted in the descriptions that follow, observations on neonates refer to 37 of these infants on their first day of life. Neonates were first examined at the beginning of the dark period and were probably from 1 to 14 hours old when first seen. All neonates retained a segment of dried umbilical cord. Frequently, bits of dried amniotic membrane were present on their bodies as well. By the second day, the cord and all dried membranes were gone, except for two infants that still retained umbilical cords. All neonates had their eyes open, and most had ear pinnae and noseleaf erect when they were first examined. A few had the pinnae and the noseleaf flattened against the head, but they became erect within six hours. Ear openings usually were apparent and the neonates twitched their pinnae or bodies in response to sounds. Six babies did not show ear openings or respond to sounds until as late as four days following birth. However, all infants examined had the ability to emit ultrasonic sounds in the 80 kHz range on their first day. A. jamaicensis emit their ultrasonic orientation calls through the nose (Griffin, 1958), and, accordingly, the mouths of infants were closed when they produced ultrasonic emissions. Their nostrils and the grooves at the base of the noseleaf on the outer edges of the nostrils quivered visibly during such vocalizations. The neonates usually were covered sparsely with dark gray fur on the dorsum and top of the head. The ears, muzzle, wings, and venter were hairless. Nonfurred skin usually was pink. However, a few neonates had dark gray skin. The feet of neonates were well developed and disproportionately large. The wings appeared short, narrow, and relatively less developed compared to the otherwise precocial physical condition of the neonates. Bones of fingers and joints were soft and flexible. The bones and skin of the wing tips were unpigmented and translucent, and the tip of digit III often curved in toward the body more than 180 degrees. Average physical measurements of neonates were taken on 22 individuals, each of which survived at least three weeks. Males and females did not differ significantly in mass, forearm length, wing span length, or wing area, although average size of female neonates exceeded that of males (Table 3-10). Phyllostomid neonates are comparatively large at birth, with neonate-to-mother mass ratios usually exceeding 0.25 and forearm ratios exceeding 0.41 (Kleiman and Davis, 1979; Figures 3-2 and 3-3; Table 3-11). A. jamaicensis compares well with these observations, with a neonate-to-mother mass ratio of 0.26 and a forearm ratio of 0.54 (Table 3-10). Neonatal Behavior Newborn young in the zoo colony typically were found hanging beside their mothers with mouth grasping one of the mother's nipples. Young thus attached kept their wings folded and did not appear to use their thumbs for clinging. Often, a mother shielded her infant, at least partially, with a wing. It was rare to find a neonate with its feet also attached to its mother while resting quietly in the roost. On the other hand, it was not unusual to find newborn infants hanging alone in the roost. They hung quietly, with both feet attached to the ceiling, their wings folded, and often had their eyes closed. They appeared to be sleeping. Neonates were able to crawl about in a slow and wobbly TABLE 3-10.—Body mass and measurements of wings of neonate Artibeus jamaicensis in the NZP colony. Variable Mass (g) Forearm length (mm) Wing span (mm) Wing area (cm2) Wing loading (g/cm2) Aspect ratio (wing span2/wing area) Female, N = 12 mean 14.1 33.9 235.8 77.2 0.187 7.41 SD 2.100 2.749 18.195 18.865 0.025 0.869 Male.W mean 13.7 33.0 227.5 71.6 0.198 7.49 = 10 SD 1.655 2.088 15.501 17.666 0.034 1.252 Combined, mean 13.9 33.5 232.0 74.7 0.192 7.45 N = 22 SD 1.878 2.453 17.159 18.120 0.029 1.034 t* 0.513 0.811 1.144 0.715 0.825 0.168 P ns ns ns ns ns ns Sex ratio (F/M) 1.031 1.026 1.036 1.078 0.944 0.990 Percent adult size 0.264 0.542 0.504 0.276 1.118 0.942 * Student's /-test for significant differences between the means of males and females (one-tailed). NUMBER 51129 1.2-1 HI i- < z o w z o o 0.40.4D 0.6 1.2 1.8 LOG ADULT MASS (g) FIGURE 3-2.—The relationship between neonate mass and adult mass for 18 species of bats. Log neonate mass (g) is plotted against log adult mass (g). The regression line for vespertilionids (solid circles) is y = -0.60 + 0.93 x, (r = 0.942, P < 0.001). Phyllostomid (open triangles) and molossid (open squares) data points are shown for comparison. Data are from Table 3-11. fashion and were active in the reattachment process upon reunion with their mothers. They showed a strong tendency to keep their heads downward. When a mother was held in a head-up position, the attached neonate readjusted its position with its feet until it grasped its mother's head. Turning the mother head-down caused the baby to resume the normal head-down posture. Neonates placed on a vertical wire mesh surface with their heads up immediately started to invert. They also tended to crawl upward (achieved by "walking" backward with the feet). Older babies tended to crawl upward until stopped by an obstacle or lack of a foothold, but neonates usually crawled upward only a few centimeters and then stopped. Autogrooming in A. jamaicensis consists of licking the body surfaces and raking or scratching body surfaces with one foot while hanging by the other. Neonates were seen licking themselves, especially their wing surfaces, but they never hung by one foot in order to use the other foot to groom. Allogrooming is known in this species to include the grooming of infants by their mothers. Occasionally an infant extended a wing to its mother as if to solicit grooming, and the mother then licked the baby's wing. Mothers frequently sniffed young other E E HI CC o U_ LJ < Z o LU o o 1.60n 1.25Q901.41.61.8 LOG ADULT FOREARM (mm) FIGURE 3-3.—The relationship between neonate and adult forearm lengths for 19 species of bats. Log neonate forearm (mm) is plotted against log adult forearm (mm). The regression line for vespertilionids (solid circles) is y = -0.24 + 0.88 x, (r = 0.879, P < 0.001). Phyllostomid (open triangles) and molossid (open squares) data points are shown for comparison. Data are from Table 3-11. than their own and often huddled with them, but they licked only their own young. We never observed allogrooming between adults at the NZP, but it must occur occasionally in wild adult A. jamaicensis. That would best explain the chewed necklace bands we noted in some individuals that were captured on BCI. When we separated a neonate from its mother's nipple, it appeared anxious and was quick to restore a holdfast with its mouth, but it often was not discriminatory about the substrate it happened to grasp. As they were being examined the newborns frequently succeeded in biting our fingers, and on two occasions babies remained attached and began to suck vigorously. By the second day postpartum, the babies were more discriminating about what they would grasp with their mouths. In order to avoid falling, they would bite our Fingers when prevented from clasping our hands with their thumbs and hind feet, but none tried to suck the skin. Infants were never found attached to females other than their own mothers. Disturbed mothers carried their babies with them in flight. 30 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY £2 Si I1* III PJ "1 „ £ ^ „ o o •» grs ^l-aS^S^^"0 « - 2. "51~. ^ ? 1 - s "- - s" I g f f Z " rf Q*o^oou:ao^^SJ2^a:&.SSQaaa 1 UUU oo — O O do — d en oo d d vn vn N w a « — oi <j a w vi <s — •» d o o N N vi CK c^ •v * c~; p — — q — IN is >n t^ do o d dddoddd ci ci o d d do ooo >no vn — vi vn vi o O vn oor~ •*Q5 vifS v N » n <f o N N ^vn^ 1 1 I 5 a qvno; qoo - - jph <* o •+ <=> P^C^^S1 £? <n ^ *^) c*> O >O c*^ c*> *o f^» ^ ^ O ^ ^ ^^ O O O M O O oo f^CJj^ £^ ** t- ^ o^ -^ oe Q oo in vi ^•o6*rirswS »n —* •a NUMBER 51131 Apparently the babies instantaneously detached their feet from the ceiling as the mother spread her wings to take off. When a mother with a baby attached to one of her nipples was netted in flight, the infant usually was found in a position parallel to its mothers's body, its feet gripping her femur or inguinal region. Rarely, an infant was found in a crosswise posture such as that noted by Kleiman and Davis (1979) for Carollia perspicillata. In this case, the infant's feet were attached to the opposite nipple region and the baby was carried across the mother's chest just behind her throat. Growth and Development of Young GROWTH OF HAIR.—Follicular activity preceding growth of fur caused the color of the skin to change from the neonatal pink to gray. Skin of the underparts darkened around seven days of age, and hair growth began at 12 days. At this time, young also began to groom themselves with their feet. The sparse, appressed fur on the head and dorsum was erect and starting to thicken on day 15. The muzzle area darkened around day 17, and facial fur growth started about day 20. By day 22 65-i 60- ^ 55E E o z UJ 50H 45UJ § 40u. 3530J 0 25 50 AGE (days) 75 100 FIGURE 3-4.—Growth of forearm in the NZP colony of Artibeus jamaicensis. Daily means, standard errors, and ranges are shown. 32 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY the ventrum was covered with sparse fur. The general and superciliary vibrissae were only red spots about the corners of the mouth and eyes until day 25, when they began to protrude. Thereafter, short, soft vibrissae were also evident. Young had full pelage by 30 days of age and whitish facial stripes were apparent on some individuals. BODY SIZE.—Growth curves for the four measured characters (forearm length, wing span, wing area, and mass) are presented in Figures 3-4 to 3-7. Daily means, standard errors, and ranges are shown for 22 young from BG-I and II. Data derived from deformed or markedly underdeveloped young that died within a few days of birth were not included in the computations. Forearm length showed the greatest development at birth relative to other measured characters and the fastest growth rate (X = 0.9 mm/day during the maximum linear growth phase; Figure 3-4). Forearm growth stabilized around 50 days of age at X = 61 mm, 1 mm less than the average forearm length of bats of the original adult colony (Tables 3-1 and 3-12). Wing span increased by X = 5.8 mm per day and wing area by X = 3.8 cm2 per day during the maximum growth phase, and both reached adult proportions in X = 70 days (Figures 3-5 and 3-6). Body mass showed the slowest rate of increase (-0.5 g per day; Figure 3-7) and took nearly 80 days to stabilize at X = 48 g, which was 5 g less than the average adult body mass of 500E 400E z Q. o z r 3002OOJ o 25 I 50 AGE (days) —r75 100 FIGURE 3-5.—Growth of wing span in the NZP colony of Artibeus jamaicensis. Means, standard errors, and ranges are shown. NUMBER 51133 TABLE 3-12.—Average asymptotic measurements by sex of Artibeus jamaicensis in the NZP colony. Variable Mass (g) Forearm (mm) Wing span (mm) Wing area (cm2) mean 48.4 61.7 466.7 264.73 Females SD 3.462 1.377 10.630 18.155 n 20 19 15 7 mean 47.6 60.7 457.1 257.77 Males SD 2.923 1.635 91.400 8.789 n 18 18 14 6 t* 0.077 2.058 2.600 0.853* Pi ns < 0.025 <0.01 ns Sex ratio (F/M) 1.017 1.016 1.021 1.027 * Student's /-test for significant differences between the means. t One-tailed. $ The variances for measurements of male and female wing area were not homoscedastic, so a Wilcoxon two-sample test was also performed (Ut = 25, ns). 300-1 250o AREA O z 200 150 10050 J r 0[ 25l 50 AGE (days) 75100 FIGURE 3-6.—Growth of wing area in the NZP colony of Artibeus jamaicensis. Means, standard errors, and ranges are shown. 34SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY the colony's original members (Tables 3-1 and 3-11). Sexual dimorphism was not as apparent in the asymptotic size of the captive-reared young as it was in the wild-caught bats. Again, females exceeded males in all measured variables, but the differences were significant only for forearm length and wing span. The magnitude of the differences between the sexes was also reduced, and females exceeded males in size by only 2%-3% (Table 3-12). The variable best correlated with age was mass (r = 0.9387, P < 0.0001; where r is the correlation coefficient and P is probability). All size variables were highly correlated with one another. The average asymptotic masses of 19 captive-born young were compared with their neonatal masses, and a significant relationship was found (r = 0.7097, t, = 4.1459, P < 0.001; Figure 3-8). Large infants tended to become large adults. INFANT-MOTHER SIZE RATIO.—There was a tendency for larger females to produce large infants (Figure 3-9), but the relationship was not significant (n = 12; r = 0.552; 0.10 > P > O) 50454035302520151OJ or 2550 AGE (days) 751OO FIGURE 3-7.—Growth of mass in the NZP colony of Artibeus jamaicensis. Means, standard errors, and ranges are shown. NUMBER 511 35 at CO 5 50H o o 46a. 2 >- < 4210 15 NEONATAL MASS (g) 20 FIGURE 3-8.—The relationship between the asymptotic mass of captive-reared young and their neonatal mass. The equation for the regression line is y = 32.94 + 1.05 x, (r = 0.710, P < 0.001). 1.25-1 v> < 5 UJ z o UJ o o 1.151.051.68 1.70 1.72 1.74 1.76 1.78 LOG FEMALE NON-GRAVID MASS (g) FIGURE 3-9.—The relationship between neonate and mother mass in the NZP colony of Artibeus jamaicensis. A mean nongravid mass was determined for each adult female. Means (solid circles) and ranges (bars) of infant mass are shown. The equation for the regression line is y = -0.14 + 0.74 x, (r = 0.552, 0.10<P<0.05). 0.05). However, relative infant size (as a percentage of the mother's nonpregnant mass) was correlated with the mother's potential weight-bearing capacity (wing area/nongravid mass). A positive relationship between relative neonatal mass and the mother's wing weight-bearing capacity has not been recorded for any other species of bat. However, Kunz (1974) noted that female wing size was larger than male wing size in Eptesicus fuscus in Kansas where the normal litter is one. He speculated that if there is any selective advantage gained by large wing size, the degree of sexual dimorphism in E. fuscus should be even greater in the more eastern states where two is the usual litter size. More attention should be paid to species' wing-loading values and other variables of wing shape and aerodynamic ability and their relationships to a variety of developmental, behavioral, physiological, and ecological factors. Wings are essential for the location, pursuit, and transport of food. Other potential functions of wings that also could be wing-size dependent are heat dissipation, evaporative water loss, transport or shielding of young, and behavioral displays. The mass of a developing fetus must be of particular significance to those bats that forage while flying. However, beyond a certain point in relative size, the larger wings themselves would represent a significant weight load with much increased drag and would probably be quite difficult to 4. Reproduction on Barro Colorado Island Don E. Wilson, Charles O. Handley, Jr., and Alfred L. Gardner Data on reproduction in bats have accumulated at an accelerating rate during the past 25 years (Racey, 1982; Wilson, 1979). Because Artibeus jamaicensis is common and widespread its reproductive cycle is better known than those of most other Neotropical bats. However, in-depth studies at a single locality over long periods, which are essential for interpreting isolated pieces of information and for elucidating details of a reproductive cycle, have been lacking. We augmented our understanding of reproduction in A. jamaicensis, gained from our colony at the National Zoological Park (NZP), by gathering information from populations on Barro Colorado Island (BCI) and vicinity. Our data on reproduction, accumulated over several annual cycles from wild-caught bats, demonstrate a high degree of reproductive synchrony in this species. In 1971 Fleming presented evidence that embryos of A. jamaicensis undergo retarded development during part of the year, and Fleming et al. (1972) demonstrated bimodal polyestry to be the basic reproductive pattern in this species. Our large data set from BCI and the adjacent mainland, together with the results of the study of our NZP colony (see Section 3, Reproduction in a Captive Colony) provide us the opportunity to examine this cycle in some detail. We base our outline of the reproductive cycle of A. jamaicensis on 4447 individual females captured between 1972 and 1980 on BCI and vicinity. The Basic Pattern We confirm the previously postulated pattern of bimodal polyestry (Fleming et al., 1972; Wilson, 1979), and Fleming's (1971) statements on delayed embryonic development (Figure 4-1, Table 4-1). In general, females are palpably pregnant in January and begin to give birth by late February or early March. They are so synchronized that normally more than half of the Don E. Wibon, and Charles O. Handley, Jr., National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560. Alfred L. Gardner, NERC, US. Fish and Wildlife Service, National Museum of Natural History, Washington, D.C. 20560. population gives birth in those two months. During a postpartum estrus most of them become pregnant again. This results in a brief period when individuals are both pregnant and lactating. Palpably pregnant, but still lactating individuals are most likely in April, May, or June, between the FebruaryMarch and July-August birth peaks. Young born during the initial birth peak (February and March) are weaned in April and May. The second cycle of births commences in July and August with the same pattern of postpartum estrus and lactation. The young from the second birth peak are then weaned in September and October. The difference between the two periods of reproductive activity occurs when the blastocyst from the second postpartum estrus implants, but does not develop at the same rate as the preceding one. The period of delayed embryonic growth covers approximately three months, and is followed by a period of essentially normal development during the remaining four. We were able to recognize individuals that were pregnant (terminal six weeks), lactating, or postlactating, and as these events tracked one another over the years of our study, we recorded peak times for each event. Some variation from year to year is obvious (Figure 4-1), but the amount of congruence from one year to the next is striking. Another way we summarized the data was to combine the information from all years into a single "average" year (Figure 4-2). Thus summarized and divided into 3-week intervals, these averages show the bimodal cycle and emphasize the 12-week diapause between July and November. Individual Variation In spite of the great number of marked bats in our sample, and even though we sampled almost continuously for three years, we were unable to accumulate a complete record of an individual female's reproductive history throughout her lifetime. To do so, we would have to capture the same female at least twice per year during the birth periods. We did catch some individuals two or more times each year, but not always at appropriate stages of reproduction. Although our data are not based on complete individual reproductive histories, it seems 43 44SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY TABLE 4-1.—Reproductive condition (by week) of adult female Artibeus jamaicensis on BCI, 1975 through 1980. Week 1975 Total 1976 Total 1977 Total 1978 9 10 11 12 10 11 12 13 41 42 43 44 45 46 47 48 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 1 2 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 24 26 27 No reproduction N 12 14 8 1 35 27 3 1 2 16 38 51 9 27 35 28 37 274 0 3 3 2 0 0 1 3 10 5 15 5 1 1 8 33 13 9 7 16 2 1 10 53 21 69 291 101 24 2 4 16 6 4 1 0 5 1 3 5 5 0 0 0 0 0 2 0 1 4 1 % 75.0 87.5 34.8 33.3 51.9 50.0 2.8 5.9 84.2 97.4 89.5 64.3 96.4 92.1 100.0 100.0 0.0 9.4 6.3 15.4 0.0 0.0 14.3 27.3 43.5 45.5 65.2 63.3 25.0 100.0 28.6 51.6 40.6 58.3 53.8 80.0 100.0 50.0 100.0 84.1 55.3 94.5 99.0 70.6 100.0 80.0 66.7 40.0 6.9 4.3 0.0 8.6 1.8 13.6 9.6 17.9 0.0 0.0 0.0 0.0 0.0 40.0 0.0 33.3 18.2 20.0 Pregnant N 4 2 11 2 19 25 3 35 32 0 0 0 0 0 0 0 0 95 1 19 13 2 0 0 2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 7 16 2 62 1 9 0 1 8 9 54 21 25 16 20 3 2 1 0 0 0 1 0 0 0 2 7 1 % 25.0 12.5 47.8 66.7 48.1 50.0 97.2 94.1 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 100.0 59.4 27.1 15.4 0.0 0.0 28.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 11.1 42.1 2.7 1.0 26.5 0.0 20.0 33.3 60.0 93.1 91.3 78.1 27.6 35.7 13.6 3.8 3.6 0.0 0.0 0.0 12.5 0.0 0.0 0.0 66.7 31.8 20.0 Lactating N 0 0 4 0 4 0 0 0 0 2 1 6 3 0 1 0 0 13 0 5 14 4 3 1 0 1 6 2 2 0 0 0 3 1 2 0 0 0 0 0 0 1 1 1 47 0 0 0 0 0 0 0 0 7 35 30 14 31 14 10 1 3 0 0 1 0 0 4 0 % 0.0 0.0 17.4 0.0 0.0 0.0 0.0 0.0 10.5 2.6 10.5 21.4 0.0 2.6 0.0 0.0 0.0 15.6 29.2 30.8 42.9 100.0 0.0 9.1 26.1 18.2 8.7 0.0 0.0 0.0 10.7 1.6 6.3 0.0 0.0 0.0 0.0 0.0 0.0 1.6 2.6 1.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 21.9 60.3 53.6 63.6 59.6 50.0 55.6 50.0 30.0 0.0 0.0 20.0 0.0 0.0 18.2 0.0 Young on teat N 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 % 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 4.5 0.0 Pregnant & lactating N 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2 0 % 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 ().() 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 9.1 0.0 Postlactating N 0 0 0 0 0 0 0 0 0 1 0 0 2 1 2 0 0 6 0 5 18 5 4 0 4 7 7 4 6 1 3 0 17 30 17 7 6 4 0 1 0 2 0 1 149 0 1 0 0 0 0 0 1 0 2 5 2 14 8 8 1 7 7 2 2 1 0 4 3 % 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 5.3 0.0 0.0 14.3 3.6 5.3 0.0 0.0 0.0 15.6 37.5 38.5 57.1 0.0 57.1 63.6 30.4 36.4 26.1 16.7 75.0 0.0 60.7 46.9 53.1 43.8 46.2 20.0 0.0 50.0 0.0 3.2 0.0 1.4 0.0 2.9 0.0 0.0 0.0 0.0 0.0 4.3 0.0 3.4 8.9 9.1 26.9 28.6 44.4 50.0 70.0 87.5 100.0 40.0 100.0 0.0 18.2 60.0 Total 16 16 23 3 58 52 6 36 34 19 39 57 14 28 38 28 37 388 1 32 48 13 7 1 7 11 23 11 23 6 4 1 28 64 32 16 13 20 2 2 10 63 38 73 549 102 34 2 5 24 15 58 23 32 58 56 22 52 28 18 2 10 8 2 5 1 3 22 5 NUMBER 51145 TABLE 4-1.—Continued. Week 28 29 30 31 32 33 43 44 45 46 47 48 49 50 51 52 Total 1979 1 2 3 4 5 7 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 26 29 36 37 38 39 40 41 42 43 44 45 46 47 48 49 '50 51 Total 1980 1 2 3 4 5 6 7 8 9 13 14 15 No reproduction N 0 1 3 2 0 1 11 4 14 13 35 16 45 7 21 59 417 28 13 8 2 2 1 8 4 3 3 7 0 4 9 4 1 3 3 4 2 1 2 0 1 12 36 34 26 9 43 36 19 11 90 95 48 5 27 70 1 677 8 59 33 6 3 17 2 17 1 1 10 3 % 0.0 100.0 9.7 100.0 0.0 14.3 91.7 80.0 82.4 86.7 92.1 94.1 95.7 100.0 95.5 98.3 96.6 100.0 29.6 22.2 33.3 33.3 47.1 20.0 3.0 18.8 15.6 0.0 3.8 7.4 6.9 100.0 25.0 25.0 10.5 3.9 3.7 3.7 0.0 9.1 52.2 56.3 72.3 78.8 69.2 74.1 92.7 95.0 100.0 100.0 94.1 98.0 83.3 93.1 98.6 100.0 88.9 96.7 97.1 85.7 75.0 94.4 50.0 63.0 100.0 2.4 8.9 2.2 Pregnant TV 3 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 185 0 0 19 7 4 2 2 3 4 0 1 0 5 11 7 0 1 5 24 38 18 39 2 1 0 0 0 0 0 0 1 0 0 0 4 1 0 2 0 0 201 1 2 1 1 1 1 2 10 0 1 6 18 % 75.0 0.0 3.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 70.4 77.6 66.7 66.7 11.8 15.0 4.0 0.0 2.2 0.0 4.8 9.1 12.1 0.0 8.3 41.7 63.2 74.5 66.7 72.2 66.7 9.1 0.0 0.0 0.0 0.0 0.0 0.0 2.4 0.0 0.0 0.0 4.0 2.0 0.0 6.9 0.0 0.0 11.1 3.3 2.9 14.3 25.0 5.6 50.0 37.0 0.0 2.4 5.4 13.2 Lactating N 1 0 16 0 3 4 0 1 0 0 0 0 0 0 0 0 175 0 0 0 0 0 0 6 10 89 12 31 2 63 71 26 0 0 2 7 2 7 13 1 9 6 6 4 3 0 1 0 0 0 0 0 0 1 0 0 0 372 0 0 0 0 0 0 0 0 0 35 82 84 % 25.0 0.0 51.6 0.0 50.0 57.1 0.0 20.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 35.3 50.0 89.9 75.0 68.9 100.0 60.0 58.7 44.8 0.0 0.0 16.7 18.4 3.9 25.9 24.1 33.3 81.8 26.1 9.4 8.5 9.1 0.0 1.7 0.0 0.0 0.0 0.0 0.0 0.0 16.7 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 83.3 73.2 61.8 Young on teat N 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 % 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Pregnant & lactating N 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 % 0.0 0.0 3.2 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 7.9 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Postlactating N 0 0 10 0 3 2 1 0 3 2 3 1 2 0 1 1 97 1 0 0 0 0 0 1 3 3 1 6 0 32 30 21 0 8 2 0 9 1 0 0 0 5 22 9 4 4 14 2 1 0 0 2 0 0 0 1 0 182 0 0 0 0 0 0 0 0 0 5 14 31 % 0.0 0.0 32.3 0.0 50.0 28.6 8.3 0.0 17.6 13.3 7.9 5.9 4.3 0.0 4.5 1.7 3.4 0.0 0.0 0.0 0.0 0.0 5.9 15.0 3.0 6.3 13.3 0.0 30.5 24.8 36.2 0.0 66.7 16.7 0.0 17.6 3.7 0.0 0.0 0.0 21.7 34.4 19.1 12.1 30.8 24.1 4.9 5.0 0.0 0.0 2.0 0.0 0.0 0.0 1.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 11.9 12.5 22.8 Total 4 1 31 2 6 7 12 5 17 15 38 17 47 7 22 60 878 29 13 27 9 6 3 17 20 99 16 45 2 105 121 58 1 12 12 38 51 27 54 3 11 23 64 47 33 13 58 41 20 11 90 101 49 6 29 71 1 1436 9 61 34 7 4 18 4 27 1 42 112 136 46SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY TABLE 4-1.—Continued. weeK 16 18 19 20 21 32 33 34 35 36 37 38 39 40 41 42 43 Total No reproduction N 0 0 9 2 7 8 15 9 13 5 26 36 14 33 59 38 6 440 % 0.0 0.0 32.1 8.3 38.9 61.5 27.3 13.8 23.2 14.3 40.0 64.3 66.7 64.7 73.8 88.4 100.0 Pregnant N 0 2 1 13 7 0 0 0 0 0 0 0 0 0 0 0 0 67 % 0.0 66.7 3.6 54.2 38.9 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Lactating N 2 0 3 7 0 4 36 20 9 8 13 3 1 1 2 0 0 310 % 50.0 0.0 10.7 29.2 0.0 30.8 65.5 30.8 16.1 22.9 20.0 5.4 4.8 2.0 2.5 0.0 0.0 Young on teat N 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 % 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Pregnant & lactating ~N 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 % 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Postlactating N 2 1 15 2 4 1 4 36 34 22 26 17 6 17 19 5 0 261 % 50.0 33.3 53.6 8.3 22.2 7.7 7.3 55.4 60.7 62.9 40.0 30.4 28.6 33.3 23.8 11.6 0.0 Total 4 3 28 24 18 13 55 65 56 35 65 56 21 51 80 43 6 1078 0) CO #•* t* mm 9 o k. o Q. 90 - 807060 - 50 - 40 - 3020101977 \ \ \ n I V '• : lv I t • "• • • 1978 a 1 *. l\ •• \ •:<- V"?i / •' ' hi, \ ••• 1 I A / / i 5 / 1979 ' I !\ K1 » Mi1 1 T ! 11 1 « i i i ii ' •' \ • 1 ! /I/M h\ 1980 # n n n 11 i i i i i i • t 1SA •r \ I v V. • * • *» : • \ *. \ '. \\\\ 10 20 30 40 50 8 18 28 38 48 6 16 26 36 46 4 14 24 34 44 Weeks FIGURE 4-1.—Percentages of pregnant (solid line), lactating (dashed line), and postlactating (dotted line) female Artibeus jamaicensis on BCL Each record represents four weeks combined. NUMBER 51147 706050 H O O) « 40c © o S 30H Q. 2010Jan Feb Mar Apr May Jun Jiil Aug Sep Oct Nov Dec FIGURE 4-2.—Average percentage of pregnant Artibeus jamaicensis on BCI (based on data from 1972 through 1980, pooled in three week increments). clear that not every female takes part in every breeding period. There are records from each week of the year of females showing no evidence of reproductive activity (Table 4-1). At the peak of each reproductive episode, the number of pregnant females averages about 70% (Figure 4-2). For females originally banded as juveniles or subadults, we have 894 capture records in the year following their birth. Of these, 547 or slightly more than 60% were recorded as nonreproductive (neither palpably pregnant, lactating, nor postlactating). From a sample of 1121 capture records of adult females (individuals that had experienced an earlier pregnancy), 449 or 40% were nonreproductive. During any given sampling period, there are almost always more reproductively active adults than there are yearlings (Figures 4-3 and 4-4). A summary graph (Figure 4-5) combining data from 1977 through 1980 shows that yearlings not only breed less frequently, they are more likely to be out of synchrony with the reproductively experienced adults. The Reproductive Cycle ESTRUS.—Each female has two periods of postpartum estrus per year. The first occurs immediately following the birth in February-March; the second immediately following the birth in July-August. Thus, the estrous cycle is easily controlled by the timing of parturition except in those females that, for whatever reason, fail to give birth during a particular reproductive period. We do not know what physiological and environmental factors cue estrus at the appropriate time. Perhaps estrus in females that failed to become pregnant is resynchronized by contact with 48 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY 0 O> (0 *1 9 O h. <D Q. 1009080706050 40 - 3020101977 • i\ i > i V i \ i L. it n • i M 1 / ^ 'i 1 V , \ , 1 1 1 t [ 1 1 t A f \/ 1978 J h li i i ft ' l ' /I ' i I M ' i i ; i r 'i 11 f! 'If I ft ; ; 1 1 I I 1 / J K 1 1979 j. T ' i i i i i i i i I i i i 1 f I •A ! x\\ I 1 Ii \[i \ /' i l ! ) ILL i 1980 i\ / \ ' i / \ ft / \ 10 20 30 40 50 8 18 28 38 48 6 16 26 36 46 4 14 24 34 44 Weeks 10 20 30 40 50 8 18 28 38 48 6 Weeks 16 26 36 46 4 14 24 34 44 NUMBER 51149 105 Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec FIGURE 4-5.—Average percentages of pregnant yearling (solid line) and older (dashed line) female Artibeus jamaicensis on BCI, based on data from 1977-1980 combined. Each record represents one week. adult female roost mates who are following a normal cycle. COPULATION.—Males have enlarged testes during the periods when females are undergoing postpartum estrus (see Section 3, Reproduction in a Captive Colony). The degree of asynchrony in the system indicates that males may be capable of inseminating females for relatively long periods overlapping both birth peaks. EMBRYONIC DEVELOPMENT.—Fertilization and implantation follow a normal sequence during each of the reproductive episodes, but subsequent development differs between the two. The normal gestation period following implantation during the first reproductive episode of the year is 3.5-4 months. FIGURE 4-3 (opposite, top).—Percentages of pregnant yearling (solid line) and older (dashed line) female Artibeus jamaicensis on BCI. Each record represents four weeks combined. FIGURE 4-4 (opposite, bottom).—Percentages of nonreproductive yearling (solid line) and nonyearling (dashed line) female Artibeus jamaicensis on BCI. Each record represents four weeks combined. Implantation during the second episode is followed by delayed embryonic development, which results in a gestation period of about seven months (Fleming, 1971). This is a unique feature of the reproductive cycle of A. jamaicensis. The cues are unknown, but because the cycle persists when animals are moved into captivity under environmental conditions unlike the natural ones (see Section 3, Reproduction in a Captive Colony), it must be under some genetic control. Eventually, the synchronization breaks down under constant conditions of captivity, suggesting that one or more environmental cues are necessary to reset the system. PARTURITION.—The two peaks of parturition are in February-March and July-August. The neonates are well-enough developed to be able to hang by themselves within a few hours of birth (see Section 3, Reproduction in a Captive Colony). Only on two occasions did we capture a female with an attached young. Apparently, the young are left behind shortly after birth while the females forage. LACTATION.—Lactation is easily detected and lasts for about two months (see Section 3, Reproduction in a Captive Colony). 50 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY I | | | I I I | I I I I jl I I I I I I I I I I I ( I I I I I ' I I ' I I ' | ' ' ' ' ' I ' I I I I I 1977 1978 1979 1980 O CO O Q. 50 • 250 - 25 50 • O (0 O) o 50 25 0 25 50 50 25 0 25 50 60 40 J FMAMJJASONDJ FMAMJJ ASON DJ FMAMJJ ASON DJ FMAMJ J ASOND 1977 1978 1979 1980 FIGURE 4-6.—Schematic representation of the reproductive cycle of Artibeus jamaicensis on BCL showing temporal variation in percentages of pregnant, lactating, and postlactating females. Mean rainfall is shown in the bar graph at the bottom. Time is in months and years. Peaks in percentages of lactating females follow about a month behind the peaks for pregnant females (Figures 4-1 and 4-6). Lactation in captivity lasts a bit longer in females that are carrying dormant embryos following the second breeding episode. This phenomenon can be inferred in our wild population (Figure 4-6). POSTLACTATION.—We were able to glean additional information about the cycle from individuals we recognized as postlactating. Apparently there is more individual variation in the length of time that nipples are classified as postlactating than in other reproductive criteria (Figures 4-1 and 4-6). Part of this variation may be an artifact of stages of refinement of our definition of postlaction. Early in the Bat Project, we presumed a bat to be postlactating if no milk could be expressed from nipples although the nipples were large, flabby, and surrounded by naked skin. Later, the category was expanded to include females showing new hair growing in around the regressing nipple, regardless of its size. NUMBER 51151 Seasonality, Rainfall, and Abundance of Food Previous hypotheses (Wilson, 1979) have focused on energetics as the primary control on reproduction, and this indeed seems the most logical explanation for the basic pattern of the cycle in A. jamaicensis. Other factors being equal, one would expect animals to achieve maximum reproductive output without sacrificing survivorship of the young. If we assume that the physiological constraints operating on an A. jamaicensis result in a litter size of one and a gestation period of 3.5-4 months, then in a world of abundant energetic resources, females simply should reproduce continually, producing up to three young per year. That they do not suggests that there must be a time when energy sources are in short supply. The time of maximum stress might occur at any of several critical points in the reproductive cycle. Pregnancy, as does lactation, carries with it an increase in energy demands to individual females (see Section 2, Physiology; Studier et al., 1972). Reproductive success should be enhanced if these stressful periods coincide with periods of peak resource availability (Bradbury and Vehrencamp, 1977b). The critical time for most bats is when the young are weaned. Not only do the growing young have high energy demands, they are forced to leam to forage at the same time, and presumably it would be easier to do so when food is readily available. Readily available food resources also are critical at the time of weaning in a species showing synchronized reproductive cycles, because the population size is markedly increased at that time. There is now a considerable body of evidence supporting the position that reproductive events in bats are timed to synchronize with seasonally abundant food resources (Dinerstein, 1983; Heithaus et al., 1975; Racey, 1982). Young born in the first reproductive episode (FebruaryMarch) are weaned in April-May, and those born in the second period (July-August) are weaned in September-October. This pattern is confirmed by our capture records, as the number of juveniles and subadults increases dramatically during these periods. On BCI this results in a general pattern in which the young are weaned during the rainy season (Figure 4-6). If the bats conceived during the second episode were to develop normally (in 3.5-4 months), they would be born in November and weaned at the beginning of the dry season in January. Foster (1982) has shown that the peaks in fruit availability on BCI occur in April-May and September-October, with a low in January at the beginning of the dry season (Figure 4-7). Therefore, it seems that delayed development has evolved as a mechanism to avoid weaning young during the dry season when fruiting trees are scarce. Thus, the reproductive cycle of A. jamaicensis seems to be an adaptation to the seasonal cycle of food abundance. Evidence from other localities supports this hypothesis (Wilson, 1979). O < u. Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul FIGURE 4-7.—Schematic representation of seasonality of fruiting activities of canopy trees on BCI. "Activity" represents the proportion of species that are fruiting (determined by seed traps). Redrawn from Foster (1982, fig. 12). s « so 25 «> o oo o _; o 2 f "i o f: » ? o ^ 2 & 2 IS o _ 00 fs| K. _ oo op ^ m 2 « _: u < a. ao ofl G yea ivel uccess c •vival ON OO ro en # 2: rked |.| 1 | en ON o v> a 00 1976 =3 en CO t-~ oo ON ON CN| E: oo vn SO 1977 =3 ti. ON VI vn oo en en B en 9t oo r~- "oo a. 00 o 00 •—1 cs cs en S3 978 all! u. r~ 2 2 P R Q. 00 a o R en 979 1 a 3 "oo o. 00 ON s o9 | cs SN o oo 78oc ON T o 3/3: SO o ? oo o re 38/3< s s 2 -79976 iia. en ON •<r •«r o 3/3: SO - R o ,~. so" o ?5 o 2 p5 <s !I c^ •«r M fl, s < en 162 >o VO 8 I/O I en ON o/6: ts s en oo s 79/9' 00 891 E: s en s r00 •o S 1 1 • u X 1 u CQ § •a I •I |u 1 g i a. ;o J= > 12 -5 ]\ 11 co co 1 CN CO O O -H ON VO ON r— vr> ON 22/31 CsSS COVOCNVOCN ** *r ^ CO CO CO CO ovr-NcjpwQoooo t-'-COOOONXSVOcN CN CN «S CN ^- CO VO 00 VO 00 ON Q OO f. r~ t— * r> i " VOr^ON00ONQ\ONON.\0 g r~ r— *~ r— •« ^ « - f. ,a ON ON ~Q ON QQ ON QQ QQ ^ u iiiilillia btu(/)U.cou.c/}c<nu.< .5 la •5 -a P 1 .3 .9h "3 BA vo II oflif c ilfyea vc hs | 3 M c 73 Surviv Number o ON OO pNO vo CO I subadulls "3 Peri< 2: 2: * 2: 2; # 2: * 2: 2; 2: marked CO r> vfi ^1 p. CN VO NO CO ON CO CO 00 227 11 1976 ~~ 00 00 cs >o vo vo cs NO CN O* CO 0 CS 1 358 II 1977 £ cs CO CO cs 0 cs CO t~ cs rcs CO vo 1 171 oc taON cs vo 5 CO CO NO 0 vo -T __ CO 1 289 ring 19 11 1978 0 CN CO cs CN vo ON CO CO 00 CO >o ; 141 nng 19 NO CN 00 NC [ 257 11 1979 — CN T O^ Q X ring 19 CO 23, pcs CO $ VO >o cs CN ro T CO 12/33 | 331 0 X X r~ ring 19 c pr7/22 7/22 NCNC 14/227 14/227 vo 'S8 - vo 00 vo 64, - 79/: oc CN CO 12/1 CO CO •<r 19/90 CO ! 1131 II 1976 £ r- >o oc vo - 0 NO VO 76/7 •«• vo s 148/1 8 s 3/1 a cs CO 3 91/1' ro >O ro vo 1/123 ro T 1462 1 < CN •I 18/400 vo 00 8 3/1 vo 00 VO vo CO 109/1 CN 209/1 00 s s co CO CN CN 99/2.' >n CO vO a ci 10 2636 3 K L If years < JZ u iccessi g irviva on 2: 2: tf c 2; 2: o 2: 2; tn Number lubadulls marked Penod es T vn Ov o> vn o oo es 00 en oo Ov 173 Fall 1976 - en t ri Cv X rs (S T M en g 108 300 FaU 1977 vn rts rs n X fS r-i s 127 Spring 1978 en CV ei m 159 r- <s ON oo O OO Ov fS — Ov en en eN si Spring 1979 FaU 1979 173 1 es 173 en /473 es vn vo r» en rs r— 2^ vo r~ Ov OO r~ es — vn •n ts 173 en en rrr •"• vn g en en oo a 3 vn vn vn 41/281 1 97/632 : O oo - oo vn vn vn O en — en en c*» oc s Spring 1980 Spring 1978FaU 1976-79 en en Ov 109 1 208/1 es en g 320/1 1174 AU periods JJ 2 3 I - i oo oo oo r» r-~ r~ r-^ f~ P^ rt -rr OO r~ TJ- ^-1 oo ^f en ts >n oo rs rs oo — oo en - en 3 s 129/ - | en 676 24/1 vn oo vn •-" oo oo vn c— — vo oo vn 164 o en s en en oo rs 104 rs oo en oo vn es 1 205 00 en 1 239 o en 1473 i o en 761/2576 oo ^S I*** en oo vi ^r O* Ol ^^ r^voooovovnovvovn — enes — en^-^oooovo vo rr-- fOv Ov O> O 00 p. — p- — — r~ 00 Ov oo ofl Ov u. oo on u. •a E ob 1.1 •£ a years after sive half' cces val in su Survi oo •> NO vn en * * Number adults marked Period C Q en TT en vn oo *"" O ts s 00 151 1976 Fall _ •"3ON o CS s 123 1 299 1977 cs s o en cs 5 en vn I 185 oo oo Sprii VO NO CS * en cs TT i 107 1978 Fall o oo '-i o — ts ts ON O Tt en cs rS 2 ig 1979 1979 'a. ^ cs 4/185 16/185 ON 40/427 64/427 vn en — cs = 8 113/ £8 >n o ig 1980 tig 1978Sprii Sprii en 4/151 CS o vn ON vn 21/450 ON 48/557 ON 82/557 cs1 156/754 en en s 1 <s | 754 1976-79 Fall en 4/151 CS o vn ON rr 25/635 ON 64/742 122/984 ON 220/1181 ON ts 103 317/11254 ye riods 62SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY LU > UJ ffi O z 40-r 3020 1040 30 20 10 UJ O 40 oc UJ Q.30 20 10 M A=cf Spring •= 9 A=cf Fall O= 9 Fall a I 3I 4I 5 I 6i 8 HALF YEARS AFTER MARK FIGURE 5-5.—Comparison of survivorship of seasonal cohorts of Artibeus jamaicensis marked on BCI and adjacent mainland from 1976 to 1980; a = adults, b = subadults, and c = juveniles. NUMBER 511 63 basis and, although producing important information on movements, these bats did not form a major part of the main population we were studying. When the A. jamaicensis from these off-island sites and their recapture records are removed from the data (presented in Table 5-9), thereby restricting the analysis to bats marked on BCI and Orchid Island, survival estimates are higher and vary from a low of 17% for adult males marked in the first half 1980 to a high of 71% for juvenile females marked in the spring of 1978 (data from Table 5-10). JUVENILES.—Overall survivorship into the second half year of life by juvenile A. jamaicensis was 31% (Figure 5-6, Tables 5-3 and 5-4). Survivorship of juveniles ranged from 25% for males marked in the falls of 1977 and 1979 to 57% for females marked in the spring of 1978 (Table 5-3). Survivorship from combined data from the fall of 1977 to the spring of 1980 suggests that approximately 68% of the juvenile cohort is lost by the time juveniles enter their second half year (Tables 5-3, and 5-4). We had anticipated high mortality during the first six months of life based on studies of temperate-zone bats (Brenner, 1968; Davis, 1967; Foster et at, 1978; Humphrey and Cope, 1970,1976, 1977; Keen and Hitchcock, 1980; Mills et al., 1975; Pearson et al., 1952; Stevenson and Tuttle, 1981; Tuttle and Stevenson, 1982). Survivorship from the second to sixth half year after marking among bats marked as juveniles averages lower than that of adults and subadults (Figure 5-6). Dispersal during the first year of life is the most likely explanation for this higher observed "mortality" or disappearance rate. Nevertheless, factors such as death, learned net avoidance, and loss of necklace (which renders the individual unrecognizable), as well as dispersal away from the area of our study, contribute to the disappearance of these bats. Survivorship by the end of the second half year of life (second half year after mark) averaged 17% (Figure 5-6, Tables 5-3 and 5-4). Thereafter, the rate of decline parallels that of subadults and adults. Necklace loss may be greater in juveniles because we had to fit bats of this age class with adult-size necklaces. However, necklace loss cannot explain the more rapid decline in the second half year of life by which time the bats have attained near adult size. Survivorship by the end of the second half year of life (second half year after mark) averaged 17% (Figure 5-6, Tables 5-3 and 5-4). Thereafter, the rate of decline parallels that of subadults and adults. We examined the number of bats caught in each half year versus the number known to be alive in that half year (Figure 5-7, Table 5-11) to understand the probable causes for the more rapid decline from the first to second half year of life in juveniles. The consistently higher percentage of males caught of those known to be alive in the first half year after being marked reflects an initially higher recapture rate among males a HI DC 40 n I30 a. o LJJ OC 20 LU O CC LJJ Q. 10 - M Artibeus jamaicensis O= Juveniles A= Subadults • = Adults O 2 3 4 5 6 HALF YEARS AFTER MARK8 FIGURE 5-6.—Survivorship curves of the three age classes of Artibeus jamaicensis marked on BCI and adjacent mainland from 1976 to 1980. SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY FaU 980 8 * en oo v> es —• •* •C? en ••« en Os so en — t— >n —• en v> in r-- en es — T o ^ ^ ^ ^* ^ &\ Cl -^ __4 C*1 s <S — CN <S =3 ». 2 CS? "? P en O en ov » go o V es — — ©' BS r— <n o — 2 fs IISO 00 \Q — en vi so Os <s en en— O OS OS so o os r~ 11 "8 •a e . 3 Ej 2 3 3 2 "S la 1.* 5 = 8-5 •SJ B.2 a I 323 (63) 247 G149 (£01) 405 193 196 (£61) 728 (801) 440 (77) 345 222 (102 261 t 126 oo en 153 1 518 s 240 H28 375 es so S?8 en (N 142 SO, 227 52 162 en oo 252 389 II {]J M 3 lla III z S I W5 NUMBER 51165 LU O QC LU Q. 80604020 80 6040- • o o • c 806040 20-F S F S F 1977 1978 1979 o • F F S F S 1976 1977 1978 1979 FIGURE 5-7.—Comparison of numbers of Artibeus jamaicensis marked on BCI and adjacent mainland and recaptured in the first and second half years after mark. Values plotted are recaptures expresssed as percentages of the numbers known to be alive in the same half year, open circles = females, closed circles = males. Adults, (a) first half year and (b) second half year of life; subadults, (c) first half year and (d) second half year of life; and juveniles, (e) first half year and (f) second half year of life. 66 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY (more recaptures within a shorter time). The same pattern among subadults (Table 5-12) and adults (Table 5-13) for the half year following their being marked suggests behavioral differences between the sexes that in turn increases susceptibility to recapture among males. The pattern is similar among the three age classes for the second half year after mark, except that the relationship is reversed for juveniles in the fall of 1977 and the spring of 1978, and for subadults in the fall of 1978. Thereafter, there is no pattern emphasizing greater catchability of males. Juveniles of either sex marked in their first half year of life are known to be alive in the second half year in about equal numbers (Tables 5-3 and 5-4; 30% for males versus 32% for females). Nevertheless, by the third and fourth half years, the survivorship rate is higher in females (21% versus 14%, X2 = 13.383, P < 0.01, for third half year; 14% versus 9%, X2 = 9.159, P < 0.02), suggesting either greater mortality in males, differential emigration, or both, along with possible learned net avoidance resulting from higher actual recapture rate during the first year after mark. Chi-square analysis of bats marked as juveniles showed some significant differences in survivorship correlated with sex and season of birth. In general, females have higher survival rates than males. Spring newborn of both sexes survive better into the second half year of life than do fall newborn with the difference between season of birth for females significant at the 0.05 level (35% versus 28%, respectively; X2 = 6.103). In subsequent half years, the survival rate is similar regardless of birth season. Spring-born young may have the advantages of long gestation (seven months versus four), birth in the dry season, weaning at the beginning of the wet season, and attainment of subadult age (and near adult size) by the time of fig scarcity at the height of the rains. Fall-born young, the product of a short (4-month) gestation period, are born in the wet season, weaned at the height of the rains when ripe figs are beginning to be scarce, and become subadults in the dry season. Preweaning events such as length of gestation and coincidence of birth with wet or dry seasons, may have little effect on survival of baby bats. Postweaning conditions, however, when young bats are on their own for the first time, may be critical. Fall-born bats are weaned in the season of heaviest rains when there is a greater probability of getting wet while foraging. The likelihood of getting wet is increased by the scarcity of figs because there is more reason to fly in the rain if quality food is harder to find. Wetting and consequent chilling could be fatal for weanling A. jamaicensis because of their restrictive, relatively inflexible energy budget (see Section 2, Physiology). Survivorship to the second half year (first half year after mark) is not significantly different for males and females marked as juveniles. However, females appear to survive at a higher rate by the third and fourth year (P < 0.01 and P < 0.02, respectively; based on data in Tables 5-3 and 5-4). Survival in the third half year (second half year after mark) favored fall-born females over fall-born males (P < 0.05). Reasons for these differential survival rates are unclear. The higher initial recapture rate characteristic of males and their higher subsequent attenuation rate in the population are both reflected in these rates. Nevertheless, we can not be sure how much of the apparent attenuation rate actually reflects greater differential emigration by males (higher probability of continued residency by females) instead of mortality. SUBADULTS.—Speculation on survival and recapture rates of bats marked as subadults must include the assumption that these bats are already in the second half year when first captured. Bats marked as subadults in the fall were born the previous spring, and those marked in the spring were born the previous fall. Therefore, the first half year after being marked is the third half year after birth. Any comparisons of these records with those of juveniles are on this basis. There is some unavoidable "slop" in aging subadults. Subadults caught late in the fall include all individuals born in the previous spring and some born early in the fall. In the first weeks of spring the pool of subadults includes all young of the previous year, and should be at its largest. Midway through spring the pool should have diminished to its lowest level because most have become adults (reproductive) regardless of age. Later in the spring, juveniles of that spring begin to cross the threshold to subadulthood. However, numbers of subadults marked in the fall greatly exceed the number marked in the spring (Tables 5-5, 5-6, and 5-9). This is simply because in a year-round capture program, most of the subadults are captured and marked in the fall, leaving fewer unmarked bats to be captured in the spring. Bats are aged as juveniles only on the basis of open epiphyses of the metacarpals and phalanges of the wings. Transition from juvenile to subadult status (ossification of the epiphyses in the wing) is relatively rapid (see Section 3, Reproduction in a Captive Colony). Assuming that subadults already are in their second half year when first captured, survival rate is slightly higher than in juveniles (Figure 5-6). Nevertheless, the proportional loss from the population of the cohort marked as juveniles through the second half year after marking must have already been absorbed by the cohort marked as subadults when these bats complete the first half year after having been marked. Reasons for that loss are the same as those outlined for losses among bats marked as juveniles, except that necklace loss may not be as important a factor. Observed overall survivorship through the first half year after mark (34%) for subadults was not significantly different from that for juveniles (32%, X2 = 3.225), whereas recapture rate for the second half year after mark was significantly lower (17% versus 23%,^ = 22.300, P < 0.001) for juveniles (Tables 5-3 and 5-5). The highest survival rate for bats marked as subadults was 43% in the fall of 1978 (males, 45%; females, 39%; Table 5-9). As we suggested for our survivorship data on juveniles, the survival rate for subadults is undoubtedly higher than the 43% we recorded during the seven half-year periods in which we accumulated survivorship data. If we restrict our data NUMBER 511 67 "3 I Si .5 I o C | c | s juvi •s • 5 § 1 s. u 1 -11 3 TAB u ca § •s 1 •a .§ a "dm 8P: •—' Os OO 00 vn vn vo oo — Os en ~ Cj QJ « — en r~- oo vn oo o en tr~ >n os — — — —. en CA oo r— — oo en t— — o oo ^c?;cis:^ — (S oo o r-. en en O Os o >n vn C CN — (; - pOO Ov so — rs m •* SO f** OS oo ^ Os Os Os Os 5jj 7v 00 ^ 00 iiiili! u. u. oo uu oo tu oo oo Os O tr55 so r** ^ oo os os os Os Os oo ^ 00 ^ oO a a •§ a '§ a •§ ^3 ^3 o.^3 Q.T3 o. pu u. oo u. oo ou oo »^; £2 s ^S2 >n en <n en en — c-> *n vi \r> oe oo — TJ- ?r o" so" r!OSN og g 5^ ros jn ^ci C ~ ?? PI * t~ VD N O» oo o^ ^ o^ jn en _ __ _ JJs p-, ir en cs so •-" Os en oo *—* oo os — ciqcifN n S3 o •< 2 * <>?~\ CO 34, oo *e 39, oo in N O JO. sg "vnn r- — •«• ^ QO ^ 00 ^ QO iliiii tu c^ uu w Uc^ O^ Os ofl Os Qg O^ oo 3=3 74 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY attack. We have attracted the Spectacled Owl (Pulsatrix perspicillata) to nets by imitating bat squeaks, and have caught these owls in nets set for bats. Handley watched a Mottled Owl (Ciccaba virgata) ambushing Myotis nigricans and Molossus coibensis at their roost at dawn when the bats were milling about before entering. The owl stationed itself in a tree off to the side and higher than the roost and made frequent sorties down through the group of milling bats, usually successfully. We caught Mottled Owls fairly frequently in nets and occasionally have caught the Vermiculated Screech-Owl (Otus guatemalae). At least seven species of owls are known to occur on BCI, and some of them are rather abundant. The Bat Falcon (Falco rufigularis), mainly a crepuscular feeder and known to be an effective predator of bats (Ridgely, 1976), is rare or absent on BCI but nests nearby. The timing of the departure from and return to day roosts by A. jamaicensis makes it susceptible to predation by the Bat Falcon. Diurnal hawks often take bats, especially when they have been disturbed in their roost during the day. Several African hawks regularly intercept bats along their flight paths (Black et al., 1979; Fenton, Cumming, et al., 1977). Habits characteristic of A. jamaicensis such as lunar phobia and picking fruit from a tree and carrying it a hundred or more meters away to be eaten at a feeding roost, are presumably defensive strategies (Howe, 1979). The frequent use of the fronds of the spiny-trunked black palm (Astrocaryum standleyanurri) and the arching fronds of the palm Oenocarpus panamanus as dining roosts by A. jamaicensis may be a defensive behavior to protect itself from predators (see Section 8, Roosting Behavior). Response to distress calls of bats that August (1979) called "mobbing" may also be a defensive mechanism. At any rate, there are enough seemingly defensive behavioral traits in A. jamaicensis to suggest that it has considerable exposure to predators. ACCIDENTS.—Many netted A. jamaicensis have damaged wings. Nevertheless, they appear to survive accidents that tear wing membranes or break fingers. These parts heal remarkably rapidly and flight is not notably impaired, as we have seen in our NZP colony. Common sources of injury are fighting among individuals and encounters with sharp plant spines or thorns. Bite wounds are most frequent in males, particulary young males trying to get access to females (observed in the NZP colony). Judging by netting results, bats may avoid areas in the forest where spines and thorns are particularly prevalent; although many, if not most, of the tears in flight membranes that we have observed must originate from this source. Broken bones may result from struggles to get free when wings get tangled. A broken forearm or humerus, rendering the bat flightless, is invariably fatal. DISEASE.—Artibeus jamaicensis is known to have been infected with rabies and trypanosomiasis on the Pacific side of the Canal de Panama area, but disease in bats is unknown on BCI. Summary On BCI A. jamaicensis made up about 66% of the total nightly catch of bats year-round. Numbers tended to be highest when young were first volant and lowest when the population was composed mostly of subadults and adults. The entire sample of A. jamaicensis marked between October 1976 and June 1980, consisted of 8907 bats captured a total of 15,728 times. Fifty-seven per cent (5061 individuals) were caught only once, leaving 3846 captured two or more times. Subsequent captures per bat diminished until we had a single individual captured 11 times. The probability of recapturing a bat was roughly the same regardless of the number of previous captures. Overall survivorship into the second half year by juvenile A. jamaicensis was 31%. Survivorship from the second to sixth half year after marking among bats marked as juveniles was lower than that of adults and subadults. Dispersal during the first year is the most likely explanation for this high rate of disappearance. More frequent recapture of males in the half year following marking suggests behavioral differences between the sexes that increase susceptibility to earlier and more frequent recapture among males. Juveniles of either sex marked in the first half year are known to be alive in the second half year in about equal numbers (30% for males versus 32% for females). But by the third and fourth half years, the survivorship rate is higher in females, suggesting that males are more likely to emigrate, have greater mortality, and possibly learn to avoid nets because of their higher actual recapture rate during the first year after mark. Overall, females have higher survival (residency) rates than males. Frequency of capture seems to have less to do with age than with sex. Apparently most bats are easier to capture the first time, regardless of age: 57% caught once, 25% caught only twice, 18% caught more than twice, and thereafter the capture curve levels out In our sample of almost 17,000 records of A. jamaicensis (October 1976 to October 1980, adult and subadult females outnumbered adult and subadult males 55 : 45, while juvenile females are outnumbered by juvenile males 48: 52. The recapture histories of bats marked as juveniles also support the contention that females have higher survival (or residency) potential than males. Among known-age bats, the mean age at last capture was 2.02 years in the females (n = 110) and 1.65 yrs in the males (n = 115). The average for the entire sample (n = 225) was 1.83 years. The oldest male was four and the oldest female was 6.5 years old. Subsequent to 1980 Handley caught two A. jamaicensis that were nine years old. With potentially great longevity, and with most female A. jamaicensis producing two young each year, the species must be under severe population controls. Death rate must be high, but its causes are speculative. Mortality factors include NUMBER 51175 starvation, predation, accidents, and disease, but we have little direct evidence and we have seen few dead bats. Young of the "spring" birth group are weaned in a season of abundant food (April-June) at the beginning of the rainy season. Young of the "summer" group, on the other hand, are weaned in a season of food scarcity (August-October) in the middle of the rainy season when they could easily starve before becoming efficient foragers. Young A. jamaicensis face a number of risks. With a high metabolic rate and a low protein, high carbohydrate diet one would expect starvation to be abrupt, perhaps in a single night if food sources are not adequate. Young A. jamaicensis and all ages of the smaller-size species of Artibeus and Vampyressa become unable to fly, lose control of body temperature, chill, and die in as little as an hour under stress and food deprivation. Rain must increase the risk of starvation in young weaned in the rainy season. Artibeus jamaicensis relies on a food source that is strongly pulsed in time and space. It must search out resources for the future while harvesting a current source. Its behavior and diet adds risk from predators that may be attracted to the food source. Snakes and owls are important predators of bats. Falcons, opossums, coatimundis, and the large carnivorous bat Vampyrum spectrum also are potential predators. There are enough seemingly defensive behavioral traits in A. jamaicensis to suggest that it has considerable exposure to predators. Artibeus jamaicensis is a strong, robust bat that usually survives accidents that tear the wing membranes or break fingers. These parts heal remarkably rapidly and flight is not notably impaired. Many netted A. jamaicensis have damaged wings. Common sources of injury are fighting among individuals and encounters with sharp plant spines or thorns. 6. Population Estimates Egbert G. Leigh, Jr., and Charles O. Handley, Jr. Between 20 October 1976 and 19 May 1980, Bat Project participants marked 8907 Artibeus jamaicensis, which had been captured a total of 15,728 times by 20 October 1980, the last night of field work covered by this section. Can this record tell us how long individual A. jamaicensis live, and how many there are on Barro Colorado Island (BCI)? For purposes of the ensuing analyses, we will assume that we are sampling the entire population of A. jamaicensis on BCI, and that it is a closed population. A Simple Estimate AVERAGE LIFETIME.—Suppose A. jamaicensis has an exponential "life table," in which the probability of a bat living past age y is emy, where e (the base of Napierian logarithms) is 2.71828, and m is the annual death rate per capita. Then, if the capture effort, averaged over a period during which a quarter of the bats are replaced, does not vary excessively from one such period to another, we may estimate the average lifetime 1/m of these bats by the average time elapsed between first and last captures of bats captured more than once. We accomplish this as follows: 1. Let cdt be the probability that a given bat, alive between time t and time t + dt, is captured during this time interval. Here, c is the capture rate and dt denotes an "infinitesimal" time interval. Then the probability of a bat that lives L years being caught n times in its life is [(cL)n/n!]exp(-cL) 2. If the probability that a bat's total lifetime lies between L and L + dL years is (mdL) exp -(mL), then the probability that a bat will be caught n times in its life is P(fl) = IT (mdL) exP (-mL)[(cL)n / n!] exp (-cL) = Jr t(cL)n / n!](mdL) exp Rrn + c)L] = (m/c)[c/(c + m)]n+1 Egbert G. Leigh, Jr., Smithsonian Tropical Research Institute, Unit 0948, APO AA 34002-0948, or Apartado Box 2072, Balboa, Republic of Panama. Charles O. Handley, Jr., National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560. 3. The average age at which a bat is first captured is Jr (ctdt) exp - (m + c)t (cdt) exp - (m + c)t = 1 / (m + c) 4. This time interval l/(m + c) is also the average time interval between captures, as long as capturing a bat does not alter its behavior. Thus, the average time between first and last captures is [P(2) + 2P(3) + 3P(4) + 4P(5) (m + c)[P(2) + P(3) = 1/m, where 1/m is the average lifetime. The average time elapsed between first and last capture of bats captured more than once is an exact estimate of average lifetime, if the life table is exponential. Although the bats we caught more than once live longer than our records suggest, the bats we captured more than once tended to be the bats that live longer than average in the first place, and the errors cancel. We have calculated the average time between first and last captures by Bat Project participants of bats first marked by Morrison (1978b) and Bonaccorso (1979) between 1972 and 1974. Judging from the average interval between first and last capture by Bat Project personnel of the 33 females and eight males originally marked by either Morrison or Bonaccorso and captured more than once during the BCI Bat Project (Table 6-1), the average lifetime of such bats was 1.5 years. If we consider only females, the estimate of average lifetime is 1.6 years. These bats were two or more years old when first caught by Bat Project personnel, but if the life table is exponential, the average expectation of further life does not depend on the initial age of the bat. The average time between first and last capture for 301 bats caught five or more times during the Bat Project is 1.6 years. This is a surprisingly low figure for bats caught so many times, because in theory the more times a bat is caught, the greater the time between first and last capture. EXPONENTIAL LIFE TABLE.—Our basis for assuming an exponential life table is as follows. The numbers n(x) of the 77 78SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY TABLE 6-1.—Interval between first and last capture by the Bat Project of Artibeus jamaicensis first marked on BCI by Morrison and Bonaccorso. Category Both sexes Females only 0.25* 6 5 0.50 9 6 1.00 2 2 Number of years between first and last capture 1.50 9 8 2.00 4 2 2.50 4 3 3.00 6 6 5.00 1 1 Total 41 33 * Minimum assumed elapsed time for bats caught for the first time and recaptured for the last time in the same half year. TABLE 6-2.—Observed and expected frequency of capture of Artibeus jamaicensis marked on BCI by Morrison and Bonaccorso, 1972-1974, and by the Bat Project, 1976-1980. MORRISON-BONACCORSO BATS Observed (N = 146) Expected (69)(0.5274)'1 BAT PROJECT BATS Observed (N= 15728) Expected (89O7)(0.4337)"1 69 69 8907 8907 >2 39 36 3846 3863 >3 22 19 1683 1675 >4 11 10 734 727 Number of times caught >5 3 5 317 315 >6 1 3 146 137 >7 1 1.5 65 59 >8 0.8 21 26 >9 0.4 6 11 >10 0.2 2 5 >11 0.1 1 2 bats marked by Morrison and Bonaccorso and caught at least x times during the bat project form a geometric series where n(x) = 69(77/146)*"1 (see Table 6-2). Moreover, one of the most striking features of the Bat Project's capture records is that the number n(x) of bats captured at least x times forms a nearly perfect geometric series: /i(x + 1) = /i(l)A\ where n(l) = 8907 is the total number of bats caught at least once, and A is the total number of captures of bats caught at least twice, divided by the total number of captures, or 6821/15728 = 0.4337 (see Table 6-2). If the probability that a bat alive at time t is caught between times t and t + dt is cdt (where capture rate c is constant and dt is the length of a "short" time interval), then if all bats live the same lifetime (L), the probability, following the law of Poisson, that a bat may be caught n times in its life is [(cL)n/n!]exp(-cL) On the other hand, if the bats have an exponential life table with an average lifetime of 1/m, the chance P(x) that a bat will be captured x times in its life is P(x) = [m/(c + m)] [c/(c + m)]\ a geometric series in x. This agrees with our observation if A = c/(c + m). If these bats have an exponential life table, their age should not affect their prospects of further life. Is an average lifetime of 1.6 years reasonable for female A. jamaicensisl Notice that if the average lifetime 1/m of females is 1.6 years, then m = 1/1.6 = 0.625. Thus, survival rate per year is exp (-m) = 53.5%, and survival rate per half year is exp (-m/2) = 73.2%. Half of the females that live long enough to reproduce can bear a young a half year after they are born (they are products of the second reproductive episode); all can bear a young every half year thereafter. If all that can bear a young do so, and if half the young are female, the expected number of young females a newborn female bat will bear in her life is (l/4)(0.732) + (1/2X0.535) [1 + 0.732 + (0.732)2 + ...] = 0.183 + (0.2675)/[I-0.732] x 1.18. If, as our records show, one of every ten mature female bats in each breeding season fails to reproduce (see Section 3, Reproduction in a Captive Colony, and Section 4, Reproduction on BCI), the population will be very nearly in balance. Therefore, the average expectation of further life for these bats probably matches the average lifetime of all female A. jamaicensis rather closely and suggests that their life table is indeed exponential. NUMBER OF BATS.—For the sake of analysis, we shall make two more assumptions, even though they are not completely true (see Section 5, Survival and Relative Abundance). (1) Capturing a bat does not affect the probability of capturing it again later. (2) Bats are sampled from a pool in which all individuals are equally liable to capture. NUMBER 511 79 Then we may estimate the total number of bats available for capture during the Bat Project from the number of bats actually marked, divided by the chance a bat in the pool will be marked, where the latter is assumed equal to the proportion of marked bats that are recaptured at least once. In other words, the total number available during the project equals the number of bats marked, divided by the chance of marking a bat (equals the number of bats marked divided by the number of bats recaptured). The number of bats available for capture at any one time is the total number available over the duration of the project, multiplied by the average bat lifetime 1/m (including both sexes), and divided by the total duration of the sampling period (4 years). Another way of saying this is: the total number of bats available at a given time equals the total available during the project, times the mean lifetime of a bat, divided by the duration of the project. If Artibeus jamaicensis has an exponential life table, and if all bats "available for capture" are equally likely to be caught, then the total number of available bats during the four years of the sampling period is (8907)/(0.4337) = 20535. The number available at any one time is the total, times the lifetime of these bats (averaged for both sexes), divided by the four years' duration of the project, or [(20535)(1.5)]/4 = 7701. An equivalent way to calculate population size of these bats is to find the capture rate c by setting A = 0.4337 equal to c/(c + m), and to assume that the mortality rate m for both sexes concurrently is 0.667. We find c = 0.5106, which implies that, on the average, (0.5106)/12 or 4.166% of the bats are caught each month. As (15,728)/48, or 328, bats were caught per month, on the average, the total population of A. jamaicensis on BCI is 328/(0.04166) = 7873. VALIDITY OF ASSUMPTIONS.—How valid are the assumptions behind these calculations? If all available bats are equally liable to capture and if all bats have equal prospects of further lifetime, regardless of current age, then the chances of recapture of all bats marked within a given half year will be the same, regardless of age or sex. This is not true. Adults marked in the fall of 1976 and 1977 were at least as likely to be recaptured as the juveniles or subadults marked at that time, although adults marked later on were recaptured much less often than were juveniles or subadults (Table 6-3; and Section 5, Survival and Relative Abundance, Table 5-9). Few of the adults newly marked in the fall of 1979 and spring of 1980 were caught again, while recapture rates were more nearly normal for juveniles and subadults. Many of these newly marked adults were caught from peninsulas on the mainland surrounding BCI, where we netted far less often than on the island. In general, the ratio of the proportion of newly marked juveniles and subadults, to the TABLE 6-3.—Percentages, by sex and age class, of Artibeus jamaicensis marked in successive half years on BCI and subsequently recaptured; proportion of young to adult recaptures; and proportion of adults marked among total captures of adults (marks and recaptures) caught in each half year. Age class Percent recaptured Juveniles Subadults Adults Ratio of percentage recaptured Juveniles and Subadults / Adults Proportion of new marks among captured bats Number adults marked / Total captures of adults Percent recaptured Juveniles Subadults Adults Ratio of percentage recaptured Juveniles and Subadults / Adults Proportion of new marks among captured bats Number adults marked / Total captures of adults Fall 1976 0.3784 0.4185 0.5225 0.7820 1.0000 0.2955 0.3526 0.3841 0.8677 1.0000 Fall 1977 0.3814 0.4413 0.5055 0.8259 0.8044 0.4215 0.4600 0.5050 0.8754 0.7910 Spring 1978 0.7273 0.5404 0.3586 1.6372 0.5380 0.6154 0.5778 0.5364 1.0966 0.4314 Fall 1978 FEMALES 0.3855 0.4879 0.2524 1.8427 0.3787 MALES 0.3093 0.3459 0.3271 1.0151 0.4632 Spring 1979 0.5551 0.5315 0.3668 1.4918 0.5303 0.5368 0.4935 0.2934 1.7762 0.5000 Fall 1979 0.4743 0.3113 0.2194 1.8373 0.2476 0.4220 0.3163 0.2893 1.3404 0.3710 Spring 1980 0.3723 0.2105 0.1237 2.9450 0.3811 0.3109 0.2154 0.2055 1.4383 0.2219 80SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY TABLE 6-4.—Total captures of Artibeus jamaicensis on BCI in successive half years by sex and age class, by proportion of young to adult females, by proportion of new marks to total captures (marks and recaptures) among adult females. Age class Juveniles Subadults Adults Juveniles Subadults Adults All bats Nonadult bats Juveniles and Subadults / Adult females Adult female marks / Adult female captures Fall 1976 76 251 213 95 201 173 1009 623 2.925 0.836 Fall 1977 247 495 587 285 421 534 2569 1448 2.467 0.624 Total fall nonadults / Spring 1978 56 241 579 48 232 522 1678 577 0.997 0.496 Fall 1978 FEMALES 85 354 296 MALES 101 200 256 TOTALS 1292 740 Spring 1979 307 197 771 308 232 514 2329 1044 PROPORTIONS 2.500 0.348 Total fall adult females = ?306 ~ 1.354 0.477 2.950 Fall 1979 397 535 666 506 409 658 3171 1847 2.773 0.233 Spring 1980 401 49 533 367 132 318 1800 949 1.781 0.364 Fall 1980 612 543 544 625 364 430 3118 2144 3.941 0.248 proportion of newly marked adults recaptured subsequently, was higher the greater the proportion of adults already marked (Table 6-3). Our assumption that all available bats are equally liable to recapture is clearly wrong. It seems, rather, that as more bats were netted, we reached a point where most of the adult bats on BCI had been marked, so most bats available for marking on the island were juveniles and subadults. Thereafter, many of the bats marked as adults resided in localities where prospects of recapture were not great or they learned to avoid the nets. Not only are bats in certain places more liable to recapture, younger bats are more liable to recapture than older ones. The ratio of the total number of juveniles and subadults of both sexes to the number of adult females handled (counting each instance of capture) in the fall of 1976 was 2.925; combining all the fall catches of the sampling period together, it was 2.950 (Table 6-4). Because adult females can have no more than two young a year, adult females must be nearly twice as hard to catch as juveniles of either sex. A Refined Estimate Population size in these bats can be better estimated after two intermediate steps. (1) Calculating mortality rates for both sexes. (2) Calculating the total numbers of marked bats of each sex alive in successive half years (Dowdeswell et al., 1940; Fisher and Ford, 1947). The first step enables the second, because the total number of individuals M(t) alive in half-year t of the Bat Project that were marked earlier is the number m'{\) marked in the first half year (fall 1976), times the proportion p(t1) surviving to half-year t, plus the number m'(2) marked in the second half year (spring 1977) times the proportion p(t-2) of those surviving to half-year t, and so on. In summary: M(t) = m'(l)p(t1) + m'(2)p(t-2) + . . . + m'(tl)p(l). MORTALITY ESTIMATES.—If we assume that the numbers of A. jamaicensis on BCI do not change substantially from year to year, the simplest way to calculate the mortality of adults is to consider the recaptures of a sample of marked bats as long after they were first marked as possible. The bats marked by Morrison (1978b) and Bonaccorso (1979) provide a suitable sample. Between 1972 and 1974 they marked 1212 A. jamaicensis, of which about half were female. Of these females, 47 were caught by the Bat Project after 1 July 1977. Because 93 of the 178 adult female A. jamaicensis marked by the Bat Project in the fall of 1976 were recaptured after 1 July 1977, it seems reasonable to assume that 93/178 of the females marked by Morrison and Bonaccorso, and still living in the fall of 1976, were recaptured after 1 July 1977. If so, then 47(178/93), or 90, of Morrison and Bonaccorso's female bats were alive in the fall of 1976, implying a survival rate of (90/606)1/35, or 58%, a year, and an expectation of further life of 1.835 years. Similarly, 16 of their marked male bats were caught after 1 July 1977. As 58 of 151 adult males marked by the Bat Project in the NUMBER 51181 fall of 1976 were caught after 1 July 1977, roughly 16(151/58), or 42, of the bats marked by Morrison and Bonaccorso were alive during the fall of 1976, implying a survival rate of 46.7% a year. If the population were declining, these estimates would be too high, and vice versa. A second way to calculate survival rate is to consider the dates of first capture of the bats recaptured in a given half year. Let n(t, x) be the number of bats captured in half-year t that were first marked in half-year x, and let m'(x) be the total number of bats marked in half-year x (Tables 6-5 and 6-6). Then m'(x)p(t-x) is the total number of bats marked in half-year x surviving to half-year t. If a fraction c(t) of the bats alive in half-year t are captured then, n(t, x) = c(t)m'(x)p(t - x). If the bats have an exponential life table, then p(t-x) = exp[-m(t-x)/2], where exp (-m) is the average survival rate per year. Moreover, n(U = c(t) exp [-m(t - suggesting that we calculate m as the coefficient of regression of log [n(t, x)/m'(x)] on x. This method of estimating m suffers from the disadvantage that bats marked later in the project are more likely to be from infrequently sampled sites. Thus, bats first marked in half-year 7 are less likely to be recaptured in half-year 8 than already marked bats caught in half-year 7. On the other hand, this estimate of m does not depend on the stability of the bat population as a whole. Nonetheless, estimates (Table 6-7) of survival rates of female bats based on recaptures for half-year 7 and half-year 9, for which the correlation between the half-year x of marking and the logarithm of the proportion n(t, x)/m'(x) of bats marked then that were recaptured in half-year t is relatively close, agree with each other, and with the estimate based on bats marked by Morrison and Bonaccorso and recaptured by the Bat Project. The average of the former two estimates is 57.4%, compared with 57.99% from the Morrison and Bonaccorso bats. On the other hand, estimates of survival rates of male bats based on recaptures for half-years 7 and 9 average 37.2% a year, markedly lower than the 46.7% estimated from recaptures of male bats marked by Morrison and Bonaccorso. Do older males survive better? The 46.7% figure is based on a rather small sample of recaptures. Another piece of evidence is the recapture in the fall of 1981 of a male marked by Morrison and Bonaccorso. If 37.2% of the males survive each year, a male has one chance in 4000 of surviving the 8.5 years to the fall of 1981, so the chances are 1 in 7 that one of their 606 bats would still be living, and perhaps half that, had it been alive, we would have caught it then. That capture record suggests, but does not prove, that at least some older males survive rather better. Finally, if the population is stable, we may calculate adult mortality rates from the proportions n(x, t)/[m'(x) + n(x)] of bats marked in a given half-year t among the total number of bats—both marks, m\x), plus recaptures, n(x)—caught in later half years. If m'(x) + n(x) = c(x)N(x), where c(x) is the proportion of the N(x) bats alive in half-year x that were caught during that half year, while n(x, t) = c(x)m'(t) exp -m(x -1)/2, then n(x, t)/[m'(x) + n(x)] = [m'(t)/N(x)] exp -m(x -1)/2. If N is constant, the regression on x of the logarithm of this proportion is m/2. If the bat population is growing by a factor exp (r/2) per half year, then N(x) = N(t) exp r(x -1)/2, and our regression gives (m + r)/2. Mortality estimates for bats, both male and female, marked in the fall of 1976 and the spring of 1978, agree with each other, and are only slightly lower than those calculated from marking dates of bats recaptured in the fall of 1979 and the fall of 1980. Mortality estimates for bats marked in the fall of 1977 are much lower, but like those for bats marked in the spring of 1978, the fall 1977 estimates are based on strong correlations, illustrating the uncertainties in our calculations. If the chances of capturing a bat depend on its age class, then this estimate requires that the age composition of bats captured in successive half years be the same. If the chances of capturing a bat depend on whether it is already marked, then the proportion of recaptures among the bats handled should also be constant. Finally, if chances of recapture vary from place to place, netting effort should be distributed over the island in the same way during successive half years. Are our figures true mortalities? Although bats do occasionally move between BCI and the surrounding mainland, this exchange does not seem to be great (see Section 7, Movements). We do not think we are mistaking emigration for mortality. However, some bats do lose their necklaces. Adding together the intervals between time of first necklacing and of last recapture for each bat concerned, our project has monitored bats carrying Morrison and Bonaccorso wing bands for over fifty bat years. During this interval, five bats lost their necklaces, suggesting that the survival rate of necklaces is 90% a year. If so, the average survival rate of A. jamaicensis is 11 % higher than our regressions suggest, perhaps between 62% and 65% a year for adult females. Given the survival rate of juveniles relative to adults, and the breeding rate of adult females, the survival rate of adult females must be near 60% a year if the population is to be in balance. To show this, we make the following assumptions and observations. Half of the females (those that are born in the second reproductive episode) can bear a young a half year after birth; all can bear a young when they are a year old, and all can breed once each half year thereafter; half their offspring are female. Twenty-eight of the 74 juvenile females caught in the fall of 1976 were recaptured in successive half years, while 87 of the 178 adult females marked then were recaptured later (see Section 5, Survival and Relative Abundance, Table 5-9). Thus, a marked juvenile female had 0.7242 times the chance of being recaptured as an adult as did an adult female marked then. If in general, the chance of a juvenile female surviving through its 82SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY TABLE 6-5.—Numbers of female Arlibeus jamaicensis marked on BCI in successive half years by age class, by numbers n(x, y) of these recaptured in later half-years y, and by number n(x) of recaptures in half-years x of these bats marked in previous half years; and based on these values, proportions n(t, x)/m'(x) of these bats recaptured in selected half-years t among those marked in earlier half-years x, and proportions n(x, t)/[m'(x) + n(x)] of bats marked in selected half years among those caught in later half years. Females Marked Juveniles Subadults Adults Total [m'(x)] Recaptures from half-year y 1 3 4 5 6 7 8Total Recaptured/Marked n(9, x)/m'(x) «(8, x)/m'(x) «(7, x)/m'(x) n(x, l)/[m'(x) + n(x)] «(x. 3)/Im'(x) + n(x)] n(x, 4)/Im'(x) + «(x)] 1 Fall 1976 74 227 178 479 16/479 23/479 33/479 3 Fall 1977 236 358 366 960 91 91 41/960 47/960 88/960 91/1051 4 Spring 1978 66 198 290 554 Recaptures n(x, 68 181 249 38/554 31/554 67/554 68/803 181/803 5 Fall 1978 83 289 103 475 Half-year x 6 Spring 1979 272 141 368 781 7 Fall 1979 331 257 155 743 8 Spring 1980 368 19 194 583 y) in half-year x of bats first caught in half-year y 27 81 61 169 46/475 32/475 100/475 27/644 81/644 61/644 36 102 74 114 326 Proportions 76/781 75/781 183/781 36/1107 102/1107 74/1107 33 88 67 100 183 471 115/743 107/743 33/1214 88/1214 62/1214 23 47 31 32 75 107 315 157/583 23/898 47/898 21/898 9 Fall 1980 500 286 135 921 16 41 38 46 76 115 157 489 16/1410 41/1410 38/1410 first year is 0.7242 times the chance (p2) of an adult female surviving a full year, and if a female has a 1/4 chance of bearing a female young when it is a year old, and a 1/2 chance of doing so each half year thereafter, then the number (Ro) of female young a newborn female A. jamaicensis can expect to bear during her life is Ro = 0.7242 = 0.1810p + 0.362/^/(1 -p). If survival rate p2 per year is 0.5625 (so p = 0.75), then Ro = 0.95; if p2 = 0.6 (sop = 0.7746), then Ro = 1.10; and ifp2 = 0.64 (sop = 0.8), then Ro = 1.30. In fact, not all females breed in each breeding season. In six reproductive episodes in our captive colony wild-caught adult females were pregnant in 66 of 72 chances (91.7%), multiparous captive-born females were pregnant in 12 of 14 chances (85.7%), and primiparous captive-bom females were pregnant in 14 of 20 chances (70%) (see Section 3, Reproduction in a Captive Colony, Tables 3-4 and 3-5). On BCI in the month of April, when all females should be reproductive, we invariably caught a few nonreproductive adult females and some subadults who failed to reproduce on time (undoubtedly 9-month-old young of the previous July-August birth-group). For example, in April 1978,9 of 102 adult females caught were nonreproductive, and we caught 22 subadult females; in April 1979, 7 of 47 newly marked adult females were nonreproductive. We can make a minimum correction for this by multiplying our values of Ro for different survival rates by 0.89, the percentage (133/149) of adult females caught in 1978 and NUMBER 51183 TABLE 6-6.—Numbers of male Artibeus jamaicensis marked on BCI in successive half years by age class, by numbers n(x, y) of these recaptured in later half-years y, and by number n(x) of recaptures in half-years x of these bats marked in previous half years; and based on these values, proportions «(t, x)/m'(x) of these bats recaptured in selected half-years t among those marked in earlier half-years x, and proportions «(x, t)/[m'(x) + n(x)] of bats marked in selected half years among those caught in later half years. Males Marked Juveniles Subadults Adults Total [m'(x)] Recaptures from half-year y 1 3 4 5 6 7 8Total Recaptured/Marked «(9, x)/m'(x) «(8, x)/m'(x) n(7, x)/m'(x) «(x. l)/[m'(x)+n(x)] n(x,3)/[m'(x)+«(x)] fl(x,4)/[m'(x)+n(x)] 1 Fall 1976 88 173 151 412 3/412 9/412 9/412 3 Fall 1977 261 300 299 860 79 79 10/860 18/860 56/860 79/939 4 Spring 1978 52 135 220 407 5 Fall 1978 97 159 107 363 Half-year x 6 Spring 1979 272 154 242 668 7 Fall 1979 410 196 197 803 8 Spring 1980 341 65 73 479 Recaptures «(x, y) in half-year x of bats first caught in half-year y 55 235 290 13/407 22/407 44/407 55/697 235/697 14 50 60 124 15/363 31/363 61/363 14/487 50/487 60/487 19 65 57 101 242 Proportions 46/668 50/668 164/668 19/910 65/910 57/910 9 56 44 61 164 334 90/803 126/803 9/1137 56/1137 44/1137 9 18 22 31 50 126 256 114/479 9/735 18/735 22/735 9 Fall 1980 496 198 101 795 3 10 13 15 46 90 114 291 3/1086 10/1086 13/1086 1979 that were actually breeding. There may be other errors in our calculations. On the one hand, spring-born juveniles should survive better than fall ones; on the other, our figures for juvenile mortality do not include deaths before the juveniles are old enough to fly into our nets. All in all, we believe the survival rate of adult female Artibeus jamaicensis is between 60% and 64% a year. ESTIMATE OF SURVIVING MARKED BATS IN SUCCESSIVE HALF YEARS.—Let M^t) represent the number of marked adult female bats alive at the beginning of half-year t, and Mrf(t) be the number of marked subadult female bats also alive then. In addition, p represents the survival rate of adult females per half year. Eighty-three of the 227 subadult bats marked in the fall of 1976 were recaptured in succeeding half years, while 87 of 178 adults marked then were recaptured later (Section 5, Survival and Relative Abundance, Table 5-9), suggesting that a subadult female was 0.8010 times as likely as an adult to survive her next half year. If this is true in any half year, then a juvenile female is (0.7242)/(0.8010) = 0.9041 times as likely as an adult female to survive her next half year. If we make the convention that a fraction -ip of the adults, a fraction 0.8010 Jp of the subadults, and a fraction 0.9041 <p of the juveniles marked in a given half year survive to the beginning of the next half year, and that juveniles become subadults, and subadults become adults, at that time, then we may assume the number Msf (t) of marked subadult females at the beginning of a half year to be 0.9041 <p m^ (t - 1), where mjf (t - 1) is the number of juvenile females marked in half-year t - 1. The marked adult females alive at the beginning of half-year 90SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY 4546 47 48 49 50 11ii1 0 12 3 4 5 kilometers FIGURE 7-1.—Distribution of locality groups (aggregations of localities) where bats were captured on BCI and adjacent mainland (from Table 7-1). The locality groups are as follows: (1) Lutz; (2) Barbour-Hood; (3) Shannon-AMNH; (4) Shannon-Balboa; (5) Lake-Wheeler; (6) Barbour Stream; (7) Miller Ridge; (8) Fuertes; (9) Conrad; (10) Plateau; (11) Drayton End; (12) Armour End; (13) Zetek 21; (14) StandQey Ridge; (15) Standley End; (16) Orchid Island; (17) Gross Point; (18) Chapman; (19) Harvard; (20) Mona Grita; (21) Gigante; (22) Frijoles Rd.; (23) Bohio; (24) Buena Vistt; (25) Pena Blanca. NUMBER 51191 46 45 I • •• •! .... I 2 3 kilometers FIGURE 7-2.—Percentage of other-captures in the total records (total individuals plus other-captures of them) in capture suites of Artibeus jamaicensis at locality groups on BCI and the adjacent mainland, 1976-1980. Data are from Table 7-2. Isolines arbitrarily enclose locality groups with similar percentages of other-captures. Marginal numbers identify kilometer squares (see Appendix, Methods of Capturing and Marking Tropical Bats). 92SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY I a. oo H B-f •*! § 5n -r i .s <- •? oo eg « T! < - n H S £ r so § -C 1 V Other _ 3 CO <s cs1 o\ oo VO vo — CO js o O\ oo c— vn CO ts captures 1 1 . NO 12 CO o O ts o - 2 CO CO O oo o § oo 230 ° s 42 106 I 5309 6 . Lutz "™ CO <s o o o o o o CS o -. CS — — 2 cs - ° 19 17 640 1 ] . B arbour CO - o o o o o - ts ,3. — CO vn vO - cs 24 19 ! 730 1 < . Shannon Vn NO o o o o o o o - o cs o CO — o NO o cs o i 165 i •a pa 6 | T NO o o o o o o o - o CO CO o o ° cs CO ° 156 ler J Lake-W - o o o o o o o - o ,^ o o ° cs « (S ° cs 1 234 1 , B arbour ON 00 - - oo o o o CO - 00 vn VO ON o - 169 CO NO NO CS 29 18 j 2499 2 u idg , Miller R •—< o o o o o o n - cs — cs CO NO cs _ NO 139 cs - NO | 1189 n CO cs o CO o o o o o ° ts NO ^- cs ° CO _ oo - - o I 578 . Conrad ^. CS - o o o o o oo —« cs vn oo CO CO 8! CO c— CO 50 13 [ 1377 1 Plateau CO CN o o o o o o o ° o Q o cs o ts o o - 138 -o W Drayton — o o in o o - o o _ cs 2 2 o o CO CS 456 I Armourcs . CO o o o o o o o CO TICS ° cs NO o o o 146 Zetek 21 to OO o •>r o o o o o - cs _ vn CO - ° — NO ON vn - CO cs cs 786d«e _u - ts o o o o o CO r~ CO o ° cs ^^ o CO o cs 649 -a t§ Slandley vn NO NO o vn cs o o o o vn vn - _ ts cs ° o ^H oo - o TT ; 249 _, J Orchid I 16. oo O O o o o o o o a o o - ° o CS o - o 145 .| Gross P( ro o cs o o o NO — o o - CO >n cs - o - cs 388 Chapmai 18. ro o - o o o o TT o o o o o ° o o CO o o o cs 125 Harvard 19. o o o o o cs o o o © o o - ° o o ° o o o o •1 Mona Gi20. CO ,—« o o o o o o o o o o o o o o o ° o o o o 140 Gigante 21. ts o o o cs o o o o o o o o o o o o ° o o o ° o 156 •3 J Friioles ] 22. NO CS ON NO ON CS o o o o o cs - vn — o vn o CO ^. o o © o 746 Bohio23. ON CO O CO vn o o o - o o o o o o o « o o o ° o 143 isti Buena V 24. CO vn O o o o o o o o o o o o - o o o 138 5 Peria Bla 25. oo oo CO _ CS ts CO CO .01 t~ ••* o >n ts - co ts cs r906 vn s oo cs 24 216 17,322* 16: Ibtal NUMBER 51193 TABLE 7-2.—Tabulation of total records, total other-captures, and percentage of other-captures (see text for explanation) of the total number of Artibeus jamaicensis in each locality group on BCI and adjacent mainland, 1975-1980. See Figure 7-1 for locations of locality groups. Localities 1. Lutz 2. Barbour-Hood 3. Shannon-AMNH 4. Shannon-Balboa 5. Lake-Wheeler 6. Barbour Stream 7. Miller Ridge 8. Fuertes 9. Conrad 10. Plateau 11. Drayton End 12. Armour End 13. Zetek 21 14. Standley Ridge 15. Standley End 16. Orchid Island 17. Gross Point 18. Chapman 19. Harvard 20. Mona Grita Point 21. Gigante 22. Frijoles Road 23. Bohio 24. Buena Vista 25. Pena Blanca (A) Total individuals 5309 640 730 165 156 234 2499 1189 578 1377 138 456 146 786 649 249 145 388 125 40 140 156 746 143 138 (B) Total other-captures 1567 233 305 65 67 57 899 421 239 524 23 92 37 248 142 66 48 111 30 4 13 2 126 39 33 (B/A) Percent other-captures 30 36 42 39 43 24 36 35 41 38 17 20 25 32 22 26 33 29 24 10 9 1 17 27 24 Mona Grita Point, 10%; Gigante, 9%; and Frijoles Road, 1%). These values show that more of the population was marked and a larger portion of the area frequented was sampled at the central localities. Bats moving in any direction from the central localities were likely to encounter other capture stations. Bats using the more isolated lakeshore localities probably foraged extensively on the mainland; hence, we sampled smaller fractions of their populations and the areas that they frequented. The effects of sampling small fractions of populations and their foraging areas are best seen on the mainland where stations were few and (except for Bohio) capture effort was low. We often netted at Bohio and it ranked sixth in numbers of A. jamaicensis caught. However, Bohio was nineteenth in the number of other-captures (17%, Table 7-2) and 83% of its A. jamaicensis were caught only once. The lowest other-capture rates (Table 7-2) on BCI were at Drayton End (17%) and Armour End (20%). Among the most isolated BCI stations, and located in high, old forest containing few fruit trees, each of these two sites is over a kilometer from the next nearest station. Bats at the southern perimeter of BCI may forage on the adjacent mainland where the forest is younger and contains many fig trees. Among the few recaptures at Gigante and Mona Grita Point on the mainland were Armour End and Drayton End bats. At the outset of the Bat Project we gridded BCI and planned to randomly net that grid. This plan was based on the perception that good (representative) catches night after night would result if we moved at random through our grid, using well-sited and well-set nets at choice netting stations. Our presumption was naive; on some nights we caught almost nothing, whereas on other nights we were overwhelmed with bats. It became obvious that the nightly distribution of fruit bats was not random, but coincided with the presence of ripe fruit, which tends to be unevenly distributed in time and space. Nets set at certain fruiting trees caught bats, nets placed elsewhere often did not, and a locality might have an abundance of bats on one date and few on another. When there was only one preferred tree with ripe fruit on BCI, clearly that was the place to be for bats. There are many factors that influence the distribution and movements of fruit bats on BCI: distribution and abundance of preferred foods; rain, wind, moon phase, and cloud cover; the size, topography, and other limitations of insularity. These factors tended to overwhelm our efforts to set up meaningful measures of capture effort involving time and netting conditions. 94SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY TABLE 7-3.—Percentages of other-captures of Artibeus jamaicensis for selected locality groups on BCI weighted by the number of total captures of A. jamaicensis at each of the locality groups where the other-captures were recorded. Adjusted percentages were derived by dividing the number of other-captures (data from Table 7-1) by the number of total captures for that locality group and reducing that value to a percentage of its column total. The name of each locality group analyzed is followed by the total number of other-captures in parentheses. Locality group 1. Lutz 2. Barbour-Hood 3. Shannon-AMNH 4. Shannon-Balboa 5. Lake-Wheeler 6. B arbour Stream 7. Miller Ridge 8. Fuertes 9. Conrad 10. Plateau 11. Drayton-End 12. Armour-End 13. Zetek 21 14. Standley Ridge 15. Standley End 16. Orchid Island 17. Gross Point 18. Chapman 19. Harvard 20. Mona Grita PL 21. Gigante 22. Frijoles Road 23. Bohio 24. Buena Vista 25. Pena Blanca Total Total captures 5309 640 730 165 156 234 2499 1189 578 1377 138 456 146 786 649 249 145 388 125 40 140 156 746 143 138 17,322* ShannonAMNH (305) % 6.35 8.06 10.04 11.51 6.95 1.17 2.71 1.60 2.82 10.26 9.82 1.79 1.87 1.38 0.84 3.28 1.87 7.00 8.68 1.98 Chapman (110) % 3.51 18.86 8.28 3.34 2.36 0.44 2.30 2.85 1.59 3.95 1.12 0.71 0.82 2.19 19.79 26.32 1.48 Other-captures Lutz (1567) % 7.09 9.71 8.60 6.75 7.52 6.77 5.39 3.50 4.87 6.09 2.55 1.03 4.02 2.46 1.17 2.36 4.45 4.99 4.69 1.47 0.84 0.79 1.64 1.27 Miller Ridge (899) % 6.05 2.74 3.35 3.55 10.02 1.25 6.60 11.50 14.04 7.02 0.70 3.00 4.01 6.34 2.70 3.13 7.41 0.51 2.34 2.35 0.68 0.70 Standley End (248) % 1.25 1.49 1.29 9.20 3.45 4.46 1.68 1.72 9.82 16.48 28.80 1.92 1.25 7.04 6.71 3.45 * 15,736 band numbers. Also, we had to adjust (prorate) capture rates because the capture effort differed among localities. We tried "recaptures per net-night" and "recaptures per net-hour," but these gave anomalous results, probably because netting effort commonly was inversely related to capture success. In other words, the poorer our catch, the more nets we maintained and the longer we worked them. After realizing that capture effort measured in units of time was unreliable, we discovered that the numbers of A. jamaicensis caught per locality provided us the simplest and most useful measure of capture effort. A. jamaicensis regularly made up about two-thirds of our catch and was usually the commonest bat anywhere we netted. Our accumulated 17,322 mark and recapture records of this species varied between localities from a maximum of 5309 bats in the Lutz Watershed to a minimum of 40 at Mona Grita Point (Table 7-3). Using capture data from A. jamaicensis (Table 7-1), we derived weighted percentages of other-captures by dividing the number of other-captures at a locality group by the total number of A. jamaicensis recorded at that same locality group. Each value was then converted to a percentage of its column total to produce the A. jamaicensis-'weighted percentages plotted in Figures 7-3 through 7-8. The efficacy of these methods for illustrating dispersion of other-capture records of A. jamaicensis around a sample center can be seen by comparing resulting percentages with the distance between each capture locality and the center (Table 7-4). When graphed (Figure 7-9), the data points approximate a line declining with increasing distance to zero indicating that the frequency of other-captures was proportional to distance from the sample center. Another way we examined the correspondence of frequency of other-captures with distance from the sample center was by plotting the adjusted values on a map containing concentric circles with radii of 1, 2, and 3 km from the sample center (Figure 7-4). The Shannon-AMNH data fit rather well NUMBER 51195 45 46 47 48 49 50 51 52 53 54 55 i.••11..,, 12 3 4 5 kilometers FIGURE 7-3.—Other-captures of individual Artibeus jamaicensis recorded at least once in the Shannon-AMNH locality group on BCI, 1976-1980 (expressed as a percentage of total other-captures of Shannon-AMNH bats and weighted by A. jamaicensis capture means; data are from Table 7-3; Shannon-AMNH is outlined with a hexagon). Isolines arbitrarily enclose locality groups with similar percentages of other-captures. Innermost line encloses core area. Marginal numbers identify kilometer squares (see Appendix, Methods of Capturing and Marking Tropical Bats). 96SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY 47 46 4545 46 47 48 49 50 51 52 53 54 • * • • i 12 3 4 5 kilometers FIGURE 7-4.—Other-captures of individual Arlibeus jamaicensis recorded at least once in the Shannon-AMNH locality group on BCI, 1976-1980 (expressed as a percentage of total other-captures of Shannon-AMNH bats and weighted by A. jamaicensis capture means). Circles at oneand two-kilometer intervals encompass 29% and 85%, respectively, of all other-captures. Shannon-AMNH is outlined with a hexagon. NUMBER 51191 47 46 45 1 .,, 2 3 kilometers FIGURE 7-5.—Other-captures of individual Artibeus jamaicensis recorded at least once in the Chapman locality group on BCI, 1976-1980 (expressed as a percentage of total other-captures of Chapman bats, and weighted by A. jamaicensis capture means). Chapman is outlined with a hexagon. Isolines arbitrarily enclose locality groups with similar percentages of other-captures. Innermost line encloses core area. Marginal numbers identify kilometer squares (see Appendix, Methods of Capturing and Marking Tropical Bats). 98SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY 55 54 »••••' 2 3 kilometers FIGURE 7-6.—Other-captures of individual Artibeus jamalcensis recorded at least once in the Lutz locality group on BCI, 1976-1980 (expressed as a percentage of total other-captures of Lutz bats, and weighted by A. jamaicensis capture means). Lutz is outlined with a hexagon. Isolines arbitrarily enclose locality groups with similar percentages of other-captures. Innermost line encloses core area. Marginal numbers identify kilometer squares (see Appendix, Methods of Capturing and Marking Tropical Bats). 99 45 1.... 2 3 kilometers FIGURE 7-7.—Other-captures of individual Artibeus jamalcensis recorded at least once in the Miller Ridge locality group on BCI, 1976-1980 (expressed as a percentage of total other-captures of Miller Ridge bats, and weighted by A jamaicensis capture means). Miller Ridge is outlined with a hexagon. Isolines arbitrarily enclose locality groups with similar percentages of other-captures. Innermost line encloses core area. Marginal numbers identify kilometer squares (see Appendix, Methods of Capturing and Marking Tropical Bats). 106SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY 49 50 51 4546 47 »••"' 0 1 2 3 4 5 kilometers FIGURE 7-13.—Overlap of core areas in West-central and peripheral portions of BCI. A = Armour End; BO = Bohio; B V = Buena Vista; CH = Chapman; C = Conrad; D = Drayton End; F = Fuertes; MR = Miller Ridge; PB = Pena Blanca. NUMBER 511107 TABLE 7-5.—Core areas on BCI and vicinity in which the central locality had a lower percentage of other-captures (see text for explanation) of Artibeus jamaicensis (captured at least once there) than some other localities in the core. Each entry includes the central locality (CAPITALIZED), percentage of other-captures* there, rank among all localities in terms of number of individual A. jamaicensis caught, and percentages of other-captures* at other localities within the core. * = A. jamaicensis-v/eighled. Locality LUTZ Barbour-Hood Shannon-AMNH Lake-Wheeler Barbour Stream Shannon-Balboa MILLER RIDGE Fuertes Lake-Wheeler Conrad Gross Point Standley Ridge Plateau PLATEAU Shannon-Balboa Shannon-AMNH Conrad Lake-Wheeler Miller Ridge Armour-End FUERTES Miller Ridge Conrad BOHIO Pena Blanca Buena Vista SHANNON-AMNH Shannon-Balboa Plateau Drayton-End BARBOUR-HOOD Chapman Harvard Shannon-AMNH Barbour Stream CONRAD Plateau Miller Ridge Fuertes Standley Ridge ARMOUR-END Mona Grita Point Shannon-Balboa CHAPMAN Harvard Barbour-Hood ORCHID ISLAND Gross Point Buena Vista Shannon -Balboa Zetek21 SHANNON-BALBOA Orchid Island GROSS POINT Orchid Island Total captures 5309 2499 1377 1189 746 730 640 578 456 388 249 165 145 Percent other-captures 7 10 9 8 7 7 7 11 10 14 7 6 7 6 14 12 11 8 7 6 13 14 11 16 27 18 10 11 10 10 11 24 14 11 10 10 15 13 12 11 13 15 12 20 26 19 9 44 15 10 9 14 19 t 70 Rank 1 2 3 4 6 7 9 10 11 12 13 15 19 TABLE 7-6.—Core areas on BCI and nearby mainland in which the central locality had a higher percentage of other-captures (see text for explantion) of Artibeus jamaicensis (captured at least once there) than other localities in the core. Each entry includes the central locality (CAPITALIZED), percentage of other-captures* there, rank among all localities in terms of number of individual A. jamaicensis caught, and percentages of other-captures* at other localities within the core. * = A. jamaicensis-weighted. Locality STANDLEY RIDGE Standley-End Conrad STANDLEY-END Standley Ridge Zetek 21 BARBOUR STREAM Lake-Wheeler LAKE-WHEELER Barbour Stream Miller Ridge Gross Point ZETEK 21 BUENA VISTA DRAYTON-END PENA BLANCA HARVARD Chapman Total captures 786 649 234 156 146 143 138 138 125 Percent other-captures 23 14 10 29 16 10 36 18 26 12 10 10 42 82 31 84 49 24 Rank 5 8 14 16 18 20 22 23 24 t No other-captures recorded. locality within each core (Tables 7-5 and 7-6). Percentages of other-captures at the central localities ranged from 6% to 84%. The lower percentages at Lutz, Miller Ridge, and Plateau must reflect the numbers of captures, which were highest at these three localities. If other-captures at a central locality are predominantly local residents, their representation out of all bats caught at that locality will be inversely related to the total number caught. The greater the number of individual bats involved, the greater the likelihood that other-captures will be widely dispersed, thus also reducing the proportion of other-captures at a central locality. The high percentages of other-captures at Buena Vista and Pefia Blanca simply demonstrate fidelity to those sites, which are day roosts in channel markers. Whenever we sampled these roosts, most of the bats caught were ones taken there before. The low percentages of these bats netted away from their day roosts suggest that they foraged where we did not net, probably on the mainland. A comparison of the Bohio core area (Figure 7-13) and the Standley End area of frequent use (Figure 7-8) shows that while these are places where Buena Vista and Pefia Blanca bats foraged, they are probably not their main feeding areas. There are extensive areas of the mainland near these roosts where we have never netted (Figure 7-13). There is a dichotomy of core areas into those whose central localities have both lower percentages and fewer other-captures than other localities in the core (Table 7-5) and those whose central localities have higher percentages and more othercaptures than do neighboring localities (Table 7-6). This division may be correlated with the number of captures. 108SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY 60 55 50 45 40 IZ 35 LJLJ O 30 cc m 25 20 15 10 5 Number of individuals in each capture category 1 5061 57 2 2163 24 3 949 11 4 417 5 5 171 2 6 81 1 7 44 0.5 8 15 0.2 Number of records in each 1 5061 32 2 4326 28 3 2847 18 4 1668 11 5 855 5 6 486 3 7 308 2 8 120 1 9 4 0.04 10 1 0.01 capture 9 36 0.2 10 10 0.01 11 1 0.01 Total 8907 category 11 11 0.01 15728 5061 . 8907 5061 15728 I 0.57 I 0.32 Individuals 3 4 5 6 7 8 CAPTURE CATEGORY 10 11 FIGURE 7-14.—Frequency of capture of Artibeus jamaicensis on BCI. Number of individuals caught 1, 2, 3, 4, and more times, and number of records in suites of individual bats in each capture category. Standley Ridge and Standley End (Table 7-6) are exceptions in that comparatively high numbers of A. jamaicensis were caught at each (they ranked fifth and eighth, respectively), yet their numbers of other-captures are also relatively high (see discussion of Fidelity, below, in this section). The distribution of fig trees probably exerts an influence on the relative numbers of other-captures at any locality. The first seven localities in Table 7-5 (Lutz through Barbour-Hood) have an abundance of fig trees. Combined, Lutz and Miller Ridge have over 200 fig trees, mostly Ficus insipida, the favored fruit of A. jamaicensis. Chapman, Standley Ridge, and Standley End have a fair number of fig trees, but not many F. insipida. The other localities in Tables 7-5 and 7-6 have few fig trees other than stranglcr figs (mainly F. obtusifolia) whose fruits arc not preferred by A. jamaicensis. If the number of A. jamaicensis harems is limited by the availability of suitable tree holes, and if trees with suitable holes are evenly distributed over BCI then there must be many harem roosts that are not near patches of fig trees. At these places we should catch the bats of resident harems and few others, thus accounting for high percentages of other-captures characterizing those sites. The bats of day roosts not located near fig trees must routinely travel to other parts of the island to feed. However, bats with day roosts near fig trees also commute to other localities to forage whenever the trees near their roosts lack ripe fruit. Nevertheless, these bats need to travel less than bats residing in tree holes in fig-poor places. Turner (1975) showed that Desmodus rotundus routinely changed roost sites to remain close to a spacially shifting resource. We assume that either roosting sites are limited on BCI or that A. jamaicensis have high fidelity to individual day roosts because we found no evidence of routine shifts even among bachelor males. Bachelor males roost in foliage and not in tree holes. Although they might be expected to concentrate in fig-rich areas and to shift concentrations to follow fruit availability, they apparently do not, judging by the unusually high ratio of males consistently found at Standley End. NUMBER 511 50 # 40Q a 30 H 109 o CO < CD 2010 019 12 6% 2029 29 15% 3039 36 19% 4049 4 2% 5059 80 42% 6069 9 5% 7079 3 2% 8099 0 100 17 9% I I I 1 I I I I 0-19 20-29 30-39 40-49 50-59 60-69 70-79 80-99 CAPTURE CATEGORIES (%) 100 FIGURE 7-15.—Fidelity of Artibeus jamaicensis to the Chapman locality group on BCI. Frequency of capture of individual bats with records of multiple captures that were caught at least once at Chapman are graphed as percentages of records in various fractional capture categories ('/6 or less of records = 0%-19%, Vs-V* = 20%-29%, V3 = 3O%-39%, 2/s = 40%-49%, lh = 50%-59%, Vs-Vi = 60%-69%, 3A = 70%-79%. 4/s5/6 = 80%-99%, all = 100%). Row A in the table lists fractional capture categories (as percentages), row B lists number of individual bats whose records at Chapman fall into each category (n = 190), and in row C the numbers are converted to percentages. See also Table 7-7. Fidelity We examined fidelity to a particular locality group by analyzing suites of multiple capture records of individual bats. For each locality group we compiled records of all A. jamaicensis having multiple captures that were taken at least once in that particular locality group between October 1976 and October 1980. For each bat, we tabulated total captures, the number of captures in that particular locality group, and percentage of its total records in that locality group. From that tabulation we grouped those having multiple captures according to the percentage of their occurrence at that locality group among total records (e.g., Chapman, Figure 7-4). The distribution of records among fractional categories is influenced by the number of records per individual suite. In all, 3846 A. jamaicensis were recaptured (Figure 7-14). Most (57%) were recaptured only once (i.e., had two captures), 24% were recaptured twice (three captures), 11% recaptured three times (four captures), and 8% recaptured four or more times. Thus, peaks would be expected (Figure 7-15) at 50% and 100% for two-record suites; 33%, 66%, and 100% for three records; and 25%, 50%, 75%, and 100% for four records, and so forth. Because most recaptured bats were caught again only once, 50% and 100% peaks per locality should occur most often. These data indicate fidelity to a particular locality group. High fidelity means a high frequency of records in the 60%-100% interval, which represents 3/s, 2/3,3A, 4/s, or more captures in the same area. The hiqher the frequency of same-site captures, the higher the fidelity. Low fidelity is revealed by a higher frequency of records in the 0%-40% interval (2/s, lfr, !/4, Vs, and !/6 of captures or less) for that locality group. A simplified summary (Table 7-7) of localitygroup fidelity, like that in Figure 7-15, shows three basic patterns among our data. Pattern I (e.g., Armour End; Figure 7-16, Table 7-7) covers most of the locality groups. There is a strong similarity among 110SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY © — Ov O tn l^ oo vd -^ co — oo >n so so rf r~ ts oo >/•> oo « N f /I - -^~* SO —* CO ^^ CO ^^— — co rco I g NUMBER 511 6050 111 o cc 0. < o CO < 00 403020100 J "O = Chapman •• • Conrad A - Fuertes •A ' Orchid Island 0-19 20-29 30-39 40-49 50-59 60-69 70-79 80-99 CAPTURE CATEGORIES (%) FIGURE 7-16.—Fidelity of Artibeus jamaicensis to Pattern I locality groups on and near BCI: Chapman, Conrad, Fuertes, and Orchid Island. See Figure 7-15 for further explanation. 100 the capture-category tabulations for locality groups with this pattern in spite of their diverse locations and the disparity in numbers in each suite of multiple capture records (from 51 to 1321). These locality groups share a low degree of fidelity (<19% in the 60%-100% column) and a high frequency of single captures (82%-99% in the 0%-40% and 0%-50% columns). Pattern II (Lutz, Bohio, and Standley End; Figure 7-17, Table 7-7) comprises three seemingly unrelated locality groups whose records of fidelity are remarkably similar. All three show higher fidelity, compared with the Pattern I locality groups (30%-32% in the 60%-l00% column), and fewer single records (68%-70% in the 0%-40% and 0%-50% columns). Their records may be similar for different reasons. We sampled little of the probable foraging range of the Bohio bats, so they had little risk of capture except at Bohio, thereby appearing to have high fidelity to that location. In contrast, the bats of Standley End may actually be sedentary because they were recaptured mostly at Standley End or nearby (see Figure 7-8 and discussion of Core Areas earlier in this section). Or, if we sampled only part of their foraging areas, the situation at Standley End may be similar to that at Bohio. We are confident that we have sampled the full extent of the foraging ranges of Lutz bats. High fidelity to Lutz may reflect a larger local resident population, or may result from catching the same individuals repeatedly because of the attraction of the fig grove in Lutz Ravine. These fig trees may adequately support local populations of A. jamaicensis and periodically attract bats from afar when ripe figs are scarce elsewhere. If this is true then some of the bats showing high fidelity are actually opportunistic transient foragers rather than Lutz residents. Pattern III (Buena Vista and Pefia Blanca; Figure 7-18, Table 7-7) shows the highest fidelity. The obvious explanation for this pattern is that we sampled only a fraction of the foraging ranges of these bats, and because we captured them mostly at their day roosts. We netted bats only at night at all other locality groups and, evidently, we did not net where the bats of Buena Vista and Pefia Blanca most often foraged. The degree of fidelity within patterns appears to be related to netting effort. On this basis we sorted Pattern I locality groups 112SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY 50-| 40Q c 30 -\ O CD 20100J O = B oh i o • = Standley End • = Lutz I | •••••• 0-19 20-29 30-39 40-49 50-59 60-69 70-79 80-99 CAPTURE CATEGORIES (%) FIGURE 7-17.—Fidelity of Artibeus jamaicensis to Pattern II locality groups on and near BCI: Bohio, Lutz, and Standley End. See Figure 7-15 for further explanation. 100 into two subpatterns (Table 7-8). One with 19 or more netting-nights per locality (subpattern la); the other with fewer (subpattern Ib). With few exceptions (low fidelity at Conrad and Lake-Wheeler; unusually high fidelity at Drayton End), fidelity and netting effort within patterns are directly and positively correlated, suggesting that the more often a locality group is netted the more likely the bats using that locality group will be recaptured. Nevertheless, although netting effort differed greatly between subpatterns la and Ib, fidelity was similar. Therefore, variation in fidelity between patterns is independent of netting effort. Subpattern la localities were netted from 20 to 52 nights and fidelity ranged from 7% to 18% (Table 7-8). Subunit Ib localities were netted only 6 to 17 nights apiece, yet the fidelity range was similar (8%-16%). Fidelity was surprisingly similar (30%-32%) among pattern II localities even though netting effort ranged from 14 nights at Bohio to 197 at Lutz (Table 7-7). However, when comparing fidelity on the basis of similar netting effort between patterns I and II the results are markedly different. For example, 14 netting nights demonstrated 6% and 14% fidelity in Pattern I, and 30% fidelity in Pattern II. Locality groups with 24 and 26 netting nights in Pattern I showed 9% and 13% fidelity; whereas fidelity was more than double (31%) after similar effort at Standley End (25 netting nights) in Pattern II. If all bats having multiple captures that were recaptured only once in a particular locality group are excluded, we should eliminate most that were probably based elsewhere. Following that logic, we considered A. jamaicensis captured two or more times in a locality group more likely to be local residents. The proportion of bats with two or more captures in a locality group to the total multiple-capture cohort from that locality group varied from 0.9% to 71.9%; although they still sorted into the same patterns I, II, and III (Table 7-7). Few individuals caught two or more times confirms low fidelity at locality groups such as Conrad (11%) and Chapman (18%). The two-plus capture fraction at Conrad may contain few local bats, judging from the lower proportion (54%) of those bats in the 60%-100% column. However, the high percentage (78%) in the 60%- 100% fractional capture category at Chapman indicates that the two-plus fraction, although small, may represent mostly local bats. The high proportion (78%) of bats at Pefla Blanca that were caught there at least twice, and the high percentage (92%) of those in the 60%-100% bracket, suggest that this exercise is a valid means of estimating fidelity to a locality because high fidelity is to be expected at a harem day roost. NUMBER 511 50 w 40Q LU cc 30 Q_ O 20 H 113 CO CD100J Buenavista Light Pena Blanca Light I 1 1 1 " 1 1 •—i 1 0-19 20-29 30-39 40-49 50-59 60-69 70-79 80-99 CAPTURE CATEGORIES (%) FIGURE 7-18.—Fidelity of Artibeus jamaicensis to Pattern III locality groups near BCI: Buena Vista and Pena Blanca. See Figure 7-15 for further explanation. 100 We found that 102 (37%) of the 275 A. jamaicensis caught at Standley End had been caught there at least twice (Table 7-7). Of these, 83% fall into the 60%-100% recapture category for the site. This is good evidence that a high proportion of the Standley End bats were local residents, a conclusion further supported when we examine their records in greater detail. Only two (7%) of the two-plus females were caught as many as two times elsewhere (at Bohio and Standley Ridge). Also, only six (8%) of the males caught two or more times at Standley End were caught as many as two times at other sites, all nearby (Fuertes, 1; Miller Ridge, 8; and Standley Ridge, 2). The sex ratio of the Standley End A. jamaicensis (102) is skewed toward males (73 : 29; Table 7-9). At Lutz, another Pattern II locality with similar fidelity statistics, the ratio is 53 :47. During three of the netting episodes at Standley End when two-plus bats were caught, males outnumbered females 30 to 1. Of males caught at least twice, 31% were also caught three or more times while only 14% of the females were caught that often. In contrast, only 21 (29%) males were caught over a relatively long period (two to four years), whereas 14 (48%) females were recaptured at Standley End during the same period. Close to three-fourths (71%) of the males were found there for a year or less (Table 7-9). Age distribution among bats captured two or more times was similar in males and females (Table 7-10); however, a larger proportion of the young males than young females stayed in the area to be caught later as adults. We believe that the males at Standley End were predominantly bachelors, and that most or all of the resident females were members of harems. Of the 20 females caught as adults during seasons of reproduction, 18 were pregnant, lactating, or postlactating at one or more of the captures. The greatest number of females caught as adults at Standley End in any half year was 11. Presuming that we caught most of the resident females, they probably represent from one to three harems. In the four harems of A. jamaicensis that Morrison (1979) studied on BCI, he found from four to 11 adult females (X= 6.5). We still do not know why fidelity was so high in the three Pattern II locality groups (Lutz, Bohio, and Standley End). Incomplete coverage of foraging ranges may be a factor at Bohio and Standley End, but not at Lutz, which is surrounded by well-netted localities that have typical low Pattern I fidelity percentages. Other peripheral locality groups such as Chapman, Standley Ridge, and Armour End show low fidelity, suggesting that peripheral location is not the reason. 114SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY TABLE 7-8.—Fidelity of Artibeus jamaicensis to a particular locality on BCI, ranked by number of netting nights. See Table 7-7 and text for explanation of categories. Locality PATTERN la (more than 19 netting nights) Miller Ridge Plateau Fuertes Shannon-AMNH Barbour-Hood PATTERNIb (less than 19 netting nights) Chapman Standley Ridge Barbour Stream Conrad Lake-Wheeler Harvard Armour End Drayton End Zetek21 Gross Point Shannon-Balboa Orchid Island PATTERN II Lutz Standley End Bohio Rank 1 2 3 4 5 1 2 3 4 5 6 7 8 9 10 11 12 1 2 3 Netting nights 52 46 26 24 20 17 16 14 14 12 11 11 10 6 5 4 4 197 25 14 60-100% 18 12 13 9 7 16 15 14 6 4 12 8 14 8 1 3 3 32 31 30 We suggest that the high capture rate at Lutz is correlated with the unusual local abundance of fig trees. That also could be true of Bohio because our netting there generally coincided with the availability of ripe figs; however, fig trees were scarce at Standley End. The abundance of fig trees at Miller Ridge rivals that of Lutz, yet fidelity was low, as was typical of other Pattern I locality groups. Geographically, the Pattern II locality groups are dissimilar: Bohio is on the mainland, Standley End is peripheral on BCI, and Lutz is more central. In terms of size, Lutz encompasses the largest area, and because of its proximity to the Laboratory Clearing, it received the greatest sampling effort. Lacking a better explanation, one could argue that the similarities in capture frequency among the locality groups in Pattern II are merely coincidental. Movements of Individual Bats Thus far we have discussed movements of bats associated with a particular locality group based on pooled data. We also examined the movements of individuals captured at two or more localities away from the central locality under consideration. To illustrate this we have mapped the movements of five female A. jamaicensis from Pefia Blanca (Figure 7-19). Each female was caught at two or more localities away from the light tower, and polygons outlining the locations where each was captured approximate their known home ranges. We then overlaid this map (Figure 7-19) with a 0.5 km grid and counted the number of times individual polygons touched each square. The resulting map (Figure 7-20) shows the frequency (from one to five) of occurrence of individual bats in each of the squares. We also mapped the same kind of information from 11 bats of the day roost in the Buena Vista light tower (Figure 7-21). This method proved effective for illustrating movements TABLE 7-10.—Age distribution of Artibeus jamaicensis captured two or more times at Standley-End, BCI. Sex Female Male Caught only when young Caught first as young later as adult Caught only as adult 14% 6% 38% 49%45% Sex FEMALE number percent MALE number percent TABLE 7-9.—Sexual variation in age, or more times at Standley-End, BCI. N J-S 29 73 1 1 Age during capture span J-A SAD 3 4 10 14 11 4 15 6 S-A 8 28 24 33 capture span, and number of captures of Artibeus jamaicensis caught two AD 14 48 33 45 1 15 52 52 71 Capture span (years) 2 3 7 4 24 14 48% 13 7 18 10 29% 4 3 10 1 1 Number of captures at Standley-End 2 25 86 50 68 3 4 5+ 3 1 10 4 14% 14 8 1 19 11 1 31% 2 11 38 31 42 Total captures 3 9 31 20 27 4 8 28 10 14 5 1 3 8 11 6+ 4 5 NUMBER 511115 45 i.•• • i.... 2 3 kilometers FIGURE 7-19.—Movements of five female Artibeus jamaicensis from the day roost in the Pena Blanca light tower to BCI and other nearby areas. 122SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY FIGURE 7-24.—Mean distance in kilometers between captures of subadult female Artibeus jamaicensis on and near BCI. Isolines enclose locality groups with similar mean distances and reflect distances between enclosed capture stations and all other capture stations. See text for further explanation. NUMBER 511123 kilometers FIGURE 7-25.—Mean distance in kilometers between captures of subadult male Artibeus jamaicensis on and near BCI. Isolines enclose locality groups with similar mean distances and reflect distances between enclosed capture stations and all other capture stations. See text for further explanation. 124SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY 45 46 47 48 49 50 51 52 53 54 55 i...ii.... 0 12 3 4 5 kilometers FIGURE 7-26.—Mean distance in kilometers between captures of juvenile male Artibeus jamaicensis on and near BCI. Isolines enclose locality groups with similar mean distances and reflect distances between enclosed capture stations and all other capture stations. See text for further explanation. NUMBER 511125 47 46 4555 i. •• •i... 0 12 3 4 5 kilometers FIGURE 7-27.—Mean distance in kilometers between captures of juvenile female Artibeus jamaicensis on and near BCI. Isolines enclose locality groups with similar mean distances and reflect distances between enclosed capture stations and all other capture stations. See text for further explanation. 126 SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY TABLE 7-12.—Catches of Artibeus jamaicensis and other bats during episodes of netting at two or more stations on the same night on BCI. Date 29 Nov 78 HJun79 7 Nov 79 8 Nov 79 9 Nov 79 11 Nov 79 12 Nov 79 13 Nov 79 14 Nov 79 15 Nov 79 16 Nov 79 18 Nov 79 12 Oct 80 13 nights Locality Lutz Creek Snyder-Molino 0-2 Barbour-Lathrop 16 Snyder-Molino 0-2 Miller 9 Armour-Zetek Jet. Armour-Zetek Jet. Armour-Conrad Jet. Miller 9 Donato 3-4 Miller 1-2 Barbour3 Donato 3-4 Barbour 3 Snyder-Molino-Lutz Standley 15 Standley 21 Miller 4-5 Standley 21 Lutz Creek Donato 3-4 Standley 16 Donato 3-4 Standley 16 Standley 16 Donato 3-4 Miller 13-14 Miller 15 Miller 8-10 Chapman 6-9 30 sites Artibeus jamaicensis 26 56 109 6 77* 83* 2 27 11 47 19 24 42 27 5 10 6 2 10 1 94 103 14 68 23 42 56 30 101 49 1170 Other bats 15 27 33 4 47 16 3 27 18 54 25 14 47 22 3 13 10 4 26 10 136 82 27 100 54 45 40 32 12 12 958 * One double capture. 60 for 128 nights; see Table 7-13), only 2% (range, 0%-6%) of individual A. jamaicensis were captured more than once at the same tree on successive nights. These recaptured bats were 5% (range, 0%-13%) of all captures of A. jamaicensis, which represents a pool of 7427 individuals amassed over 128 nights. In all, only 181 were caught at the same tree on two successive nights, and seven were captured on three successive nights. Individual A. jamaicensis were caught more than once in 35 of 46 multinight netting episodes at single ripe trees. The frequency of repeats (bats caught more than once in a night at the same station) is similar to the frequency of recaptures on successive nights. At least some A. jamaicensis reentered a net within minutes of release, probably because of disorientation resulting from the handling process and because of proximity of the processing station to nets. Nevertheless, the impression was that, once captured, these bats avoided the nets. During sampling periods totaling 133 nights in 1979 and 1980 we had 158 repeats of A. jamaicensis on 59 nights; an average of 2.7 repeats on the nights with repeats. On the nights with repeats, we caught 4022 A. jamaicensis, 3.9% of which represented repeated captures. Repeated captures represented 2.7% of all A. jamaicensis caught on all nights repeated (133 nights, 5946 bats). If we could estimate the proportion of marked and unmarked bats in a foraging aggregation we might be able to estimate the number of A. jamaicensis that come to a fruiting tree. We know that the recapture rate is low (<2%), but we do not know what our success rate is in capturing unmarked bats. To judge from the nightly ratio of unmarked to recaptured bats (Table 7-13), these variables must be changing from night to night. If we assume that many of the same bats return night after night to a choice tree as long as fruit remains, and that we recapture only a small fraction of marked bats, then a tally of TABLE 7-13.—Results of netting episodes involving consecutive nights at single localities on BCI and the adjacent mainland. The columns of capture records, multiple captures, and mean captures per night pertain to Artibeus jamaicensis alone. The last column, total captures all species, includes A. jamaicensis and all other bats. Date Jan 78 Apr 78 Jun78 Nov 78 May 79 Nov 79 Dec 79 Apr 80 Oct 80 Total Feb78 Scp79 Nov 79 Total Nights N 3 4 3 2 2 2 4 4 3 27 3 2 4 9 N 291 97 85 106 79 94 253 145 319 1469 99 83 109 291 Capture records Marks 179 37 72 75 48 27 76 38 120 672 102 58 87 247 Recaps 117 63 16 32 32 67 182 109 208 826 201 141 196 538 Total Multiple captures* Individuals N% Records N « Miller Ridge (9 multinight episodes) 296 100 88 107 80 94 258 147 328 1498 5 3 3 1 1 0 5 2 9 29 2 3 4 1 1 0 2 1 3 2 10 6 6 2 2 0 10 4 18 58 Standley Ridge (3 multinight episodes) 12 4 4 20 25 8 8 41 12 6 4 8 188 137 192 517 3 6 7 2 3 0 4 3 6 4 6 3 2 4 Mean captures per night 99 25 29 54 40 47 65 37 109 56 67 71 49 60 Total captures all all species 529 148 132 150 140 179 321 273 402 2274 274t 205 448 927 NUMBER 511127 TABLE 7-13.—Continued. Date Dec 78 Apr 79 Sep80 Oct80 Total Jan 80 Aug80 Oct80 Total Sep79 Oct79 Nov79 Aug80 Aug80 Total Mar 78 Mar 78 Nov-Dec 78 Mar 79 Oct79 Oct79 Nov79 Nov79 Nov79 May 80 Aug80 Sep-Oct 80 Oct80 Total Aug80 Dec 78 Sep79 Nov78 Sep80 Sep80 Apr 79 Apr 79 Apr 80 Total Total for table Nights N 2 2 2 3 9 2 4 2 8 2 2 2 3 2 11 2 2 3 3 3 2 2 2 2 13 2 3 2 41 2 2 4 3 2 3 2 2 3 23 128 N 76 149 101 121 447 50 270 285 605 138 105 82 107 98 530 67 63 120 145 332 108 89 54 108 767 147 169 104 2273 147 99 326 86 199 283 165 151 130 1586 7427 Capture records Marks 64 106 88 56 314 11 176 142 329 89 65 19 75 64 312 16 19 88 54 183 45 34 15 40 637 96 93 49 1369 93 52 225 70 100 203 163 147 113 1166 4453 Recaps 13 46 18 69 146 39 102 146 287 51 41 63 35 34 224 52 44 35 99 161 64 55 39 68 157 51 80 55 960 56 52 119 19 105 86 2 6 27 472 3162 Total Multiple captures* Individuals N % Records N Bohio (4 multinight episodes) 77 152 106 125 460 1 3 5 3 12 1 2 5 3 3 2 6 10 7 25 Fuertes (3 multinight episodes) 50 278 288 616 0 8 3 11 0 3 1 2 0 16 6 22 Plateau (5 multinight episodes) 140 106 82 110 98 536 2 1 0 3 0 6 1 1 0 3 0 1 4 2 0 6 0 12 Lutz (13 multinight episodes) 68 63 123 153 344 109 89 54 108 794 147 173 104 2329 1 0 3 8 12 1 0 0 0 27 0 4 0 56 2 0 3 6 4 1 0 0 0 4 0 2 0 3 2 0 6 16 24 2 0 0 0 54 0 8 0 112 Miscellaneous (9 multinight episodes) 149 104 344 89 205 289 165 153 140 1638 7615 Barbour-Hood 2 Chapman 5 1 5 RCS-AMNH 155 Lake-WMW 3 Conrad 6 4 3 Armour End 6 2 Orchid Island 0 0 Frijoles Road 2 Gigante 8 47 181 1 6 3 2 4 10 33 6 12 12 0 4 18 99 369 % 3 4 9 6 5 0 6 2 4 3 2 0 6 0 2 3 0 5 11 7 2 0 0 0 7 0 5 0 5 3 10 10 7 6 4 0 3 13 6 5 Mean captures night 39 76 53 42 51 25 70 144 77 70 53 41 37 49 49 34 32 41 51 115 55 45 27 54 61 74 58 52 57 75 52 86 30 103 96 83 72 47 71 60 Total captures all species 160 177 156 176t 669 119 325 342 786 181 393 101 131 112 918 159 143 178 260 419 149 189 90 271 969 171 230 172 3400 170 134 419$ 240 234 351 1% 220 224* 2188* 11162 * Individual A. jamaicensis captured two or more times during a multinight netting episode. t One caught three times. X Three caught three times. * Two caught three times. • Seven caught three times. 128SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY TABLE 7-14.—Large concentrations of Artibeus jamaicensis at single netting stations on BCI on successive nights. Locality Miller Ridge Shannon-AMNH Lutz Miller Ridge Lutz Fuertes Armour End Miller Ridge Fuertes Date Jan 78 Sep79 Oct79 Dec 79 May 80 Aug80 Sep80 Oct80 Oct80 Nights N 3 4 3 4 13 4 3 3 2 Artibeus jamaicensis N 291 326 332 253 767 270 283 319 285 Average number per night 99 86 115 65 61 70 96 109 144 the number of individuals captured during a multinight netting sequence should suggest the number of bats that flock to a tree (Table 7-14). We believe that, on occasion, these numbers must be large, and even if we are catching only a third of the bats, the total at a fruiting tree may exceed a thousand. Our usual impression at the nets during a highly successful night is that while the air seems full of bats, relatively few are getting into the nets. Group Movements The number of bats at a netting station, especially if it is at an active feeding roost near a fruiting tree, varies during the night, usually abruptly. This is evidence that the bats are moving about in groups. Whether these groups are associations of individuals that forage together night after night (and could be considered true flocks), or are merely aggregations of bats that form at a feeding roost and shift together from one feeding site to another during the night and then disband for indefinite periods of time is not known. There is no doubt about the existence of aggregations. Anyone who has much experience working with mist nets in the American tropics has to be impressed with abrupt and dramatic shifts in numbers of bats during the night, signaling the arrival and departure of groups. Heithaus et al. (1974) commented that large bats appeared to arrive in groups to feed and that visits by groups to food sources was pulsed. We refer to these pulses as surges, and we define a surge as an abrupt increase in numbers of bats captured, followed by an abrupt decrease. When and whether there are surges depend on such influences as rain, presence of ripe fruit, the amount of light, and the phase of the moon. Our data on surges were compiled for three-month periods at the height of the rainy season (August-November) in 1979 and 1980 (n = 133 nights; means summarized in Table 7-15). Admittedly these data are biased and incomplete. On a night with few bats we might stop netting after a few hours, thus recording a night without a surge. On similar nights we persisted, perhaps waiting for the moon to set and still did not record a surge; or, had the first surge commence as late as 2300 h. We undoubtedly missed some second surges by terminating too soon after the first. During the weeks of a full and new moon, moonrise and moonset often coincided with the beginning and end of surges. Bats are least active in the week following the first quarter of the moon (the week with most light in early evening). Captures were few (average 35.4 per night), surges infrequent and short (2.5 h), and only half of the nights had any surge at all (Table 7-15). Only 4% (one in 28 nights) had a second surge. The first surge was early, beginning at approximately 1915 h and ending at about 2045 h. Presence of fruit and absence of rain seemed to be of relatively little consequence during nights with bright moonlight. In spite of increasing light in early evening, A. jamaicensis was most active in the week following the new moon. Capture rate was high (averaging 53.8 per night) and surges occurred on 80% of the nights. Probably because of the early evening light, second surges occurred on 20% of the nights. The first surges were of long duration, lasting about 3.5 h, and they began about 1945 h and ended near 2215 h. Morrison (1978a) used the term lunar phobia to describe these correlations of activity and moon phase. Summary The recapture rate in our 8907 A. jamaicensis averaged 1.8 per bat, indicating that less than 50% were recaptured at all, and less than 20% were recaptured three or more times. The low recapture rate, long distances between recapture sites, and potential home ranges that include the entire island and parts of the adjacent mainland complicate movement analyses. We adjusted for some of these difficulties by pooling the data for each locality. We tabulated captures by species and by half years on a matrix of locality groups and then summarized the entire marking interval (October 1976 through October 1980). Although neither factor completely explains the variation, proximity of netting localities and differences in netting effort led to recapture rates varying from 1 % to 43%. Localities in the center of the island consistently yielded better recapture rates than did marginal localities, suggesting that bats living near the edge of the island also forage on the surrounding mainland. It proved impractical to use netting sites randomly on a grid on BCI. Capture success was influenced by rain, wind, cloud cover, moon phase, topography, and the distribution and abundance of preferred foods. Thus, we had to try to standardize capture effort by the use of such measures as recaptures per net-night, or net-hour. We settled on using the number of A. jamaicensis caught per locality as the most useful and effective measure of capture effort. The population of A. jamaicensis on BCI is not an a <- 5 <S 8-1 S J5 OOVO ONVO ON^* TfcS Q oo t-^ «—4 «-* \O ^-< NONO VOCN ONON Oco «O « CO CN CO ~H \O O O >ri CN f~ CO ON CO CO 00 a u g A 8.1 S JJ l-sl i8 O CN -H o 5 R8 O co •a 00 "> _, OO "J tS 3 ° 3 3° — on 2 a, on 2 Ci cent ;urge £ oo 2 OS 130SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY amorphous unstructured mass of mobile bats flying randomly throughout the island. Individuals are more likely to be recaptured close to where they were marked. Bats netted at a particular locality include those that have day roosts in the area as well as those coming to forage at fruiting trees, and those caught as they travelled between day roosts and feeding sites elsewhere. We believe that the distribution of fig trees is a major influence on the relative number of recaptures at any locality. We examined fidelity to a particular area (locality group) by analyzing suites of multiple capture records of the 3846 bats recaptured and found three patterns. Highest fidelity occured where we sampled day roosts, rather than foraging sites. Fidelity is at least partially dependent upon netting effort. In locality groups with unusually high fidelity, the sex ratio was skewed toward males. In one such group, most of the males were bachelors, and most of the females were members of harems. Data on movements of individual bats was restricted by the relatively small number of multiple recaptures. The average bat flies on about 700 nights during its lifetime, yet we captured a bat only an average of 1.8 times; a very narrow window through which to examine its movements. Radio tracking and capture of bats with known day roosts showed average movements of 1-4 km between day roost and feeding site. Mean distances between captures show no correlation with sex or age. Proportionately, records of movements by juveniles were much more frequent than expected and records of adults are about the same or a little less than expected, perhaps reflecting the relative difficulty in recapturing older adults. The longest distance we recorded between captures was about 6 km for three individuals. However, these were not single-night movements; therefore, these distances may represent dispersion instead of actual foraging distances. Artibeus jamaicensis often visit more than one tree in an evening. We recaptured one individual at a fruiting tree 1.2 km away from another tree where she had been captured earlier the same evening. Recapture rates were only 2% for bats captured more than once at the same tree on successive nights, and similarly low for bats recaptured the same night at the same station. The number of bats at a netting station shifted during the night, usually abruptly, suggesting that the bats are moving and foraging in groups. These surges of activity seemed influenced by rain, presence of ripe fruit, and phase of the moon. Clearly, bats were least active in the week following the first quarter of the moon (most light early in the evening), and most active in the week following the new moon. 8. Roosting Behavior Douglas W. Morrison and Charles O. Handley, Jr. In this section we present a synthesis of our mark-recapture, radio-tracking, and night-viewing-scope observations pertinent to day roosts of Artibeus jamaicensis on Barro Colorado Island (BCI). For comparison, we have included data on other species of bats where appropriate. We present a similar synthesis of observations on night roosts as part of the discussion in Section 9, Foraging Behavior. Mark-recapture and radio-tracking techniques have inherent strengths and weaknesses. With radio-tagging, the behavior of individuals can be intensely monitored for days or weeks. However, sample sizes tend to be small and there is no independent way to measure how much the behavior of an individual has been altered by the transmitter. In contrast, long-term netting studies can generate huge samples, but the data are sometimes difficult to interpret. Taken together, the two techniques complement each other in important ways. Questions provoked by radio-tracking observations can provide a framework around which to analyze mark-recapture data and netting data provide independent confirmation that the behavior of radio-tagged individuals is typical of the population as a whole. Using the two techniques in concert has increased our confidence in the accuracy of the picture that has emerged. Bats show diverse roosting behaviors, often using different kinds of roosts for different periods of their daily and annual cycles (Kunz, 1982). Day roosts offer protection from predators and the elements and are generally used for extended periods. In contrast, night roosts may be less protected, temporary sites chosen for their proximity to food sources. Individual bats also may change roosts on a seasonal basis, with certain types of roosts favored for mating, rearing young, or other activities. Day-roosting Sites A. jamaicensis apparently is opportunistic in its selection of day-roosting sites. Where caves are available, it roosts in large groups (Dalquest, 1953; Kunz, 1982; Tuttle, 1968). In forested Douglas W. Morrison, Department of Zoology and Physiology, Rutgers University, 195 University Ave., Newark, NJ. 07102. Charles O. Handley, Jr., National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560. habitats, the largest groups roost by day in tree hollows, while smaller groups and individuals are found in foliage (Morrison, 1979). In the moist tropical forest of BCI, we located 25 day roosts used by 18 radio-tagged A. jamaicensis. Sixteen were in foliage and nine in tree hollows. DAY ROOSTS IN FOLIAGE.—Foliage roosts were used primarily by males. Seven radio-tagged males made transient use of a variety of sites in the foliage, typically occupying a site for 3-5 days (range 1-13 days) before moving to another foliage site 100 m or farther away. In contrast, females normally roosted in tree hollows and rarely used foliage sites. Of 11 radio-tagged females, two roosted in foliage but only after being captured and radio-tagged as they emerged from their day roosts in tree hollows. The kinds of foliage roosts used by A. jamaicensis (a 45 g bat) are typical of those used by canopy-foraging fruit bats (Goodwin and Greenhall, 1961) and were indistinguishable from those used by A. lituratus (70 g) and Vampyrodes caraccioli (36 g) on BCI (Morrison, 1979, 1980a). Foliage roosts used by radio-tagged A. jamaicensis on BCI included shelters under a long, arching frond of a palm, Oenocarpus panamanus; under an "umbrella" formed by a single, wilted leaf of a broadleaf epiphyte; in the crown of a spiny black palm, Astrocaryum standleyanum; and in a shallow shelter formed by the forest canopy. Radio-tagged V. caraccioli (nine sightings) invariably roosted in groups of three or four adults, 7-12 m above the ground, under the umbrella-like crowns of understory trees (12-20 cm Diameter at Breast Height (DBH)). The day roosts of radio-tagged A. lituratus (26 sightings) were more variable, from 2.7-28 m above the ground in a variety of situations, such as under broken or crossed fronds of Oenocarpus panamanus, in vine-tangled crowns of subcanopy trees, in cavelike recesses on the underside of the crowns of canopy-height trees, and in branches overhanging the water along the lakeshore. As variable as these roosts might seem, all had two important features in common: they were difficult to see from the ground and they had unstable supporting structures. Despite their open appearance, foliage roosts may be almost as effective as tree hollows for reducing the exposure of adult bats to predation and rain. The dense leaf cover forming the roof of these recesses is 131 138SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY detector recordings of activity (Fenton, Boyle et al., 1977). Eckert (1982) argued persuasively that all such effects are adaptations for the avoidance of visually orienting predators. Longer-term Foraging Patterns Patterns of fruit tree use were most apparent on bright moonlit nights because flying was limited and feeding passes seemed especially regimented. Occasionally on such nights, a radio-tagged A. jamaicensis made what appeared to be a feeding pass in the wrong direction. In some cases, these flights were simply feeding passes to another fruiting tree, demonstrating that bats sometimes use a single night roost while feeding on more than one tree. Or, if these were nonfeeding passes they may have been reconnaissance flights to gain information about future feeding sites. Although bats returned to feed on the same fruit tree for as many as eight consecutive nights, they discontinued feeding visits when the density of the figs declined to less than one or two per square meter (n = 6), unless a fig tree was less than 150 m from the day roost. In switching fruit trees, a bat typically made no more than the first feeding pass of the night to the previously used tree and then immediately, without search flying, switched its feeding roost or feeding passes to a new tree. In 11 instances this shift was to an area that the bat had visited from 1-8 (4.2 ± 2.2) nights previously. In five of these 11 cases, the shift was to a fig tree that had almost certainly been the object of earlier reconnaissance flights. Flying by radio-tagged A. jamaicensis during the bright half of the lunar month was reduced to the minimum length of time required to get to a feeding roost (usually from one to five minutes). Even on moonless nights, the bats often made only short flights and spent most of their time hanging at the feeding roosts. Occasionally, however, bouts of sustained flying that lasted 10-45 minutes were recorded. Six of 40 all-night tracking sessions contained at least one such period of prolonged flying. All but one of those longer flights occurred during the dark half of the lunar month. The one exception occurred in the evening before a late (2345 hours) moonrise. Prolonged flights began either from night feeding roosts (n = 3) or upon emergence at dusk from day roosts (jn = 3). The area covered by these flights could only be approximated, as it was not possible to triangulate the bats' constantly changing position. Usually the flying was confined to the watershed in which it began (n = 4), but once it shifted into an adjacent watershed and another time into a nonadjacent watershed. These long search flights undoubtedly increase the chances of encountering newly fruiting trees and might have been initiated for that purpose. Feeding Roosts Triangulation by radio of foraging A. jamaicensis suggests that feeding roosts usually are in one or a few favored trees, within a 0.25 to 0.5 ha area, 25 to 200 m from a fruiting tree. Morrison (1975) was able to pinpoint and confirm, by on-site inspection, two feeding roosts used by his radio-tagged A. jamaicensis. One was in the crown of a 13-m-tall, spiny black palm (Astrocaryum standleyanum). The other was about 3 m above the ground under the long, arching frond of an Oenocarpus panamanus. Both sites could have been selected for their reduced accessibility to terrestrial predators (see Section 8, Roosting Behavior). After Alfred Gardner discovered that bats drop pellets of chewed pulp when feeding on figs, Bat Project personnel routinely used accumulations of pellets to recognize sites used as feeding roosts. Depending on the height from which they are dropped, the pellets may be in small piles, distinct clusters, or broadly dispersed. The area covered can be quite extensive, sometimes exceeding 100 m2. These feeding roosts are often located along ridges above fruiting trees, but may be found elsewhere, especially where young palms or subcanopy trees are numerous. The single common feature is an open understory that does not restrict flight. This may explain why we have found so many feeding roosts along or adjacent to trails, which also function as open flight paths for bats. Radio-tagged A. jamaicensis were relatively faithful to feeding roosts. Invariably the same feeding roost area was used night after night for as long as a bat fed from the same fruit tree (n = 14 fruiting trees). Four radio-tagged bats made foraging passes from the same feeding roost to two simultaneously fruiting trees 35-220 m apart, and two bats used the same feeding roost areas when they returned after three and eight nights, respectively, to get fruit from another tree in the neighborhood. There are several possible, but untested, explanations for this apparent fidelity to feeding roost areas. It might simply be the result of the patchy distribution of vegetation preferred for feeding roosts. Or it might be that A. jamaicensis is a creature of habit, preferring to use familiar sites along well-known flight paths. The latter hypothesis is supported by our observation that radio-tagged A. jamaicensis frequently did not use (or failed to find) a fig tree producing ripe fruit close to their day roosts. Instead they commuted two to three times farther to another tree. This suggests that many fruiting trees are found because they stand along familiar foraging routes. Feeding roosts may have other functions not related to feeding. For example, if they are closer to a fruiting tree than the day roost, they may serve as convenient havens during bright moonlight. This at least seems to be the case with A. lituratus and V. caraccioli, two fruit bats that have day roosts in foliage (Morrison, 1980a). Feeding roosts also may be sites for social interaction. Adult bachelor males may try to copulate with females at feeding roosts away from the day roost and its defending harem male because they probably lack other access to females (Morrison and Morrison, 1981). However, females may be less receptive to males encountered outside the day-roost hollow if competition for hollows selects for males NUMBER 511139 who sire stronger, more aggressive, or otherwise fitter young. Also, day roosts are probably safer sites for copulating bats than are night roosts. Group Foraging Feeding roosts might be "information centers" for food finding (sensu Ward and Zahavi, 1973). In theory, A. jamaicensis is a likely candidate for group foraging information sharing. A prerequisite for the evolution of information centers is that food be found in short-lived, locally superabundant patches. This sort of food distribution means that users must continually find new food patches. However, because the patches contain so much food (and for so short a time), nothing is gained by trying to conceal or defend a food patch once it is found. Fig trees fit this description. A single fig tree on BCI may bear 40,000 figs, of which a single A. jamaicensis would likely carry away no more than 100 (Morrison, 1978d) before the fruits were gone or spoiled. A second characteristic that makes A. jamaicensis a potential user of information centers is that foraging bats aggregate at night (feeding) roosts. A night roost may be occupied by dozens, if not hundreds, of bats, and they may be making feeding passes to several different fruiting trees. A simple mechanism for information exchange is plausible. If an A. jamaicensis, which has been successful in finding a fig tree, returns to its feeding roost with a fig or the odor of figs, any roost mates who had been unsuccessful in finding food could follow the successful bat out on its next feeding flight. The information sharing would not need to be intentional, but the information would greatly benefit the unsuccessful bat and would not significantly deplete the food at the tree for the successful bat. In a group that remained together over a period of time, it is possible that tonight's successful bat could be tomorrow night's unsuccessful bat, and vice versa, so all would gain over the long term. Note also in this context the probability of food finding reconnaissance flights initiated in advance of need. Despite the theoretical potential, it is still unknown whether A. jamaicensis on BCI forages in cohesive groups. Certainly this bat does forage in groups, as the nightly ebb and flow of bats at a fruit tree demonstrates (see Section 7, Movements). However, whether the flocks that swirl about a fruit tree stay together for hours, all night, several nights, or for long periods of time is unknown. Morrison and Morrison (1981) tracked a group of three and another of four radio-tagged female A. jamaicensis from two different harems for eight complete nights and found no evidence of cohesiveness of harem members away from the day roost. These females left and reentered their day roost hole individually. Because emergences were typically more than a minute apart, each female emerged after the previous one had disappeared from the day roost area. On three of the eight nights, three of four females did visit the same ripe-fruitbearing Ficus insipida in the course of the night, but they moved between fruit trees independently and did not roost anywhere near each other while feeding. This radio-tracking study suggests that harem females do not forage in groups, at least not during the circumstances operant when our observations were made. Nevertheless, it is still possible that other sex and age classes forage in groups, or that A. jamaicensis forage together in habitats where fruiting trees are not as abundant and easy to find as they are on BCI. In Mexico and Costa Rica, for example, clumped mist-netting capture times have been interpreted to mean that the A. jamaicensis there forage in groups (Dalquest, 1953; Heithaus et al., 1975). Group foraging has been reported in flower-feeding bats in Brazil (Sazima and Sazima, 1977) and Arizona (Howell, 1979). It should be possible to detect foraging-group associations by analysis of the occurrence of "double recapture pairs" in the mark-recapture data. A "double recapture pair" consists of two bats that were captured at the same netting site on the same date and were subsequently recaptured together at another site some days, months, or years later. Morrison (1975) found 22 such pairs among the 259 recaptures of 1472 A. jamaicensis he and Bonaccorso marked over a 14-month period in 1972 and 1973. What is the probability that this number of double recapture pairs might occur simply as a result of chance associations in a population of independently foraging bats? Given that two or more recaptures are made on night n, let Pn stand for the conditional probability that any two of these recaptures were captured together previously (Morrison, 1975). If (n - 1) = the number of previous capture nights, \{ = the number of bats captured on previous capture night i, where l<i<n - 1, and y = the total number of bats banded to night n, then P = y fr1) n S S y (y-D The number of double recapture pairs Cj that could be expected to occur by chance on any night n is simply Pn times the total number of pair combinations of recaptures on night n, or Pa times c2r, where zn is the number of recaptures on night n. The total numer of recaptures expected to occur by chance during the entire 14-month netting program is the sum of the P n. c/ values of all 60 of the 131 sampling nights in which two or more recaptures were made. This sum was calculated and found to equal 14.2, the number of double recapture pairs expected to occur simply by chance. The observed value of 22 was not significantly greater than that expected from random assortment as determined by a one-tailed binomial test, P = 0.195 (the a priori expectation for higher than random assortment, justified the use of a one-tailed test). Thus, Morrison concluded that A. jamaicensis does not forage in groups, at least not in groups with memberships sufficiently constant over the long term to be detected by this method. However, this conclusion was based on the unrealistic 140SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY assumption that there is no mortality of marked bats. The chance that a marked bat will die reduces the expected number of double recapture pairs and so increases the significance level of the observed value. To reflect this reality we used the equation a~ Here it is reasonable to make s equal to 0.57, the annual survivorship for adult A. jamaicensis (see Section 6, Population Estimates), and for t£ we use 0.32 years, the average interval between recaptures for Morrison's (1975, table 1) data. The expected value now drops from 14.2 to 9.9 [14.2 x (0.57)0-64]. The observed value of 22 double recapture pairs is significantly greater than this revised expected value (one-tailed binomial test, P < 0.05). Unfortunately, even if these revised calculations reveal an incidence of double recapture pairs significantly greater than expected, we still cannot be sure that this difference is due to the existence of foraging groups. Other factors could produce the same result. For example, the probability of observing a double recapture pair would be increased if independently foraging bats frequented the same feeding area. Two bats might frequent the same area because it is especially attractive (e.g., the great Lutz fig patch) or because it is near the day roosts of both bats. The calculations of site fidelity (see Section 7, Movements) indicate that the capture sites were not in equally attractive areas and that individual bats were not equally likely to be captured at all sites. Furthermore, we suspect that after the first capture, bats tend to avoid mist nets because they are seldom recaptured (see Section 6, Population Estimates). Recapture pairs might be together at fruiting trees many times before one or both are recaptured again. These inequalities tend to increase the expected number of double recapture pairs. Without some estimate of the magnitude of these biases, any test for foraging groups based on the incidence of double recapture pairs will be inconclusive. Thus, although clustered capture times and double recapture pairs suggest group foraging, we lack sufficient evidence to show that these associations are anything more than coincidental. A powerful test for group foraging could be based on direct observation of the foraging movements of a large number of radio-tagged bats captured together at the day roost or captured in "clusters" while foraging. This remains to be done. Summary The usual pattern for an A. jamaicensis is to leave the day roost half an hour after sunset, fly to a fruit tree, take a fruit to a nearby feeding roost, consume it (dropping dry pellets to the ground beneath), make several more feeding passes to the fruit tree, then return to the day roost half an hour before sunrise. On dark moon nights, a bat may visit as many as five feeding areas. On nights with bright moonlight, bats visit only one or two feeding areas, and several hours may be spent back in the day roost during the brightest hours. Both lunar phobia and the use of feeding roosts separate from the fruiting tree may reduce exposure to predation. Radio-tracking observations suggest that bats sometimes use the same feeding roost to make feeding passes to more than one fruiting tree. The bats also appear to make brief reconnaissance flights to assess the condition of previously located fruit trees. The bats spend most of their time hanging in the feeding roost and seem to minimize feeding time even on dark moon nights. Prolonged flights of 10-45 minutes (searching for new trees?) were rare and occurred only in the absence of bright moonlight. Feeding roosts tend to be groups of favored trees 25-200 m from a fruit source, and most are located in areas of open understory. Bats are faithful to individual feeding roosts as long as the nearby fruiting tree remains productive. This fidelity may result from patchy distribution of preferred roosting trees, or may simply reflect the habitual use of a familiar area. The bats clearly congregate at food sources, but whether cohesive, long-term group foraging occurs is unknown. Clustered capture times and double recapture pairs suggest group foraging, but direct observation of the phenomenon is lacking. Radio-tracking data on females from the same harem show no evidence of such flocking. Whether or not A. jamaicensis forms foraging groups, the potential exists for an exchange of food location information among individuals at feeding roosts. Foraging theory suggests that feeding roosts could be information centers for finding trees with ripe fruit (short-lived, locally superabundant food patches), especially if the bats using a night roost are making feeding passes to different fruiting trees. 10. Food Habits Charles O. Handley, Jr., Alfred L. Gardner, and Don E. Wilson Initially, the Bat Project focused entirely on capturing and marking bats, and little attention was given to food habits and other aspects of natural history. Gradually it became apparent that capture rate usually was linked to the foraging behavior of the bats and to the location of nets in relation to food sources and feeding roosts. Consequently, Bat Project personnel took increasing interest in the content of the bats' feces, in evidence and location of feeding roosts, in food items carried by bats into nets, and in the location and phenology of fruiting trees. As a result, capture effort became more productive in number of bats marked and recaptured. Although our data on food habits lack the precision of the smaller-scale studies of Bonaccorso (1979) and Heithaus et al. (1975), they do offer insights on food preferences, seasonality of food availability, and the foraging habits of Artibeus jamaicensis on Barro Colorado Island (BCI). Project personnel eventually established a protocol that included searching for evidence of bat activity before selecting sites as capture stations. Along the trails we watched for pellets ("bat chop") that stenodermatine bats spit out after chewing and pressing the juice from the pulp of fruits such as Ficus insipida, F. yoponensis, and Spondias radlkoferi. The pellets appear in discrete piles or loose clusters on the ground when the bats drop them from roosts low in the subcanopy, or they may be scattered, as though broadcast, when they are dropped from the canopy. Because the bats return again and again during the night to the same perch, piles of pellets may represent many fruits. We also watched for fragments of partly eaten fruits; large seeds such as those of Spondias mombin, S. radlkoferi, and Dipteryx panamensis discarded by bats after they had scraped off the pulp; cast off skins of fruit such as Quararibea asterolepis; and trees dropping ripe fruit of kinds known to be favored by bats. Congregations of noisy diurnal frugivores such as monkeys (Alouatta palliata and Cebus capucinus), guans Charles O. Handley, Jr., and Don E. Wilson, National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560. Alfred L. Gardner, NERC, US. Fish and Wildlife Service, National Museum of Natural History, Washington, D.C. 20560. (Penelope purpurascens), and parrots (Amazona spp.) often led to the discovery of trees with ripening fruit. We left the trails to search under certain trees (e.g., Oenocarpus panamanus Bailey and Gustavia superba (Humboldt, Bonpland, and Kunth) (Berg) particularly favored by bats as feeding roosts. We found that patches of these trees associated with several fruit trees were used repeatedly by A. jamaicensis as feeding roosts. Thus, in the course of a year, the bats may use the same feeding roost many times. Taking this into account, we routinely inspected known feeding roosts in addition to walking the trails in search of new sites. Fruits Used as Food Our observations on fruits eaten by A. jamaicensis, mostly late in the rainy seasons (August to November) of 1979 and 1980, are summarized here (Figure 10-1 and Table 10-1). Ficus insipida Willdenow: Bonaccorso (1979), Fleming (1971), Morrison (1978d), and our own observations agree that in central Panam£ F. insipida is the favorite food of A. jamaicensis (Figure 10-1), A. lituratus, and perhaps Vampyrodes caraccioli, and is eaten in lesser amounts by several other bats. Its fruit production is aseasonal, but the crop is limited in October and from January through March (Table 10-1). Few fruits reach maturity in August and September, two months when F. insipida does not seem to be a significant food for bats. Large concentrations of bats, mostly large stenodermatines, were found near F. insipida bearing ripe fruit in March, April (four sites), May, June, October (two sites), November (two sites) and December. Normally the bats pick soft, fragrant, fully mature fruit, but occasionally (as was noted on 11 January, 8 February, 21 October, and 7-9 November when few suitable F. insipida were available) A. jamaicensis carried partly eaten small, hard, and latex-laden unripe fruits into the nets. Ficus yoponensis Desvaux: Similar to F. insipida in abundance and seasonal availability on BCI (Table 10-1), the small-fruited F. yoponensis is a favorite of smaller frugivores such as Uroderma bilobatum, Vampyressa pusilla, V. nymphaea, and Vampyrops helleri and is consumed in great 141 142 Ficus inslpida Ficus yoponensis Ficus sp. Ficus trlgonata Ficus obtusifolia Ficus costaricana Ficus popenoel Dlpteryx panamensis Quararlbea asterolepis Spondias mombin Spondias radlkoferi Poulsenla armata Anacardium excelsum Calophyllum longifolium Cecropla sp. Unidentified fruit pulp Unidentified seed Pollen on fur JAN FEB MAR SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY APR MAY AUG SEP OCT NOV DEC 3 A FIGURE 10-1.—Seasonal distribution of foods of Artibeus jamaicensis on BCI observed during the period 1977-1980. Squares represent seedless feces, circles represent feces with seeds, and triangles represent fruit carried into nets. Numbers represent observations per month. quantities by larger stenodermatines when F. insipida is scarce. Although we caught large numbers of A. jamaicensis at fruiting F. yoponensis, we often had the experience of netting under a tree dropping ripe figs and finding the area devoid of bats, or of catching A. jamaicensis carrying F. insipida from some more distant tree. Bats with feces containing seeds of F. yoponensis were outnumbered by about two to one by bats with feces containing F. insipida. However, large concentrations of bats were found at ripe F. yoponensis in May, October (two sites), and November (three sites). Ficus dugandii Standley: On the breezy nights of 25 and 26 October 1979, a giant F. dugandii, with a crown emerging above the canopy near the highest point of the island, attracted (perhaps by wind-wafted odor) large flocks of unmarked stenodermatines, presumably from the mainland. In mist nets beneath the tree, which was dropping many ripe fruits, we captured only 106 A. jamaicensis, 7 A. lituratus, and 2 V. caraccioli, while the catch of smaller stenodermatines totalled an astonishing 265, including 138 Uroderma bilobatum, 90 Chiroderma villosum, 15 Vampyressa nymphaea, 6 V. pusilla, 1 Vampyrops helleri,! Artibeus phaeotis, 1 A. watsoni, and the only A. hartii ever taken on BCI. Ficus trigonata Linnaeus: The red-spotted, small-seeded fruit of Ficus trigonata was an important food of A. jamaicensis and C. villosum at the height of the rainy season in September when fruit of F. insipida and F. yoponensis was scarce (Table 10-1). Many fruits of F. trigonata were carried into the nets and its seeds were common in feces (Figure 10-1). F. trigonata also occasionally was found in the feces of Carollia perspicillata and Phyllostomus hastatus. Although commonly seen in September, the lean month, the fruit of F. trigonata seemed not to dominate the diet of any bat in that month save possibly Chiroderma villosum. Spondias mombin and Quararibea asterolepis were eaten more commonly. One F. trigonata was NUMBER 511 143 1'«? -2 2 3 •S -3 .s Si i~ Si •c oo bear 3 tre .s 1 §• bers i z ON .2P •i :cies §• o u | 3 eg here i 0000^0000 00000000000 000000 00000—0—0 00000000000 000000 — — — cSOO fS — O OOOtS — — —ifS — tS — <S tS —< — fS •-•CS — 00000 — 000 000000000 — 0 000000 000000000 00000000000 000000 (SOO — 0000 — 0000 0000000000 0000000 000000 — 00 00000000 — 0 0000000 — en — — csen — OtSencs-^aocScscn — O — — — OenenTf — en N O O >• O O >. — ts mOOOO es — •—> OOOOOOO OOOOOOOOOOO OOOOOOO OOOOO~~-O OOOOOOOOOOO OOOOOOO f— — — OOCS Tf r- •"-• O o en cCS fS IS — — OOOOOOtS O — ~~ fS en S 5 i 1 144SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY carried into a net by an A. phaeotis in March. Ficus obtusifolia Humboldt, Bonpland, and Kunth: The large velvet-skinned fruits of F. obtusifolia were carried into nets occasionally by A. jamaicensis, A. lituratus, and C. villosum. The fruits seemed to be more attractive to the smaller frugivores such as C. villosum. We found bats concentrated at an F. obtusifolia with ripe fruit in March. Ficus popenoei Standley: The velvety, oblong fruits of F. popenoei are uncommon on BCI, and we found only a few trees with fruit. We netted near one that was dropping great quantities of ripe fruit, and although we caught a number of small stenodermatines of various species, the tree was ignored by A. jamaicensis and the other larger fruit eaters. Spondias mombin Linnaeus: An important food of bats (Gardner, 1977), S. mombin, ripens in August and September when fig production is low. It is found in scattered patches in the young forest on the lower-lying areas of the island. Apparently A. jamaicensis ingests the pulp of this fruit without trying to extract the juice and, thus, does not make pellets. Although S. mombin fruit seldom was carried into the nets, those nets set nearest to trees with ripe fruit invariably caught the most A. jamaicensis and C. perspicillata. A large concentration of bats, mostly A. jamaicensis, was found in a patch of S. mombin between 12 and 17 September. Spondias radlkoferi Donnell Smith: Widespread on BCI, but neither forming patches nor producing such great quantities of fruit as S. mombin and Q. asterolepis, S. radlkoferi is an important source of food for bats in the season of fig scarcity at the height of the rains. Fruit of this tree is ripe from September to November and is much sought by A. jamaicensis, A. lituratus, A. phaeotis, A. watsoni, and V. caraccioli. In 1979, ripe fruit began to fall on 19 September. Many fruits were carried into the nets on 26 September, and by 30 September most A. jamaicensis and A. lituratus that we captured were eating it. A large concentration of bats was associated with a S. radlkoferi with ripe fruit on 7 October. The last time we noted feces containing fruit pulp of this species was on 7 November. Seeds of S. radlkoferi with much or all of the pulp scraped off by bats were common under feeding roosts throughout October. Quararibea asterolepis Pittier: The beautiful Q. asterolepis grows in patches, which frequently contain many individuals, mostly in the old forest on the elevated interior of the island. We found it at lower elevations at Snyder-Molino 2, Shannon 1, mouth of Barbour Creek, Standley 16, and Wheeler 26.5. When its fruit ripens in August, September, and October it is the most abundant tree on the Plateau with fruit eaten by A. jamaicensis. Individual trees drop large quantities of ripe fruit over a period of two or three weeks or more. We found concentrations of bats at Q. asterolepis with ripe fruit on three occasions in September, and A. jamaicensis and C. perspicillata frequently carried the fruit into our nets. The fibrous fruit shell remains for weeks on the ground under feeding roosts, persisting long after the fruiting season has passed. Q. asterolepis and S. mombin, together with the much less abundant F. trigonata, are the fruits of choice for A. jamaicensis in September, and possibly in August as well, when fruits of F. insipida and F. yoponensis are scarce or absent. It is of interest that Q. asterolepis and S. mombin, which fruit simultaneously and are visited by the same bats, have complementary, nearly nonoverlapping distributions on BCI. Both occur in patches and produce much more fruit than the vertebrate frugivores can consume. Near fruiting Quararibea astrolepis, A. jamaicensis usually was caught in small groups rather than in such large concentrations as often appeared at favored fig trees at other seasons. Some Q. asterolepis, even when dropping quantities of fruit, almost seemed to be ignored by bats on some nights when we netted near them. Dipteryxpanamensis (Pittier) Record and Mell: During the dry season when fig productivity is low, the fruit of D. panamensis is an important food of A. jamaicensis. On BCI Dipteryx is a widespread and abundant tree,'sometimes forming large patches (as at Zetek 22 and east of Standley 19). Its fruit is ripe from December to March and occasionally at other seasons of the year (fruit was carried into a net in midNovember). We found large concentrations of bats, principally A. jamaicensis, at ripe D. panamensis in January 1979 and February 1978, when numerous fruits were carried into the nets. Brosimum alicastrum Swartz: Robin Foster (pers. comm.) regards bats as the principal dispersers of B. alicastrum seeds, and Croat (1978) believed that B. alicastrum is second only to figs in importance as a food for forest animals. We found a large concentration of bats, including almost 100 A. jamaicensis, at a Brosimum bearing ripe fruit in May. However, a large tree raining ripe fruit in October, near the end of the fruiting season, was attractive to kinkajous (Potosflavus) and peccaries (Tayassu tajacu), but apparently was of little or no interest to bats. Anacardium excelsum Bertero and Balbis: Abundant and widespread, and fruiting from March to May, A. excelsum is an important food of C. perspicillata. We often caught this bat, and A. phaeotis on one occasion, carrying the fruit of A. excelsum. Except for one fruit carried into a net in March, we have no evidence that A. jamaicensis eats the fruit of this species. Poulsenia armata Miguel: The spiny fruits of P. armata occasionally were carried into the nets by A. phaeotis (January, September, and October), A. watsoni (October), and once by a A. jamaicensis (November). Calophyllum longifolium Willdenow: In September and October we often found the large seeds of C. longifolium under dining roosts of A. jamaicensis, and occasionally we found remains of pulp and skin in the feces of this bat. The pellets from the fruit of C. longifolium dropped by A. jamaicensis and A. lituratus seem extremely resistant to decay and persist on the ground for several weeks. Although the large round seeds are conspicuous, the pellets are much smaller and darker than those NUMBER 511145 dropped by A. jamaicensis when feeding on figs, and consequently we often overlooked them. Fruits of C. longifolium that were carried into nets by bats were offered to A. jamaicensis, A. lituratus, and V. caraccioli temporarily held in captivity on BCI. Both species of Artibeus ate the fruit, but V. caraccioli did not, although it readily ate figs when those were offered. Cecropia spp.: Occasionally we found seeds of unidentified species of Cecropia in feces and under feeding roosts of A. jamaicensis. Guettarda foliacia Standley: Partly eaten fruits of G. foliacia were found beneath feeding roosts, probably of A. jamaicensis, in September 1979 and October 1978. Hura crepitans Linnaeus: Numerous fleshy staminate flower stems of Hura crepitans were carried into nets by Phyllostomus discolor, A. jamaicensis, and A. lituratus in November. Unidentified Pollen: In November and December, numerous bats of the species Phyllostomus hastatus, Glossophaga commissarisi, Uroderma bilobatum, U. magnirostrum, A. jamaicensis, and A. lituratus were stained yellow with unidentified pollen (probably mainly from flowers of the Bombacaceae). Tree Selection We know that A. jamaicensis will eat a variety of fruits, and sometimes other foods, but most of its meals are figs. On BCI it prefers F. insipida over all other figs, and it must be a rare night on the island when there are no ripe fruits of this species available. We have no idea why A. jamaicensis occasionally chooses the fruits of Brosimum, Calophyllum, Dipteryx, Quararibea, or Spondias over Ficus, but we have learned much about why and when they choose a particular fig tree. The normal massive crop of figs of each tree goes through several stages of harvest as it ripens. First come the howler monkeys (Alouatta palliata), which begin to eat the fruit long before it is ripe, while the pulp is hard and the skin is still full of latex. The monkeys are followed by the local bats from nearby roosts. They also begin to harvest a crop before it is fully ripe, probably selecting scattered ripe fruit, and sometimes picking unripe fruit. We have caught A. jamaicensis carrying unripe figs, but we do not know whether they actually eat such fruit, and if they do, whether they eat all or only part of it. Perhaps they are sampling crops to determine the stage of ripeness. Pellets dropped by local bats may be conspicuous during several nights before enough of the fruit crop is ripe to attract large groups of bats. When the big crop finally is fully ripened, bats at the tree may number in the hundreds or even thousands. They come in surges, and some may stay in the vicinity all night (see Section 7, Movements and Section 9, Foraging Behavior). Groups will return to a tree that has a big crop for three or four nights. In the final stage of harvest, monkeys, other diurnal frugivores, and local bats still return to the tree after its crop has been too depleted to continue to attract large flocks of bats. Altogether, local bats may harvest figs from a tree over a period of eight or nine nights. Some trees produce large crops that ripen a few figs at a time during a prolonged period (up to 10 to 15 days) rather than ripening abruptly as figs normally do. These trees are attractive to local bats and nonchiropteran frugivores, but not to groups of bats, except for brief passes. The remnants of fig crops aborted because of lack of pollination by fig wasps, or because of infestation with larvae of beetles and flies, as well as the crops produced by small, young trees, might not attract many bats. Size of the fruit crop has an important bearing on its use by bats. Small crops may be harvested by local bats, but usually they are not attractive to groups. Because monkeys {Alouatta palliata) begin to harvest a crop before bats do, they may strip a small crop before it is ripe enough to attract bats. Local bats may help finish harvesting a crop that is too small to attract a group. Or, a group may make one or two feeding passes and then leave. Radio-tracking revealed that a bat may visit three to five trees in a night (Morrison, 1978a). The fruits of figs are subject to destructive processes such as infection with fungi and infestation by insects that make them unattractive to bats. During the rainy season we often saw F. insipida infected with a fungus manifested in fully developed ripe fruit that have a bearded appearance while still on the tree. As falling fruit accumulates, the ground beneath the tree becomes carpeted with the fuzzy white fruit. Nothing eats it, in the tree or on the ground, and eventually it rots away. Fungi, as well as invertebrates of the litter, accelerate the breakdown of the pellets of chewed fig pulp dropped to the ground as bats feed. Only the most discerning eye will spot traces of these pellets four or five days after they have been dropped. The fruit of both F. insipida and F. yoponensis can be infested by the larvae of flies and beetles. If the infestation is heavy, the tree aborts part or all of its crop at about the sixth week (about three-fourths of the way through the developmental cycle). These hard, latex-laden, and insect-riddled fruits are eagerly eaten by monkeys (both Alouatta palliata and Cebus capucinus), frugivorous birds, and terrestrial frugivores such as pacas (Agouti paca), peccaries (Tayassu tajacu), and tapirs (Tapirus bairdii). However, although they may be large and appear to be ripe, they are not sweet, and they are ignored by bats. Until we learned the nature of this kind of fruit-crop abortion, we sometimes wasted effort by netting at these trees. We were misled by the abundance of falling fruit and the feverish activity of many frugivores into believing that bats also would flock to the tree. Summary Because capture rate is clearly linked to foraging behavior, we gathered data on fecal contents, feeding roosts, food carried 146SMITHSONIAN CONTRIBUTIONS TO ZOOLOGY into nets, and on the location and phenology of fruiting trees. The bats' habit of dropping small pellets of chewed fruit pulp under favorite feeding roosts frequently dictated our choice of netting sites. Congregations of diurnal frugivores led us to trees with ripening fruit that would be visited at night by bats. Certain species of trees favored as feeding roosts are used repeatedly, especially in the vicinity of fruiting trees. Fruit of the fig Ficus insipida is the favorite food of A. jamaicensis as well as several other bats on BCI. Fruits of other species of figs, including F. yoponensis, F. dugandii, F. trigonata, and F. obtusifolia, are sometimes eaten by A. jamaicensis, but the uncommon F. popenoei is ignored. Spondias mombin, S. radlkoferi, and Quararibea asterolepis are important food sources late in the rainy season when figs are scarce. Dipteryx panamensis is used in the dry season as an alternate food source. The fruits of Anacardium excelsum and Poulsenia armata and the flowers of Hura crepitans occasionally are eaten by A. jamaicensis. The normal massive crop of a F. insipida is used by local bats for several days before the entire crop is fully ripe, at which time hundreds or even thousands of bats may visit the tree for three or four nights until the crop is depleted. Then local bats continue to use the tree for several nights until there is no more fruit. Trees with small crops and trees producing crops that ripen asynchronously over a period of a week or more are visited by local bats but are largely ignored by groups. 11. Diet and Food Supply Charles O. Handley, Jr., and Egbert G. Leigh, Jr. Bonaccorso (1979) argued that food supply limits populations of bats on Barro Colorado Island (BCI). Late in the rainy season when fruit is scarce, he found more fruit bats with empty stomachs than he did at other times of year. About 83% of the frugivores he netted in October and November had empty stomachs, in contrast to 71% of those netted in March and April when fruit was far more abundant. However, these data must be interpreted with caution. Bonaccorso kept his data free of bias by netting the same localities each month, without regard for presence or absence of fruit near the netting stations. The bats he caught with empty stomachs must have been on their way to someplace else where there was fruit They could not endure empty stomachs for more than a few hours without starving to death. Moreover, Bonaccorso (1979) observed that birth (and breeding) of bats is timed to coincide with seasons of fruit abundance, with one birth peak in March and April coinciding with the fruiting peak at the onset of the rainy season and another in July and August coinciding with the fruiting peak of August and September (Figures 4-6 and 4-7). Few frugivorous bats give birth between November and mid-March when fruit is least abundant. Bonaccorso (1979) also found that the diets of various species of bats that forage for fruit in the canopy differed to the extent one would expect if these animals were food limited. He found that larger bats ate larger figs such as Ficus insipida and F. obtusifolia, whereas smaller bats concentrated on smaller figs such as F. popenoei and F. yoponensis. For the three larger frugivores (Artibeus lituratus, A.jamaicensis, and Vampyrodes caraccioli), he found that the regression of the mass Y of a fruit carried by a bat on the mass X of its carrier was Y = 0.23X - 3.92 g (r2 = 0.46, n = 27). Presumably, smaller bats carried fruit into the nets so rarely that he could not extend the regression. Reading the mass of these fruits from Bonaccorso's graph and calculating the mean and standard deviation of the logarithms of the mass of the fruits these bats carried, we Charles O. Handley, Jr., National Museum of Natural History, Smithsonian Institution, Washington, D.C. 20560. Egbert G. Leigh, Jr., Smithsonian Tropical Research Institute, Unit 0948, APO AA 34002-0948 or Apartado Box 2072, Balboa, Republic of Panama. derived values of 2.30 ± 0.24 (n = 6) for A. lituratus, 2.04 ± 0.33 (n = 17) for A. jamaicensis, and 1.54 ± 0.31 (n = 4) for V. caraccioli. The standard deviations are roughly equal to the differences between neighboring means. May and MacArthur (1972) showed that in an idealized competitive community, species could coexist securely if the sizes of foods eaten by the various species differed to this extent. Excepting Artibeus phaeotis, which eats few figs, the ratio of the mass of each species of canopy frugivore to that of its next smaller competitor was roughly the same as in these three species, about 1.4 : 1 (Bonaccorso, 1979). It is tempting to assume that the relation between the sizes of the smaller stenodermatines and the sizes of fruit they eat is the one Bonaccorso inferred from the few data on his largest frugivores. Furthermore, we could conclude from the elegant theory developed by May (1974) and May and MacArthur (1972) on niche overlap, that the canopy frugivores coexist by virtue of the differences Bonaccorso observed in the sizes of fruit these bats eat Ideally, if all sizes of figs were available at all times and were uniformly distributed, the smaller bats would usually take the smaller figs and the larger bats would usually take the larger figs. However, small-fruited and large-fruited species of figs are not uniformly distributed either in time or space, and the bats are adaptable enough to take what is available. Of course, in standardizing their diets, it is also necessary to take into account the foraging behavior of these bats. The energy expended in commuting from dining roost to fruiting tree makes it energetically imprudent for the large bats to routinely feed on small figs from each of which they would extract a comparatively tiny amount of nutrients (see Section 2, Physiology). Fig Production When we became concerned about the relation between food availability and the bats' energy requirements (see Section 2, Physiology), we began to gather information on estimates of fig production on BCI. Each individual fig tree bears fruit to its own rhythm (Morrison, 1978d) so that at any season some trees are bearing ripe figs (Table 10-1). Fig production peaks early in 147