University of Kansas Publications MUSEUM OF NATURAL HISTORY The University of Kansas Publications, Museum of Natural History, beginning with volume 1 in 1946, was discontinued with volume 20 in 1971. Shorter research papers formerly pub- lished in the above series are now published as Occasional Papers, Museum of Natural History. The Miscellaneous Publica- tions, Museum of Natural History, began with number 1 in 1946. Longer research papers are published in that series. Monographs of die Museum of Natural History were initiated in 1970. All manuscripts are subject to critical review by intra- and extra- mural specialists; final acceptance is at die discretion of the publications committee. Institutional libraries interested in exchanging publications may obtain die Occasional Papers and Miscellaneous Publica- tions by addressing die Exchange Librarian, University of Kan- sas Library, Lawrence, Kansas, 66045. Individuals may pur- chase separate numbers of all series. Prices may be obtained upon request addressed to Publications Secretary, Museum of Natural History, University of Kansas, Lawrence, Kansas 66045. Editor: E. O. Wiley PRINTED BY UNIVERSITY OF KANSAS PRINTING SERVICE LAWRENCE, KANSAS -AJft-L **> SEP 1 9 1977 HARVARD OCCASIONAL PAPERS MUSEUM OF NATURAL HISTORY of the The University of Kansas Lawrence, Kansas NUMBER 69, PAGES 1-28 AUGUST 26, 1977 STRUCTURE, MOVEMENTS AND REPRODUCTION IN THREE COSTA RICAN BAT COMMUNITIES By Richard K. LaVal1 and Henry S. Fitch2 The ecological role of bats in tropical forest communities has only recently come to be appreciated by biologists and is still far from being fully documented or evaluated. In Costa Rica, varied physiography results in drastically changing rainfall and temper- ature regimes and consequently, a diversity of different floral and faunal communities within a relatively small geographical area. The adaptive patterns within bat communities to these various asso- ciations are of great interest in view of the important involvements of bats as insect predators, seed disseminators and pollinators within tropical communities. Ecological investigations of Neotropical bats, especially in Costa Rica, have greatly enhanced our knowledge of chiropteran biology and distribution there. Recent contributions include Gardner, et al. (1970), LaVal (1970), Fleming, et al. (1972), and Heithaus, et al. (1975). From February 1973 through March 1974 we gathered data on several of the most common fmgivorous and insectivorous spe- cies of bats at three sites in Costa Rica. Data were obtained for 78 of the approximately 100 species of bats known to occur in Costa Rica. 1 Organization for Tropical Studies, San Jose, Costa Rica (present address: Missouri Department of Conservation, Fish and Wildlife Research Center, Co- lumbia, Missouri 65201, U.S.A.). 2 Natural History Reservation and Department of Systematics and Ecology, The University of Kansas, Lawrence, Kansas 66015. 2 OCCASIONAL PAPERS, MUSEUM OF NATURAL HISTORY Materials and Methods Objectives and Techniques. — This study is a segment of a larger project, the objectives of which were to correlate seasonal cycles in the reproduction, food habits, and behavior of Neotropical insect predators with environmental cycles as reflected by plant phenology and changes in prey abundance. In the present paper, we have attempted to resolve the following questions: How are bats dis- tributed among three major habitats at our study sites? How does community structure at three sites of contrasting climate and vege- tation differ as indicated by diversity indices and niche matrices based on size and feeding behavior? How far do bats move in their foraging activities, and how are these movements related to food habits and to the real size of activity ranges? Finally, to what extent are seasonal climatic changes reflected in the reproductive strategies of bats? Data were obtained from three sites in Costa Rica — La Selva, Monteverde and La Pacifica. Ten- and 14-meter mist nets and 1.7 m X 1.5 m harp traps (Tuttle, 1974a) were erected to block trails through undisturbed forest, second growth forest, and plantations of bananas, coffee, palms, and cacao. The traps were in regular use only at La Selva and Monteverde. Mark, release, and recapture techniques were employed, using the metal or plastic bat bands issued by the National Fish and Wildlife Laboratory or purchased from A. C. Hughes, Hampton Hill, England. At La Selva, nets were set in a different location on each night of sampling. Netting activity was alternated between the western and eastern halves of the area, with distances of up to 2300 m be- tween netting sites. The trap was erected on the same trails, but usually on the half of the area opposite the half where the nets were set. Owing to the patchwork character of the forest, coverage at Monteverde was spotty. Most of the netting sites were clumped within a circle of one kilometer diameter. The greatest distance between any two sites was 1500 m. Our data from La Pacifica, COMELCO, and other sites in Guanacaste do not lend themselves to the elucidation of movement patterns. During the study period 1184 bats of 52 species (excluding Molossus sinaloae at a roost site) were banded at La Selva, and 613 bats of 20 species were banded at Monteverde. An additional 678 individuals of 35 species were captured in Guanacaste. The biases inherent in some of these techniques were discussed by LaVal (1970) and Fleming, et al. (1972). Use of the harp trap in the tropics has been mentioned only briefly in the literature (Tuttle, 1974b). The trap was designed specifically to catch bats of the genus Myotis (M. D. Tuttle, pers. comm.) but is effective for a variety of other small bats up to the size of Carollia brevicauda COSTA RICAN BAT COMMUNITIES 3 (15-20 g). Larger bats such as Trachops cirrhosus (30-40 g) and Artibeus jamaicensis (40-60 g) seldom were caught. Some of the species trapped apparently were adept at avoiding mist nets and thus have been poorly represented in previous studies. For ex- ample, Myotis riparius ( La Selva ) and Myotis keaysi ( Monteverde ) were among the bats caught most commonly in the trap, but they rarely have been taken in mist nets. Though not effective in heavy rain, the trap seems to be superior to mist nets in light rain because mist nets eventually become waterlogged, resulting in a decline in catch-rate. Bats also were captured in a variety of other ways, mostly by hand in hollow logs and trees. Several species were caught only in a mist net set outside a Molossus sinaloae roost in a building, and another building served occasionally as a roost for two species of Myotis. Bats were preserved in alcohol when specimens were needed to confirm field identifications or to examine reproductive tracts, and are deposited in the Museum of Natural History, The University of Kansas. All other individuals captured were released after record- ing sex, age class (juvenile, subadult, adult), reproductive condi- tion, forearm length and weight. Fecal samples containing insect remains or seeds were retained. The names used for bat species follow Hall and Kelson ( 1959 ) except in cases where more recent taxonomic revisions exist. At La Selva, the three species of Carollia with which Fleming, et al. (1972) encountered difficulty in Panama (castanea, brevicauda and perspicillata) can be readily separated in the field (see Pine, 1972). Also at La Selva, Myotis riparius and M. elegans broadly overlap in several external characters that supposedly separate them (see LaVal, 1973 ) , and we could find no one set of characters that would ensure correct field identifications. The best single character ( often requiring the use of a hand lens ) is the relatively crowded condition of the small upper premolars in M. riparius. Artibeus phaeotis and A. watsoni at La Selva appear identical externally. The individuals lacking the third molars were identified as A. phaeotis and those having them as A. watsoni, following Davis (1969). Hylonycteris underivoodi and Choeroniscus godmani were difficult to separate at La Selva and Monteverde, but the bicolored fur of Choeroniscus usually distinguished it from Hylonycteris, which has tricolored fur. At Monteverde all small Artibeus were identified as A. toltecus. Although there is a possibility that a few were actually of the very similar species, A. aztecus, all preserved examples proved to be A. toltecus. Study Areas. — The three study sites are characterized by con- trasting climate and vegetation: La Selva, ca. 100 m elevation, Tropical Wet Forest (Holdridge, 1967), in the Caribbean lowlands; 4 OCCASIONAL PAPERS, MUSEUM OF NATURAL HISTORY Monteverde, 1400-1600 m elevation, Premontane Moist and Pre- montane Wet forests, on the Pacific slope of the Cordillera de Tileran; La Pacifica, COMELCO Ranch, and nearby sites, ca. 100 m elevation, Tropical Diy Forest, in the Pacific lowlands of Gua- nacaste Province. The La Selva site has been described by Orians (1969, and see references in Fleming, 1974), and the La Pacifica site by Fleming (1974). La Selva and COMELCO were described by Frankie, et al. (1974). Monteverde is a dairy-farming commu- nity on moderately-sloping terrain near the continental divide. Pas- tures typically are separated by strips or tracts of forest, mostly undisturbed, amounting to about 30 percent of the total land area. The community is bordered on two sides by the Monteverde Cloud Forest Preserve and the Monteverde watershed, which together constitute about 1551 hectares of continuous, undisturbed forest. The dessicating effects of the harsh Pacific diy season are moder- ated by low temperatures (Table 1), higher rainfall (Fig. 1), cloud cover, and mists blown over the divide from the wet Caribbean slopes (see also Rentz, 1975). As a result, the forest is largely ever- green and supports many faunal and floral elements more typical of the Caribbean slopes and lowlands. The remainder of the biota consists of a mixture of montane species, species characteristic of the Pacific lowlands, and species of wide altitudinal range. Netting and trapping were carried out in or near undisturbed forest at La Selva, in both disturbed and undisturbed forest at Monteverde, and in riparian forest at La Pacifica and at the nearby COMELCO Ranch. Nets and traps were set at ground level across trails. In the La Pacifica area, netting efforts were concentrated on a few insectivorous species because of the extensive amount of data already compiled there on other species by Fleming, et al. (1972) andLaVal (1970). At all localities, netting and trapping periods were confined to about eight to ten nights a month, with the trap set each night but the nets only on the most favorable nights. Scheduling problems and periods of unfavorable weather reduced the actual number of sampling periods to ten at each site. Table 1. — Temperature (°C) and Rainfall (mm) at Three Sites in Costa Rica from Records of the Costa Rican Meteorological Service; Ranges Given in Parentheses. La Selva Monteverde La Pacifica X monthly temperature 24.1(22.8-24.9) 14.9(14.1-15.6) 27.8(26.2-29.0) X annual rainfall 3961(2921-5659)2482(1716-3180) 1528(959-2267) Months of dry season Feb.-Mar. Jan.-Apr. Dec-Apr. X monthly rainfall during dry season 172.5 47.5 21.3 COSTA RICAN BAT COMMUNITIES ui 5 500-1 400- 300 200 100- La Selva r J F M A M J J A Fig. 1. — Rainfall at three sites in Costa Rica. Monteverdi: La Pacifica Usually, nets were left in place until about 2200 hours. The trap was left out all night, and checked at about 2100 and again at about 0600 hours the next morning; its effectiveness sometimes was re- duced by heavy rains. Typically eight to twelve nets were set diag- onally across trails. Often nets were set end to end; equally often, they were spaced with one to three empty net-lengths between. A total of 261 net-nights were tallied for La Selva, and 189 for Monte- verde. The trap was set for 76 nights at La Selva, 58 nights at Monteverde (Table 2). At COMELCO a few nets were used in special situations: two or three nets each, once a month, at a small cave below a large waterfall; over water at a ford; and two nets at the entrance to the Cueva del Tigre, near the Palo Verde Field Station. The Palo Verde netting effort was abandoned after Ptero- notus parnellii ceased using the cave. At La Pacifica the trap was set occasionally, mostly during the dry season. On a few dates dur- ing the diy season, two to four nets were set over water at the nearby Rio Abangares and Rio Congo. Acknowledgments. — Margaret L. LaVal provided invaluable as- sistance in every phase of the study. J. Campabadal and the staff of the Organization for Tropical Studies supplied excellent logistical support. We are grateful to the W. Hagnauer family (La Pacifica), the D. Stewart family (COMELCO), and the many landowners of Monteverde (especially the D. Green family, the W. James family OCCASIONAL PAPERS, MUSEUM OF NATURAL HISTORY s o 3 'o ~ 00 m CO ■4-J in CM CO 03 Q «1 >.T3 c« JS O CD „ -c eg '5 .SO ? 3^ CT> CO CO « s£ P4 05 a,; >- m §• a CO CO CO Cv i—, CD CO CO 3 d d c & C 05 -r t- »£3 O s cp ^2 H"0 « sU i— i : CD ~* t/3 "os CC_r 93 "3 c. ^ bC^ C,f>u CD E 4-> M < a c O CO OJ i-H 1> c •3 s .S 3 U 01 o Z CD 5 s H E u =^ CO >. o < 60 2 TF CO in co . v> CD •4-* WO a*B £, a S ft V C C CD 05 CD 05 X *j U C '— c § a 4-> V3 0) 4-1 co 4-* s 4-1 u O y. 03 - fe •= -C P H 4-J 4-* C/5 C in c 4-* CD *o CC S V T3 4-) &H d - 03 2 4^ CD ^ *^ 03 Qj OS ft > Q CO 03 o c c CD '•Z C OS CD t« CD < "3 ■ fH o CD "ft o £ £ Cu ft W3 H X 1 CD C5 *c/: 1 T3 a OS "33 CD > CD cc c^ OS CD "2 M < CO 05 c — CD C H J s J — COSTA RICAN BAT COMMUNITIES 7 and M. Moss) for their warm hospitality and permission to conduct our studies on their property. T. H. Fleming generously made available to us his records of bats banded at La Selva in 1970-71 and provided a critical review of an early draft of this paper. L. de la Torre and J. Pefaur assisted in preparation of the Spanish sum- mary. Thanks also go to the numerous visiting OTS researchers and students who assisted us in the field at La Selva. Financial support was provided by NSF Grant No. GB25592 to Henry S. Fitch. Results and Discussion Distribution Among Habitats. — At La Selva bats were captured in several different habitats, as follows: primary forest, cacao groves (shaded by scattered old-growth forest trees) and second- growth forest; whereas, at Monteverde bats were captured in pri- mary forest, undisturbed riparian forest and second-growth forest. Netting and trapping success was evaluated for the more common species in each of these habitats at each site. At La Selva, primary forest accounted for more than half of all net- and trap-nights, whereas at Monteverde second-growth forest accounted for more than half. Relative overall abundance in the habitats is indicated by num- bers of bats caught per net-night and trap-night (Table 3). Trap- ping success at La Selva and both trapping and netting success at Monteverde were highest in second-growth forest. Netting success at La Selva was highest in cacao, primarily because of the large numbers of Carollia taken there. Preliminary examination of seed- loads carried by Carollia shows that Piper, a common early suc- cessional plant along trails in disturbed areas, was the major food item in the diet of these bats. It is hardly surprising that more bats were caught at ground level in disturbed habitats, simply because the canopy is lower and thus ground level nets sample a higher percentage of the flyway area through which bats must pass. Fur- thermore, disturbed habitats usually contain concentrations of Piper, Cecropia, Solatium, and other plants frequently used by frugivorons bats; whereas food sources in the primary forest, though numerous Table 3. — Mean No. of Bats Captured Per Net-Night and Trap-Niyhl in Three Habitats at Two Costa Rican Sites. LA SELVA MONTEVERDE net trap net trap Primary Forest 3.92 5.01 3.74 5.82 7.63 10.33 Primary Forest .... 1.93 ' .VI Cacao Riparian Forest Second-growth 2.65 3.85 3.45 Second-growth 6.27 OCCASIONAL PAPERS, MUSEUM OF NATURAL HISTORY and diverse, have a patchy distribution, both horizontally and ver- tically. It should not be concluded, however, that the actual density of bats is greater in disturbed forest, as opposed to primary forest; our observations lead us to suspect that the converse is true at these two sites. Foliage-gleaning bats of the genera Micronycteris, Trachops, Mimon, Tonatia and Vampyrum were taken mainly in primary for- est. This is surprising because the large arthropods and small vertebrates on which they feed appear to be quite common in cacao and second-growth. Other species for which data are adequate seem to be equally distributed in all three habitats. At Monteverde, bats of different species seemed to be evenly distributed among the habitats, except that 91 percent of the Myotis keaysi captured were in second-growth. Movements. — Two-hundred and fifty-one bats of 22 species were recaptured at La Selva (23.9%) and 93 bats of ten species at Monte- verde (15.9%). This rather substantial difference in recapture per- centage is probably a function of both the relatively spotty netting coverage at Monteverde and the difference in species composition of the bat faunas at the two sites, i.e., Carollia brevicauda and C. castanea, with their relatively high recapture rates, were the most common bats at La Selva, whereas Sturnira ludovici and Myotis keaysi, the two most common species at Monteverde, had lower recapture rates. At La Selva recapture rates ranged from a high of 75 percent for Mimon cozumelae to a low of 1.8 percent for Furi- pterus horrens (Table 4). At Monteverde rates varied between 36.4 percent for Carollia brevicauda and 3.8 percent for Myotis nigricans (Table 5). The species with five or more individuals recaptured (Tables 4 and 5) were subjected to chi-square tests, separately for the two localities, and were found to differ significantly in percent- ages of recaptured individuals (X2=16.29, 0.05>P>0.02 for La Selva; X2=10.73, P<0.001 for Monteverde). Though the two sites shared relatively few species, it is interesting to note that only in the case of one species, Carollia brevicauda, were the recapture percentages of similar values. For Glossophaga commissarisi, Ptero- notus parnellii and Myotis nigricans there were wide discrepancies. Mean recapture distances at La Selva (for species with three or more recaptures) ranged from 750 m (Desmodus rotundus) to 133 m (Myotis nigricans), whereas at Monteverde they ranged from 650 m (Glossophaga commissarisi) to 163 m (Artibeus toltecus). Mean recapture time intervals for La Selva ranged from a high of 281.3 days for Mimon crenulatum to a low of 49.4 days for Mimon cozumelae. At Monteverde the high was only 104.8 days (Vam- pyrops vittatus), the low 59.3 days ( both Glossophaga commissarisi and Artibeus toltecus). At La Selva, mean recapture distances, as well as mean recapture intervals, differed significantly among the COSTA RICAN BAT COMMUNITIES 9 species when subjected to an analysis of variance (F=2.26 and 2.17; P<0.01). At Monteverde the same was true (F=3.13, P<0.01; F=2.18, Pr=0.05). However, the grand mean recapture distance for La Selva was found not to differ significantly from the corre- sponding distance at Monteverde (t=0.37; P>0.5). Likewise the grand mean recapture time did not differ significantly between La Selva and Monteverde (t=1.33; P=0.2). A Mann-Whitney U-test gave similar results (U=39 with critical value of 17; P>0.05 for distance; U=79 with critical value of 17; P>0.05 for time interval). For the four species common to the two sites, mean recapture dis- tance and time values were disparate, excepting mean recapture time of Carollia brevicauda (Tables 4 and 5), which differed by only two days at the two sites. The significantly higher recapture rate at La Selva (23.9%, as opposed to 15.9% for Monteverde and 12.9% for La Pacifica) suggests that the behavior of the bats must expose them to nets and traps more frequently at La Selva than at the other sites. This behavior may be related in part to the year-round availability and diversity of food sources at La Selva. Thus, the less seasonal climate at La Selva may necessitate fewer seasonal movements there than at Monteverde or La Pacifica, resulting in fewer individual bats leav- ing the study area to search for food. If this is true, one would expect a higher number of multiple recaptures at La Selva than at Monteverde and La Pacifica. Such proved to be the case, as the average numbers of times recaptured was 1.55 at La Selva, 1.38 at Monteverde, and 1.35 at La Pacifica. From the conclusions of Fleming, et al. (1972), or on the basis of time required to move, one would predict mean recapture per- centage (and perhaps also mean recapture time interval) to be inversely proportional to mean recapture distance. Further, we would expect to find large negative correlation coefficients when the grand means from Tables 4 and 5 are subjected to correlation analysis (except for recapture distance and recapture time, which should be positively correlated). Such is not the case; none is sig- nificant at the 5 percent level at either La Selva or Monteverde. Therefore, we cannot lend support to the hypothesis of Fleming, et al. (1972), that recapture percentages should be inversely related to mean recapture distance, based on these data. The mean recap- ture distances found in this study are in many cases different from those found for the same or related species by Fleming, et al. (1972) or Heithaus, et al. (1975) at La Pacifica. Although these differences undoubtedly result partly from differences in the design of the studies (for example, in widely spaced vs. closely adjacent net sites, and in use of the bat trap), it is possible that they also reflect real behavioral differences among populations of the same or related species at the three sites, which in turn reflect differences in 10 OCCASIONAL PAPERS, MUSEUM OF NATURAL HISTORY CO - -t_» CD t. 03 1/5 "^; 3 *"0 tc ^ a * ' §"3 S"| J S r£ fc a o 03 r- Q 13 < 3 03 Q, C ™ £ u CD 'O 3 (LI 3 cd 3 — CI) a 03 Oh « TJ 0) o lH 3 4-1 c a c« X u m cd ;. « +j ^< <: T3 cd ~ T3 •M _ Q 03 ;- PC CO 1— 1 ** "# l—i CM 02 CO "tf 1— 1 o CM CO co CO CO -— - rp -~- LO 00 00 CO i-H -^ i— I OS I _ I , , CD CO - — . i— i ' — » oa . — - CO , — * \r. , — . Tf , — „ i— i , , CD o "*. oa CO in CM CO CO © t^ O LO CI oa CD CO — 1 oa CM i ^H i— i i CO 1— 1 1 t^ co oa CO 1 00 co ■* CM — CO ^r O oo oa 00 CM oa CD CD — 1 co l^- o i-H o ^~^ CM CM t- ""*~" " V ' v-' CO 1—1 — 1 o © in c cm CM -^ -— ' O c O CM o -— o ^ CM co «? L0 o io 1> --' O ^-. 00 °° i CM o ~~v 00 ^^ o CO c lO in O CO o o O CM C ^ co ^^ o 2 in i— i , CO CO o ^ CM in ^h co CM CO CD CM O o c o o o CD o c o o CM o CO CO 00 o CO CM t- o o 00 CD i-H in o CD c CM T}H -, CM " 00 CO ■* CO CD CO CD i-H t^ CM in i-H CD CO CO CM oa co CD co CD 00 oa CM "* a o y CJ O O s o o s o &H oc ir; 00 -v CD r" o CM T 1 — 1 l- CM l> co oa •"- cc ■««* ~ a <* .00 •^ "5 Q HA «3 5 _2 5 -5 s s ~2 _Cj o 2 3 a a ^D a i^ g 61 t£ ~ a CO Q CJ u -C ~s CJ ■~- CO e a q ^ .2 B 53 s 5" o "C^ -^ £*j ^ CA Co — £; CO Co ca O Q ~i »^ o o k — . s u o * o w o « « o COSTA RICAN BAT COMMUNITIES 11 © (M co CM 'T co o © CO fM co -^ — in CO ^ W ^ Tf t> co co , . ""7 i-H °? CO CO 00 O t^ CO -~- CO i-H ^ CM --i CO CO fM CO "7 eq ^ in 0,' ©" co co i— i b- co co • — ■ ■— - co ■— ' CM O © ^-. ,-v H2o o Tin io o co >— i in —* •— ' — i co fM CO CO © co CO © © © © © © o © in © © o © CO © fM i-h io in co co «aj i- © © oq cm CD <* co © in co i— i ^H r- © CM -t CO cm CM © co CM fM -H "tf CO i—i in in (M in fM > i/. CO CO 6 © © CO fM £ "5b ID %* 03 s 03 ID 4-< s o 03 PC d o 03 03 Q ID •— 5 a 03 O V PC I > 03 31 ■M a 03 03 O 0) > PC ID -W c H* ^^ C« I) >> 03 £ T5 a 03 O 13 ID C PC — PC H3 4-J ID <-l t- M D p O a ID 03 a O V PC — o OS K o 03 o 0) X u_» tz< £ H3 ID ~ 13 4-1 3 o 03 H PC 00 in CM TP o CD 1—1 CO >—l CO I- os ^- in -— CD CSI CO CO C-l N -^ in i - © ^-- in CM O ^ © 00 ■* w o^ inn ■* j "t t± ^ © "7 ©' "7 o t> in rt< i— i cd i — i i — i com ©^ i—l — - — - — - — ' — ' 00 ^ 05 CO ^ °0 """ _ l> i 05 CO CM ^~- CD in o o o o o © ©_© oooo in ^wT(©o^cMr7'©^cot7^H©F7cD on omin © —< ooo oco oos o o © o © i— I CDCO ■<# CM i— I COOO i—l © ^ -~- ^ ~~^ ^ -^- -^ ^ -w co 05 CI ■<# —I t> 00 CO oo CM o © © © CM © rr CM "* © © © o co co © © 00 CO © i—l CO -* 1— 1 © 00 © in 00 © © CM CM i—l CO © CO CO 1—1 CM r~ 00 r— 1 O CO CM i—l co in i-H CM ~H c-1 CO CO o o (D bo 03 i- 3 O a o o 03 be 03 a o CO o s « CO CO CO d <~ 03 as 1 .2 O !-. -3 03 u u * co co o 00 CO o © © 1-1 "tf CM lO 00 © CM CO © oo in > o 3 C 3 •4-t 00 o 3 O > a 3 « il ID i - 03 M a >> £ •* o 3 O 0.05) and much smaller at La Selva (r= -0.09, P>0.05). It seems probable to us that increasing diversity in food habits and foraging behavior will lead to decreas- 14 OCCASIONAL PAPERS, MUSEUM OF NATURAL HISTORY ing correlation between range and body size, as has been observed at La Selva. Whatever the pattern, the problem deserves further study. The exceptionally high recapture rate and high mean recapture distances of the two species of Mimon, which are probably foliage- gleaning insectivores, are unexplainable given present knowledge, but must reflect some unusual behavioral traits. Extremely low re- capture rates in several species, especially Tojmtia bidens, Artibeus famaicensis, A. lituratus, Vampyressa pusilla, Ectophylla alba and Myotis nigricans ( Monteverde only) suggest that these species either may be nomadic or else may be unusually proficient at avoid- ing mist nets and traps on their second encounter. The recapture percentage of Furipterus horrens was lowest of all, but most of the individuals were caught by hand in a log. None was taken in mist nets, and few were caught in the trap. At La Selva, of 79 bats banded by T. Fleming in November and December 1970 and January and February 1971, 14 (17.7%) were recaptured at least once in this study, compared with our re- capture rate of 23.9 percent for the bats banded in this study. The recapture rate for Fleming's bats would have been much higher had it not been for the large numbers of Artibeus jamaicensis and Desmodus rotundus he banded. Only two Desmodus and no Arti- beus were recaptured. The recaptured bats were still within several hundred meters of the original banding site after three years. On 30 December 1975, on a brief visit to Monteverde, eight nets were erected at a site in undisturbed forest. Four (11.8%) out of 34 bats netted were recaptures, similar to the original overall re- capture rate (Table 5) of 15.9 percent. These results suggest that many of the bats banded in 1973 still survive in the same general area after nearly three years. Community Structure. — Data on the numbers of species en- countered and numbers of individuals netted or trapped are com- pared in Table 2 with similar data from La Pacifica. The diversity in these three communities is compared using an index of species diversity based on information theory, H'=-^Pik>g,.pi, and the "eveness" or equitability component of diversity, ]=H' /Hmax (Pie- lou, 1969). Capture-rate is given for La Selva and Monteverde only. The species composition and relative numbers of individuals in our study were augmented by use of the trap in addition to mist nets. Thus Peropleryx kappleri, Rhynchonycteris naso, Centro- nycteris maximiliani, Micronycteris nicefori, M. brachyotis, Barti- conycteris daviesi, Macrophyllum macrophyllum, Lichonycteris obscura, Furipterus horrens and Myotis elegans were caught only in the trap (of the two means of capture) at La Selva, and Ptero- notus suapurensis and Myotis oxyotus at Monteverde. Furthermore, the trap captured over 50 percent of the individuals of each of COSTA RICAN BAT COMMUNITIES 15 these additional species at La Selva: Saccopteryx leptura, Ptero- notus parnellii, Micronycteris megalotis, Mimon crenulatum, Choer- oniscus godmani, Hylonycteris underwoodi, and Myotis riparius, plus three more speeies at Monteverde: Pteronotus parnellii, Mi- cronycteris megalotis and Myotis keaysi. Had the trap not been used, diversity indices calculated From these data would have been substantially lower. We believe that our indices are therefore a more accurate measure of bat diversity than previously published indices based solely on mist netting data. The high species diversity of La Selva is clearly indicated by the high value of //'. Also, the higher / value for La Selva indicates that individuals are distributed more uniformly among the species. If animal diversity is correlated with structural diversity of the en- vironment, then the high W values for La Selva and the compara- tively low H' values for Monteverde and La Pacifica should be reflected in high or low structural diversity. Humphrey (1975:321) outlined the causal chain for this relationship as follows: ". . . struc- tural diversity results in diversification of food resources, permitting invasion by new species, specialization or generalization of feeding habits regulated by the resources' patchiness, and finally a decrease in niche breadth and increase in competitive exclusion." The fact that Tropical Wet Forest, as exemplified by the forest at La Selva, exhibits significantly higher structural diversity than Premontane Moist and Wet forests (Monteverde) or Tropical Dry Forest (La Pacifica) is unchallenged, to the best of our knowledge. Although the structural differences have not all been quantified, the greater structural diversity at La Selva was obvious from this study and from numerous other studies, published or in progress. The in- crease in bat species diversity and concomitant decrease in niche breadth is apparent not only from the diversity indices, but also from the niche matrices (Table 6) discussed below. Another indi- cator of differences in diversity is the fact that at La Pacifica and Monteverde most of the individuals captured belonged to 4 and 3 species (Appendix I), respectively; whereas at La Selva most be- longed to 7 species. Bats of the genera Carollia and Artibeus were common at all three localities. Niche matrices were constructed for La Selva and Monteverde following the reasoning of McNab (1971) and Fleming, et al. (1972). We used length of forearm as a crude index of body size, although weight would have been a better index if it had been available for all species. For example, Artibeus jamaicensis and Pteronotus parnelli have forearms of 64 mm and 62 mm, respec- tively, whereas their weights are 55 g and 24 g, so they are sepa- rated into different cells only by food habits if forearm length is the size criterion. Similarly, Carollia brevicauda (FA of 41 mm) is in the same cell as Artibeus phaeotis (FA of 39 mm) on this basis. 16 OCCASIONAL PAPERS, MUSEUM OF NATURAL HISTORY Table 6. Distribution of Bat Species by Body Size (As Estimated by Forearm Length) and Food Preferences From Three Costa Rican Localities. Mean Forearm Length ( mm ) 30-34 35-43 44-54 55-68 69-86 87+ TOTALS La Selva insects-aerial 4 6 4 3 17 insects-foliage- gleaning -17 2 3 13 fish 1 1 flesh 1 1 2 plants-fruit 2 7 4 11 15 plants-pollen/nectar 4 1 5 blood 1 1 omnivore 1 1114 TOTALS . 12 21 10 10 3 2 58 cells occupied 5 4 3 6 3 2 23 MONTEVERDE insects-aerial 14 3 1 9 insects-foliage-gleaning 1 1 fish 0 flesh 1 1 plants-fruit 4 12 7 plants-pollen/nectar 2 2 1 5 blood 0 omnivore 1 1 TOTALS 4 11 5 3 0 1 24 cells occupied 3 4 3 2 0 1 13 La Pacifica insects-aerial 2 3 3 2 10 insects-foliage-gleaning 14 5 fish 1 1 flesh 1113 plants-fruit 5 111 8 plants-pollen/nectar 12 3 blood 1 1 omnivore 112 TOTALS 5 14 4 6 3 1 33 cells occupied 4 4 2 5 3 1 19 but their weights are quite different ( 19 g vs. 12 g ) . Although the bias inherent in the use of forearm length is substantial in these and certain other cases, it does not entirely negate the value of the niche matrices and must be used until weights for the rarer species become available. Niche matrices for La Selva, Monteverde and La Pacifica (La Pacifica data modified from Fleming, et al., 1972) are shown in Table 6. The La Pacifica matrix is augmented by the addition of 5 species caught there in the trap (that were not taken by Fleming, COSTA RICAN BAT COMMUNITIES 17 et al.) — Saccopteryx leptura, Glossophaga commissarisi, Choero- niscus godmani, Hylonycteris underwoodi and Myotis elegans. Also included is Vampyrum spectrum, a group of which roosted in a hollow tree on the bank of the Rio Corobici near La Pacifica. The assignment of species to food categories is based on a va- riety of information. In this study fecal samples were collected from as many species of bats as possible. Although analysis of these samples is incomplete, the presence of insect remains or seeds sug- gests an insectivorous or frugivorous diet. The division of insecti- vores into aerial and foliage-gleaning is based on morphology, plus field observations of several species, in the present as well as past studies by LaVal. A detailed food study of one foliage-gleaner (Micronycteris megalotis) was carried out. The consumption of fish, flesh, and blood by some bats is well documented both in the liter- ature and by our unpublished observations. The separation of plant-feeders into frugivores and nectarivores is based on fecal samples, morphology, and the literature. As shown by Heithaus, et al. (1975), many frugivores use pollen and nectar (at least season- ally), and at least one nectarivore (Glossophaga soricina) also eats fruit seasonally and probably should be regarded as an omnivore. Many other species of pollen- and nectar-feeding bats exhibit a greater degree of morphological specialization for nectarivory than Glossophaga, and probably they will prove to have correspondingly specialized habits (as in Mexico; see Alvarez and Quintero, 1970). Although most of the species occurring at two or three sites are assigned to the same food category, we recognize that detailed studies of feeding behavior may establish that some species have different food habits at each of the sites. Of the 48 cells in our matrices, 23 were occupied at La Selva, 13 at Monteverde, and 19 at La Pacifica. La Selva differed from La Pacifica primarily in the presence of more species per cell. Additionally, quite a few more species of foliage-gleaning insecti- vores lived at La Selva. Monteverde differed from the other local- ities in having fewer species per cell and in having zero or only one cell for foliage-gleaners, piscivores, carnivores, sanguinivores, and omnivores. Among pollen- and nectar-feeders Monteverde had as many species as the other sites and more occupied cells. At all three sites, greatest occupancy is in the 35-43 mm insect-eatinu cells, followed by the 35-43 mm plant-eating cells, suggesting either a finer degree of niche separation among these species or simply a greater variety of potential niches available to bats of that size range. Although it seems amazing that 13 species of insectivorous bats can co-exist in the 35-43 mm cell at La Selva, most of these species have strikingly different morphologies, and, insofar as ob- served, they differ substantially in feeding behavior, especially in terms of vertical stratification. However, certain species pairs, such 18 OCCASIONAL PAPERS, MUSEUM OF NATURAL HISTORY / C E U !C — ^ rt w > 13 5 Q c w s o n3 eg 0) o #~ cd Sh - j X j Jh o n V — W3 H es — O - « £ — M •— - o -e 5 5j .w — c 2 ID C ^ — i s .5 * 0) — M C/. — 1 ShJ ' — CQ - c ^ ni - 2 2 < <"rt w "2 rt u 8 OJ '-m •— * '-' v. +J .s s u -c u ns £ — C/2 fl o^ c 5< y: ^ •J- ;r 0 J£ W r~ in o Z — 0 •— J - - u ^f PC c i^ - - - J < - s 4-< La o a c/) CM O" © c. (M 05 —I \ i— I l- "V. r-4 O CO CM o in i-H — © -H < W3 jH < -J \ en — '© —H — I cm o o O i-l O i— I O i-H -h CM C/5 J < r* M J \ t- Ph ►— , C/5 > i— .\PL, _T» O-H-Hi-4-hO-Hq^ — i if) i-i'*! h co © i— I i—i © O co cm ^r co CM -H CM fM CM CO i— I CO -H CO H in © ^ CO CO © © o —I CO \ "v. C -H \ © © — I C/3 C/5 o O C co c o -=: u CC <3 3 C_ 4) 5j 5 < Q i^- a 550 o cyT 2v C75 -J C/5 CA CT) C/5 cc ^ — — •— ^ •**» ^ ^ Hot u ^ O 1—* _x t> o tS a s o 5 •S Q U 53 « a icr W2 ( ^i u fi fc£ ,5 ^- o '■-* ~ 53 .2 .2 u <— 05 a o "5 •4o> C 5 o *— e g o o B-4 o Eh <5 § COSTA RICAN BAT COMMUNITIES 19 CM —. H CO H ^ CM \ PL, >— -v.. "v \ CD ,-H ,-H r-l _| rt © _ CO « CI pJ > \ Ph © ,-H ^H -H CI CM "v. C (M "t CO J <:\\pl, i— I r- © ^H CM J \ Ph CM c o ^H CI Ph 10 ^ ci C cl © O-l \ J CD i— 1 CI < CD i— 1 ^ O CM o Pi i— I CD 1— c i— 1 f-H I— 1 CM f— t CO CM << CD ^ i-H CM I— 1 l-J .-H \. Ph \. hJ C i— i © i— i \PH >— > i-H Cl ^H CM ^H M 1 ^ ^ \ J Ph CD CI I>1 H n O co l> < "-v CD O r-H CM CM — I < h« \ CD vS O CM o O -t © ^H O .-H )C co CO CM CM CD CD < S hJ CD CD CD CDCD CDCD — CD CD CD o o o 8 hJ o •2 § 5 w -.- CD 3 4j co o — - 4i J ~ U5> O *fc3 63 •5 = =o o =o « 3 -2 C 3 2 5 - A; 3 ~ a tH 3 -3 W B a -p: — -_ -— _ o o is O -2 20 OCCASIONAL PAPERS, MUSEUM OF NATURAL HISTORY as Myotis riparius and M . elegans, as well as Micronycteris brachy- otis and M. nicefori, are so similar morphologically that a careful study of their food habits and feeding behavior is needed to deter- mine how their niche requirements might differ. That similar-sized bats do indeed differ in food habits, at least in the relative amounts of the different species of plants utilized, has been established at La Pacifica (Heithaus, et al., 1975), has been suggested for a num- ber of other Costa Rican species (Howell and Burch, 1974), and will probably also prove to be the case at Monteverde and La Selva, as suggested by the preliminary examination of seeds collected from feces of frugivorous bats in this study. Reproduction. — Only three of the four different reproductive patterns described by Fleming, et al. (1972) were observed in this study, at least among adequately sampled species. Seasonally monestrous: Pteronotus parnellii seems to conform with this pattern, at least in Guanacaste (Table 7), as does Myotis riparius at La Selva. Seasonally polyestrous: At La Selva this pattern included the three species of Carollia (Fig. 2), plus Artibeus phaeotis and Glos- sophaga commissarisi (Fig. 3). At Monteverde, Sturnira ludovici, Artibeus toltecus, and Myotis keaysi (Fig. 4) were seasonally polyestrous. Although the generalizations about bimodal birth peaks made by Fleming, et al. (1972) seem to apply to all of them (ex- cept perhaps Myotis keaysi) the actual timing of the peaks varies one to three months among the different species, perhaps in re- sponse to different fruiting and flowering periods among their food plants. All species seem to exhibit minimum reproductive activity7 (no pregnancies) late in the wet season, from October to early January. In Carollia brevicauda at La Selva this nonreproductive period is relatively brief, but in Sturnira ludovici and Artibeus tol- tecus, in the dryer climate of Monteverde, it is relatively long. Although no data were obtained for the latter two species during February of the study period, examination of bats captured on 10 February 1977 revealed that 21% (6 of 28) female S. indovici and 50% (13 of 26) female A. toltecus were pregnant. Table 7 suggests that the seasonally polyestrous pattern is com- mon among the frugivorous stenodermine and the nectarivorous glossophagine bats, but less frequent in the foliage-gleaning phyl- lostomine bats and molossids. Nine species of phyllostomatid bats south of the equator in Brazil (Taddei, 1976) were also seasonally polyestrous. Year-round activity: Tentatively, Molossus sinaloae (Fig. 3) at La Selva has been placed in this category, but there is a three- month gap in our data. Occurrence of pregnant females before the gap and the capture of juveniles following the gap suggest that there is some reproductive activity throughout the year; but most COSTA RICAN BAT COMMUNITIES 21 Co ro 1 1 i o brevicgud 23 ."■ 19 7 II 17 24 28 24 37 36 68 15 21 42 57 46 49 wc-v^ C o r o 1 1 i g cast g n e g 9 4 2 9 6 6 4 19 22 M 100-1 ,8 ^ ll 26 23 22 8 12 55 68 42 90- 80- / \ 70- / \ 60- / \ 50- 40- / >*s /-V 30- 20- 10- / • /' o-i I ■ ■ -Jf- —i- k."^-^/ Co ro I I ig p e r s p i c i I lo to 1 973 Months 1 974 Fig. 2. — Indicators of reproductive cycle timing of three species of CaroUia at La Selva, Costa Rica. The "dash-dot" line indicates percentages of adult females that were pregnant. The solid line indicates percentages of adult fe- males that were lactating. The "slanted-dash" line indicates percentages of all bats captured that were juveniles. The lower of the small figures below the species name is the total number of bats caught each month. The upper of the small figures is the number of adult females captured each month. pregnancies were in the late dry season and early wet season when females of most other species of bats we examined were also pregnant. These reproductive cycles are related to fruiting, flowering, and insect abundance (Fleming, et al., 1972) and will be dealt with in a future paper (LaVal and Fitch, MS). 22 OCCASIONAL PAPERS, MUSEUM OF NATURAL HISTORY Sturnirfl ludovi e i 23 12 27 16 35 28 x*V 19 33 32 50 9 22 M v o t i s kensyi 100-1 90- 80- 70- 60- 50- 40- 3 0- 20- 10- o- M 12 29 13 19 20 34 6 28 / v A 9 23 M \ / "X. ■ X ^ r* 5 18 O. N •—I A 1973 Months 1974 FIG. 3. — Indicators of reproductive cycle timing of three species of hats at La Selva, Costa Rica. See explanation in caption for Fig. 2. Summary Harp traps and mist nets were used to sample bats at monthly intervals for one year at three sites of contrasting climate and nat- ural vegetation in Costa Rica. Bats were examined, banded, and released in order to document community structure, distribution among habitats, movements, and reproductive strategies. Community structure at the Tropical Wet Forest site is charac- terized by high species diversity (if '=2.69) and high occupancy rate (48%) of niche matrices (body size plotted against food prefer- ences), with as many as 7 species per cell. At the Tropical Dry Forest site species diversity is lower (H'=2.07), as is the occu- COSTA RICAN BAT COMMUNITIES 23 100 -. 90- 80 70 60- 50- 40- 30- 20- 10- 0 Glo isophoqg com misso risi i 8 2 / 2 9 \ \ * 3 0 5 5 3 6 3 5 l U 7 8 15 12 / \ ^W^ 5 2 1 5 i 7 1 1 6 II 21 2 17 16 32 19 13" M olossus sinaloae 100- 90- 00- 70- 60- 50- 40- 30- 20- 10- 0- 22 20 18 16 6 5 25 27 25 28 19 13 ♦ +' * ~T 1 1 i 1 1 1 1 1 1 1 1 1 FMAMJ J ASONDJ FM 1973 Months 1974 Fig. 4. — Indicators of reproductive cycle timing of three species of bats at Monteverde, Costa Rica. See explanation in caption for Fig. 2. pancy rate (40%), with no more than 5 species per cell. At the Premontane Moist Forest site, both species diversity (H'=1.94), and occupancy rate (27%), are even lower, with no more than 4 species per cell. Bats were unequally distributed among habitats at the two sites where pertinent data were collected. At La Selva trap results showed nearly twice the catch-rate in second-growth, as compared to primary forest, whereas net results at the two sites differed little. At Monteverde trap data showed a somewhat greater catch-rate in second-growth forest, whereas netting results gave much higher catch-rates in second-growth than in primary forest. 24 OCCASIONAL PAPERS, MUSEUM OF NATURAL HISTORY Mean movements of banded bats were high at La Selva (332 m), lower at La Pacifica (258 m), and still lower at Monteverde (193 m). As there is no reason to believe some bats banded in this project did not move beyond the radius of the study areas during the study period, mean distances moved are minimal. Because this sequence is the reverse of the predicted sequence of mean distances moved, it is not possible to relate them to food habits and activity ranges in any meaningful way based on our data. However, recap- ture rates suggested that activity ranges are smaller at La Selva (recapture rate of 23.9%), as compared to Monteverde (15.9%) and La Pacifica (12.9%). All bats in this study except Molossus sinaloae at La Selva, were found to have seasonal reproduction, with the shorter, more sharply delineated reproductive seasons characteristic of La Pacifica, with its relatively brief wet season; longer reproductive seasons, as evi- denced by polyestry, were more typical of La Selva, with its very long wet season. Monteverde, with its wet season of intermediate length, was more like La Pacifica than it was like La Selva, in terms of reproductive cycles. Resumen Trampas de arpa y redes de nylon fueron usadas para obtener murcielagos una vez al mes durante un ano en tres lugares de clima y vegetation diferentes en Costa Rica. Los murcielagos fueron procesados, marcados, y soltados para documentar el movimiento, distribution ecologica, estructura de comunidad, y estrategias re- productivas. El area de forrajear, indicada por medio de las distancias de recaptura, fue mas grande en lugares de bosque muy humedo ("Wet Forest"). En los tres sitios, las areas de forrajear de las di- versas especies variaron en amplitud. El mejor resultado se obturo con las redes en las plantaciones viejas de cacao en lugar del bosque muy humedo, pero en el bosque montaneso ("Premontane Moist Forest") el resultado fue mejor que en los bosques secundarios. Con las trampas, el resultado fue mejor en el bosque secundario en dos lugares. La diversidad de la comunidad indicado por IV fue mucho mas alta en el lugar de bosque muy humedo. El numero de especies en el sitio de bosque muy humedo (58) fue casi el doble al de los otros dos sitios. Las matrices de los nichos ("niche matrices") basados en los tres sitios indicaron que 23 celulas fueron ocupadas en el sitio de bosque muy humedo, solamente 13 en el sitio de bosque montaneso y 19 en el sitio de bosque seco ("Dry Forest"). Tres modelos de reproduction fueron observados en este es- tudio. Pocos murcielagos insectivoros fueron "monoestros"; en cam- COSTA RICAN BAT COMMUNITIES 25 bio, la mayor parte de los murcielagos fmgivoros y nectarivoros y unas pocas especies insectivoros fueron "poliestros." Una de las especies insectivoras parece estar aetiva reproductivainente todo el ario. Literature Cited Alvarez, T., Gonzalez Quintero, L. 1970. Analisis polinico del contenido gastrico de murcielagos Glossophaginae de Mexico. An. Es. Nac. Cienc. Biol. Mexico, 18:137-165. Davis, W. B. 1969. A review of the small fruit bats (genus Artibeus) of Middle America. Southwest. Nat., 14:15-29. Fleming, T. H. 1974. The population ecology of two species of Costa Rican heteromyid rodents. Ecology, 55:493-510. Fleming, T. H., Hooper, E. T., Wilson, D. E. 1972. Three Central American bat communities: Structure, reproductive cycles; and movement pat- terns. Ecology, 53:555-569. Frankie, G. W., Baker, H. G., Opler, P. A. 1974. Comparative phenological studies of trees in Tropical Wet and Tropical Dry forests in the lowlands of Costa Rica. J. Ecol., 62:881-919. Gardner, A. L., LaVal, R. K., Wilson, D. E. 1970. The distributional status of some Costa Rican bats. J. Mamm., 51:712-729. Hall, E. R., Kelson, K. R. 1959. The mammals of North America. Ronald Press, New York, l:xxx-f 546+79 pp. Heithaus, E. R., Fleming, T. H., Opler, P. A. 1975. Foraging patterns and resource utilization in seven species of bats in a seasonal tropical forest. Ecology, 56:841-854. Holdridge, L. R. 1967. Life zone ecology. Tropical Science Center, San Jose, Costa Rica, 206 pp. Howell, D. J., Burch, D. 1974. Food habits of some Costa Rican bats. Rev. Biol. Trop., 21:281-294. Humphrey, S. R. 1975. Nursery roosts and community diversity of Nearctic bats. J. Mamm., 56:321-346. McNab, B. K. 1971. The structure of tropical bat faunas. Ecology, 52:352- 358. LaVal, R. K. 1970. Banding returns and activity periods of some Costa Rican bats. Southwest. Nat., 15:1-10. LaVal, R. K. 1973. A revision of the Neotropical bats of the genus Myotis. Sci. Bull., Los Angeles Co. Mus., 15:1-54. Orians, G. H. 1969. The number of bird species in some tropical forests. Ecology, 50:783-801. Pielou, E. C. 1969. An introduction to mathematical ecology. Wiley-Inter- science, New York, 286 pp. Pine, R. H. 1972. The bats of the genus Carollio. Tech. Monogr., Texas A & M Univ., No. 8, 125 pp. Rentz, D. C. 1975. Two new katydids of the genus Melanonotus from Costa Rica with comments on their life history strategies (Tettigoniidae: Pseudophyllinae). Ento. News, 86:129-140. Taddei, V. A. 1976. The reproduction of some Phyllostomidae (Chiroptera) from the northwestern region of the state of Sao Paulo. Boletim Zool., Univ. Sao Paulo, 1:313-330. Tuttle, M. D. 1974a. An improved trap for bats. J. Mamm., 55:475-477. Tuttle, M. D. 1974b. Unusual drinking behavior of some stenodermine bats. Mammalia, 38:141-144. 26 OCCASIONAL PAPERS, MUSEUM OF NATURAL HISTORY APPENDIX 1: Summary Data on 78 Species of Bats from Three Localities in Costa Rica Species mean mean Numbers of indivic luals captured 1 Ulcal III length Monte- Guana- ( mm ) weight(g) La Selva verde caste Food' 37.3 4.1 7 40 IA 45.2 7.6 17 12 IA 40 7 1 IA 48 16 1 IA 43.7 5 4 IA 41 1 IA 65.0 37.8 1 44 IA 82 25 F 47.5 10 IA 54.4 14.6 2 o IA 61.2 23.7 25 7 269 IA 47 3 IA 38 9.4 3 5 IFG 42.3 14.7 4 1 IFG 34.2 6.7 21 13 3 IFG 34.8 9 1 IFG 37.6 10.2 6 IFG 35 1 1 IFG 56.2 1 IFG 50 1 IFG i 35 3 3 IFG 57.5 31.8 15 IFG 35.2 7.4 8 IFG 52 1 IFG 50.3 15.0 10 IFG 57.5 25.9 16 IFG 64 1 7 O 90 1 O 77 1 O 61 32.3 35 2 F 107.7 166.0 4 4 F 34.5 9.3 81 34 10 PN 34.0 10.5 23 6 51 O 43.3 16.3 12 PN 43.6 1 PN 42.5 15.2 13 PN 32.9 7.9 o 8 5 PN 33.3 8.0 15 7 3 PN 32.4 7.0 3 PN 41.3 18.5 461 53 75 PF 36.5 13.8 302 PF 44.9 22.5 101 PF 40 2 22 PF 42.4 19.2 1 287 PF 48.1 26.9 8 PF 43.9 19.6 17 1 PF 39.8 15.4 16 PF 63.4 54.7 23 PF 53.1 32.8 1 PF Rhynchonycteris naso Saccopteryx bilineata Saccopteryx leptiura Pewpteryx kappleri Centron ycteris maximilian i Balantioptcryx plicata Noctilio labialis Noctilio lepoiinus Pteronotus davyi Ptcronotus suapu rensis Pteronotus pamellii Pteronotus personata Micronycteris brachyotis Micronyctetis hirsuta Micron ycteris megalotis Micronycteris minuta Micronycteris nicefori Micronycteris schm idtoru m Bart icon ycteris daviesi Lonchorhina aurita Macrophyllum macrophyllum Tonatia bidens Tonatia minuta Tonatia sylvicola Minion crenulatum Mitnon cozumelae Ph yllosto mus discolor Phyllostomus hastatus Phylloderma stenops Trachops cirrhosus Vampyrum spectrum Glossophaga commissarisi Glossophaga soricina Lonchophylla robust a Anoura eultrata Anoura geoffroyi Choeroniscus godmani Hylonycteris underwoodi Lichonycteris obscura Carollia brevicauda Carollia castanea Carollia perspicillata Stumira lilium Sturnira ludovici Stumira mordax Uroderma bilobatum Vampyrops helleri Vampyrops vittatus Vampyrodes major COSTA RICAN BAT COMMUNITIES 27 Vampyressa nymphaea Vampyressa pusilla Chirodcrma villosu m Ectophylla alba Artibcus jamaicensis Artibeus lituratus Artibeus phaeotis Artibcus toltecus Artibeus watsoni Enchisihenes luuii Desmodus rotundus Furipterus horrens Tlnjroptera tricolor Myotis albescens Myotis elegans Myotis keaysi Myotis nigricans Myotis oxyotus Myotis riparius Eptesicus andinus Eptesicus furinalis Rhogeessa tumida Lasiurus ega Molossops grecnlialli Molossus ater Molossus bondac Molossus molossus Molossus pretiosus Molossus sinaloac 37.7 2 PF 31.9 8.2 27 PF 46 1 1 PF 28.1 5.6 38 PF 64.2 54.9 111 24 PF 66.2 63.8 16 4 PF 38.8 12.4 133 5 PF 40.6 15.3 161 PF 39 4 PF 40.1 15 1 PF 59 29 B 34.3 3.3 59 IA 35 1 IA 36.3 6.3 7 IA 33.0 4.0 20 4 IA 36.3 5.0 226 IA 35.0 4.7 14 40 IA 40.7 2 IA 35.7 5.3 78 4 IA 44.2 1 1 IA 38.8 1 2 IA 29.6 1 9 IA 45.2 11 1 IA 539.4 9 36.9 3 IA 48 6 IA 5 40.7 22.9 17 IA 9 39.2 16.8 35 7 IA 544.9 29.9 47 IA 9 43.8 24.3 548.8 28.7 76 IA 9 47.2 23.1 1 Food includes: I A, insects — aerial feeders; IFG, insects — foliage-gleaners; PF, plants — fruit; PN, plants — pollen and nectar; F, flesh or fish; O, onmivore; B, blood.