Recurrence of Mexican long-tongued bats (Choeronycteris mexicana) at historical sites in Arizona and New Mexico

The Mexican long-tongued bat (Choeronycteris mexicana) is 1 of 3 migratory, nectarivorous bats that seasonally occur in the extreme southwestern United States (US); the other 2 species are Leptonycteris curasoae and L. nivalis. Unlike the species of Leptonycteris, C. mexicana is not known to form large maternity colonies and is rarely encountered in groups of more than 12 individuals (Hoffmeister 1986). Possibly because of a propensity to form small roosting groups, the number of C. mexicana historically encountered is relatively low compared to other bat species. Although the range of C. mexicana extends from the southwestern United States into Honduras, less than 1500 individuals have been documented since its discovery in 1844 (Petryszyn and Cockrum In Press). Roosting and habitat needs of C. mexicana are poorly understood and it is unclear how such requirements might influence the apparent scarcity of these bats.

Choeronycteris mexicana is known to roost in a variety of situations, typically in shallow caves or near the entrances of more extensive structures (Arroyo-Cabrales et al. 1987). Roost sites have been reported from various vegetation zones, including tropical deciduous forests at southern latitudes (Davis and Russell 1954), but roosts are frequently found in oak-conifer woodlands in the northern part of its range (Hoffmeister 1986). Mexican long-tongued bats are known to feed on nectar, pollen, or fruit of various flowering plants throughout their range (Gardner 1977). Although mutualistic relationships likely exist between C. mexicana and its food plants, very little is known about the role that this species plays as a pollinator or seed disperser of such plants. The identification and elucidation of mutualistic relationships are necessary steps toward effectively conserving ecosystems in the southwestern US (Allen-Wardell et al. 1998). Given the potential importance of C. mexicana as a pollinator and its apparent scarcity in the southwest US, current status and habitat requirements of the species need to be determined. Furthermore, the majority of C. mexicana historically encountered north of the Mexican border were adult females and young (Petryszyn and Cockrum In Press), indicating that the southwestern US is an important breeding area. The objective of this study was to assess recurrence of C. mexicana at historical roost sites in Arizona and New Mexico, count numbers of bats present, and gather data on roost and habitat characteristics.

MATERIALS AND METHODS

Between 13 April and 9 August 1999, we surveyed sites in New Mexico and Arizona historically occupied by C. mexicana. Historical locality records were compiled from Petryszyn and Cockrum (In Press) and by communicating with other researchers in the region. Specific locations of historical records were often not described in detail and, in some cases, multiple locality descriptions were used for what we suspected was the same location. To account for such ambiguities, we defined a "site" as any point within a 5 km radius of where we estimated an observation of C. mexicana was originally recorded. For example, if the historical record listed "2 mi. W, 5 mi. N of Patagonia" as a locality, we considered any potential or occupied roost within a 5 km radius of that point as the same site. Field visits to sites coincided with the date (month and day) of original observation. If bats were not observed at a site on the initial visit, return visits were made between 14 and 30 days after the initial visit.

We approached potential roosts as quietly as possible and inspected roost entrances with binoculars from a distance > 10 m. If bats were not observed from > 10 m we then entered the potential roost. Bats were visually identified within roosts using lights and binoculars; the distinct profile of C. mexicana renders identification possible without capture (see photos in Hoffmeister 1986). We chose to not capture and handle bats due to the scarcity of known roosts and given the potential for roost abandonment. Historically occupied roosts at which we did not encounter bats during the initial search were monitored for 30 minutes in case we flushed bats upon our arrival. Individuals seen attached to adult bats were classified as young-of-year.

Locality data for each roost were recorded in UTM (NAD27 datum) coordinates using a global positioning system (GPS -- Rockwell PLGR, Rockwell Collins Inc., Cedar Rapids, Iowa, USA). Because C. mexicana is listed as threatened by the states of Arizona and New Mexico, specific locality data are restricted and available from the authors. In addition to roost locality data, we classified vegetative communities directly surrounding sites (Brown 1994). Distance from the nearest source of water and the presence of flowering plants were recorded for each site. We characterized roosts by measuring the height and width of entrances, depth of roost, depth of bats within roost, height of bats above ground, and aspect of roost opening.

RESULTS

We located and accessed 24 historical sites from an initial list of 39 in Arizona and New Mexico (Table 1). We found C. mexicana occupying 75% (n = 18) of the sites we visited. On several occasions C. mexicana returned to roosts after being flushed by our approach. A total of 104 bats was observed and average group size was 4.5 (range 1-17). Multiple individuals were observed at 83% (n = 15) of the occupied sites, and young-of-year were observed at 71% (n = 12) of occupied sites. Young-of-year bats comprised 23% (n = 24) of the total number encountered. Young were first observed on 17 June and we first observed evidence that young were flying on 26 July (indicated by presence of pollen on muzzle). Mean number of roosts encountered at each site was 1.6 (range 1-3) and mean roost elevation was 1,477.5 m (range 975-1,846). C. mexicana roosted in mine adits (n = 12), wide rock crevices (n = 6), caves (n = 6), and abandoned buildings (n = 3). Fifty-two percent of occupied structures had multiple entrances. Bats typically roosted in relatively well-lit areas close to entrances (mean = 2.7 m). All occupied sites were in Madrean evergreen woodlands (n = 11 -- sensu Brown 1994) or semidesert grasslands (n = 7 -- sensu Brown 1994). All sites were within 1 km of streams and, with the exception of a single site, within or near (< 0.5 km) riparian deciduous vegetation. Species of Agave were present in the vicinity of all occupied roosts; flowering A. schottii was observed near all but 1 of the occupied sites prior to the blooming of A. palmeri in mid-June. Sites where we did not encounter C. mexicana were either frequently disturbed (e.g., recreational caves with high visitation), difficult to search (e.g., extensive caves with high, inaccessible ceilings), or sites historically occupied by single individuals.

DISCUSSION

The number of bats we observed at historical sites was approximately half the number of individuals originally recorded at those sites (Table 1). However, comparison of present occurrences to historical records is problematic because in many cases the number of bats originally occupying a site was totaled over multiple visits. The total number of individuals we encountered during our survey is roughly equal to one-fifth the number previously documented north of the Mexican border (Petryszyn and Cockrum In Press). Considering the number of individuals we encountered and the relatively high rate of recurrence at historical sites, we do not have sufficient evidence to conclude that C. mexicana populations have increased or decreased in recent years.

Our observations of young-of-year bats at 71% of the sites we visited suggest that the species is successfully reproducing in the northern part of its range. Although nearly one-quarter of the bats we observed were young-of-year, this count is likely conservative due to the limitation of our survey method in assessing the age of bats. We classified individuals as young only during the time in which they were non-volant and attached to their mothers, hence independent juveniles were not counted. Furthermore, mothers with attached young were frequently observed rotating as they hung by a single leg so that their ventral surfaces were opposite the observer, thus obscuring their young from view. Such behavior may have limited our ability to accurately assess the presence of young.

In concordance with the observations of previous researchers, we found C. mexicana forming relatively small groups (Goodwin 1946; Hoffmeister 1986). Bats were rarely seen clustering and we did not observe adults in close contact (< 5 cm) with each other. On several occasions, individuals were flushed from roosts and led us to other nearby sites occupied by conspecifics. Furthermore, searches of rock crevices and shelters in historical roost areas often revealed multiple roosting groups of C. mexicana. These observations suggest that aggregations of C. mexicana are dispersed among several proximate sites.

Groups of C. mexicana have been observed occupying the same roost within (Mumford et al. 1964) and between years (Campbell 1934). The relatively high incidence of C. mexicana we observed at historical localities suggests year-to-year site fidelity. The first record of occupancy at sites for which we had data (n = 16; some historical records did not include year of capture) ranged between 3 and 77 years, with a mean span of 21 years elapsing since the first recorded occurrence. This apparent long-term fidelity indicates that some need is fulfilled at occupied sites which is not met elsewhere.

Choeronycteris mexicana is opportunistic in its roosting habits and there is no clear indication that it is dependent on a particular type of roost structure. Unlike species such as L. curasoae, which favor deeper caves and mines that conserve the metabolic heat of roosting bats (Fleming et al. 1998), roosts used by C. mexicana were shallow, relatively well-lit, and typically exposed to external ambient temperatures. In contrast to the limited availability of deeper roosts in suitable caves and mines where L. curasoae congregate, C. mexicana is apparently able to exploit a wider range of structures as roosts, possibly resulting in the dispersed nature of their aggregations. Based on our observations, we believe it is unlikely that the availability of roost structures limits the distribution of C. mexicana.

We believe that C. mexicana may be selecting roosts in relatively mesic areas. Nearly all occupied sites we encountered were within or very near (< 0.5 km) areas of riparian vegetation. In the southwest US, riparian vegetation likely buffers associated microclimates against fluctuations in temperature and humidity relative to surrounding desert habitats. By occupying riparian zones bats could utilize spatially abundant, shallow roost structures without exposing themselves to lethal temperatures or desiccation. The distribution of riparian zones, in conjunction with the distribution of plants on which C. mexicana feeds, may limit the distribution of these bats in Arizona and New Mexico. Given the wary nature of C. mexicana and its propensity to move between roosts in a small area, we believe that loss of riparian vegetation may be a greater threat to the species than disturbance at a particular roost.

In Mexico, C. mexicana is known to feed on the nectar and pollen of various cacti (e.g, Lemaireocereus, Myrtillocactus), Agave, and other flowering plants such as Ipomoea and Ceiba (Villa-R 1967, Alvarez and Gonzalez-Q 1970). In Arizona, Van de Water and Peachey (1997) found that C. mexicana near Tucson fed predominantly on cactus and Agave species. While flowering cacti were present near (< 10 km) some of the occupied sites we visited, species of Agave were the consistent floral characteristic of all sites. Agave schottii was the only species of that genus observed blooming at occupied sites before mid-June, after which blooming A. palmeri was also encountered. The co occurrence of early- and late-blooming species of Agave may be another factor influencing site occupancy by C. mexicana.

This survey established a body of information that will be useful to future population monitoring efforts. Considering their ecological importance and susceptibility to population decline, it is crucial that bat populations are effectively monitored (O'Shea and Bogan 2000). Because of the limited number of roost sites involved in this study, we recommend that these areas be surveyed every 2 to 3 years and protected from disturbance. In addition, further research into the habitat needs, food habits, pollination role, survivorship, and distribution of C. mexicana would enhance our understanding of these potentially important migratory pollinators.

ACKNOWLEDGMENTS

This project was funded through the United States Geological Survey "Species at Risk" program awarded to MAB. We would like to extend special thanks to B. Alberti, J.S. Altenbach, B. Brown, M. Chew, N. Cline, E.L. Cockrum, W. Cowan, C. Cowan, L. Cryan, V. Dalton, T. Deeken, R. Ferdon, C. Geiselman, E. Gering, G. Helbing, R. Hill, K. Hinman, J. Koloszar, G.P. Nabhan, W. Peachey, K. Peterson, Y. Petryszyn, C. Rau, H. Richter, S. Schmidt, S. Schuetze, S. Schwartz, R. Sidner, S. Spofford, S. Stone, T. Snow, J. Tyburec, and M. Weesner for their generosity and assistance. Without the cooperation of this diverse group of individuals, this survey would have been impossible. We thank T.J. O'Shea, T.H. Fleming, P.M. Gorresen, B.L. Cypher, and an anonymous reviewer for offering helpful comments on earlier drafts of this manuscript. Thanks to J.C. Richardson for providing logistical support throughout the project.

LITERATURE CITED

Allen-Wardell, G. et al. 1998. The potential consequences of pollinator declines on the conservation of biodiversity and stability of food crop yields. Conservation Biology 12:8-17.

Alvarez, T., and L. Gonzalez-Q. 1970. Análisis polínico del contenido gástrico de murciélagos Glossophaginae de México. Anales del Escuela Nacional Ciencias Biologia, México 18:137-165.

Arroyo-Cabrales, J., R.R. Hollander, and J.K. Jones, Jr. 1987. Choeronycteris mexicana. Mammalian Species 291:1-5.

Barbour, R.W., and W.H. Davis. 1969. Bats of America. The University Press of Kentucky, Lexington, 286 pp.

Brown, D.E. (Editor). 1994. Biotic communities: southwestern United States and northwestern Mexico. University of Utah Press, Salt Lake, 342 pp.

Campbell, B. 1934. Notes on the bats collected in Arizona during the summer of 1933. Journal of Mammalogy 15:241-242.

Davis, W.B., and R.J. Russell, Jr. 1954. Mammals of the Mexican state of Morelos. Journal of Mammalogy 35:63-80.

Fleming, T.H., A.A. Nelson, and V.M. Dalton. 1998. Roosting behavior of the lesser long-nosed bat, Leptonycteris curasoae. Journal of Mammalogy 79:147-155.

Gardner, A.L. 1977. Feeding habits. Pages 293-350 in R. J. Baker, J. K. Jones, and D. C. Carter, editors, Biology of the bats of the new world family Phyllostomatidae. Part II. Special Publications of the Museum, Texas Tech University, Lubbock.

Goodwin, G.C. 1946. Mammals of Costa Rica. Bulletin of the American Museum of Natural History 87:271-474.
Hoffmeister, D.R. 1986. Mammals of Arizona. University of Arizona Press, Tucson, 602 pp.

Mumford, R.E., L.L. Oakley, and D.A. Zimmerman. 1964. June bat records from Guadalupe Canyon, New Mexico. The Southwestern Naturalist 9:43-45.

Petryszyn, Y., and L.E. Cockrum. In preparation. The geographic distribution of the Mexican long-tongued bat (Choeronycteris mexicana). University of Arizona, Tucson.

Van de Water, P.K., and W.D. Peachey. 1997. Dietary analysis of the Mexican long-tongued bat (Choeronycteris mexicana) using pollen analysis of guano collected in Cienega Creek Nature Preserve. Bat Research News 38:133.

Villa-R., B. 1967. Los murciélagos de México. Instituto de Biología, Universidad Nacional Autónoma de México, Distrito Federal, México, 491 pp.

Table 1. List of sites historically occupied by Choeronycteris mexicana. Sites listed by state and county with information pertaining to the number of C. mexicana historically encountered (Historic n =), whether the site was found during the 1999 survey (Visited -- "X" indicates site was visited), the number of C. mexicana encountered during this survey (1999 n =), and whether young were present during this survey (Young -- "X" indicates young were present). Values marked by * reflect the number of bats observed on multiple visits.

Figure 1. Map showing the survey area and distribution of sites at which C. mexicana have been encountered in Arizona and New Mexico. Dashed lines represent county boundaries. Circles indicate sites visited during the summer of 1999 when new data were obtained. Triangles indicate previously occupied sites that were not visited during 1999.