Author: Henterly, Andrew C
Date published: April 1, 2011
In most ecological or behavioral studies, accurate species identification of the animals is critical. In the field, mammalogists typically use a combination of external morphological characters, such as pelage coloration and patterning, the length of the tail relative to the body, the length of the tail relative to the hind foot, the number of plantar tubercles and the number of mammae to distinguish between species (e.g., Gottschang, 1981; Hall, 1981). However, in some cases, external morphology may be unreliable, either because the two species in question are extremely similar in appearance, or due to intraspecific variation in supposedly diagnostic morphological characters. For example, in a region of sympatry in the eastern United States, subspecies of Peromyscus maniculatus (deer mouse) and P. leucopus (white-footed mouse) are so morphologically similar that accurate species determination in the field is almost impossible, and a sizeable literature is devoted to methods for distinguishing between the two species (e.g., Rich et al, 1996; Bruseo et al, 1999; Lindqtiist et al, 2003). Where external morphology is unreliable, dental characteristics may allow the determination of species (e.g., Hall, 1981). However, the observation of dental characters typically requires sacrificing the animal, making the use of dental characters unsuitable for studies in which the observation of individuals through time is necessary, such as many field studies that utilize catch-mark-recapture techniques. Thus, reliable identification of all individuals for some sympatric species may be extremely difficult in certain field situations.
When the identification of morphologically similar sympatric species in the field is difficult because of intraspecific variation or interspecific overlap in external morphology, species-specific differences at the molecular level may provide a more accurate means for identifying species. For example, Aquadro and Patton (1980) demonstrated that speciesspecific differences in salivary amylase electromorph phenotypes could be used to unambiguously distinguish between morphologically similar Peromyscus leucopus and P. manimlatus. More recently, several DNA-based techniques have utilized species-specific nucleotide polymorphisms to identify a variety of species of mammals (Beriy and Sarre, 2007; O'Reilly et al, 2007; Moran et al, 2008). These molecular diagnostics have proven extremely useful in cases where it is difficult or impossible to distinguish between morphologically similar species in the field.
Microtus ochrogaster (prairie voles) and M. pennsylvanicus (meadow voles) are morphologically similar species of arvicoline rodents that occur sympatrically in some parts of the midwestern United States (Reich, 1981; Stalling, 1990). Numerous field studies have examined the ecology, demography and social behavior of these two species in areas of symapatry (e.g., Krebs, 1977; Getz et al, 2001, 2005), and positive species identification is critical. Unfortunately, intraspecific morphological variation may confound the ability of researchers to reliably identify the species of some individuals in the field using external characters as diagnostics (e.g., DeCoursey, 1957; Fitch, 1957; Oppenheimer, 1965; Reich, 1981; Stalling, 1990; Schwartz, and Schwartz, 2001).
The neuropeptide arginine vasopressin influences social and reproductive behavior in male Microtus ochrogaster and Ai. pennsylvanicus by directly regulating the neural expression of arginine vasopressin la receptors (Vl aR; Lim et al, 2004; Hammock and Young, 2005) . Recent molecular studies have shown a species-specific difference between M. ochrogaster and Ai. pennsylvanicus in the length of microsatellite DNA within the gene (avprla) encoding the VIaR (Young et al, 1999; Fink et al, 2006) and the potential effects of this interspecific difference in avprla length on male social behavior have been extensively studied (Hammock and Young, 2005; Young and Hammock, 2007). Microtus ochrogaster have an expanded region of repetitive microsatellite DNA [approximately 430 base pairs (bp) J in the regulatory region of the avprla gene that is absent in M. pennsylvanicus. This length difference between species in the microsatellite DNA of the avprla gene has been found in M. ochrogaster sampled from populations in Illinois (IL), Kansas (KS) and Tennessee (Fink et al, 2006; Ophir et al, 2008; Solomon et al, 2009) and M. pennsylvannicus populations from IL, Montana and Pennsylvania (Fink et al, 2006). We used this apparent species-specific polymorphism in microsatellite length within the regulatory region of the avprlagene to determine species identity in order to assess the likelihood of species misidentification using the morphological characters typically employed for distinguishing between sympatric M. ochrogaster ?mà M. pennsylvanicus in the field.
Study areas and fiekl procedures. - Our study sites were located at the University of Kansas' Nelson Environmental Study Area (NESA), 12 km northeast of Lawrence, KS (39°03'07''N, 95°11'27''W), and at the Indiana University Bayles Road Preserve (BRP), 5 km north of Bloomington, IN (39°13'0''0, 86°32'27''W). At each site, the study area was 1.5 ha, with live traps placed in a grid pattern and spaced at 10 m intervals. Both sites consisted of old-field habitat, containing a variety of grasses interspersed by patches of forbs, shrubs and tree seedlings and were mowed annually to discourage the growth of invasive species and woody vegetation. Only Microtus ochrogaster was present at the KS site, which is outside the species range of M. pennsylvanicus (Reich, 1981; Stalling, 1990; Slade, pers. comm.); both vole species occurred at the IN site.
Fieldwork was conducted in KS during May-early Jul. 2008 and in IN during mid JuLSept. 2008. At each site, voles were trapped twice daily, 4-5 d per week for 8 wk, using singlecapture (Sherman; H. B. Sherman, Tallahassee, Florida) and multiple-capture (Ugglan; Grahnab, Hillerstorp, Sweden) live traps baited with cracked corn. Trapping occurred either at grid points (weeks 1, 2, 5 and 6) or known Microtus ochrogaster nest sites (weeks 3, 4, 7 and 8). Upon initial capture, all voles were given a unique toe-clip pattern for individual identification and the toes were frozen at -20 C for subsequent genetic determination of their avprla genotype.
All research procedures involving live voles followed the guidelines of the American Society of Mammalogists for the use of wild animals in research (Gannon et al, 2007) and were approved by the Institutional Animal Care and Use Committees of Miami University, the University of Kansas, and Indiana University.
Assessment of species identity in the field. - For every capture event of a vole at the IN study site, a field assessment of species identity was made based on differences between Microtus ochrogaster and M. pennsylvanicus in number of plantar tubercles and pelage as described by Reich (1981), Gottschang (1981), Stalling (1990) and Schwartz and Schwartz (2001) and summarized in Table 1. During the course of the 8 wk study, field assessments of the species identity of voles were made by five people who differed in their level of experience in identifying Microtus species in the field.
Preserved specimens. - Differences in dental features of the maxillary molars are considered the most reliable diagnostic characters for distinguishing these two species of voles (Reich, 1981; Stalling, 1990; Schwartz and Schwartz, 2001). Since these dental characteristics can best be examined in preserved skulls, a subset of males (n = 15) were euthanized at the completion of fieldwork in IN to permit analysis of molar cusp patterns. In addition, preserved skulls of three Microtus ochrogaster and four M. pennsylvanicus were acquired from the Miami University Hefner Zoology Museum in Oxford, OH. All seven voles from the museum were collected in the area around Oxford, OH. Molar cusp patterns of these 22 skulls were viewed through a dissecting microscope. The diagnostic character states of the second (M^sup 2^) and third (M^sup 3^) maxillae molars for each species are summarized in Table 1, as described by Reich (1981) and Stalling (1990).
The number of plantar tubercles and the ratio of tail length:hind foot length from fieldcollected specimens was also recorded. Microtus ochrogaster typically have five plantar tubercles and a tail length:hind foot length ratio <2, while in M. pennsylvanicus the typical number of plantar tubercles is six with a tail length:hind foot length ratio >2 (Table 1; Schwartz and Schwartz, 2001). When data on tail and hind foot lengths were available from specimen tags of preserved skulls, this information from museum specimens was also recorded. We only present data on the tail length/hind foot length ratio for the males collected from the IN site and the museum specimens because of the difficulty of accurately measuring these characters in living animals. We were unable to determine the number of plantar tubercles from museum specimens because we only had access to the skulls. All of the data on dentition, plantar tubercle number and tail length.hind foot length ratio of the preserved specimens was collected by one of the authors (ACH). The avprla genotypes of the museum specimens were determined from tissue collected by scraping the interior of brain cases with a sterilized sewing needle.
avprla genolyping. - Genomic DNA was extracted using DNeasy kits (Qiagen, Valencia, CA) and the avprla genotypes of voles were determined by polymerase chain reaction (PCR), with primers specifically designed to amplify the microsatellite region of the avprla gene in prairie voles (Hammock and Young, 2005). Our avprla genotyping procedure was identical to that followed by Solomon el al (2009), who confirmed that these primers amplified the targeted region of the avprla gene by sequencing the resulting PCR product and comparing the sequence to the prairie vole avprla gene sequence published in GenBank. Previous studies using the same primers as in our investigation have demonstrated that PCR amplification of this microsatellite region of the avprla gene in known Microtus pennsylvanicus yields PCR fragments that are 200-300 bp, while in known M. ochrogaster the resultant PCR fragments are 600-800 bp (Young el al, 1999; Fink el al, 2006).
Data analysis. - All means are reported ± 1 se. For the voles from the IN field site, we compared the concordance of species identification based on morphological characteristics observable in the field (pelage and plantar tubercles) with identifications based on avprla genotypes for all adult male voles trapped only once and all adult males trapped four or more times. Thus, we used two different levels of contact with live-trapped voles. Data from animals that were only trapped once were used to assess the accuracy of species identification when each animal was observed by a single handler. Data from animals with 5:4 captures increased the chances that an individual vole would be identified by multiple observers and allowed us to examine if rapture history may be helpful in clarifying misidentification issues.
Species identity in the field. - We determined the avprla genotypes of 70 adult males from the KS study site and 279 adult males from the IN study site. At the KS site, where only Microtus ochrogaster occurs, allele lengths ranged from 686 to 797 bp (Fig. 1). The distribution of avprla allele lengths at the IN site, where both vole species occur, was bimodal, with 50 individuals possessing two alleles ranging from 289 to 294 bp and 229 voles possessing two alleles between 696 and 798 bp (Fig. 1). None of these voles possessed one allele between 289-294 bp and the other between 696-798 bp. Thus, all the voles from the KS site as well as the 229 voles from the IN site with avprla allele lengths between 696 and 798 bp had avprla genotypes diagnostic of M. ochrogaster, while the 50 voles at the IN site with alleles lengths between 289 and 294 bp had avprla genotypes diagnostic of M. pennsylvanicus. The avprla génotype data from IN included the 15 males collected from the population at the end of the study, of which 1 3 had M. ochrogaster avprla genotypes and two had M. pennsylvanicus avprla genotypes.
Voles trapped only once. - There were 58 male voles at the IN site that were captured only once. Based on their avprla genotypes, 1 7 of these voles were Microtus pennsylvanicus and 41 were M. ochrogaster. Among the voles with M. pennsylvanicus avprla genotypes, 5 (29%) were identified in the field as M. ochrogaster. All five of these individuals had five plantar tubercles as is expected for M. ochrogaster. For the voles with M. ochrogaster avprla genotypes, one individual (2.5%) was identified as M. pennsylvanicus and had six plantar tubercles that is characteristic of M. pennsylvanicus. An additional vole (2.5%) identified as M. pennsylvanicus liad an indeterminate number of plantar tubercles, and was later genetically identified as M. ochrogaster. Overall, 12% (7/58) of the voles captured only once were misiden tilled in the field and in six of the seven cases of misidentification the vole had the number of plantar tubercles characteristic of the species for which it was mistaken.
Voles trapped four or more limes. - For the 221 voles caught four or more times at the IN site, voles with Microtus ochrogaster avprla genotypes were misidentified in 4.9% of all capture events, and animals with M. pennsylvanicus avprla genotypes were incorrectly identified in 4.2% of all capture events. Considering the total capture history of each individual, 28% (52/188) of individuals with M. ochrogaster avprla genotypes were recorded as M. pennsylvanicus at least once. Similarly, 30% (10/33) of voles with M. pennsylvanicus avprla genotypes were identified as M. ochrogaster at least once. Thus, 28% (62/221) of all adult male voles trapped s4 times at the IN site were misidentified to species during at least one capture. In 87% of these cases the number of plantar tubercles recorded for the misidentified animal was characteristic of the species that it was incorrectly identified as in the field.
Presented specimens. - Of the 22 vole skulls examined, one of the 16 (6%) voles with a Microtus ochrogaster avprla genotype deviated from the expected dental characteristics for this species, possessing an atypical molar cusp pattern at M^sup 3^ that did not resemble the expected pattern of either species (Fig. 2). Among the six voles with a M. pennsylvanicus avprla genotype, one individual (17%) deviated from the expected dental features for this species at both M^sup 2^ and M^sup 3^ , possessing typical "ochrogaster" M^sup 2^ morphology and irregular M^sup 3^ morphology (Fig. 2). Overall, the molar morphology of 9% (2/22) of skulls examined did not match that expected based on their avprla genotype.
The average tail length/hind foot length ratio among voles collected from the IN site and the museum specimens was 1.96 ± 0.1 (n = 14) for individuals with Microtus ochrogaster avprla genotypes and 2.19 ± 0. 1 (n = 4) for individuals with M. pennsylvanicus avprla genotypes. Only 57% (8/14) of these voles with M. ochrogaster avprla genotypes had tail length/hind foot length ratios less than two, as expected for this species (Table 1; Schwartz and Schwartz, 2001). For voles with M. pennsylvanicus avprla genotypes, 25% (1/4) had tail length/hind foot length ratios characteristic (<2) of M. ochrogaster. Among all the preserved vole specimens, 39% (7/18) of the individuals had tail length/hind foot length ratios that were inconsistent with their species diagnostic avprla genotypes.
For the males collected from the IN site, only one of 13 voles (8%) with a Microtus ochrogaster avprla genotype had an atypical number of plantar tubercles (six rather than five), while both of the voles with M. pennsylvanicus avprla genotypes had the species diagnostic six plantar tubercles. Overall 7% (1/15) of these voles had an atypical number of plantar tubercles.
When considering dentition, tail length/hind foot length ratio and plantar tubercles together, nine of the 22 preserved voles (41%) had one morphological feature that exhibited a character state that was not consistent with their avprla genotype and one vole (5%) had two morphological features that displayed character states not consistent with its avprla genotype. Thus, 46% of the preserved voles had at least one morphological feature exhibiting a character state that was inconsistent with their avprla genotype.
Our avprla genotype data for the voles we sampled from KS and IN were completely consistent with the previously reported species-specific genetic polymorphism in avprla allele length between Microtus ochrogaster and M. pennsylvanicus, with M. ochrogaster containing an expanded segment of repetitive microsatellite DNA in the avprla gene that is absent in M. pennsylvanicus (Young et al, 1999; Fink et al, 2006). Since only M. ochrogaster occurs at our KS study site, we expected and found all the individuals we sampled to have two avprla alleles between approximately 600 and 800 bp in length. At our IN study site, where the two vole species are sympatric, all individuals had two avprla alleles characteristic of either M. ochrogaster (696-798 bp) or M. pennsylvanicus (289-294 bp). No voles possessed avprla microsatellite lengths that were characteristic of both species, or atypical for either species. Therefore, we were able to use avprla allele lengths to unambiguously identify the species of all adult male voles from the IN study site genotyped at this locus.
Although pelage and number of plantar tubercles are the primary phenotypic characters used for distinguishing between Microtus ochrogaster and M. pennsylvanicus in the field, we found that, in our hands, they were not completely reliable when compared to the species diagnostic avprla genotypes of individuals. Species misidentification at our IN site ranged from -5% per capture event for M. ochrogaster, regardless of the number of times captured, to 29% for M. pennsylvanicus captured only once. Categorizing pelage color is very subjective, and subject to inter-observer differences. Moreover, both species display intraspecific variability in pelage color within and between populations (Gottschang, 1981; Reich, 1981; Stalling, 1990). We found some individuals with M. ochrogaster avprla genotypes to possess M. ftennsylvanicus-iike pelage color, and vice versa. While plantar tubercle number is a quantitative character and less subjective, it is not invariant in either vole species (Fitch, 1957; Schwartz and Schwartz, 2001). We found that for almost 90% of the cases of species misidentifications in the field, the misidentified vole had an atypical number of plantar tubercles. In addition, the number of plantar tubercles on an individual may be ambiguous due to wear, differential pigmentation or the amount of dirt on the feet.
Il might be expected that preserved specimens of Microtus ochrogaster and M. pennsylvanicus would be less prone to species misidentification since it should be easier to examine species diagnostic morphological characters, including dentition, on these voles than individuals in the field. However, for the three morphological characters we examined, 41% of the preserved specimens had one character state indicative of M. ochrogaster and another characteristic of M. pennsylvanicus, making species identification of these individuals ambiguous (one vole had dentition atypical for either species).
Features of the second and third maxillary molars are considered the most reliable morphological character for distinguishing between Microtus ochrogaster and M. pennsylvanicus (Reich, 1981; Stalling, 1990; Schwartz and Schwartz, 2001), but even with our small sample size (n = 22) we encountered two (9%) instances where molar morphologies deviated from that expected based on the avprla genotype of the animals. Interestingly, the percentage of voles where the expected number of plantar tubercles was not consistent with an individual's avprla genotype was similar to that for molar dentition (7%), suggesting that plantar tubercle number may be as reliable a character as dentition for species identification. The tail length/hind foot length ratio was the least reliable diagnostic morphological character we examined as 43% of the voles with M. ochrogaster avprla génotypes and 25% of voles with M. pennsylvanicus avprla genotypes deviated from their expected ratio. These results suggest that the tail length/hind foot length ratio would be even less reliable when used in the field, when measurements must be made on live animals.
The data from our study demonstrates that using morphological characteristics to distinguish between Microtus ochrogaster and M. pennsylvanicus results in cases where species identity is erroneous or ambiguous. Although genotyping voles at their avprla locus is more time consuming and expensive than using morphological characters, we believe there are situations where such genetic analysis may be warranted to obtain accurate species identification of all individuals where M. ochrogaster and M. pennsylvanicus are sympatric. The species identity of voles caught in the field >4 times was misidentified in 4-5% of all capture events, but for these animals an individual's capture history may be used to identify, and possibly correct, cases of species misidentification given the low error rate. Approximately 5% of M. ochrogaster and 29% of Ai. pennsylvanicus captured only once were misidentified. In particular, the level of misidentification of M. pennsylvanicus captured only once was substantial and 7-8 times that of M. pennsylvanicus captured a4 times. Determining the avprla genotype of voles captured just once may be the only way to detect instances of species misidentification and ensure the accurate species identification of all individuals. Our data indicate that species identification is not just a problem in mark-recapture field studies. The species identity of about 40% of the preserved specimens we examined exhibited diagnostic characteristics of both M. ochrogaster and M pennsylvanicus, but any ambiguity about the species identity of these specimens could be eliminated by genotyping these animals at the avprla locus. In general, we suggest that genotyping putative M. ochrogaster and M. pennsylvanicus at their avprla locus can be a useful diagnostic for eliminating erroneous or ambiguous species identification in many studies.
Acknowledgments. - We thank J. Edwards, M. Hirschauer, S. Schmits andj. Smith for assistance in the field, and G. Piumati and K. Clay for logistical support at our field sites. We also thank M. Wright of Miami University's Hefner Museum of Zoology for his assistance in obtaining museum specimens and S. Hoffman and Z. Taylor for their help with obtaining DNA samples from preserved skulls. Funding was provided by a National Science Foundation (NSF) award (1OB0614015) and a NSF REU supplement to BK and NGS that supported ACH.
AQUADRO, C. F. AND J. C. PATTON. 1980. Salivary amylase variation in Peromyscus: use in species identification. J. Mammal, 61:703-707.
BERRV, O. AND S. D. SARRE. 2007. Gel-free species identification using melt-curve analysis. MoI Ecol. Notes, 7:1-4.
BRUSEO, J. A., S. H. VESSEV AND J. S. GRAHAM. 1999. Discrimination between Peromyscus leiicopus novebinacensis and Peromyscus maniculatus nubiterrae in the field. Acta Theriol, 44:151-160.
DECOURSEV, G. E, 1957. Identification, ecology and reproduction of Microtus in Ohio. /. Mammal, 38:44-52.
FINK, S., L. EXCOFFIER AND G. LIECKEL. 2006. Mammalian monogamy is not controlled by a single gene. P. Natl Acad. Sci. USA, 103:10956-10960.
FITCH, H. S. 1957. Aspects of reproduction and development in the prairie vole (Microtus ochrogaster). Miscellaneous Publication of the Museum of Natural History, University of Kansas, 10:129-161.
GANNON, W. L., R. S. SIKES AND THE ANIMAL CARE AND LISE COMMITTEE OF THE AMERICAN SOCIETY OF MAMMALOGISTS. 2007. Guidelines of the American Society of Mammalogists for the use of wild mammals in research. J. Mammal, 88:809-823.
GETZ, L. L., J. E. HOFFMANN, B. MCGUIRE AND T. W. DOIAN. 2001. Twenty-five years of population fluctuations of Microtus ochrogaster and M. pennsylvanicus in three habitats in east-central Illinois. J. Mammal. 82:22-34.
_____ . M. K. OLI, J. E. HOFFMANN, B. MCGUIRE AND A. OZGUL. 2005. Factors influencing movement distances of two species of sympatric voles. J. Mammal., 86:647-654.
GOTTSCHANG, J. L. 1981. A Guide to the Mammals of Ohio. The Ohio State University Press, Columbus. Ohio.
HALL, E. R. 1981. The Mammals of North America. 2nd ed. John Wiley and Sons, New York, New York.
HAMMOCK, E. A. D. AND L. J. YOUNG. 2005. Microsatellite instability in brain and sociobehavioral traits. Science, 308:1630-1634.
KREBS, CJ. 1977. Competition between Microtus pennsylvanicus and Microtus ocrogaster. Am. Midi. Nat, 97:42-49.
LIM, M. M., Z. WANG, D. E. OLAZABAL, X. REN, E. F. TERWILI.IGER AND L.J. YOUNG. 2004. Enhanced partner preference in a promiscuous species by manipulating the expression of a single gene. Nature, 429:754-757.
LINDQUIST, E. S., C. F. AQUADRO, D. MCCLEARN AND K. J. MCGOWAN. 2003. Field identification of the mice Peromyscus leucoftus noveboracensis and P. maniculatus gracilis in Central New York. Can. Field Nat, 117:184-189.
MORAN, S., P. D. TURNER AND C. O. O'REILLY. 2008. Non-invasive genetic identification of small mammal species using real-time polymerase chain reaction. MoI. Ecol. Resimi., 8:1267-1269.
OPHIR, A. O, P. CAMPBELL, K. HANNA AND M. S. PHELPS. 2008. Field tests of (»-regulatory variation at the prairie vole atiprla locus: association with VIaR abundance but not sexual or social fidelity. Harm. Behav., 54:694-702.
OPPENHEIMER, J. R. 1965. Molar cusp pattern variations and their interrelationships in the meadow vole, Microtus pennsylvanicus (Ord). Am. Midi. Nat, 74:39-49.
O'REILLY, C., M. STATHAM, J. MULLINS, P. D. TURNER AND D. O'MAHONY. 2007. Efficient species identification of pine marten (Martes martes) and red fox (Vulpes vulpes) scats using a 5' nuclease real-time PCR assay. Conser. Genet, 9:735-738.
REICH, L. M. 1981. Microtus pennsylvanicus. Mammal. Spec, 159:1-8.
RICH, S. M., C. W. KJLPATRICK, J. L. SHIPPF.E AND K. L. CROWELL. 1996. Morphological differentiation and identification of Peromyscus leucopus and P. maniculatus in northeastern North America. J. Mammal, 77:985-991.
SCHWARTZ, C. W. AND E. R. SCHWARTZ. 2001. The Wild Mammals of Missouri, 2nd ed. University of Missouri Press, Columbia, Missouri.
SOLOMON, N. G., A. R. RICHMOND, P. A. HARDING, A. FRIES, S. JACQUEMIN, R. L. SCHAEFER, K. E. LUCIA AND B. KEANE. 2009. Polymorphism at the avprla locus in male prairie voles correlated with genetic but not social monogamy in field populations. MoI. Eroi, 18:4680-4695.
STALLING, D. T. 1990. Microtus ochrogaster. Mammal. Spec, 355:1-9.
YOUNG, L.J. AND E. A. D. HAMMOCK. 2007. On switches and knobs, microsatellites and monogamy. Trends Genet, 23:209-212.
_____, R. NILSEN, K. G. WAYMIRE, G. R. MACGREGOR AND T. R INSEL. 1999. Increased affiliative response to vasopressin in mice expressing the Vhl receptor from monogamous vole. Nature, 400:766-768.
SUBIMITTED 7 JANUARY 2010 ACCEPTED 17 SEPTEMBER 2010
ANDREW C. HENTERLY, KAREN E. MABRY,1 NANCY G. SOLOMON and ADRIAN S. CHESH
Center for Animal Behavior and Department of Zoology, Miami University, Oxford, Ohio
Center for Animal Behavior and 'Department of 'Zoology, Miami University, Hamilton, Ohio
1 Present address: Department of Biology, MSC 3AF, New Mexico State University, P.O. Box 30001, Las Cruces, New Mexico 88003
2 Corresponding author: e-mail: firstname.lastname@example.org