MOLECUUUl EVOLUTION IN NEW WORLD LEAF-NOSED BATS
OF THE FAiMILY PHYLLOSTOMIDAE WITH COMMENTS
ON THE SUPERFi^ILY NOCTILIONOIDEA
by
RODNEY L. HONEYCUTT, B.A., M.S.
A DISSERTATION
IN
BIOLOGY
Submitted to the Graduate Faculty of Texas Tech University in Partial Fulfillment cf the Requirements for the Degree of
DOCTOR OF PHILOSOPHY
Approved
Accepted
May, 1981 '^"^ ^ ACKNOWLEDGMENTS
Two people have had a large impact on my professional career, my thinking, and my thesis research. I thank my major professor. Dr.
Robert J. Baker, whose personal integrity, insight into problems, and guidance of students will always serve as an inspiration to me during my career. I also thank Dr. Vincent M. Sarich (Departments of
Biochemistry and Anthropology, University of California at Berkeley) who opened his home and his mind to me, and who also taught and assisted me with the immunology research.
I am indebted to Professors Raymond C. Jackson, John S. Mecham,
Francis L. Rose, and J. Knox Jones, Jr. for direction and valuable interaction during the final phases of this thesis. I thank Dr. Allan
C. Wilson (Department of Biochemistry, University of California at
Berkeley) for providing access to his laboratory and equipment for the inmiunology research. Ellen M. Prager provided helpful suggestions concerning various aspects of the immunological research. Hugh H.
Genoways, Curator of Mammals, Carnegie Museum of Natural History, provided me with an opportunity to collect in Suriname, and without his help this thesis could not have been completed. J. A. Groen, I. F.
Greenbaum, S. L. Williams, A. Capparella, C. Davidson, R. J. Baker,
H. H. Genoways, and J. Knox Jones, Jr. assisted in collecting specimens.
Stuart L. Pimm and Michael P. Moulton gave invaluable assistance and advise concerning statistical analyses. Michael L. Arnold provided interaction as well as unpublished electrophoretic data, which I used
ii in my comparisons of albumin and electrophoretic rates of divergence.
The collection of specimens and laboratory equipment were sponsored by the following sources: 1) NSF grants DEB-80-04293 and DEB-76-20580;
2) a graduate studies grant from the Graduate School, Texas Tech
University; 3) M. Graham Netting Research Fund through a grant from
Cordelia S. May Charitable Trust; 4) Institute of Museum Research;
5) John D. Archbold Family Trust; 6) a grant from the Alcoa
Foundation, Charles L. Griswald, President; 7) International
Environmental Sciences Program awarded to John Eisenberg.
This thesis is dedicated to three ladies—my wife, Dierdre, who not only typed this thesis but who has continually supported my career goals, my daughter, Heather, who provided me with enjoyable distraction, and my mother-in-law, Laginia Hale, who has been an inspiration to me as well as one of my strongest supporters.
iii TABLE OF CONTENTS
ACKNOWLEDGMENTS ii
LIST OF TABLES v
LIST OF FIGURES vii
INTRODUCTION 1
METHODS AND MATERIALS 9
RESULTS 20
DISCUSSION 60
CONCLUSIONS 80
LITERATURE CITED 84
iv LIST OF TABLES
Table Page
1. Albumin immunological distance values for mormoopid and noctilionid species 21
2. Corrected albumin immunological distances among representatives of the subfamilies Phyllostominae, Desmodontinae, Brachyphyllinae, Glossophaginae, and Carolliinae 24
3. Albumin immunological distances comparing Artibeus (Stenodermatinae) to major lineages depicted in Fig. 1 27
4. Albumin immunological distances among certain lineages depicted in Fig. 1, Artibeus, and two outside reference points, Pteronotus parnellii and the Pteropidae 28
5. Albumin immunological distances resulting from one way comparisons of certain phyllostomine genera 32
6. .\lbumin immunological distance values for six Tonatia species, three Phyllostomus species, and Phylloderma 33
7- Albumin immunological distance values for five species of Micronycteris 36
8. Albumin immunological distances resulting from one way comparisons of Erophylla and additional glossophagine species 38
9. Albumin immunological distances comparing Rhinophylla to the major lineages depicted in Fig. 2 41
10. Alleles and frequencies for the loci that show genie variation among the three genera of vampire bats, Glossophaga, and Stumira 42 Table Page
11. Coefficients of genie similarity and distance among the three genera of vampire bats 45
12. Chi-Square tests for regular rates of albumin change among lineages of noctilionoid bats 50
13. Relative amounts of albumin change along all phyllostomid lineages 54
14. Electrophoretic distance (Nei's D) and albumin immunological distance (AID) estimates for 55 species pairs 56
VI LIST OF FIGURES
Figure Page
1. i\ phylogeny of the Noctilionoidea lineages based on the albumin immunological distances of Table 1 22
2. An albumin phylogeny for the family Phyllostomidae based on the values in Table 2 25
3. An albumin phylogeny for the family Phyllostomidae based on the distance matrix in Table 4 29
4. An albumin phylogeny of the Brachyphyllinae, Glossophaga, and Monophyllus 37
5. A phylogenetic tree depicting the relationships among the three vampire bat genera 47
6. An albumin phylogeny for the vampire bats 48
7. A bar graph of the amount of albumin change per lineage (ordinate) versus the number of lineages (abscissa) demonstrating a particular range of albumin change 53
8. Nei's electrophoretic distance (dependent variable) plotted against albumin immunological distance (independent variable) 58
9. Albumin immunological distance (dependent variable) plotted against Nei's electrophoretic distance (independent variable) 59
Vll INTRODUCTION
The family Phyllostomidae is one of the most taxonomically diverse families of bats, comprised of six subfamilies (Phyllostominae,
Glossophaginae, Carolliinae, Stenodermatinae, Brachyphyllinae, and
Desmodontinae), 49 genera, and about 137 species (Jones and Carter,
1976). Adaptations for a wide array of feeding strategies including sanguivory, camivory, omnivory, frugivory, and nectivory also are reflective of the magnitude of diversity within the family (Smith, 1976;
Gardner, 1977^). The systematics and classification of bats within the family have been based primarily on classical morphological characters such as dentition, cranial features, and the postcranial skeleton.
Current ideas concerning the evolution and phylogenetic relationships of phyllostomids are uncertain (Smith, 1976). Smith further suggests that this uncertainty may stem from convergent or parallel evolution of morphological features in response to similar feeding strategies.
Chromosomal data generated from both standard karyotypes and, more recently, G- and C-banding have contributed to our understanding of phyllostomid evolution, and in several cases they have provided alternatives to those posed by classical morphology (Baker, 1967; Bass,
1978; Patton and Baker, 1978; Baker, 1979; Baker and Bass, 1979;
Baker et al., 1979).
Although the application and usefulness of molecular techniques
(electrophoresis and immunology) in evolutionary studies on vertebrates is well documented (Sarich and Wilson, 1967; Sarich, 1969a,b; Sarich,
1 2
1973; Avise, 1974; Maxson and Wilson, 1974; Cronin and Sarich, 1975;
Maxson and Wilson, 1975; Sarich and Cronin, 1977; Gorman et al.,
1980), except for a few cases (Gerber and Leone, 1971; Greenbaum and
Baker, 1976; Straney et al., 1979). data concerning molecular evolution within the Phyllostomidae are lacking. More complete data on molecular evolution within the Phyllostomidae could provide valuable insight into the evolutionary relationships within the family as well as information necessary for the eventual development and testing of broad concepts which attempt to explain the overall processes of evolution. The purpose of this thesis is to provide information on both molecular evolution in phyllostomid bats and an independent assessment of current morphological and chromosomal data. The data are primarily immunological and concern albumin evolution. Sections of this thesis already have been submitted for publication; therefore, I have taken the liberty of incorporating excerpts from those papers. The major topics to be addressed using albumin immunology are outlined below.
Superfamilial Associations
The superfamily Noctilionoidea is comprised of three families,
Phyllostomidae, Mormoopidae, and Noctilionidae. The genera Mormoops and Pteronotus classically have been placed either in the family
Mormoopidae or (formerly) in the subfamily Chilonycterinae within the
family Phyllostomidae (Smith, 1972). The systematic association of
Mormoops with Pteronotus, suggesting a shared common ancestor for these
genera following their separation from all other taxa in the superfamily,
has been recognized since the work of Dobson (1875). Noctilio 3
(representing the family Noctilionidae) has presented more of a problem
to chiropteran systematists. Some workers have associated Noctilio with the Emballonuroidea (Dobson, 1875; Trouessart, 1897; Miller,
1907), whereas others (Winge, 1892; Smith, 1972; Patton and Baker,
1978) have indicated a mormoopid-phyllostomid association.
Chromosomal change between the Mormoopidae and Noctilionidae
constitutes the least amount of divergence thus far documented between
two mammalian families (Patton and Baker, 1978). This disparity in
chromosomal change is in direct contrast to the degree of morphological divergence (Smith, 1972). Chromosomal data have been interpreted as
supporting a common ancestor for the two families, Mormoopidae and
Noctilionidae (five synapomorphies) relative to the Phyllostomidae.
Patton and Baker (1978) further suggested that the mormoopids and noctilionids are related to the phyllostomids. The two mormoopid
genera, Pteronotus and Mormoops, share two synapomorphic elements
(Patton and Baker, 1978), with one more rearrangement distinguishing
Mormoops from Pteronotus (Baker and Bickham, 1980).
Subfamilial .Associations
On morphological grounds, most investigators agree that the
Phyllostominae is the most primitive of the phyllostomid subfamilies
(Walton and Walton, 1968; Slaughter, 1970; Smith, 1976). Chromosomal
data also suggest that the primitive karyotype for the family is like
that found in the extant phyllostomine species Macrotus waterhousii
(Patton and Baker, 1978). Xs reviewed by Smith (1976), the remaining
subfamilies have radiated from certain phyllostomine lineages. 4
.\1though this radiation is a point of agreement among chiropteran
systematists, the phylogenetic relationships among the subfamilies are
not well defined, and the viable alternatives seem to be limited only
by the mathematical possibilities.
Subfamily Phyllostominae
This subfamily seems to be comprised of several lineages, but
their composition is not clear (see Smith, 1976, for a review). A working hypothesis is that the Phyllostominae is a monophyletic group,
and the genera therein share a common origin relative to genera from
other subfamilies.
Patton and Baker (1978) have derived several lineages within the
Phyllostominae. However, intrageneric relationships in two genera,
Tonatia and Micronycteris, were unclear due to the lack of presumed
chromosomal homology between certain congeneric species. Baker and
Bickham (1980) recently have proposed a concept termed "karyotypic
megaevolution" to explain the lack of homology between congeneric
species. The question as to whether species within these genera are
monophyletic is important to the assessment of such a concept;
therefore, intrageneric relationships in the genera Phyllostomus,
Tonatia, and Micronycteris will be tested using immunology.
Subfamily Glossophaginae
The Glossophaginae seemingly are comprised of several distinct
lineages (Baker, 1967; Phillips, 1971). Certain authors (Baker, 1967;
Gerber and Leone, 1971) have proposed a hypothesis which suggests that 5
the glossophagine genera are polyphyletic in origin. This stems mainly from Baker's (1967) indication that Leptonycteris and Glossophaga shared karyotypic affinities with Phyllostomus (subfamily Phyllostominae) whereas Choeroniscus shared affinities with Carollia (subfamily
Carolliinae). Using G- and C-bandlng, Stock (1975) indicated that
Choeroniscus and Carollia shared no chromosomal homologies; however, as indicated by Baker and Bickham (1980), this could be due to karyotypic megaevolution. Immunology should provide a valid test of the polyphyletic origin hypothesis.
Subfamily Carolliinae
Classical morphology aligns two genera (Rhinophylla and Carollia) in this subfamily (Miller, 1907; Walton and Walton, 1968; Slaughter,
1970). In contrast, initial chromosomal data from G- and C-banding suggest an alignment of one species, Rhinophylla pumilio, with the phyllostomine genera Phyllostomus, Tonatia, and Mimon (Baker, unpublished). Intergeneric and intrageneric relationships are complicated by differential or rapid rates of chromosome evolution in
Carollia and Rhinophylla fischerae (Baker and Bickham, 1980; Baker et al., manuscript); therefore, the chromosomal data cannot be compared to the morphological data.
Subfamily Brachyphyllinae
The evolutionary affinities and number of genera in the
Brachyphyllinae has been a persistent source of debate (Sllva-Taboada and Pine, 1969; Baker and Bass, 1979). One of the three genera. 6
Brachyphylla, currently placed in the subfamily Brachyphyllinae has been placed in four different subfamilies (Baker and Bass, 1979). Data from G-band chromosomes clearly show that relative to the proposed primitive karyotype for the family Phyllostomidae (Patton and Baker,
1978), Brachyphylla shares a highly derived karyotype with Erophylla and Phyllonycteris (subfamily Brachyphyllinae) and Glossophaga and
Monophyllus of the subfamily Glossophaginae (Baker and Bass, 1979).
Therefore, the chromosomal data cannot document a closer alignment of
Brachyphylla to Erophylla and Phyllonycteris than to the two glossophagine genera. The karyotypic data can be interpreted in two ways. First, the karyotype of the brachyphyllines, Glossophaga and
Monophyllus, is plesiomorphic (primitive) for all glossophagines as well as the brachyphyllines. Second, this karyotype is the result of synapomorphies (shared derived) which document that some glossophagine genera (Monophyllus, Glossophaga, and possibly others) and the
Brachyphyllinae shared a common ancestor after diverging from other glossophagines (as proposed by Gardner, 19772)- ^^® systematic implications of the two alternative hypotheses are radically different.
In the first case, recognition of the taxon Brachyphyllinae would be justifiable from a phylogenetic standpoint, with the Glossophaginae as a sister taxon. However, the second case would imply that despite the classical anatomical and dental differences which serve as the basis for recognition of the two higher taxa, there would be no phylogenetic basis for recognition of the Brachyphyllinae as a sister taxon to the
Glossophaginae. Subfamily Desmodontinae
To date, two points of agreement about the evolution of vampire bats exist. First, among living mammals, obligate blood feeding is fovmd only in the true vampire genera (Desmodus, Diaemus, and Diphylla), and it is most parsimonious to conclude that these genera evolved from a common obligate blood feeding ancestor. Second, the vampires, among all living bats, are most closely related to phyllostomids (Forman et al., 1968). The specific nature of the phyllostomid-vampire relationship as well as the relationships among the three genera remain unclear. For example, classic morphological evidence indicates that
Desmodus and Diaemus fomn a clade relative to Diphylla (Miller, 1907;
Slaughter, 1970; Cadena, 1977). Two authors (Handley, 1976; Koopman,
1978) go so far as to place Diaemus in synonomy with Desmodus and retain
Diphylla as a separate genus. In contrast, chromosomal data from G- and C-bandlng studies indicate that Desmodus and Diphylla form a clade relative to Diaemus (Bass, 1978). Using starch-gel electrophoresis and immunology, I propose to examine relationships among the three vampire genera and test the two alternative hypotheses of morphology and chromosomes.
Rates of Molecular Evolution
Two basic patterns of molecular evolution, both of which reflect the time-dependent nature of changes in amino acid sequences, have been observed for most vertebrate groups. The first pattern indicates that albumin, detected with quantitative immunology, accumulates amino acid substitutions along lineages at a constant or regular rate (Wilson et 8
al., 1977). Although deviations from regularity have been identified
in certain instances (Cronin and Sarich, 1975; Maxson et al., 1975;
Sarich and Cronin, 1977), general regularity has been demonstrated for
a large number of vertebrates including anurans (Wallace et al., 1973;
Maxson and Wilson, 1975; Maxson et al., 1975; Case, 1978), urodeles
(Maxson and Wilson, 1979), reptiles (Gorman et al., 1971; Gorman et al., 1980), and mammals (Sarich and Wilson, 1967; Sarich, 1969a,b;
Cronin and Sarich, 1975; Maxson et al., 1975; Sarich and Cronin,
1977). The second pattern involves the amounts of change in albumins and other estimates of sequence divergence including non-repeating DNA
sequences and electrophoresis. These estimates are linearly related when the same species pairs are compared (Sarich, 1977; Sarich and
Cronin, 1977; Wilson et al., 1977; Maxson and Wilson, 1979). This
second pattern led Sarich (1977) to suggest that time was probably responsible for such a linear relationship. Thus, Sarich devised a means of estimating time since divergence from Nei's electrophoretic distance using an extrapolation from time estimates calculated from albumin immunological distances. Recently, Corruccini et al. (1980) have proposed fitting macromolecules to a non-linear model. According to these authors, the demonstration that two molecules vary in a curvilinear fashion relative to one another would indicate that both molecules could not be linearly related with time.
The albiamln immunological data in this thesis plus existing electrophoretic data allow an opportunity to examine both these patterns in phyllostomid bats. METHODS AND MATERIALS
Immunology
Albumins of Macrotus waterhousii, Vampyrum spectrum, Micronycteris sylvestris, Micronycteris davlesi, Phyllostomus hastatus, Tonatia sllvlcola, Carollia perspicillata, Artibeus jamalcensis, Glossophaga sorlclna, Monophyllus plethodon, Phyllonycteris aphylla, Brachyphylla cavemarum, Desmodus rotundus, Diphylla ecaudata, Mormoops megalophylla,
Pteronotus parnellii, and Noctilio leporinus were purified from tissue or serum according to the following procedure. For each albumin, 5 g of tissue was homogenized in 15 ml isotris buffer, dlalyzed against
50 mM Trls (pH = 8.0) and centrlfuged. To the supernatant was added
20 mg Rlvanol (2-ethoxy 6, 9 diaminoacrldine lactate) in 1 ml of the above buffer and the resulting precipitate separated by centrifuging.
The precipitate (mainly a Rlvanol-albumin complex) was then treated with 5 ml of 0.5 M Trizma-HCL until the fine yellow particles of free
Rlvanol were regenerated. The Rivanol was then separated by centrifuging and the supernatant dlalyzed against 0.2 M Tris-sulfate
(pH = 8.9) and vacuum-dialyzed to a volume of 0.7 ml. Fifty percent glycerol (0.2 ml) was then added and the albumin was isolated by preparative polyacrylamide gel electrophoresis using the above- mentioned Trls-sulfate buffer for the gel and a Tris-borate electrode buffer (3.2 gm Tris, 0.45 gm boric acid per liter). The acrylamlde concentration was 7.5%. The albumin band was Identified using ANS
(8-Anlllno-l-Naphthalenesulfonic Acid Magnesium Salt), cut out of the 10 gel, and eluted in 12 ml isotris buffer. Albumins from serum were purified in the same way except that the original dialysis was done on a six-fold diluted serxmi sample. It is also useful (assuming an appropriate antiserum is available) to test the completeness of the albumin precipitation by an immunodiffusion (Ouchterlony) comparison of the regenerated albumin and Rivanol supernatant prior to electrophoresis.
Antisera to all the purified albumins were prepared in rabbits
(three to four Dutch Belted rabbits per albumin) according to the schedule of Sarich (1969a). The immunization of at least three rabbits to the same antigen is important to counter the individual variation in specificity of the antiserum each rabbit produces (Prager and Wilson,
1971^). In each case, the resulting antisera to a particular bat antigen were titered using the mlcrocomplement fixation (MC'F) procedure and pooled in inverse proportion to their titers (Prager and
Wilson, 1971a^) . These pooled antisera were used for all subsequent analyses. All antigens used for cross-reactions with different antisera were extracted from samples of whole serum or tissue diluents.
Immunological cross-reactions (antigen-antibody reactions) for all comparisons within the family Phyllostomidae were measured by quantitative mlcrocomplement fixation according to the procedure of
Champion et al. (1974). Interfamllial comparisons (among the
Mormoopidae, Noctilionidae, and Phyllostomidae) were measured by the quantitative precipitin technique (Sarich and Wilson, 1966; Prager and
Wilson, 19712)- As Indicated by Prager and Wilson (19712). these two
techniques are interconvertible, and the degree of cross-reaction in 11 both techniques can be expressed quantitatively as albumin immunological distance units (AID). Each unit is approximately equivalent to one amino acid substitution (Prager and Wilson, 19712; Maxson and Wilson,
1974; Wilson et al., 1977). Champion et al. (1974) provide explicit details concerning the calculation of immunological distances from immunological cross-reactions.
With the exception of one-way comparisons, immunological distance estimates represent the mean of reciprocal immunological comparions.
Ideally, reciprocal comparisons should yield identical immunological distances; however, this ideal is seldom obtained (Champion et al.,
1974; Cronin and Sarich, 1975; Sarich and Cronin, 1977). Therefore, an internal test of the immunological data is the degree of nonreciprocity (Sarich and Cronin, 1977)- Nonreclprocity can be defined as
(anti-A with B - anti-B with A) ^ (anti-A with B -I- anti-B with A)
Generally, nonreciprocity varies from 4-10%, and values which are higher can be corrected as indicated by Cronin and Sarich (1975). In this study reciprocal comparisons were corrected prior to cladistical analyses following the standard procedures of Cronin and Sarich (1975).
Electrophoresis
Genie variation for structural genes encoding enzymatic and nonenzymatic proteins was assayed for the Desmodontinae using starch gel electrophoresis. Techniques of tissue preparation, electrophoresis, and biochemical staining were similar to those described in Selander 12 et al. (1971) as modified by Greenbaum and Baker (1976). All gels were made using Electrostarch (Lot 371, Electrostarch Co., Madison,
Wisconsin). The following gel buffer systems were used: 1) continuous tris-cltrate I for isocitrate dehydrogenase (Idh-1, Idh-2), lactate dehydrogenase (Ldh-1, Ldh-2), malate dehydrogenase (Mdh-1, Mdh-2), and malic enzyme (Me); 2) continuous tris-cltrate II for a- glycerophosphate dehydrogenase (a-Gpd), alcohol dehydrogenase (Adh), glutamate oxalate transaminase (Got-1, Got-2), and leucine aminopeptidase (Lap); 3) discontinuous tris-citrate (Poulik) for glucose-6-phosphate dehydrogenase (G6p); 4) phosphate (pH 6.7) for albumin (Alb), sorbitol dehydrogenase (Sdh), and phosphoglucose isomerase (Pgi-1); 5) tris-hydrochloric acid for esterase (Est-1,
Est-3); 6) tris-malate for indophenol oxidase (Ipo-1, Ipo-2),
6-phosphogluconate dehydrogenase (6Pgd) and phosphoglueomutase (Pgm-1).
Alleles at each presumed locus were designated alphabetically in order of decreasing mobility. The most anodal locus in a system was designated number "1", with more cathodal loci receiving progressively higher numerical designations. Alleles were compared using direct side-by-side migrations of all variants on the same gel, and bands with identical mobility were assumed homologous.
Coefficients of genie similarity and distance were calculated for paired combinations of all samples. Estimates of genie similarity were based on S (Rogers, 1972). Genetic distance defined by Nei (1972) as
D = In I was also calculated. 13
Construction of Phylogenetic Trees
Since immunological and electrophoretic distance values provide indirect estimates of sequence divergence, the selection of appropriate procedures for phylogeny reconstruction are of paramount importance.
As indicated by Wilson et al. (1977), the method of analysis should provide information concerning branching sequences, taxonomic relationships, and relative rates of evolution. Two methods of analyses meet these criteria. The Fitch and Margoliash (1967) method is an iterative averaging procedure which apportions amounts of change along branch lengths in a manner such that the differences between the distances calculated from the resultant tree (output data) and the distances from the original matrix (input data) are minimized. Thus, all possible trees can be assessed. The distance Wagner procedure
(^Farris, 1972) is not an iterative averaging procedure but rather an approach which estimates the most parsimonious tree with the assumption
that molecular distance data are minimum estimates of actual distance.
Since proponents of the Fltch-Margollash method and the distance
Wagner procedure disagree on the best approach to use in tree construction (Farrls, 1972; Prager and Wilson, 1978; Farris et al.,
1979), I have constructed the most probable trees from my immunological data using both approaches. In all cases the branching sequences were
Identical. However, the estimate for goodness-of-fit of the output values were consistently better using the Fltch-Margollash method.
Therefore, all trees presented in this thesis were constructed using
the Fitch-Margoliash method. I might add that both approaches have 14 two major strengths. First, neither approach has an assumption of homogeneous rates of change. Second, the phylogenetic trees so constructed can be evaluated by comparing original input values (from the distance matrices) and output values (from the branch lengths along the trees). The evaluation used in this study was the F value introduced by Prager and Wilson (1978), which is defined as lOOf/I, where f is the sum of the absolute values of difference between input and output and I is the sum of the input values. The most probable tree then becomes the tree with the lowest F value. As indicated by
Cronin and Sarich (1975) and Sarich and Cronin (1977), a second internal test of the accuracy of the resultant tree is the percent nonreciprocity. Empirically, the F value of the best tree generally approximates the nonreciprocity values of the input matrices; therefore, most of the disagreement between input and output values can be accounted for by the nonsensitivity of immunological method (Sarich and
Cronin, 1977).
It cannot be overemphasized that objectivity in tree construction and the application of the molecular clock concept require that no a^ priori assumptions concerning equal rates of change be involved. In
tact, homogeneous or heterogeneous rates of change are internally evaluated by analysis (Sarich and Wilson, 1967; Cronin and Sarich,
1975; Sarich and Cronin, 1977; Wilson et al., 1977). The means of
testing for constant or unequal rates of change involves an outside
reference point; therefore, rates of change along the lineages of
interest are apportioned relative to the reference point (Sarich and
Wilson, 1967; Sarich, 1969a.; Cronin and Sarich, 1976; Sarich and 15
Cronin, 1977). The only assumption in this approach is that the group being evaluated forms an independent or monophyletic group relative to the outside reference point. Thus, this outside reference point can be used to determine both branching sequences and rates of evolution.
In addition, the reference point provides a means of rooting the tree
(Farris, 1972).
The electrophoretic data for the Desmodontinae were also analyzed cladlstleally using a locus-by-locus approach as advocated by Baverstock et al. (1979). In the locus-by-loeus approach, Glossophaga sorlclna and Stumira lllium were used as outside reference species. Any allele at a given locus present in either Glossophaga or Stumira and any vampire bats species was considered primitive, and all other alleles at this locus were considered to be either unique derived or shared derived.
Specimens Examined
Specimens used in this study were collected at the following localities. Noctilio leporinus - Mexico: Chiapas; Pijijiapan, (1).
Mormoops megalophylla - Mexico: Guerrero; 24.1 ml N Rio La Union on
Hwy 200, (1); Queretaro; 8.2 mi S Plna Blanca on Hwy 120, (2).
Pteronotus parnellii - Mexico: Sinaloa; 1 ml N El Dorado, (3); 2 mi
NE Rosarlo, (1); Nayarit; 0.4 ml E Hwy 15 on road to Acaponeta, (1);
Guerrero; 24.1 ml N Rio La Union on Hwy 200, (1). Micronycteris megalotls - Suriname: Brokopondo; 1 km N Rudi Kappelvliegveld, 3° 48'
N, 56° 48' W, (1); 3 km SW Rudi Kappelvliegveld, 3° 46' N, 56° 10' W,
(2). M. schmldtorum - Nicaragua: Zelaya; Muella de las Vueyes on Rio 16
Nico, (1). M. mlnuta - Suriname: Brokopondo; 1 km N Rudi
Kappelvliegveld, 3° 48' N, 56° 8' W, (1); Commewijne; Nleuwe Grond
Plantation, 5° 53' N, 54° 54' W, (1). M. hirsuta - Nicaragua: Zelaya;
3 km NW Rama, (1). M. nicefori - Suriname: Brokopondo; 1 km N Rudi
Kappelvliegveld, 3° 48'N, 56° 8' W, (1); Commewijne; Nleuwe Grond
Plantation, 5° 53' N, 54° 54' W, (1). M. sylvestris - Suriname:
Brokopondo; Brownsberg Nature Park, 8 km S, 2 km W Brownsweg, 4° 55'
N, 55° 11' W, (7). M. daviesi - Suriname: Saramacca; Raleigh Falls,
4° 44' N, 56° 12' W, (1). Macrotus waterhousii - Mexico: Guerrero;
42 mi SW Iguala on Hwy 95 in Zapilote Canyon, (2). Lonchorhina aurita
- Mexico: Chiapas; 23.2 km SE Tonala on Hwy 200, (3). Maerophyllum macrophyllum - Costa Rica: Limon Province; 2.5 km S Limon, (13).
Tonatia bldens - Suriname: Saramacca; Voltzberg, 4° 40' N, 56° 12' W,
(1); Nlckerie; Grassalco, 4° 46' N, 56° 46' W, (1); Brokopondo; 1^5 km W Rudi Kappelvliegveld, 3° 47' N, 56° 10' W, (1). T. brasiliense -
Suriname: Commewijne; Nieuwe Grond Plantation, 5° 53' N, 54° 54' W,
(1). T_. nicaraguae - Nicaragua: Zelaya; 3 km NW Rama, (1). T^. venezuelae - Venezuela: Guarico; 45 km S Calabozo Gallery Forest, (1).
T. carrikeri - Suriname: Saramacca; Voltzberg, 4° 40' N, 56° 12' W,
(1). T. sllvlcola - Suriname: Brokopondo; Brownsberg Nature Park,
8 km 3, 2 km W Brownsweg, 4° 55' N, 55° 11' W, (6). T_. schulzi -
Suriname: Brokopondo; 3 km SW Rudi Kappelvliegveld, 3° 46' N, 56° 10'
W, (1). Mimon crenulatum - Suriname: Brokopondo; 3 km SW Rudi
Kappelvliegveld, 3° 46' N, 56° 10' W, (6). Phyllostomus discolor -
Suriname: Brokopondo; Brownsberg Nature Park, 8 km S, 2 km W
Brownsweg, 4° 55' N, 55° H' W, (1); Brokopondo; Rudi Kappelvliegveld, 17
3° 47' N, 56° 8' W, (1); Saramacca; Raleigh Falls, 4° 44' N, 56° 12'
W, (1). 2- hastatus - Suriname: Saramacca; Raleigh Falls, 4° 44' N,
56° 12' W, (3). 2- elongatus - Suriname: Brokopondo; Brownsberg
Nature Park, 8 km S, 2 km W Brownsweg, 4° 55' N, 55° U' W, (1);
Brokopondo; Rudi Kappelvliegveld, 3° 47' N, 56° 8' W, (1), 2- latlfollus - Suriname; Brokopondo; 1 km N Rudi Kappelvliegveld, 3° 48'
N, 56° 8' W, (1); 3 km SW Rudi Kappelvliegveld, 3° 46' N, 56° 10' W,
(2); iJj km W Rudi Kappelvliegveld, 4° 47' N, 56° 10' W, (1).
Phylloderma stenops - Suriname: Brokopondo; 3 km SW Rudi
Kappelvliegveld, 3° 46' N, 56° 10' W, (2); Nlckerie; Grassalco, 4° 46'
N, 56° 46' W, (1). Traehops eirrhosus - Suriname: Nlckerie;
Sipaliwini Airstrip, 2° 2' N, 56° 7' W, (2); Saramacca; Raleigh Falls,
4° 44' N, 56° 12' W, (3). Chrotopterus auritus - Suriname: Nlckerie;
Sipaliwini Airstrip, 2° 2' N, 56° 7' W, (1); Saramacca; Raleigh Falls,
4° 44' N, 56° 12' W, (1); Brokopondo; 3 km SW Rudi Kappelvliegveld,
3° 46' N, 56° 12' W, (1). Vampyrum spectrum - Suriname: Brokopondo;
Rudi Kappelvliegveld, 3° 47' N, 56° 8' W, (1). Glossophaga sorlclna -
Venezuela: Guarico; 45 km S Calabozo, (4). Monophyllus redmani -
Jamaica: St. Catherine Parish; 0.2 mi E Watermount, (4);
Westmoreland Parish; Bluefields, (10); St. Ann Parish; Orange Valley,
(10). M. plethodon - Dominica: St. Joseph Parish; 1 mi from mouth of
Layon River, (1); St. Paul Parish; Springfield, (9). Leptonycteris sanbomi - Mexico: Sonora; 6 mi NW Alamos, Kinas Arraalillo, (1).
Lonchophylla thomasi - Suriname: Brokopondo; Brownsberg Nature Park,
8 km S, 2 km W Brownsweg, 4° 55' N, 55° 11' W, (1). Lionycteris spurrelll - Suriname: Nlckerie; Grassalco, 4° 46' N, 56° 46' W, (1). 18
Anoura caudlfer - Suriname: Brokopondo; 1 km Rudi Kappelvliegveld, 3°
48' N, 56° 8' W, (1). Hylonycteris underwoodi - Mexico: Chiapas; 8 ml
N Berriozabal, (1). Choeroniscus minor - Suriname: Nickerie;
Grassalco, 4° 46' N, 56° 46' W, (1). Carollia perspicillata - Suriname:
Nickerie; Sipaliwini Airstrip, 2° 2' N, 56° 10' W, (4). Rhinophylla pumilio - Suriname: Brokopondo; 3 km SW Rudi Kappelvliegveld, 3° 46'
N, 56° 10' W, (1), Stumira lllium - Venezuela: Miranda; 24 km N
Altagracla, (5); Suriname: Saramacca; Bitagron, 5° 6' N, 56° 4' W,
(1). Artibeus jamalcensis - Trinidad: St. George Co.; Simla, 4 mi N,
1 mi E Arlma, (3). Brachyphylla cavemarum - Dominica: St. Paul
Parish; Stinking Hole, (5); Springfield, (1); Montserrat: St.
Anthony Parish; Belham River, 0.5 mi from mouth of river, (4).
Erophylla sezekomi - Jamaica: St. Catherine Parish; 0.2 mi E
Watermount, (1); Westmoreland Parish; Bluefields, (1); St. Ann
Parish; Orange Valley (5); Portland Parish, 0.8 mi W Drapers, (5).
Phyllonycteris aphylla - Jamaica: St. Ann Parish; Orange Valley, (14).
Desmodus rotundus - Mexico: Queretaro; 6.5 ui NE Pinal de Amoles,
(10); Guatemala: Santa Rosa; 1 mi E Taxisco, (3); Honduras: Valle;
10 ml SSW Nacaome, (2); Morazan; 0.3 ml SE Sabana Grande, (5);
Nicaragua: Zelaya; 4.5 km NW Rama, (3); 3 km NW Rama, (3); 9.5 ml
NW Rama, (4); Costa Rica: Puntarenas; 2.1 mi S, 1.1 mi E San Vito,
(2); San Jose; 41.2 mi SW Canas, (10). Diaemus youngii - Nicaragua:
Managua; 0.75 mi N Massaehapa, (1); Zelaya; 3 km NW Rama, (1);
Costa Rica: San Jose; 41.2 ml SW Canas, (1). Diphylla ecaudata -
Mexico: Queretaro; 6.5 mi NE Pinal de Amoles, (2); Honduras:
Morazan; 0.3 ml SE Sabana Grande, (1). All specimens are deposited 19 in the Carnegie Museum of Natural History except for the following:
1) Tonatia nicaraguae, T_. venezuelae, Brachyphylla cavemarum,
Glossophaga sorlclna, Monophyllus plethodon, Desmodus rotundus, Diaemus youngii, and Diphylla ecaudata are deposited in The Museum, Texas Tech
University; 2) Micronycteris hirsuta and M. schmldtorum are deposited in the Texas Cooperative Wildlife Collection, Texas A&M University;
3) Noctilio leporinus, Macrotus waterhousii, Lonchorhina aurita,
Macrophyllum maerophyllum, Carollia perspicillata, Artibeus jamalcensis,
Hylonycteris underwoodi, and Leptonycteris sanbomi are deposited in the Museum of Vertebrate Zoology, University of California at Berkeley. RESULTS
Noctilionoidea
Antisera to the purified albumins of two mormoopid species
(Mormoops megalophylla and Pteronotus parnellii), one noctilionid species (Noctilio leporinus), and several phyllostomid species (Macrotus waterhousii, Vampyrum spectrum, Glossophaga sorlclna, Carollia perspicillata, and Desmodus rotundus) were used to investigate the phylogenetic relationships among the three noctilionoid families,
Noctilionidae, Mormoopidae, and Phyllostomidae. The results of the immunological comparisons are given in Table 1, and a phylogenetic tree calculated from the data matrix in Table 1 is presented in Fig. 1. The goodness-of-fit of the tree in Fig. 1 as measured by the "F value" is
Mormoops, Pteronotus, and the various phyllostomid albumins are more or less equidistant from one another. Indeed, the albumins of the two mormoopid genera (Mormoops and Pteronotus) are, on the average, somewhat more distant from each other than either is from the albumins of the phyllostomids. If Mormoops and Pteronotus are associated, then one would expect to find that their albumins have changed to a greater degree than those of the phyllostomids. Otherwise, the albumin data could not be readily interpreted as representing mormoopids as a derived unit relative to other bats. Using the family Pteropidae as an outside reference point, the relative amount of change along lineages can be assessed. Mormoops at a distance of 140 units from the pteropid
20 21 Table 1. Albumin immunological distance values for mormoopid and noctilionid species. The phyllostomidae sample consists of Macrotus,
Vampyrum, Glossophaga, Carollia, and Desmodus. The Pteropidae sample consists of Syconycteris, Pteropus, Dobsonia, Nyctimene, and
Paranyctlmene.
Antisera
M.m. P.p. N.l. Ph. Pt.
Mormoops megalophylla 0
Pteronotus parnellii 95 0
Noctilio leporinus 146 150 0
Phyllostomidae 87 80 135 0
Pteropidae 140 178 180 172 0 22
45 Pteronotus 14 1 ^1 35 1 A Phyllostomidae 37 Mormoops 83 Noctilio t, 02 ^^ Pteropidae
Figure 1. A phylogeny of the Noctilionoidea lineages based on the albumin immunological distances of Table 1. The numbers along the lines represent the amount of albumin change allocated to a particular lineage. The Pteropidae are used as an outside reference point to root the tree and apportion the amounts of change among lineages. 23 reference is the least changed albumin, whereas Pteronotus is 178 units and the phyllostomid mean is 172 units (Table 1). These data may be most parsimoniously explained (as indicated in Fig. 1) by the phyllostomids and Pteronotus sharing a common ancestor (separate from
Mormoops) in which there was an albumin change of approximately 30 AID units.
The albumin of Noctilio is divergent in the same way as that of other noctlllonoids, being some 180 units from the pteropids (Table 1); however, it is quite different from those of other noctilionoid lineages.
The distances separating Noctilio from Pteronotus, Mormoops, and the phyllostomids are 150, 146, and 135 units, respectively (Table 1). The phylogenetic implications of these data are not equivocal. However, the apportionment of the Noctilio-phyllostomid distance of 135 units, using the pteropid information, would suggest a Noctilio separation somewhere along the common noctilionoid lineage as shown in Fig. 1.
Subfamilies of Phyllostomidae
Antisera to the albumins of 14 species of phyllostomid bats were used to assess subfamilial relationships within the Phyllostomidae.
The data presented in Table 2 are the result of reciprocal Immunological comparisons among 13 of the 14 antisera. The nonreciprocity was 11.1% before correction and 7.4% after correction. Figure 2 is the phylogenetic tree constructed from the data matrix in Table 2. The estimated goodness-of-fit for the tree in Fig. 2 is F = 7.2%.
The placement of Artibeus jamalcensis (representing the subfamily
Stenodermatinae) relative to the lineages in Fig. 2 was determined by 24 a. CL3
CO o CM C3
CO CO 03 o •H (U en 1-1 ft B U t-t« X CQ CM on u-l 3 09 H CO U tJ CO *J "O O en a\ U-) 00 CO CM Q* -^ en sf en r^ r». 03c CcO CO »OcMp>.u-irvivooocM u u en<-u-iu-ienu-im^in CcO M XJ CO C •cH •cH B u 3 C (0 U3 o ^ TS M CO o u •H e CO 10 •V CO 0) 0) >, 03 (1) CO •H .c U 3 H> 0) U a > a O CO 3 CO m X C>O! "O i CO CO 3 a C CO •H 3 V4 C>^ B> , -O 03 r-to >i X 4J >N .c o r-t r-t o Mo Uo o o. X a o o o^ >, u CJ a CO >N CO X CO X •H •uH e c CO H (U o U X CO CD S CO > s g c I Micronycteris sylvestris Micronycteris daviesi Vampyrum spectrum Phyllostomus hastatus Tonatia silvicola Carollia perspicillata Q c ^'^•Brachyphylla cavemarum 11 Phylloncyteris aphylla Glossophaga soricina Monophyllus plethodon Desmodus rotundus Diphylla ecaudata Macrotus waterhousii Figure 2. An albumin phylogeny for the family Phyllostomidae based on the values in Table 2. The numbers along the lines represent the amount of change allocated to a particular lineage. Artibeus (representing the Stenodermatinae) can be aligned with Carollia from the values in Table 3. 26 reciprocal immunological comparisons of the Artibeus jamalcensis antisera to the antisera of the major lineages in Fig. 2. Again, percent nonreciprocity was estimated, and the values were 8.0% before correction and 5.3% after correction. The resulting data matrix from these reciprocal comparisons is shown in Table 3. From these data, Artibeus can be shown to associate with Carollia perspicillata (representing the Carolliinae) by an intemode length of 5 AID units with an additional 23 units of change along its lineage after the divergence from Carollia. The goodness-of-fit relative to this association is F = 2.1%. Next, a relative rate test of certain lineages depicted in Fig. 2 plus the Artibeus lineage was applied using Pteronotus parnellii as the outside reference point. The immunological distance estimates are presented in Table 4 and were derived from reciprocal comparisons. The percent nonreciprocity before correction was 11.5% and after correction was 5.7%. Figure 3 represents the phylogenetic associations derived from the distance matrix in Table 4. The F value for this tree is 5.5%. As can be seen from Fig. 3 the phylogenetic associations depicted in Fig. 2 are the same, and the amounts of change along the lineages in Fig. 3 approximate the degree of change along the same lineages in Fig. 2. Using the Pteropidae (values in Table 4) as another reference point indicates identical associations as well as similar rates of change along the lineages examined. The F value for the tree involving pteropids as the outside reference point was 2.1%. Overall, the data from Figs. 2 and 3 indicate the following: 1) an intemode length of 1.3 AID units suggests a weak association of 27 Table 3. Albumin immunological distances comparing Artibeus (Stenodermatinae) to major lineages depicted in Fig. 1. Antisera M-D P-T V-M G-B C A Macrotus-Desmodus 0 Phyllostomus-Tonatla 73 0 Vampyrum-Micronycterls 49 47 0 Glossophagines-Brachyphylllnes 59 51 38 0 Carollia 79 65 48 56 0 Artibeus 64 64 43 49 53 0 28 Table 4. .\lbumin immunological distances among certain lineages depicted in Fig. 1, Artibeus, and two outside reference points, Pteronotus parnellii and the Pteropidae. The Pteropidae sample consists of the same genera indicated in Table 1. Antisera Ma Va Gl Ar Ca De Ph Pt Macrotus 0 Vampyrum 69 0 Glossophaga 83 50 0 Artibeus 80 41 54 0 Carollia 90 62 67 48 0 Desmodus 48 43 84 63 83 0 Phyllostomus 103 65 70 62 87 96 0 Pteronotus 98 74 81 67 96 79 101 Pteropidae 174 158 172 186 29 7 I Artibeus jamaicensis 34 1.7 Carollia perspicillata 44 5.5 'Phyllostomus hastatus 29 Glossophaga soricina 30 Macrotus waterhousii 16 18 t Desmodus rotundus 48 16 Vampyrum spectrum Pteronotus parnellii Figure 3. An albumin phylogeny for the family Phyllostomidae based on the distance matrix in Table 4. The tree is rooted with Pteronotus, and the amounts of change (number along lineages) are apportioned among the phyllostomine lineages in reference to Pteronotus. 30 Micronycteris sylvestris, M. daviesi, and Vampyrum spectrum (all of the subfamily Phyllostominae); 2) Phyllostomus hastatus and Tonatia sllvlcola, representing the Phyllostominae form a clade relative to other genera with an intemode length of 8.5 AID units; 3) Carollia perspicillata (Carolliinae) and Artibeus jamalcensis (Stenodermatinae) are associated by an intemode length of 5.5 AID units and are most closely aligned to the Phyllostomus-Tonatla clade; 4) the Brachyphyllinae (represented by Brachyphylla cavemarum and Phyllonycteris aphylla) are associated with the Glossophaginae (represented by Glossophaga soricina and Monophyllus plethodon); 5) the Brachyphyllinae-Glossophaginae clade is aligned with the Phyllostomus-Tonatia-Carollia-Artlbeus clade by an internode length of 4.4 AID units; 6) Macrotus waterhousii (Phyllostominae) and the Desmodontinae (represented by Diphylla ecaudata and Desmodus rotundus) form a clade relative to the other genera. Phyllostominae In addition to the phyllostomine genera examined during the evaluation of subfamilial relationships, unidirectional tests were performed to determine the placement of five additional genera of phyllostomines (Mimon crenulatum, Lonchorhina aurita, Macrophyllum maerophyllum, Traehops eirrhosus, and Chrotopterus auritus) relative to the tree depicted in Fig. 2. Although these unidirectional tests are one-way comparisons and do not determine relationships among the genera being placed in this manner, they do provide information concerning the phylogenetic affinities of these genera relative to the 31 existing lineages on the already constructed phylogenetic tree. The immunological distances of these genera to eight representative antisera are presented in Table 5. The placement of these five genera can be summarized as follows: 1) Mimon is associated with the Phyllostomus-Tonatla clade by an intemode length of 1 AID unit (F = 6.1%). Ihe amount of change along the Mimon lineage is 11 AID units; 2) Lonchorhina is associated with the Phyllostomus-Tonatla clade (F = 4.8%); however, this association is no closer than that of Carollia; 3) Macrophyllum is associated with the Phyllostomus-Tonatla clade by an intemode length of 4 AID units with 12 units of change along its -ineage (F = 4.5%); 4) Traehops is associated with the Vampyrum lineage by an intemode length of 3 AID units (F = 6.3%). Six units of change have occurred along the Traehops lineage; 5) Chrotopterus is strongly associated with Vampyrum by an intemode length of 8.5 AID units (F = 4.6%). The amount of change along the Vampyrum and Chrotopterus lineages after their divergence from one another is 2 and 1 unit, respectively. Unidirectional tests were used to assess intrageneric relationships in the three phyllostomine genera, Tonatia, Micronycteris, and Phyllostomus. Immunological distances were calculated for each Tonatia and Phyllostomus species relative to the antisera of Tonatia silvicola and Phyllostomus hastatus, and the amounts of change along these lineages were apportioned using an outgroup (mixed pool) consisting of eight reference antisera (Table 6). The results of this test indicate the Tonatia bldens, T^. carrikeri, and T_. schulzi cannot be equivocally associated with the T. silvicola lineage. Tonatia bldens is appreciably 32 u CJ\ CM en o CO •a CO •a o 1 (U tH CU r-l r-t • SI CO pO t^ rH en u-l en c CO X>, CO •H i-i en en en en en a 3 00 -H i-i distant from T_. silvicola and associates better with the Phyllostomus lineage by an internode length of 5 AID units (F = 1.5%). The amount of change along the I_. bldens lineage is 16 units. Although T^. carrikeri (F = 1.0%) and T^. schulzi (F = 3.0%) do not fit on the T_. silvicola lineage, an association with the Phyllostomus lineage is not apparent. The rates of change along the T^. carrikeri and T^. schulzi lineages are 8 units in both cases. Immunological distances of T^. venezuelae, T_. brasiliense, and T^. nicaraguae relative to Phyllostomus are not available, but an immunological distance of 14 relative to T^. sllvlcola implies a close association of these taxa to T^. silvicola. .\s estimated from the distance matrix in Table 6, all the species of Phyllostomus including Phylloderma stenops are associated with the Phyllostomus lineage. The results are as follows: 1) Phyllostomus discolor is connected to the P^. hastatus lineage by an internode length of 10.5 AID units and has 17 units of change along its lineage (F = 3.3%); 2) P^. elongatus is strongly associated with P^. hastatus by an intemode length of 19.5 AID units (F = 4.1%). The amount of change along the P. hastatus and P^. elongatus lineages since their divergence from each other is 11 and 4 units, respectively; 3) P^. latifollus is connected to the ?_. hastatus lineage by an Intemode length of 8 units with 9 units of change along its lineage (F = 9.2%); 4) although most distant from ^. hastatus, Phylloderma stenops is associated with P. hastatus by an internode length of 7 units and has accumulated 24 units of change since the divergence from P^. hastatus (F = 2.4%). Species of Micronycteris were compared to the antisera of Vampyrum spectrum, Micronycteris sylvestris, and M. daviesi as well as an 35 outgroup pool consisting of Macrotus waterhousii, Desmodus rotundus, Glossophaga soricina, and Carollia perspicillata (Table 7). Nearly all species of Micronycteris are clearly associated with the M. sylvestris- >I. daviesi clade. Micronycteris schmldtorum (F = 2.8%) and M. megalotls (F = 5.0%) associate with the M. sylvestris-M. daviesi clade by an intemode length of 5 units. The units of change along the M. schmldtorum and M. megalotls lineages are 12 and 23, respectively. The intemode lengths connecting H. nicefori (F = 3.8%) and M. hirsuta (F = 3.0%) to the M. sylvestris-M. daviesi clade are 2.5 and 2 units, respectively. Micronycteris nicefori has accumulated 19 units of change along its lineage whereas M. hirsuta has accumulated 29 units. Micronycteris mlnuta is the most divergent taxon and cannot be associated with the M. sylvestris-M. daviesi clade. Glossophaginae and Brachyphyllinae Reciprocal measurements among the albumins of Monophyllus plethodon, Phyllonycteris aphylla, Brachyphylla cavemarum, and Glossophaga soricina were carried out using antisera to each of the four. The nonreciprocity value was 6.6% before correction and 5.5% after correction. Amounts of change along the four lineages were assessed using antisera from an outgroup consisting of six reference albumins - Carollia perspicillata, Micronycteris daviesi, M. sylvestris, Tonatia sllvlcola, Vampyrum spectrum, and Phyllostomus hastatus. The additive tree developed from these data is presented in Fig. 4. In Table 8, I present the albumin immunological distances (AIDs) of Erophylla sezekomi and several glossophagine genera to the four 36 03 a X 3 u O CJ\ u-l CO \o en CO en -3^ • M B P G O Monophyllus 0 32 26 35 37 o Brachyphylla 31 0 24 36 47 u Phyllonyaeris 25 24 0 30 42 T Glossophaga 33 36 30 0 50 P Outgroup 39 47 43 51 0 INPUT Phyllonyaeris—Erophylla 1.7 1 "^1-^.7 4.4 Brachyphylla 19.3 Glossophaga 11 Monophyllus t ! •^ 26.J Outgroup Figure 4. An albumin phylogeny of the Brachyphyllinae, Glossophaga, and Monophyllus. The input values represent the original distance matrix, and the output values represent distances calculated from the tree. The tree is rooted by an outgroup (see text), and the amounts of change are apportioned in reference to the outgroup. 38 aj B O U 1H CO 44 B >< —4 r>» CI 1H 1H 00 B o\ CM rH O CM cn T-^J 1-1 CO >% "O X C "3. CO >> CJ CO CD rH U rr ea X> ! a. oa« 0 ns>. 00 3 B lH V 44 s B CO B > > o c B Cfl 0 CJ (U ta rl E>> 0 4= >. rl o u CD 03 0 44 CJ B> "a X o "3 O r-l a 3 B O 0 « rH B r» >, lU O 0 •H tH H 00 CO ij x ,.9 •4. This procedure indicates an association of Erophylla to Phyllonycteris (F = 5.5%) resulting in a Phyllonyeterls-Erophylla lineage as indicated in Fig. 4. The results of this procedure for the glossophagine genera are as follows: 1) Leptonycteris sanbomi (lineage length = 12.4 units) is associated with the Glossophaga lineage by an intemode length of 6.5 units (F = 7.0%); 2) Choeroniscus minor (lineage length = 4.1 units; F = 3.0%) and .'^oura caudlfer (lineage length = 9.5 units; F = 2.5%) are associated with the Monophyllus lineage by intemode lengths of 4 and 3.5 units, respectively; 3) the position of Hylonycteris is more tenative and can be placed near the base of either the Glossophaga (F= 3.3%) or Monophyllus (F = 4.0%) lineage, and in both cases, the lineage length is 8 units; 4) both Lonchophylla thomasi (lineage length = 29 units) and Lionycteris spurrelll (lineage length = 29.5 units) represent independent lineages being separated from all four glossophagine and braehyphylline genera (F = 2.8% and 2.1%, respectively). Although the relationships of these two genera cannot be assessed, rate test data from outgroup comparisons suggest that the degree of immunological divergence cannot be attributed to differential rates, but rather to a branching off of these genera prior to the common ancestry of the other glossophagine-brachyphylline lineages. 40 Carolliinae One-way comparions of Rhinophylla pximlllo to the major lineages depicted in Figs. 2 and 3 were used to determine the phylogenetic affinity (or lack of affinity) between the two Carolliinae genera, Rhinophylla and Carollia (Table 9). Rhinophylla can be placed on two lineages, and in both eases, the F value is 5.1%. The first placement indicates that Rhinophylla is aligned with Carollia by an intemode length of 3.5 AID units, whereas the second placement aligns Rhinophylla with Micronycteris by an Intemode length of 5 AID units. Both placments indicate that Rhinophylla has experienced 19 units of change along its lineage. If Rhinophylla is compared to the 13 independent antisera from Table 2, the alignment with Carollia has an F value of S.IZ whereas the association with Micronycteris is 8.6%. In both cases, the F value is higher than the percent nonreciprocity (7.2%) as determined for the antisera represented in Table 2. Desmodontinae Twenty-two presumptive loci encoding 15 enzymes and one non enzymatic protein were assayed for the three vampire bat genera. Of the loci examined, five (G6p, Idh-2, Ldh-1, Ldh-2, and Mdh-2) are monomorphic for the same allozyme in all three genera of vampire bats. The other 17 loci (Table 10) can be divided into the following three groups: 1) four loci are monomorphic for a different allozyme in each genus; 2) five loci segregate for the same allozyme in two genera with an alternate allozyme for the third genus; 3) several loci 41 Table 9. Albumin immunological distances comparing Rhinophylla to the major lineages depicted in Fig. 2. Antisera M V Mi. B-G C P-T Macrotus 0 Vampyrum 63 0 Micronycteris 59 23 0 Brachyphyllines-Glossophagines 70 36 39 0 Carollia 85 54 45 56 0 Phyllostomus-Tonatla 85 45 48 51 65 0 .\rtlbeus 86 41 40 49 53 64 0 Rhinophylla 80 39 26 54 45 55 52 CO 42 0) CD 1H rH 00 CJ 0) cd 0 rH X' I-I r-l Ok CtH CD a o Ll (0 O 03 CO CO viH 43 o u 44 r-t 03 CO C <4H o 03o 00 fto o 03 CD 44 03 iH iH q o o U > rH h X 0) lH f 44 U X o o X o g y 03 3 CO CO CO CO X X s I X CD 4H I 44 <-t CA •a CO 00 a T3 c CD O X Bo 00 4J CD 0) CD CO r^ en <-l U r-t vo en B a. r-t o B CO >•, o o iH CO X V O CO CU 0) 0) •ff N*." 44 CO ca o. X CJ T) CO c CJ u B vO p^ r^ •H 00 -rl CO vO —< .rt e 3 0) 01 C S o o o 00 X 00 u CD >*r' \^ v*./ iH •a u CSO I I •a CJ 43 » CO lH o u o X o ^^\ CO <4H 1 44 CO CO lU vO >a- 44 44 iH cyi o lH 44 • • lH r-t CO • o o CJ CO 00 >> < CD CO •a o CJ 0) 44 I •o U -o .A r-to lH I o a r-t rJ 03 iH < /—V ^-^ X 43 .—N f-\ 4-1 m un un u^ un CD o ON o ON o VI CO. s a . u iH r-t rH « • o o o v-o^ ^o^ MH C CO e' CO CO s-^ v«/ Vi • CO CO en r«» en vO o o X vo .O CD a en r^ en r* r^ cn CO -rt CO rH vo en CO 3 £ o o o o o o CU .O U 43 CD CU 43 CO O u-l m CM 00 en r^ o\ o o CO -H CO rH . . • • • • • V—o / o^^ s-o^ S.o^ N.o^ Vwo . ^wo^ 0) T3 y CO y •o o ^-v y—\ ^-\ /—V ^s ^-N •—V •* vO vo en rH U-l u-| ON o u^ en rH ON O ...... ••—yo -^-oz Vwo ' >wo^ S.o^ ^o^ s.o^ 0) 73 CO y 43 •O y CO 3 T3 lO O u-l cri u-l O vo rrt CM ON O £ CO O • o. o. o• o• V v.-/ \m0^ Q ^-cf^l T3 ^—^y -o ^.y^ /-\ ^—s /-s •—\ en t-~ u-l u-l ON O ON O T3 CD CD 03 VMd/ Nr^ v-/ v*d/ 3 CO y S T3 y iH 44 B O y CO rH r-l en CM n CM (3. 3 1 1 1 I I I I X y 42 44 44 4= X X vO CD lU CO CO o I H »oJ •a Cd -a T3 2 2 u a J M rJ CJ 44 exhibit varying degrees of polymorphism. Intraspecific similarities for Desmodus are within the range reported for interpopulatlon variation. Genetic similarity values for all pair-wise comparisons between Desmodus samples averaged 0.98, and no geographic discontinuities were discernable. No electrophoretic variation was found among the three specimens of Diphylla, and those of Diaemus had an average S = 0.98. Since similarity values among samples of all three species were high, and common alleles were shared at all assayed loci, samples were pooled for analysis of genie similarity and distance among the three genera as follows: 1) Desmodus from Mexico; 2) Desmodus from Guatemala and Honduras, 3) Desmodus from Nicaragua; 4) Desmodus from Costa Rica; 5) all Diaemus; 6) all Diphylla. The two measurements of genie divergence, Rogers' S (Rogers, 1972) and Nei's D (Nei, 1972) are presented in Table 11. The similarity and distance values indicate that Diaemus and Diphylla are nearly equidistant from Desmodus. The average value for Desmodus to Diphylla is S = 0.547 (D = 0.578), and between Desmodus to Diaemus, S = 0.540 (D = 0.581). The value between Diaemus and Diphylla is relatively lower with S = 0.350 (D = 1.05). Seventeen loci (Est, Adh, Pgm, and Me being excluded) were examined for the three vampire genera, Glossophaga, and Stumira (Table 10). Two loci (Mdh-1, Mdh-2) appear to be highly conservative and to share the same symplesiomorphic state in all five genera examined. The Diaemus sample also displays an autapomorphic condition (allele a) at the Mdh-1 locus. Two alleles (allele b at the Ldh-1 locus and allele b at the 45 Table 11. Coefficients of genie similarity and distance among the three genera of vampire bats. Above the diagonal, Rogers' coefficient, S (Rogers, 1972); Below the diagonal, Nei's distance, D (Nei, 1972). Desmod us Diaemus Diphylla 1 2 3 4 5 6 1 0.9843 0.9732 0.9829 0.5486 0.5468 2 0.0033 0.9866 0.9821 0.5340 0.5493 3 0.0059 0.0012 0.9758 0.5313 0.5469 4 0.0017 0.0021 0.0035 0.5456 0.5443 5 0.5720 0.5871 0.5898 0.5744 0.3501 6 0.5805 0.5760 0.5758 0.5804 1.0506 46 Idh-2 locus) can be considered symplesiomorphic for Glossophaga and the three vampire bat genera. Allele d at the Ipo-2 locus is symplesiomorphic for Glossophaga. Desmodus, and Diphylla with Diaemus and some individuals of Desmodus having an autapomorphic condition (alleles a and c, respectively). The b allele at the ldh-1 locus is shared among Diaemus, Diphylla, and Stumira. Thus, the ldh-1 d and e alleles could therefore be considered synapomorphic for Diaemus and Desmodus. The vampire genera are linked by nine biochemical characters. Three loci, G6p, Ldh-2, and Pgi-1 (Desmodus also has a unique condition, allele e, at this locus), display the same condition in all three genera. Six additional loci, Got-1, Got-2, a-Gpd, Lap, ldh-1, and 6Pgd, reveal the same condition in two of the three genera and a inique condition in the third genus. The unique condition is found in either Diaemus or Diphylla but not Desmodus. Using Nei's D values and the Fitch and Margoliash method, a cladogram can be constructed as seen in Fig. 5. This analysis indicates a Desmodus-Dlaemus clade relative to Diphylla with a more rapid change along the Diaemus lineage. Rate tests using Stumira^ lllium, Brachyphylla cavemarum, Erophylla sezekomi, Phyllonycteris aphylla, and Monophyllus redmani as outside reference species to the vampires reveal the same associations to the one depleted in Fig. 5. The immunological distances among the vampire albumins were measured using Desmodus and Diphylla antisera, and a tree was constructed from the resulting data matrix (Fig. 6). These data indicate that Desmodus and Diaemus are more closely aligned to each 47 (1) (2) (3) (4) (1) Desmodus 0 56 49 118 O U (2) Diaemus 56 0 81 150 T P (3) Diphylla 45 85 0 126 U T 23 147 127 0 NPUT 12 Desmodus 8.5 44 Diaemus 2i5. 5 Diphylla \ ^ 97.2 * Glossophaga Figure 5. A phylogenetic tree depicting the relationships among the three vampire bat genera. This tree was derived from Nei's electrophoretic distance values. The input values represent the original Nei's D values (multiplied by 100). The output values are calculated frcjm the derived tree. The numbers along each lineage represent the amount of electrophoretic change and are apportioned relative to Glossophaga. The F value for this tree is 2.9%. 48 Desmodus Diphylla Desmodus 0 10 Diaemus 5 12 Diphylla 14 0 Desmodus Diaemus 10 Diphylla Figure 6. An albumin phylogeny for the vampire bats. The matrix indicates albumin immunological distances calculated from albumin cross-reactions to antisera of Desmodus and Diphylla. The numbers along the lineages are amounts of albumin change (measured in immunological distance units). 49 other than either is to Diphylla. Rates of Albumin Evolution A test for regular rates of albumin evolution along lineages was applied in the following manner. First, immunological distance values (AID) along independent lineages were determined by calculating distances back to a common ancestor. The common ancestor represents the primary node connecting all lineages being compared. Second, using a null hypothesis of equal rates, the expected mean amount of change for all lineages was compared to the observed amount of change along each lineage and tested by Chi-Square. When all the Noctilionoidea lineages (Fig. 1) are examined in this manner, albumin evolution has not occurred at a regular rate (Table 12). The rate change along the Noctilio lineage is high, and when Noctilio is excluded, the three remaining lineages (Mormoops; Pteronotus, and Phyllostomidae) demonstrate a more regular rate of change. Tests for regular rates of change within the Phyllostomidae were applied in three ways. First, I examined only those lineages which weije rate tested using an outside reference point (Fig. 3, Table 12). When all lineages are considered, the rate of change is not regular; however, if Vampyrum is excluded, the other lineages are similar in rates. This suggests a slowdown along the Vampyrum lineage. The clade consisting of the Artibeus, Carollia, Glossophaga, and Phyllostomus lineages does not demonstrate regular rates. Phyllostomus and Carollia have more change along their lineages than Artibeus and Glossophaga. The Macrotus and Desmodus lineages are similar in rate. 50 B lU un u-l u-l u-l CO o o ITl un m un CO. CU o O O O o o 44 u o CO a. d O O o o C-) d d /N X 03 Oi V V A A. PLH a, PL, 04 u •K V a, OH •a 04 •K •ie iH CO * o 03 B CO o lU iH X rH 44 iH B 44 CU u-l y U en on CM rH rH rH o CO B O. X! rl un rH t4H B • un un u-l un Ul un u-l O o O un un u-l o O O O o o O o O o d A a, 04 9 ^ ?/ R o. 04 OH OH 04 OH 04 •K •K •K OO r^ O sr CM en ON CM o "^ o ON m CM o 00 r-- u-l en o p~ en ON vO 00 CM B CO 03 S 71 ~! CO -H ON ^ 00 en CO CM o rH sr sr CM en en en en 0 0 o 0 o o 0 0 44 4-1 ^\ < 44 44 44 44 44 CM rH CN en CM CM CM 00 CM CM A tfl 0) 0*) 0) tn., •rl CD ca CO lH rH B B M r-t l^H •rt « •H Cfl lU o rH ^ cfl r-t tfl r-t 4-1 u rH rH lH rH •H ft y tfl >^r-t 44 u X> > X f^ 4>=> Cfl X>, fi> > n. o. o a B Cu 0 >x U >v o iH M tn# . u .>B^ X Cfl 4= to Cfl H Q y 3 tu y y L3 y iH iH •H 44 4-1 01 44 CO tfl cfl " 44 44 f. o to 00 y u ^4 « )H CO CO to tf.l. tn s u dl >> n « cn oa cn •rl B B •H 3 „ o H G 3 3 r-t o O r-t •o £ tfl O ». QJ .. OJ r-t H EH r-t 0 3 S CO IH 0)•» 0) 42 03 X o O £ u 03 y to CO iH CO iH u M U cn M ac 7) B B 44 CO B 4-1 CO to tn-. tfl 03 a>^. CO V c: iH iH M 3 lH ^4 U 3 3 U a £ 3 3 03 50 10 00 < E 00 < £ £ cfl £ CO CO Cfl o CO « 0 o n M o 03 03 £ X 4= •> 44 43 - to 44 4-1 tn cn > 44 CO s 3 a. CU CO CO (X Cfl 3 tn CO 3 3 44 CO •a M o O •n 0 0 •H o; o 0 01 44 3 0 O >^ CO cn 4-1 r-t cn 44 X ft r-t X O r-t iH o a a CO CO cfl r-t CO cfl iH r-t r-t •rt u X r^ a o 0 B O B 44 >S 4-1 o 44 >, o £ r-t r-t O rH 0 44 X>> M 4>3> 43 IH jfl •H 43 CO tj CJ) H 04 O H 04 04 CO < < Z 3 PH O > X CO X H 04 52 Second, rates of change along the lineages in Fig, 2 plus the Artibeus lineage were compared (Table 12). When all lineages are considered, disparity in rates of change is apparent. The Micronycteris and Vampyrum lineages are similar in rate and are slow relative to other lineages. The Glossophaginae (Glossophaga and Monophyllus) and Brachyphyllinae (Phyllonycteris and Brachyphylla) are also similar in rates. When all the lineages in the clade consisting of the Glossophaginae, Brachyphyllinae, Tonatia, Carollia, Artibeus, and Phyllostomus are examined, the rates are not similar. However, the exclusion of Phyllostomus and Carollia (which are similar in rate) suggests similar rates of change. The Phyllostomus lineage relative to Tonatia has a higher rate of change; nevertheless, when Phyllostomus, Tonatia, Carollia, and Artibeus are considered together, the rates are similar. Unlike the rate tested lineages (in Fig. 3), the vampire bat lineages (Desmodus and Diphylla) are different in rate from Macrotus. When Vampyrum (slow rate), Micronycteris (slow rate), Phyllostomus (fast rate), and Macrotus (fast rate) are excluded, the remaining lineages are similar in rate. My third approach was to graphically demonstrate rates of change in the phyllostomids thus far examined. I grouped relative rates of all lineages (both reciprocal and unidirectional comparisons) as in Fig. 7. The values for these lineages are presented in Table 13. Wilson et al. (1977) suggest that uniformity in rates should be indicated if data plotted in this manner fit a unimodal distribution. Clearly, three and possibly four modes can be seen in Fig. 7. Even more interesting is the lack of phylogenetic continuity in the distribution of rate changes. 53 8 10 14 17 21 25 29 33 37 41 45 48 Amounts of Change Per Lineage Figure 7. A bar graph of the amount of albumin change per lineage (ordinate) versus the number of lineages (abscissa) demonstrating a particular range of albumin change. The amount of albumin change per lineage is shown in Table 13, and these values are from both reciprocal and unidirectional comparisons. 54 a un en en 38 . cn en en en 39 . ••3 sr sr 48 . U U CO tn 3 rH > as to i) •H iH 3 tn tfl 44 •H B to 3 ~l 44 M cn V4 0 rH CO 44 Cfl cn lH 44 r-t tfl 0 3 •H 4-1 0 r-t o 3 fi •rl cn tfl r-t r-t rH cn 4-1 .•H 43 CO B 00 to C8 ..J tn ati f ric i min u hir s cen s (0 hom a lon g ici l urr e isc o to enop i asta ; hous : D -a o •H 4J tn 0) OH "O 44 X Cfl undu s u 0£ pumili o CO to ^ CO 3 tn cn Cfl c tn cn to •J CO 4-1 •rt iH CO •H to iH 0) £ tfl 03 CO ^4 tn at 4J 03 Cfl U U 3 y tfl tfl rH 3 tn 01 3 tfl 3 tfl o rH u u •v fi IH 03 03 03 03 0) 00 •r-) •H £ •rt ex £ E 3 •n 4-1 44 e 4-1 4-1 CO >i 43 U o ^ y y 4o4 o u 5 to >> y tfl y 43 tn 43 44 01 tfl .LJ 0) 44 cn 3 .B >. >. CO >- rH >, a. 3 a tfl CO 44 •rl tn "0 cn 3 •o O. c fi o fi rH c o 03 0 •H 0 y r-t o o o 44 o 0 0 0 r-t o o to X 43 4J r-t r-l r-t r-t r-t o B 44 M r-t IH X> > )H cn •H y tfl r-l fi> , 0 r-t r-t r-t )H E3] iH a y >v y a. y o 4-1 B fi o u S^ y OJ 43 0) •H 43 •H •H •rt r-t u 0 0 4>3^ iH tfl 4>3 , 43 X>- , cfl 2 ai X s OH 2 Q s o < J H Ol hJ CJ O. a. OH s U -3 :Uj 3 fi Cfl 00 en oo r^ oo sr —I en vO 00 ON (3N 00 00 CM en a eM en un vo CO 00 ^\ rl —I rf ^ 1—1 CM vO rH CM CM CM CM CN e^J sr CM CM a. .fi^ OJ £ 3 -3 CO ! 3 O lH iH iH r-( E c en )H 0 •T3 rH to iH 3 Cfl 11 4-1 iH B 44 0 >v rH U M ;c tn CO to 0 TS o 43 r^ 0 to ty to 3 0) OJ Ta •H •rl fi. >, MH B OJ tn 44 w -rt tfl 0 £ e c O .B y U > 3 lH E > 0 4-1 43 43 OJ M U 0. •H y OJ CO ^ > 44 CO to U 3 ?^ c Cfl •H E y •o 0 y CO M iH > rH OJ o 3 IH tn •rl TS U 3 y cn M fi 4 Each mode incorporates representatives from several different clades. Next, albumin immunological and Nei electrophoretic distance values for 55 species pairs (Table 14) were used to assess linear as opposed to a curvilinear relationship between these two estimates of sequence divergence. For these analyses, I used the General Linear Models Procedure (GLM) in the Statistical Analysis System (Helwig and Council, 1979). Since neither Nei's D nor albumin (AID) are obviously the dependent variable, I performed two analyses with each variable as the dependent one. The statistical models used were ^1 = ^ ^2 ^ h h^ X„ = B.. X.. -(- B„ X. where X is Nei's D and X„ is albumin AID. Because X.. = 0 implies X„ = 0 and vice versa, the intercept was zero in both eases. The significance of the curvilinear terms (B„) indicates that there is a strong non-linear relationship between D and AID irrespective of which variable is considered the dependent one. These curves (Figs. 8 and 9) are significantly curvilinear at the 0.0001 and 0.0006 levels, respectively. Residual variation about the model was examined, and in both cases, was not found to differ from that expected from a normal distribution with a fixed variance. The pattern does not change if mormoopid and noctilionid pairs are considered; indeed, the lines are even more curved. 56 Table 14. Electrophoretic distance (Nei's D) and albumin immunological distance (AID) estimates for 55 species pairs. Species P^alrs Nei's D AID 2- sil.-T. bid. 1.283 47 I- sll.-T. sch. 0.793 33 T. sil.-T. carr. 0.313 28 T. sil.-T. bra. 0.577 19 2- sil.-T. nic. 0.957 19 T^. sil.-T. ven. 1.612 19 T. s11.-Mimon 0.694 42 T. sil.-P. dis. 1.207 56 M. dav. -M. syl. 0.714 13 \A _ dav. -M. meg 1.065 52 M. dav. -M. nic 0.825 37 M. dav. -M. min 1.649 49 M. dav.-M. hir. 1.194 49 M^. dav.-M. sch. 1.253 51 M. dav.-Rhinophylla 1.658 23 M. dav.-Vampyrtmi 2.351 26 M. syl.-M. meg. 1.192 23 >1. syl.-M. nic. 1.007 27 M. syl.-M. min. 1.927 48 M. syl.-M. hir. 1.167 40 M. syl.-M. sch. 1.359 42 M. syl.-Rhinophylla 1.312 28 M. syl.-Vampyrxjm 2.324 26 Vampyrum-Rhinophylla 1.099 42 Carollla-Phyllostomus 1.493 77 Carollia-Glossophaga 0.848 62 Carollla-Desmodus 1.136 72 Carollla-Phyllonyeteris 1.541 54 Carollla-Rhlnophylla 2.234 45 57 Table 14 continued Species Pairs Nei's D AID Desmodus-Rhlnophylla 2.398 55 Phyllonycteris-Rhinophylla 2.398 68 Phyllonyeteris-Phyllostomus 1.252 61 Phyllonycteris-Desmodus 1.705 38 Phyllonycteris-Glossophaga 0.529 30 Desmodus-Phyllostomus 1.657 71 Desmodus-Glossophaga 0.857 73 Desmodus-Dlaemus 0.589 5 Desmodus-Diphylla 0.485 12 Dlphylla-Diaemus 0.861 12 Glossophaga-Brachyphylla 0.636 36 Glossophaga-Monophyllus 0.429 33 Glossophaga-Erophylla 0.528 32 Phyllonyeterls-Braehyphylla 0.343 24 Phyllonyeterls-Monophyllus 0.526 25 Phyllonycteris-Erophylla 0.269 11 Monophyllus-Braehyphylla 0.631 31 Monophyllus-Erophylla 0.523 28 Braehyphylla-Erophylla 0.437 23 Diphylla-Glossophaga 1.027 73 Phyllostomus-Glossophaga 1.252 70 Phyllostomus-Rhinophylla 2.350 59 Glossophaga-Rhinophylla 1.705 49 Noctilio-Pteronotus 1.504 150 Noetilio-Mormoops 1.099 146 Pteronotus-Mormoops 1.280 95 58 AID Figure 3. Nei's electrophoretic distance (dependent variable) plotted against albumin immunological distance (Independent variable). The comparisons are between 55 species pairs (Table 14). The two estimates are significantly curvilinear at the 0.0001 level. 59 2.2 2.4 Figure 9. Albumin immunological distance (dependent variable) plotted against Nei's electrophoretic distance (Independent variable) for the 55 species pairs shown in Table 14. These two estimates are significantly curvilinear at the 0.0006 level. DISCUSSION Superfamilial Relationships Immunological, classical morphological, and chromosomal data support the idea that the three families, Mormoopidae, Noctilionidae, and Phyllostomidae, form a monophyletic group. A salient feature of almost all noctilionoid albumins is their immunological distinctiveness relative to those of other bats. The assimiption of microehlropteran monophylly or the use of phylogenetic analyses involving non-bat reference species (Sarich, manuscript) leads to the conclusion that an event producing the antigenic equivalent of 30 to 40 units of albumin immunological distance must have occurred early in noctilionoid history. This event provides an effective phenetic marker that positions the aoctllionoids into a single clade. As revised from Smith (1972) by Arnold et al. (manuscript), several morphological characters can be considered synapomorphic (shared derived) for the noctlllonoids. These characters are as follows: 1) premaxillary bones complete and fused to the maxillary and palatine bones; 2) os penis absent; 3) marked sella turcica; 4) similar orientation of foramen ovale; 5) wart like bumps and ridges on the lower lip; 6) lanceoleate tragus; 7) strong and heavy upper incisors. Chromosomally, the noctlllonoids share 30 pairs of homologous autosomal arms as ascertained from chromosome G-banding (Patton and Baker, 1978). Thus, these authors suggested that the primitive karyotype for the superfamily is essentially like that of the phyllostomid species, Macrotus waterhousii. 60 61 The phylogenetic relationships within the Noctilionoidea clade are unclear. The alignment of Mormoops with Pteronotus (family Mormoopidae) has been maintained since Dobson (1875). The morphological justification for a clade including Mormoops and Pteronotus is not unequivocal, and Mormoops is considered quite morphologically divergent from Pteronotus (Smith, 1972). Nevertheless, certain characters associated with the distal portion of the humerus, the proximal end of the femur, and the tragus may be synapomorphic (Smith, 1972; Arnold et al., manuscript). Although the primitive karyotype for the superfamily has not been tested with appropriate outgroups, the chromosomal data do indicate that Mormoops and Pteronotus are linked by one fission event (Patton and Baker, 1978; Baker and Bickham, 1980). In contrast, the immunological data question the alignment of Mormoops with Pteronotus. Subsequent to the 30 to 40 units of change associated with the noctlllonoids in general, a second event similar in magnitude occurred in the Pteronotus and phyllostomid lineages with Mormoops remaining the least changed. Thus, this second event suggests an alignment of Pteronotus with the Phyllostomidae. It seems premature and somewhat unjustified for me to contend that the immunological data are conclusive in the associations or lack of associations between Mormoops and Pteronotus. The eladlstle significance of the immunological data could be misleading. It is just as premature to suggest that the morphology and chromosomes are conclusive; however, I accept as the most probable hypothesis that the association of Pteronotus and Mormoops is evolutlonarlly valid. At this time, I think it more appropriate to simply provide an immunological alternative and suggest that further 62 research be conducted. Noctilio has a markedly divergent albumin, and most of the change must have occurred along the Noctilio lineage after the separation of Noctilio from the other noctlllonoids. Although Noctilio can be placed along the noctilionoid lineage, the eladlstle associations with any one noctilionoid lineage is not unequivocal. Smith (1972) indicated that Noctilio was the morphologically least derived of all the noctlllonoids; however, he alluded to a Noetlllonidae-Mormoopidae association. Based on a cladistical analysis of Smith's (1972) data, Arnold et al. (manuscript) indicated that a closer association of the Noctilionidae to the Mormoopidae was not apparent. Assuming the primitive karyotype for the superfamily to be correct, five fusion events are shared by the Mormoopidae and Noctilionidae (Patton and Baker, 1978). Thus, the chromosomal data indicate the most definitive placement of Noctilio. In order to resolve the questions of interfamllial relationships among the noctlllonoids as well as the associations of Mormoops and Pteronotus, further research is needed. First, a morphological assessment of the superfamily using qualitative and possibly quantitative characters should be conducted in a eladlstle framework. Second, the primitive karyotype of the superfamily should be tested with an outgroup (possibly the Emballonuroidea). Third, molecular data including transferrin immunological distances and estimates of DNA sequence divergence are needed for the noctlllonoids. As indicated by Cronin and Sarich (1975) and Sarich and Cronin (1977), additional molecular data provide both Independent molecular tests for concordance and a more complete assessment of eladlstle relationships. 63 Subfamilial Relationships Historically, agreement as to the subfamilial relationships within the Phyllostomidae is the exception rather than the rule, and an assessment as to which relationships are validly supported with hard data and which are the products of conjecture is difficult to attain. Nevertheless, several alternatives have been proposed based on morphological characters (dental and skeletal), standard karyotypes, and more recently chromosome banding, and these alternatives need to be addressed in light of the albumin data. The overall phylogenetic associations as revealed by the immunological data (Figs. 2 and 3) do not agree in entlrlty with either the morphological or chromosomal data. One major point of agreement between karyology and morphology is that the Phyllostominae as currently conceived consists of several independent lines of evolution (Walton and Walton, 1968; Slaughter, 1970; Smith, 1972, 1976; Patton and Baker, 1978). The immunological data support the conclusions of morphology and karyology. Furthermore, the albumin data are compatible with the morphology in suggesting that at least two of the Phyllostominae lineages are associated with certain other subfamilies; however, the immunological associations cannot be equated with the Macrotus-type and Phyllostomus-type of divergence as proposed by several authors (Walton and Walton, 1968; Slaughter, 1970; Smith, 1972, 1976). In contrast to the idea of a Macrotus-type giving rise to at least two if not more subfamilies, the albumins indicate that Macrotus is divergent from the other lineages. The immunological status of Macrotus 64 is more compatible with the chromosomal data (Patton and Baker, 1978), but the close immunological association of Macrotus with the Desmodontinae is not supported by either morphology or karyology. Chromosomally, the Desmodontinae is associated with the Glossophaginae, Brachyphyllinae, and Stenodermatinae (Bass, 1978; Johnson, 1979). Based on morphology, the Desmodontinae and Carolliinae are considered to share a common origin (Slaughter, 1970; Smith, 1976). One might suggest that the immunological association of Macrotus and the Desmodontinae is the result of a rate slowdown along the Desmodontinae lineage relative to the other lineages. The relative rate test (Table 4, Fig. 3) is not suggestive of either possibility. It could be that the desmodontlne and Macrotus association is the result of equally divergent albumins relative to other phyllostomid albumins. With the exception of the desmodontlne-Macrotus distance of 45 AID units, all other albumins are quite divergent from Macrotus. However, the distances between the Desmodontinae and several other taxa are similar to the desmodontlne-Macrotus distance (Table 2). The lack of an equallv divergent albumin along the desmodontlne lineage suggests that Macrotus and the vampires shared a common evolutionary origin. Furthermore, the placement of the vampires near the base of the phyllostomid radiation suggests an early origin for the Desmodontinae. This is a rather intriguing hypothesis considering that the desmodontlnes' adaptation to blood feeding is associated with several derived morphological features. The immunological results suggest a major clade (Fig. 2), which can be loosely compared to the Phyllostomus-type as suggested by 65 morphology, but the components of the immunological clade are quite different. This major clade is comprised of three smaller clades as follows: 1) Phyllostomus-Tonatia; 2) Stenodermatlnae-Carollilnae; 3) Brachyphyllinae-Glossophaginae. The Carolliinae-Stenodermatinae clade is weakly associated with the Phyllostomus-Tonatla clade. The immunological association of the Brachyphyllinae to the Glossophaginae and Phyllostomus to Tonatia is in total agreement with the chromosomal data (Gardner, 1977_a ; Patton and Baker, 1978; Baker and Bass, 1979). The chromosomal alignment of the Brachyphyllinae-Glossophaginae clade with the Stenodermatinae (Johnson, 1969) is not supported by immunology. The morphology does not resolve this problem, and in fact, contributes more alternatives (Smith, 1976). Two explanations seem to make the albumin and chromosomal data somewhat more compatible. First, the Carolliinae are chromosomally so derived that G-band homologies cannot be determined by comparison to the primitive (Baker and Bickham, 1980). Second, the Stenodermatlnae-Carollilnae clade and Brachyphyllinae- Glossophaginae clade are associated by the immunology; however, an Intemode length of 1 AID unit places the former clade closer to the Phyllostomus-Tonatia clade. If one allows for the relatively small Intemode length (1 unit), then the common lineages leading to the Stenodermatlnae-Carollilnae clade and the Brachyphyllinae-Glossophaginae clade could have arisen at approximately the same point in time. A somewhat more complicated case of discordance between the immunological and chromosomal data is the association of a Phyllostomus-Tonatia clade with the other subfamilies. The chromosomal data do not suggest such an association (Johnson, 1979), but the morphology at least alludes to 66 the possibility of this association (Walton and Walton, 1968; Smith, 1976). At this time, I have little explanation for this discrepancy. The immunological data suggest yet a third Phyllostominae lineage comprised of Micronycteris and Vampyrum (Fig. 2). An intemode length of 1.3 AID units suggests that this association is rather weak, and as suggested by Cronin and Sarich (1975, 1976), taxa linked by small distances should be considered somewhat tenative. The primary reason for the tenative nature of this association is that both Vampyrum and Micronycteris have undergone slower rates of albumin change relative to other phyllostomids. Therefore, the association may simply reflect conservatism of albumin evolution along both lineages rather than a close eladlstle relationship. The chromosomal data support the idea that Micronycteris represents an independent lineage relative to other phyllostomids (Patton and Baker, 1978). Unfortunately, the chromosome G-banding patterns of Vampyrum are unknown so no direct comparison to Micronycteris can be made. The morphological data are even less clear and cannot be used to resolve this dilemma. When the chromosomal, morphological, and immunological data are considered in reference to the subfamily Phyllostominae, all three data sets suggest that this subfamily is not a natural or monophyletic assemblage. Although the immunological and chromosomal data are not totally concordant in all respects, the major discrepancies concern the Interpretation of morphological relationships. Discordance among the three data sets is somewhat harder to explain. Maxson and Wilson (1975) provided three reasons for discordance. First, one or more of the phylogenles generated from different sources of data could be in error. 67 As suggested by Sarich and Cronin (1977), a test of the albumin phylogeny could best be applied by examining phylogenles produced from other molecules. In the ease of the Phyllostomidae, this is an obvious course of action. Another approach would be to examine morphological evolution in a manner compatible with procedures used in immunological and chromosomal tree construction. Second, disagreement at least with morphology may be due to convergent evolution. Smith (1976) suggests that this is a valid explanation for some of the patterns seen in phyllostomid bats. Third, organlsmal, molecular, and chromosomal evolution may not be proceeding at the same rate. Differential rates of evolution in all three characters within the same group of organisms have been well documented (Wilson et al., 1974; Wilson et al., 1975; Maxson and Wilson, 1979). Thus, one can envision cases where two or more taxa could be evolutlonarlly related yet demonstrate entirely different rates of phyletic change relative to morphology, chromosomes, and possibly molecules. The chromosomal and albumin data reveal an interesting pattern which may have some bearing on the interpretation of organlsmal evolution within the Phyllostomidae. Both sets of data imply that diversification within the Phyllostomidae may have proceeded rather rapidly with three and possibly four lineages originating close to the basal phyllostomid radiation. This idea may explain in part the reason for initial recognition of the Phyllostominae as a monophyletic group as well as explain some discrepancies observed by different Interpretations of morphological evolution. If one considers the possibility of differential rates of organlsmal evolution since this 68 initial radiation, then genera maintaining the more primitive or typical phyllostomine morphology might be grouped together on the basis of retained primitive characters rather than characters which suggest a common ancestry. Obviously, this interpretation is somewhat simplistic, and the possibility of convergence in morphology cannot be excluded from any assessment of organlsmal evolution within the Phyllostomidae. Relationships Within the Phyllostominae All remaining genera of the subfamily Phyllostominae can be associated with either the Vampyrum or Phyllostomus-Tonatia lineages. Although G-band chromosomal data do not exist for Chrotopterus, Vampyrum, and Traehops, Gardner (1977£) suggested on the basis of standard karyotypes that Chrotopterus was more similar to Tonatia with Vampyrijm and Traehops being more closely associated. On the other hand. Slaughter (1970) indicated that Chrotopterus, Vampyrum, and Traehops were morphologically similar. The albumin data concur with Slaughter and indicate that Chrotopterus and Vampyrum are strongly associated. The immunological distance of 3 is the lowest value found between genera of phyllostomid bats. The immunological association of Mimon with the Phyllostomus-Tonatla clade is substantiated by G-band chromosomal data (Patton and Baker, 1978) as well as by morphology (Slaughter, 1970). Although Lonchorhina is associated with the major clade which includes the Phyllostomus-Tonatia group, the placement of this genus is approximately at the main node connecting all three groups including Carolliinae-Stenodermatinae, Phyllostomus-Tonatia, and Brachyphyllinae-Glossophaginae. Unpublished G-band chromosomal data 69 and morphological data (Slaughter, 1970) are more definitive in that both suggest a close association of Lonchorhina with the Phyllostomus- Tonatia group. Gardner (1977a_) indicated that Lonchorhina and Macrophyllum were similar in standard karyotype, and the morphology supports this association (Smith, 1976). Unfortunately, the immunological data cannot address this association; however, Macrophyllum is associated with the Phyllostomus-Tonatia clade. From the foregoing accounts, the chromosomal, immunological and morphological data sets demonstrate a large degree of concordance in the placement of the remaining phyllostomine genera. Although from a eladistic standpoint, the Phyllostominae cannot be viewed as monophyletic, all the existing phyllostomine genera can be associated with one of the four major lineages of phyllostomine bats. This adds some support to the rapid radiation hypothesis discussed in the last section. Chromosome G-banding studies of intrageneric relationships within two genera (Tonatia and Micronycteris) have indicated that certain species cannot be related to their congeners due to numerous amounts of chromosomal rearrangements associated with certain species (Patton and Baker, 1978; Baker and Bickham, 1980). In the ease of Tonatia, the immunology adds little resolution. Tonatia bidens can be placed better along the Phyllostomus lineage with T^. schulzi and T^. carrikeri being appreciably distant from the T^. silvicola lineage. The possibility of differential rates of albumin change along the Tonatia lineages cannot be ruled out, and T_. bldens has obviously experienced more change along its lineage relative to T_. silvicola. In light of the morphological similarity of these genera (Goodwin, 1942), I think it 70 best at this time to suggest that the lack of association of these species with T_. silvicola is probably an immunological artifact. A suggestion for future work would be to make an antiserimi to the T. bidens albumin. This antiser\jm should allow for a test of the monophyletic grouping of these species more accurate. All the Micronycteris species form a monophyletic group, and immunological data suggest a closer relationship of M. sylvestris and M. daviesi relative to the other species. The close alignment of M. sylvestris and M. daviesi is also supported by morphology (Handley. 1976). Species of Phyllostomus also form a monophyletic group relative to other genera. Phyllostomus elongatus is closely aligned with P^. hastatus, and P^. discolor is one of the more divergent Phyllostomus species. The immunological data also agree with both chromosome banding data (Patton and Baker, 1978) and morphological data (Smith, 1976) in associating Phylloderma with Phyllostomus lineage. Another interesting aspect of these unidirectional tests is the weaker reaction of the P^. latifollus albimiln to P^. hastatus than that seen for P^. elongatus. As suggested by Jones and Carter (1976), P^. elongatus and P^. latifollus may be conspeciflc; however, the albumins seem somewhat divergent. Brachyphyllinae and Glossophaginae Relationships Within the Brachyphyllinae The use of molecular data in providing plausible solutions to rather longstanding systematic problems is well documented (Sarich, 1969 a,b; Sarich, 1973; Maxson and Wilson, 1974; Maxson and Wilson, 1975; 71 Cronin and Sarich, 1976). In the case of the placement of Brachyphylla and the determination of inter-brachyphylline relationships, the albumin immunological data support the hypothesis that the Brachyphyllinae as currently conceived form a natural evolutionary unit. These data define a braehyphylline clade with Erophylla and Phyllonycteris being more closely aligned to each other than either is Brachyphylla (Fig. 4). Recent electrophoretic data add further support to this association (Baker et al., manuscript). Of greater interest is the contrast in degree of difference which distinguishes Brachyphylla from Erophylla and Phyllonycteris at the anatomical and biochemical levels. Brachyphylla is anatomically divergent and has been aligned in the past with several different subfamilies (Miller, 1907; Silva-Toboada and Pine, 1969; Slaughter, 1970; Baker and Bass, 1979). At the molecular level Brachyphylla is as close to Phyllonycteris and Erophylla as Monophyllus is to Glossophaga. Anatomically, Monophyllus and Glossophaga are sufficiently similar that Varona (1974) considered them congeneric. Clearly, the magnitude of anatomical distinctiveness is not mirrored by the molecular data. Disparities between anatomical and molecular rates of evolution are well documented (Wilson et al., 1977), and in several eases, these disparities point to major adaptive shifts in morphology (Maxson and Wilson, 1975). These disparities coupled with differential rates of chromosomal evolution even have prompted some authors (Wilson et al., 1974; Wilson et al., 1975) to suggest a cause-and-effeet relationship between rates of chromosomal and anatomical evolution. In the case of the Brachyphyllinae, the G- and C-banded chromosome patterns are 72 identical (Baker and Bass, 1979); however, the morphological diversity is rather large. Whether the morphological divergence seen in Brachyphylla is due to an adaptive shift in either Brachyphylla or the Phyllonycteris-Erophylla lineage remains to be seen. Nevertheless, this seems to be at least a viable hypothesis. Because the braehyphylline genera were sufficiently unique to be accorded subfamilial status, and because of an Implied long period of isolation. Baker and Genoways (1978) suggested that these genera were the most likely candidates of any of the Antlllean bat fauna for having reached the area by vlcariance (Rosen, 1978) rather than over water dispersal. However, distance values from both the electrophoretic (Baker et al., manuscript) and albumin studies are not of the magnitude that would be anticipated if the intrasubfamillal radiation was extremely ancient. Comparative data will be needed from other Antillean bat taxa before the significance of values from the Brachyphyllinae can be understood. Chromosomal Relationships The two alternative explanations suggested by G- and C-banding chromosomal data can be interpreted in light of the immunological data. The albumin data do not support the grouping of the Brachyphyllinae, Monophyllus, and Glossophaga lineages into a derived clade possessing a similar karyotype of 2n = 32, FN = 60. At least four other glossophagine genera (Choeroniscus, Hylonycteris, Leptonycteris, and Anoura) are more closely aligned to the Glossophaga and Monophyllus lineages than either of these two lineages are to the Brachyphyllinae. 72 identical (Baker and Bass, 1979); however, the morphological diversity is rather large. Whether the morphological divergence seen in Brachyphylla is due to an adaptive shift in either Brachyphylla or the Phyllonycteris-Erophylla lineage remains to be seen. Nevertheless, this seems to be at least a viable hypothesis. Because the braehyphylline genera were sufficiently unique to be accorded subfamilial status, and because of an implied long period of isolation. Baker and Genoways (1978) suggested that these genera were the most likely candidates of any of the Antillean bat fauna for having reached the area by vlcariance (Rosen, 1978) rather than over water dispersal. However, distance values from both the electrophoretic (Baker et al., manuscript) and albumin studies are not of the magnitude that would be anticipated if the intrasubfamillal radiation was extremely ancient. Comparative data will be needed from other Antillean bat taxa before the significance of values from the Brachyphyllinae can be understood. Chromosomal Relationships The two alternative explanations suggested by G- and C-banding chromosomal data can be interpreted in light of the immunological data. The albumin data do not support the grouping of the Brachyphyllinae, Monophyllus, and Glossophaga lineages into a derived clade possessing a similar karyotype of 2n = 32, FN = 60. At least four other glossophagine genera (Choeroniscus, Hylonycteris, Leptonycteris, and Anoura) are more closely aligned to the Glossophaga and Monophyllus lineages than either of these two lineages are to the Brachyphyllinae. 73 These data indicate that the 2n = 32, FN = 60 karyotype may be primitive for the entire glossophagine-brachyphylline clade as suggested by Baker and Bass (1979). The Anoura caudlfer (2n = 30, FN = 60), Choeroniscus minor (2n = 20, FN = 36), and Hylonycteris underwoodi (2n = 16, FN = 24) karyotypes are then seen as derived from this primitive pattern. An alternative to this primitive pattern would be to consider the Lonchophylla thomasi (2n = 32, FN = 38) or Lionycteris spurrelll (2n = 28, FN = 50) karyotypes as primitive. The immunological data indicate that these two genera are the most divergent of all the glossophagine and braehyphylline genera examined. Although the exact relationships between Lionycteris and Lonchophylla cannot be ascertained with immunology at this time, Phillips (1971) did suggest an alignment of the two genera based on cranial and dental features. It is also Interesting to note that Gardner (1977^) suggested a Lonchophylla-like karyotype as primitive for the glossophagines, and this alternative cannot be ruled out. The immunological and ehrcmiosomal data raise a question as to the coordinate (sister-group) status of the subfamilies Brachyphyllinae and Glossophaginae. A viable alternative classification would be to simply relegate the Brachyphyllinae to the status of a tribe within the Glossophaginae, thus emphasizing the fact that the brachyphyllines are simply island forms deriving from the basal glossophagine radiation. Polyphyly and the Glossophagines As indicated by Maxson and Wilson (1975), unidirectional tests cannot be used to construct complete phylogenetic trees due to the lack 73 These data indicate that the 2n = 32, FN = 60 karyotype may be primitive for the entire glossophagine-brachyphylline clade as suggested by Baker and Bass (1979). The Anoura caudlfer (2n = 30, FN = 60). Choeroniscus "tinor (2n = 20, FN = 36), and Hylonycteris underwoodi (2n = 16, FN = 24) karyotypes are then seen as derived from this primitive pattern. An alternative to this primitive pattern would be to consider the Lonchophylla thomasi (2n =32, FN = 38) or Lionycteris spurrelll (2n = 28, FN = 50) karyotypes as primitive. The immunological data indicate that these two genera are the most divergent of all the glossophagine and braehyphylline genera examined. Although the exact relationships between Lionycteris and Lonchophylla cannot be ascertained with immunology at this time, Phillips (1971) did suggest an alignment of the two genera based on cranial and dental features. It is also interesting to note that Gardner (1977^) suggested a Lonchophylla-like karyotype as primitive for the glossophagines, and this alternative cannot be ruled out. The immunological and chromosomal data raise a question as to the coordinate (sister-group) status of the subfamilies Brachyphyllinae and Glossophaginae. A viable alternative classification would be to simply relegate the Brachyphyllinae to the status of a tribe within the Glossophaginae, thus emphasizing the fact that the brachyphyllines are simply island forms deriving from the basal glossophagine radiation. Polyphyly and the Glossophagines As indicated by Maxson and Wilson (1975), unidirectional tests cannot be used to construct complete phylogenetic trees due to the lack 74 of reciprocal comparisons, but the relative placement of taxa can be determined with such tests. Phillips (1971) provided a rather intricate phylogenetic tree for the glossophagines. Although I cannot address all the intricacies indicated by Phillips, I can address the central issue of a polyphyletic origin for the glossophagines. As first proposed by Baker (1967). the glossophagines were shown to be chromosomally related to two and possibly three groups as follows: 1) Leptonycteris and Glossophaga aligned with Phyllostomus; 2) Choeroniscus and Choernycteris aligned with Carollia; 3) Anoura affinities were unclear. Contrary to Baker's proposal. Stock (1975) using chrcDmosome G-banding indicated that Choeroniscus and Carollia shared no chromosomal homologies. Gerber and Leone (1971), using immunology, also suggested a basic dicotomy in the glossophagines. Although these immunological relationships differed from Baker's (1967) groups, Gerber and Leone (1971) agreed with the idea that the Glossophaginae were of polyphyletic origin. The weakness of this immunological study stems mainly from the total exclusion of any consideration of differential rates of molecular evolution; therefore, I place little stock in the conclusions. The albumin immunological data in this study do not support the proposition that several glossophagine lineages arose independently from a non-nector feeding stock. Choeroniscus and Anoura are not seen as outside the major glossophagine clade, but rather strongly associated with the Glossophaga and Monophyllus lineages. Although rather incomplete, electrophoretic data also indicate an association of Anoura and Glossophaga (Straney et al., 1979). Continued G- and C-chromosome banding studies coupled with 75 more complete morphological and immunological data should add further resolution to the exact relationships of the glossophagine genera. Relationships Within the Carolliinae The Immunological data are not unequivocal in associating Rhinophylla with Carollia, and the status of this relationship cannot be resolved without additional tests using an antiserum to Rhinophylla albumin. The morphological evidence would seem to be more definitive since these two genera have been retained in the same subfamily at least since Miller (1907). In eonstrast, G-band chromosomal data are harder to assess. Rhinophylla pumilio and ^. fischerae can neither be related to each other nor to Carollia based on G-band homologies; hc3wever, R. pumilio does share presumably derived chromosomes with the Phyllostomus-Tonatla clade (Baker, unpublished). Carollia has already been shown to be chromosomally divergent (Baker and Bickham, 1980), and if Rhinophylla is aligned with Carollia, an unusual case of chromosomal evolution could be documented for this subfamily. In this case, R. piimjlio may be chromosomally conservative, whereas R.. fischerae and Carollia have evolved at a much faster rate. This is an interesting hypothesis to test with albumin data. Relationships Within the Desmodontinae The albimiln immunological and electrophoretic data indicate that Desmodus and Diaemus are more closely aligned to each other than either is to Diphylla. Thus, these data support the phylogenetic hypothesis suggested by classical morphology (Miller, 1907; Slaughter, 1970; 76 Cadena, 1977). No other phyllostomid albumin tested gave a distance of less than 40 units to the vampires; therefore, the usual rate tests would not have sufficient sensitivity to provide useful resolving power for the apportionment of the low values seen in intra-vampire comparisons. Nonetheless, the striking immunological similarity (15 AID units) of the albumins of Desmodus and Diaemus is almost unique for intergeneric comparisons among mammals. Such low values occur elsewhere only among some of the very closely related, and often interfertlle, cercopitheelne genera (Macaea, Papio, Mandrillus, Cercocebus) and between the proboscidean genera, Loxodonta and Elephus (Sarich, unpublished). The albumin immunological distance between Vampyrum and Chrotopterus (3 AID units) is the only case where similar low values were found between genera of phyllostomid bats. It remains possible, of course, that the albumins of both Desmodus and Diaemus have experienced very little change through their independent existences, with essentially all the measured immunological distances within the vampires contributed by the Diphylla lineage, but this seems much less likely than a eladlstle association of Desmodus and Diaemus subsequent to the divergence of the Diphylla lineage (Fig. 5). The one eleetrophoretieally detectable synapomorphy (ldh-1) for Diaemus and Desmodus is not conclusive evidence that these two genera are cladlstleally close. However, the Fitch and Margoliash method substantiates this association when one considers the possibility of Diaemus experiencing a more rapid rate of evolution relative to Desmodus. A rate slowdown along the Desmodus lineage also provides an explanation for Diaemus and Diphylla being equidistant from Desmodus as indicated 77 by the values of genetic similarity and distance. Ultimately, all characters should lead to taxonomic congruence. The molecular and morphological data are congruent; therefore, only the chromosomal data are inconsistent which leads one to question the possibility of homoplasy or misidentificatlon of homologous segments. Rates of Albumin Evolution In general, albumin evolution (as assessed immunologically) has been documented to proceed at a relatively constant rate within most vertebrate groups, thus reflecting the time-dependent nature of sequence divergence. The documentation of this regularity was not derived by a priori assumptions. To the contrary, certain lineages have been shown to depart from a pattern of regularity as a result of internal rate tests. Examples of individual departures are Actus, Caluromys, Marmosa, and Ursus in the "slow" side, and Rousettus and Phaner on the "fast" side (Sarich, 1969_b; Cronin and Sarich, 1975; Maxson et al., 1975). In some instances, departures from regularity have been documented for several lineages comprising a major group of vertebrates. This seems to be the ease for all anthropoid albumins where subsequent to the anthropoid/prosimian split, about 25 to 30 units of change in excess of the average accumulated along the anthropoid lineage (Cronin and Sarich, 1975). Therefore, all anthropoid albumins have changed to a greater degree than those of most prosimlans. On a larger scale, Prager et al. (1974) indicated that bird albumin change relative to other vertebrates is appreciably slower, but Wilson et al. (1977) suggest that this rate difference may be an artifact of both 78 the fossil record and the immunological approach. Relative to other vertebrates as well as other bats (Sarich, unpublished), the Noctilionoidea appear to have undergone a rate destabilization for subsequent albumin evolution in the group (at least as it is assessed immunologically). Within the Phyllostomidae, there is more disparity in amounts of change along different lineages than has ever been demonstrated in any other vertebrate group. This disparity suggests that a molecular clock concept might never have been formulated on the basis of noctilionoid albumin evolution. The rate destabilization of albumin change within the Noctilionoidea lineage should not be construed as evidence to totally refute the molecular clock concept. As Wilson et al. (1977) have indicated, "It is quite possible that future studies will uncover more unequivocal examples of deviations from constant rates of sequence evolution. This does not necessarily mean that the evolutionary-clock phenomenon is invalid. Moreover, such exceptions may provide useful information about the forces that act on a particular macromoleeule or organism at the molecular level." Another interesting result derived from an assessment of rates of molecular evolution within the Noctilionoidea (specifically the Phyllostomidae) is the curvilinear relationship between albumin immunological and electrophoretic distance values. In view of albumin evolution within the superfamily, one would expect such a relationship. As suggested by Corruccini et al. (1980), such a relationship implies three things - 1) different molecules cannot be linearly related with time; 2) disparity between the rates at which one molecule evolves 79 relative to another is substantiated; 3) each point along the curve implies an increase or decrease in rate of one molecule relative to the other; therefore, the rate changes systematically according to the location on the curve. Although rates of albumin change may be contributing greatly to this curvilinear relationship, proteins assessed by electrophoresis also are demonstrating some disparity in rates relative to the albumin data. Several examples are as follows: 1) the Tonatia sllvieola-T. venezuelae Nei's distance relative to T^. brasiliense and T_. nicaraguae is appreciably higher; however, the albumin distance is identical; 2) the Nei's distance between T^. silvicola and T^. carrikeri is 0.313 whereas the former species has a distance of 0.793. The immunological distance values of 28 and 33 are quite similar; 3) the Micronycteris daviesi-Vampyriim Nei's distance is 2.351, whereas the AID distance is 26. In lighc of this curvilinear relationship, one must conclude that time since divergence estimates from electrophoretic distances using immunological distances as a means of calibration (sensu Sarich, 1977) must be approached cautiously for phyllostomid bats. Therefore, I question the time estimates suggested by Straney et al. (1979) for phyllostomid bats. In a group where no immunological data exist, it might also be wise to refrain from overemphasizing time since divergence estimates based on electrophoresis. CONCLUSIONS The most significant conclusion from the data set on which this work is based is that albumin evolution in noctilionoid bats is not clock-like in its rate. The disparity of rates in these bats is greater than that found for any other vertebrate group. If noctilionoid bats had been used as the primary source of data for albumin evolution studies, it is safe to conclude that the molecular clock hypothesis would never have been formulated. The magnitude of importance of this observation is clear when one realizes that for the past 15 years major portions of such journals as Science, Nature, Evolution, and Systematic Zoology have been devoted to debate over the validity of the molecular clock hypothesis. The major strength of the immunological approach rests on the fact that evolutionary relationships can be determined without a^ priori reference to nonmolecular data, yet the agreement with nonmoleeular data is strong. The areas of agreement among the data sets are as follows: 1) the Noctilionoidea form a distinct unit relative to other bat families; 2) Brachyphylla is associated with Erophylla and Phyllonycteris; therefore, the proper subfamilial name for this group is the Brachyphyllinae; 3) the Brachyphyllinae and Glossophaginae form a unit relative to other subfamilies; 4) Phyllostomus and Tonatia form a clade, and this clade is aligned with several other phyllostomine genera including Lonchorhina, Maerophyllum. aaid Mimon; 5) Phylloderma is closely aligned with Phyllostomus; 6) Chrotopterus 80 81 is aligned with Vampyrum; 7) species of Phyllostomus and Micronycteris can be related to their congeners; 8) the vampire bats are properly classified as members of the family Phyllostomidae. The most significant area of disagreement between the immunological data and the currently accepted systematics of the family Phyllostomidae is that the subfamily Phyllostominae is not a natural assemblage. This subfamily represents several independent lines of evolution, each of which is more closely related to another subfamily than to other members of the subfamily Phyllostominae. Although Walton and Walton (1968) and Smith (1976) alluded to this Idea, they provided little data to substantiate their hypothesis. The second area of disagreement is that the Glossophaginae are not polyphyletic as suggested by Baker (1967) and Gerber and Leone (1971). At this time, the causes of inconsistencies among the morphological, chromosomal, and biochemical data sets are not clear; however, I think that several steps can be taken to resolve some of the controversies. In terms of morphology, a method of analysis (cladistics) that can detect differential rates of change, convergence, parallelism, and primitive (as opposed to derived) character states is virtually nonexistent for noctilionoid bats. The lack of such an approach leads to an inability to compare directly chromosomal and molecular phylogenles to those generated by morphology. In fact, an appropriate quantitative approach (void of size relationships) to morphological relationships within noctlllonoids does not exist. This is especially apparent at the generic level and above. Hopefully, research on morphological divergence in noctlllonoids will continue, and our 82 understanding of this evolution will i:.prove in the future. The chromosomal data, especially G- and C-banding data, are Incomplete, and several key genera need to be examined before definitive conclusions can be made concerning relationships. The immunological studies are fairly complete except for certain intrageneric comparisons and the Stenodermatinae. In the future, the albumin immunological relationships should be tested using other molecular data such as transferrins and DNA. These data would not only provide independent molecular tests, but also would provide additional information on rates of molecular evolution. Other discrepancies such as Tonatia and the placement of Rhinophylla and Lonchorhina dictate more complete albumin immunological data using reciprocal tests. Intergeneric relationships also could be verified by these same reciprocal tests. Seemingly, intrageneric relationships and some intergeneric relationships could also be approached using starch-gel electrophoresis. From the existing data, one can conclude that our current concepts concerning the classification of the Noctilionoidea will have to be revised. The degree of revision will depend on whether one classifies according to a straight eladlstle approach which emphasizes branching order or whether one considers a phyletic approach which emphasizes magnitude of divergence. The classification might seem an end in itself and rightly so; however, I think the information obtainable from an objective and consistent evolutionary framework is more important. The closer this framework approximates the "true" phylogeny, the easier it will be to address some important evolutionary questions. I can do no better than cite Hull (1979) in regard to this matter. 83 "Perhaps the establishment of lineages is a risky business, but it is also an extremely important undertaking, not for the purposes of classification but for the purposes of formulating and testing hypotheses about the evolutionary process. As important as scientific classifications are, scientific theories are even more important." Evolutionary biology is a broad field and has reached a point where basic concepts concerning the overall processes of evolution are being developed. The development and testing of such broad concepts, which attempt to explain not only organlsmal evolution but also chromosomal and molecular evolution, demands the integration of numerous approaches. It is in the spirit of this idea that I submit this thesis for evaluation by the scientific community. LITERATURE CITED .\vise, J. C. 1974. Systematic value of electrophoretic data. Syst. Zool., 23:465-481. Baker, R. J. 1967. Karyotypes of bats of the family Phyllostomidae and their taxonomic implications. Southwestern Nat., 12:407-428. • 1979. Karyology. Pp. 157-176, in Biology of bats of the New '.vorld family Phyllostomatidae, Part III (R. J. Baker, J. K. Jones, Jr., and D. C. Carter, eds.). Spec. Publ. Mus., Texas Tech Univ., 16:1-441. Baker, R. J., and R. A. Bass. 1979. Evolutionary relationship of the Brachyphyllinae to the glossophagine genera Glossophaga and Monophyllus. J. Mamm., 60:364-372. Baker, R. J., R. A. Bass, and M. A. Johnson. 1979. Evolutionary implications of chromosomal homology in four genera of stenodermine bats (Phyllostomatidae:Chlroptera). Evolution 33:220-226. Baker, R. J., and J. W. Bickham. 1980. Karyotypic evolution in bats: evidence of extensive and conservative chromosomal evolution in closely related taxa. Syst. Zool., 29:239-253. Baker, R. J., and H. H. Genoways. 1978. Zoogeography of Antillean bats. Pp. 53-97, ^ Zoogeography in the Caribbean: the 1975 Leidy Medal Symposium (F. G. Gill, ed.). Philadelphia Acad. Nat. Scl., Spec. Publ., 13:1-128. Bass, R. 1978. Systematics of the Desmodontinae and Phyllonyeterinae (Chlroptera:Phyllostomatidae) based on G-band chromosomal homologies. M. S. thesis, Texas Tech University, Lubbock, 53 pp. Baverstock, P. R., S. R. Cole, B. J. Richardson, and C. H. S. Watts. 1979. Electrophoresis and cladistics. Syst. Zool., 28:214-219. Cadena, A. 1977. A systematic study of vampire bats, Desmodontinae (Phyllostomidae, Chlroptera). Ph.D. thesis. University of Kansas, Lawrence, 189 pp. Case, S. M. 1978. Biochemical systematics of members of the genus Rana native to western North America. Syst. Zool., 27:299-311. Champion, A. B. , E. M. Prager, D. Wachter, and A. C. Wilson. 1974. Micro-complement fixation. Pp. 397-416, in Biochemical and immunological taxonomy of animals (C. A. Wright, ed,). Academic Press, London, 490 pp. 84 85 Corruccini, R. S M. Baba, M. Goodman. R. L. Glochon, and J. E. C._ronln_ . iyau. Non-lmear macromolecular evolution and the molecular clock Evolution 34:1216-1219. Cronin, J. E., and V. M. Sarich. 1975. Molecular systematics of the New World monkeys. J. Human Evol., 4:357-375. 1976. Molecular evidence for dual origin of mangabeys among Old World monkeys. Nature 260:700-702. Dobson, G. E. 1875. Conspectus of the suborders, families, and genera of Chlroptera arranged according to their natural affinities. Ann. Mag. Nat. Hist., ser. 4, 16:345-357. Farris, J. S. 1972. Estimating phylogenetic trees from distance matrices. Amer. Natur., 106:645-668. Farris, J. S., A. G. Kluge, and M. F. Mlckevleh. 1979. Paraphyly of the Rana boylii species group. Syst. Zool., 28:627-634. Fitch, W. M., and E. Margoliash. 1967. Construction of phylogenetic trees. Science 155:279-284. Forman, G. L. , R. J. Baker, and J. D. Gerber. 1968. Comments on the systematic status of vampire bats (family Desmodontidae). Syst. Zool., 17:417-425. Gardner, A. L. 1977a_. Feeding habits. Pp. 293-350, in Biology of bats of the New World family Phyllostomatidae, Part III (R. J. Baker, J. K. Jones, Jr., and D. C. Carter, eds.). Spec. Publ. Mus., Texas Tech Univ., 13:1-364. . 1977b^. Chromosomal variation in Vampyressa and a review of chromosomal evolution in the Phyllostomatidae (Chlroptera). Syst. Zool., 26:300-318. Gerber, J. D., and C. A. Leone. 1971. Immunologic comparisons of the sera of certain phyllostomatid bats. Syst. Zool., 20:160-166. Goodwin, G. G. 1942. A summary of recognizable species of Tonatia, with descriptions of two new species. J. Mamm., 23:204-209. Gorman, G. C., D. G. Buth, and J. S. Wyles. 1980. Anolis lizards of the eastern Caribbean: a ease study in evolution. III. A eladistic analysis of albumin immunological data, and the definition of species groups. Syst. Zool., 29:143-158. eorman, G.* C, A. C. Wilsoa, and M. Nakanlshi. 1971. A biochemical approach towards the study of repfilian phylogeny: evolGtion of eiriim albumin and.lactic dehydrogenase. Syst..Zool., 20:167-185. 86 Greenbaum. I- F., and R J. Baker. 1976. Evolutionary relationships in Macrotus (Mammalia:Chlroptera): biochemical variation and karyology. Syst. Zool., 25:15-25. Handley, C. 0. 1976. Mammals of the Smithsonian Venezuelan project. Brigham Young Univ. Scl. Bull., Biological series 20:1-89. Helwig, J. T., and K. A. Council. 1979. SAS user's guide. SAS Institute, Raleigh, North Carolina, 495 pp. Hull, D. L. 1979. The limits of cladism. Syst. Zool., 28:416-440. Johnson, M. A. 1979. Evolutionary implications of G- and C-banded chromosomes of 13 species of stenodermine bats. M. S. thesis, Texas Tech University, Lubbock, 49 pp. Jones, J. K., Jr., and D. C. Carter. 1976. Annotated checklist, with keys to subfamilies and genera. Pp. 7-38, in Biology of bats of the New World family Phyllostomatidae, Part I (R. J. Baker, J. K. Jones, Jr., and D. C. Carter, eds.). Spec. Publ. Mus., Texas Tech Univ., 10:1-218. Koopman, K. F. 1978. Zoogeography of Peruvian bats with special emphasis on the role of the Andes. Amer. Mus. Novitates, 2651:1-33. Maxson, L. R. , V. M. Sarich, and A. C. Wilson. 1975. Continental drift and the use of albumin as an evolutionary clock. Nature 255:397-400. Maxson, L. R., and A. C. Wilson. 1974. Convergent morphological evolution detected by studying proteins of tree frogs in the Hyla exlmla species group. Science 185:66-68-, . 1975. Albumin evolution and organlsmal evolution in tree frogs (Hylidae). Syst. Zool., 24:1-15. . 1979. Rates of molecular and chromosomal evolution in salamanders. Evolution 33:734-740. Miller, G. S., Jr. 1907, The families and genera of bats. Bull. U. S. Nat. Mus., 57:xvli-l-l-282. Nei, M. 1972. Genetic distance between populations. Amer. Natur., 106:283-292. Patton, J. C, and R. J. Baker. 1978. Chromosomal homology and evolution of phyllstomatoid bats. Syst. Zool., 27:449-462. 87 Phillips, C. J. 1971. The dentition of glossophagine bats: development, morphological characteristics, variation, pathology, and evolution. Misc. Publ. Mus. Nat. Hist., Univ. Kansas, 54:1-138. Prager, E. M. , A. H. Brush, R. A. Nolan, M. Nakanlshi, and A. C. Wilson. 1974. Slow evolution of transferrin and albumin in birds according to micro-complement fixation analysis. J. Mol. Evol., 3:243-262. Prager, E. M., and A. C. Wilson. 1971a.. The dependence of immvinologlcal cross-reactivity upon sequence resemblance among lysozymes. I. Mlcrocomplement fixation studies. J. Biol. Chem., 246:5978-5989. . 1971b_. The dependence of immunological cross-reactivity upon sequence resemblance among lysozymes. II. Comparison of precipitin and mlcrocomplement fixation results. J. Biol. Chem., 246:7010-7017. . 1978. Construction of phylogenetic trees for proteins and nucleic acids: empirical evaluation of alternative matrix methods. J. Mol. Evol., 11:129-142. Rogers, J. S. 1972. Measure of genetic similarity and genetic distance. Univ. Texas Publ., 7213:145-153. Rosen, D. E. 1978. Viearlant patterns and historical explanation in biogeography. Syst. Zool., 27:159-188. Sarich, V. M. 1969a.. Pinniped origins and the rate of evolution of carnivore albumins. Syst. Zool., 18:286-295. . 1969b^. Pinniped phylogeny. Syst. Zool., 18:416-422. 1973. The giant panda is a bear. Nature 245:218-220. 1977. Rates, sample sizes, and the neutrality hypothesis for electrophoresis in evolutionary studies. Nature 265:24-28. Sarich, V. M., and J. E. Cronin. 1977. Molecular systematics of the primates. Pp. 141-170, in Molecular anthropology, genes and proteins in the evolutionary ascent of the primates (M. Goodman and R. E. Tashian, eds.). Plenum Press, New York, New York, xlv-l-466. Sarich, V. M., and A. C. Wilson. 1966. Quantitative immunochemlstry and the evolution of primate albumins: microcomplement fixation. Science 154:1563-1566. 1967. Rates of albumin evolution in primates. Proc. Natl, Acad. Sci., 58:142-148. 88 Selander R. KM. H. Smith, S. Y. Yang, W. E. Johnson, and J, B. i.entr>- 1971. Biochemical polymorphism and systematics in the genus Peromyscus. I, Variation in the old-field mouse (Peromyscus polionotus). Studies in Genetics VI, Univ. Texas Publ., 7103:49-90. Sllva-Taboada, G., and R. H. Pine. 1969. Morphological and behavorial evidence for the relationship between the bat genus Brachyphylla and the Phyllonyeterinae. Blotropica 1:10-19. Slaughter, B. H. 1970. Evolutionary trends of chiropteran dentitions. Pp. 51-83, in About bats (B, H, Slaughter and D. W. Walton, eds.). Southem Methodist Univ. Press, Dallas, Texas, 319 pp. Smith, J. D. 1972. Systematics of the chiropteran family Mormoopidae. Misc. Publ. Mus. Nat. Hist., Univ. Kansas, 56:1-132. • 1976. Chiropteran evolution. Pp. 49-69, in Biology of bats of the New World family Phyllostomatidae, Part I (R. J. Baker, J. K. Jones, Jr., and D. C. Carter, eds.). Spec. Publ. Mus., Texas Tech Univ., 10:1-218. Stock, A. D. 1975. Chromosome banding pattern homology and its phylogenetic implications in the bat genera Carollia and Choeroniscus. Cytogenet. Cell Genet., 14:34-41. Straney, D. 0., M. H. Smith, I. F. Greenbaum, and R. J. Baker. 1979. Biochemical genetics. Pp. 157-176, ±n Biology of bats of the New World family Phyllostomatidae, Part III (R. J. Baker, J. K. Jones, Jr., and D. C. Carter, eds.). Spec. Publ. Mus., Texas Tech Univ., 16:1-441. Trouessart, E. L. 1897. Catalogue mammalium tarn viventum quam follllium. Nova edito (prima completa). Berlin, R. Friedlander and Sohn, 1:1-218. Varona, L. S. 1974. Catalogo de los mamiferos vivientes y extinguidos de las Antillas. Acad. Cien. Cuba, viil-l-139. Wallace, D. G., M. -C. King, and A. C. Wilson. 1973. Albumin differences among ranid frogs: taxonomic and phylogenetic implications. Syst. Zool., 22:1-14. Walton, D. W., and G. W. Walton. 1968. Comparative osteology of the pelvic and pectoral girdles of the Phyllostomatidae (Chlroptera: Mammalia). J. Grad. Res. Center, Southem Methodist Univ., 37:1-35. Wilson, A. C, G. L. Bush, S. M. Case, and M. -C. King. 1975. Social structuring of mammalian populations and the rate of chromosomal evolution. Proc. Nat. Acad. Sci., 72:5061-5065. 89 Wilson, A. C, S. S. Carlson, and T. J. White. 1977. Biochemical evolution. .\nn. Rev. Biochem., 46:573-639. Wilson, A. C, V. M. Sarich, and L. R. Maxson. 1974. The importance of gene rearrangements in evolution: evidence from studies on rates of chromosomal, protein, and anatomical evolution. Proc. Nat. Acad. Sci., 71:3028-3030. Winge, H. 1892. Jordfundne og nulevende Flagermus (Chlroptera) fra Logoa Santa, Mines Geraes, Brasilien, Copenhagen, Museo Lundii, 2:1-65. rj 3 PH 3 U to 3 CO u -o lH CO 3 OJ 43 VH )H rH rH •H y M cn OJ y 03 fi fi 3 OJ 3 U rH 44 y CO OJ r-t iH 4-1 _J 01 44 to •H OJ rH 44 to OJ CO rH y to o 44 >N tn cn 44 £ y iH y 43 iH r-t y y 44 r-t >1 y 43 O a a 3 >N B >i M y y rH 43 fi> , tfl cfl fi. cfl 43 )H >. •^ >. o 0 B O fi O X fi cfl ?^ a o •rt iH fi •H r- 4-1 O IH O 43 fi a o M fi> . 4= o rH 44 4-1 0 4>3 , 4J >! )H £ "y o a U OJ y O 0 M 3 o a V4 r^ tfl CO u y tfl X O cfl u £ y O y fi £ B y o rH o y >-, B fi y CO fi cfl u u X CO iH X •H O •H O •rl ^4 CO 43 0 0 •rt M o H H tj i-j ^ Cd 04 H H M-l > S u S 2 s 2 ^ tn 2 2 m H 55