FOSSIL BATS OF THE AMERICAS
Phylogenetic relationships of the Natalidae
The discovery of Primonatalus prattae, a new genus and species of Natalidae from the early Miocene of Florida, prompted us to examine the phylogenetic relationships of all species within the Natalidae (Morgan and Czaplewski 2003). Although our primary aim was to determine the systematic position of Primonatalus, we also examined the relationships among the six extant species of natalids because no formal hypothesis of relationships had been proposed for the family. Our phylogenetic analysis also incorporated as outgroups representatives of two other closely related Neotropical families included in the Nataloidea by Simmons (1998) and Simmons and Geisler (1998), the Furipteridae (Furipterus horrens) and Thyropteridae (Thyroptera tricolor), as well as two vespertilionids, Myotis lucifugus and Kerivoula picta.
Other workers have commented on the close relationship between the Natalidae, Furipteridae and Thyropteridae. Miller (1907) placed the genera Furipterus and Amorphochilus in the Furipteridae and Thyroptera in the Thyropteridae, and considered both families to be closely related to the Natalidae. Miller's arrangement of these three families has been followed by most subsequent chiropteran systematists. Van Valen (1979) expanded the definition of the Natalidae to include the Furipteridae, Thyropteridae, Myzopodidae, and Kerivoulidae as subfamilies. This is similar to the concept of the superfamily Nataloidea proposed by Simmons (1998), including Natalidae, Furipteridae, Thyropteridae, and Myzopodidae. Van Den Bussche and Hoofer (2001) recently presented molecular evidence that Nataloidea is not a monophyletic lineage. Their mitochondrial DNA sequence data suggested instead that Myzopodidae represents a basal microchiropteran lineage and that Furipteridae, Thyropteridae, and Natalidae are successive outgroups to the Noctilionoidea.
Based on Van Valens (1979) concept of the Natalidae, Beard et al. (1992) referred five extinct genera to the Natalidae, including Ageina and Stehlinia from the Eocene of Europe, Honrovits from the Eocene of Wyoming, Chadronycteris from the Eocene of Nebraska, and Chamtwaria from the Miocene of Africa. We follow Simmons and Geisler (1998) who referred Ageina, Chadronycteris, Honrovits, and Stehlinia to Nataloidea incertae sedis, because these genera lack the derived cranial, dental, and postcranial features of the Natalidae sensu stricto (i.e., the six living species plus the Miocene Primonatalus prattae). Simmons and Geisler (1998) did not discuss Chamtwaria because this genus is Miocene and they reviewed only Eocene taxa. McKenna and Bell (1997) referred Chamtwaria to the subfamily Kervoulinae within the Vespertilionidae.
Parsimony analysis of our data set of 50 characters (see character list and data matrix below) finds strong support for the monophyly of the family Natalidae (sensu stricto), including Primonatalus prattae (see cladogram ). Among the extant taxa, the three species of Natalus are monophyletic and this clade is strongly supported. Within this clade, Natalus stramineus and Natalus tumidirostris are sister taxa, with strong support for the inclusion of Natalus major. The two species of Chilonatalus are monophyletic, supporting the idea that Chilonatalus is a distinct genus from Natalus. Chilonatalus and Natalus cluster together, but support for this clade is moderate. There is moderate support for distinction of the traditional subgenera Natalus, Chilonatalus, and Nyctiellus.
All living members of the Natalidae are traditionally placed in the genus Natalus (e.g., Koopman 1993); however, our phylogenetic analysis suggests that the six extant species should be separated into three genera: Natalus (including N. major, N. stramineus, and N. tumidirostris), Chilonatalus (including C. micropus and C. tumidifrons), and Nyctiellus (including N. lepidus), each previously recognized as a subgenus of Natalus (Dalquest 1950). The fossil Primonatalus is the most primitive member of the family and is distinct from other natalids, but its membership in the family is strongly supported. These results are remarkably concordant with aspects of the traditional classification scheme for the living natalids, and suggest there may be a phylogenetic signal in the dental-cranial-postcranial morphology in this family.
The fossil material of Primonatalus prattae is not nearly so complete as that available for the six living species of natalids. Although many of the derived characters of the Natalidae are unknown in Primonatalus, enough of the cranial and dental anatomy of Primonatalus is preserved to confirm that this taxon is unquestionably referable to the Natalidae. The combination of the derived character states listed below unite all members of the Natalidae and distinguish them from other chiropteran families.
The derived features of the Natalidae are as follows (characters in our data set are identified by their character number in parentheses followed by the character state in brackets; our characters and character states for the Natalidae are listed below): elongated, dorsoventrally flattened rostrum; reduced orbit with anterior edge located far posterior, above M2; deep funnel-shaped fossa at anterior edge of orbit enclosing lacrimal foramen and posterior opening of infraorbital foramen (39); greatly elongated infraorbital canal originating above M2 and opening anteriorly above P3 (38); elongated styliform process of ectotympanic that extends anteromedially and connects with pterygoid process (33); presence on P4 of cingular or stylar cusp labial to primary cusp (14); presence on M1 and M2 of small curved crest extending anteriorly from mesostyle into anteroexternal valley (17); manubrium (presternum) very broad with strong posteriorly oriented median keel (43); last thoracic and all lumbar vertebrae except the last 1 or 2 fused into rigid, laterally-compressed structure with prominent dorsal and ventral ridges (41); exceptionally elongated caudal vertebrae 3 through 6; femur and tibia extremely elongated and slender; proximal end of femur strongly bent laterally (46); proximal end of femur with greater and lesser trochanters reduced (48); presence of natalid organ, a subcutaneous gland-like organ on forehead and muzzle of males. Simmons (1998) listed the presence of the natalid organ as an apomorphy diagnosing the Natalidae.