CHIROPTERA, Blumenbach, 1779

Phillips, MJ, 2015, Four mammal fossil calibrations: balancing competing palaeontological and molecular considerations, Palaeontologia Electronica (Basel, Switzerland) 1 (7), pp. 1-16 : 9-11

CROWN CHIROPTERA

Node Calibrated. Chiroptera , the divergence between Yangochiroptera (e.g., pipistrelles, sheathtail bats) and Yinpterochiroptera (e.g., flying foxes, horseshoe bats). See Figure 1.4 .

Fossil Taxon. Tanzanycteris mannardi ( Gunnell et al., 2003) .

Specimen. TNM MP-207 (Tanzanian National Museum), holotype and only specimen of Tanzanycteris mannardi is a partial skeleton including skull, mandibles, vertebral column anterior to the sacrum, shoulder girdle, partial humeri, and left radius. Teeth are unknown .

Phylogenetic Justification. Gunnell et al. (2003) identified a suite of characters that place T. mannardi within Yinpterochiroptera, specifically with Rhinolophoidea. These include extremely enlarged cochlea, broadened first rib, and a dorsally flared iliac blade (this later character is also shared with some probable stem chiropterans). In my combined parsimony analysis of morphological data from Gunnell and Simmons (2005) and DNA sequences from Meredith et al. (2011), T. mannardi groups with rhinolophoids with 58% bootstrap support, while its placement within crown Chiroptera receives 83% bootstrap support. In this analysis the enlarged cochlea and broadened 1 st rib are unambiguous apomorphies for Rhinolophoidea, including T. mannardi . These characters are highly conserved among bats. The enlarged cochlea is otherwise only known from one species of mormoopid (although, without the enlarged cochlea fenestra) and a similar rib morphology is known from one other genus, Nycteris .

Hard Minimum Age. 45.0 Ma.

Soft Maximum Age. 58.9 Ma.

Age Justification. Tanzanycteris mannardi was recovered from the lacustrine Mahenge locality in north-central Tanzania. Zircon at the base of the Mahenge sequence (~ 1.2 m below the fossil) was 206 Pb/ 238 U dated by Harrison et al. (2001) to 45.83 ± 0.17 Ma. The authors also considered sedimentation rates, for which minimum estimates and error on the 206 Pb/ 238 U dates allow a minimum bound of 45.0 Ma for T. mannardi and the crown chiropteran divergence. This mid-Eocene age is also consistent with the Mahenge fossil fish fauna (e.g., Murray, 2000).

Several possible crown bats with putative yangochiropteran affinities occur in Early Eocene localities ( Eiting and Gunnell, 2009). Bats are remarkable among mammals in that accepted crown fossil records are closely bracketed by older stem fossils from all continents except Antarctica ( Ravel et al., 2011). Although no stem bats are known from prior to the Eocene, some of these records that may be used to bracket the calibration are very close to the Thanetian-Ypresian boundary and so I use the base of the Thanetian (no older than 58.9 Ma) as a soft maximum for the age of crown Chiroptera .

Discussion. Modern bats have traditionally been divided morphologically into the mainly frugivorous or nectivorous Megachiroptera ( Pteropodidae , flying foxes, etc.) and the primarily insectivorous Microchiroptera (all other families). Many of the oldest fossil bats share features associated with echolocation and general body form with microchiropterans to the exclusion of megachiropterans and other mammals. This in turn has resulted in some of the earliest bats (e.g., Icaronycteris , Australonycteris ) being linked phylogenetically to microchiropterans ( Simmons and Geisler, 1998) and used to calibrate the chiropteran crown divergence (e.g., dos Reis et al., 2012). Analyses of multiple nuclear genes ( Teeling et al., 2000, 2005) and mitochondrial genomes ( Lin et al., 2002) now provide overwhelming evidence for microchiropteran paraphyly, with rhinolophoids grouping with pteropodids. The use of molecular phylogenetic scaffolds have resulted in all Early Eocene bats that have been included in matrix-based cladistic analyses falling as stem chiropterans ( Teeling et al., 2005; Simmons et al., 2008), with their “microchiropteran” traits found to be plesiomorphic for bats.

The finding that incorrect placement of pteropodids distorts character covariation on the tree for inferring the placement of fossil bats ( Teeling et al., 2005) is also relevant here. Hermsen and Hendricks’ (2008) molecular scaffold analysis of Gunnell and Simmons’ (2005) morphological matrix clearly favoured rhinolophoid affinities for Tanzanycteris . In contrast, their combined data analysis found this rhinolophoid placement to be equally parsimonious with exclusion of Tanzanycteris from crown Microchiroptera, but with this fossil taxon still falling within crown Chiroptera . By replacing the Teeling et al. (2005) molecular matrix with the nearly 3-fold longer Meredith et al. (2011) DNA matrix, pteropodids fall back into their expected placement and most parsimonious trees again favour rhinolophoid affinities for Tanzanycteris . Hence, Tanzanycteris might yet prove to be appropriate for calibrating the younger Yinpterochiroptera node. However, only 58% bootstrap support in the present combined analysis is reason for caution. Additionally, it would have to be shown that exclusion of pteropodids from the Tanzanycteris / Rhinolophoidea grouping is not also an artefact of the morphological homoplasy that attracts pteropodids towards the chiropteran root.

Given the historical difficulties for inferring relationships among bats from morphological data, some authors may reasonably consider that a soft, rather than hard minimum bound of 45 Ma is warranted, based on the non-dental Tanzanycteris alone. Support for the minimum bound firms however, when considered in the wider fossil record context. A Tunisian rhinolophoid (dental) taxon described by Sigé (1991) adds slightly older dental evidence for Rhinolophoidea on the same continent. Chiroptera is also very likely constrained to be at least 47 Ma by Tachypteron , which is generally regarded as an emballonurid ( Storch et al., 2002), placed on the opposite side of the chiropteran root to Tanzanycteris .

More material is needed for the Tunisian rhinolophoid, and Tachypteron is yet to be tested with formal matrix-based phylogenetic analyses, but both push the balance of evidence substantially in favour of Chiroptera being at least as old as the hard minimum bound suggested here. Complete mitochondrial genome molecular dating for the chiropteran crown divergence (~54 Ma, Phillips et al., 2009), independent of chiropteran calibrations further supports the hard minimum suggested here being conservative.

Bats have relatively low fossil record completeness at the genus level ( Eiting and Gunnell, 2009). This may be due in part to often sparse diagnostic characters available at this taxonomic level from mandible fragments and isolated teeth, as well as the restriction of many genera to regions with low preservation potential or that are poorly sampled. At the ordinal level, however, the global bat fossil record is exceptional among mammals in its potential for providing a tight maximum bound. As flying mammals that filled new ecological space, bats first appear early in the Ypresian almost simultaneously in North America, Asia, Europe, Australia, Africa, and South America. In all cases (except for fragmentary dental material of uncertain affinities) these first appearances have been assigned stem, rather than crown placements ( Simmons et al., 2008, Tabuce et al., 2009; Ravel et al., 2011; Smith et al., 2007). Hence, it is likely that crown Chiroptera originated in the Ypresian, although I use the base of the Thanetian as a more conservative soft maximum.

More primitive stem bats are likely to have been geographically restricted without strong flight and difficult to distinguish from archaic insectivores ( Gunnell and Simmons, 2005), but this is not directly relevant to the question of crown Chiroptera calibration. Substantially earlier absence of crown bats is also unlikely to be explained by sampling artefacts. From their first appearance, bats occur in every subsequent sub-epoch in the fossil records of North America, Eurasia, and in at least one of the Gondwanan continents ( Gunnell and Simmons, 2005).

Most molecular dates for the origin of crown bats fall inside or very close to the bounds suggested here (e.g., Nikaido et al., 2001; Jones et al., 2005; dos Reis et al., 2012). The best estimate from the present study ( Table 1, excluding large-bodied clade calibrations) is 55 Ma, coincident with the initial radiation of bats in the fossil record. Several studies have dated crown Chiroptera more than 5 Ma older than the 58.9 Ma maximum suggested here. These studies typically employ multiple calibrations among large-bodied groups with far slower rates of molecular evolution and use possible stem bats (e.g., Ageina , Honrovits ) as crown calibrations (e.g., Bininda-Emonds et al., 2007; Meredith et al., 2011) or place a mean prior on the age of Chiroptera that is far older than any recognised bats (e.g., 65 Ma in Teeling et al., 2005).

Even without the questionable bat calibrations, re-including the large-bodied clade calibrations in the present study and removing the maximum bound for Chiroptera increases the crown age of bats to 60.2 Ma, with only 13% of the posterior distribution younger than the maximum bound. Employing the soft maximum bound buffers the influence of the large-bodied clade calibrations and pulls the mean estimate and 65% of the posterior distribution within the chiropteran calibration bounds ( Table 1). This underlines the potential importance of the chiropteran maximum bound for informing models of substitution rate variation in mammalian molecular dating studies.

Figures

FIGURE 1. The phylogenetic placement of reference fossils (full red arrows) relative to modern groups for 1, Macropodoidea; 2, Monotremata; 3 Caviomorpha-Phiomorpha; and 4, Chiroptera. These clades are shaded grey and black circles identify the calibrated nodes. Relationships among modern taxa are taken from the DNA analysis of Meredith et al. (2011), although each is supported by a broad consensus of molecular phylogenetic studies. Dotted red arrows (1, 3, 4) indicate possible alternative placements for the reference fossil that if they can be confirmed, would offer (tighter) calibration of a slightly younger node.