CICHLIDAE

Benton, MJ, Donoghue, PCJ, Vinther, J, Asher, RJ, Friedman, M & Near, TJ, 2015, Constraints on the timescale of animal evolutionary history, Palaeontologia Electronica (Florence, Italy) 15 (1), pp. 1-107 : 43-44

CROWN CICHLIDAE (37)

Node Calibrated. Divergence between Etroplinae and Pseudocrenilabrinae.

Fossil Taxon and Specimen. Mahengechromis plethos from Mahenge, Singida Plateu, Tanzania (holotype NMT WM 339 /96a, b, National Museum of Tanzania, Dar es Salaam, Tanzania) .

Phylogenetic Justification. The most compelling evidence for cichlid monophyly derives from aspects of soft-tissue anatomy and other details unlikely to be preserved in fossils (Stiassny, 1981), but placement of Mahengechromis within Cichlidae is supported by the structure of the lower pharyngeal jaw, details of squamation, and meristic counts of the vertebral column and median fins (Murray, 2000, 2001a, b). The derived presence of a single supraneural and ctenoid scales indicate that Mahengechromis belongs to crown-group Cichlidae generally and Pseudocrenilabrinae specifically (Murray, 2000: figure 11; Murray, 2001b: figure 4).

Minimum Age. 45.46 Ma

Soft Maximum Age. 100.5 Ma

Age Justification. The fossil deposits at Mahenge represent lacustrine sediments that accumulated within a crater formed by intrusion of kimberlite. U- Pb dating of a zircon crystal provides a well-constrained age estimate of 45.83 Ma ± 0.17 Myr for the kimberlite at Mahenge ( Harrison et al., 2001), but the lacustrine sediments yielding Mahengechromis must postdate this intrusive event. Harrison et al. (2001) provide a convincing argument, based on other crater lakes, that kimberlite emplacement likely predates lacustrine deposition and complete basin infill by only 0.2 to 0.1 Ma. We therefore define a minimum age of 45.46 Ma for Mahengechromis and crown-group Cichlidae .

A soft maximum age for crown Cichlidae can be derived from the earliest fossil occurrences of more inclusive acanthomorph clades containing the family. The oldest well-substantiated remains of Ovalentaria only predate the earliest fossil chichlids by a few million years (see node 36). We prefer to adopt a more generous soft maximum based on the rich fish faunas of Cenomanian age that yield a diversity of acanthomorphs belonging to deeply diverging groups like polymixiids, lampridiforms, paracanthopterygians, and beryciforms, but only questionable percomorphs (faunas reviewed in Forey et al., 2003; see calibration for node 35 for a discussion of uncertainties surrounding early Late Cretaceous tetraodontiforms). The base of the Cenomanian is dated as 100.5 Ma (Ogg et al., 2012b), which we propose as a soft maximum age for crown Cichlidae .

Discussion. A series of articulated cichlids have been described from the ‘Faja Verde’ level of the lower Lumbrera Formation of northwestern Argentina (Malabarba et al., 2006, 2010; Perez et al., 2010). Published age estimates for these fossils range from latest Paleocene-earliest Eocene (Sempere et al., 1997) to middle Eocene (Malabarba et al., 2006, 2010; Perez et al., 2010). These proposed ages are particularly remarkable due to the fact that one of the cichlids described from the ‘Faja Verde’ has been interpreted as nesting within an extant genus (Malabarba et al., 2010).

Historically, the Lumbrera Formation has been assigned to the Casamayoran South American Land Mammal Age, although this deposit shares no taxa in common with reference horizons in Patagonia (del Papa et al., 2010). Attribution to this land mammal age provides only the coarsest chronostratigraphic resolution; the Casamayoran is correlated with the Ypresian through early Priabonian stages of the international timescale, and therefore spans most of the Eocene (ca. 19 Myr duration). The top of the Casamayoran can be dated only approximately through correlation, yielding an estimate of 37.5 Ma (Vandenberghe et al., 2012). The most robust minimum age constraint for the lower Lumbrera Formation cichlids comes from U/Pb zircon date of 39.9 Ma for the upper Lumbrera Formation (del Papa et al., 2010). López-Fernández et al. (2013) have recently used cichlids from the ‘Faja Verde’ to calibrate a molecular clock for neotropical cichlids. They assigned a hard minimum age of 49.0 Ma to these taxa, which is nearly 10 Ma older than the more robust minimum constraint endorsed here. The source of this older estimate provides no numerical estimates for the age of the Lumbrera Formation and does not mention this deposit specifically (White et al., 2009). If an age of 49.0 Ma for the Lumbrera cichlids can be substantiated, these fossils would provide the appropriate minimum age constraint for crown Cichlidae .

It is commonplace for molecular clock analyses to specify minimum ages for cichlid divergences based on biogeographical scenarios invoking vicariant events associated with the mid- Mesozoic breakup of Gondwana ( Genner et al., 2007; Azuma et al., 2008; López-Fernández et al., 2013), but we do not advocate this calibration strategy at present. These hypothesized divergences substantially predate the earliest fossil evidence for not only cichlids, but also a series of more inclusive teleost clades (e.g., Ovalentaria, Percomorpha, Acanthomorpha, Ctenosquamata; Lundberg, 1993). Additionally, molecular timescales based on well-justified calibrations derived from marine fish fossils with narrowly constrained ages indicate a Late Cretaceous-early Paleogene origin for Cichlidae (Santini et al., 2009; Near et al., 2012, 2013; Betancur-R. et al., 2013; McMahan et al., 2013; Friedman et al., 2013), which is more consistent with the observed paleontological record of the group (Murray, 2001b) and statistical estimates of times of origin based on the distribution of fossil horizons and outgroup ages ( Friedman et al., 2013).