Pulanesaura, AND

PULANESAURA AND THE MEANING OF SAUROPODA

The relevance of the morphologies outlined above to the possible grouping relations of Pulanesaura is unfortunately obscured by the non-preservation in the latter of many of the structures in question. Although specific features of the humerus and ulna of Pulanesaura were proposed in McPhee et al. (2015a) as presenting a shift towards a more sauropod-like form of forelimb parasagittalism, their disassociation and poor manner of preservation severely limits their contribution to our understanding of that particular transition. For these reasons, Vulcanodon continues to represent perhaps the best exemplar of the basal-most sauropod condition, an interpretation reflected in the retention of the node-based definition of Salgado et al. (1997, see above) preferred here.

If the above discussion has emphasized anything it is that, with respect to available information (e.g. there is currently a near-total absence of cranial information known for early sauropods), it is apomorphies of the forelimb that provide the most cogent approximation of the traditional conception of sauropods as ‘columnar-limbed quadrupeds’. This condition is represented by an integrated suite of modifications that clearly reflect a shift towards parasagittalism of the structure [see Remes (2008), McPhee et al. (2015a) and above; this transition was also potentially correlated with an increase in the relative elongation of the metacarpus, as suggested by surviving information on the fragmentary manus of Vulcanodon ( Raath, 1972; Remes, 2008)]. Moreover, in contrast to the hindlimb modifications discussed above, which are potentially more sensitive to size increase and hence susceptible to misleading convergences, the derived sauropodan forelimb configuration appears to have evolved only once – at the base of what can historically be considered Sauropoda ( Salgado et al., 1997; Wilson & Sereno, 1998; Sereno, 1999b). (N.B. Although these features were undoubtedly acquired in a stepwise fashion as part of a multi-taxic transformation series, current phylogenetic hypotheses nonetheless suggest that the ultimate distillation of the complex occurred relatively rapidly.)

The tendency for specific sets of features (i.e. complex apomorphies) to regularly optimize at or around a specific phylogenetic locus is intriguing, and is potentially emblematic of interesting evolutionary events that mark a genuine departure from the historical dynamics that had held sway prior [e.g. adaptive radiations, vicariant cladism ( Assis & de Carvalho, 2010)]. This has already been discussed in more general terms with respect to sauropod evolution, with Barrett & Upchurch (2007: 105) suggesting that the initial radiation of Sauropoda is consistent with a ‘correlated progression’ model of evolution whereby the development of one character(s) facilitates the elaboration/origin of another within a positive feedback loop of mutual fitness (e.g. herbivory favours increasing stomach capacity favours increasing size favours quadrupedality favours a columnar forelimb stance favours further specializations of the feeding apparatus; see also McPhee et al., 2015a). This idea is also mirrored in recent ‘cascade’ theories pertaining to the origins sauropod giganticism ( Sander et al., 2011; Sander, 2013) [see Button, Barrett & Rayfield (2017) for a recent synthesis of these two concepts].

As a further conceptual consideration, it is worth asking if, in certain instances, the morphological tokens of such events should be reflected in the codification/classification of the resultant clade? The relevance of apomorphies to our classificatory schemes has a long and rather fraught history, with the primary criticism being the perceived lability – and potential homoplastic expression – of character data compared to that of taxon-based specifiers (see discussion in, e.g., Benton, 2000; Nixon & Carpenter, 2000; Langer, 2001; Lee, 2001; Sereno, 2005; Bertrand & Härlin, 2008). Nonetheless, the mutability of most cladistic topologies renders the choice of ‘node-fixing’ object a relatively arbitrary one ( Härlin, 1998; Bertrand & Härlin, 2008). In this sense, the repeated tendency for a subset of characters to concentrate upon a particular point of phylospace suggests that the ‘stability’ criterion ( Cantino & de Queiroz, 2010) is fulfilled no more or less successfully via apomorphies than is the case with a particularly ‘dependable’ taxonomic specifier ( Franz, 2005). A potential corollary of incorporating character information (when it is warranted, e.g., in pectinate grades) into classificatory considerations is the closer epistemic link between character observations (homology hypotheses) and the manner in which those observations are co-opted into phylogenetic reconstructions. This suggestion has been addressed (although not always explicitly) in the recent literature on natural kinds (homeostatic property cluster natural kinds sensu Boyd, 1991, 1999). The explanatory strength of this line of theorizing, it is argued, extends from treating properties not as traditional essences, but as integrated clusters that co-occur with better than chance probability (projectability), thus contributing to scientific induction and generalization ( Griffiths, 1999; Keller, Boyd & Wheeler, 2003; Rieppel, 2005, 2006, 2008; Rieppel & Kearney, 2007; Assis & Brigandt, 2009). The homeostatic mechanism in this sense refers to the causal processes that maintain the boundary and integrity of the kind and, although generally historical (i.e. genealogical) in nature, can also include, for example, developmental, modular and/or functional influences (see Wagner & Schwenk, 2000; Wagner, 2001; Rieppel, 2005).

Although this form of character (or, more appropriately, topological) congruence could also potentially be couched in an individualist paradigm as a form of ‘recurrence of concordance’ [sensu Haber (2016), see that work for further discussion, cf. ‘consilience of inductions’ ( Ruse, 1987)], it is likely that natural kinds qua similarity classes will remain of greater pragmatic (i.e. empirical) relevance to most palaeontologists given the nature of our data [although these approaches are not entirely mutually exclusive ( Pleijel & Härlin, 2004; Brigandt, 2009)]. Thus, the defining features of the sauropod forelimb meet the epistemic (inductive and explanatory) requirements of an integrated locomotory complex clearly distinguishable from that of the plesiomorphic sauropodomorph condition ( McPhee et al., 2015a). A more focused treatment of character information with respect to our naming practices also represents a potential compromise between positions that treat names as (often theory-free) tokens of ostensive reference (to, e.g., hypotheses: Härlin, 1998; Bertrand & Härlin, 2008; cf. Fitzhugh, 2008) and those that view names as emblemizing biologically interesting groupings accessible to scientific description, understanding and communication ( Keller et al., 2003; Franz, 2005; Rieppel, 2005, 2006, 2008; Rieppel & Kearney, 2007; Assis, 2011). In summary, although these issues warrant an independent, more in-depth future treatment, it is nonetheless clear that Sauropoda has a genuine meaning formed of the intensional properties generally associated with it. Greater consideration of the properties that render organismal groupings of interest in the first place is required in order to reconcile classificatory practice with the explanatory aims (e.g. repeatability, communication) of empirical science.