CONCLUSIONS

Sphenodon is the best available outgroup for inferring soft tissue homologies and soft tissue character polarity in squamates (e.g., Schwenk 1986; Bryant and Russell 1992; Witmer 1995, 1997; Abdala and Moro 2003; Tsuihiji 2005, 2007; Holliday and Witmer 2007). However, Sphenodon's status as a representative of the ancestral condition must be approached with caution as squamates and rhynchocephalians separated from one another approximately 240-250 million years ago (Evans 2003; Vidal and Hedges 2005; MEHJ unpublished data), and Sphenodon is certainly not representative of Rhynchocephalia as a whole (e.g., Evans 1980; Whiteside 1986; Fraser 1988; Reynoso 2000; Apesteguía and Novas 2003; Jones 2008). Incidentally, it is not correct to refer to Rhynchocephalia as being basal to Squamata (e.g., Reilly et al. 2001, p. 403) because, by definition as sister taxa, both groups originated at the same time (Evans 2003; MEHJ unpublished data). Correspondingly, Squamata can be used as an outgroup for Rhynchocephalia, and Sphenodon is not "the most basal member of the extant lepidosaurs" (contra Tsuihiji 2005, p. 176; see Baum et al. 2005). Despite frequent claims to the contrary (e.g., Sharell 1966; Crook 1975; Dawbin 1982; Daugherty and Cree 1990; Finch and Lambert 1996; Pough et al. 2005; Alibardi and Toni 2006; Reilly et al. 2006), there is very little direct evidence that Sphenodon has remained "unchanged" for 140 million years or more (Whiteside 1986; Jones 2008, Jones et al. 2009; Evans and Jones in press). Sphenovipera, from the Early Jurassic of Mexico, known from a single lower jaw (Reynoso 2005), has Sphenodon-like dentary teeth but the proportions of the lower jaw are notably different. Its adductor fossa is very short, suggesting its temporal region was also short. The most plesiomorphic rhynchocephalians, such as Diphydontosaurus from the Triassic of England, appear to have employed some kind of propalinal jaw movement as in Sphenodon (Evans 1980; Fraser 1982; Whiteside 1986; Jones 2008), but differences in skull shape suggest that they may have differed in muscle arrangement (Evans 1980; Whiteside 1986; Jones 2008).

Here we have provided an up-to-date overview of the head and neck musculature with novel contributions from direct examinations of wet material. The musculature is complicated and composed of several distinct groups themselves amenable to further subdivision (e.g., Figure 13, Figure 21, Figure 22, Figure 42, and Figure 43).

Several aspects of the muscle arragement are interesting from a functional point of view, for example, the fact that the large m Adductor Mandibulae Externus Superficialis attaches to a strong sheet of fascia held within the lower temporal fenestra (Figure 20). This has previously attracted little attention but in other circumstances it has been noted that where skeletal structures are in net tension through all functional loadings, bone is often replaced by ligament or membrane (Oxnard et al.1995; Witzel and Preuschoft 2005]) whereas net compression ensures bone deposition (Gregory and Adams 1915; Adams 1919; Case 1924; Olsen 1961). It is possible that the lower temporal fenestra (and fascia) exists because there is net tension as a result of the functioning of the m Adductor Mandibulae Externus Superficialis (when the lower jaw meets resistance). Another point of interest involves the position of the m. Pterygoideus Atypicus (Figure 31 and Figure 37). Because the vertebrate jaw joint is usually closer to the jaw muscles than are the teeth contacting the prey item, joint reaction forces are expected to be higher than those resulting from the bite. It has been argued that mammals reduce this problem by the development of an anteriorly placed masseter muscle (Crompton and Parker 1978; Russell and Thomason 1993, p. 351). The m. Pterygoideus Atypicus in Sphenodon, and other reptiles that posses it, may similarly offset reaction force to some extent.

As previously described, the tongue morphology of Sphenodon is very similar to that of iguanian squamates (Oelrich 1956; Schwenk 1986, 1988; Smith 1988). Cladistic analyses based on morphological data (e.g Estes 1988; Conrad 2008) suggest this shared tongue morphology may represent the ancestral condition for lepidosaurs. By contrast, phylogenetic topology based on molecular data (Townsend et al. 2004; Vidal and Hedges 2005) favour this tongue morphology having been acquired independently by Sphenodon and iguanians. Extinct rhynchocephalians demonstrate different palatal tooth arrangements compared to Sphenodon that may also reflect differences in tongue structure (Jones 2008).

Discrepancies in previous descriptions with regards to the details of origins and insertion are probably due to a number of factors including specimen quality, descriptive accuracy, and criteria for homology (e.g., see Haas 1973, p. 293-296). The problem is exacerbated by mistakes such as those found in the figure labelling of Haas (1973), Holliday and Witmer (2007) and possibly Gorniak et al. (1982). Nevertheless, individual variation is probably a 'real' and significant factor (Ostrom 1962; Haas 1973; Wu 2003). In fact, individual variation is probably greater than reported because when confronted by uncertain morphology authors may, on occassion, have chosen to follow previous descriptions. The level of intraspecific variation in muscle structure found in Sphenodon may be no greater than that in any other taxon, being evident only because of the number of descriptions carried out. Alternatively the isolated island existence may have allowed intraspecific variation to increase under relaxed selection pressures (Whiteside 1986, p. 425). Ostrom (1962) suggested that individual variation might be related to environmental factors such as favoured diet but Haas (1973) considered ontogeny to be the main contribution to morphological variation.

To reduce the possibility of future misunderstandings we advise that further descriptions of muscle arrangement in reptiles should use previously established colour coding schemes and abbreviation formatting (e.g., Holliday and Witmer 2007; Tsuihiji 2005, 2007) as far as possible.

The work presented here will form the basis of future computer modelling work (e.g., Curtis et al. 2008; Moazen et al. 2008, 2009) to evaluate the complex relationships between muscle arrangement and skull shape in diapsid reptiles.