DISCUSSION
The rationale of this study was the expectation that the biogeographic adaptive and genetic divergences between northern quolls should be evident across their 5000-kilometre longitudinal range. Our aim was to improve our understanding on whether the species represent morphological or functional conservation units, which could be used alongside population genetic approaches in the conservation management of the species. Surprisingly, however, we found little structure in northern quoll shape variation (~76% of shape variation remains unexplained) and no strong evidence that any of the populations have evolved into discrete, possibly locally adapted, morphotypes. In particular, populations and sex differences have low effect sizes and explain less variation than size, meaning that similarly-sized individuals from opposite ends of the biogeographic distribution are likely to be similar in shape, even if they are of different sex. It also seems that most variation is evenly distributed within each of the populations, such that more localized western populations appear just as disparate in shape as individuals across the length of the eastern Queensland seaboard.
There is some limited support for our concern that developmental constraints might reduce the adaptability of northern quoll skulls, as size explains the most shape variation and the populations seem to differentiate by size. However, given the low amount of variation (~16%) that size explains, and the broad overlap of populations in size and shape, any such constraint is unlikely to be strong. However, it is an intriguing indication that at least some of the larger-scale evolutionary association between skull shape and size among marsupials may be visible at the within-species level. This is contrary to findings in other marsupials (Mitchell, Sherratt, Sansalone, et al., 2018; Weisbecker et al., 2019), and might represent one of several ways to shape morphological traits within the species.
Although differentiation of skull shape between populations had low effect sizes, populations and sex are statistically distinguishable, even when size is taken into account. This might suggest a stochastic, possibly heritable, shallow divergence between populations which, however, does not appear to reflect local adaptation. These effects also demonstrate the ability of skull shape to vary independently of size based on genetic factors related to sex or population, again contradicting the developmental constraints hypothesis. Thus, the population divergences do not appear to coincide with adaptive morphological differentiations. This provides an indication that genetic fitness benefits of outbreeding populations (Cardoso et al., 2009; Kelly & Phillips, 2019) would not risk any adverse effect due to differential local adaptation, although of course this would need to be further investigated based on non-morphological (behavioural or physiological) traits.
To better understand the functional implications of shape divergence among the northern quolls, the displacement between landmark configurations according to size, population, and sex can be visually interpreted. The landmark displacement predicted by allometry identifies two main regions of variation: larger skulls tend to have overall smaller braincases relative to the rest of the skull, a larger sagittal crest, a more anteriorly positioned masseteric scar and associated dorsally-oriented zygomatic arch. Differences between sexes include males with larger sagittal crests, smaller braincases, shorter nasals and wider zygomatic arches.
Rearrangements of the zygomatic arch and the muzzle appear to distinguish some northern quoll populations. This variation mirrors a well-known evolutionary adaptation to changes in mastication (Meloro, 2011; Mitchell, Sherratt, Sansalone, et al., 2018; Mitchell & Wroe, 2019; Weisbecker et al., 2019). It is therefore possible that the variation in muzzle length and zygomatic arch placement between some populations could be interpreted as adaptation to a particular diet or feeding habit within each population. For instance, populations in drier environments, such as Pilbara or Groote Eylandt, show a shortening in the muzzle, when compared to wetter environments, such as Northern Territory or Queensland. Shorter faces are known to imply greater bite forces, and thus might indicate the mastication of tougher foods consisting of more vertebrates (Wroe & Milne, 2007). However, whether this is truly an adaptive effect is doubtful due to the abovementioned low effect sizes and extensive overlap in shape between populations, as well as the lack of association between climatic factors and shape. In addition, although precipitation is a main predictor of quoll diets (Fancourt et al., 2015), this variable explains little variation in shape.
Aside from hypotheses of microevolutionary adaptation, it is also possible that much of the variation in the northern quoll skull derives from a re-modelling process based on individual uses of the skull. Perhaps the best example of this is the sagittal crest, which varies widely in length among northern quoll individuals. It is common for mammals – particularly males – to display larger sagittal crests with age (Flores et al., 2006), but this is a purely behavioural consequence of the pulling action of the temporalis muscle (Washburn, 1947). Similarly, bone re-modelling during an individual’s lifetime has been suggested for the zygomatic arch of mammals (Abdala & Giannini, 2000; Ravosa, 1991), and is also suspected in wombats and kangaroos (Mitchell, Sherratt, Ledogar, et al., 2018; Mitchell, Sherratt, Sansalone, et al., 2018; Weisbecker et al., 2019). It is therefore possible that the emphasis of shape variation on the masticatory apparatus, found along PC1 and the allometric pattern, arise from masticatory behaviours and possibly the “mating bite” (Braithwaite & Griffiths, 1994; Oakwood, 2000) of the larger males. This is also consistent with the weak tendency of males to have shorter muzzles than females, regardless of size, as shorter muzzles are related to higher bite forces (Wroe & Milne, 2007). This again suggests that the differences between populations are not a decisive factor to the individuals’ survival, and rather originate from potentially non-adaptive factors such as genetic drift or individual variation in feeding (Weisbecker et al., 2019) between populations.
The geometric morphometric analyses of northern quoll skull shape adds useful, quantitative, phenomic data to assessments of variation across the distribution of an endangered marsupial. The overarching find of low morphological differentiation, and very high levels of unexplained variation, has two important implications. First, it suggests that individuals of different populations are not locally adapted to the point where a separation of population phenomes is indicated (although it needs to be investigated if there might be behavioural or physiological reasons to do so). On the other hand, the lack of differentiation across the diversity of biomes, climatic conditions, or populations is a concern because it suggests a low adaptability of the species to environmental change. The concentration of shape variation in the masticatory apparatus suggests individual plasticity is a major response mechanism in the determination of northern quoll skull shape, suggesting that there is little scope for larger-scale, heritable variation within the species. A similar pattern of potentially high within-species plasticity in the masticatory apparatus has also been suggested for the living wombat species as well as kangaroos; together, the concerning suggestion is that marsupial mammals might have a scope for individual plasticity, but not evolve specific adaptations within short time spans. Further research should be directed into identifying the scope of shape variation in other threatened marsupials, investigating other climatic variables or patterns as predictors, and adding biomechanical and developmental studies to further dissect the variation that exists in this clade; in addition, a comparison with ecologically similar placental species would be useful to identify if marsupials show less intrinsic capacity of shape variation than placental mammals.