1. Introduction
Glossina morsitans (Diptera: Glossinidae) is a savannah tsetse species of the subgenus Glossina (morsitans group) whose distribution is restricted to savannah woodlands (Leak et al., 2008) and is correlated with that of wildlife (Vreysen et al., 2013). Three allopatric subspecies occur, namely, G. m. submorsitansNewstead, G. m. centralis Machado, and G. m. morsitansWestwood (Jordan, 1993), all of which are efficient vectors of trypanosomes (Kinetoplastida: Trypanosomatidae), which cause human and animal trypanosomiasis in sub-Saharan Africa (Rogers, 2000). The geographical distribution of G. m. submorsitans is from Western to Central Africa, while G. m. centralis and G. m. morsitans occur in Eastern, Central, and Southern Africa (Rogers and Robinson, 2004). In Zambia, G. m. centralis and G. m. morsitans are predicted to occupy 151,353 km2 or 20% of the land mass (Muyobela et al., 2023).
In conformity with most insect species, the distribution of G. morsitans within its geographic range is generally discontinuous (Krafsur, 2009; Muyobela et al., 2023), being strategically arranged based on the availability of food sources, reproductive needs, dispersal capacity, and adaptation to local environmental conditions (Dujardin, 2008). Due to this spatial arrangement, adaptation to local conditions can give rise to population-level phenotypic variation that may result in the structuring of populations into biogeographical islands or subpopulations (Dujardin and Le Pont, 2004; Getahun et al., 2014; Mbewe et al., 2018). In the presence of significant barriers to the exchange of individuals between them, these subpopulations become isolated and can undergo rapid evolutionary changes in morphological traits due to phenotypic plasticity, founder effects, and genetic drift (Dujardin, 2011; Ostwald et al., 2023). The identification of isolated tsetse populations has been deemed crucial for the successful and sustainable implementation of area-wide integrated vector management (AW-IVM) (Bouyer et al., 2010; Kgori et al., 2006), guiding the decision whether to undertake suppression or elimination campaigns (Bouyer et al., 2007).
A relatively low-cost approach for investigating tsetse population structure is the use of landmark-based geometric morphometrics (GM), defined as the statistical analysis of shape variation and its covariation with other variables (Rohlf and Bookstein, 2003). Unlike traditional morphometrics, GM is a powerful technique that captures the geometry of the morphological structure under study and retains this information throughout the analysis (Zelditch et al., 2004). The procedure is accomplished through the Procrustes paradigm (Adams et al., 2013) in which a set of two-dimensional landmark coordinates recording the relative positions of homologous anatomical points are obtained and then subjected to generalised Procrustes analysis (GPA) (Rohlf and Slice, 1990). This least-squares superimposition technique produces a set of shape variables whose geometric dissimilarity is expressed as the Procrustes distance between the homologous points of two configurations (Zelditch et al., 2004) and whose pattern of variation can be visualised by graphical methods (Baken et al., 2021). An additional output of this analysis is centroid size (CS), defined as the square root of the summed squared distance of each landmark from the centroid of the form (Tatsuta et al., 2018). This isometric measure of size is used as an estimator of the global size of the form under study in GM studies (Dujardin, 2008).
Conspecific size variability within and among insect populations is generally known to be an environmentally induced and reversible character (Jirakanjanakit et al., 2007). In G. morsitans , size variability has been attributed to seasonal effects (Hargrove et al., 2019) with temperature being the major source of variation (Glasgow, 1961; Phelps and Clarke, 1974). High heritability values for insect size have however been reported (Lehmann et al., 2006) and the transgenerational effects of size among the Glossina spp have been demonstrated (Mbewe et al., 2018). Therefore, heritable size variation can be used to discriminate populations. Size-corrected or allometry-free shape is known to be a polygenic character and strong evidence of its genetic determinism has been provided (Klingenberg and Leamy, 2001; Patterson and Klingenberg, 2007). Allometry-free shape has also been shown to be a powerful discriminator of groups (Dujardin, 2008) and is, therefore, a very useful tool in taxonomic studies (Klingenberg, 2016).
The insect body part most subjected to GM studies is the wing (Tatsuta et al., 2018). This is due to several reasons. Firstly, insect wings are almost entirely two-dimensional structures, a fact that greatly reduces digitisation errors (Dujardin, 2008). Secondly, the arrangement and branching patterns of insect wing veins contain taxonomic information that has been used to construct classification schemes, infer phylogeny (Bybee et al., 2008), elucidate evolutionary patterns (Debat et al., 2003), and evaluate fluctuating asymmetry – deviations from perfect symmetry that indicate developmental noise (Klingenberg et al., 2001). Lastly, the geometric shape of insect wings has been shown to exhibit high environmental canalisation – the ability of a genotype’s phenotype to remain relatively invariant when exposed to different environments (Henry et al., 2010). These attributes, therefore, make the geometric shape of insect wings, a suitable phenotypic character to distinguish conspecific populations and species using GM (Dujardin, 2011). Insect wing shape is captured by placing homologous landmarks on the intersection of wing veins.
Geometric morphometrics has been used to study natural population variation in several species of Glossina . These include G. palpalis gambiensis (Bouyer et al., 2007; Solano et al., 1999),G. p. palpalis (Ebhodaghe et al., 2017; Kaba et al., 2012b),G. m. submorsitans (Achukwi et al., 2013), G. pallidipes(Getahun et al., 2014), G. austeni (De Beer et al., 2019),G. fuscipes fuscipes (Mbewe et al., 2018), G. tachinoides(Mustapha et al., 2018) and G. brevipalpis (De Beer et al., 2019). However, phenotypic variation in natural populations of G. m. centralis and G. m. morsitans has not been investigated. Therefore, this study aimed to use landmark-based wing geometric morphometrics to investigate phenotypic variation and determine the level of population structuring in G. m. centralis and G. m. morsitans populations in Zambia.