4. Discussion
We employed a geometric morphometrics framework to elucidate the intraspecific phenotypic variability of the two subspecies of G. morsitans that occur in Zambia. Population-level variability in centroid size (CS) and wing morphology can serve as a useful proxy for assessing the extent of divergence between conspecific populations (Ostwald et al., 2023) and may further provide preliminary data for the diagnosis of isolated populations (Dujardin, 2008). This information has important implications for the area-wide integrated vector management (AW-IVM) of G. morsitans in Zambia and further provides insights into the population differentiation status in its entire geographical range. Broadly, these results provide evidence for microevolutionary change in both CS and wing morphology in G. m. centralis andG. m. morsitans populations in Zambia.
Our results are consistent with the long-held observation that size sexual dimorphism is well established in tsetse as female G. morsitans were found to be larger than male flies. The estimated CS difference between the two sexes (nine percent) was similar to that reported by Hargrove et al. (2019), who found wings of female G. morsitans to be eight percent longer than those of males. This observation provides further evidence that size studies based on wing measurements as described by Hargrove et al. (2019) and CS generated by geometric morphometric analysis, produce comparable results. Therefore, both measures are reliable estimators of mean wing size inGlossina spp .
This study has demonstrated that the mean wing size of G. m. centralis is larger than G. m. morsitans . It has been reported that the size of tsetse is largely dependent on the nutritional state (Bursell, 1966) and temperature (Hargrove, 2001) experienced by the female. High temperatures exceeding 32°C result in tsetse entering cooler dark refuges such as rot holes in trees and antbear holes in the ground (Vale, 1971), a behaviour that reduces their metabolic rate but also reduces feeding opportunities (Lord et al., 2018). As such, female tsetse have reduced fat levels and produce progressively smaller pupae as temperature increases (English et al., 2016). Hargrove et al. (2018) showed that small pupae have lower fat reserves which results in the emergence of smaller-sized adults. Thus, the smaller fly size ofG. m. morsitans may be an adaptation to its occupation of a hotter environment than that of G. m. centralis as reported by Evison and Kathuria (1982) and Muyobela et al. (2023) and reaffirmed by our results. Location differences in mean wing size were observed in both subspecies’ ranges and temperature is again implicated as the major source of fly size variation.
We postulate that the observed environmentally driven fly size variation between the two subspecies may be explained by the hypotheses of phenotypic plasticity and genetic assimilation (Dujardin, 2011). Phenotypic plasticity is defined as the occurrence of phenotypic variation of a single genotype interacting with different environments (Pigliucci et al., 2006). The observed within species differences in fly size are probably adaptive to the different ecotopes where G. morsitans occurs, with plastic responses facilitating the enlargement of its ecological range. Consequently, phenotypic plasticity may have aided G. morsitans to survive in both warm (G. m. morsitans ) and cooler (G. m. centralis ) environments within its range, by providing both small and large-sized flies upon which natural selection has acted. It is conceivable that selection has resulted in fly size being genetically determined at the subspecies level through the process of genetic assimilation (Flatt, 2005), and has now become a heritable trait. Heritability for insect size has been demonstrated inAnopheles mosquitoes (Lehmann et al., 2006) and its transgenerational effects were shown in G. f. fuscipes (Mbewe et al., 2018).
Although fly size differences within the subspecies G. m. morsitans are known to occur (Bursell, 1966) and are reported in this study, it is unlikely that these within subspecies differences are heritable. This is because temperature variability within a subspecies range is expected to be less variable than across the subspecies range. Therefore, other factors that affect size variability such as host availability affecting the nutritional state of females, ovarian age, and capture month and year (Hargrove et al., 2019) are likely to be more important. Since these factors are highly variable within the subspecies range and consequently do not exert selection in any specific direction, fly size change concerning these factors is unlikely to result in heritable change (Jirakanjanakit et al., 2007). As such, size is expected to be a poor discriminator of G. morsitans subspecies population structure.
Our results showed that allometry and environmental variability accounted for 11.6% and 10.7% of shape variation in G. morsitans . As such, we estimate that 77.7% of wing shape variation could be attributed to genetic effects, a finding in support of the suggestion by Patterson and Klingenberg (2007) that shape exhibits high genetic determinism. The low contribution of environmental variability to allometry-free wing shape variation suggests that G. morsitanswing shape exhibits high environmental canalization, in agreement with results from other Diptera, sand flies (Dujardin and Le Pont, 2004) and mosquitoes (Henry et al., 2010).
We found that wing shape in G. morsitans varies according to gender, subspecies, and location of origin. The detection of allometric-free shape sexual dimorphism indicates that the phenotypic expression of wing shape in this tsetse is sex-specific. Shape sexual dimorphism has been reported in other Dipteran families such as Drosophilidae (Gilchrist et al., 2000) and Culicinae (Virginio et al., 2015). Gilchrist et al. (2000) suggest that the gender regulation of shape in the Diptera represents a developmental constraint during morphogenesis rather than adaptive change. Tsetse biology appears to support this view as female flies reproduce by adenotrophic viviparity (Vreysen et al., 2013) which may present a different aerial dynamic challenge to pregnant females compared to males, hence the need for female wings to be designed differently. Evidence of strong genetic determinism of wing-shape sexual dimorphism in the Diptera has been presented by Cowley et al. (1986).
Subspecies wing shape variation in G. morsitans may be an adaptive trait as G. m. centralis and G. m. morsitans in different habitats with different aerodynamic conditions due to temperature differences. Temperature is known to significantly affect aerodynamic lift (Jun et al., 2015). As air temperature increases, its density decreases leading to a decrease in the amount of lift generated by the wings. Therefore, selection may be acting on the wing phenotypes of the two subspecies differently as G. m. centralis occupies a cooler environment than G. m. morsitans , thereby producing wing shapes aerodynamically suitable for their specific environments. Ray et al. (2016) showed that selective pressure resulting in large and small changes in the wing shape of Drosophila can lead to significant changes in key flight performance metrics, leading to improved manoeuvrability and agility.
Significant wing shape variation was also observed within the subspecies ranges of both G. m. centralis and G. m. morsitans . As environmental conditions within the specific subspecies range are similar, it is unlikely that aerobatic capability selection is the primary cause of this variation. Since shape is known to be the output of polygenic genes (Patterson and Klingenberg, 2007), it is more likely that this within subspecies variation is non-adaptive and a result of random genetic drift. Several field studies have implicated genetic drift as a source of shape variation among geographic isolates of conspecific populations (Dujardin, 2011; Henry et al., 2010; Kaba et al., 2012a; Ravel et al., 2006). Shape change due to genetic drift has also been demonstrated in the laboratory (Jirakanjanakit et al., 2007).
Based on the allometry-free shape variation data, our results suggest that both G. m. centralis and G. m. morsitans populations in Zambia are highly structured and exhibit significant morphological divergence. Within the G. m. centralis range, structuring is probably the result of the physical separation of KNP, KSP, and SNP populations (Fig 1) by large areas of unsuitable habitat (Muyobela et al., 2023). Such geographic isolation in natural conditions tends to quickly generate wing shape changes due to genetic drift. Physical separation between the two G. m. centralis locations in KNP and all sample locations within the G. m. morsitans does not occur (Fig 1). The observed population structuring at these locations could be attributed to the notion that tsetse is essentially a local insect and the interchange of individuals between contiguous parts of the general population is limited (Bursell, 1966).
The observed structuring of G. morsitans populations in Zambia suggests that the implementation of tsetse population management technologies that target an entire isolated population may be technically feasible. However, to categorically designate populations as isolated, it is essential to estimate the number of migrants per generation or the levels of gene flow between them (Bouyer et al., 2007), and methods using morphometric variation are not suited for these tasks. Therefore, the results presented in this study only provide preliminary information justifying further investigation using molecular techniques to conclusively identify genetically isolated populations (Dujardin, 2008). This is particularly crucial in the G. m. morsitans range where physical separation of sample location was not apparent. It should be noted however, that some authors have suggested that results from geometric morphometric studies are comparable to those of molecular studies using microsatellite markers (Bouyer et al., 2010, 2007; Solano et al., 1999).
We conclude that G. morsitans populations in Zambia exhibit significant population-level variation in body size and allometry-free wing shape. This variation suggests high levels of population structuring that may be indicative of population isolation. Molecular studies to estimate the levels of gene flow between these populations and determine their levels of genetic isolation will be able to shed even more on G. morsitanspopulation structure in Zambia and its under lying drivers.