Adults of T. infestans included in this study were collected from domestic and peridomestic habitats in an urban area contiguous to San Juan city (Rawson department) (31° 39′ 48” S; 68° 27′ 11” W) and from rural areas in Sarmiento department (32° 08′ 46” S, 68° 42′ 56” W), San Juan province, Argentina, during 2018 and 2019. In each habitat, 20 to 30 females and males were collected and taken to the laboratory in Buenos Aires, Argentina, where they were stored in 70% ethyl alcohol.

At the laboratory, adults were dissected by removing the head and pronotum. The thoracic cavity where flight muscles are located was observed under a stereomicroscope (Zeiss SV11, Germany), and flight muscles were classified into three different categories: Type I: developed flight muscles (11–21% of total body mass), completely filled thoracic cavity, and firm consistency and clear development of longitudinal and transverse muscles (Figure 1A). Type II: almost empty thoracic cavities (3–9% of total body mass), except for the small whitish thoracic contents of almost liquid consistency (Figure 1B). Type III: thoracic cavity completely empty, dark brown and without any liquid content (not shown). According to our previous observations, in T. infestans adults collected in the field or under experimentation, type III corresponds to insects that died of starvation and presumably consumed their muscles during the starved period. This situation was described for other insect species (Marden, 1987). Hence, adults that present Type III flight muscles were not included in this study. After the classification of insects according to flight muscles, a total of 78 females (55 and 23 with and without flight muscles, respectively) and 83 males (56 and 27 with and without flight muscles, respectively) were included in this study. Insects, both with and without developed flight muscle, exhibited macropterous forewings. Left and right forewings of the 161 insects were mounted between microscope slides and cover slips, and photographed with a digital camera (Nikon S9900) connected to the stereomicroscope using a 6× magnification. All images included a reference scale.

To quantify variation in wing CS and shape, we used landmark-based geometric morphometrics (Bookstein, 1997). We collected 10 type-I landmarks positioned at vein intersections as described by Schachter-Broide et al. (2004) using the free software tpsDig2 (version 2.31) (Rohlf, 2017). Both right and left wings were digitized twice to estimate measurement error (ME), which is of critical importance in the analysis of FA (Palmer, 1994). We used a generalized full Procrustes fit to superimpose all configurations and remove the effects of position, orientation, and size (Rohlf and Slice, 1990). To analyze matching symmetry, this procedure superimposes the left and right wings, thus allowing the separation of symmetric and asymmetric variation components (Mardia et al., 2000; Klingenberg et al., 2002). We computed CS (i.e., the square root of the sum of squared distances from each landmark to the centroid of the configuration) as a measure of wing size (Dryden and Mardia, 1998). Allometric relationships between wing shape and CS for females and males with and without developed flight muscles were checked with a multivariate regression of Procrustes coordinates on CS. Regression results confirmed significant allometric effects for all groups (p < 0.001 in all cases), but with relatively small allometric effects (less than 7% in all cases).

In addition, the maximum wing length and the total wing area were measured for all sampled individuals. These measurements were taken using the ImageJ software, version 1.52 (https://imagej.nih.gov/ij/).

A principal component analysis (PCA), based on the covariance matrix of landmark coordinates after the Procrustes fit, was performed to examine dimensionality of shape variation and the global shape data for the complete data set. We compared CS between females and males with and without developed flight muscles using ANOVA tests.

Canalization and developmental stability, as estimators of developmental buffering, were examined in each of the four groups (females and males with and without flight muscles) considering the amount of individual and FA variation in wing CS and shape. We used two-way ANOVAs for size and Procrustes ANOVAs for shape (Klingenberg and McIntyre, 1998), considering individual variation, side, interaction between individual and side as effects, and the ME. The effect of individuals represents the variation among individuals corrected for any asymmetry effects and is considered the symmetric component of shape variation. The side represents the variation due to differences between the right and left wing and is useful to evaluate directional asymmetry. Finally, the interaction between individual and side captures the variability of right–left differences among individuals (FA) and is interpreted as the asymmetric component of shape variation (Klingenberg et al., 2002). The mean square (MS) related to the interaction (individual × side) was used to compute the FA10 index as FA10 = (MS interaction–MSME)/2, following Palmer (1994). This index accounts for measurement error and provides a reliable estimation of FA. Variance estimates were then compared among groups using standard F-tests. To determine the relationship between canalization and developmental stability, the correlation between symmetric and asymmetric variance covariance (VCV) matrices was evaluated within each group for both females and males. Furthermore, we tested whether symmetric and asymmetric wing shape variation components were differently structured in the two groups for each sex. For this purpose, we examined the correlation between VCV matrices examining matrix association across groups. Afterwards, we examined if these differences were reflected in wing shape. We first examined whether individuals without flight muscles exhibited a higher total wing shape variance than individuals with flight muscles. We calculated shape variance as the trace of the within-population-type VCV matrix (Zelditch et al., 2004) and then compared variances across groups within each sex with F-tests.

To examine whether different areas of the landmark configuration correspond to different developmental and/or functional modules, we conducted a modularity analysis of wing shape configuration. We first formulated a developmental hypothesis (hypothesis 1) stating that the landmarks of the stiff proximal region and the membranous apex are two separate modules. This hypothesis was formulated according to the patterns of expression of the genes nub and vg, which specify these two parts of the wing in Hemiptera (Medved et al., 2015; Figure 2A). For Hemiptera, no previous studies have defined the functional modules into which the hemelytra wing can be divided. Therefore, the different modularity hypotheses proposed here are exploratory hypotheses to test functional modularity. We formulated three additional hypotheses of modularity (hypotheses 2, 3 and 4): (1) proximal, central and distal wing compartments are distinct modules (Figure 2B), (2) the anterior and posterior compartments of the wing are distinct modules (Figure 2C), and (3) the anterior, central and posterior compartments of the wing form separate modules (Figure 2D). The modularity hypotheses were evaluated by computing the strength of covariation between the hypothetical modules with alternative partitions of landmarks into subsets of the corresponding sizes (Klingenberg, 2009). If a hypothesis of modularity were correct, the covariation between those subsets of landmarks should be weaker than the covariation between different partitions of the landmarks into subsets (Klingenberg, 2009). We computed the RV coefficient (Escoufier, 1973), which is a multivariate measure of covariation between two sets of landmarks (hypotheses 1 and 3) and the multiset RV coefficient (Klingenberg, 2009) when three compartments were proposed (hypotheses 2 and 4). To test the modularity hypotheses, we used a single Procrustes fit for all landmarks jointly and then examined the covariance of subsets of landmarks within the overall configuration (simultaneous-fit). Spatially contiguous partitions were considered and defined with an adjacency graph (Supplementary material 1): a partition of landmarks is spatially contiguous if all resulting subsets of landmarks are connected through the edges of the adjacency graph (Klingenberg, 2009). All morphometric analyses were performed with the free software MorphoJ (version 1.07a) (Klingenberg, 2011).

Finally, we compared the morphological aspect of wings among females and males with and without developed flight muscles. For this, the estimator known as wing aspect ratio: 4 × (wing length)²/wing area (Betts and Wootton, 1988) was compared with one-way ANOVAs tests among groups.

The integrated assessment of various approaches that gaged the contribution of the two measured components of developmental homeostasis, i.e., canalization and developmental stability, revealed that adults without developed flight muscles suffer impaired wing development. The common mechanism underlying the variation of canalization and developmental stability was reduced in males and females without flight muscles. These results evidence that a decrease in developmental buffering can be expected in the lack of muscle development, probably associated with reduced wing stability. In flight muscle polymorphic species, female flightless morphs reach sexual maturity more rapidly and attain higher fecundity compared with females with flight muscles (Johnson, 1969; Roff, 1986; Roff, 1990). For males, reproduction was also found to be enhanced in flightless morphs (Crnokrak and Roff, 1995). These evidences suggest that individuals investing more in reproduction would have experienced reduction in wing stability and individuals that migrate by flight would have invested in a flight migratory syndrome that involved traits, such as morphology, hormone titers, development time and growth rate, distribution of energy stores, flight propensity, and age-specific reproduction (Roff and Fairbairn, 2007).

Our findings indicate that the developmental stability process was affected in individuals without developed flight muscles, with very high increases in fluctuating asymmetry both in shape and size being recorded in comparison with individuals with developed flight muscles. However, for individuals without flight muscles, the canalization process remained active, preventing the final morphology of the wings from being highly polymorphic and maintaining ontogenetic patterns similar to those of individuals that developed under more favorable conditions. For T. infestans, wing FA associated with different host-feeding sources and habitats determines different patterns and amount of FA (Nattero et al., 2015a,b). Another study, in which T. infestans females and males were collected before and 4 months after a community-wide pyrethroid spraying campaign, showed consistently small levels of FA for those individuals collected 4 months after spraying (Nattero et al., 2019). This pattern was suggested to be related, among other possible causes, to the lower triatomine densities after insecticide spraying and the concomitant increase in feeding success of these insects.

We found a reduction of a common mechanism underlying the variation of canalization and developmental stability, which was confirmed by the lack of significant correlations between the VCV matrices corresponding to FA and among-individual variation for females and males without developed flight muscles. For T. infestans, the study of canalization and developmental stability related to the effects of a sublethal application of a pyrethroid insecticide suggested that treatment affected only individual variation matrices, i.e., insecticide treatment impaired environmental canalization. These results suggest differences in the processes involved in the buffering of variation within and between individuals (Nattero et al., 2021). Breno et al. (2011) compared different growth trajectories and reported a similar result, i.e., no common mechanism between canalization and developmental stability of the skull of Mastomys natalensis Smith, 1834 (Rodentia, Muridae). Moreover, other studies suggest the existence of different molecular mechanisms acting on canalization and developmental stability for different traits and influenced by different stressors (e.g., Debat et al., 2006; Takahashi et al., 2010).

Our results showed an alteration of the patterns of developmental regulators, since covariation of symmetric and asymmetric components between females and males with and without developed flight muscles revealed incongruent symmetric patterns for wing shape variation but a congruent pattern for asymmetric components. This finding may support the hypothesis that canalization and developmental stability are not necessarily related and that variation among individuals is not necessarily regulated by the same mechanisms. Debat et al. (2009) reported a similar result for D. melanogaster, and showed that the effect of temperature affected among-individual wing variation but not FA. The amount of symmetric and asymmetric variation in individuals with and without developed muscle flight varied greatly. These results suggest that variation spontaneously produced by the developmental system of the wing is higher than the variation among individuals that lack flight muscles. Moreover, shape variance was higher for females and males that lack flight muscles than for those that have flight muscles, although variance was statistically significant only for females. These results might support the idea that those individuals that lack developed flight muscles exhibited a disruption in homeostasis and an increase in wing shape variation. When environmental conditions are suboptimal, buffering mechanisms can be challenged, increasing phenotypic variance (Willmore et al., 2007). All this evidence contributes to the hypothesis that there are multiple signs of increased phenotypic variation related to the lack of flight muscle. T. infestans individuals exposed to stressful environments during developmental ontogeny showed higher phenotypic variation than those developed under more stable conditions (Nattero et al., 2021).

Visualization of wing shape patterns of variation in the comparisons of groups reveals that there is no consistency in shape variation across groups or between the patterns for symmetric and asymmetric components. Both stiff proximal and membranous distal regions of the hemelytra wing exhibited high variation. Indeed, both intra- and inter-individual shape differences include deformations of the wing areas corresponding to elongated and short wings and to thin and wide wings. To our knowledge, this is the first study that used these approaches to explore the variation of hemelytra wing in a Hemiptera. However, it is difficult to assess if this is a common feature of the wings of Hemiptera, Triatominae or T. infestans.

The lack of association between symmetric and asymmetric components, at least for those individuals without flight muscle, suggests a different pattern of morphological integration throughout the wing. This result is consistent with what is expected under the developmental hypothesis proposed, i.e., wing subdivision in proximal and distal portions, and a functional hypothesis, i.e., proximal, central and distal portions; both hypotheses are in agreement with the stiff and membranous part of the hemelytra wing. Interestingly, these hypotheses are significant only for females with and without developed flight muscles considering the variation among individuals and only for females with developed flight muscles. These modularity hypotheses were significant for FA, which implies that the pattern of covariation considering the wing subdivision for FA might derive from spontaneous variation in developmental processes (Klingenberg, 2008). The wing subdivision that showed significant covariation was tissue-bound, and coincides with two different genes that structure these portions of the wing in Hemiptera (Medved et al., 2015). The evolution of morphological structures is influenced by both the processes involved in the development of the structure and its function in the organism’s interactions with the environment (Breuker et al., 2006; Laubichler and Maienschein, 2009). The genetic and developmental structure would evolve to match functional modules (Cheverud, 1984; Wagner and Altenberg, 1996). Recent results considering the functional roles of wing structural components in dragonflies, which belong to the order Odonata, suggest the need for a balance between flexibility and stiffness in the wings (Rajabi and Gorb, 2020). The flexibility of wings should be kept within a suitable range. Our results of the morphological aspect of wings showed that females with developed flight muscles exhibited bigger wings, with an aspect more elongated and thin than that of females without flight muscles. For Triatoma guasayana Wygodzinsky and Abalos, 1949, depending on the period of the year, the individuals present elongated and thin wings, which were suggested to allow sustained long-duration flights, and short and wide wings, which would be related to short flights (Fiad et al., 2022). The study of wing modularity and morphological aspect provides coincident results in that females with developed flight muscles have certain differential characteristics in their wings that suggest an enhanced flight performance.

The origin of the insects studied in this work is heterogeneous, with adults collected in different habitats from urban and rural contexts. In the rural context, peridomestic habitats such as goat corrals, pig corrals and chicken coops usually displayed extreme and widely variable conditions of temperature, relative humidity, and insolation, which may act as environmental stressors for the developing bugs (Vázquez-Prokopec et al., 2002). Furthermore, domestic habitats had near-optimum temperatures for T. infestans and more capacity to damp variations in external conditions than other frequently infested peridomestic ecotopes (Vázquez-Prokopec et al., 2002). For T. dimidiata (Latreille, 1811), another triatomine species, the comparison of urban and rural insects showed evidence of adaptation to urban areas for morpho-functional related traits (Montes de Oca-Aguilar et al., 2022). For T. infestans the effect of habitat had been demonstrated for different traits such us wing, head and antennal phenotype (Schachter-Broide et al., 2004; Hernández et al., 2011). Body size and sexual dimorphism had also been demonstrated to decrease in domestic populations of this species (Dujardin et al., 1997, 1999); these observations suggest that peridomestic triatomine populations can be differentiated from domestic ones. Also, adults collected in distinct ecotopes differed in the amount and pattern of FA (Nattero et al., 2015). The proportion of the individuals without developed flight muscle collected in the different habitats of urban and rural context varied greatly across habitats, with goat corral and chicken coops being those that exhibited the greatest proportion of these insects (Nattero et al., unpublished data). However, despite the heterogeneity of the insect collection sites, we observed the same expression of the pattern of developmental regulators and morphological wing aspect between individuals with and without flight muscles in T. infestans.