Immature Instars of Three Species of Rhodnius (Hemiptera, Reduviidae, Triatominae): Morphological and Morphometric Studies

Gustavo Lázari Cacini (  cacini.gustavo@gmail.com ) Universidade Estadual Paulista (Unesp) Jader de Oliveira Universidade de São Paulo (Usp) Tiago Belintani Universidade Estadual de Campinas (Unicamp) Éder dos Santos Souza Fundação de Medicina Tropical (FMT-HVD) Nicoly Olaia Universidade Estadual Paulista (Unesp) Mara Cristina Pinto Universidade Estadual Paulista (Unesp) João Aristeu da Rosa Universidade Estadual Paulista (Unesp)


Morphological study
To generate images, ve specimens of each of the ve instars of R. marabaensis, R. prolixus, and R. robustus were used. Images of the dorsal sides of head, thorax, and abdomen as well as the complete images of each of the ve instars from dorsal and ventral view were obtained using a Leica M205 stereoscopic microscope and the Leica Application Suite X software.

Morphometric study
Fifteen specimens of 1st, 2nd, 3rd, 4th, and 5th instar nymphs of R. marabaensis, R. prolixus, and R. robustus were measured using a Leica MZ APO stereoscopic microscope and the Motic Advanced 3.2 plus image analysis system. Nymphs of all instars had their total length (TL), head length (HL), thorax length (XL) and abdomen length (AL) measured.
Following Dujardin et al. [27], interocular (IO), anteocular (AO) and postocular (PO) distances were measured, as well as the three segments of the proboscis. The four antennal segments were also measured, according to Rosa et al. [28]. The length of all the parameters was measured in millimeters.
The obtained data were analyzed by descriptive statistics, using t-test for mean, standard deviation. An ANOVA and Tukey´s pairwise to evaluate the degree of differentiation of the three species using the Past software.
Geometric morphometrics of heads Geometric morphometrics was used to evaluate variations in head shape and size using Cartesian reference coordinates.
Variations among the heads of all nymphal instars of the studied species were evaluated. Fifteen heads of each instar were selected, and the images were obtained by means of a stereoscopic magnifying glass coupled to the Motic Advanced 3.2 plus scanning system. The coordinates of the reference points were selected according to Bookstein [29]. Ten anatomical landmarks of Type 1 (for example, intersection between veins) were collected and processed using the modules available in the tps Dig v.1.18 software [30]. Four anatomical landmarks were digitized for the 1st instar of development and ve anatomical landmarks were selected for the other instars of development; the landmarks being digitized using the CLIC package (https://xyom-clic.eu/the-clic-package/). Then the le with the raw coordinates was used for a generalized Procrustes analysis (GPA). GPA is a method that allows eliminating all the information related to size, position, and orientation of previously digitized anatomical frames [30]. The matrix of form was held in a Euclidean space to generate a set of marks known as partial warps [29]. All the additional statistical forms were performed using Procrustes residues to analyze differences in the size and shape of the heads of each nymphal instar. Procrustes ANOVA (p<0.0001) [31] is used to infer differences between species. Procrustes ANOVA is a method for quantifying relative amounts of variation at different levels.
These differences in size were assessed using an isometric estimator de-ned as centroid size (CS) [32]. Mahalanobis distances between pairs of species were calculated for measurements of shape and signi cance was assessed using a nonparametric test based on permutations (bootstrap, 10000 replications) using MorphoJ [33]. In addition to that, using distance dice from Mahalanobis, neighbor-joining trees (NJ) were recovered using PAST v.3.25 [34]. To determine the relationships between species, canonical variable analysis (CVA) was performed using MorphoJ [33]. The CVA was performed associated with a resampling method (bootstrap, 10000 replications) to build regions of trust in relation to the median size of the species centre. A factorial map of the rst two canonical fathers was created using MorphoJ, version 1.0.7a [33].

Results
Morphological description of the ve nymphal instars of R. marabaensis, R. prolixus, and R. robustus by optical microscopy First instar: the head of the nymphs has a dark-brown cuticle covering all its granular extension due to the presence of tubercles with small sensilla, whose colour is darker than that of the cuticle. Gena and juga showed no signi cant differences among the species. Regarding the postocular area, the species shows Y-shaped cephalic sutures ( Figure 6A-C). On the thorax there are tubercles with sensilla in the three segments, located mainly in the centre. The prothorax has a trapezoidal shape and is the segment with larger external borders, followed by the metathorax and the mesothorax. The three segments are well delimited by the dividing lines, but the line separating the mesothorax from the metathorax shows a sinuous protuberance that overlaps the metathorax in about one-third of its size. The metathorax is broad on the sides and narrow in the central portion ( Figure 6D-F), while, near segment I of the abdomen of R. robustus, there is a marked convexity that is not seen in R. marabaensis and R. prolixus. The abdomen of 1st instar nymphs has a lighter colour in comparison with the thorax and the head. There are many tubercles with sensilla lighter-coloured than the cuticle. A lighter median longitudinal strip can be seen all over the abdomen. Darker spots are noticed along the connexivum ( Figure 6G-I). The abdomen of the three species is different in shape: R. prolixus has a larger abdomen than R. marabaensis and R. robustus, particularly between segments III and VI. The abdomen of R. robustus widens gradually from segment I to IV, which is broader, then it starts to narrow, also gradually. In R. marabaensis segments IV and V are the broadest of the abdomen. In R. marabaensis and R. prolixus segment I of the abdomen has the same length as the metanotum, whereas in R. robustus it is longer than the metanotum ( Figure 6G-I).
Second instar: the general aspects of the head of 2nd instar nymphs follow the pattern described for 1st instar nymphs. However, some differences are noticeable, such as the increase in the granulation grade and the number of sensilla in the three species, as well as the lighter colour of the cuticle ( Figure 7A-C). Tubercles with sensilla are present in the three segments of the thorax, located mainly in the central portion. The prothorax has the form of a trapezium and is the segment with the largest external borders, and it is longer in R. prolixus than in R. marabaensis and R. robustus. It is not possible to identify the difference in size between the mesothorax and the metathorax, but, as in the 1st instar, the metathorax is broad on the sides and narrow in the central portion. The three segments are well delimited by the dividing lines, but the line separating the mesothorax from the metathorax has a sinuous protuberance that overlaps the metathorax in about one-third of its size. At the central limit between the mesonotum and the metanotum, the dividing line is straight in R. marabaensis and R. robustus, whereas in R. prolixus it is concave ( Figure 7D-F). The abdomen has a median spot lighter than the cuticle in the dividing line of each of the urotergites, resembling a strip. The connexivum spots become more evident in this instar. The abdomen of R. prolixus has the maximum width in segment V. The external limits of the abdomen of R. marabaensis between segments III and VI are straight. The abdomen of R. robustus is like the 1st instar, i.e., it widens gradually from segment I to IV and from this segment on it narrows gradually ( Figure 7G-I).
Third instar: for the three species, in the third instar gena are more rounded and extend until the end of the clypeus. Postocular cephalic sutures are also more rounded and roughly have a U-shape ( Figure 8A-C). From dorsal view the head of R. marabaensis shows a prominent white strip that starts at the antennal tubercle and extends until the neck ( Figure 8A). In R. prolixus this strip is not so evident, and in R. robustus it is visible between the clypeus and the antennal tubercles; posteriorly it is visible in the intermediate portion and between the posterior region of the eye and the neck ( Figure 8B-C). Tubercles with sensilla are present in the three segments of the thorax, distributed across them. The prothorax has the shape of a trapezium and is the segment with the largest external borders, followed by the mesothorax and the metathorax. The three segments are well delimited by the dividing lines, but the line separating the mesothorax from the metathorax has a sinuous protuberance that overlaps the metathorax in about one third of its size. The mesonotum of R. prolixus shows a concavity in the central portion, near the metanotum, whereas this concavity is not seen in R. marabaensis and R. robustus ( Figure 8D-F). The three species possesses 2+2 dark strips across the abdomen. An increase of spots in the connexivum can also be observed. In R. marabaensis abdominal segments III and IV are the broadest. The segments widen from segment I to III and gradually narrow from segment VI onwards. A central strip of straw colour is also easily visible on the abdomen of this species, and from the sides there is another straw strip, located between two black strips. These three strips are arched and have the same shape as the abdomen, but they are not continuous, as they are interrupted in the intersegmental sutures. In R. prolixus and R. robustus the largest abdominal segment is the IV. The central strip that travels across the abdomen also distinguishes the three species in this instar: in R. marabaensis this strip has straw-colored marks elongated in each segment; in R. prolixus such marks are neither so elongated nor clear, whereas in R. robustus they are not elongated, yet clear ( Figure 8G-I).
Fourth instar: 4th instar nymphs present some peculiar characteristics, such as triangular juga, rounded gena surpassing the clypeus, and a higher granulation grade near the eyes. As in the third instar, the white strip on the head is very clear in R. marabaensis. The strip is present in R. prolixus, but it is narrower, and in R. robustus it is less clear in the intermediate portion ( Figure 9A-C). The three segments of the thorax have tubercles with sensilla. The mesothorax is the largest segment in this instar due to the presence of the rst pair of wing pads. The second pair of wing pads originates from the metathorax. The three segments are well delimited by the dividing lines. The line separating the prothorax from the mesothorax has a slight sinuosity overlapping a tiny part of the mesothorax in R. marabaensis and R. prolixus. In R. robustus the line separating the mesothorax from the metathorax has a concave aspect, the same for the line between the metathorax and the beginning of the abdomen ( Figure 9D-F). In all three species the dark strips on the abdomen are more evident, which gives the area a striped aspect. In this instar the connexivum spots become more rounded. The central strip on the abdomen has the same aspect as in the 3rd instar and differentiates the three species: in R. marabaensis the three side strips on the abdomen, a straw strip between two black ones to the right and left, are similar to what is observed in the 3rd instar ( Figure 9G-I).
Fifth instar: in this instar all three species also have a quite visible white strip on the head. The gena of R. marabaensis reach the initial limit of the clypeus, which is straight; the anteclypeus is curve ( Figure 10A). In R. prolixus the gena surpass the clypeus, which is curve at the beginning, and the anteclypeus has a trapezoidal shape ( Figure 10B). The gena of R. robustus reach the limit of the clypeus, which initially has a concave form, whereas the anteclypeus has a semicircular shape ( Figure   10C). There are tubercles with sensilla in the three segments of the thorax. The posterior pair of wing pads can be seen overlapping, projecting from the mesonotum through the anterior pair, which in turn projects from the mesothorax ( Figure   10D-F). It is possible to see only the central area of the metathorax as a result of this large overlapping. The anterior pair of wing pads reach the beginning of the 3rd urotergite. The side limit of the anterior wing pads is determined by a clear line in R. prolixus and R. robustus, while in R. marabaensis the line is broad and diffuse ( Figure 10D-F). The abdomen of R. prolixus and R. robustus is more elongated than that of R. marabaensis. In this instar the central line of the abdomen retains the characteristics observed in the 3rd and 4th instars. In R. marabaensis the three side strips on the right and left (one straw and two black) noticed in the 3rd and 4th instars are still present. There is also an increase in the number of tubercles with sensilla ( Figure 10G-I).
Morphometric study of the ve nymphal instars of R. marabaensis, R. prolixus, and R. robustus With the acquired data it was possible to calculate the mean for each parameter and species, and then compare them to evaluate the degree in which the three Rhodnius species differ.
In the rst and second instars, none of the parameters were statistically signi cant to evaluate the degree of the differences among the three Rhodnius species. As for the third instar, the parameter of the 3rd segment of the antenna (F (2,42) 23.12, p= 1.693) was signi cant ( Table 1). The 4th instar only the postocular distance stands out (F (2,42) 13.64, p= 2.718) ( Table 2).
Lastly, on the 5th instar just the 2nd segment of the antenna (F (2,42) 36.32, p= 6.965) ( Table 1) made it possible to evaluate the degree of the difference between R. marabaensis, R. prolixus and R. robustus. ***Insert tables here**  Geometric morphometrics of the ve nymphal instars of R. marabaensis, R. prolixus, and R. robustus By ontogenetic geometric morphometry of the heads of nymphs, it was possible to describe the differences in shape and size of the ve instars of R. robustus, R. prolixus, and R. marabaensis. Centroid size (CS) measures show variability in head size of the species. Furthermore, by the isometric measurement of the centroid size, the size gain among immature shapes can be clearly seen (Figure 11). Analysis of centroid size shows that differences among the size means are signi cant (p<0.0001, supplementary material). Rhodnius robustus and R. prolixus have larger size means compared to R. marabaensis ( Figure 11). Differences can also be explained as a percentage of the total variance among groups in the Eigenvalues (auto values), the percentages being: 89% for the 1st instar, 83% for the 2nd, 98% for the 3rd, 93% for the 4th, and 90% for the 5th. Mahalanobis distance was used as a metric estimator. The estimator considers the variations and correlations among groups de ned a priori and allows pairwise comparison. Mahalanobis distances were signi cant among the pairs of the assessed species (p<0.001, supplementary material). Dendrograms were built based on the values recovered for Mahalanobis distances and Neighbor joining (NJ). The topology is identical for all instars (Figure 12). It was possible to delimit the proximity between R. prolixus and R. marabaensis ( Figure 12). Procrustes ANOVA test also recovers signi cant values, showing shape differences among the species (p<0.0001, supplementary material).
The projection of the three species in the space de ned by canonical axes 1 (CVA1) and 2 (CVA2) provides a description among the speci ed groups in the set of multivariate data. The analyses of the canonical variables resulted in 10 variables and explain 100% of the discrimination among the species (Figure 13). The rst two variables (CVA1 and CVA2) generated the following percentages: 85.2% and 22.49% for the 1st instar; 47.76% and 21.48% for the 2nd; 97.1% and 3% for the 3rd; 92.8% and 3% for the 4th; 85.2% and 22.49% for the 5th (Figure 13). The grouping in the space of the canonical axis shows an overlapping relationship between R. prolixus, R. robustus and R. marabaensis in the 1st and 4th instar, however the separation of populations in the 2nd, 3rd and 5th stages is clear. R. marabaensis is the species that was best separated in the CVA analysis.

Discussion
A striking feature of Triatominae is that males, females, and nymphs of all instars can transmit T. cruzi if infected [35,36]. Therefore, studies about nymphal instar have not only taxonomic and phylogenetic interest but also epidemiological importance. Speci cally, about the genus Rhodnius, the following works can be mentioned: Mascarenhas [37], which studied the ve instars of R. Morphological characters are useful tools for taxonomic and systematic studies in Triatominae, in addition to being useful for epidemiological surveillance. The morphology analyses show the separation of the three species by characters was observed at head, thorax, and abdomen shape. This made it possible to separate them in all ve nymphal instars and characterized for the rst time the development stages of R. marabaensis. In the chapter on the nymphal instars, Lent & Wygodzinsky [2] mentioned that R. prolixus and R. robustus do not have sub median tubercles or aggregations of granules along midline, but such characters were noticed in all ve nymphal instars of those species, as well as in R. marabaensis. Rosa et al. [36], studying 1st and 2nd instar nymphs of Triatoma arthurneivai Lent & Martins, 1940, distinguished the two instars by morphological characters of the thorax. Thus, by scanning electronic microscopy, they noticed the absence of collar, glabrous areas, and tubercles in the 1st instar of T. arthurneivai, which were present in the 2nd instar. Nevertheless, the differentiation among 3rd, 4th, and 5th instars of R. marabaensis, R. prolixus, and R. robustus was made using the same characters observed by Rosa et al. [36] in nymphs of the previously mentioned instars of T. arthurneivai, i.e., the formation and conformation of the two pairs of wing pads located on the thorax.
In this study, the results of the morphometry of characters from the abdomen, antenna, head, proboscides, and thorax show little discrimination between the three species. In general, the compared averages are little or no signi cant, the morphometric study is not suitable for identi cation. However, R. marabaensis had its nymphs characterized morphometrically and morphologically for the rst time.
The relative length of the four antennal segments in R. marabaensis shows the same pattern for the rst three instars, another for the 4th instar, and a third pattern for the 5th instar, whereas R. prolixus and R. robustus show the same pattern for the 1st and 2nd instars, another for the 3rd and 4th instar, and a third one for the 5th instar. Santos [40], measuring R. colombiensis, R. ecuadoriensis and R. stali, found two patterns of relative length for antennal segments of the ve nymphal instars. For R. milesi the author found three patterns: one for the 1st and 2nd instars, another for the 4th and 5th instars, and a third one for the 3rd instar, hence different patterns from the ones observed in R. marabaensis, R. prolixus and R. robustus.
Rosa et al. [41] carried out a morphometric study of the four antennal segments of nymphs of the ve instars and adults of Panstrongylus megistus (Burmeister, 1835), R. neglectus, R. prolixus and Triatoma vitticeps Stäl, 1859. The patterns identi ed in R. neglectus and R. prolixus were the same found for R. prolixus and R. robustus in this work. Rosa et al.
Geometric morphometry allows evaluating the variation of shape in relation to causal effects [42]. The technique allows us to quantify biological forms and discuss the evolution of phenetic patterns [33]. The technique is used in paleontological, anthropological, ecological, zoological, and botanical studies [29,33]. In triatomines, geometric morphometry is used to assess the shape and size variables of hemelytra [43,44], heads [12,45], and eggs [46]. Also, for ontogenetic studies [47,48,49].
Recently two subcomplexes of the genus Triatoma were studied by means of geometric morphometrics, which indicated the potential of the technique to study specimens that are phylogenetically close [43,45]. Geometric morphometrics allowed describing the differences in head shape and size of the ve nymphal instars. In relation to the CS, all the values obtained were signi cant and allowed to differentiate the three species in the ve nymphal instars. A variation among the instars is noticed but considering the general aspect R. robustus is easily characterized by the geometric pro le of the heads of nymphs. The 2nd and 4th instar showed less discrimination potential, i.e., only approximated size means were recovered.
The metric estimator of Mahalanobis distance was used to recover NJ dendrograms, where it is possible to visualize that in all evaluated instars R. robustus is distant, whilst R. prolixus and R. marabaensis form a single clade. However, CVA ellipses show that in the 1st and 2nd instars R. marabaensis and R. robustus remain close, while groups are clearly separated in the 3rd, 4th, and 5th instars. Regarding the shape, the values of the Procrustes ANOVA test reveal differences among the cephalic capsules, allowing discrimination. We show that the multivariate morphometric technique is more e cient to discriminate against the studied species when confronted with linear morphometric data.

Conclusion
In this study, the morphological and morphometric differences of three Rhodnius species were evaluated. It was also provided new data for R. marabaensis. Furthermore, was show that the morphology of the head (3rd, 4th, and 5th), thorax (2nd and 5th instar) and abdomen (1st, 2nd, 3rd, and 5th instar) are useful to discriminate the studied species. Through morphometric analysis of the head, it was veri ed that the postocular distance of the 4th instar and the lengths of the antennal segments of the 3rd and 5th instars distinguish the three species. Lastly, geometric morphometry proves to be useful for these species.
The size and shape variables clearly show the differences between R. marabaensis, R. prolixus, and R. robustus.

Declarations Acknowledgments
The authors would like to thank the collaborators of the

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.