Can body traits , other than wings , refl ect the fl ight ability of Triatominae bugs ?

Introduction: Insects of the subfamily Triatominae are vectors of Trypanosoma cruzi, the Chagas disease parasite, and their fl ying behavior has epidemiological importance. The fl ying capacity is strikingly different across and within Triatominae species, as well as between sexes or individuals. Many Triatoma infestans individuals have wings but no fl ying muscles. In other Triatominae species, no clear relationships were found between wing length and fl ying behavior. If wing presence or size is not refl ective of the fl ying behavior, which other parts of the body could be considered as reliable markers of this important function? Methods: The genus Mepraia has exceptional characteristics with invariably wingless females and wingless or winged males. We calculated the porous surface exposed to odorant molecules to estimate the olfactory capacity of Mepraia spinolai. The head shape and thorax size were estimated using the geometric morphometric approach and traditional morphometric techniques, respectively. Results: Alary polymorphism in M. spinolai was signifi cantly associated with consistent modifi cation of the thorax size, head shape, and notable change in the estimated olfactory capacity. The macropterous individuals had a larger olfactory surface and thorax size and signifi cantly different head shape compared to those of the micropterous individuals. Conclusions: We concluded that these structural changes could be associated with the fl ying potential of Triatominae. Thus, morphological attributes not found on wings could help determine the likely fl ying potential of the bugs.

Dispersal capacity by fl ight is an important feature in insects and plays a key role in evolutionary processes (1) (2) .Several insect species are polymorphic in their dispersive capacity, by having a discontinuous variation either in their wing length or the fl ight muscle mass (3) (4) (5) (6) .The loss or reduction of wings in pterygote insects is not fortuitous, but is related to environmental and geographical causes such as gender, metamorphosis types, taxa, and reproductive strategies.
Polymorphism in wing size is common among different insect taxa (1) (4) (5) (7) .In some cases, the loss or reduction of wings occurs in a single sex, mainly in females, since it allows them to divert the energy used for wings and fl ight muscles to the production of eggs, which has a high energy demand.Mate location can be an important factor for the retention of wings or fl ight capacity and is more frequent in males.The fl ightless frequency is more common in paurometabolous insects such as Triatominae (Hemiptera, Reduviidae) than hemimetabolous or holometabolous insects, since in the former, nymphs and adults occupy the same environment.Triatominae insects are vectors of Trypanosoma cruzi, the etiological agent of Chagas disease.Currently, the dispersal of these insects continues to remain an epidemiologically important issue, because of its connection with the geographical distribution of the species and the spread of the disease (8) .All adult triatomines are winged insects, except for females and some male specimens of two species, Mepraia spinolai and Mepraia gajardoi (9) .The Mepraia genus is endemic to the arid and semiarid regions of north and central Chile and is the only one within Triatominae to present a well-marked alary polymorphism.Hence, this genus could be used as a model to analyze the morphological adaptations associated with the development of wings and fl ight capacity in Triatominae.
Mepraia spinolai lives in the arid regions of north and central Chile and in some coastal islands (10) .It has been found to be associated with stony areas and resting sites of different mammals such as foxes, rabbits, hares, rodents, marsupials, and domestic animals (11) and is occasionally associated with terrestrial bromeliads (e.g., Puya sp.) (12) .Females of M. spinolai are invariably wingless, whereas males might be wingless or winged, with wings having normal length or extending up to 2mm beyond the terminalia.The wingless forms need to be METHODS strictly termed micropterous, because tiny wingpads are present both on the meso-and metathoracic segments (13) .Similarly, the two forms of winged males are more correctly described as submacropterous and macropterous because, although the long-winged macropterous forms have similar wing shape and venation to that of other species of Triatominae, the normallength submacropterous wings sometimes show slightly truncated venation of the membrane both of the meso-and metathoracic wings (11) .Campos et al. (14) detected a new wing phenotype in Mepraia with vestigial wings and shape features intermediate between that of M. spinolai and M. gajardoi, associated with a coastal area in Northern Chile.
Shortening of wings has been observed in laboratory individuals and/or geographical populations of Triatominae (15) (16) (17) (18) ; however, whether this modifi es their ability to fl y is not known.
The habitat where an insect develops is known to cause behavioral, physiological, and morphological variations that might affect the dispersal capacity and infl uence the selection of a new target habitat (19) (20) (21) .The organs where these changes have been quantitatively detected are wings, head, thorax, and antennae (22) (23) (24) (25) (26) (27) (28) (29) (30) .Geometric morphometry allows decomposition of the variation resulting from the physiology of an individual (size) from that of the genetic component (variation of shape) (31) .In addition, because triatomine insects occupy a wide variety of habitats, their sensory system displays different specializations reflected by their antennal phenotype (32) .Therefore, both the geometric morphometry of the head and the analysis of antennal phenotypes are good tools to identify individuals from environmentally different sites (26) (27) (33) (34) and are very useful for analyzing the changes related to the acquisition of wings such as in the transition from nymph to adult (26) (35) , or in the case of wingless insects, where geometric morphometry of wings cannot be used.In Mepraia spp., a marked variability was found in the antennal phenotype related to the wing condition, resulting in a greater number and increased length of olfactory sensilla on the pedicel in macropterous males than in micropterous males (36) .
Earlier observations on many Triatominae species (Catalá, unpublished) allow us to hypothesize that the head and thorax undergo signifi cant modifi cations related to changes in the wings and could be indicators of their ability to fl y.Moreover, the presence of wings might be associated with the modifi cation of other variables related to fl ight such as antennal olfactory receptors (35) .
Several studies have attempted to characterize the fl ight ability of Triatominae, but ethical issues prevent experiments from being conducted in the natural habitat where these vectors develop and live.Markers proposed to detect insects with greater fl ight ability proved of very limited applicability and did not allow comparison between species.Studies determining the maximum distance reached by fl ight in Triatoma infestans do not refl ect the actual ecological and/or physiological conditions to which the insects are subjected (37) (38) .Physiological indicators of fl ight potential, such as the weight/length index (39) (40) (41) and the qualitative and quantitative conditions of the thoracic muscles (42) (43) (44) (45) , have limited applicability since they refl ect a particular physiological state in the life history of the insect.Focusing only on the wing traits might be misleading, because old laboratory colonies of T. infestans (46) -as well as a proportion of fi eld-collected specimens (45) (47) (48) -have been shown to have wings but no thoracic muscles.
We think that other fi xed characters of adults, which are not modifi ed during the adult life, might be useful indicators for comparative studies of Triatominae species and, with the addition of particular characteristics as weight/length ratio and thorax muscle mass, could allow the determination of population differences within a species.During the fi nal molt and the acquisition of wings, morphological changes related to newly acquired reproduction and fl ight activities are known to occur.Changes in thorax, wings, head, and antennas in the nymph-adult transition as well as in insects with wing polymorphism (22) (23) (35) would provide evidence of the ability of insect fl ight.Therefore, in order to infer fl ight potential in T. cruzi insect vectors, identifying quantitative markers not only those recorded from the wings themselves is necessary.In this regard, we examined the head, thorax, and antennal olfactory sensilla and compared them according to the presence or absence of wings in M. spinolai.Such markers might be useful in predicting the fl ight ability in more domestic triatomine bugs, especially within the context of house reinfestation studies.

Insects
Sixteen micropterous males, 7 winged males (5 submacropterous and 2 macropterous), and 28 micropterous females of M. spinolai were used.A group of 4 fi fth instar nymphs of the same species was used for thorax width measurements.Insects used in this study are the same specimens used by Moreno et al. (36) .
Early analyses on M. spinolai were conducted considering the three male wing conditions-micropterous, submacropterous, and macropterous.Since submacropterous and macropterous insects did not show substantial differences in their antennal phenotype and size and shape of heads and because macropterous insects are infrequent (49) (5% of adult males), we grouped M. spinolai males into two categories: macropterous males (which included macropterous and submacropterous) and micropterous males.

Geometric morphometrics
The heads were separated at the collar and mounted on a pin attached to a metal bracket.Photographs of the ventral surface of the heads were obtained using a Kodak C613 digital camera (6.2 MP) and a stereomicroscope (10× magnifi cation).Ten type II landmarks were selected on the image of the ventral side of the adult heads (Figure 1).Average landmarks on both the sides of the head (5 landmarks) were used to reduce the intraindividual variation and digitizing errors.The shape variables were obtained using generalized procrustes analysis and the subsequent projection into a Euclidean space by computing partial warps (PWs) (50) .Partial warps describe shape variation as deviations from the average landmark confi guration.For head size comparison, the centroid size (CS) was used; it is derived from the coordinates data.It is defi ned as the square root of the sum of the squared distances between the centroid of the landmark confi guration and each individual landmark (31) .
The variance of head shape was determined using the metric disparity (MD) index.This index provides a single value measuring the variety of forms in a sample; it is based on the Euclidean distances of each form with the consensus one (51) .
Statistical comparisons for head shape, CS, and MD were made using non-parametric permutation-based tests (5,000 cycles).

Thorax width and wing length
Thorax width, defi ned as the distance between humeri of the fi fth instar nymphs and adults, and the wing length of M. spinolai adults were measured using an ocular micrometer on a Leica MZ12 stereomicroscope.For micropterous adults, the length of wing pads was considered as the basis for comparison of wing length.Since a portion of the wing chorion had been cut for previous analysis by other researchers, wing measurement was obtained from the intersection between the radial and median veins at the most distal edge of the wing.Thorax width and wing length were used as a proxy of the thoracic muscular competence.Although the thoracic cavity might be empty, it is useful as an indicator of the maximum amount of muscle mass that it might contain.

Antennal olfactory sensilla
Thin-walled trichoids (THs) are involved in the remote perception of volatiles, and thus in flight activity for the identifi cation of habitats and/or hosts.These receptors have greater length in winged male adults of Mepraia sp.than in micropterous males and females and adults of other triatomine species (36) .By using these previous data on length and number of TH sensilla, and knowing that pores are uniformly distributed over them (52) (Catalá, unpublished), we calculated the porous surface exposed to odorant molecules.For the purposes of this study, we defi ned the olfactory estimated capacity (OEC) as the number of thin-walled trichoids on antennal pedicel (PTH) × average length of TH in microns (µ) according to Moreno et al. (36) .
The OEC between insect groups was compared using analysis of variance (ANOVA).

Data collection and analysis
For the digitization of landmarks, the analysis of morphometric data and multivariate analysis with antennal variables, we used the CLIC 97 package (53) .In addition, MorphoJ 1.05f (54) was used for the analysis and graphical representation of the variations in the conformation of heads.The ANOVA statistics was performed using the Statistica package (55) .
Alary polymorphism in M. spinolai was significantly associated with consistent modifi cation in the thorax size, head shape, and a notable change in the OEC.Below, we describe the results obtained for each structure analyzed.

Head
The head shape analysis (performed using 6 principal components, representing 92% of total shape variance) showed that all 3 groups (macropterous and micropterous males, and females) had statistically different head shapes (p < 0.0001).Because of the strong sexual size dimorphism in our data, a consistent allometric effect was recorded in the shape-based discrimination of the groups (55% for the fi rst canonical factors).Micro-and macropterous males showed signifi cant (p < 0.05) allometry-free differences in head shape.The main changes affecting head shape was the difference in the outer distance between the eyes and the anteocular distance.The distance between eyes increased with the development of wings, being larger in macropterous males than in micropterous males and smaller in females.In contrast, the anteocular distance decreased in macropterous males and was signifi cantly larger in females.The anteocular distance in micropterous males was intermediate between macropterous males and females.Landmarks 3, 4, 9, and 10 (related to the distance between eyes) and 1 and 2 (related to anteocular distance) were those that showed major displacements in comparison between groups (Figure 2).
Mepraia spinolai micropterous males showed the largest MD of head shape, which was signifi cantly larger (p < 0.05) than that of macropterous males.Females showed lower MD than that in micropterous males, although not signifi cantly different (Table 1).
The CS of heads was not signifi cantly different between M. spinolai males with different wing conditions (p > 0.05).Sexual dimorphism of head size was detected; CS of females was larger than that of males (Table 1).

Thorax and wing length
The fi fth instar nymphs showed lower humeral distance than adults (females and both male phenotypes).The thorax width was larger (p < 0.05) in macropterous than in micropterous males.The thorax width was signifi cantly lower in females  than in males.The thorax width (humeral distance) increases with wing length (r = 0.82, p < 0.05; Figure 3).The wing pad length of females and micropterous males did not differ, although it was different from the wing length of macropterous males (Table 1).

Antennal olfactory sensilla
The OEC of the TH sensilla of the pedicel increased by 92.6% in macropterous males of M. spinolai than in micropterous males.Sexual dimorphism of this variable was observed, confi rming that, even in micropterous males, OEC is greater than that in females (p < 0.05; Figure 4).

DISCUSSION
In Mepraia spinolai, the alary polymorphism was associated with other body attributes: I) the OEC, II) thorax size, and III) head shape.We believe that these changes are morphological attributes probably related to the fl ying potential of Triatominae.Winged aphids (Hemiptera) exhibit stronger sclerotization of the head and thorax, further development of the eyes and ocelli, longer antennae, and more rhinaria, which have an olfactory function (56) (57) (58) .They are also equipped with an elaborated sensory system to search for host plants and show greater resistance to fasting than their wingless counterparts (59) (60) .
The ability to fl y has obvious advantages in insects that need to disperse rapidly to remote distances, although the presence of wings does not guarantee fl ight.The dispersal behavior of insects is known to depend on multiple factors (local landscape conditions, intraspecifi c competition by limited food resources, mating, etc.) and can be highly variable even among individuals of the same species (61) (62) .Larger wing length does not always suggest greater fl ight ability.et al. (63) found that the hybrid between Triatoma sherlocki (short wings, no fl iers) and Triatoma juazeirensis (wings that reach the seventh or eighth urostergite, fl iers) showed more fl ight ability than T. juazeirensis despite having intermediate size wings.Furthermore, the presence of wings, as shown for T. infestans, does not necessarily account to fl ying capacity (45) .Nonetheless, this does not discard the wing metrics as a key element in fl ying performances (63) .
Habitat quality is one of the local conditions that might directly affect dispersal, by modifying the physiological Olfactory capacity in microns (υ) condition of individuals or by producing changes in their phenotype (19) .The habitat dimensionality hypothesis states that habitats distributed at ground level can be reached by walking insects, whereas dispersal is more diffi cult without fl ight in three-dimensional habitats (with trees) (64) .The desert-like habitat of M. spinolai is mainly two-dimensional, and alary polymorphism is expected to occur.In this species, micropterous and macropterous individuals are mainly observed, with females always being micropterous allowing them to spend their energy mainly in reproduction (65) .Males could be either unwinged or winged; the latter disperse over longer distances (66) .Although fl ying M. spinolai are not usually captured, there are fl ight records under natural fi eld conditions (67) ; therefore, changes associated with its wing phenotype would be related to dispersal by fl ight.
The important role of antennal olfactory receptors during dispersal is well known.Although Moreno et al. (36) have already observed and quantified intraspecific variation in the number and length of TH, the combination of these two variables to estimate olfactory capacity allows a more accurate assessment of the role played by these sensory receptors and would allow the comparison among species.The OEC showed remarkable intraspecifi c differences between micropterous and macropterous individuals, even within the same sex, suggesting that this variation is not due to sexual dimorphism, but because of the ability to fl y.Regarding the dispersal ability, multiporous sensilla would be particularly important.These olfactory receptors are capable of capturing odor molecules from considerable distances (68) .In bloodsucking insects, a relationship was found between fl ight capacity and olfactory sensilla.Antennas lice, fl eas, and bedbugs contain olfactory receptors, unlike mosquitoes and other Diptera species that have signifi cant ability to fl y (69) (70) .
The sexual dimorphism in the OEC, even between micropterous males and females, indicates that TH receptors are also involved in the search of mates.The increased number of TH sensilla in males has been frequently observed in other species of Triatominae (25) (26) (28) (32) (71) .Insects use various strategies to expand their olfactory surface in order to increase contact with volatiles used for communication and orientation.In some hemimetabolous insects such as Ephemeroptera, Plecoptera, and Blattodea, the length of the antennae is increased to accommodate a greater number of olfactory sensilla (72) .In holometabolous insects, antennae also increase in size, but their shape is more sophisticated, allowing a greater exposure surface, such as feathery antennae in mosquitoes (73) and lepidopterans (74) .Another strategy to extend the olfactory surface is to increase the distal area of the antenna, such as in the case of capitate antennae of some hymenopterans and coleopterans (75) .In Triatominae, no direct relationship has been found between the antenna length and sensilla number (76) , except during development.However, Catalá (32) found that triatomines with a broad habitat range had more chemo-sensilla on the pedicel than species with a more restrictive habitat range.In Mepraia spp., the PTH length in macropterous males is the highest recorded for Triatominae (36) .Our results suggested that greater length of TH indicated greater olfactory surface, with a larger porous surface for exposure to compounds involved in chemical communication.
Our morphometric analyses indicated that the size, but not the shape, of the head in males did not vary with the presence/ absence of wings.Macropterous males showed less variability of head shape than micropterous males (lower metric disparity), which could suggest the importance of this variable in the fl ight processes.The head of an insect contains most of the sense organs, which allows it to interact with the environment; therefore, variability in head shape could be of great relevance in all activities, including fl ight.According to our comparisons, the ability to fl y would be associated with a tighter shape of the head.
Both simple and compound eyes are known to infl uence the fl ight activities of insects in general (56) (77) (78) (79) (80) , and specifi cally in Triatominae (81) .The sensory information input in insects can occur in a composite and multimodal manner, combining the information from the eyes, ocelli, antennae, and cephalic sensilla, which translates into a composite signal used to control fl ight (82) .Compound eyes alone (that is, the ablation of ocelli) are capable of mediating visual stability of the head and the directionality behavior of fl ight under a wide range of conditions.Ocelli act synergistically with the eyes and increase the responses related to fl ight control (83) .Our results showed that the association between fl ight and compound eyes is due to the further development of the anterior portion of the eyes in macropterous individuals, which probably intervene in fl ight directionality (Figure 2).
The morphological changes found in M. spinolai could be used to determine differences in fl ight potential through morphological variables in other Triatominae species or subpopulations within the same species.In this study, only few insects available in the entomological collection were used for analysis.Although some samples might seem small, they can serve as a good indicator of the trends observed in the morphotypes.When analyzing natural populations (not reared in the laboratory), the sample size can be low even with high sampling effort.Schofi eld et al. (84) collected very few adult M. spinolai after six years of sampling.Moreover, in studies of reinfestant Triatominae, low number of insects collected after an insecticide spray can also raise doubts about the technique used (27) (85) .The results of these studies can provide evidence for further long-term studies, which might increase the number of insects collected and enhance the power of statistical tests.
For T. infestans, the main vector of T. cruzi in Latin America, sexual differences have been found in the fl ight dispersal activity.In La Rioja (Argentina), many male T. infestans showed differential allocation of food and sexual dimorphism in dispersion (86) .
The thorax of M. spinolai also showed an increase in size according to the development of the wings.These results are consistent with those reported for T. sherlocki, where its reduced wing size corresponded to less developed fl ight muscles and thorax size compared to that in T. juazeirensis, a closely related fl ying species (63) .Similarly, Hoffmann et al. (24) found that, in Drosophila melanogaster, wing length and thorax size changed according to fl ight capacity.
Our results showed that macropterous M. spinolai have a larger olfactory surface, greater eye size, smaller anteocular distance of the head, and larger thorax size compared to those of micropterous specimens.The studied phenetic characteristics can serve as variables to determine differences in fl ight potential in this as well as other insect groups.Furthermore, they allow elucidation of the behavioral aspects related to insect dispersal and determination of how morphology accompanies behavioral alterations with changes in accordance to the requirements of insects in a particular environment.The identifi cation of morphological characteristics associated with the presence/ absence of wings in insects is of major importance since it allows the determination of their dispersal potential.This is very useful, particularly in the case of insect disease vectors or crop pests, where the identifi cation of dispersant morphotypes contributes to their effi cient control.

6 FIGURE 1 -
FIGURE 1 -Points of reference (landmarks) measured as coordinates of the heads in Mepraia spinolai (Porter).The numbers indicate the order of landmark capture.