Movement and olfactory signals: Sexually dimorphic antennae and female flightlessness in moths

Johnson, Tamara L.; Elgar, Mark A.; Symonds, Matthew R. E.

doi:10.3389/fevo.2022.919093

ORIGINAL RESEARCH article

Front. Ecol. Evol., 26 August 2022

Sec. Behavioral and Evolutionary Ecology

Volume 10 - 2022 | https://doi.org/10.3389/fevo.2022.919093

This article is part of the Research Topic Signals in Motion View all 10 articles

Movement and olfactory signals: Sexually dimorphic antennae and female flightlessness in moths

$\r\nTamara L. Johnson$ Tamara L. Johnson¹

Mark A. Elgar^1*

Matthew R. E. Symonds^2*

¹School of Biosciences, The University of Melbourne, Melbourne, VIC, Australia
²Centre for Integrative Ecology, School of Life and Environmental Sciences, Deakin University, Burwood, VIC, Australia

Darwin argued a role for sexual selection in the evolution of male sensory structures, including insect antennae, the strength of which will depend upon the importance of early arrival at receptive females. There is remarkable variation in the nature and degree of sexual dimorphism in moth antennae, with males of some species having spectacular, feathery antennae. Although it is widely assumed that these elaborate structures provide greater sensitivity to chemical signals (sex pheromones), the factors underlying the interspecific diversity in male antennal structure and size are poorly understood. Because male antennal morphology may be affected by several female life–history traits, including flight ability, we conducted a phylogenetic comparative analysis to test how these traits are linked, using data from 93 species of moths across 11 superfamilies. Our results reveal that elaborate antennae in males have evolved more frequently in species where females are monandrous. Further, female loss of flight ability evolved more frequently in species where males have elaborate antennae. These results suggest that elaborate antennae have evolved in response to more intense male competition, arising from female monandry, and that the evolution of elaborate antennae in males has, in turn, shaped the evolution of female flightlessness.

Introduction

Signalling is a crucial component of reproduction for mobile diecious species, playing a role in both bringing mates together and in facilitating mechanisms of pre–mating sexual selection (e.g., Darwin, 1871; Andersson, 1994; Rosenthal, 2017). The latter is responsible for the evolution of an extraordinary diversity of conspicuous, sexually selected signals, across the range of sensory modalities, and has attracted very extensive research interest. In many moths, females use sex pheromones (olfactory signals) to advertise their location to mate searching males. Although long–distance location–revealing sex pheromones are typically not regarded as being subject to sexual selection [see Johansson and Jones (2007)], they may cause sexual selection to act on male receptor organs (Darwin, 1871; Elgar et al., 2019).

Darwin (1871) suggested that sexual selection will favour improvements in “organs of sense” if that improves the likelihood of male mating success in a competitive environment. Darwin did not explicitly mention antennae as “organs of sense,” because pheromones and the odour receptors located on antennal sensilla (e.g., Hansson, 1995) were not known at that time (Elgar et al., 2019). Nevertheless, his perspective suggests that selection will favour males with antennal features that improve the speed of detection of sex pheromones if that allows males to locate females more quickly than rival males. The taxonomically widespread sexual dimorphism in insect antennal morphology (e.g., Schneider, 1964; Elgar et al., 2018), together with several lines of empirical evidence, are consistent with this perspective (Elgar et al., 2019). Field experiments on a sexually dimorphic moth, the gum-leaf skeletonizer Uraba lugens, demonstrated that males with longer antennae were more likely to detect lower amounts of pheromone (Johnson et al., 2017b), and laboratory experiments with the same species revealed that larvae developing in higher densities (thereby anticipating greater reproductive competition as adults) resulted in males with larger antennae and testes (Johnson et al., 2017a), and in females releasing more attractive sex pheromone (Pham et al., 2020).

Sexual dimorphism in some species of moths is remarkably striking, with the elaborate, feathery, pennate antennae of males contrasting with the simple threadlike or filiform antennae of females (Schneider, 1964; Young, 1997). Phylogenetic comparative analyses indicate that while elaborate antennae are linked with larger body size, suggesting a functional cost to these structures, this pattern is not necessarily consistent with a greater capacity to detect chemical signals, because larger females might be expected to release larger quantities of pheromone (Symonds et al., 2012). On the other hand, in species where males have elaborate antennae, there is a negative correlation between male abundance and male antennal length. This result suggests that lower population densities with concomitantly lower concentrations of pheromone, may select for larger antennae in males, at least in species where males have elaborate antennae (Symonds et al., 2012). Regardless, males of most moth species possess relatively simple filiform antennae, whilst females still emit long–distance sex pheromones. This suggests that the strength of selection on male antennal morphology is linked to other factors that determine the importance of quickly locating a female, and thus rapidly detecting her sex pheromone. A previous comparative analysis of male elaborate antennae in geometrid moths (Javoiš et al., 2019) found that they were more likely to be found in species where females were capital breeders (i.e., that eclose with greater body reserves already available for breeding). Javoiš et al. (2019) suggested that such a strategy may be associated with traits that make them more difficult for males to locate quickly, such as reduced mobility.

There is emerging interest in the effects of movement on the production and detection of signals and cues, although research is largely confined to visual and auditory sensory modalities (Tan and Elgar, 2021). Nevertheless, movement may be consequential for chemical signalling: female moths may adjust the detectability of their sex pheromones by selecting different kinds of locations where they call (emit pheromones). For example, pheromones released in closed habitats may be less easily detected, and females may compensate by moving to a more open location (Murlis et al., 1992). Clearly this option is not available to less mobile, flightless female moths. Females are flightless in roughly 1% of lepidopteran species, although it is taxonomically widespread, occurring in 25 families (Sattler, 1991). Female wing loss in these species varies from a complete loss of wings (aptery) to retaining full sized wings but with a loss of function (Tweedie, 1976; Sattler, 1991). Flightlessness is associated with winter–active adults and spring feeding, as well as high host breadth (Hunter, 1995). In most flightless species, the females are unable to move far and remain on or near their pupation site throughout their typically short adult life (Sattler, 1991). Hackman (1966) noted that these locations may not necessarily be optimal for calling and suggested that this selects for greater sensitivity to the pheromone in males. The converse may also be possible: that loss of flight evolved in part due to the speed with which highly sensitive males can locate females.

Here we use a phylogenetic comparative approach to examine several potential selection pressures favouring male antennal complexity in moths, including flightlessness and other life–history traits. Specifically, we ask whether certain female traits that potentially reduce their reproductive window (i.e., monandry, short lifespan, and the stage of egg development) are associated with the evolution of elaborate male antennae. Our hypothetical framework is that rapid location of a female would be favoured under female monandry, or when females eclose with a full complement of eggs (proovigeny): slower males may arrive after a female has mated and is no longer receptive, or after she has commenced ovipositing, and thus their sperm will fertilise fewer eggs (Jervis et al., 2005; Shuker, 2014). We hypothesise that these female life–history traits would likely increase the level of competition among males and increase the strength of selection favouring elaborate antennae, because such antennae should increase the male’s likelihood of locating the female more quickly.

Subsequently, we explore the links between male antennal morphology, female life-history and the evolution of female flightlessness. We specifically test the hypothesis proposed by Hackman (1966) that female flightlessness will be associated with the need for greater male sensitivity to sex pheromone, and hence with male elaborate antennae.

Materials and methods

Data collection

We collected information on female mating strategy, male antenna type, female flight ability, egg development, oviposition behaviour, lifespan, and fecundity for 93 species of moths from 11 superfamilies. The species were selected based on the availability of these data, which were collated from various sources including published literature, field guides and online lepidopterist resources (see Table 1).

TABLE 1

Table 1. Classification of species for traits examined.

We characterised male antennae as either simple or elaborate: the former is filiform or ciliate, whereas elaborate antennae have one or more side branches (pectinate, bipectinate, quadripectinate). Information on male antennae was obtained from published descriptions or images from various resources (see Table 1 for a full list). Females were designated as capable of flying (macropterous) or flightless: flightless females may still have wings, but do not fly (e.g., Lymantria dispar) or may be brachypterous (small remnant wings) or apterous (wingless).

Mating strategies of females were classified as either monandrous or polyandrous, using information on female remating frequencies (mating frequency of males was not taken into account, as less information is available on this). Monandrous species are those with a remating frequency of 30% or less. This is a conservative value compared with rates previously used to categorise monandry [for example, <50% in Arnqvist et al. (2000), and <40% in Torres–Vila et al. (2002)]. Where detailed data on remating frequency were not available, we followed published qualitative descriptions of species as being either monandrous (including mostly monandrous) or polyandrous.

Females were classified as either synovigenic (continuing to form eggs during their adult life) or proovigenic (eclosing as an adult with a full complement of mature eggs). Species where females eclose with some mature eggs but produce more as an adult were classified as synovigenic. We also distinguished between females that oviposit eggs singly, in multiple small clutches of eggs, or in a single clutch. We then made this classification binary by combining the data for species in the latter two categories who did not lay eggs singly, as this was necessary for statistical analysis (see below). The fecundity and lifespan of the females were obtained from published papers, and we included the midpoint if a range was given. Fecundity was the total reported number of eggs laid over the lifespan. Female lifespan was the average number of days as an adult, as observed in natural field populations where possible. Finally, as body size may affect aspects of mating behaviour (Blanckenhorn, 2000) and male antennal morphology (Symonds et al., 2012), we also obtained measures of the maximum wingspan to use as a proxy for body size (Miller, 1977, 1997). We included male wingspan if the wingspan of both sexes were reported, and the midpoint of the range if a range was reported.

Phylogenetic comparative analysis

We used phylogenetic comparative methods to control for common ancestry (Harvey and Pagel, 1991) in our analysis of evolutionary associations between traits. There is no single phylogeny that incorporates all the species in our sample, and genetic data covering these species is sufficiently poor to make phylogenetic estimation unreliable. Following recommendations by Beaulieu et al. (2012), we compiled a composite phylogeny from the published phylogenies of the families included in our sample. The full tree, along with character mapping for binary traits, is presented in Figure 1.

FIGURE 1

Figure 1. Composite phylogeny of the moth species used in the analysis, with character states of categorical traits indicated (missing blocks indicate the character state was unknown). Branch lengths are for illustrative purposes only, and do not represent the lengths used for analysis.

The main framework for the phylogeny was the superfamilial tree constructed by Heikkilä et al. (2015). The superfamilies Hepialoidea, Sesoidea, and Zygaenoidea each contained two species only, and so further resolution was not necessary. Any species from the same genus were also grouped together. We used the phylogeny from Regier et al. (2012b) to resolve relationships within the Pyraloidea. No further details were available for the Phycitinae family, leaving this group of eight species as a polytomy except for those species from the same genus, which were grouped together. The Tortricoidea were resolved to species level where possible following Regier et al. (2012a). Within the Tortricoidea, Paralobesia viteana was absent from all published phylogenies and therefore its position was left unresolved as a basal polytomy within this clade. The Archipini was also lacking detail on species from our analysis leaving those four species unresolved. Sohn et al. (2013) was used to fully resolve the Yponomeutoidea phylogeny, whereas Löfstedt et al. (1991) was used to resolve the three species in the genus Yponomeuta. The Bombycoidea phylogeny was taken from Mutanen et al. (2010). Mutanen et al. (2010) also provided further resolution to the Psychidae, although two species were missing rendering this group not fully resolved. Yamamoto and Sota (2007) and Sihvonen et al. (2011) were used to resolve the Geometroidea to subfamily, but no further details were available for all species present in our phylogeny leaving the six species of the Ennominae unresolved as well as the three species in the genus Operophtera. The Noctuidae were resolved using Zahiri et al. (2013), and the Lymantria phylogeny was taken from Sutrisno (2014). The superfamily Gracillariidae was fully resolved using the phylogeny from Regier et al. (2013). Relationships within the Gelechiidae were taken from Kaila et al. (2011) and Heikkilä et al. (2014).

In the absence of branch length information, all branches were assigned equal length (=1). The exception to this rule was cases where there remained uncertainty over the branching pattern; any polytomies were arbitrarily resolved but the resolved branches were assigned minimal branch lengths of 0.00001, giving them negligible weight in the analyses. This resolution was necessary because the phylogenetic comparative analysis approach requires fully dichotomously resolved phylogenies.

Following Ives and Garland (2014), we ran multiple tests to improve the strength of our inferences. These tests included generalised linear mixed models with Bayesian estimation (MCMCglmm) (Hadfield and Nakagawa, 2010) and the concentrated changes test (Maddison, 1990). The MCMCglmm models were conducted using the MCMCglmm package in R (Hadfield, 2010; R Core Team, 2014). To determine which traits are linked to our dependent variables when controlling for phylogenetic relatedness, all independent variables of interest were tested using MCMCglmm in separate models where wingspan was included as the covariate to control for body size. Specificially we performed one set of tests with antennal morphology as the dependent variable, coded as simple (1) and elaborate (2) antennae. In these tests the independent variables were female mating system (monandrous/polyandrous), egg development (proovigenic/synovigenic), and lifespan. In the second set of tests the dependent variable was female wing type coded as macropterous (1) or flightless (2). In these tests the independent predictors were female mating system, lifespan, antenna type, egg development, oviposition strategy (single eggs laid/multiple eggs clutches), and fecundity.

These models cannot incorporate missing data; therefore, the data sets were reduced to include only those species where all data were available for the variables being tested in each analysis. These tests require a prior to be set, and when the response variable is categorical in nature the prior for the residual variance should be fixed. For our analyses, we set the residual variance to 1 and used a χ² distribution with 1 degree of freedom (v = 1, nu = 1000, alpha.mu = 0, alpha.V = 1) for the random effects variance, as suggested by Villemereuil et al. (2013). We used the categorical family response, and used the slice function to improve mixing. To obtain adequate mixing with low autocorrelation, 5 × 10⁷ iterations were used with a 10,000 burnin period, followed by sampling every 2,500 iterations (thinning). The R code used in this procedure is provided in the Supplementary material.

The Concentrate Changes Test (CCT) was used to investigate the co–evolution of two binary discrete characters across a phylogeny, using the computer program MacClade 4 (Maddison and Maddison, 2005). CCTs examine changes from one state to another in a character of interest (the dependent variable) and whether these are more concentrated in lineages that have evolved a particular state of interest in another character (the independent variable) than would be expected by chance (determined by comparison with 1,000 permutations of the dependent variable on the phylogeny). When characters are mapped onto the phylogeny, the most parsimonious reconstruction of the character is applied for all species, including those with missing data, allowing us to leave all 93 species in each analysis. CCTs require a fully resolved phylogeny and we used MacClade to randomly resolve any polytomies 20 times in our composite phylogeny, performing the analysis on each tree individually, thereby yielding a mean p-value ± SEM. For each pair of traits, the tests were run in both directions, with the independent and dependent traits switched, to determine the order of evolution. We also applied the test only to changes where the parsimonious resolution of the evolution of the dependent was unequivocal. These tests lack the statistical power of MCMCglmm, and cannot control for body size, but unlike the MCMCglmm, they do provide information about the direction of evolution (i.e., which correlated trait preceded the evolution of the other), and possible patterns of causality.

Results

We found a significant association between antennal type and female mating system (MCMCglmm analysis: Table 2), with the concentrated changes test indicating weakly (p = 0.062) that elaborate male antennae evolve more frequently in species where females are monandrous (CCT analysis: Table 3). Indeed, most species in our analysis with elaborate male antennae have monandrous females, whereas most species with simple male antennae have polyandrous females (Figure 2A). Similarly, male antenna type is also associated with egg development pattern, with elaborate male antennae being associated with proovigenic females (Table 2). Indeed, within our sample of species with elaborate male antennae, only one is synovigenic, producing eggs as an adult (Figure 2B). However, the CCT analysis suggests (again weakly: p = 0.081) that proovigeny has evolved more frequently in species where males have elaborate antennae (Table 3). The type of male antennae was not associated with female lifespan (Table 2).

TABLE 2

Table 2. Results from the MCMCglmm analysis comparing the dependent variable (male antennal type, female wing type) to the independent variable, controlling for body size (male wingspan).

TABLE 3

Table 3. Results from the concentrated changes test investigating evolutionary associations between discrete characters with the state of interest listed.

FIGURE 2

Figure 2. (A,B) Percentage of species with particular traits relative to the type of male antennae.

We found a link between female flight ability and male antenna type (Table 2), with the CCT analysis suggesting that female flightlessness is more likely to have evolved in species where males have elaborate antennae, rather than vice versa (Table 3 and Figure 3C). Female flightlessness is also associated with a shorter female lifespan (Table 2), and the association between flight ability and oviposition strategy (single eggs vs. batches) approached significance in the MCMCglmm analysis (p = 0.056, Table 2 and Figure 3D). The CCT results suggest that flightlessness has evolved more often in species where females lay eggs in clutches (Table 3). While all the flightless species in our sample are proovigenic (Figure 3B), the lack of transitions from synovigenic to proovigenic severely limits the power for the MCMCglmm analysis, and these analyses did not converge, so we do not include these tests in our analysis. Although most flightless species have monoandrous females (Figure 3A), the phylogenetic comparative analyses found no significant association between female flight ability and mating system, or between flight ability and fecundity (Tables 2, 3).

FIGURE 3

Figure 3. (A–D) Percentage of species with particular traits relative to the flight ability of female.

Discussion

Prevailing wisdom states that elaborate antennae in males evolved to increase their ability to detect odours (specifically the sex pheromone of females). However, this hypothesis is contradicted by the considerable number of moth species where males do not possess elaborate antennae, despite most species utilising long-distance female sex pheromones. Our results suggest a more nuanced version of the hypothesis where ability to detect and locate females quickly is advantageous in some species: elaborate male antennae being more common in species where females are monandrous, proovigenic, or flightless. Concentrated changes tests indicate that it is more likely that the evolution of elaborate male antennae is concentrated lineages where females are monoandrous rather than vice versa (suggesting that the mating system evolved prior to the antennal morphology). Similarly, that the CCTs suggest that female flightlessness is more likely to have evolved in lineages where elaborate male antennae had already evolved rather than the opposite, which argues against the Hackman (1966) hypothesis that female flightlessness selects for males with more sensitive antennae. These trends provide some insights into the conditions leading to the evolution of both male elaborate antennae and female flightlessness, and more generally highlight how olfactory signal perception can be linked with movement.

Darwin’s (1871) conjecture that sexual selection favours male sensory receptor traits that improve the ability to detect and locate females more quickly is supported by the strong association between complex male antennal structure and female monandry. Female monandry places a premium on males being able to rapidly locate virgin females, whose numbers may decline during the male’s adult lifespan, and larger or more elaborate male antennae apparently facilitate this process [see also Johnson et al. (2017b)]. This interpretation implicitly assumes that elaborate antennae bestow greater sensory capabilities, and this has been widely assumed in previous analyses of antennal morphology in moths (Symonds et al., 2012; Javoiš et al., 2019). However, while elaborate antennae might facilitate the trapping of air flow and hence chemical compounds (Loudon and Koehl, 2000), the direct evidence that they have higher sensitivity is surprisingly rare. In this context exceptions to the patterns outlined above can be informative. For example, selection may still favour elaborate male antennae in the polyandrous satin moth Leucoma salicis because females initially oviposit a large clutch of eggs, and subsequently lay smaller clutches (Wagner and Leonard, 1979), which places a premium on males locating virgin females. In this context, it is interesting that there is a relationship between egg development strategy and male antennal morphology, with proovigeny (where females eclose with their full complement of eggs) being associated with elaborate male antennae.

Although the large majority (>90%) of species with elaborate male antennae have monandrous females, males have filiform antennae in 17 of the 30 monandrous species in our sample. In these species, perhaps males with filiform antennae utilise other means of improving sensitivity in detecting female pheromones. For example, detectability may be improved by increasing the number and/or density of sensilla, changing the distribution of the sensilla (Keil, 1989; Lee and Strausfeld, 1990), or adjusting the angle of antennal scales (Wang et al., 2018). Additionally, features of the adult population, such as high population density (Symonds et al., 2012) or male flight speed, may relax selection on rapid detection. Finally, male mating system may also affect the evolution of antennal morphology, although the nature of the relationship between typical male mating frequency and the strength of selection for greater detection capacity seems unlikely to be independent of other factors, including female mating strategy. Consequently it may be that it is overall mating strategy (monogamy vs. polygamy) that is the stronger determinant of selection on male antennal morphology.

Female flightlessness is thought to allow females to invest more resources in egg production, thereby yielding higher fecundity (Tweedie, 1976; Roff, 1990; Sattler, 1991). Our analyses revealed that female flightlessness is associated with a shorter adult lifespan and confining oviposition to a single clutch of eggs, an unsurprising pattern because laying eggs singly or across multiple clutches would require an ability to disperse to different oviposition sites (Sattler, 1991). However, there was no significant correlation between female flightlessness and fecundity after controlling for phylogeny and body size. Hunter (1995) reported a positive correlation between female flightlessness and fecundity, but this pattern was phylogenetically constrained, and the correlation was no longer significant after taking account of phylogeny. Our analysis supports that finding. Nevertheless, we cannot rule out the possibility that females can improve reproductive success by increasing egg size rather than egg number. Indeed, the flight ability of female tussock moth Orgyia thyellina varies seasonally: individuals eclosing in autumn have reduced wings whilst those emerging in summer have fully functional wings (Kimura and Masaki, 1977). The flightless females produce much larger eggs than females that emerge in summer, a strategy that may improve the survival of eggs that diapause over winter. Males of O. thyellina have elaborate, bipectinate antennae (Table 1), and it would be interesting to know if there is a similar seasonal pattern in antennal size.

While our data broadly support the idea that the presence of elaborate antennae in males favoured the evolution of female flightlessness, there are seven species within our data set where females are flightless, and yet males have simple, filiform antennae (Figure 3). Interestingly, females in five of these species are still mobile, either by walking or hopping (Contant, 1988; Sattler, 1991; Medeiros and Gillespie, 2011), and females of the other two species have unusual biological features. Females of Acentria ephemerella are aquatic and the loss of flight is likely to be an adaptation that supports this lifestyle (Miler, 2008), highlighting that the evolution of female flightlessness is not linked exclusively with issues associated with mate search. The second species indicates that the strength of selection through mate search can override the relationship between elaborate antennae and flightlessness: around 80% of Alsophila pometaria females are pseudogamous asexuals (Mitter and Klun, 1987), who produce asexual offspring after mating, resulting in a strongly female–biased population. So, while the loss of flight has likely evolved due to similar pressures affecting other species in our data set, the pseudogamous nature of this species may relax selection favouring elaborate male antennae.

To conclude, our findings suggest that the communication and mating systems of moths are inherently associated, with elaborate male antennae being associated with lineages where females are monandrous, suggesting that necessity to detect females quickly has selected for more sensitive males. In turn male elaborate antennae may have subsequently driven the evolution of female life-history and flightlessness where having sensitive males promoted selection for complete development of eggs at the expense of movement capability. A key aspect of this narrative is that elaborate antennae in males is associated with greater sensitivity, an assumption that still needs to be more thoroughly tested in moths.

Data availability statement

The original contributions presented in this study are included in the article/Supplementary material, further inquiries can be directed to the corresponding authors.

Author contributions

TJ assembled the comparative data set. TJ and MS conducted the comparative statistical analyses. All authors conceived and designed the research, interpreted the data, prepared and edited the manuscript, and approved the submitted version.

Funding

We thank the Australian Research Council (DP0987360) for financial support.

Acknowledgments

We thank Nina Wedell and the late Matthew Gage for their helpful comments on an earlier version of this manuscript.

Conflict of interest

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fevo.2022.919093/full#supplementary-material

References

Abdullah, M., Sarnthoy, O., Chaeychomsri, S., and Sarnthoy, O. (2000). Comparative study of artificial diet and soybean leaves on growth, development and fecundity of beet armyworm, Spodoptera exigua (Hübner) (Lepidoptera: Noctuidae). Nat. Sci. 34, 339–344.

Google Scholar

Abernathy, R. L., Teal, P. E. A., and Tumlinson, J. H. (1994). Age and crowding affects the amount of sex pheromone and the oviposition rates of virgin and mated females of Helicoverpa zea (Lepidoptera: Noctuidae). Ann. Entomol. Soc. Am. 87, 350–354. doi: 10.1093/aesa/87.3.350

PubMed Abstract | CrossRef Full Text | Google Scholar

Akter, T., Jahan, M., and Bhuiyan, M. (2013). Biology of the angoumois grain moth, Sitotroga cerealella (Oliver) on stored rice grain in laboratory condition. J. Asia. Soc. Bangladesh Sci. 39, 61–67. doi: 10.3329/jasbs.v39i1.16034