Investigation: Habitat Selection in Flour Beetles - Biology


In this laboratory, you will observe the behavior of an insect and design an experiment to investigate its responses to environmental variables.


After doing this laboratory you should be able to measure the effects of environmental variables on habitat selection in a controlled experiment.

Pre-Lab Questions

1. What is meant by an environmental variable?

2. What is meant by habitat selection?

3. Provide two examples of variables you could test with any animal study.

Examine sample data from an experiment on pillbug. The pillbugs were placed in the choice chamber, 5 on each side. They were then observed for 10 minutes and the number on each side was recorded.

4. Graph the data shown and write a brief summary of what the data shows about pillbug behavior.

Time(minutes)# In Wet Chamber# In Dry Chamber
Conduct Your Experiment

Choose the variable you want to test from the following options. (Choose ONE)

  1. Wet vs Dry
  2. Water vs Vinegar
  3. Sand vs Paper
  4. Light Vs Dark

5. Write a hypothesis about what environmental condition will be preferred by your specimen. Write the hypothesis as a complete sentence.

Experiment Set-Up

For this lab, you will use flour beetles. Set up your choice chambers to test the variable of your choice. Generally, filter paper or paper towels can be used to test chemicals such as alcohol and vinegar. Your instructor will have the materials you need. Each side of the chamber should have a different variable.

6. Label the diagram to show how yours is set up.

7. Collect Data: You will need to collect data for at least 10 minutes. A graph will not be necessary this time.

Time (minutes)Side LSide R

8. Summarize Data: Examine the data you collected and write a statement that summarizes the data and answers the experimental question (hypothesis.) This statement should be a complete and thoughtful sentence that is supported by the data you gathered.

Instructor Notes:

Choice Chambers available form Amazon or Carolina Biological Supply Company

Cut filter paper in matching shapes to save time. You can just soak the filter paper before putting it into cage. Paper towels that aren’t flush with containers will likely result in the animals crawling under them.

Students may need additional assistance with drawing the graph on the first page, I recommend a bar graph with colors indicated numbers on the wet and the dry side.

Other animals can be used, I maintain a mealworm colony in my class to feed lizards and turtles. The adult beetles perform well in this experiment. You could potentially substitute crickets or have students collect pillbugs. I didn’t find the larval mealworms to be very good in this experiment, they seem to have limited senses and do not move very quickly.

Beetles in choice chambers. Filter paper allows for changing habitat properties.

Ecology of kin and nonkin larval interactions in Tribolium beetles

The larvae of flour beetles Tribolium castaneum and T. confusum were reared in two kinds of groups: full siblings and unrelated individuats. These kin and nonkin groups were reared in open cultures, in which emigration was permitted (both species) and in closed cultures, in which emigration was prohibited (only T. confusum). We measured larval development time and survivorship, weight of pupae, and time of larval emigration from open cultures. The effects of age structure were investigated by establishing open cultures of larvae of uniform age (larvae hatched from eggs laid within 72 h) and cultures of larvae of variable age (eggs laid within 240 h). In closed cultures of siblings, T. confusum larvae pupated on an average 2.2 days earlier than larvae reared in nonsibling groups. In T. castaneum, more small and medium size and fewer large size larvae emigrated from groups of siblings compared to groups of nonsiblings. Males that emigrated and pupated remained with their sibs for a shorter time than did similar males raised with unrelated larvae. In T. castaneum, age structure variation reduced the sibs tendency to migrate, but did not influence interactions among unrelated larvae. The genetical effects of kinship and the ecological effects of age structure were shown to affect the interactions of Tribolium larvae reared in groups. Reducing the similarity between individuals, either genetically or demographically (using mixed broods or mixed age cohorts), changed the pattern of larval interactions. Upon occasion, the effects of kin interactions may well be the mechanical consequences of the coexistence of similar individuals rather than the effects of altruistic behavior.

This is a preview of subscription content, access via your institution.


2.1 Inhomogeneous Poisson SPP and habitat selection

Spatial point processes (SPP) are models describing the probability of a point event (e.g. animal presence) across space (Cressie, 2015 ). A spatial point pattern, such as the collection of points where a species occurs, is a stochastic realisation of a SPP. Under a homogeneous Poisson SPP, the probability of observing exactly N events within a spatial unit of size m is: , where λ is the (homogeneous) Poisson SPP intensity (the expected density of events anywhere within the spatial domain Cressie, 2015 ). The inhomogeneous Poisson SPP is the simplest alternative to the complete spatial randomness assumed by the homogeneous Poisson SPP and is hence a fundamental model in spatial ecology (Cressie, 2015 Hooten et al., 2017 ). Under an inhomogeneous Poisson SPP, intensity varies across spatial domain Ω, so that the local SPP intensity in location x (ϵΩ) is: (1) Here μ is a normalisation constant ensuring that the mean (homogeneous) intensity across Ω is . The βk's are parameters linking the relative SPP intensity at location x to the local values of q habitat dimensions, Hk=1:q(x), such as resource availability (e.g. nesting sites) or autecological conditions (e.g. solar radiation). The inhomogeneous Poisson SPP thus provides a crucial statistical link between observed spatial occurrence patterns and the underlying drivers of spatial heterogeneity in population density (i.e. habitat quality) for habitat unit of size m centred at x, λ(x) = E[N(x)]/m (Cressie, 2015 Hooten et al., 2017 Renner et al., 2015 Velázquez et al., 2016 Warton & Shepherd, 2010 ).

Habitats are sectors of environmental space differing along one or more of its dimensions, such as resource availability (e.g. food, water, or shelter) or autecological conditions (e.g. temperature, pH or salinity the constituents of the species' fundamental niche), and hence in their quality (Aarts et al., 2008 Morris, 2003 ). Habitat units are discrete, environmentally homogeneous, sectors of geographical space. Habitat selection is the usage of a given habitat more often than would be expected based on the availability of its units (Avgar et al., 2017 Lele et al., 2013 Matthiopoulos et al., 2020 ). The exponential Habitat Selection Function (eHSF) is a common statistical model of animal space use, yielding values that are proportional to the usage of a habitat unit centred at x: (Manly et al., 2002 ). In the context of eHSF, βk is termed the selection coefficient for Hk, and it is readily estimable using standard logistic regression with presence-background (also known as used-available /-control /-pseudoabsence) data. Although other HSF formulations exist, eHSF is by far the most popular among field biologists due to its ease of application and the prevalence of presence-only and remote-sensing data (Avgar et al., 2017 McDonald et al., 2013 ).

In recent years, several authors have shown that the use of logistic regression with presence-background data to fit an eHSF corresponds to estimating an inhomogeneous Poisson SPP (Aarts et al., 2012 Fithian & Hastie, 2013 Johnson et al., 2013 McDonald et al., 2013 McDonald, 2013 Renner et al., 2015 Warton & Aarts, 2013 ). This can be demonstrated by expressing p(x is used) as the Poisson probability that at least one event occurs within a habitat unit of size m (centred at x): , providing a simple formal link between ‘abundance’ and ‘occupancy’. As long as (because either μ or m are very small), is well approximated by . Based on Equation 1, we can now write: (2) Assuming an inhomogeneous Poisson SPP, and for any two habitats, i and j (each a unique location in environmental space with coordinates and ), with corresponding total areas Mi and Mj (the sum of the areas of their respective habitat units), we can express the ratio of expected habitat-specific densities, ρi and ρj, using Equation 1: (3) (4) Equation 4 is the mathematical definition of the eHSF coefficient, βk, also known as the log relative-selection strength (Avgar et al., 2017 ) it is the natural logarithm of the inter-habitat (expected) density ratio divided by the inter-habitat distance along the focal habitat dimension, k. Note that, for categorical habitat partitions (e.g. Hk,i = ‘forest’ and Hk,j = ‘meadow’), the inter-habitat distance in environmental space is always 1 ( where is an indicator function, accordingly valued 0 or 1). In other words, βk is a measure of the proportional population density change across the k'th dimension of environmental space.

Assuming IFD, the expected density ratio, and hence βk, can be explicitly calculated as the fitness-equalising solution across a two-habitat system (see next section). Hence, under the prevalent assumptions of an adequately defined availability domain (Johnson, 1980 Matthiopoulos et al., 2020 ), and log-linear relationship between λ and Hk=1:q (Renner et al., 2015 ), βk is a simple function of the IFD if the IFD is consumer- and/or resource density dependent, so is exponential habitat selection strength.

2.2 A multidimensional model of the ideal free distribution

Our approach is based on the premise that ‘habitat quality’ should be more generally regarded as a function of multiple factors (multiple dimensions in environmental space), with various degrees of density-dependent payoffs (Lampert et al., 2003 Matthiopoulos et al., 2015 Tyler & Gilliam, 1995 ). We refer to this extension as the multidimensional Ideal-Free Distribution (hereafter, mIFD) as it can be applied to arbitrarily complex habitat-dependent payoff functions, in contrast to the common perspective of the IFD as emerging from a single density-dependent relationship.

(5) Here is the direct (additive) fitness gain from local autecological conditions, whereas is the contribution to fitness obtained from acquisition and assimilation of resources, which is a function of habitat-specific resource density, consumer density and a vector of parameters, θ, that may or may not vary with autecological conditions. These parameters might include, for example, habitat-specific variation in consumer search rate or net-energy gain from a single consumption event. Hence, to keep our formulation as general as possible, we allow autecological conditions to affect ωi via two different pathways, a direct (additive as in Matthiopoulos et al., 2015 ) pathway, and an indirect (multiplicative as in Pulliam, 2000 ) pathway operating through autecological effects on the parameters of . For example, temperature or habitat structure (e.g. vegetation cover or ruggedness) may directly influence consumer fitness (by affecting metabolic rates, thermoregulation costs or predation risk), but could also have multiplicative influences by affecting the parameters of , such as search rate and intraspecific competition intensity (e.g. due to temperature- or habitat-dependent movement costs). (6) To investigate the ecological conditions leading to density-dependent mIFD and eHSF we must first define an explicit functional form for (Van Der Meer & Ens, 1997 ). We set to one of the two commonly studied consumer-dependent Type-II formulations either a Beddington–DeAngelis functional response (Beddington, 1975 DeAngelis et al., 1975 ), or an Arditi–Akcakaya functional response ( Arditi & Akçakaya, 1990 ). Both versions have been well-documented under experimental conditions (Kratina et al., 2009 Novak et al., 2017 Prokopenko, Turgeon, et al., 2017 ). Under the Beddington–DeAngelis model, resource consumption rate is given by: (7) (8)

Substituting Equations 7 or 8 into Equation 6, and solving for ρi given ρj, Ri, Rj, θ[Ai] − θ[Aj], and g[Ai] − g[Aj], allows us to explore the expected relationship between mIFD (habitat selection strength) and the system-wide densities of consumers and resources. For the sake of simplicity and gerality, we avoid specifying functional forms for θ[A] and g[A]. Instead, we solve the mIFD for pairs of habitats (i and j) that differ by a single unit of one parameter or variable (a[A], b[A], c[A], s[A], R or g[A] dimensions in environmental space), but are identical in all other aspects. Our analysis thus considers 12 ecological scenarios, six types of between-habitat differences under each of the two functional response formulations. Given the simplest imaginable ecological setting of a single type of resource, two habitats may minimally differ by their associated resource abundance, search rates, handling times, resource assimilation efficiencies, intensities of consumer interference or autecological fitness gains. Our aim here is to provide a theoretical perspective that is ‘as simple as possible, but not simpler’ by providing an exhaustive account of all possible configurations of a two-habitat system under a generic fitness function. In Appendix S1 we provide the explicit terms for ρ1 for each of 12 scenarios the six types of between-habitat differences (a[A], b[A], c[A], s[A], R or g[A]) under each of the two functional response formulations considered here. In Figures 1–3-1–3 we plot the mIFD-based habitat selection strength (βk = ln[ρi] − ln[ρj] Equation 4 where Hk,iHk,j = 1) as a function of mean consumer and resource densities ((ρi + ρj)/2 and (Ri + Rj)/2 respectively). As ours is a qualitative investigation, parameter values, as well as variable magnitudes, are arbitrary. In the plots shown below, unless stated otherwise, all parameters were set to 1 except for assimilation efficiency (s[A]), which was set to 10.


Maternal immune challenge impacts offspring development times dynamically in laboratory and wild beetle strains

In the laboratory beetle experiment, maternal treatment, batch order and their interaction significantly affected larval development time (Table 1). Peri-primed (occurring immediately after maternal challenge) larval offspring from batch 1 developed significantly faster than batch 1 unprimed and saline control larvae (Cox proportional hazards regression on individual development time with breeding group as a random effect and contrast tests on a predetermined subset of possible comparisons, Table S1, Supporting information), and also developed significantly faster than all batch 2 treatments (Fig. 2a, representing breeding group median development). Batch 3 saline larvae also developed significantly faster than unprimed batch 3 larvae and trended towards faster development than batch 3 primed larvae. Across batches, there was no significant difference in development time for unprimed larvae.

Factor d.f. Sum Sq Mean Sq F value P (>F)
Factorial anova on laboratory Tribolium castaneum
Bt treatment 2 117 58·5 12·212 <0·0001
Batch 1 2 1·67 0·348 0·555
Treatment:Batch 2 245 122·27 25·525 <0·0001
Residuals 1496 7166 4·79
Factorial anova on wild T. confusum
Gregarine treatment 2 842 421 45·38 <0·0001
Batch (Bt treatment) 3 85 28·2 3·039 0·0285
Treatment:Batch 6 357 59·6 6·42 <0·0001
Residuals 600 5566 9·3

In the wild beetle experiment, parents were Bt-challenged prior to batch 2 rather than batch 1, and batch 1 took on the role of the unprimed group. Consistent with laboratory beetle experiments, larval development time significantly and transiently increased (Fig. 2b, representing breeding group median development time) immediately after the Bt priming event for all parental gregarine treatment conditions (Cox proportional hazards regression and planned contrasts on individual beetle data, Table S1, Supporting information), returning to pre-priming levels by batch 3 for all treatments except cc, which remained significantly accelerated.

Maternal challenge improves offspring survival against infection except when co-infected

Maternal treatment significantly impacted batch 3 laboratory T. castaneum adult offspring survival (Fig. 3a) after Bt infection. Primed beetles were significantly more likely to survive infection than unprimed offspring (Cox hazard ratio (95% CI) = 0·347 (0·18, 0·67), χ 2 = 9·58, d.f. = 1, P = 0·002), while offspring from saline control mothers were not more likely to survive than unprimed offspring (Cox hazard ratio (95% CI) = 0·65 (0·37, 1·11), χ 2 = 2·44, d.f. = 1, P = 0·12). Primed individuals trended towards a higher survival rate than offspring from saline control mothers, but the difference was marginally non-significant (χ 2 = 3·44, d.f. = 1, P = 0·065). Bacterial density over time for unprimed and primed treatment groups (Appendix S1 and Fig. S1, Supporting information) suggests that primed individuals are better able to resist bacterial growth over the interval preceding the period of peak disease-induced mortality (12 h).

For the wild T. confusum, the probability of surviving Bt infection depended upon whether parents were co-infected with gregarines at the time of the priming challenge and batch order, which co-varies with parental Bt challenge (Cox proportional hazards regression with parental treatment and batch as main effects and breeding group as a random effect, Table 2). Overall, offspring from cc, cd, and dc treatments had nearly indistinguishable survival rates (Fig. 4), but offspring from the dd treatment had a significantly lower probability of survival. Comparison of survival across batches (Fig. 3b, from unprimed batch 1 to peri-primed batch 2 to fully primed batch 3) demonstrates that wild beetles are capable of priming (Table 2), as offspring from both batches 2 and 3 were significantly more likely to survive Bt infection than batch 1 offspring.

Level Coef SE (coef) Robust SE Exp (coef) Lower 0·95 Upper 0·95 z P value
cd (treat) 0·10 0·22 0·31 1·11 0·61 2·03 0·33 0·741
dc (treat) 0·15 0·23 0·27 1·16 0·69 1·96 0·56 0·578
dd (treat) 0·84 0·23 0·31 2·32 1·26 4·27 2·71 0·007
Batch 2 −0·57 0·26 0·25 0·57 0·35 0·92 −2·27 0·023
Batch 3 −0·63 0·25 0·27 0·53 0·31 0·90 −2·37 0·018
  • Bt treatment co-varies with batch order as parents were challenged prior to batch 2. coxph model formula is survival

The impact of larval development time on survival post-infection

In the laboratory T. castaneum, the median development time of its breeding group cohort was not a significant predictor of individual adult survival, regardless of priming status (Fig. 5a, Table 3). In the wild beetle experiment, however, the median development time significantly predicted individual survival (Table 3), although the slope of the effect depended on its interaction with priming status. Faster development strongly predicted survival in primed beetles, but it does not adequately explain survival in unprimed beetles and has even less influence on survival in peri-primed beetles (Fig. 5b).

Effect Estimate SE z value P (>|z|)
Binomial GLMM on laboratory Tribolium castaneum
All treatments
Intercept −0·03574 0·9861 −0·036 0·97
Development −0·12222 0·18733 −0·652 0·51
Naïve only
Intercept 10·7636 9·9385 1·083 0·279
Development −0·4019 0·3658 −1·099 0·272
Saline only
Intercept −7·8276 9·2815 −0·843 0·399
Development 0·269 0·3473 0·775 0·439
Primed only
Intercept −5·9867 9·7178 −0·616 0·538
Development 0·1657 0·3493 0·474 0·635
Binomial GLMM on wild T. confusum
All treatments
Intercept −4·23438 1·32683 −3·191 0·00142
Development 0·10857 0·03722 2·917 0·00354
Unprimed (batch 1)
Intercept −4·083 3·844 −1·062 0·288
Development 0·114 0·1028 1·109 0·267
Peri-primed (batch 2)
Intercept −1·87088 3·51483 −0·532 0·595
Development 0·04137 0·1038 0·399 0·69
Primed (batch 3)
Intercept −5·25314 1·70586 −3·079 0·00207
Development 0·13159 0·04699 2·8 0·0051


The results of the study indicated that the rearing of two IUCN red-listed (endangered) species of flat bark beetles (C. cinnaberinus and C. haematodes) from larvae to adults is possible in laboratory conditions and both species at the examined developmental stages clearly prefer dead prey. This information appears to highly relevant for management and the conservation of the species.

Both larvae and adults of Cucujus prey on dead individuals, refusing living ones. It is likely that the refuse to forage on live individuals due to the active movement of live prey. These findings partially support previous data considering the Cucujus species as scavengers only occasionally preying on live larvae and pupae of other beetles ( Mamaev et al. 1977). The under bark environment is apparently populated by living individuals that could be potential prey, but in this subcortical environment, many trophic interactions may occur. Thus, for Cucujus species, it is possible to find eggs, dead larvae, pupae, and adults of other arthropod species.

We predicted that remnants of prey killed by other predator species and individuals killed pathogens might represent a suitable food for scavengers (e.g., Horák 2011). In some periods of the year (e.g., in winter), this food availability could be very important for the survival of Cucujus individuals, especially when low temperatures and snow limit the food search. Cucujus larvae are known to be highly freeze-tolerant ( Sformo et al. 2010). Therefore, they can move all year long and search for prey killed by extreme abiotic conditions (e.g., low temperatures).

The investigated Cucujus species spend their immature life stages exclusively under bark, slowly moving inside this microhabitat until pupation and adult emergence. In addition, the adults spend most of their life inside the trunk crevices ( Horák and Chobot 2011), where they lay the eggs after the adults mate (Bonacci, personal observations).

Since in nature the developmental time of Cucujus beetles is over a year long, the choice of dead prey (belonging to the scavenger functional group) can be more advantageous than predation. The choice of prey taxa in this study was based on the recent data indicating that the community associated with C. cinnaberinus and C. haematodes included species of the family Scolytidae, ants of the genus Lasius and dipteran larvae ( Horák et al. 2012b, Horák 2015, Mazzei et al. 2018). A historical study also mentioned the presence of these arthropods in the environment of C. cinnaberinus ( Palm 1941). The absence of significance in the choice of sets of dead prey indicated opportunistic feeding behavior on dead arthropods. Our investigations showed that larvae and adults of Cucujus spp. in the laboratory refused fungi as a food source.

These results agree with previous data indicating Cucujus preference of decaying bast (phloem Přikryl et al. 2012) and may also explain why fungi are mentioned as a food in the study as insect pathogens, fungi may directly cause death or develop on dead arthropods immediately after death. Beetles may therefore accidentally ingest the hyphae while feeding on dead individuals covered by moulds and other fungal species.

In conclusion, the data obtained in this study support the hypothesis that flat bark beetles may feed on dead insects ( Horák 2011). The analysis of food preference in laboratory conditions showed a statistically significant preference for dead prey in comparison to living ones, supporting the scavenger feeding habits of adults and larvae of genus Cucujus. Therefore, the species is most probably saprophagous in the wild. However, many other aspects of the biology and behavior of these endangered beetles still require detailed investigation.

Species Conservation in the Wild

The two flat bark beetle species examined in this study live in highly similar microhabitat ( Horák et al. 2011, 2012b Mazzei et al. 2018, Jaworski et al. 2019). However, C. cinnaberinus probably has adapted to lands affected by anthropogenic uses ( Jaworski et al. 2019), and it exhibits a higher dispersal ability ( Della Rocca and Milanesi 2020). Therefore, C. haematodes appears to be affected by the conversion of traditionally managed forests into tree plantations ( Jaworski et al. 2019), while C. cinnaberinus apparently manages to survive, with some limits, in less favorable conditions ( Bełcik et al. 2019)

The main problem concerning the conservation of these beetle species is that potentially dangerous trees and their remnants (e.g., snags or high stumps) in the European landscape are frequently removed. These trees in open areas (such as riparian greenings or old avenues) or urban locations (e.g., parks) are under threat ( Horák 2018). Such trees are sometimes inhabited by the two studied species (especially, C. cinnaberinus), and their larvae are often gregarious and abundant in one microhabitat ( Mazzei et al. 2011). Therefore, the loss of these larval aggregations may be unfavorable for the local population of these beetles species, protected in pan-EU Natura 2000 network ( To avoid loss of the population of these species when the inhabited tree is felled and has to be replaced or destroyed (e.g., due to public safety or other reason), the option for the larvae is their relocation in the laboratory and rearing by artificial feeding. This study shows for the first time that the rescue of the local populations of the two beetle species is possible through laboratory rearing and provides new data about diet preferences, life cycle duration, and mortality. Artificial feeding may be successfully applied for population increase or reintroductions. In conservation projects, the laboratory rearing of Cucujus individuals for future release could be very useful in rescue programs or reintroduction to habitats where the species have disappeared.


The P. hirtus adult head transcriptome

We chose to perform RNA sequencing, which enables highly accurate qualitative and quantitative assessment of gene expression (Nagalakshmi et al., 2008 Wilhelm and Landry, 2009), on the P. hirtus adult head transcriptome to generate a database that could be probed for the conservation and expression of vision-related genes. Poly-A RNA was isolated from 25 dissected head capsules of CO2-anesthetized animals of both sexes after removal of the antennae. Using the Illumina sequencing platform, we generated 27,428,409 sequence reads of 75 bp length. Close to 20% of these (5,476,803) produced significant alignments by blastx to RefSeq protein sequences of the distantly related red flour beetle T. castaneum (Tenebrionidae) and could be sorted into 9888 putatively ortholog-specific sequence read groups. Assembly of sequence reads in each group yielded a total of 67,079 contigs with an average length of 163.3 bp. Contig orthologies were re-investigated by reciprocal BLAST against the T. castaneum and D. melanogaster RefSeq databases. This step identified 6259 and 5649 high confidence orthologs in T. castaneum and D. melanogaster, respectively. These numbers were consistent with the closer phylogenetic relationship of P. hirtus to T. castaneum. The only moderately higher number of insect-conserved loci previously identified in genome sequence comparisons (Richards et al., 2008) validated the representational depth of the P. hirtus transcript database.

Expression of phototransduction genes

Using a previously established compilation of conserved insect phototransduction genes as the query sequence source (Bao and Friedrich, 2009), we searched the P. hirtus adult head transcript database for orthologs of insect vision-related genes. This approach revealed the presence of transcripts of all critical components of the phototransduction protein network as characterized in Drosophila (Table 1) (for reviews, see Katz and Minke, 2009 Wang and Montell, 2007). The first step in the phototransduction signaling cascade involves the absorption of photons by the G-protein-coupled receptor complexes (GPCRs), which consist of a member of the 7-transmembrane photopigment opsin family and the G-protein subunits dGαq, Gβe and Gγe, all of which were found in the P. hirtus transcript database. Most significant was the recovery of dGαq and Gβe, which are photoreceptor specific in Drosophila (Dolph et al., 1994 Schulz et al., 1999 Scott et al., 1995 Yarfitz et al., 1991), as well as the recovery of a single opsin ortholog, which was supported by 369 single reads yielding 1008 bp of contiguous coding sequence. Gene tree analysis revealed that the P. hirtus singleton opsin was a member of the long wavelength (LW)-sensitive opsin subfamily (Fig. 2). Within this subfamily, the P. hirtus opsin was most closely related to the LW opsin of T. castaneum, as expected on taxonomic grounds. Interestingly, the selective conservation of the broadband-sensitive LW opsin in P. hirtus corresponds to the discovery of LW opsins in other species adapted to low light level ecologies such as cave crayfish (Crandall and Hillis, 1997).

Further, consistent with the preservation of functional phototransduction in P. hirtus, we recovered orthologs of key genes specifically involved in opsin deactivation and recycling, including Arrestin 1 (Arr1), Arrestin 2 (Arr2), G-protein-coupled receptorkinase 1 (Gprk1) and the Rhodopsin phosphatase retinal degeneration C (rdgC) (Table 1). We also detected the ortholog of Rab protein 6 (Rab6), which is thought to be specifically involved in the transport of rhodopsin in the Golgi complex (Shetty et al., 1998 Wang and Montell, 2007).

The second major step in Drosophila phototransduction involves the interaction of activated dGαq with the phospholipase NO RECEPTOR POTENTIAL A (NORPA) (Bähner et al., 2000 Bloomquist et al., 1988). Drosophila possess two differentially spliced isoforms of norpA, one of which, type I, is specific to the retina while the other, type II, is found in a variety of tissues outside the retina (Bloomquist et al., 1988 Kim et al., 1995 Zhu et al., 1993). Only norpA type II was detected in the P. hirtus transcript database, although both are conserved in Tribolium (Table 1). Interestingly, the Drosophila norpA type I photoresponse phenotype can be almost completely rescued by retinal misexpression of norpA type II (Kim et al., 2003). It is therefore tempting to speculate that a loss of norpA type I was compensated by norpA type II in P. hirtus.

The activation of NORPA is followed by the transient activation of the signalplex (Montell, 1998), the core of which is formed by the Ca 2+ channel protein TRANSIENT RECEPTOR POTENTIAL (TRP) and the scaffolding protein INACTIVATION NO AFTERPOTENTIAL D (INAD). Orthologs of both corresponding genes were present in the P. hirtus transcript database. In addition to TRP, Drosophila photoreceptor cells express the related protein channel genes TRPL and TRPγ (Niemeyer et al., 1996 Xu et al., 2000). Only the former was represented in the P. hirtus transcript database although TRPγ is conserved in Tribolium. Also noteworthy was the recovery of the protein kinase neither inactivation nor afterpotential C (ninaC), which is specifically required for the calmodulin (CaM)-mediated localization of the signalplex in Drosophila photoreceptor cells (Porter et al., 1993).

Differential preservation of structural photoreceptor gene expression

Studies in Drosophila have shown that the phototransduction protein machinery is localized in the highly condensed array of microvillar extensions of the photoreceptor rhabdomere, the formation of which requires specific proteins. This includes the photoreceptor cell-specific homophilic transmembrane protein CHAOPTIC (CHP), which confers adhesion between the microvilli, and the antagonistic membrane protein PROMININ (PROM) (Zelhof et al., 2006). In addition, Drosophila photoreceptors secrete the extracellular proteoglycan protein SPACEMAKER (SPAM), also known as EYES SHUT (EYS), which results in the creation of an inter-rhabdomeral space between the Drosophila photoreceptors (Husain et al., 2006 Zelhof et al., 2006). We detected intermediate levels of chp expression in P. hirtus (Table 1), further supporting the presence of canonical insect photoreceptor cells in the adult head of this species. We also detected a low level of spam expression but found no evidence of prom in the P. hirtus transcript database (Table 1). As spam is also involved in mechanical stress protection of sensory neurons outside the visual system (Cook et al., 2008), these findings are indicative of photoreceptors in P. hirtus that are not associated with inter-rhabdomeral space formation.

Phylogenetic analysis of P. hirtus long wavelength (LW) opsin evolution. The putative P. hirtus (Phir) opsin protein sequence was aligned with the retinal opsin sequences from Drosophila melanogaster (Dmel), the honeybee Apis mellifera (Amel), the tobacco hornmoth Manduca sexta (Msex) and Tribolium castaneum (Tcas). Ambiguous regions in the resulting multiple alignment were removed, leaving 285 sites subjected to gene tree analysis. Numbers at branches indicate Tree-Puzzle branch support values.

Phylogenetic analysis of P. hirtus long wavelength (LW) opsin evolution. The putative P. hirtus (Phir) opsin protein sequence was aligned with the retinal opsin sequences from Drosophila melanogaster (Dmel), the honeybee Apis mellifera (Amel), the tobacco hornmoth Manduca sexta (Msex) and Tribolium castaneum (Tcas). Ambiguous regions in the resulting multiple alignment were removed, leaving 285 sites subjected to gene tree analysis. Numbers at branches indicate Tree-Puzzle branch support values.

Differential preservation of eye pigmentation gene expression

Insect compound eyes are usually deep brown or black in appearance because of pigment granule formation in the photoreceptors and their accessory pigment cells for optical insulation between ommatidia and the regulation of light influx within ommatidia. To investigate whether the lack of overt pigmentation in the presumptive P. hirtus lateral eyes was correlated with a reduction or loss of eye pigmentation gene transcripts, we screened the P. hirtus adult head transcriptome database for the presence of Drosophila eye pigmentation gene orthologs. While the large majority of 25 queried genes were present in P. hirtus (Table 2), we failed to detect the expression of orthologs of the ABC transporter protein genes white (w), brown (bw) and scarlet (st), which are essential and specific for membrane passage of eye pigment compounds in Drosophila (Mackenzie et al., 2000). The lack of bw corresponded to the absence of this gene in Tribolium, which expresses only one type of eye pigment (ommochrome) (Lorenzen et al., 2002). The additional lack of w and st in the P. hirtus head transcriptome suggested a gene loss-associated complete reduction of eye pigmentation in this species. This conclusion was reinforced by the absence of the eye pigmentation-associated phosphoesterase prune (pn) and kynurenine 3-monooxygenase cinnabar (cn) in the P. hirtus transcript database, both of which are conserved in Tribolium (Table 2).

Analysis of photoresponse behavior

The evidence of a functional visual system prompted the question of the significance of photic stimuli in the biology of P. hirtus. To explore the possibility of phototactic behavior in P. hirtus, we investigated dispersal behavior in a light (L) versus dark (D) choice model (Fig. 3) as detailed in Materials and methods. Strikingly, after only 2 min, less than 30% of the tested animals were recorded in the L-area. For the rest of the experiment, the proportion of animals in the L-area varied between 10 and 30%. This translated into an average phototaxis index of –0.658 or a probability of 0.17 of observing an animal in the L-area (Maria Camassa, 2001). Continuous observation of five independent animals in the same test conditions further revealed that animals tended to move faster in the L-area while resting more often in the D-area.

To evaluate how the phototactic behavior of P. hirtus compared with that of species with a fully developed visual system, we subjected the closely related eutroglophile P. cavernicola to the same L/D choice test. Ptomaphagus cavernicola animals withdrew on average twice as quickly into the D-area such that an average of less 30% were scored in the L-area 1 min after the start of the experiment. Ptomaphagus cavernicola also exhibited a stronger average negative phototaxis index (–0.889) over the subsequent 8 min of the experiment (Fig. 3A), corresponding to a probability of 0.05 of being observed in the L-area (Fig. 3B). Besides revealing a shared disposition of P. hirtus and P. cavernicola to avoid exposure to strong light, the similarity in the response of the two species was highly suggestive that P. hirtus utilized peripheral photoreception in phototaxis, which is a reasonable assumption for P. cavernicola. Continuous behavorial observation of P. cavernicola individuals revealed similar D-area-biased resting patterns to those in P. hirtus but also a faster mobility of this larger species (3–4 mm body length), contributing to the faster withdrawal into the D-area.

To explore how the phototactic behavior of the Ptomaphagus species compared with that of non-cave-adapted Coleoptera, we also tested T. castaneum, which is characterized by crepuscular digging behavior, relatively small eyes and the lack of blue-sensitive opsin (Jackowska et al., 2007). In contrast to the Ptomaphagus species, T. castaneum demonstrated no evidence of Lor D-area preference. The proportion of animals present in the L-area decreased more slowly after onset of the experiments, never dropped below 30% and averaged 49%, translating into equal probabilities of being encountered in the Dor L-area. These findings supported the notion that both P. hirtus and P. cavernicola are characterized by cave adaptation-associated enhancement of negative phototaxis.

Conservation of circadian clock gene expression

Luminosity-based measuring of day length can play a role in the entrainment of oscillating processes through the circadian clock gene network, alone or in addition to temperature or other external zeitgeber stimuli. Remarkably, eutroglophilic and troglobiotic ground beetles exhibit day length entrainment of diurnal activity rhythms (for a review, see Lamprecht and Weber, 1992). To explore the potential for a role of vision in circadian entrainment, we also probed the P. hirtus transcript database for the presence of clock gene orthologs because the expression of clock genes is concentrated in neuronal cell clusters in the brain of insects (for a review, see Tomioka and Matsumoto, 2010). This search revealed the expression of all components of the insect clock gene core network that have been identified in Tribolium (Table 3) (Cortes et al., 2010), including the bHlH-PAS transcription factor genes period (per), cycle (cyc), tango (tgo), clock (clk) and timeless (tim) as well as interacting genes like Par-domain protein 1 (Pdp1) and vrille (vri). Previous studies reported the preservation of the photolyase chryptochrome 2 (cry2) but the likely ancestral absence of its paralog cryptochrome 1 (cry1) in Tribolium (Yuan et al., 2007 Zhu et al., 2005). Consistent with this, only cry2 was detected in the P. hirtus transcript database.

Investigation: Habitat Selection in Flour Beetles - Biology

Polyphenism is the phenomenon where two or more distinct phenotypes are produced by the same genotype. Examples of polyphenism provide some of the most compelling systems for the study of epigenetics. Polyphenisms are a major reason for the success of the insects, allowing them to partition life history stages (with larvae dedicated to feeding and growth, and adults dedicated to reproduction and dispersal), to adopt different phenotypes that best suit predictable environmental changes (seasonal morphs), to cope with temporally heterogeneous environments (dispersal morphs), and to partition labour within social groups (the castes of eusocial insects). We survey the status of research on some of the best known examples of insect polyphenism, in each case considering the environmental cues that trigger shifts in phenotype, the neurochemical and hormonal pathways that mediate the transformation, the molecular genetic and epigenetic mechanisms involved in initiating and maintaining the polyphenism, and the adaptive and life-history significance of the phenomenon. We conclude by highlighting some of the common features of these examples and consider future avenues for research on polyphenism.


Experimental evolution

We used three polyandrous (P A, P B, P C) and three monogamous (M A, M B, M C) selection lines of Tribolium castaneum flour beetles (see [44] for details) after 39 non-overlapping generations of selection. M-lines evolved in absence of sexual selection and conflict, as one mate was randomly assigned to single females (i.e. 20 pairs estimated N e = 40), whereas in P-lines sexual selection and conflict were present as single females were housed together with five males (i.e. 12 groups estimated N e = 40). Beetles were able to reproduce for seven to ten days in separate 5 cm Petri dishes with ca. 10 g flour-yeast mix (organic white wheat flour with 10% brewers yeast). Then the adults were removed and offspring were pooled within lines with ample flour-yeast mix to avoid crowding. Pupae were collected randomly from these pools to separate males and females for the next generation (started when beetles were at least ten days old). Our standard rearing temperature was 30°C. T. castaneum is commonly fed with flour supplemented with yeast, because white wheat flour does not provide sufficient amounts of certain necessary amino acids, and productivity is greatly reduced without yeast [51].

Reproductive success (RS) of between line crosses in two different environments

To ensure virginity, all animals were separated by sex as pupae (generation 39). Single sex groups (ca. 20 beetles) had access to flour-mix ad libitum (10 g per 5 cm Petri dish). Equal numbers of pupae were placed in one of two different food treatments. We used flour-mix with 10% yeast as our standard quality food treatment (= benign environment). In contrast, low quality food contained only 1% yeast (= stressful environment). After reaching maturity (all beetles >10 d post emergence) we crossed M- and P-lines within and between sexual selection regimes, without crossing individuals from the same line. Only the ♂ × ♀ crosses indicated in Figure 1 were performed, so that each line was only used once per cross type. For example, line MA females were used in a single M × M cross and a single M × P cross, rather than being used in all possible combinations. Sample sizes are shown in Figure 1. Pairs were used in order to allow us to generate comparable individual fitness measures across treatment combinations. The pairs were allowed to mate and lay eggs in Petri dishes (5 cm) containing ca. 10 g flour-yeast mix (1% or 10% yeast) for 14 days. To avoid crowding of the larvae, the pairs were transferred to a new Petri dish with fresh flour-yeast mix for a further 14 days (again ca. 10 g with the same yeast content as in the first period according to treatment). Adults were removed after four weeks in total, and RS was measured as the total number of offspring produced over this period. We deliberately chose a period of four weeks as a compromise between a time window close to the selection period (seven to ten days) and the natural average life span of this species (several months). Of a total of 447 pairs in the experiment, eleven pairs had no offspring, which were distributed across all crosses and treatments.

Performed crosses. The numbers in the shaded boxes indicate the pairs per cross among selection lines: left = low food quality treatment (1% yeast stressful environment), right = standard food quality treatment (10% yeast benign environment).

Data analysis

We analysed the influence of selection history of males and females (M vs. P) and food quality (low vs. standard) on RS with a linear mixed model using the lme function (nlme package) in R (version 2.13.0, R Development Core Team 2011). The explanatory variables included male selection history, female selection history, food quality and all possible interactions. Furthermore, we included a random factor cross (three population crosses for each ♂ × ♀-type (M × M, M × P, P × M or P × P), i.e. 12 different crosses, see Figure 1). The analysis of RS was performed on the full dataset including all pairs and on a reduced dataset excluding pairs, which did not produce offspring. Qualitatively these alternative analyses were equal, as the same factors and interactions were significant, and significance increased when the eleven pairs with no offspring were excluded. Thus, in the interests of being conservative and, because a lack of offspring may be biologically meaningful, results based on all pairs are shown. The residuals of the presented model were inspected visually and were normally distributed.

Journal publishes doctoral candidate's findings on beetle promiscuity

Elizabeth Droge-Young has long been fascinated by the mysteries and motivations behind sexual selection. But the promiscuity among females of one particular species--the red flour beetle--had her particularly stumped. These beetles would mate multiple times over the course of a day, sometimes multiple times an hour. Were they getting something more out of the mating process than the sperm they needed to reproduce? If so, what?

A doctoral candidate in biology at the College of Arts and Sciences, Droge-Young believes she has found some answers, and her findings have been published online by Behavioral Ecology, the official journal of the International Society for Behavioral Ecology.

"Theory suggests that there should be an optimal level of mating for females, after which point they should no longer be interested in mating," says Droge-Young. "But that doesn't seem to be the case in these beetles, even though they get enough sperm in a single mating to lay fertilized eggs for many months. So mating multiple times an hour can't just be to get sperm."

Her research focused on four possibilities: that mating benefits the female beetles by providing them with moisture with nutrients in the ejaculate with proteins that support egg laying or with additional sperm. The findings led her to conclude that it was the need for additional moisture that fed the beetles' drive to mate so frequently--even to the point where they would sometimes coerce a reluctant male.

"That was my favorite hypothesis," Droge-Young admits. "It just makes sense intuitively. Thirsty females in dry conditions benefit from a moist ejaculate." The red flour beetle, she says, originated from humid areas around India, but today the beetles often live in a much drier environment--grain storage facilities. The hydration provided through mating, she says, serves to enhance reproductive success in low-humidity environments.

The benefits do not appear to be reciprocal. While female fitness increased with frequent mating, the research showed that male longevity significantly declined with increasing exposure to the females.

Droge-Young, who has a B.S. in biological science from Colorado State University and expects to complete her Ph.D. this spring, traces her fascination with sexual selection back to her childhood. "I grew up on Mutual of Omaha's Wild America and nature specials and was always captivated by the bizarre traits and behaviors that males used to get females to mate with them," she says. "During my undergrad work, I was introduced to a bit more of the theory behind sexual selection and was fascinated by different reasons why a female might choose to mate with particular males--and how that choice might play into the quality of offspring she produces."

Droge-Young says her findings underscore the important role environment can play in species' mating habits. "Few studies on reproduction patterns consider that factor," she says. "I hope this study inspires others to consider how changes in the environment can influence how mating systems work."

While she is listed as the principal investigator for the study--titled "Extreme ecology and mating system: discriminating among direct benefits models in red flour beetles"--Droge-Young credits the contributions of, and lists as co-authors on the paper, her advisor, biology professor Scott Pitnick biology professor John Belote and undergraduate research assistant Anjalika Eeswara.

To read the complete article on her findings, click here:

Disclaimer: AAAS and EurekAlert! are not responsible for the accuracy of news releases posted to EurekAlert! by contributing institutions or for the use of any information through the EurekAlert system.

Supporting information

S1 Fig. Similarity of the assembled transcript to public sequences.

(A—C) Similarity search to NCBI nr database. The proportion of assembled transcripts that show the highest similarity to each group is indicated in phylum (A), order (B) and species (C) levels. (D and E) Similarity search to the fruit fly Drosophila melanogaster (D) and the red flour beetle Tribolium castaneum (E) sequences in OrthoDB5. The number of transcripts or genes belonging to each section and the percentage of T. dichotomus transcripts that have putative orthologous genes in OrthoDB5 database are indicated. (F) Results of BUSCO analysis against either Metazoa or Insecta database are shown.

S2 Fig. Count data is distinctly distributed among samples.

A multi-dimentional scaling plot of count data.

S3 Fig. Developmental genes are differentially expressed between head horn and thoracic horn.

Enriched GO terms in each data set comparison that clustered based on semantic similarity of GO terms using REVIGO. The size of each point reflects the number of genes assigned to a GO term. Color indicates enrichment false discovery rate (FDR) in GO enrichment analysis using ErmineJ. GO terms enriched at FDR < 0.5 for each comparison in ErmineJ anlaysis are plotted. Descriptions of 10 GO terms that have the lowest FDR are shown on each plot.

S4 Fig. Molecular phylogeny of SP family genes.

A phylogenetic tree (A) and multiple amino acid sequence alignment (B) of SP family genes in representative insects. The number at nodes indicate boot strap values in A. Amino acid identity and similarity among sequences are indicated by asterisks and dots, respectively.

S5 Fig. Sox14 RNAi beetle showed defects in overall segment formation.

Dorsal view of adult male beetles. Scale bar is 1 cm.

S6 Fig. RNAi for BarH1, Sox21b, dac, Optix and Tbx20 caused no reduction in horn length.

Lateral view of head and prothorax in RNAi beetles are displayed. Horn length and body length are plotted, and linear regression lines are drawn for each gene in red. Gray dots and gray regression lines are for EGFP RNAi beetles. P-values show significance of each RNAi treatment using Wald test on logistic linear regression.

S7 Fig. RNAi caused differential effects on head, but similar effects on appendage between Tribolium castaneum and Trypoxylus dichotomus.

(A—D) Dorsal view of T. castaneum adult heads after RNAi treatments. (E) A box plot shows decreased number of ommatidium in Optix RNAi beetle compared to control. (F) Dorsal view of an SP8 RNAi beetle. Arrowheads in F indicate malformed antennae in SP8 RNAi beetles. (G and H) Frontal view of adult heads in EGFP (G) and SP8 (H) RNAi beetles. Arrowheads point to corresponding regions between G and H, which is fused by SP8 RNAi treatment (H). The dorsal part of the labrum is in red. (I) Adult metathoracic legs in T. castaneum after RNAi treatments. Scale bar is 0.1 mm in A—D, F—H, 0.5 mm in I.

S8 Fig. Comparison of mechanisms regulating head and thoracic horn growth in Trypoxylus dichotomus and Onthophagus spp.

Both thoracic (A) and head (B) horns appear to have arisen through the repeated evolutionary co-option of parallel mechanistic processes, including the pathway regulating sexually dimorphic amounts of weapon growth (alternative splice forms of doublesex), the embryonic locations of horn outgrowth, and the co-option of portions of the insect appendage patterning pathway. However, the specific genes within the patterning pathway with the most pronounced effects on horn size differ somewhat between rhinoceros and dung beetle horns, consistent with their presumed independent evolutionary origins. Results summarized from [2], [5–6], [9–10], [21], [49].