The impact of adult weight, age, and density on reproduction of Tenebrio molitor L. (Coleoptera: Tenebrionidae) was studied. The impact of adult weight on reproduction was determined by: (1) counting the daily progeny of individual adult pairs of known weight and analyzing the data with linear regression and (2) creating 5 weight classes of 10-mg intervals starting at 60 mg (60 - 69.9mg) and ending at 100 mg (100 - 109.9 mg), then the progeny of 10 groups of 5 males and 5 females of each weight class were compared using ANOVA. To determine the impact of adult density on reproduction, adults were grouped at 8 different densities by increasing numbers per box (at 1:1 sex ratio). Weekly progeny production of 8 groups per density treatment was compared using ANOVA. There was no significant relationship between female weight and progeny production in the individual pair analysis. Fecundity was significantly different among weight classes, but the relationship was not linear. Adult density had a significant impact on progeny per female and progeny per unit area. Reproductive output per female decreased as adult density increased. Progeny per unit area increased to a maximum at a density of 14 adults/dm2 and then declined sharply. Adult age had a significant impact on reproduction. The highest reproductive output occurred at 2 and 3 wk of age and was significantly higher than at any other age. Adult density and age may be manipulated to maximize production of T. molitor larvae.
The yellow mealworm, Tenebrio molitor L. (Coleoptera: Tenebrionidae), is mass produced and sold in the United States for a variety of purposes. The larvae of T. molitor are one of the most common foods for captive mammals, birds, reptiles, and amphibians because they are easy to propagate, harvest, and feed (Martin et al. 1976, Barker et al. 1998, Finke 2002). Tenebrio molitor has also been proposed as a source of protein for catfish (Ng et al. 2001), chicken production (Ramos-Elorduy et al. 2003, Klasing et al. 2000), and even for human consumption in a variety of forms (Ramos-Elorduy 1997, DeFoliart 1999, Aguilar-Miranda et al. 2002).
Some applications of T. molitor in biological control have been suggested as factitious prey for heteropteran predators. Some species reared using T. molitor include: Perillus bioculatus (F.) (Pentatomidae) (Saint-Cyr and Cloutier 1996), Podisus maculiventris (Say) (Pentatomidae) (De Clercq et al. 1998), P. mucronatus Uhler (Costello et al. 2002), P. nigrispinus (Dallas) (Lemos et al. 2003, De Bortoli et al. 2011), and Pristhesancus plagipennis (Walker) (Reduviidae) (Grundy et al. 2000). Tenebrio molitor also has been used with some success to rear other predators such as Dichochrysa prasina Burmeister (Neuroptera: Chrysopidae) (Pappas et al. 2007).
Another application of T. molitor in biological control is as a host for in vivo mass production of entomopathogenic nematodes (Shapiro-Ilan et al. 2002, 2008). In some cases, or for certain species, the quality (e.g., virulence or efficacy) of in vivo-produced nematodes may be superior to those produced in vitro (Gaugler and Georgis 1991, Shapiro and Lewis 1999, Cottrell et al. 2011). The main limitation of in vivo mass production of entomopathogenic nematodes is cost. One way to reduce costs is by optimizing the rearing system of the host, in this case T. molitor.
Indeed, increasing the efficiency of T. molitor production would provide benefits to the various applications that use the insect. In this study, we focused on optimizing reproductive output. Maximizing the reproductive output of T. molitor could improve the capabilities of the rearing system by producing more beetles per unit of rearing space. Our specific objectives were to study the effects of female weight, adult density and female age on progeny production by T. molitor.
Materials and Methods
Tenebrio molitor rearing. A new system for rearing T. molitor was used to maintain a stock colony at the USDA-ARS NBCL, Stoneville, MS. The T. molitor colony was started from stock donated by Southeastern Insectaries, Inc. (Perry, GA) in 2005.
First instars were collected weekly using a modified two-tray stacked system divided by nylon screen (system I) (Fig. 1). The top tray measuring 29 × 55 × 14.5 cm housed 500 adult beetles and 1 kg of wheat bran as food and oviposition substrate. The top tray was modified by cutting out the fiberglass bottom and replacing it with standard No. 20 (with 850 μm openings) nylon screen and by adding 3 circular screened windows (5.5 cm diam.) along both of the long sides of the tray (Fig. 1). Adults oviposited and glued the eggs to the bran flakes and sides of the top tray. Eclosed first instars fell through the screen openings to the bottom tray. The bottom tray was unmodified and measured 29 × 55 × 8 cm. Adults were provided with approx. 50 ml of water twice a week.
First instars were collected weekly from the bottom tray and placed in plastic boxes (950 ml, 220 × 150 × 52 mm) where they remained for a period of 30 d to allow them to grow larger than the screen openings of the next set of trays. After 30 d larval production of 2 wk was transferred to a fiber glass tray 41 × 62.5 × 15 cm modified by replacing the bottom with nylon screen (standard No. 35 with 500 μm openings) and adding 3 screened windows (5.5 cm diam.) in each of the longer sides. The 500 μm mesh allowed frass particles to fall through each of 4 stacked trays to a lower unmodified frass collecting tray, maintaining a cleaner environment (Fig. 2). Continuous movement of the larvae in screened trays caused separation of the frass particles from the bran flakes. Larvae in screened trays were provided with 1 kg of wheat bran per tray, which was replenished as needed. No water was provided to the larvae.
Tenebrio molitor development time varies greatly due to their developmental plasticity and instar variation (Morales-Ramos et al. 2010). As a consequence, larvae of multiple instars coexist within each cohort. Pupating last-instar larvae must be separated from other instars to prevent cannibalism by earlier instars. Larvae were separated when they reached 4 months of age using a 3-screen separator (Midwest Industries Inc., Macon, GA) that divided larvae into 4 groups according to their size. The openings of the 3 screens were rectangular and measured 12.7 × 1.854 mm, 12.7 × 1.65 mm, and 12.7 × 1.09 mm. The first group of larvae that could not pass through the larger screen openings (12th instar or older) was transferred to plastic trays (52 × 39.5 × 12 cm) having solid bottoms provided with 200 g of wheat bran and allowed to pupate. The second and third groups of smaller larvae (11th instars and 9 - 10th instars, respectively) were placed in screen bottom trays as described above (grouped by size), provided with 1 kg of wheat bran, and returned to the rearing room to continue development.
Pupae were collected daily from the largest larvae trays by using a sifter standard No. 8 (3.5 mm openings). Pupae were allowed to complete development in a plastic tray with a solid bottom lined with tissue paper. Emerging adults were collected daily and separated into groups of the same age. Three to 4 days after emerging, dark-colored adult beetles were transferred to oviposition boxes as described above. The dark color of the adult beetles is an indicator of age and is usually correlated with the onset of reproductive readiness (J. A. Morales-Ramos, unpubl. data).
Adult weight and fecundity. Two experiments were designed to test the effect of adult weight on fecundity. The first experiment recorded the fecundity of individual females of known weight to measure the relationship between weight and fecundity. In the second experiment, fecundity was compared among 5 different adult weight classes.
In the first experiment, a total of 24 pairs (male and female) of adult T. molitor were placed individually in modified stacked Petri dishes (58 × 20 mm). Top and bottom dishes were separated by a standard No. 20 screen (850 μm openings) to allow first instars, eclosing from eggs glued by the females to the substrate, to fall from the top to the bottom dish. The weight of each female was recorded using a precision balance (Mettler Toledo PB303-S/FACT, Switzerland). Pairs were placed in an environmental chamber at 28 ± 1°C, 75 ± 5% RH and continuous darkness. Each pair was provided once with 2 g of wheat bran and 25 μL of reverse osmosis (RO) water twice a week. The progeny produced per female was determined by counting and recording the number of fallen first instars for a period of 70 days.
In the second experiment, newly-emerged male and female T. molitor adults from the stock colony were individually weighed and grouped in 5 different weight classes (60 - 69, 70 - 79, 80 - 89, 90 - 99, and 100 - 109 mg). Groups of 10 adults (1:1 sex ratio) from each weight class were randomly selected, and each group was placed in stacked plastic boxes (140 × 103 × 36 mm) separated by No. 20 screen to allow first instars to fall through as described above. Boxes were placed in an environmental chamber at the same conditions described above. Each weight treatment consisted of 10 groups (replicates) of 10 adults each. Each group was provided once with 15 mg of wheat bran each and 125 μL of RO water twice a week. Each week first instars from each group were collected from the bottom box of each stack, counted, and recorded for a 20 wk period.
Age-dependent fecundity. Pooled data from all treatments of experiment 2 were used to determine age-dependent fecundity. Weekly progeny production per group for a period of 20 wk was recorded and used to calculate mean age-dependent weekly progeny production per female. Weeks 1 - 20 were each designated as individual age classes. Adult mortality was recorded daily, and dead adults were sexed to determine the number of living adult females at the end of each age class.
Cumulative progeny produced during each age class was used to calculate the percentage of potential reproduction (PR) at each age class as:
The total progeny produced at the end of week 20 was used as the 100% mark to calculate the increasing percentage of cumulative reproduction at each age class. The curve of cumulative reproduction percentage (percent realized reproduction) was compared with the adult percent survival curve to determine the optimal reproductive age range for mass production.
Adult density and fecundity. The density of adults was varied to create density treatments by adding an increasing number of adults at a 1:1 sex ratio to plastic boxes of constant dimensions (140 × 103 × 36 mm). Density was measured as the number of adults per area in dm2. The total surface area in each box was 1.428 dm2. The volume of food provided to the adult beetles increased in proportion to the number of adults per box (with one exception), but the surface where the beetles could walk remained constant.
In a third experiment, newly-emerged T. molitor adults from the stock colony were randomly selected and placed in stacked plastic boxes divided by No. 20 nylon screens as described above. Eight density treatments consisted of 2, 4, 8, 12, 20, 34, 70, and 120 adults (1:1 sex ratio) per box were created (Table 1), and the experiment was replicated 8 times. All the adult groups inside stacked boxes were placed in an environmental chamber at the same conditions described above. Groups were provided initially with 1 g of wheat bran per adult beetle except for those in treatment 8, which were provided with a total of 70 g per group due to the lack of space. Wheat bran was added to treatment 8 boxes as the beetles consumed the food available. Boxes also were provided with the equivalent of 25 μL of RO water per adult twice a week except for treatments 7 and 8, which were provided with a total of 1.5 ml of water per box. Water volume was reduced in treatments 7 and 8 to prevent uncontrollable growth of fungi on the wheat bran within the boxes. First instars falling through the separating screen to the bottom box were collected weekly, counted, and recorded for a period of 4 weeks.
Data analyses. The impact of adult weight on fecundity in experiment 1 was analyzed using linear regression analysis (SAS Institute 2007). Experiment 2 data were analyzed using ANOVA and means of progeny produced per adult group were compared among weight classes using the Tukey-Kramer HSD test. In experiment 3, weekly progeny production was compared among density treatments using ANOVA, and means were compared using the Tukey-Kramer HSD test. Similarly, ANOVA was used to analyze progeny produced in the different age classes (wks) and means were compared using the Tukey-Kramer HSD test. General linear model (GLM) was used to analyze progeny produced per female per wk among weight classes. Weight classes were compared as nominal variables in the linear model and age in weeks was included in the model as a numerical variable to account for age effects.
Results and Discussion
Adult weight and fecundity. Results from experiment 1 showed no evidence that female weight impacts fecundity. Linear regression analysis showed no significant correlation of weight with progeny production per female (R2 = 0.01; F = 0.23; df = 1, 23; P = 0.631). This indicates that the weight of the females does not appear to have a linear effect on their reproductive potential.
Results of experiment 2 were consistent with those obtained in experiment 1. Results from the ANOVA showed no significant difference in total progeny production per box among weight classes for the entire experimental period (F = 1.76; df = 4, 45; P = 0.1526). Also, weekly progeny production per box among the 5 different adult weight classes were not significantly different (F = 0.6; df = 4, 967; P= 0.662). Similar results were obtained when the weekly progeny production per female was compared among weight classes (F = 0.478; df = 4, 967; P = 0.751). These results confirm the observations of experiment 1 showing no effect of adult weight on progeny production.
Significant differences among weight classes were observed when data of progeny production per female per week were analyzed using GLM and including female age (in weeks) in the model (F= 755.65; df = 5, 966; P < 0.00001) (Fig. 3). Most of the variability of the data was explained by female age (F ratio = 3,768.885, P < 0.00001), but weight class showed some significant impact on the dependent variable (progeny/female/wk) (F ratio = 5.789, P = 0.0001). Least squares means comparisons showed that progeny / female / wk in weight classes 100 and 70 (14.25 ± 0.37 and 13.95 ± 0.37 first instars / female / wk, respectively) was significantly higher than that in weight classes 60 and 90 (12.39 ± 0.38 and 12.26 ± 0.37 first instars / female / wk, respectively) (Fig. 3). Progeny in weight class 80 (13.19 ± 0.37 first instars / female / wk) was not significantly different from any other weight class. Differences in the results from ANOVA and GLM analyses are explained by the account of female age variability in the GLM model. Although the results of the GLM analysis could mean that female weight has an effect on progeny production, this effect does not show a linear pattern. Lack of fit of the model was significant (F = 7.81; df = 94, 874; P < 0.0001) indicating that the relationship among the variables was not linear. Results from the GLM analysis seem to contradict earlier results, but without a consistent linear relationship between weight and progeny production the predictive value of female weight is lost.
The apparent lack of a linear and positive female weight impact on the progeny production of T. molitor was unexpected. Female weight has been shown to have a positive linear effect on fecundity in a variety of insect species from different orders. Numerous examples of insects in which female weight significantly impact fecundity include: Peregrinus maidis (Ashmead) (Homoptera: Delphacidae) (Wang et al. 2006), Magarcys signata (Hagen) (Plecoptera: Perlodidae) (Taylor et al. 1998), Aphodius ater DeGeer (Coleoptera: Scarabaeidae) (Hirschberger 1999), Carcinops pumpilo (Erichson) (Coleoptera: Histeridae) (Achiano and Giliomee 2004), Callosobruchus maculatus (F.) (Coleoptera: Bruchidae) (Colegrave 1993), Tomicus piniperda (L.) (Coleoptera: Scolytidae) (Amezaga and Garbisu 2000), Antheraea mylitta (Drury) (Lepidoptera: Saturnidae) (Rath 2005), Chilo partellus Swinhoe (Lepidoptera: Pyralidae) (Ochieng-Odero et al. 1994), Streblote panda Habner (Lepidoptra: Lasiocampidae) (Calvo and Molina 2005), Quadricalcarifera punctatella (Motschulsky) (Lepidoptera: Notodontidae) (Kamata and Igarashi 1995), Bombyx mori L. (Lepidoptera: Bombycidae) (Saha and Bajpai 2009), Homotriaxa alleni Barraclough (Diptera: Tachinidae) (Allen and Hunt 2001), Trichopoda giacomellii (Blanchard) (Diptera: Tachinidae) (Coombs 1997), Sirex nitobei Matsumura (Hymenoptera: Siricidae) (Fukuda et al. 1993), Tetrastichus incertus (Ratzeburg) (Hymenoptera: Eulophidae) (Pitcairn and Gutierrez 1992), and Catolaccus grandis (Burks) (Hymenoptera: Pteromalidae) (Greenberg et al. 1995).
Pupal weight in T. molitor is correlated with the number of stadia it takes to complete larval development (J. A. Morales-Ramos, unpubl. data) and the number of stadia is correlated with development time (Morales-Ramos et al. 2010). Thereby, adult weight is affected by development time in T. molitor. It is reasonable to assume that under constant conditions of nutrient availability, larval density, and controlled environmental factors (such as in this study), adult weight could be genetically determined. However, there is evidence that adult size in T. molitor is significantly affected by environmental factors such as moisture (Urs and Hopkins 1973), nutrition (Morales-Ramos et al. 2010), oxygen content (Greenberg and Ar 1996) and density during development (Weaver and McFarlane 1990). To determine the true effect of adult size on fecundity it is important to separate the genetic from the environmental effects causing adult weight variability. Future research will be focused on determining how fecundity is impacted by environmentally induced adult weight variation.
Age-dependent fecundity. Females reached their peak reproductive potential in the second week after emergence and remained at the highest level through week 3. Progeny production during weeks 2 and 3 was significantly higher than progeny production at any other age class (F = 386.34; df = 19, 952; P < 0.00001) (Fig. 4). After week 3, progeny production declined steadily from 33.93 ± 0.8 - 4.0 ± 0.44 at 14 wks of age. After week 14, progeny production remained at a very low level (< 4 first instars per female/wk) and continued for 6 more weeks (Fig. 4).
Females were fertile for a period of 20 wk, but continued to live for another 10 wk. Female 20-wk oviposition potential reached 91.15% after 11 wk. Adult survival at this age was 93.28% (Fig. 5). Adult mortality reached 50% at 135 d or 19.3 wk. At this age females were producing almost no progeny. This information can be applied to colony optimization practices. Adults selected for reproduction could be replaced every 58 - 74 d, when they reach 80 - 90% of their oviposition potential (Fig. 5 bar), to maximize the production of larvae for nematode infection. However, adult replacement could be initiated a week earlier based on progeny production per female/wk. Mean progeny production during weeks 8, 9, and 10 (50 - 70 d of age) was 14.9 ± 0.62, 13.35 ± 0.55, and 11.33 ± 0.66 first instars per female/wk, respectively. Progeny production during weeks 8 - 10 was significantly lower than that observed during week 7 (18.43 ± 0.67 first instars per female/wk) (Fig. 4). Replacing adults at 8 wk (50 d) of age could improve the colony's progeny production significantly because progeny production during the first week was significantly higher at 25.7 ± 0.79 first instars per female/wk (Fig. 4).
Adult density and fecundity. Adult density had a significant effect on progeny production (F = 44.99; df = 7, 248; P < 0.0001). Total progeny production per group was significantly higher in treatments 4 and 5 (8.4 and 14 adults / dm2, respectively) than in any of the other density treatments (Fig. 6 A). However, mean progeny production per female was significantly higher in treatments 1 and 2 (1.4 and 2.8 adults / dm2, respectively) as compared with treatments 4 and 5 (Fig. 6 B). Progeny production of individual females continuously declined as adult density increased, whereas progeny production per box increased to a maximum and then declined (Fig. 6).
Adult density had a negative impact on progeny production per female in T. molitor, and this is consistent with reports in other beetle species. Increasing adult density in Callosobruchus rhodesianus (Pic) and C. maculatus significantly reduced fecundity (Giga and Smith 1991). Fecundity of the histerid beetle Carcinops pumpilo (Erichson) decreased with increased adult density (Achiano and Giliomee 2004). Increasing adult density significantly reduced fecundity in Tribolium castaneum (Herbst.) (Tenebrionidae) (Longstaff 1995). The negative impact of adult high density has been identified as a detrimental factor in the culturing of some insect species. Fecundity of the coffee berry borer, Hypothenemus hampei (Ferrari) (Scolytidae), reared in artificial diet decreased significantly as the number of females per vial increased (Vega et al. 2011). Fecundity was not affected by adult density in Euscepes postfasciatus (Fairamaire) (Curculionidae), but increased egg cannibalism reduced egg production significantly in mass rearing conditions (Kuriwada et al. 2009). Increased egg cannibalism associated with increased adult density may be the cause of reduced progeny production in T. molitor. Egg cannibalism by adult beetles has been observed in T. molitor on several occasions during this study (J. A. Morales-Ramos, unpubl. data). Longstaff (1995) suggested that egg cannibalism may have been the major factor reducing egg production as adult density increases in T. castaneum.
Adult density is a factor impacting progeny production that can be easily manipulated to optimize egg production in a mass rearing facility. In a commercial production of T. molitor, egg production should be maximized whereas rearing space and labor should be minimized. Our data show that single couples are the most productive in terms of progeny produced per female: however, isolated pairs require more space and labor for egg collection, watering, and feeding. The optimal situation for mass production is to maximize progeny production per unit of rearing area even if progeny production per female is not the highest. Our results indicate that the optimal reproductive output occurs at adult densities between 8.4 and 14 adults per dm2 (Treatments 4 and 5). Progeny production per box was not significantly different between treatments 4 and 5, but progeny production per female was significantly higher in treatment 4. Thus, we conclude that a density of 8.4 adults / dm2 as depicted in treatment 4 is the most efficient for mass-production purposes. In theory, the optimal adult density of 8.4 adults per dm2 of rearing space can be extrapolated to any tray size available.
The authors thank the USDA-National Institute of Food and Agriculture (NIFA) for financing a portion of this research through the Small Business Innovation Research (SBIR) program (grant No. 2007-33610-18416/proposal No. 2007-03695). We also thank Scott Lee for producing the drawings for Figures 1 and 2.
3Southeastern Insectaries, Inc., Perry, GA 31069.
4USDA-ARS Southeastern Fruit & Nut Tree Research Lab, Byron, GA 31008.