Developmental Plasticity in Tenebrio molitor (Coleoptera: Tenebrionidae): Analysis of Instar Variation in Number and Development Time under Different Diets
The variation in instar number and the pattern of sequential instar development time of Tenebrio molitor L. (Coleoptera: Tenebrionidae) was studied under 4 different diet regimes. Addition of dietary supplements consisting of dry potato or a mix of dry potato and dry egg whites significantly reduced the number of larval instars and total development time. The pattern of sequential stadia length showed regularity and low variability in instars 5 - 9. Stadia length continuously increased between instar 10 and the last instar before pupation (P−1). Instar number significantly impacted total development time and was also significantly correlated with stadia length. The length of the 1st, 2nd, 3rd, and 4th stadia was not significantly affected by diet indicating that they may represent a stable part of the life cycle of T. molitor. The length of the last stadium (P−1) was regular within food treatments regardless of sequential instar number, but P−1 was also significantly affected by nutrition in a similar way as all the sequential stadia. The pattern of instar variation is discussed in regard to the insertion of new instars based on variation of stadia length.
The yellow mealworm, Tenebrio molitor L. (Coleoptera: Tenebrionidae), is a commonly mass-produced beetle that is currently 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 (Klasing et al. 2000). Tenebrio molitor also has been proposed as a source of protein for catfish (Ng et al. 2001) and broiler chicken production (Ramos-Elorduy et al. 2003). Additionally, T. molitor can be used as a host for in vivo mass production of entomopathogenic nematodes (Shapiro-Ilan et al. 2002, 2008).
Tenebrio molitor displays plasticity in the number of instars needed for its development (Esperk et al. 2007). The number of instars can vary from 9 to more than 20 (Cotton and St. George 1929) depending on several factors including temperature (Ludwig 1956), humidity (Murray 1968, Urs and Hopkins 1973), photoperiod (Tyshchenko and Sheyk Ba 1986), oxygen concentration (Loudon 1988, Greenberg and Ar 1996), population density (Connat et al. 1991), parental age (Ludwig 1956, Ludwig and Flore 1960), and food quality (Stellwaag-Kittler 1954). The number of instars in T. molitor increases in response to adverse conditions (Esperk et al. 2007).
Larval development time is reported to be affected by environmental factors in a similar way as the number of instars (Ludwig 1956, Murray 1968, Urs and Hopkins 1973, Tyshchenko and Sheyk Ba 1986, Loudon 1988). Larval development time is probably correlated with the number of instars in T. molitor, but this has not been statistically determined or specifically addressed by previous research. Also, information on development time of each instar and the impact of instar number variation on stadia length are lacking. It has been reported that nutrition has an effect on instar number and total development time (Stellwaag-Kittler 1954), but the effect of specific nutrients on instar number and stadia length has not been fully explored. Davis (1969, 1971a, 1971b, 1974, 1975, 1978) conducted extensive studies on the protein and amino acid requirements of T. molitor, but his studies reported the effects of nutrition on growth rather than the developmental plasticity of this insect. The objectives of this study were to determine the development time of each instar of T. molitor, to induce variation in number of instars and development time by supplementing the diet, and to use the observed variation to identify characteristic patterns of development using regression analysis.
Materials and Methods
The rearing stock of T. molitor was provided by Southeastern Insectaries, Inc. (Perry, GA). The mother colony was maintained at 27 ± 2°C and 50% RH on a supplement-free diet of wheat bran. Water was provided twice a week in the form of saturated polyacrylamide crystals, which was presented to adults and larvae by directly mixing the crystals in the wheat bran. Adult beetles were reared in 40 × 30 × 20 cm fiberglass pans. These pans were replaced every 2 wks to recover the eggs that adult beetles had glued to the bottom of the pan. Fresh wheat bran was added to the pans with eggs as food for hatching larvae. Eggs and subsequent hatching larvae were allowed to develop in the same pan under the same conditions until pupal stage was reached approximately 3 months later. Wheat bran was added as needed, and water was provided twice a week as described above. Pupae were separated from the diet by the use of a standard No. 6 sieve (3.35 mm openings) and allowed to complete development in plastic boxes (295 × 190 × 80 mm, L × W × H) (Rubbermaid®, Servin' Saver # 7, Rubbermaid Inc., Atlanta, GA) lined with tissue paper.
Experimental design and data collection. Tenebrio molitor pupae from the mother colony were selected, sexed and allowed to complete development. Eight groups of 100 adult beetles, half of them females, were each placed in plastic boxes (110 × 110 × 35 mm). Boxes with adult beetles were divided into 4 treatment groups receiving different food supplements or no supplements (control). Three food supplements included (1) dry potato (Idahoan®, Idahoan Foods, Lewisville, ID), (2) 1:9 mix of dry egg white (Just Whites®, Deb-EI Foods Corporation, Elizabeth, NJ) and dry potato, and (3) 1:4 mix of dry egg white and dry potato (hereafter referred to as diet 1, 2, and 3, respectively). Dry food supplements were mixed with wheat bran in a 1:4 proportion (20% supplements) to prepare the 3 diet treatments. The supplement component of diet 1 (dry potato) was mixed in the form of flakes directly with the wheat bran without any processing. Supplement components of diets 2 and 3 were mixed with deionized water and then dried in a vacuum oven at 58°C and −50.84 cm of mercury and ground (particle size 0.5 - 2.0 mm) before mixing with the wheat bran. Diet treatments had relatively small differences in nutritional content, and all of them allowed development of T. molitor (Table 1). These formulations were designed to induce variation in development time and number of instars without affecting immature survival. Some important differences among the food treatments included the content of starch and vitamin C, which were absent in the wheat bran but present in the dry potato and the content of sodium, which was relatively high in the dry egg white resulting in 20, 220, 460, and 690 mg per g in the control, and diets 1, 2, and 3, respectively.
Each treatment and control consisted of 2 boxes with adult beetles. Each box was provided daily with 10 g of diet of the corresponding treatment and kept in an environmental chamber at 27 ± 1°C, 50 ± 5% RH, and 14:10 h (L:D) photoperiod. Adult beetles were allowed to oviposit in the bottom of the boxes for a period of 7 d. Every week, adult beetles were removed from the boxes and introduced to new boxes containing 10 g of new diet of the corresponding treatment; the old diet was discarded.
Boxes containing eggs glued to the bottom were dated and placed in the same environmental chamber to allow the development of eggs at the same environmental conditions. Boxes were monitored daily for hatching of first instars. First instars were collected and grouped according to the collection date and parental diet treatment in small Petri dishes (35 × 12 mm), provided with 100 mg of the corresponding diet treatment, and allowed to continue development at the same conditions. These groups of between 100 - 300 first instars were monitored daily to detect changes of instar. Second instars were easily distinguishable from first instars by their dark yellow coloration, which contrasted with the white coloration of first instars. Second instars were collected and held individually in new small Petri dishes. Each second instar was assigned a number, provided with 100 mg of diet of the corresponding treatment and allowed to develop under the same conditions. Each larva also was provided daily with 50 μL of de-ionized water dispensed in the diet by a 1-ml syringe.
Larvae were monitored daily for the presence of the exuvium that signaled the change of instar. The collection of second instars continued until 72 third instars were obtained for each treatment. This was done because mortality occurring during the first and second stadia was disproportionally high as compared with mortality occurring during older stadia. The total number of instars and the duration of each stadium were recorded for each larva until pupation. Pupae were sexed and weighed using a precisions balance (Mettler Toledo PB303-S, Switzerland) and allowed to complete development under the same environmental conditions.
Data analysis. Data consisting of number of instars, developmental time from first instar to adult, duration of stadia, and pupal weight were compared among treatments by analysis of variance (ANOVA) and Tukey-Kramer's HSD test. The number of observations varied among treatments (between 49 and 56) depending on the survival of immature stages in each treatment. Survival from first instar to adult and survival of each stage and instar were compared among treatments by Z-test. Linear regression analysis was used to determine the effect of number of instars on total development time. Regression lines of instars versus developmental time from the different food treatments were compared for differences in slope and intercept using the F statistic method for multiple lines comparisons (Zar 1999).
The pattern of stadia length among different instars was modeled with linear and quadratic regressions using 2 methods: (1) by sequential instar number (1st, 2nd, …kth) versus stadia length, and (2) inversely by instar previous to pupation versus stadia length. Stadia length patterns obtained with these 2 methods were graphically compared to look for consistencies in stadia length. Instars previous to pupation were named as P−1 for the last before pupation, P−2 for the instar 2 molts away from pupations, and P−k for the instar “k” molts away from pupation. A multiple linear regression model was used to test for interactions between sequential instar development time and the length of stadia previous to pupation. The presence of interactions would show inconsistency of stadia length analyzed by the 2 methods, which would help to identify the location of changing instar sequences.
Regression analysis also was used to determine if correlation existed between total number of instars and stadia length. This analysis was done instar by instar using methods of analysis 1 and 2. All statistical procedures except the Z-test were done using JMP software release 7 (SAS Institute 2007). The Z-test calculations were done in a Microsoft Excel spreadsheet using the formula reported by Zar (1999).
Results
General food treatment effects. The highest number of instars was 16 and was observed in the control group. The lowest number of instars was 10 and was recorded in the diet 3 treatment. The number of instars observed was significantly different among food treatments (F = 100.39; df 3, 209; P < 0.0001) (Fig. 1A) inducing sufficient instar variation to provide valuable data for the analyses. The control group showed a mean of 13.90 ± 0.68, which was significantly larger than that observed in the 3 diet groups. The number of instars observed in the 3 food treatments was 12.41 ± 0.69, 12.06 ± 0.59, and 11.43 ± 0.61 for diets 1, 2, and 3, respectively. Diet 3 had significantly fewer instars than the other treatments, but the difference in the number of instars between diet 1 and 2 was not statistically significant (Fig. 1A). Individuals completing development through 14 instars were more common in the control group, whereas 11, 12, and 12 - 13 instars were more common in diet 3, diet 2, and diet 1, respectively (Fig. 2).
Citation: Journal of Entomological Science 45, 2; 10.18474/0749-8004-45.2.75
Citation: Journal of Entomological Science 45, 2; 10.18474/0749-8004-45.2.75
Pupal weight was significantly higher in diet 2 (149.38 ± 22.17 mg) than in all the other treatments (132.06 ± 21.75, 134.89 ± 20.26, and 134.84 ± 21.98 mg in the control, diet 1, and diet 3, respectively)(F = 7.22; df = 3, 209; P = 0.0001) (Fig. 1C).
Despite the relatively small differences in nutritional value among the diet treatments (Table 1), diet had a significant effect on the survival rate from first instar to adult. Survival from first instar to adult was significantly higher in the control group than in diet 3, and larvae on diet 1 had significantly better survival than those on diets 2 and 3 (|Z | = 2.79, 2.81, and 4.01; df = 501, 447, and 409; P = 0.0052, 0.005, and < 0.001, respectively). Most of the mortality occurred during the first instar (within the first 5 days after eclosion) resulting in 0.462, 0.547, 0.325, and 0.285 survival rates of second instars for the control, and diets 1, 2, and 3 respectively. Because the adult parents were fed with the same diet treatments, diet may have impacted first instar survival through parental nutrition. The survival rates from 1st instar to adult were 0.266, 0.320, 0.202, and 0.163 for the control and diets 1, 2, and 3, respectively.
Development time was significantly different among all treatment groups (F = 227.62; df = 3, 209; P < 0.0001) (Fig. 1B). Stadium by stadium length comparisons among treatment groups and control revealed significant differences in the 3rd (F = 4.99; df = 3, 209; P = 0.0023), 5th (F = 4.69; df = 3, 209; P = 0.0034), 6th (F = 24.75; df = 3, 209; P < 0.0001), 7th (F =43.18; df = 3, 209; P < 0.0001), 8th (F = 14.01; df = 2, 209; P < 0.0001), 9th (F = 48.33; df = 3, 209; P < 0.0001), 10th (F 32.69; df = 3, 209; P < 0.0001), 11th (F = 17.93; df = 3, 208; P < 0.0001), 12th (F = 6.00; df = 3, 171; P = 0.0007), and 13th (F = 5.99; df = 3, 96; P = 0.0009) stadia (Table 2).
Relationship between number of instars and total development time. There was a significant positive regression impact of the number of instars on the total development time (r2 = 0.87; F = 1410.52; df = 1, 211; P < 0.0001). Linear regression analysis provided an estimated value of −115.03 ± 5.82 for the intercept (a) and 17.366 ± 0.46 for the slope (b). This regression was significant even when the data were analyzed by diet treatment (Table 3) (Fig. 3). The values of the parameters “a” and “b” showed some differences among the diet treatments (Table 3). The parameter “b” (slope) was significantly smaller in the diet 3 regression line after T-test paired comparisons with regression lines of the other 2 diets and the control group. No significant differences were observed in the slope among diets 1 and 2 and the control group. The differences observed in the parameter “a” (intercept) among these last 3 groups were significant (F = 37.81; df = 1, 113; P < 0.0001) and consistent with results from the analysis of variance (Table 3).
Citation: Journal of Entomological Science 45, 2; 10.18474/0749-8004-45.2.75
Pattern of sequential instars and stadia length. The sequential pattern of stadia length was consistent among the 4 food treatments. This pattern showed that the oldest stadia (10 - 15) were longer than the middle stadia (4 - 9). Stadia 1, 2, and 3 lengths showed the most consistent pattern among the food treatments (Fig. 4). Instar effects on stadia length fit a quadratic regression model (F = 1900.96; df = 2, 2666; P < 0.0001; r2 = 0.588) with parameters a = 1.98, b1 = 0.59, and b2 = 0.117.
Citation: Journal of Entomological Science 45, 2; 10.18474/0749-8004-45.2.75
A different pattern was evident when the data were studied backward as remaining molts to pupation instead of sequential instars. In this case, the stadium previous to pupa (P−1) was always the longest in all treatments regardless of the number of instars (Fig. 5). Data analyzed in this manner also fit a quadratic model (F = 1900.96; df = 2, 2666; P < 0.0001; r2 = 0.47) with parameters a = 9.58, b1 = −0.588, and b2 = 0.106.
Citation: Journal of Entomological Science 45, 2; 10.18474/0749-8004-45.2.75
The length of the P−1 stadium was not significantly different among the different instars when analyzed within food treatments (Fig. 6). However, significant differences in developmental time among sequential instars were evident in the length of the P−2 stadium when analyzed within food treatment (F = 3.77, 18.14, 9.88, and 18.37; df1 = 2; df2 = 46, 50, 51, and 47; P = 0.0305, P < 0.0001, P = 0.0002, and P < 0.0001 for control, diet 1, diet 2, and diet 3, respectively) and the P−3 stadium (F = 11.45, 51.28, 20.31, and 22.52; df1 = 2; df2 = 46, 50, 51, and 47; P < 0.0001 for control, diet 1, diet 2, and diet 3, respectively) (Fig. 6). Multiple linear regression analysis showed that stadia length was significantly impacted by sequential instar (F = 806.35; df = 1, 2060; P < 0.0001), molts before pupation (P−k) (F = 18.71; df = 1, 2060; P < 0.0001), and their cross product (F = 2431.99; df = 1, 2060; P < 0.0001), which confirmed the existence of a significant interaction between these 2 independent variables. Food treatment also significantly affected stadia length when analyzed as molts previous to pupation regardless of sequential instar (Table 4).
Citation: Journal of Entomological Science 45, 2; 10.18474/0749-8004-45.2.75
Relationship between number of instars and stadia length. Stadia length was significantly correlated with total number of instars (F = 150.77; df = 1, 2667; P < 0.0001), but the r2 value of 0.054 indicated that the model only explains 5.4% of the variability of stadia length. Adding instar and its quadratic as a second and third independent variable improved the model performance (F = 1287.4; df = 3, 2665; P < 0.0001; r2 = 0.59), which indicated that the number of instars was correlated with stadia length as a function of the sequential instar. This meant that the length of each stadium was affected differently by the total number of instars. Simple linear regressions between each stadia length and number of instars showed different degrees of correlation and slope values (Table 5).
Discussion
Although the objectives of this study were not to quantify the effects of nutrition on the development of T. molitor, our results showed that the addition of a diet supplement that only made up 20% of the total diet can significantly affect development time, instar number, and survival of T. molitor larvae. One of the diet supplements (diet 2) also impacted significantly the resulting pupal weight. This showed that the addition of certain nutrients that may be lacking in the wheat bran basic diet (such as starch, protein, and vitamin C) had a favorable effect on the growth and development of this beetle. These results agree with those reported by Davis (1969, 1970, 1971a, b), although he tested different protein sources in his diets and did not mix them with wheat bran.
Diet also impacted significantly total development time, number of instars, and the length of most of the stadia of T. molitor. These 3 variables showed different degrees of relationship with each other between and within diet treatments. For instance, the total development time of T. molitor was significantly affected by the number of instars observed. This indicated that developmental plasticity in the number of instars plays an important role in the physiological response of T. molitor to suboptimally nutritious diets. Differences observed on the length of stadia among the different diet treatments showed that plasticity on stadia length, especially in stadia 9 - 15, played an equally important role. The plasticity on these 2 variables (number of instars and stadia length) appeared to explain most of the developmental response of T. molitor to nutritional supplements. Based on these findings, we recognized these 2 variables affecting developmental plasticity as different phenomena. The variables of number of instars and stadia length seemed to be significantly correlated, and this correlation was evident within most, but not all, of the sequential instars. Developmental plasticity appeared to be responding simultaneously in these 2 ways to favorable versus unfavorable nutrition.
Differences among slope values indicated that development time in the diet 3 treatment was impacted by a lesser magnitude by number of instars as compared with the other 3 diet treatments. A possible explanation for this phenomenon may be toxicity induced by the excessive increase in the content of sodium in this treatment. Immature mortality in diets 1, 2 and 3 appeared to be correlated with sodium content, which increased from 220 mg/g in diet 1 - 460 mg/g in diet 2, and 690 mg/g in diet 3. No other nutritional component showed similar extreme changes among these 3 diets.
When the data were analyzed backward as stadia previous to pupation, the results showed that the last stadium before pupation (P−1) was of a consistent length within food treatment groups. Because of the variability in developmental plasticity shown on the number of instars, P−1 stadium represented different sequential instars even within each food treatment group. However, the length of the last stadium did not significantly change between the different sequential instars. This is an important fact because the development time among sequential instars 9 - 12 was significantly different in all food treatments, except diet 1, and instars 11 and 12 were the last before pupation in many individuals from groups fed on diets 1, 2, and 3. In the control group, instars 12 - 15 showed no difference in development time (Fig. 4), and no individuals in this treatment group completed larval development with less than 13 instars. Further more, the P−1 stadium was significantly longer than the P−2 stadium in all treatments except the control group, and the development time among different instars representing the P−2 stadium showed significant differences. The same can be said about the P−3 stadium (3 molts away from pupation).
This pattern seemed to suggest that the P−1 instar is a basic biological characteristic of the life cycle of T. molitor, and instar insertion resulting from developmental plasticity may be occurring before this last instar. Connat et al. (1991) were able to differentiate these last instars by morphological differences in the cuticle and eye developmental stages. They also determined that the last instar of T. molitor was able to complete development and molt to pupa in absence of food contrasting with earlier instars, which suppressed molting under starvation. Urs and Hopkins (1973) also were able to differentiate the P−1 instar and referred to it as a prepupa.
The pattern of stadia length previous to pupation observed suggested that the P−1 instar may be replicated in extreme cases, but usually instar insertion may occur earlier during the development between the 9th sequential instar and the P−1 instar. Another possibility of instar insertion may be by duplication of earlier sequential instars. There was no significant difference in length among the 5th, 6th, 7th, 8th, and 9th sequential stadia within food treatment, and their variability was comparatively small. The regularity in length among these stadia appears to indicate that they may represent the same instar, which has been duplicated several times.
The length of the P−1 stadium was significantly affected by food treatment, which indicated that stadia length played a significant role in the developmental plasticity of this last stadium. Plasticity in stadia length was evident in almost every sequential instar and was affected by food treatment in all of them with the notable exception of sequential instars 1 through 4. This may indicate that the 1st, 2nd, 3rd, and 4th sequential instars constitute a stable section of the life cycle of T. molitor.
The length of individual stadia was significantly correlated with total number of instars in all but the 1st and 4th instars. The degree of correlation measured as a function of the parameter “b” was dramatically different in P−2 and P−3 as compared with sequential instars and P−1. This is an indication that P−2 and P−3 are not true instars but represent a mix of several instars. The correlation between the total number of instars and the development time of P−2 and P−3 most likely correspond to the last portion of the pattern presented in Fig. 4.
Although the mechanism of instar insertion in T. molitor is still not clear, data presented herein suggest that it may be occurring in 2 ways: (1) instar duplication between the 5th and 9th sequential instars, or (2) insertion between instar 9 and P−1. Future research will focus on determining how head capsule measurements change in each instar and how its size pattern is impacted by nutrition. A better knowledge of the mechanisms of stadia length and instar number variation in T. molitor could improve manipulation techniques of the development time of this insect. These techniques may consist of rich diet supplements to accelerate development during periods of high demand and the use of impoverish diets to slow development during periods of low demand. This could be advantageous for the commercialization of this insect by providing industry with the production flexibility to match changes in the demand of the product.
Effect of diet on the number of instars (A), total development time (B), and pupal weight of Tenebrio molitor. Boxes represent the 25th to the 75th percentiles, the lower brackets represent the 10th percentiles and the upper brackets represent the 90th percentiles. Dashed line represents the mean and continuous line represents the median. Dots represent outlying points. Means with the same letter are not significantly different after Tukey-Kramer's HSD test at α = 0.05.
Frequencies in number of instars observed in T. molitor fed on 4 different diets.
Regression of number of instars versus total development time of T. molitor developing in 4 different diets. Parameter estimates are presented in Table 3.
Pattern of sequential instar development time of T. molitor feeding on 4 different diets. Circles represent means of development time and brackets represent standard deviation. Means within diet treatment with the same letter are not significantly different after Tukey-Kramer's HSD test at α = 0.05.
Reversed pattern of stadia length as molts before pupation of T. molitor feeding on 4 different diets. Circles represent means of development time and brackets represent standard deviation. Means within diet treatment with the same letter are not significantly different after Tukey-Kramer's HSD test at α = 0.05.
Development time of instars represented within stadia 1, 2, and 3 molts before pupation (P−1, P−2, and P−3, respectively) (columns) of T. molitor feeding on 4 different diets (Rows). Symbols represent means of development time and brackets represent standard deviation. Means with the same letter (within graphs) are not significantly different after Tukey-Kramer's HSD test at α = 0.05.
Contributor Notes
3USDA-ARS Southeastern Fruit and Tree Nut Research Laboratory, Byron, GA 31008.
4Southeastern Insectaries Inc., Perry, GA 31069.