Reproductive Parameters

Pre-oviposition, oviposition, post-oviposition periods, and fecundity for E. lignosellus were determined at nine constant temperatures [13, 15, 18, 21, 24, 27, 30, 33, and 36 °C (± 0.05 °C)] at 14:10 (L:D) and 65–70% RH in temperature controlled chambers (Custom made, Department of Entomology and Nematology, University of Florida, Gainesville, FL) to construct time-specific life tables. This study was conducted in 2006–2008. Using two chambers, two temperatures were tested simultaneously with pairs of temperatures tested staggered over time. The range of tested temperatures was selected based on those experienced by lesser cornstalk borer in sugarcane in southern Florida. The experiment design was a randomized complete block with three cohorts at each temperature and it was replicated through time; repeated for three generations at each temperature. Adults were obtained from immatures reared on sugarcane used for companion developmental studies (Sandhu et al. 2010) conducted at the same temperatures and relative humidity as indicated above. Details on materials and methods used to rear immature stages are available in Sandhu et al. (2010). Ten male:female pairs of newly emerged adults (< 12 h old) were first released into each of three oviposition cages (17 × 17 × 17 cm) to facilitate mating. Adults were provided with a 10% honey solution for free feeding (ad libitum feeding), because sugarcane does not produce a food source for adults. After 24-h, individual pairs were moved to transparent plastic cylinders (one pair per cylinder) (11 cm length and 5 cm diameter; Thornton Plastic Co., Salt Lake City, UT) lined with tubular synthetic stockinette (Independent Medical Co-Op, Ormond Beach, FL) as an oviposition substrate as the females prefer rough, dry substrates for oviposition. We used stockinette because females preferred it over the Handy Wipes (The Chlorox Co., Oakland, CA) used by previous workers (e.g., Chalfant 1975). Adults were observed daily to record the limits of the reproductive parameters. The stockinette was replaced daily during oviposition periods. The orange-colored eggs were easily observed against the white background of the material and were counted using a hand lens. Fecundity was reported as the number of eggs deposited by an individual female during her entire life period. Age-specific female survival (l_x, percentage of females alive at specific age x) and age-specific fecundity (m_x, number of female offspring produced by a female in a unit of time) were recorded for each day (x) they were alive. Age specific fecundity was calculated as (f / (m + f)) × n, where f = number of females, m = number of males, and n = number of offspring. The l_x and m_x were calculated for each cohort of 10 females. Age-specific survivorship curves were constructed using adult l_x and m_x values for cohorts at each temperature treatment.

Life Table Parameters

The age-specific life table method was used to calculate the life table parameters for E. lignosellus (Birch 1948). The values of lx used in life table calculations were obtained by combining age specific adult female survival with the survival of immatures reported under same environmental conditions in a concurrent study (Sandhu et al. 2010). Similarly, the age of adults (x) used in these calculations also includes the development time of immatures reared on sugarcane under the same environmental conditions (Sandhu et al. 2010). The intrinsic rate of increase (r) was calculated through iteration of the Euler-Lotka equation (Σe-rx l_xm_x = 1). The l_x and m_x values were used to calculate the net reproductive rate (R₀ = Σl_xm_x, mean number of female offspring / female) and the mean generation time (T = Σ(xl_xm_x) / Σ(l_xm_x), mean age of the mothers in a cohort at the birth of female offspring). The values for r at each temperature were used to calculate the finite rate of increase (λ = e^r, the number of times the population multiplies in a unit of time) and population doubling time (DT = ln (2) / r, the time required for the population to double). To compare the thermal sensitivities of r and R₀, we plotted these alternative fitness measures against the tested temperatures.

TABLE 1.

MATHEMATICAL EQUATIONS OF THE MODELS TESTED TO DESCRIBE THE RELATIONSHIP BETWEEN TEMPERATURE AND INTRINSIC RATE OF INCREASE IN LESSER CORNSTALK BORER ON SUGARCANE.

Model Evaluation

A non-linear distribution was observed when r was plotted against the temperature treatments. To find an equation that best fit the observed relationship between r and temperature, eight non-linear models (Table 1) previously used (Roy et al. 2003; Bonato et al. 2007; Sandhu et al. 2010; Shi & Ge 2010) to describe temperature dependent development or other population parameters of different insects were tested. The parameters of the nonlinear models were estimated with the nonlinear regression model of Marquardt (1963) using SAS (SAS Institute 2008). Sigma Plot (Systat Software, Inc., San Jose, CA) was used to plot the regressions of the non-linear models. The models were evaluated based on the coefficient of determination (r²), the adjusted coefficient of determination (r²_adj, a modified r² that adjusts for the number of explanatory terms in the model), the residual sum of squares (RSS), and the corrected Akaike Information Criterion (AICC) (Burnham and Anderson 2004). The r² and r²_adj. indicate better fits with higher values, whereas RSS and AIC_C indicate better fits with lower values. The corrected AIC value was calculated using the formula:

Statistics

PROC MIXED (SAS Institute 2008) was used to analyze the variance due to the potential covariance structure associated with taking repeated measures over time at each temperature. Normality of the data was tested with the Shapiro-Wilk normality test (Shapiro & Wilk 1965). The oviposition cages were treated as cohorts and replications through time were treated as generations for data analysis. Temperatures, cohorts, generations, and their interactions were used in the analysis of variance models. Generations were used as the repeated variable and the cohorts were nested under temperature in the repeated measures statement. Several covariance structures were fitted to the data. The unstructured covariance type fit well and was used for the analysis (Littell et al. 1998). Data for each pair of adults were used for analysis of effects of temperature, cohort, and generation for reproductive parameters. Daily values by cohort were used for analysis of effects of temperature and generation on l_x and m_x. The percentage of females alive at age x (l_x) was arcsin square root transformed for normality purposes before analysis and retransformed for presentation purposes. The Tukey's HSD test (SAS Institute 2008) was used for means separation with α = 0.05.

Reproduction

Temperature significantly affected the lengths of the lesser cornstalk borer pre-oviposition, oviposition, and post-oviposition periods (Table 2). Cohorts, generations and the modeled interactions were not significant sources of variation in the models for any of these periods. Therefore, data were pooled across cohorts and generations to calculate means for these periods. The mean pre-oviposition period decreased with an increase in temperature from 9.7 d at 13 °C to 2.3 d at 33 °C (Table 3). The mean oviposition period was longest (5.6 d) at 27 °C and decreased with an increase or decrease in temperature from 27 °C. The post-oviposition period became progressively shorter from a maximum at 13 °C to a minimum at 36 °C (Table 3).

TABLE 2.

ANALYSIS OF VARIANCE FOR EFFECTS OF TEMPERATURE, COHORT AND GENERATION ON REPRODUCTIVE PARAMETERS OF ELASMOPALPUS LIGNOSELLUS ON SUGARCANE.

TABLE 3.

MEAN (± SEM) PRE-OVIPOSITION, OVIPOSITION, POST-OVIPOSITION PERIODS AND FECUNDITY FOR ELASMOPALPUS LIGNOSELLUS ON SUGARCANE.

Fecundity was also significantly affected by temperature (Table 2). Cohort, generation and modeled interactions were not significant sources of variation in the fecundity model. Therefore, the fecundity data were pooled across cohorts and generations to calculate mean fecundity at each temperature. Fecundity increased with an increase in temperature from 13 to 30 °C and decreased at 33 and 36 °C. Mean fecundity ranged from 29.2 (13 °C) to 165.3 eggs (30 °C) (Table 3) which indicates that warm temperatures are more favorable for fecundity than cool temperatures.

Model Evaluation

The model evaluation parameters, and the predicted optimum temperatures (temperatures for maximum intrinsic rate of increase) are presented in Table 5. The predicted optimum temperature ranged from 28.8 °C (Taylor model) to 31.3 °C (Logan model). In general, all the tested models showed satisfactory fitted results with r2> 0.9312 and r2adj > 0.8623. The polynomial (fourth order) and Wang-Lan-Ding models showed the best fit to the data based on their higher r2and r2adj and lower RSS values than other models (Table 5). However, the greater number of fitted parameters in the polynomial model resulted in greater AICC value (-39.09) than the other models and the AICC for Wang-Lan-Ding could not be calculated, because (n-K-1) in the AICC formula equaled zero. Based on the lowest AICC (-62.43) values, the Brière-1 was found to be the best model to describe the relationship between r and temperature.

Fig. 1.

Relationships between the adult age (d) and age-specific survival l_x (solid line) and age specific daily fecundity m_x (dashed line) for Elasmopalpus lignosellus at 9 constant temperatures.

The fitted curves for all the tested models representing the relationship between r and temperature for the lesser cornstalk borer on sugarcane are presented in Figs. 3a and 3b. In general, the shape of these curves estimated by different models is quite similar. However, some models had different lower and upper temperatures where the curve intersected x-axis. For example, in Brière-1, Brière-2, Lactin, Ratkowsky and Wang-Lan-Ding models, curve intersected the x-axis at approximately 12–13 °C which indicates their estimated lower temperature threshold. However, in Logan and Taylor models, curve did not intersect x-axis at lower temperature which indicate the estimated lower temperature threshold to be < 10 °C. Similarly, the estimated upper temperature threshold was 36–37 °C in most of the models, but in Taylor model it was > 40 °C, because curve did not intersect x-axis even at 40 °C.

TABLE 4.

LIFE TABLE PARAMETERS FOR ELASMOPALPUS LIGNOSELLUS ON SUGARCANE AT NINE CONSTANT TEMPERATURES.

Fig. 2.

Effect of temperature (°C) on Elasmopalpus lignosellus intrinsic rate of increase (r) and net reproductive rate (R₀ ).

Reproduction

The values for reproductive parameters on sugarcane fell mostly within the ranges of those determined for E. lignosellus on other crops such as southern pea and soybean. The mean (± SEM) pre-oviposition period found in this study (2.3 ± 0.05 d at 33 °C to 9.7 ± 0.06 d at 13 °C) is similar to the value of 2.8 d reported by Stone (1968) for E. lignosellus on an artificial diet at 27 °C. The mean oviposition period on sugarcane (2.2 ± 0.05 d at 13 °C to 5.6 ± 0.06 d at 27 °C) was shorter than those reported on artificial diets (10.4 d, Luginbill and Ainslei 1917; 11.8 d, Stone 1968; and 6.4 d, Simmons and Lynch 1990), but within the range determined by Dupree (1965) on southern pea (mean: 4.1 d, range 1 to 9 d). The 4.7 d post-oviposition period reported by Leuck (1967) on soybean is consistent with that found on sugarcane (2.5 ± 0.05 to 5.9 ± 0.07 d).

Fecundity on sugarcane (29 to 165 eggs per female) mostly fell within the range reported by others on leguminous crops and artificial diets. The lesser cornstalk borer mean fecundity (number of eggs/female) reported in earlier studies was 192 on cowpeas (Luginbill & Ainslei 1917), ranged from 124 to 129 on soybean (King et al. 1961; Dupree 1965; Leuck 1967), and ranged from 67 (Calvo 1966) to 419.5 on artificial diet (Stone 1968). The results of our study are similar to those of Mack and Backman (1984) who reported an increase in fecundity with an increase in temperature from 17 to 27.5 °C, peaks at 27.5 and 30.5 °C, and large decreases at 17 and 35 °C.

Life Table Parameters

Elasmopalpus lignosellus reached its maximum reproductive rate within the same temperature range where generation time and population doubling time were the lowest (30 to 33 °C). Life table parameters for lesser cornstalk borer on other hosts are not available for comparison, but such values have been published for other stalk boring Lepidoptera pests of monocots (Table 6). Intrinsic rate of increase (r) at 25 °C on corn and artificial diet for Diatraea lineolata (Walker) (Lepidoptera: Crambidae), a neotropical cornstalk borer (Rodríguez-del-Bosque et al. 1989), were slightly lower than those found in this study for E. lignosellus that developed as larvae on sugarcane at 24 °C. The sugarcane borer, D. saccharalis (F.), and the Mexican rice borer, Eoreuma loftini (Dyar) (both Lepidoptera: Crambidae), recorded lower r and λ, and higher T and DT parameters at 30 °C on an artificial diet that included dried, ground sugarcane leaf sheaths (Sétamou et al. 2002) than those reported in our study for E. lignosellus. The Mexican rice borer had greater R₀ than the lesser cornstalk borer that indicates its high reproductive potential, but the r for the Mexican rice borer was lower on an artificial diet than for the lesser cornstalk borer on sugarcane. Our studies determined that r and λ for E. lignosellus reared on sugarcane were slightly lower than D. saccharalis reared on corn and on sugarcane at 27 °C (Bessin & Reagan 1990). However, the net reproductive rate (R₀) for D. saccharalis on both of these host plants was approximately seven times greater than for E. lignosellus on sugarcane. The results of Bessin & Reagan (1990) and Sétamou et al. (2002), who found very different r and R₀ values for D. saccharalis that developed on corn and sugarcane compared to an artificial diet containing <10% sugarcane leaf sheaths, provide further evidence that larval host or diet affects the intrinsic rate of increase.

TABLE 5.

COMPARISON OF DIFFERENT NON-LINEAR MODELS TESTED TO DESCRIBE THE RELATIONSHIP BETWEEN INTRINSIC RATE OF INCREASE (r) AND TEMPERATURE (°C).

Fig. 3a–b.

Relationship between temperature (°C) and intrinsic rate of natural increase (r) for E. lignosellus on sugarcane described by different non-linear models.

TABLE 6.

LIFE TABLE PARAMETERS FOR SELECTED STEM-BORING LEPIDOPTERA PESTS OF SUGARCANE AND CORN.

In most ectotherms, r reaches its maximum value at a greater temperature than for R₀. Huey & Berrigan (2001) attributed this to r (but not R₀) being sensitive to the accelerating effect of high temperature on generation time. Because r is inversely related to generation time and R₀ is independent of it (Cole 1954; Lewontin 1965), the normal shortening of the generation time associated with rising temperatures will work to increase r without affecting R₀. This leads to a shift to the right in the thermal fitness curve for r relative to R₀, as was observed in this study. This right shift of the r curve relative to the R₀ curve observed for E. lignosellus on sugarcane may be due to the shortest generation time occurring at a temperature 3 and 6 °C greater than the maximum r and R₀ values, respectively.

Model Evaluation

Mathematical models used in this study were also tested to describe the relationship between temperature and r for Tetranychus mcdanieli McGregor (Acarina: Tetranychidea) and Stethorus punctillum Weise (Coleoptera: Coccinellidae) on red raspberry (Rubus idaeus L.), Sitotroga cereallela (Olivier) (Lepidoptera: Gelechiidae) on corn (Zea mays L.), and Halyomorpha halys (Stal) (Hemiptera: Pentatomidae) on green beans [Vigna radiata (L.) Wilczek] (Roy et al. 2003; Hansen et al. 2004; Nielsen et al. 2008). High r² values were used to choose the Lactin-2 model for T. mcdanieli and the Brière-1 model for S. punctillum, S. cereallela, and H. hays. In the present study, the Wang-Lan-Ding model provided the best fit between temperature and r based on r² and RSS. However, Angilletta Jr (2006) reported that the selection of model based on r² and RSS can be erroneous to select the best model. Based on AICC values, the Brière-1 model was the best to describe the relationship between r and temperature. Shi and Ge (2010) also reported the Brière-1, Brière-2 and Performance models to be the most suitable for explaining temperature-dependent development rates of insects. Additional factors such as soil moisture and natural enemies have been reported to affect E. lignosellus populations under field conditions. For example, elevated moisture levels at the soil surface play an important role in reducing oviposition and larval survival under field conditions (Smith & Ota 2002). Larval parasitoids and predators may also play an important role in regulating E. lignosellus population growth in sugarcane (Falloon 1974).