MATERIAL SOURCES FOR EXPERIMENTS

Diaphorina citri nymphs used as hosts for D. aligarhensis development were obtained from Candidatus Liberibacter asiaticus-free colonies maintained at the University of California-Riverside, Insectary and Quarantine Facility, Riverside, California, USA (Hoddle & Pandey 2014; Bistline-East et al. 2015; Bistline-East & Hoddle 2016). The initial population of Candidatus Liberibacter asiaticus-free D. citri were collected from backyard citrus trees in Riverside and Los Angeles counties, California, USA (California Department of Food and Agriculture permit number 2870). Diaphorina citri colonies were reared on 1.5-yr-old Citrus volkameriana V. Tenore & Pasquale (Rutaceae) plants at 27 ± 2 °C, a 14:10 h (L:D) photoperiod, and 40 ± 5% relative humidity (RH), in riser cages (15 × 15 × 30 cm [W × D × H]) constructed from 2 U-shaped acrylic risers (S&W Plastics F2191, Riverside, California, USA) with the back covered with no-see-um netting (Skeeta, Bradenton, Florida, USA), and the front covered with a 30 cm long no-see-um sleeve to allow access into cages (Bistline-East et al. 2015). Plants were grown in greenhouses at the University of California-Riverside, Agricultural Operations (27 ± 2 °C, 50 ± 6% RH, and natural sunlight) in 0.5 L pots with University of California-Riverside Type III potting soil mix, with daily watering, and Osmocote Pro 19:6:12 granular fertilizer (The Scotts Company LLC, Marysville, Ohio, USA) applied every 3 mo. Plants were pruned regularly until new shoots (i.e., flush growth) emerged; D. citri requires newly developing young citrus foliage for oviposition and nymph development (Milosavljević et al. 2018).

To establish experimental cohorts of D. citri nymphs on which to raise D. aligarhensis (and subsequently P. diaphorinae), individual citrus plants were isolated in riser cages and transferred from greenhouses to University of California-Riverside, Insectary and Quarantine Facility for D. citri inoculation. Twenty non-sexed D. citri adults were introduced into each riser cage, and females were allowed to mate and lay eggs for 8 to 10 d before removal. Diaphorina citri nymphs were left to develop to second through fourth instars, which required an additional 7 to 10 d at 27 ± 2 °C, 40 ± 5% RH, and a 14:10 h (L:D) photoperiod.

Diaphorencyrtus aligarhensis and P. diaphorinae adults used for experiments were taken from colonies established in University of California-Riverside, Insectary and Quarantine Facility. These colonies originally were established from specimens reared from parasitized D. citri nymphs collected from citrus orchards in Punjab, Pakistan (USDA-APHIS permit number P526P-11-00103; Hoddle 2012; Bistline-East & Hoddle 2016). Diaphorencyrtus aligarhensis were reared on potted C. volkameriana infested with second through fourth instar D. citri nymphs and kept in 15 × 15 × 30 cm (W × D × H) riser cages at 27 ± 2 °C, 40 ± 5% RH, and a 14:10 h (L:D) photoperiod following the procedure described above (Bistline-East et al. 2015). Psyllaphycus diaphorinae were reared on D. aligarhensis pupae developing inside parasitized D. citri nymphs, and P. diaphorinae colonies were maintained under similar conditions as described above for D. aligarhensis (Bistline-East & Hoddle 2016). Newly emerged D. aligarhensis and P. diaphorinae adults were collected daily from colonies, sexed, labeled, and held in individual 200 µL O-ring centrifuge tubes (Micrewtube brand, Simport Scientific Inc., Vancouver, British Columbia, Canada) provisioned with honey droplets and maintained at 16 °C and a 4:10 h (L:D) photoperiod until used for experiments. Single female D. aligarhensis and P. diaphorinae were confined with 2 males of the same species for 24 h within each vial to ensure they were mated before being used for experiments.

EFFECTS OF TEMPERATURE ON HYPERPARASITOID DEVELOPMENT AND LONGEVITY

The effects of 5 fluctuating temperatures that averaged 15, 18, 20, 25, and 32 °C, over 24 h on the development times, and longevity of male and female P. diaphorinae were quantified and compared to 5 corresponding temperatures that were held constant over 24 h. At the beginning of experiments, individual citrus plants infested with second through fourth instar D. citri nymphs were chosen randomly from the stock colony and placed inside riser cages at 27 °C, 40% RH, and a 14:10 h (L:D) photoperiod. To obtain a synchronous cohort of hosts for P. diaphorinae parasitism, 10 honey-fed male and 5 female D. aligarhensis were released from O-ring centrifuge tubes into each riser cage, and females were allowed to oviposit for 3 d. Thereafter, adults were removed, and D. aligarhensis larvae were left to develop into pupae, which required approximately 8 to 10 d (Bistline-East & Hoddle 2016), after which they were exposed to P. diaphorinae.

After 8 to 10 d, developing D. aligarhensis were exposed to 5 male and 15 female P. diaphorinae that were released into each riser cage, and females were allowed to oviposit for 24 h before removal. Immediately after removing P. diaphorinae, the cages with plants and D. aligarhensis pupae exposed to P. diaphorinae were placed into growth chambers (model I-30BLL; Percival Scientific, Perry, Indiana, USA) set to 1 of the 10 experimental temperature regimens described above (Table 1). For constant temperature regimens, growth chambers were programmed to hold the target temperature static (± 0.2 °C) for the entirety of the study. For fluctuating temperature regimens, growth chambers were programmed with 9 temperature steps to incrementally increase and decrease temperatures around target temperature averages. The mean values of fluctuating temperatures were equivalent to the corresponding constant temperatures (± 0.2 °C) over a 24 h period. Each fluctuating temperature profile was modeled after field recorded temperatures that had the desired average target temperature (Table 1; CIMIS 2018). Thus, with respect to fluctuating temperature regimens, these temperatures represent realistic thermal exposure scenarios (McCalla et al. 2019; Milosavljević et al. 2019a).

Table 1.

Stepwise settings used in temperature cabinets for fluctuating temperature regimens calculated from an average of 2 yr of data (Jan 2013–Dec 2014) from the California Irrigation Management Information System (station #44) weather station at the Riverside Citrus Experiment Station, Riverside, California, USA.

The incremental temperature increase and decrease steps for fluctuating temperature regimens were modeled over 2 yr (Jan 2013–Dec 2014) of daily temperature data from a meteorological station located at the Riverside Citrus Experiment Station, Riverside, California, USA (CIMIS 2018). Daily temperature averages within 0.2 °C of each of the target constant temperatures were selected for each respective time period and date for each yr. After initial selection, hourly air temperatures (compiled from CIMIS 2018) for each selected d were averaged per h for every target mean temperature. These data were used to program temperature fluctuations within chambers over a 24 h period that when averaged were equivalent to the target mean temperature (Table 1). Target temperature regimens and relative humidity in each of the chambers were confirmed using HOBO Pro V2 Temperature/RH loggers (Onset Computer, Pocasset, Massachusetts, USA) set to record temperature and humidity every 15 min. The photoperiod setting for all experiments was 14:10 h (L:D), with light intensity of 100 µE m^-2 s^-1. All growth chambers were maintained at 45 ± 5% RH regardless of the experimental temperature regimen, which was the observed average relative humidity for selected time periods in Riverside, California, USA (CIMIS 2018).

Four riser cages were assigned to each of 10 temperature regimens to provide sufficient development data for a combined minimum of 40 emerged adult male and female P. diaphorinae. Cages were rotated daily within incubators to counteract potential thermal gradients (see Ashmore & Janzen 2003 for more details). Riser cages were inspected for adult P. diaphorinae that emerged from initial D. aligarhensis cohorts within each temperature profile, beginning 7 d after experimental trials were initiated. Subsequent checks for adult emergence at every experimental temperature were performed daily, and inspections continued for 14 consecutive d after the last recorded emergence in each cage. Emerged hyperparasitoids were isolated daily into individual 200 µL O-ring centrifuge tubes (Micrewtube brand, Simport Scientific Inc., Vancouver, British Columbia, Canada), sexed, labeled, and maintained in their respective incubators. All vials were provisioned with raw honey droplets as a source of carbohydrates, and replenished on alternating d for the duration of the experiment. Emerged male and female P. diaphorinae were monitored at 24-h intervals until death to determine adult longevity under each temperature regimen. Adult P. diaphorinae that died prematurely (e.g., trapped in honey droplets) were removed from statistical analyses.

No development was observed from cages of P. diaphorinae maintained at constant 15 °C. Sixty d after exposure to P. diaphorinae, 10 D. aligarhensis pupae were harvested randomly from riser cages held at constant 15 °C and dissected under a dissecting microscope (Leica MZ12.5, Meyer Instruments, Houston, Texas, USA); 8 of the 10 pupae contained live larvae. An additional 10 D. aligarhensis mummies were then collected and reared at 27 °C, 40% RH, and a 14:10 h (L:D) photoperiod to identify the species (i.e., D. aligarhensis or P. diaphorinae). After 7 d at 27 °C, 8 adult P. diaphorinae emerged from the 10 harvested D. aligarhensis mummies. The remaining 2 mummies were considered dead when no emergence occurred 4 wk after the initiation of the species identity test. The remaining D. aligarhensis exposed to P. diaphorinae and held at constant 15 °C were maintained for 90 d post-exposure. No hyperparasitoids emerged, and replicates at this temperature were discontinued (Table 2).

STATISTICAL ANALYSES

To analyze whether P. diaphorinae development and longevity were affected by temperature (treated as a continuous variable), each temperature profile (i.e., 15, 18, 20, 25, or 32 °C) was considered separately. For each temperature profile, a 1-way analysis of variance (ANO-VA) was performed using the 4 treatment combinations (i.e., females constant, females fluctuating, males constant, and males fluctuating) in SAS (2009). Response variables were log-transformed as necessary prior to analysis of variance to meet assumptions of normality and homoscedasticity. If the analysis of variance was significant at the 0.01 alpha level, individual pairwise comparisons were made for gender within regimens (i.e., comparing females vs. males for fluctuating and constant regimens separately) as well as regimen within gender (i.e., comparing fluctuating vs. constant for females and males separately), a total of 4 comparisons were made, between treatment means using the DIFF option in the LSMEANS statement of the GLM procedure in SAS. For all experimental temperatures, there were no significant differences between the development rates of male and female P. diaphorinae, and the values were pooled for subsequent model fitting (see Results; Tables 2 & 3).

Table 2.

Mean development time in d (± SE) for female and male Psyllaphycus diaphorinae reared under constant and fluctuating temperatures.

MATHEMATICAL MODELS

To assess the relationship between temperature and development rate (r = 1/d, where d is mean development time in d) of male and female P. diaphorinae under constant and fluctuating temperature regimens, 1 linear (Campbell et al. 1974) and 7 nonlinear (i.e., Beta [Yin et al. 2003; Auzanneau et al. 2011; Shi et al. 2015], Lactin-2 [Logan et al. 1976; Lactin et al. 1995], Brière-2 [Brière et al. 1999], Performance-2 [Shi et al. 2011; Wang et al. 2013], Lobry-Rosso-Flandrois [Lobry et al. 1991; Rosso et al. 1993], Weibull [Angilletta 2006], and Ratkowsky [Ratkowsky et al. 1983]) development rate models were fit to datasets. Temperatures for which no P. diaphorinae emergence was recorded were not used for model fitting.

Linear regression was used to calculate thermal constants (i.e., degree-d necessary for completion of development), theoretical lower development thresholds (i.e., T_min), and adjusted coefficient of determination (R² _adj) was used as a measure of model goodness-of-fit. Linear models with high values of R² _adj indicated a good fit to the data for each response variable (Table 3; Campbell et al. 1974). The linear relationship between development rate of P. diaphorinae and temperature was assessed for each response variable using the PROC REG procedure in SAS (2009). In each analysis, Cook's D metric was calculated to identify influential outliers with a critical value of Cook's D > 4/n, where n represents the number of observations (Bollen & Jackman 1990). Once identified, outliers exhibiting large externally studentized residuals and leverage (i.e., greater than 3 standard deviations away) were removed from analyses. For the linear model, temperatures equal to or less than 25 °C (i.e., 18, 20, and 25 °C) were used in analyses because these data points aligned with the linear portion of the dataset.

Development rate data of P. diaphorinae were used to parameterize nonlinear models (Table 3). Of the 7 nonlinear models evaluated, the Beta, Lactin-2, Brière-2, Lobry-Rosso-Flandrois, and Weibull models failed to fit to development rate data for P. diaphorinae (see Results section) regardless of temperature regimen, and were removed from further analyses. The performance of the remaining 2 nonlinear models, Performance-2 and Ratkowsky, were evaluated based on the goodness-of-fit to observed data (i.e., residual sum of squares [RSS]), and the biological relevance of model estimations (i.e., T_min [lower theoretical temperature bound for development], T_opt [optimal temperature for development], and T_max [upper theoretical temperature bound for development]). The 2 nonlinear models tested have 4 parameters and thus, the same degrees of freedom [df], and 2 ‘conceptual’ lower (i.e., T_min) and upper development (i.e., T_max) thresholds (Mirhosseini et al. 2017; Ratkowsky & Reddy 2017).

Table 3.

Mathematical models and parameter estimates for 3 performance functions describing the relationship between development rate (D_r) and temperature for Psyllaphycus diaphorinae reared under constant and fluctuating temperature regimes.

The coefficient of determination (R²) and its adjusted value (R² _adj), although commonly used in the literature to describe the fit of nonlinear models to development data, were not used to assess goodness-of-fit for nonlinear models because they do not describe accurately the validity of a nonlinear fit to data (see Spiess and Neumeyer 2010 for further explanation). Instead, nonlinear model goodness-of-fit was assessed with the residual sum of squares (Shi et al. 2015; Ratkowsky & Reddy 2017; McCalla et al. 2019; Milosavljević et al. 2019a):

where n represents the number of observations and and denote the observed and expected development rate at the i-th temperature, respectively. Nonlinear models with small values of the residual sum of squares (e.g., RSS < 0.001) indicated a good fit to the data.

The Performance-2 model was fit with development rate in its untransformed form as the response variable, because data satisfied homogeneity assumptions (see Ratkowsky 2004 for more details). The Ratkowsky model was fitted with the square root of development rate as the response variable, because no statistical comparisons were made with the Performance-2 model (Ratkowsky & Reddy 2017). All temperature-based development models were fit for each response variable using the PROC NLIN procedure in SAS (2009), and graphed in SigmaPlot vers. 12.0 (Systat 2008).

EFFECT OF TEMPERATURE AND REGIMEN TYPE ON HYPERPARASITOID DEVELOPMENT

Psyllaphycus diaphorinae males and females completed egg-to-adult development under experimental temperature profiles from 18 to 32 °C in the constant regime and 15 to 32 °C in the fluctuating regime (Table 2). Average temperature and regimen type (i.e., constant or fluctuating) significantly affected mean development time (regimen effects: P < 0.001 for all temperature profiles; Table 2). Nevertheless, male and female hyperparasitoids responded equivalently to the temperature treatments (gender effects: P > 0.44 for all temperature profiles; Table 2). Egg-to-adult development time of P. diaphorinae decreased with increasing temperature to a minimum of 10.00 ± 0.26 d (males) and 10.23 ± 0.23 d (females) at 32 °C constant and 13.71 ± 0.35 d (males) and 14.46 ± 0.40 d (females) at 32 °C fluctuating (Table 2). Hyperparasitoids completed development faster at constant 20 to 32 °C than the corresponding fluctuating temperatures (Table 2), and development times did not vary between constant and fluctuating temperature regimens at 18 °C (Table 2).

For the linear and 2 nonlinear (i.e., Ratkowsky and Performance-2) models fitted to P. diaphorinae development rate data parameter, R² _adj (linear models), and residual sum of squares (nonlinear models) estimates are presented in Table 3. The linear model had R² _adj values > 0.98 for both constant and fluctuating temperature regimens (Table 3). The Beta, Lactin-2, and Brière-2, Lobry-Rosso-Flandrois, and Weibull nonlinear models failed to converge for both constant and fluctuating datasets, a finding which could be partially attributed to the limited number of experimental temperatures being evaluated. Therefore, these models were removed from further consideration. The remaining 2 nonlinear models, Ratkowsky and Performance-2, had small residual sum of squares values (< 0.00008) for fluctuating temperature regimens, indicating a good fit to datasets (Table 3). For constant temperatures, however, Ratkowsky and Performance-2 both failed to converge, and again this may have been a consequence of too few experimental temperatures being evaluated.

Linear regression estimates indicated the total development time in degree d for immature P. diaphorinae was significantly lower (i.e., 36% lower) under the constant temperature regimen than the corresponding fluctuating regimen (Table 3; Fig. 1). The Ratkowsky and Performance-2 models fit to fluctuating temperature data estimated similar values for the optimum ([T_opt]: about 30 °C) and upper ([T_max]: about 34 °C) development temperatures for P. diaphorinae females and males (Table 3). The theoretical critical lower temperatures (T_min) above which development of female and male P. diaphorinae occurs were not estimated equally by the linear and 2 nonlinear models. For the fluctuating dataset, estimates of T_min for the Ratkowsky model (–3.74 °C) were lower than those estimated by the linear (8.63 °C) and Performance-2 (8 °C) models (Table 3; Fig. 1).

EFFECT OF TEMPERATURE AND REGIMEN TYPE ON HYPERPARASITOID LONGEVITY

Mean longevity of P. diaphorinae adults was significantly affected by average temperature and thermal variation (regimen effect: P < 0.001 for all temperature profiles; Table 4). Maximum longevity of 56.08 ± 2.66 d (males) and 83.55 ± 2.89 d (females) was observed at 15 °C fluctuating and 41.20 ± 1.15 d (males) and 52.72 ± 4.81 d (females) at 18 °C constant (Table 4). Above these temperatures, longevity steadily decreased with rising temperature, reaching minimums of 2.29 ± 0.28 d (males) and 7.61 ± 0.74 d (females) at 32 °C fluctuating and 11.57 ± 1.63 d (males) and 14.15 ± 1.58 d (females) at 32 °C constant. Despite the similarity in overall trend, average lifespan duration was significantly different between the 2 genders (gender effect: P < 0.001 for all temperature profiles) across all experimental temperatures except fluctuating 20 °C and constant and fluctuating 25 °C (Table 4). The magnitude of this effect was greatest near the lower and upper extremes of experimental temperatures (Table 4). No significant temperature effects were observed between genders at 25 °C. At intermediate temperatures, P. diaphorinae females and males reared under fluctuating profiles completed development more slowly (20 and 25 °C) and survived longer (18–25 °C) than those reared under constant temperatures with equivalent means (Table 4). In comparison with constant regimens, fluctuating regimens produced P. diaphorinae males with a lifespan approximately 5 times shorter and P. diaphorinae females with a lifespan approximately 2 times shorter at 32 °C.