Introduction

The mealybug destroyer, Cryptolaemus montrouzieri Mulsant (Coleoptera: Coccinellidae), is native to Australia and was among the first species to be used in the (classical) biological control of mealybug pests (Pseudococcidae) (Fisher 1963; Clausen 1978). Several attempts have been made to introduce C. montrouzieri in different parts of the world, but particularly in moderate climates most of these attempts have failed because of the ladybird’s low capacity to survive cold winters (DeBach and Hagen 1964). Therefore, the biological control of mealybugs in fruit trees and other perennial crops largely relies on the periodical release of mass cultured C. montrouzieri ladybirds (Fisher 1963; Hodek and Honĕk 2009). Also the augmentative use of C. montrouzieri in protected cultivation and interior landscaping requires large numbers of commercially produced ladybirds (Chong and Oetting 2007; Muştu et al. 2008; Hodek and Honĕk 2009). The relatively high cost price of commercial natural enemies is an important drawback of augmentative biological control and slows down its adoption in most agricultural systems (van Lenteren 2012). This obstacle can be in part overcome by rationalizing the mass production techniques for arthropod natural enemies thus lowering their production cost (van Lenteren and Tommasini 2003).

Given its specialist feeding habit, C. montrouzieri is traditionally produced using mealybugs reared on plant materials like potato sprouts (Fisher 1963) or pumpkins (Babu and Azam 1987). Not only is this natural rearing method time-consuming but the seasonal availability of these plant materials also complicates the continuous supply of adequate quantities of mealybugs. In order to improve the economy of mass production, factitious food sources have been employed for the rearing of different coleopteran predators with notable success (Hodek and Honĕk 2009; Riddick and Chen 2013). For example, eggs of different Lepidoptera species have shown to be a suitable factitious food for the predatory ladybirds Adalia bipunctata (L.) (De Clercq et al. 2005), Propylea japonica (Thunberg) (Hamasaki and Matsui 2006) and Harmonia axyridis (Pallas) (Berkvens et al. 2008b). C. montrouzieri larvae can also be successfully reared on eggs of the Angoumois grain moth Sitotroga cerealella (Olivier) (Lepidoptera: Gelechiidae) and the Mediterranean flour moth Ephestia kuehniella Zeller (Lepidoptera: Pyralidae) (Pilipjuk et al. 1982; Attia et al. 2011).

Previous studies indicated that C. montrouzieri females largely depend on mealybug wax filaments to stimulate oviposition and that contact chemical cues seem predominant in inducing the search for an oviposition site. Besides, mechanical and thigmotactic cues, such as those provided by mealybug ovisacs, must be met (Merlin et al. 1996; Finlay-Doney and Walter 2012). Therefore, the implementation of factitious prey in the mass culture of C. montrouzieri is only useful if an artificial oviposition substrate is available that triggers oviposition in the absence of mealybug ovisacs. To our knowledge no artificial oviposition substrate has been reported in the literature for C. montrouzieri.

In the present study a semi-artificial rearing system using a factitious prey and an artificial oviposition substrate for C. montrouzieri was evaluated. Developmental and reproductive performance of the predator reared on E. kuehniella eggs as a food source and synthetic wadding as an oviposition substrate was compared to that in a natural rearing method based on the citrus mealybug Planococcus citri (Risso) (Hemiptera: Pseudococcidae) as prey and P. citri ovisacs as a substrate for oviposition. Because of the relatively high cost price of E. kuehniella eggs, we also investigated the nutritional value of frozen bee pollen to sustain the development and reproduction of C. montrouzieri and whether the production cost of the predator can be lowered further by mixing the flour moth eggs with bee pollen.

Materials and methods

Insect cultures

Laboratory colonies of C. montrouzieri and P. citri were established in 2010 with larvae acquired from Katz Biotech AG (Baruth, Germany) and Koppert B.V. (Berkel and Rodenrijs, The Netherlands), respectively. The mealybugs were cultured on potato sprouts and kept in a dark room at ambient conditions. Potatoes infested with mealybugs and covered with ovisacs were transferred to the stock colony of C. montrouzieri maintained at 25 ± 1 °C, 75 ± 5 % relative humidity (RH) and a 16:8 h (L:D) photoperiod.

Experiments on oviposition substrate and diet

In a first laboratory experiment, the potential of E. kuehniella eggs as a food source and the suitability of Rolta®Soft synthetic polyester wadding (1 × 1 cm) as an artificial oviposition substrate for C. montrouzieri (semi-artificial rearing method 1) was tested. E. kuehniella eggs were offered to larvae and adults in a 1.5 cm (Ø) plastic dish. Developmental parameters (immature survival, developmental time, adult body weight), reproductive traits (preoviposition period, total oviposition, egg hatch), sex ratio and longevity of predators reared under these conditions were compared to those of a control group offered P. citri nymphs as a diet and mealybug ovisacs as an oviposition substrate (natural rearing method). Because oviposition may be triggered by the presence of food for the developing larvae in the oviposition substrate, a treatment in which the females were offered polyester wadding sprinkled with E. kuehniella eggs in addition to the moth eggs supplied in a plastic dish was also included (semi-artificial rearing method 2).

In a second experiment, four food sources were tested. The value of frozen moist bee pollen as a (supplemental) food source to sustain the development and reproduction of C. montrouzieri was assessed. The performance of the ladybird on a mixture of E. kuehniella eggs and bee pollen (1:1 proportion) (food source 1) and on bee pollen alone (food source 2) was compared with that on P. citri nymphs (food source 3) or E. kuehniella eggs (food source 4). Both flour moth eggs and bee pollen were supplied by Koppert B.V. (Berkel and Rodenrijs, The Netherlands). Polyester wadding served as an oviposition substrate in all treatments, including the mealybug treatment.

In all treatments of both experiments, newly emerged first instar larvae (±50 larvae in both experiments depending on the availability of ladybird larvae and mealybugs) were taken out of the stock colony and placed individually in polystyrene Petri dishes (diameter 9 cm, height 2 cm). All foods used in both experiments were offered ad libitum and replenished every other day. Water was provided to larvae and adults by way of a moist wadding plug fitted into a 1.5 cm (Ø) plastic dish. Survival and development of C. montrouzieri larvae was monitored daily. Newly emerged adults were sexed and weighed using a Sartorius Genius ME215P balance, after which they were randomly paired. Adults received the same food as during their larval stage. Oviposition substrates were checked daily for eggs to determine the preoviposition period. Once the first egg was laid, substrates were replaced three times a week for a time period of one month (second experiment) or until the female died (first experiment). Oviposition rate and egg hatch were monitored during the first 30 days of egg laying (second experiment) or during the entire lifetime of the female (first experiment). Longevity of both males and females was determined. All experiments were conducted in a climatic chamber set at 25 ± 1 °C, a 75 ± 5 % RH and a 16:8 h (L:D) photoperiod.

Statistical analysis

All data were analysed using SPSS 21.0 (SPSS Inc. 2009). In both experiments, survival rates, egg hatch and sex ratios were compared by means of a logistic regression. This regression is a generalized linear model using a probit (log odds) link and a binomial error function (McCullagh and Nelder 1989). P-values below 0.05 were considered significant.

For the first experiment, the developmental parameters of predators reared according to the two semi-artificial systems were pooled as the larvae received the same treatment. A Kolmogorov–Smirnov test indicated that developmental time and adult body weight of naturally and semi-artificially reared individuals were normally distributed. These parameters were analysed using Student’s homo- or heteroscedastic t-tests when the Levene test indicated equal or unequal variances, respectively. The preoviposition period and longevity parameters were also normally distributed and analysed using a one-way analysis of variance (ANOVA). Means were separated using Tukey’s test as a Levene test proved homoscedasticity. A non-parametric Kruskal–Wallis test was used to evaluate differences in the number of deposited eggs among treatments.

For the second experiment, a Kolmogorov–Smirnov test again indicated that adult body weight and total number of deposited eggs were normally distributed. These parameters were therefore analysed using a one-way ANOVA and means were separated using Tukey’s test, after homoscedasticity was confirmed using a Levene test. Developmental times and preoviposition periods were not normally distributed and thus analysed using a Kruskal–Wallis test followed by Mann–Whitney U tests (SPSS Inc. 2009).

Results

Predator larvae reared in the first experiment on E. kuehniella eggs or P. citri mealybugs developed successfully with survival rates above 90 % (Table 1). Developmental time (t = 10.035, df = 25.887, P < 0.001 for males and t = 10.828, df = 62, P < 0.001 for females) and adult body weight (t = −2.542, df = 62, P = 0.014 for males, t = −3.038, df = 62, P = 0.003 for females) were affected by diet. Ladybirds fed E. kuehniella eggs developed nearly two days faster and weighed approximately 10 % more than those fed mealybugs. Sex ratios were proportionally divided in both groups (χ2 = 0.037, df = 1, P = 0.848).

Table 1 Survival, developmental time, adult body weight and sex ratio of C. montrouzieri ladybirds when reared according to a semi-artificial rearing system (with E. kuehniella eggs) or a natural rearing system (with P. citri mealybugs)
Full size table

Although the preoviposition period of females offered synthetic wadding as an oviposition substrate was prolonged by about eight days (F = 12.615, df = 2, 62, P < 0.001), they produced a similar number of eggs as naturally reared females (χ2 = 3.605, df = 2, P = 0.165). However, the hatching rate of the eggs laid in artificial substrates was significantly lower than that of eggs deposited near mealybugs (χ2 = 18.920, df = 1, P < 0.001 for semi-artificial 1 and χ2 = 21.659, df = 1, P < 0.001 for semi-artificial 2) (Table 2). No significant differences were recorded in reproduction parameters between the group that was offered clean synthetic wadding and the group that was offered wadding with imbedded E. kuehniella eggs (preoviposition period: P = 0.925 (Tukey post-hoc test); no. of deposited eggs: χ2 = 3.605, df = 2, P = 0.165; egg hatch: χ2 = 0.148, df = 1, P = 0.700; longevity: F = 0.477, df = 2, 125, P = 0.622).

Table 2 Reproduction and female longevity of C. montrouzieri reared according to a natural rearing system (food: mealybugs, oviposition substrate: mealybug ovisacs), semi-artificial rearing system 1 (food: E. kuehniella eggs, oviposition substrate: polyester wadding) or semi-artificial rearing system 2 (food: E. kuehniella eggs, oviposition substrate: polyester wadding with imbedded E. kuehniella eggs)
Full size table

The developmental parameters of C. montrouzieri fed the diets tested in experiment 2 are presented in Table 3. Larvae developed successfully on all diets (survival >90 %) except on bee pollen alone allowing only 20 % of the larvae to reach the 4th larval stage and only one male to reach the adult stage. Diet influenced both developmental time (χ2 = 39.834, df = 2, P < 0.001 for males and for χ2 = 38.033, df = 2, P < 0.001 for females) and adult body weight (F = 5.848, df = 2, 70, P = 0.006 for males and F = 6.314, df = 2, 68, P = 0.003 for females). Developmental rate (all P < 0.001; Mann–Whitney test) and female body weight (both P < 0.05; Tukey post-hoc test) of ladybirds fed E. kuehniella eggs or a mixture of E. kuehniella eggs and pollen were superior to those of their counterparts offered mealybugs. Body weight of males fed the mixture of E. kuehniella eggs and pollen was similar to that of males fed mealybugs (P = 0.182; Tukey post-hoc test). Except for pollen alone, sex ratios were essentially equal on all diets (χ2 = 0.015, df = 2, P = 0.993).

Table 3 Survival, developmental time, adult body weight and sex ratio of C. montrouzieri fed different food sources
Full size table

Preoviposition period was affected by diet (χ2 = 8.481, df = 2, P = 0.014) and was over four days shorter on mealybugs than on diets containing E. kuehniella eggs (Table 4). The diet offered to the ladybirds, however, had no significant effect on the number of eggs deposited by the predator in the synthetic wadding over a 30-day period (F = 0.155, df = 2, 68, P = 0.857) and on egg hatch (χ2 = 0.017, df = 2, P = 0.992).

Table 4 Reproduction (over 30 days from the onset of oviposition) of C. montrouzieri offered different food sources and synthetic wadding as an oviposition substrate
Full size table

Discussion

The semi-artificial rearing method based on E. kuehniella eggs as a factitious food and synthetic wadding as an artificial oviposition substrate successfully supported development and reproduction of the specialist mealybug predator C. montrouzieri. E. kuehniella eggs appeared to be an even better food source for the developing larvae of the predator than one of P. citri mealybugs: larvae reared on flour moth eggs developed faster and reached a higher adult weight than their counterparts reared on mealybugs. These findings are consistent with the results of Attia et al. (2011), suggesting that E. kuehniella eggs may constitute a nutritionally superior food for C. montrouzieri larvae than its natural prey, mealybugs. Earlier studies showed that a number of other predatory insects generally perform better when reared on E. kuehniella eggs than when offered their natural prey, which might be attributed to the well balanced amino acid and fatty acid composition of this factitious prey (Fauvel et al. 1987; Cocuzza et al. 1997; Specty et al. 2003; De Clercq et al. 2005; Berkvens et al. 2008b; De Clercq et al. 2013).

Artificially reared females were not only capable of producing viable eggs but also deposited the same number of eggs as naturally reared females. To our knowledge, this is the first report of C. montrouzieri ladybirds reproducing in the absence of mealybugs. Chemical cues produced by their natural prey were always considered to be a necessity to stimulate oviposition. For example, C. montrouzieri ladybirds reared on eggs of the flour moth E. kuehniella or of the grain moth S. cerealella depended on mealybug egg masses to trigger oviposition (Pilipjuk et al. 1982; Attia et al. 2011). Merlin et al. (1996) reported that C. montrouzieri females reared on P. citri mealybugs and offered cotton wool pellets as an oviposition substrate refrained from ovipositing and withheld their eggs until egg masses were met. Finlay-Doney and Walter (2012) noted that C. montrouzieri females were able to oviposit in the absence of ovisacs as long as mealybugs were offered as food source. In the latter study, the presence of mealybug egg masses had a significant impact on the number of deposited eggs, with significantly fewer eggs overall for females that were not provided with ovisacs but only with mealybug nymphs. Likewise, Attia et al. (2011) showed that ladybirds reared on E. kuehniella eggs and offered ovisacs deposited 30 % less eggs compared to females provided with mealybugs and ovisacs. In contrast to abovementioned studies, our results prove that C. montrouzieri can effectively be reared in the absence of mealybugs without a loss in fecundity.

The equally good reproductive rates observed in C. montrouzieri females reared according to the semi-artificial method as compared to the natural rearing system using mealybugs, can be attributed to the use of a nutritionally adequate food source for the larvae and adults of the predator and the presence of an acceptable substrate for egg laying by the adult females. The adequacy of both the factitious food and the artificial substrate for rearing the predator is confirmed by the fact that 15 consecutive generations could be produced using this semi-artificial method and in the complete absence of mealybugs in the rearing environment without evidence of loss in fitness. After 15 generations, female ladybirds had an average developmental time of 23.1 ± 0.2 days and reached an average adult body weight of 11.5 ± 0.2 mg. They laid an average of 313 eggs over a time period of 30 days. These data are comparable with (for developmental time and body weight) or even better than (for fecundity) obtained in the first generation of rearing on E. kuehniella eggs and synthetic wadding (Tables 3, 4).

It is unclear why the synthetic wadding used in this study was accepted as oviposition substrate while the cotton wool pellets used by Merlin et al. (1996) were refused. Latter authors asserted that thigmotactic cues are key for triggering the oviposition in C. montrouzieri, which stresses the importance of the physical properties of the oviposition substrate. A possible explanation for the acceptance of the polyester wadding used in the present study is that it is more successful in mimicking the filamentous structure of mealybug ovisacs than cotton wool pellets.

The semi-artificial rearing system proposed here does have its limitations. A first drawback is a prolongation in preoviposition period from seven days for the natural rearing method to 15 days for the semi-artificial method. When the ladybirds were fed mealybugs and provided with synthetic wadding an intermediate preoviposition period of ten days was found. These findings suggest that the oviposition of C. montrouzieri is postponed in the absence of ovisacs and potentially associated chemical cues produced by the mealybugs, but that these cues do not seem required per se if a suitable oviposition substrate is available. A significant decrease in egg hatch is the second drawback of the semi-artificial rearing system. When the ladybirds were provided with E. kuehniella eggs and synthetic wadding instead of mealybug ovisacs in the first experiment, egg hatch monitored over the entire lifetime of the female decreased from 72 to 60 %.

Imbedding E. kuehniella eggs into the oviposition substrate with the objective to accelerate reproduction, proved unsuccessful. Apparently, the availability of food for newly emerged larvae in the oviposition substrate does not trigger oviposition by the females of C. montrouzieri.

Ephestia kuehniella eggs were found to be an effective factitious food source to sustain the development and reproduction of C. montrouzieri and show promise to be implemented in a large scale production of this ladybird. As a downside E. kuehniella eggs are relatively expensive (De Clercq et al. 2013). The search for cheaper alternatives is recommended to lower the production cost of this and other insect predators further. Different pollens have been shown to be suitable alternative or supplementary food for the development and reproduction of various predatory arthropods (De Clercq et al. 2005; Berkvens et al. 2008a; Bonte et al. 2010; Vandekerkhove and De Clercq 2010; Nguyen et al. 2013; Vangansbeke et al. 2014). The frozen, moist bee pollen tested in the current study was not an adequate food source for immature C. montrouzieri as only one individual reached the adult stage, which indicates a lack or shortage of certain nutritional components essential for optimal development of C. montrouzieri. However, 60 % of the larvae reached the 3rd larval stage when fed pollen alone and no loss in fitness was recorded when the flour moth eggs were mixed with pollen. Dissected ladybird larvae fed the mixture of E. kuehniella eggs and pollen had high amounts of pollen in their gut, indicating that the larvae not only actively fed on the lepidopteran eggs but also on the pollen. Therefore, mixing E. kuehniella eggs with pollen may contribute to an increase in cost-effectiveness of mass production by reducing the input of the nutritionally superior but highly expensive flour moth eggs (De Clercq et al. 2005; Berkvens et al. 2008a; Bonte et al. 2010; Lundgren and Weber 2010; Pilorget et al. 2010; Lundgren et al. 2011; Weber and Lundgren 2011).

In conclusion, C. montrouzieri can be successfully reared on E. kuehniella eggs as a food source and synthetic wadding as an oviposition substrate. The major drawback of the semi-artificial rearing method is a delay in the onset of oviposition by about eight days. However, when the shorter developmental time of larvae fed E. kuehniella eggs is taken into consideration (two days), the semi-artificial rearing system yields a delay of only about six days compared to the traditional rearing method. Besides, the semi-artificial rearing method has an important advantage: C. montrouzieri can be reared without the need to maintain cultures of mealybugs and potato sprouts or other plant materials. This should make the mass rearing of C. montrouzieri less time-consuming, less labour-intensive and more cost-effective. A possible next step in rationalizing the rearing system for this ladybird is the development of an artificial diet (Riddick and Chen 2013). Chumakova (1962) described an artificial diet based on casein and amino acids for C. montrouzieri but larvae suffered from high mortality (Hodek 1967). Recent advances in the development of artificial diets for predatory arthropods (Morales-Ramos et al. 2013) may be helpful in designing an effective artificial diet for this ladybird as well.