The egg parasitoid, Trichogramma brassicae Bezdenko (Hymenoptera: Trichogrammatidae) is the most important and widely distributed species of Trichogramma in Iran. It attacks eggs of several lepidopterous pests, and is a major biological control agent. Rearing parasitoids is necessary for experimental work, and, potentially, for mass release in the field. Selecting a suitable host is critical for developing a successful rearing method. If other conditions are the same, the rate of population increase will be a suitable indicator of parasitoid performance on different hosts. However, conclusions based on a single generation can be misleading because of the learning ability of parasitoids. Life history parameters of T. brassicae were studied on two hosts easily reared in the laboratory, Anagasta kuehniella Zeller, and Plodia interpunctella Hübner (Lepidoptera: Pyralidae). All the experiments were carried out at 24 ± 1°C, 65±10% RH, and 16:8 L:D photoperiod. Eight parameters including gross and net reproductive rates (GRR and R0 respectively), intrinsic rate of natural increase (rm), finite rate of population increase (λ), intrinsic birth and death rates (b and d respectively), cohort generation time (T), and doubling time (DT) were compared between two hosts for two generations. All parameters showed a highly significant difference (α = 0.01) between hosts. GRR, R0, rm, λ, and b were higher, while d, T, and DT were lower in Anagasta than Plodia. The intrinsic rate of natural increase was 0.2912 and 0.2145 female/female/day and net replacement rate was 45.51 and 19.26 female/female/generation in Anagasta and Plodia respectively. Differences between generations were significant except for rm, λ, and d. The net replacement rate was 28.56 and 39 in the 1st and 2nd generations respectively. These results showed that A. kuehniella was a better host than P. interpunctella. Higher reproduction occurred in the second generation that may be due to increased adaptation to experimental conditions.
Introduction
Insect eggs belonging to 11 orders are attacked by different species of Trichogrammatidae parasitoids. Lepidopterous insects are the most preferred hosts (Sorokina 1999). Augmentation of parasitoids is used against some serious pests throughout the world (Li 1994; Smith 1996). Trichogramma brassicae Bezdenko (Hymenoptera: Trichogrammatidae) is the most important and widely distributed species of Trichogramma in Iran (Ebrahimi et al. 1998). Many Trichogramma species can be easily reared on laboratory hosts such as Anagasta kuehniella, Sitotroga cerealella, Plodia interpunctella, Galleria mellonella, and Corcyra cephalonica (Smith 1996; Shojai et al. 1998; Ebrahimi 2004). Furthermore, major crop pests such as the rice stem borer Chilo suppressalis, the European corn borer, Ostrinia nubilalis, the cotton bollworm Helicoverpa armigera, and some related species are target hosts of T. brassicae and related species (Li 1994; Dutton et al. 1996; Greenberg et al. 1998; Ebrahimi et al. 1998).
Success in biological control by Trichogramma depends on identification of the best host species, as well as a good understanding of the ecological requirements of the parasitoid wasp (Parra et al. 1987; van Lenteren et al. 1997). One of the most important aspects in ecology of the parasitoid is a suitable host. The size and age of the species chosen are important characteristics that determine the quality of a host (Hiehata et al. 1976; Houseweart et al. 1982; Bournier 1982; Brower 1983; Ahmad and Sivapragasam 1984; Pak 1986; Lewis and Redlinger 1989; Hintz and Andow 1990; Reznik and Umarova 1990; Ruberson and Timothy 1992; Monje et al. 1999; Jeffry and Robert 2000; Mansfield and Mills 2002; Roriz et al. 2006; El-Wakeil 2007). There are a few ecological studies on local populations in Iran that show trichogrammatids are affected by these factors (for example Karimian 1999; Attaran et al. 2000; Dadpour Moghanlou 2002; Haghani and Fathipour 2004; Alizadeh and Ebrahimi 2004; Karimi Malati et al. 2004; Shirazi 2004; Hosseini Bai et al. 2006). However demographic traits such as the intrinsic rate of natural increase or net reproductive rate have been estimated only in a few studies (Dadpour Moghanlou 2002; Haghani and Fathipour 2004). According to Andrewartha and Birch (1954) demographic parameters are the best indicators of fitness of a population and are suitable criteria for comparing physiological states of different species, populations, etc. or even as bioclimatic or nutritional indices (Messenger 1964; Dent and Walton 1997). Also, in biological control programs, population growth rate is an essential criterion for preliminary screening and choice of potential biocontrol agents (van Lenteren and Woets 1988).
To achieve a successful laboratory culture of T. brassicae it is necessary to study the effects of different ecological factors on important biological parameters of local populations of T. brassicae. As van Driesche and Bellows (1996) point out, rearing natural enemies on natural hosts is often difficult and expensive. Augmentation using alternative hosts is therefore usually necessary. Stored products moths often have been used traditionally for rearing Trichogramma species (Smith 1996). S. cerealella, A. kuehniella, and P. interpunctella are not only alternative hosts but also obvious target pests (Smith 1996; Shojai et al., 1998; Ebrahimi 2004). Availability due to worldwide distribution, as well as an inexpensive rearing method (Sepasgozarian 1966; Behdad 2002) are the most important advantages of these species for mass production purposes.
Natural enemies are able to learn (Bigler 1994; van Driesche and Bellows 1996). This ability allows them to improve their response to their host when they are reared sequential generations on a host (van Driesche and Bellows 1996). This occurs via enhancing their skills in orientation, host finding, host detection and acceptance (Noldus et al. 1990). Van Bergeijk et al. (1989) observed different responses in T. brassicae to European corn borer when it was reared previously on the same host compared to A. kuehniella.
In this study, life history statistics are examined for T. brassicae on two laboratory hosts, A. kuehniella, and P. interpunctella in order to explore if there is any advantage to each one. As the null hypothesis we assumed that no differences were present between these hosts so that both are similar in quality for their host. If this is the case, then either host can be used equally well for production purposes, assuming all other conditions are the same. Rejection of the null hypothesis means that the hosts are not of the same value for rearing and the best host can be used. As mentioned earlier, continuous rearing on a host can improve the impact of a parasitoid. If this is the case, then considering only one generation in host evaluation may be misleading. For example if continuous rearing on a host leads to improvement in parasitoid performance on one of the hosts, then it may be recommendable using it in sequential cultures in spite of its initial low performance. Thus the second null hypothesis is that the parasitoid has equal performance in sequential generations.
Materials and Methods
Host cultures
A culture of the Indian meal moth, Plodia interpunctella Hübner (Lepidoptera: Pyralidae), was reared on pistachio in the Department of Plant Protection, Faculty of Agriculture, University of Tabriz. To adapt the culture to the laboratory conditons it was reared two generations on single cross hybrid of maize kernels at 24±1°C, 65±10% RH, and 16:8 L:D photoperiod in an incubator. Plastic rectangular containers (201×4×6 cm) were used for rearing larvae.
A culture of the flour moth, Anagasta kuehniella Zeller(Lepidoptera: Pyralidae) was reared on wheat flour in the same location. It was reared two generations on wheat flour cultivar Omid under the conditions described above. Similar containers were used for this purpose.
Cylindrical containers (30cm in height, 11 cm in diameter), were used to confine moths to obtain eggs. Ten pairs of adult moths were removed by an aspirator and transferred to a cylindrical dish. The top of the dish was covered by a 10 mesh cloth net and set upside down on a piece of paper. Deposited eggs were collected every day and used in experiments.
Preparation of parasitoids for experiments
The Trichogramma brassicae used in this study, were collected from eggs of the cereal moth, Sitotroga cerealella (Olivier) (Lepidoptera: Gelechiidae), obtained from the Agricultural Research Center of East Azerbaijan, Tabriz. These wasps had been already reared on the S. cerealella for five generations. To adapt it to experimental conditions it was reared for two generations under the conditions described above using S. cerealella eggs as the host. Honey droplets were used to feed wasps.
One-day-old eggs of both hosts were offered to newly emerged females of T. brassicae for 24 hours. Female wasps were then removed from host eggs and they were held under the same conditions until symptoms of parasitism appeared. Forty parasitized eggs from each host were selected randomly and the parasitoids were reared to adults. Each surviving female was then confined with a male that had been reared under the same conditions. Males were replaced when they died. Fifty eggs of the same host species that parents had been reared on were offered daily to each pair of wasps up to the time of death of the female. The number of eggs offered daily to the wasps was determined based on a functional response experiment under the same conditions (Farazmand and Iranipour 2006) in which the maximum attack rate was 40 hosts/wasp/day obtained for A. kuehniella. Cohorts of the second generation were selected at the peak of oviposition.
Table 1.
Summary of analysis of variance (F and p-values) for eight life history parameters of T. brassicae on two host species A. kuehniella, and P. interpunctella for two generations*

Experimental design and data analyses
The experiment was designed as a 2×2 factorial with completely randomized design. One of the factors was host with two levels A. kuehniella and P. interpunctella. The other factor was generation with two levels. Due to unequal mortality among treatments, the analysis was converted to an imbalanced design. Eight parameters including gross reproductive rate (GRR), net replacement rate (R0), intrinsic rate of natural increase (rm), finite rate of increase (λ), intrinsic birth rate (b), intrinsic death rate (d), cohort generation time (T), and doubling time (DT) were calculated as described by Carey (1993). Variances and standard errors were measured using the jackknife method of Meyer et al. (1986). Analyses of variances and comparisons of means were carried out using SAS software (SAS Institute Inc.). Survivorship curves were drawn using data in column lx and type was determined by entropy (Demitrius 1978; Carey 1993).
Results
A summary of the analyses of variances for the eight above-mentioned statistics are shown in table 1. All the parameters differ highly significantly (p <0.01) between hosts, as GRR, R0, rm, λ, and b are higher while d, T, and DT are lower in A. kuehniella than P. interpunctella (Table 2). These results reveal that A. kuehniella is a more desirable host than P. interpunctella for T. brassicae. There are also significant differences in all parameters except for rm, λ, and d in the second generation. All parameters but doubling time were higher in second generation (Table 2). Because R0 and T have opposite effects on rm, the latter parameter showed no significant difference in the two generations. These data suggest that T. brassicae has habituated to hosts as well as other experimental conditions during the first generation. Interactions between two factors were significant in all but three parameters (GRR, λ and d). This means that level of adaptation during the second generation was not equal between the two hosts. Components of life history statistics in each host and generation are shown in Table 3.
Table 2.
Summary statistics of life history of T. brassicae on two hosts in two generations (mean ± 95% confidence interval).

Table 3.
Life history statistics of T. brassicae in each host and generation separately (mean ± 95% confidence interval).

Gross reproductive rate
The gross reproductive rate (GRR) was higher in A. kuehniella than P. interpunctella by a factor of 1.4 (F = 1265.1, P <0.01) (Table 2).
Net reproductive rate
The net reproductive rate (R0) of A. kuehniella was more than twice as high as P. interpunctella (F = 4709.9, P <0.01). A 36.55% increase occurred between the first and second generations (Table 2).
Intrinsic rate of natural increase
The intrinsic rate of natural increase (rm) was significantly higher for A. kuehniella than P. interpunctella (F = 3618.8, P<0.01). There was no significant difference between generations (F = 0.05, P = 0.82) (Table 2).
Finite rate of population increase
The finite rate of population increase (λ) is calculated easily from the intrinsic rate of natural increase. Therefore its changes follow the same pattern.
Intrinsic birth rate and death rate
The balance between these statistics (b and d) determines population growth rate. They changed in an inverse direction, as the birth rate was higher (F = 5781.3, P <0.01) and the death rate was less (F = 7.64, P <0.01) in A. kuehniella than P. interpunctella. So the intrinsic rate of natural increase (rm) was higher in A. kuehniella (Table 2). The intrinsic birth rate also was higher in second generation (F = 9.52, P <0.01), while the intrinsic death rate was similar in both (F = 3.45, P = 0.06) (Tables 2 and 3).
Cohort generation time
Cohort generation time (T) was significantly longer in P. interpunctella than A. kuehniella (F = 460.8, P <0.01). There was also a significant difference between the two generations (F = 863.3, P <0.01), as in second one was one day more (12.93 ± 0.066 vs. 13.92 ± 0.222) (Table 2). However, the differences were small and significance of such small differences means a low variance in this statistic due to the limited reproductive period.
Doubling time
In T.brassicae the population doubles every 2.38 ± 0.022 days for A. kuehniella and every 3.23±0.026 days for P. interpunctella. The differences were significant (F = 3944.5, P <0.01). The difference in doubling time between generations was small but significant (F = 5.2, P = 0.02) (Table 2).
Survivorship curves
A survivorship curve similar type I of Slobodkin (1980) was observed in all treatments (Figure 1). In order to determine curve type, entropy (Carey 1993) was used as a criterion. An entropy amount below 0.5 shows a survivorship curve near to type 1. During two generations, entropy was 0.28 and 019 in A. kuehniella, and 0.31 and 0.30 in P. interpunctella. Both survivorship and age specific fecundity are shown in Figure 1.
Discussion
The results of this study revealed that A. kuehniella is a more acceptable host for T. brassicae than P. interpunctella. A higher GRR, R0, rm, λ, and b suggest higher reproduction of T. brassicae on A. kuehniella, whereas lower d, T, and DT may be due to an accelerated development of A. kuehniella. Differences in total fecundity, developmental time and other statistics related to life history have observed among hosts in different Trichogramma species (e.g. Shirazi 2004; Hosseini Bai 2006; Roriz et al. 2006; El-Wakeil 2007). Shirazi (2004) found a shorter developmental time, and a higher daily fecundity in females of Trichogramma chilonis on Corcyra cephalonica compared to its natural host, Helicoverpa armigera. Larger T. brassicae females with higher fecundity were obtained when reared on the cereal moth compared to the Indian meal moth (Hosseini Bai et al. 2006). An average of 34.55 and 21.61 eggs were laid under non-feeding conditions respectively. Significant differences in fecundity and developmental time were observed in Trichogramma cordubensis on different noctuid hosts (Roriz et al. 2006). El-Wakeil (2007) also found a higher longevity, parasitism and emergence rates on target host H. armigera than three factitious hosts.
The gross and net reproductive rates, and intrinsic rate of natural increase in T. brassicae that were found in the present study are well in the range of Trichogramma embryophagum and Trichogramma pintoi respectively as reported in the Haghani and Fathipour (2004) and Dadpour Moghanlou (2002) studies. GRR, R0 and rm in T. embryophagum were 55.24, 48.88, and 0.238 reared on the flour moth, Ephestia kuehniella and 41.74, 37.63, and 0.218 when reared on S. cerealella respectively (Haghani and Fathipour 2004). On the other hand, the above mentioned parameters in T. pintoi were estimated to be 46.30, 45.30, and 0.257 when reared on A. kuehniella, and 49.19, 45.68, and 0.281 when reared on S. cerealella respectively (Dadpour Moghanlou 2002).
Pratissoli and Parra (2000) calculated these statistics for Trichogramma pretiosum and Trichogramma acacioi in five constant temperatures, 15, 20, 25, 30 and 35°C. In those temperatures, R0 was determined to be 13.98, 39.44, 31.53, 54.97, and 15.54 for T. pretiosum and 11.85, 62.89 20.64 20.42, and 9.36 for T. acacioi at each temperature respectively. Furthermore, rm was calculated as 0.05, 0.21, 0.32, 0.47, and 0.36 in former species while it was 0.05, 0.22, 0.35, 0.34, and 0.30 in later species at the same temperatures respectively. Their results do not resemble T. brassicae in this study totally, but there is some similarity at intermediate temperatures (20–30°C). For example R0 in T. acacioi at 25 and 30 °C, as well as rm in both species at 20 °C is near to their values in T. brassicae on P. interpunctella in the present study. Haile and Hassan (1999) found an intrinsic rate of increase of 0.309 in Trichogramma bournier reared on S. cerealella that is very near to T. brassicae reared on A. kuehniella in our study.
Birth rates in T. embryophagum and T. pintoi found by Haghani and Fathipour (2004) and Dadpour Moghanlou (2002) were similar to T. brassicae in this study, while death rates were a little higher in our study. Both T and DT in our study resemble T. pintoi found by Dadpour Moghanlou (2002), as one generation took 14.81 and 13.59 days on E. kuehniella and S. cerealella respectively (Dadpour Moghanlou 2002). It was considerably longer (16.37–16.49 days) in T. embryophagum (Haghani and Fathipour 2004). Doubling time in their studies was between 2.46 and 3.77 days in different experiments. Differences observed in all the statistics between the present study and those of the other studies may be due to variation in species and population sources of parasitoid and/or hosts, rearing background, nutritional state, physical conditions and even analytical approaches.
Lower entropy can also result in a higher survival rate when T. brassicae is reared on A. kuehniella. On the other hand, a significant increase in GRR, R0, and b may due to increased reproduction and adaptation of T. brassicae to hosts following one generation of rearing. This is in full agreement with van Bergeijk et al. (1989). It should not to be forgotten that the degree of adaptation is not equal in the two hosts and this is why interaction between host and generation was significant in reproductive rates. Absolute change in R0 was higher in A. kuehniella. Indeed, the net replacement rate increased significantly with generation whereas rm did not suffer parallel changes. This suggests a trade off between R0 and T. In other words generation time lengthened while fecundity increased and population growth rate remained unchanged.