The eastern spruce budworm, Choristoneura fumiferana (Clemens) (Lepidoptera: Tortricidae), is commonly reared on artificial diet (McMorran Reference McMorran1965) for both laboratory and field experiments. Certain differences exist between foliage-fed and artificial diet-fed larvae, such as higher survival of larvae, increased pupal weight, adult longevity, fecundity and development times (McMorran Reference McMorran1965). Poirier and Borden (Reference Poirier and Borden2000) observed that foliage-fed larvae did not respond to the oral exudates of artificial diet-fed larvae, whereas exudates from foliage-fed larvae had a deterrent effect on the feeding behaviour of both foliage-fed and artificial diet-fed larvae. Several studies used spruce budworm larvae reared on artificial diet for exposure in the field to investigate parasitism and parasitoid behaviour in endemic budworm populations (Doucet and Cusson Reference Doucet and Cusson1996a, Reference Doucet and Cusson1996b; Fidgen and Eveleigh Reference Fidgen and Eveleigh1998; Fidgen et al. Reference Fidgen, Eveleigh and Quiring2000; Cusson et al. Reference Cusson, Laforge, Régnière, Béliveau, Trudel and Thireau2002).

It is well known that variation in the food consumed by herbivores can change their chemical composition and affect their susceptibility to predator attack (reviewed by Price et al. Reference Price, Bouton, Gross, McPheron, Thompson and Weis1980). For instance, several plant compounds such as toxins assimilated by Lepidoptera can alter parasitoid performance (reviewed by Ode Reference Ode2006). However, few studies have dealt with influences of artificial diets compared to natural diets of Lepidoptera on parasitism (but see Song et al. Reference Song, Bourchier and Smith1997; Havill and Raffa Reference Havill and Raffa2000). So far only the effect of host diet on parasitism of spruce budworm eggs had been tested (Song et al. Reference Song, Bourchier and Smith1997) but no attempt has been made to test if larval parasitoids of the spruce budworm are influenced by the artificial diet ingested by their host. Therefore, both field and laboratory experiments were conducted to test the use of diet-fed spruce budworm larvae for parasitism studies.

The field experiment was conducted in summer 2011 in the context of a larger study on the influence of partial cutting on parasitism (Seehausen Reference Seehausen2012), in Laval University's Montmorency experimental forest (47°20′N, 41°30′W). This boreal forest is located in the Laurentian Mountains, 70 km north of Quebec City, Québec, Canada, in a balsam fir (Abies balsamea (Linnaeus) Miller; Pinaceae) – white birch (Betula papyrifera Marshall; Betulaceae) forest. Four plots, each ∼4 ha in size, were located in unaltered, mature forests. Plot elevations ranged from 650 to 800 m on a north-eastern slope.

Diapausing second instar spruce budworm larvae were obtained from the Canadian Forest Service, Sault Ste. Marie, Ontario, Canada, and reared at 23 °C and a 16-hour photoperiod on either antibiotic-free artificial diet (modified from McMorran Reference McMorran1965) or foliage of balsam fir harvested in the study area. Foliage was stored at 4 °C for not longer than one week in buckets filled with water and offered to larvae as small branches in water-filled tubes. Branches in rearing containers were changed twice a week to provide larvae with fresh food. Larvae were held in plastic containers with food ad libitum until the desired instar for the experiment was reached (7–20 days).

To compare parasitism rates of budworm larvae, fourth to sixth foliage-fed and diet-fed instars were implanted twice a week from 21 June to 19 July 2011 in each of the four plots. Before implantation, larvae were held in small plastic containers without food for 24 hours at 4 °C in growth chambers. They were transported to the study area in cooling boxes and installed on current-year balsam fir shoots at eye level in groups of five larvae with the same treatment distributed on two to five young trees. Trees and branches were marked with coloured flags to facilitate recovery of larvae after exposure. Every group was randomly distributed in each plot and at least 15 m apart from each other. To synchronise the implantation of the larvae with their natural occurrence and the occurrence of the parasitoids in the study area, a model of spruce budworm seasonality (Régnière et al. Reference Régnière, St-Amant and Duval2012) was used. Larvae were recovered at nine dates after a seven-day exposure period (Fig. 1). Shoots containing the exposed budworms were cut so that immature stages of parasitoids (larvae or cocoons) hidden in the foliage in the vicinity of their host were also collected. After transport to the laboratory in a cooling box, larvae were placed in individual containers on artificial diet, reared at room temperature, and checked three times a week for parasitoid emergence. The proportion of parasitised recovered larvae was analysed using logistic regression, with treatments (diet) as fixed effects, plots as random effects, and time as repeated measures (PROC GLIMMIX, SAS Institute 2003). Repeated measures were needed because of the likely autocorrelation of the plots effect error. Individuals dying from causes other than parasitism were not included in the analysis.

Fig. 1 Mean rates of (A) overall attack; (B) attack by Tranosema rostrale; (C) attack by Elachertus cacoeciae; and (D) attack by other parasitoid species, as a function of time (ordinal date) in 2011 (black: foliage-fed larvae; dashed grey: diet-fed larvae).

A laboratory choice test was conducted in 2011 and 2012 with a total of 33 female specimens of the parasitoid Tranosema rostrale (Brischke) (Hymenoptera: Ichneumonidae) obtained from field implanted spruce budworm larvae in and near the study area. One foliage-fed and one diet-fed larva of the same instar (fourth or fifth) and size were placed at opposite sides in a 100 × 15 mm petri dish. Using an aspirator, one inexperienced wasp (a wasp that had never before parasitised a larva) was transferred to the petri dish for 10 minutes or until it had attacked one host larva. A larva was considered as attacked if the wasp's ovipositor was clearly inserted in its skin. If no parasitism occurred, the larvae were transferred to a new petri dish where the experiment was repeated using another inexperienced wasp. The choice test was analysed using a two-tailed sign test (PROC UNIVARIATE, SAS Institute 2003).

The overall parasitism rate in the field experiment was 86.2%, most of which was done by two species: T. rostrale and Elachertus cacoeciae (Howard) (Hymenoptera: Eulophidae). Phytodietus vulgaris Cresson (Hymenoptera: Ichneumonidae) (a probable determination) and a few unidentified individuals were also present (hereafter referred to as other species; Table 1; Fig. 1). No effect of larval diet on parasitism was found, whether overall or by species (Table 1). Overall parasitism decreased significantly throughout the season (F _8,19.67 = 2.94; P = 0.0244), although this effect was not significant when parasitism was analysed separately for each parasitoid species (Fig. 1). These changes over time reflect the different phenology of each species, T. rostrale being more active early in the summer, and E. cacoeciae later (J. Régnière, unpublished data).

Table 1 Percent parasitism (mean and SEM, n = 4 × 9) and statistical analysis of parasitism of foliage-fed (n = 325) and diet-fed (n = 347) spruce budworm larvae.

*Two larvae were multiparasitised by T. rostrale and E. cacoeciae; they were excluded from the species-specific analysis but included in total parasitism.

In the choice test with T. rostrale females, spruce budworm larvae fed on balsam fir foliage were chosen more than twice as often (70%, 23/33) as diet-fed larvae (30%, 10/33; P = 0.0351). While the sample sizes in these tests are small, the difficulty of obtaining sufficient biological material precluded access to larger number of female parasitoids. We hypothesise that T. rostrale uses olfactory cues from the budworm's host plant to find its host, such as volatiles liberated by feeding larvae or their frass, and therefore find or recognise foliage-fed larvae more readily. It is possible that in the first few hours of exposure in the field, foliage-fed larvae are also attacked more often than diet-fed larvae by T. rostrale females. However, an exposure period of seven days as described above would mask such an effect as the diet-fed larvae feed on foliage.

We conclude that implanting in the field of spruce budworm larvae reared on antibiotic-free McMorran (Reference McMorran1965) artificial diet can be used without concern that the food source used in rearing would affect parasitism rates.