Testing the impact of laboratory reared indigenous leafroller (Lepidoptera: Tortricidae) parasitoids (Hymenoptera: Ichneumonidae, Braconidae) on sentinel hosts in controlled orchard releases

Leafrollers can experience high levels of indigenous parasitism in organically managed apple orchards and the augmentative release of specific parasitoid species to suppress these secondary pests may be advantageous in orchards converting to nonchemical pest management. Caged and uncaged releases of two ichneumonid [Apophua simplicipes (Cresson) and Glypta variegata Dasch] and two braconid (Macrocentrus linearis Nees and Apanteles polychrosidis Viereck) koinobiont endoparasitoids of the obliquebanded leafroller, Choristoneura rosaceana (Lepidoptera: Tortricidae) on host-infested potted apple trees were conducted to assess the parasitoids’ abilities to find and successfully parasitize sentinel hosts under orchard conditions. Seasonal timing of the trials varied for each parasitoid species, based on their relative performance under simulated summer/fall conditions in laboratory trials. After the release of five or fifty parasitoid females, the mean percent parasitism of leafroller larvae collected from infested trees ranged from 0 to 75% depending on the parasitoid species involved. Although caged releases tended to increase the percentage of live parasitized hosts in release treatments, uncaged releases provided a more realistic assessment of the parasitoid’s ability to seek and find hosts within an infested area over a longer period. Release of the large, solitary A. simplicipes, had the most significant impact on the host population density.


INTRODUCTION
Apple orchards in the interior of British Columbia, Canada are part of an area-wide sterile male codling moth, Cydia pomonella (L.) (Lepidoptera: Tortricidae) release program (Dyck & Gardiner, 1992).Where this biological control strategy has been successful, there has been a substantial reduction in the use of codling mothtargeted chemical insecticides that also suppressed other susceptible secondary orchard insects such as leafrollers (Lepidoptera: Tortricidae).High levels of obliquebanded leafroller, Choristoneura rosaceana (Harris) (Lepidoptera: Tortricidae) parasitism have been recorded in certified organically managed orchards within this program (Cossentine et al., 2004a) and the orchardists frequently rely on high parasitism to suppress leafroller populations.It would be advantageous to augment the incidence of efficacious leafroller parasitoid species in formerly chemically managed orchards which no longer require multiple codling moth targeted chemical sprays and do not host leafroller parasitoids, however studies are required to assess the ability of specific parasitoid species to effectively find and parasitize hosts in orchard releases.
Apophua simplicipes (Cresson), Glypta variegata Dasch (Hymenoptera: Ichneumonidae), Macrocentrus linearis Nees and Apanteles polychrosidis Viereck (Hymenoptera: Braconidae) are common indigenous, koinobiont endoparasitoids of early-instar obliquebanded leafroller larvae in organically managed orchards in the southern region of the above codling moth area-wide con-trol program (Cossentine et al., 2004a).All four parasitoid species emerge from leafroller larvae in the spring, after the potentially bivoltine early instar host has overwintered and emerged to feed and mature on apple blossoms, foliage and developing apples (Chapman et al., 1968;Cossentine et al., 2004a).Adult obliquebanded leafrollers from overwintering larvae produce a summer generation and a second generation occurs in late summer and early fall (Madsen & Procter, 1982).All of the above parasitoids, except possibly the G. variegata, are also bivoltine as they have been recorded to emerge from the summer leafroller generation (Cossentine et al., 2004a).
The biology of each of the four parasitoids has been studied under laboratory conditions (Cossentine et al., 2004b(Cossentine et al., , 2005(Cossentine et al., , 2007)), but laboratory derived data may not accurately reflect the parasitoids' potential performance when released into real agro-ecosystems (Van Lenteren & Woets, 1988).A recommended step in the evaluation of parasitoid species for augmentative release is to test the ability of each candidate to find the host on the correct crop under realistic conditions (Van Driesche & Bellows, 1996).Obliquebanded leafrollers are not easily communally reared, making mass-rearing of the parasitoids expensive and wild host larval population densities are unreliable.Both these factors limit the success of orchard release trials.In this study, a more controlled field release was designed that required reasonably low numbers of parasitoids and ensured a sufficient number of available hosts.Laboratory-reared female parasitoids were released at two rates on caged and uncaged obliquebanded leafroller infested potted trees in order to assess the ability of each laboratory species to find and parasitize the host under realistic semi-orchard conditions.Although the response of parasitoids to host density is important in determining their efficacy (Bezemer & Mills, 2001) it was not evaluated in these trials as it is difficult to accurately determine the number of early instar leafroller that are established per tree without causing larval movement and mortality.

Infested potted trees
Fifty-four (and one week later, an additional 27) two to four year-old potted Royal Gala apple trees, similar in size to trees found in newly established high-density orchard plantings, were infested with laboratory colony derived neonate obliquebanded leafrollers before each trial to allow them to develop to second instars.Each tree either had a leafroller egg mass (75 to 150 eggs per mass, Madsen & Procter, 1982) stapled to a leaf from which neonates emerged within one or two days, or neonates were manually placed on the tree with a brush.Successful establishment was similar for both life stages.The trees were infested outdoors inside a 2.4 × 3.7 × 2.4 m screen cage to avoid exposure to parasitoids and predators.Obliquebanded leafroller density between trees was assumed to be relatively uniform due to the close proximity of the trees (touching) while infested and the active dispersive nature of the neonates (Carrière, 1992).Three 2.4 × 3.7 × 2.4 m screen cages with zippered doors, supported by metal frames, were placed adjacent to a pesticide-free apple orchard.Nine randomly chosen leafroller-infested potted apple trees were placed inside each cage in an X pattern.Three groups of nine randomly chosen trees were placed uncaged in a similar pattern within the orchard; each group was widely separated over an one ha area.Four 59.2-ml plastic cups, containing cotton wicks soaked with 10% honey water and supported by 1 m sticks, were placed as parasitoid feeding stations within each treatment.The honey water was replenished as required.

Parasitoids
All female parasitoids used in the releases were from laboratory colonies reared on obliquebanded leafrollers on an agarbased pinto-bean diet (modified from Shorey & Hale, 1965).They were all five day old adults, allowed to mate and one day prior to release, all were exposed to early instar leafrollers feeding on apple leaves.During this pre-release exposure the female parasitoids actively antennated and probed the infested leaves.This pre-release conditioning was included to maximize the recognition of hosts under field conditions, as parasitoids reared on artificial diets do not come into contact with all wild host chemical clues (Noldus, 1989;Lewis & Martin, 1990).
On the first day of the trials, 0 (control), 5 (low release) or 50 (high release) parasitoid females (and an equivalent number of males) were released on the centre caged and uncaged trees.One week later, all leaves on the potted trees within each treatment were inspected and those with evidence of leafrollers were removed.The trees were replaced with new second instarinfested potted trees in the uncaged half of the study only.There was potential for surviving parasitoids to enter and exit cages while trees were exchanged, therefore infested potted trees were not placed inside cages for a second week.Movement between treatments, natural parasitism and predation would be expected to occur in the uncaged treatments.Additional parasitoids were not released and seven days later, all leaves with evidence of leafrollers were removed.

Data collection
All removed leaves were examined under a dissecting microscope to find associated leafroller larvae.The large dark eggs of A. simplicipes and G. variegata are visible through the host integument, therefore for these assessments, collected leafrollers were classified as parasitized (one or more egg visible), unparasitized (no egg visible), dead or dead and parasitized.In the two braconid trials, collected leafrollers were classified as dead or alive.All living larvae (parasitized or not) were placed on the meridic diet mentioned above, in separate 30 ml plastic cups and held at 25°C until host or parasitoid pupation.Parasitoid emergence constituted successful parasitism.

Timing of trials
The trial was replicated twice over time for each parasitoid species over a three-year period with no temporal overlap among the trials.Test dates for each species were based on laboratory assessments and when early instar wild obliquebanded leafrollers would be encountered in orchards.In previous laboratory trials, A. simplicipes fecundity was not significantly different when the females were held at a constant 25 versus 15°C (Cossentine et al., 2004b).Apophua simplicipes field trials were carried out September 6 to October 11, 2002 (average maximum and minimum temperatures: 20.2 ± 4.7°C and 5.3 ± 3.1°C) to coincide with the natural presence of wild second instar larvae in the orchards.To date, G. variegata have only been found emerging from overwintering hosts in the region of the trials, therefore the field test was conducted June 13 to July 14, 2003 (average maximum and minimum temperatures: 27.1 ± 4.4°C and 12.8 ± 3.2°C), when wild adult G. variegata would be expected to be seeking hosts.In previous laboratory trials, parasitism of obliquebanded leafrollers by M. linearis was not significantly different when the females were held under summer conditions (fluctuating daily between 30.3 to12.0°C)versus fall conditions (Cossentine et al., 2005); M. linearis field trials were carried out August 1 to September 3, 2003 (average maximum and minimum temperatures: 30.3 ± 2.9°C and 14.9 ± 2.0°C) when wild summer generation larvae would potentially be present in the orchard.In previous laboratory trials, A. polychrosidis fecundity was significantly higher when the females were held under fall conditions (fluctuating daily between 17.5 to 4.0°C) versus summer conditions (Cossentine et al., 2005), therefore A. polychrosidis field trials were carried out September 13 to October 4, 2004 (average maximum and minimum temperatures: 18.1 ± 3.0°C and 7.8 ± 2.3°C).

Data analysis
The effects of replication, cage and the parasitoid release treatments on total hosts recovered, percent dead host larvae and percent parasitism by the released species as well as wild species were determined with a split-plot analysis of variance (ANOVA) (SAS, 2000) for each week of the study, with the effect of trees nested in treatments.Data were also analyzed separately for caged or uncaged status.Percentages were arcsine-transformed before analysis.The difference between individual treatment means were determined using Tukey's Studentized range test (SAS, 2000) (Tables 1-3).

Total sentinel host larvae recovered
In the first week of the trials, neither replication nor the presence or absence of the cages had significant (P < 0.05) impact on the total number of obliquebanded leafroller larvae collected in the parasitoid trials.The parasitoid release treatments however did have significant effects on the total number of host larvae collected in the two ichneumonid trials: A. simplicipes (F2,78 = 31.3,P < 0.0001) and G. variegata (F2, 78 = 9.2, P = 0.0003).In both cases, generally fewer larvae were collected in the release treatments than in the controls in the first week (Table 1).The number of larvae establishing on the potted trees in the first week of the A. simplicipes trials was higher than for all other trials and it is possible that these high artificial infestations may have influenced the significant reduction in the number of leafroller larvae collected, however, the G. variegata trials had host levels that were more consistent across trials.The parasitoid release treatments had no significant (P > 0.05) effect on the total number of obliquebanded leafroller larvae collected in either of the two braconid release trials.

Caged trials
When only the data from the caged trials were analyzed, in week one, significantly fewer leafroller larvae were found on the trees exposed to the high A. simplicipes (F2,26 = 20.5, P < 0.0001) and G. variegata release treatment (F2,26 = 9.3, P = 0.0009) compared to the number found in control cages (Table 1).

Uncaged trials
Significantly fewer leafroller larvae were collected from the low and high A. simplicipes release trees in week one in the uncaged trees (F2,26 = 10.5, P = 0.0005) than from control trees.The release of the other parasitoid species had no significant effect on the total number of leafroller larvae collected in uncaged treatments.

Parasitism
In the first week of the trials, replication had no significant impact (P > 0.05) on the level of parasitism in any of the parasitoid trials and the presence or absence of cages only had a significant impact on parasitism in the G. variegata (F1,1 = 3,220; P = 0.01) trials with more parasitism occurring in the cages (Table 2, Fig. 1).
The mean percentages of the collected hosts that were parasitized by A. polychrosidis (low release: = 6 and x high release = 14%) and M. linearis (low release = 0 x x and high release = 12%) were relatively low in the x caged trials (Table 3, Fig. 2).Death of host larvae due to parasitism was not discernable for these two braconids as parasitized larvae did not show evidence of their parasitized status as early instars.However, the percentage of dead hosts at the time of collection was not significantly larger (P > 0.05) in the release treatments compared to the control for either parasitoid species.

Uncaged treatments Caged treatment
Mean % of collected OBLR larvae per release treatment Time collected and host status post-release TABLE 2. Mean percent of obliquebanded leafroller (OBLR) (± SE) collected that were alive and parasitized (parasitoid developed), dead and parasitized (egg visible), dead and parasitized by other species one and two weeks post-release of 0 (control), ment versus the control in the caged treatment in week one (F2,26 = 9.6, P = 0.0008) and in the uncaged treatment in week two (F2,26 = 4.0, P = 0.03).

Dead hosts
When the cage effect was included in the analysis of each release trial, neither replication nor the presence or absence of a cage had a significant effect (P < 0.05) on the number of dead, unparasitized leafroller host larvae found in any of the parasitoid release trials, although a high number of dead hosts were found on uncaged control and release trees in both weeks of the M. linearis trial (Table 3).This may have been due to the warmer summer temperatures as the cages did buffer temperature and humidity which may account for the lower number of dead larvae collected in caged treatments.In the first week of the trials, only caged A. simplicipes releases (F2,26 = 8.3, P = 0.002) and in the second week, uncaged A. polychrosidis releases (F2,26 = 3.6, P = 0.04) had significant impacts on the percent of dead hosts with more dead hosts being found in the control trees.This indicates that some of the dead and parasitized hosts in the release treatments probably would have died even if not parasitized.

Parasitism by other species
Parasitism by wild species in the uncaged releases was highest in the spring G. variegata trials at 2 to 32% (Table 2, Fig. 1), lowest in the summer M. linearis trials at 0 to 4% (Table 3, Fig. 2) and ranged from 1 to 14% in the A. simplicipes and from 7 to 20% in the A. polychrosidis trials in the fall.In the fall release trials of A. simplicipes and A. polychrosidis, 70 and 74%, respectively, of the wild parasitism was by an unidentified ichneumonid species, 18 and 14%, respectively by unidentified eulophid ectoparasitoids, 21% by Apanteles species in the Apophua trials, and 0.1% by wild Apophua in the Apanteles trials.In the spring Glypta trials, only 3% of the wild parasitism was by unidentified eulophid ectoparasitoids, whereas 69% was by Apanteles species, 8% by an unidentified ichneumonid species, 15% by Apophua species and 5% by Meteorus species.Of the wild parasitism that did occur in the Macrocentrus uncaged releases, 51% was by an unidentified eulophid, 22% by an unidentified ichneumonid species, 24% by Apanteles species and 2% by Meteorus species.

DISCUSSION
In this study there were at least 5 to 36 hosts per tree available for parasitism, host feeding, consumption and/or intimidation.If the parasitoids were responding in a direct host density dependent fashion when released in the caged or uncaged trees, it would be within the theoretical capacity of 50 females of each parasitoid species released to parasitize all hosts.However, the percent parasitism in only the A. simplicipes trials came close to reaching the high levels expected, at least in the caged release trials.

245
1 Means within week, row and caged and uncaged classification followed by the same lowercase letter are not significantly (P > 0.05) different as determined by Tukey's Studentized range test (SAS, 2000). 2 Not applicable, trial not conducted.

Uncaged treatments Caged treatment
Mean % of collected OBLR larvae per release treatment Time collected and host status post-release TABLE 3. Mean percent of obliquebanded leafroller (± SE) collected that were parasitized, dead or parasitized by other species one and two weeks post-release of 0 (control), 5 (low release) or 50 (high release) M. linearis or A. polychrosidis females onto caged and uncaged infested potted apple trees.(N = 9 trees, 2 replications).
Honey-water was provided and most females would be expected to survive for at least the first week of the study, assuming that they could find this food, as females of all four parasitoid species survived for means of at least 20 days in laboratory trials (Cossentine et al., 2004b(Cossentine et al., , 2005(Cossentine et al., , 2007)).Conditions would be less amiable in the orchard environment and it is acknowledged that laboratoryreared parasitoids may have modified intrinsic foraging behaviour and may not respond as effectively to hosts on leaves as they do to hosts on artificial diet (Vet, 2001).Any of the four parasitoid species may not have been responding to kairomones resulting from obliquebanded leafroller larvae feeding on apple foliage and this aspect of the trials needs to be better understood.The fecundity of the two braconids as measured in laboratory trials (Cossentine et al., 2005) suggests that parasitism should have been higher in the field trials.It is possible that the time of year the trials were carried out may have been inappropriate for these two species, resulting in poor parasitism and the relatively warm temperatures under which the M. linearis were released would not be recommended for future trials.Similarly, the age and experience status of the released females require greater attention (Hougardy & Mills, 2006).Although all four parasitoid species were old enough to begin parasitizing hosts based on laboratory data under summer conditions, the cooler or warmer conditions encountered in the field may have affected their fecundity (Foerster & Butnariu, 2004;Castillo et al., 2006).
Although the total numbers of leafrollers per tree in this study were not counted prior to the parasitoid releases due to the instability of the early obliquebanded leafroller larvae after they are disturbed from feeding sites, the significantly lower numbers found in the A. simplicipes and G. variegata trials post-release implies that leafroller larvae are leaving the trees in response to parasitoid probing or are being consumed.Females of the two ichneumonid species cause leafroller mortality by hostfeeding and in a previous study, A. simplicipes have also been observed to chew early instar larvae, usually leaving the damaged remains (Cossentine et al., 2004b).Host-fed larvae in the A. simplicipes and G. variegata field trials may have died, however, most cadavers would be expected to remain associated with the leafroller webbing and consequently would be found under the dissecting microscope.In the laboratory, early instar leafrollers have frequently been observed exiting the silk webbing where they were feeding in response to ovipositor probing by both A. simplicipes and G. variegata.Neither of the two braconid species used in these release trials have been observed to host feed, consume or intimidate hosts and there was no significant impact of the release of either of these parasitoids on the total number of hosts found on the trees.
Host escape reactions to probing parasitoids have been recorded in other lepidopteran larvae in the field.Lightbrown apple moth, Eiphyas postvittana (Walker) (Lepidoptera: Tortricidae) larvae, exit webbing on leaves and spin down or crawl over the leaf surface when probed by the parasitoid Dolichogenidia tasmanica (Hymenoptera: Braconidae) (Suckling et al., 2001) and diamond-back moth, Plutella xylostella (Linnaeus) (Lepidoptera: Plutel-lidae) larvae, have been observed to hang by silk from plant leaves after attack by Diadegma spp.(Hymenoptera: Ichneumonidae) (Waage, 1983).Host escape behaviour in response to parasitoid probing in the orchard may cause a leafroller larva to move away from its feeding site, making it more vulnerable to movement off the tree and mortality from predation, starvation or desiccation.It is unfortunate that the number of leafrollers were not counted prior to the release of the parasitoids to confirm this hypothesis.Whether either of the ichneumonid parasitoid species used in this study do indeed have an intimidating impact on early host instars movement from feeding sites should be investigated as it may influence the overall impact of the indigenous species on host population densities.
Although the cages provided confinement and protection from wild predation and parasitism and thereby increased the potential for each released parasitoid species to maximize its impact on the sentinel hosts, the uncaged releases provided a more realistic assessment of each parasitoid potential efficacy.It is apparent from this study that it is important to consider the relative impact of 247 Fig. 2. Mean percent of total obliquebanded leafroller (OBLR) larvae found to be parasitized, dead or parasitized by other species after collection from second instar OBLR-infested potted apple trees placed inside or outside screen cages, one week after release of 0, 5 or 50 female Macrocentrus linearis or Apanteles polychrosidis.There was no temporal overlap of the two parasitoid trials.Nine trees per release rate; replicated twice over time.
release rates when considering a caged versus uncaged parasitoid release.The higher mortality of parasitized larvae at the higher female A. simplicipes and G. variegata release rate within cages may have been due to extreme superparasitism, predation, or to a limited extent, host feeding.The increased survival of ichneumonidparasitized host larvae under the lower pressure of five versus 50 released parasitoids was not seen in the uncaged releases supporting this hypothesis.By releasing the parasitoids in an unconfined area, these unrealistic results were avoided.
The uncaged releases also allowed wild parasitism which may support the accuracy of the timing of the release trials in that other host parasitoids were naturally present and actively searching for early instar hosts at the time of the release.The relatively low rate of natural parasitism that occurred in the M. linearis trials may indicate that the timing was sub-optimal for natural parasitism.The incidence of wild Apanteles parasitizing sentinel hosts in the A. simplicipes trials also indicates that the similarly timed A. polychrosidis releases were probably affected by unacknowledged wild Apanteles parasitism.
Results of controlled field releases need to be correctly interpreted in order to assess the potential for released parasitoids to parasitize or otherwise influence a target orchard pest.The results from this study indicate that releases of laboratory reared A. simplicipes may have a higher impact on obliquebanded hosts population densities than the other parasitoids tested; however, the differences in the levels of parasitism between the species indicate that many species specific variables such as responses to host-density and the importance, length and age-effect of learning host-related kairomones cues prior to release need to be better understood before this type of orchard trial can be used to compare the potential of individual parasitoid species to suppress host populations in orchard ecosystems.

Fig. 1 .
Fig.1.Mean percent of total obliquebanded leafroller (OBLR) larvae found to be alive and parasitized, dead and parasitized, dead or parasitized by other species after collection from second instar OBLR-infested potted apple trees placed inside or outside screen cages, one week after release of 0, 5 or 50 female Apophua simplicipes or Glypta variegata.There was no temporal overlap of the two parasitoid trials.Nine trees per release rate; replicated twice over time.